WO2017171606A1 - Method for the detection of molecules being in close proximity - Google Patents

Abstract

The invention relates to a method for determining if two species are located in close proximity to each other. It comprises immobilization of an object 111 to a solid support 110, contacting the solid support with a labelled first species dissolved in liquid 120, measuring the rate of formation of complex of object and species 151, supplementing the liquid with a quenching second species 121, measuring the rate of formation of complex of object and first species in the presence of second species 152, followed by determining change in rate of complex formation before and after addition of second species. A reduction of the rate of complex formation due to the addition of second species indicates that first and second species are located in close proximity to each other when binding to the object. An exemplary application area comprises biological objects like cells and macromolecular species.

Description

METHOD FOR THE DETECTION OF MOLECULES BEING IN CLOSE

PROXIMITY

FIELD OF INVENTION

The present invention relates to the field of analysis of species and objects. More in particular, it relates to a method to detect if two molecular

recognition elements are located in close proximity to each other on or in a biological object. Even more in particular, it relates to determination of proximity of two or more molecules on recombinant proteins, purified proteins, organelles, micelles, lipid vesicles, viruses, bacteria, cells, tissue sections and similar biological material.

BACKGROUND OF THE INVENTION

In basic biology science as well as in the development of pharmaceuticals and diagnostic agents, it is sometimes meaningful to determine if two structures are located in close proximity to each other. Proximity assays can for example be used to determine if two cell surface receptors form a dimer, or it can be used to determine if two different antibodies bind to the same antigen in a non-competitive manner, to mention two non-limiting examples.

One known method for determining proximity is the cross-linking method. In brief, the biological object under study is exposed to a reactive chemical which attaches closely related molecules to each other. After exposure, the object is homogenized and analyzed using western blot. This method is cumbersome and semi- quantitative, and is described in the report "Gefitinib Induces Epidermal Growth Factor Receptor Dimers Which Alters the

Interaction Characteristics with 125I-EGF." By Bjorkelund and co-authors as published PLoS ONE 6(9): e24739. doi: 10.1371 /journal.pone.0024739 (which is incorporated by reference herein).

Another known method for determining proximity is Forster Resonance Energy Transfer (FRET) assays. In such an assay, one molecule is labeled with an acceptor dye, and another with a donor dye. Upon illumination with a monochromatic light, the donor dye transfers the energy from the illumination to the acceptor dye, followed by the acceptor dye emitting light in a completely different color than the illumination. A requirement for FRET to work is that the donor dye and acceptor dye is located in close proximity, i.e. a few nanometers from each other. By detecting only the light from the acceptor dye, FRET will reveal if acceptor and donor dye are located in very close proximity. One of many disclosures discussing FRET is

US2015/ 0045253 by D Jaga.

Another known method for determining proximity is in-situ Proximity Ligation Assay (PLA) as described in the publication "Direct observation of individual endogenous protein complexes in situ by proximity ligation" by Soderberg and co-authors as published in Nature Methods 3, 995 - 1000 (2006) which is incorporated by reference herein. In this assay, PLA-probes recognize structures in a biological sample. PLA-probes are antibodies that have a DNA sequence conjugated to them. When two PLA probes are in close proximity their DNA sequences can be joined by an enzymatic reaction. The joined DNA sequences are amplified in a second enzymatic reaction and the amplified DNA can then be made visible by fluorescently labeled DNA molecules designed to bind to the amplified sequences.

Another known method for determining proximity is the use of quenchers. A quencher is a molecular label that absorbs light of a particular wavelength and hence has a capacity to reduce performance of other dyes located in close proximity. The typical use of a quencher is to label one molecule with it, and then label another molecule with a fluorophore which is known to be affected by the quencher. This results in a reverse assay result: If

fluorescence can be detected, the quencher is not in close proximity of the fluorescently labeled molecule. If the fluorescence intensity is reduced, both labeled molecules are located in close proximity. A typical quencher assay has been disclosed by Buijs and co-authors in "Human Growth Hormone Adsorption Kinetics and Conformation on Self-Assembled Monolayers" as published in Langmuir (1998; 14, 2, 335-341) which is incorporated by reference herein. In brief, the assay by Buijs and co-authors was conducted in the following manner:

1. A first entity having a fluorescent moiety attached was put in contact at known concentration with an object to which the first entity binds.

2. A second entity having quenching properties, assumingly being in close proximity to the first entity, was then added at known

concentration to the object.

3. Fluorescence intensity was measured

4. Steps 1-3 were repeated multiple times using the same known

concentration of first entity but different concentrations of second entity.

5. The fluorescence intensity results were plotted in a Stern- Vollmer plot so as to calculate the Stern- Vollmer quenching constants.

A typical quencher assay is relying on molecular interactions to reach equilibrium.

Quenching assays have been described in the context of time resolved Forster Resonance Energy Transfer (FRET) assays, as exemplified in the disclosure "Time -Re solved Fluorescence Quenching in Micellar Assemblies" by Gehlen and Schryver as published in Chem Rev (1993, 93, 199-221), which is incorporated by reference herein. The time resolved aspect utilized in the disclosure is the decay of fluorescence signal after a short burst of illumination, a process with a typical lifespan of a few nanoseconds. Hence the context of time resolution is applied on the physical properties of the fluorescent dyes only, and not for the purpose of determining any functional aspect of the material under study in a time-frame of regular biological events (i.e. seconds, minutes, hours or days).

Common for all known methods for determining proximity are that they are cumbersome to conduct. Not related to proximity assays, but relevant for the present invention, is a method for quality control of molecules as disclosed by M Malmqvist in WO2008/ 088288. This method relies on detection of both magnitude and slope in binding measurements shortly after putting molecules of interest into contact with a target for the purpose of determining if binding occurs. Since the molecules of interest are typically labeled with fluorescence or radioactive markers in this assay, it is common that disturbances occur at the time of putting molecules and target into contact, and for that reason the use of both slope and magnitude of binding allows to determine if molecule and target interacts earlier in time.

SUMMARY OF THE INVENTION

One object of the present invention is to facilitate the detection of two species being in close proximity. The method comprises providing a solid support having an object, typically a biological object, immobilized thereon and a solution wherein the first of the species is dissolved. First species is labeled to enable detection of both the presence of interaction between first species and object and of the rate of formation of complexes of first species and object. When the rate of formation of complexes has been determined, a second species is added to the liquid holding the first species, said second species being labeled with a quencher. The detection of both of the presence of interaction between first species and object, and of the rate of formation of complexes of first species and object is then continued, and a change in rate of formation is indicative of first species and second species interacting with object in a manner that brings first and second species into close proximity.

In a preferred embodiment, the invention provides a method for determining if a first species and a second species are located in close proximity. First species is labelled with a detectable label, and second species is conjugated to a quencher capable of quenching the label attached to the first species. The initial part of the method comprises the steps of (a) providing an object under study that is presenting target structures being capable of interacting with the first species to form species- object complexes, (b) immobilizing the object on a solid support, (c) providing a solution of the first species of interest and (d) bringing the solution in contact with the object immobilized on the support. Thereafter, (e) presence of interaction between first species and object and the binding profile or the rate of formation of first species- object complexes are detected so as to determine an initial binding profile or apparent rate of complex formation. Next, (f) the solution is supplemented with second species followed by (g) detecting the second binding profile or the apparent rate of formation of first species-object complexes in presence of second species so as to determine a second binding profile or a second apparent rate of complex formation. By (h) comparing the initial binding profile to the second binding profile it can be determined if said second binding profile is inducing apparent loss of signal compared to the initial binding profile, and if so is the case, this is indicative of first species being in close proximity of second species, while being bound to object. In a similar manner, when comparing the apparent rate of complex formation to the second apparent rate of complex formation, it can be determined if said initial rate is greater than said second apparent rate and if so is the case, this is indicative of first species being in close proximity of second species, while being bound to object.

In a further embodiment, the method relies on detection to be performed without bringing the detector in contact with said solid support.

In still another embodiment, the method includes use of a fluorescence detector, first species being labelled with a fluorescent dye, and second species being labelled with a quencher that quenches the fluorescent dye of the first species. In yet another embodiment, the method includes use of a method for detecting interactions between species in solution and object on a solid support which comprises:

immobilizing the object on a selected portion of the solid support; reducing the amount of solution covering the selected portion of the support prior to performing said measurement;

performing a reference measurement on a portion of said support where no interaction takes place. In still another embodiment, the method includes a data processing step wherein a difference between detection and reference measurements is calculated.

In yet another embodiment, the method includes use of a solid support which is an essentially flat dish capable of holding a solution confined within its boundaries.

In still another embodiment, the method includes use of objects that are of biological or chemical origin.

In yet another embodiment, the method includes use of species that are macromolecules (e.g. proteins, DNA, RNA) or other chemical compounds that can be dissolved in a liquid. In still another embodiment, the method includes use of species that are inherently fluorescent, inherently radioactive or inherently quenching.

In yet another embodiment, the step of detecting the presence of interaction between the first species and the object in the method is conducted in a time-frame of 1 minute to 10 hours or even 24 hours. In still another embodiment, the method includes a step for estimating the quantity of first and second species being in close proximity through fitting a molecular interaction model to obtained data. The molecular interaction model includes a term representing quenching effect. The fitting process is preferably implemented in a computer.

In an embodiment, the invention provides a method for determining if a first species is bound to a second species. An object under study that produces said first species is provided, the object is immobilized on a solid support, and a baseline level and baseline slope of signal from first species in said object is detected. Next, a solution of second species is provided and brought into contact with the object immobilized on the support, after which detection of a post-addition level and a slope of signal of first species in presence of second species is conducted. The initial level and slope of signal from first species are compared to post-addition level and slope of signal so as to determine if the initial level and slope is different than the post- addition level and slope. This is conducted using a first species which is a fusion molecule comprising a target protein of interest and a fluorescent entity, and a second species which is conjugated to a quencher capable of quenching the fluorescent entity fused to said first species. If a smaller second level and/ or a change in slope compared to the initial level and slope is detected, it is indicative of first species being in close proximity of second species, which is in turn indicative of first species binding second species.

BRIEF DESCRIPTION OF THE DRAWINGS

The present invention will be disclosed in closer detail in the description and example below, with reference to the accompanying drawing, in which

Figure 1 shows a schematic of the present invention; and Figure 2 shows a device known in prior art suitable for use with the present invention; and

Figure 3 shows an example of how molecular interactions may progress over time; and

Figures 4, 5, 6, 7, 8, 9, 10, 1 1 , 12, 13, 14, and 15 all shows results obtained when determining if two antibodies are binding to structures located in close proximity to each other on using a variety of cell lines, antibodies, antibody concentrations, incubation times, labels and quenchers; and

Figure 16 shows results obtained when determining if heterodimer occurs in a cell; and Figure 17 shows results obtained when determining if one antibody is binding to a structure which forms clusters on B-cells; and

Figure 18 shows one measured curve and one fitted curve from an

evaluation using a non-linear fitting algorithm to fit a binding model to captured data.

DETAILED DESCRIPTION OF THE INVENTION For the purpose of the present application, and for clarity, an "object" refers to a biological or chemical object immobilized onto a solid support. Possible objects include, but are not limited to, tissue samples, embedded tissue samples and sections thereof, cells, bacteria, viruses, solid particles, magnetic particles, macromolecules (e.g. proteins, DNA, RNA) and other chemical compounds, chemically modified surfaces, surface coatings (e.g. paint) or any combination thereof. The term "target" refers to a structure to which molecules can bind. Targets include, but are not limited to, macromolecules (e.g. proteins, DNA, RNA) and other chemical compounds. An object typically contains one or more targets. The smallest possible object is in itself a target.

A species present in a liquid may be denoted "ligand". The present invention uses two types of species: A first species and a second species. Possible first species for use in the present method include macromolecules (e.g. proteins, DNA, RNA), other chemical compounds and any species that can be dissolved in a liquid or even cells, viruses, bacteria, organelles or organisms that can be suspended in a liquid. The first species are either inherently fluorescent or radioactive, or have some sort of label attached. Suitable labels include, but are not limited to, radioactive labels and fluorescent labels. Second species are macromolecules (e.g. proteins, DNA, RNA) or other chemical compounds. Second species are either inherently quenching, or have some sort of quencher attached. The label of the first species and the quencher of the second species must match each other.

A fluorescent tag or fluorescent entity is a fluorescent molecule which can be attached, conjugated to or fused with a ligand or any other molecule. Suitable fluorescent tags include FITC, Texas Red, Rhodamine, Cy3, Cy4, Cy5, and other Cy dyes, Alexa Fluor 488, Alexa Fluor 633 and other Alexa Fluor dyes, Dylight dyes, to mention a few non-limiting examples. Green fluorescent protein, Red fluorescent protein and similar fluorescent protein constructs are suitable fluorescent entities in fusion protein constructs.

A quencher is a molecule or an atom capable of disturbing the function of another label. For example, the commonly used label fluorescein

isothiocyanate (FITC) can be quenched or disturbed in function using the quencher ATTO Q540 as supplied by ATTO-TEC GmbH (Siegen, Germany). ATTO Q540 absorbs light in the wave length span 500-580 nm, and FITC is emitting light of approximately 535 nm. Hence, when a FITC molecule and an ATTO Q540 molecule are brought into close proximity, the fluorescent ability of FITC is canceled due to the presence of ATTO Q540. Quenchers are common and known in the field of fluorescence detection, but the concept of quenching as used in the present invention is broader than fluorescence. Boron cages are for example known to absorb neutron radiation and have been evaluated for use in medical applications as evident in the publication "Development and in vitro studies of epidermal growth factor- dextran conjugates for boron neutron capture therapy." By Gedda and co-authors as published in Bioconjug Chem. 1996 Sep-Oct;7(5):584-91 , which is incorporated by reference herein. This means that a boron cage in essence is a quencher for neutron radiation. Another method to quench a label is to physically bind to it, as described by Gostring and co-authors in

"Quantification of internalization of EGFR-binding Affibody molecules:

Methodological aspects" as published in INTERNATIONAL JOURNAL OF ONCOLOGY 36: 757-763, 2010, which is incorporated by reference herein.

The term "close proximity" is defined as a distance which is smaller than 100 nm. Different quenchers have different working ranges, and each quencher used will induce a practical upper limitation of the concept close proximity. Many quenchers for use in fluorescence have a working range of up to 30 nm, or about 10 nm.

The term "binding profile" is defined as the shape of the measured binding signal over time as experimentally assessed, possibly through repeated measurements at identical conditions. Upon having a binding profile established, it is possible to conduct a similar measurement wherein a perturbation with an external stimuli (a perturbing molecule, a change in environment, or any other similar stimuli) is included. The binding profile of the perturbed measurement can be compared to the established one, and deviations may be used as characterization of the perturbation. A binding profile is typically curvilinear. The term "apparent rate of complex formation" is defined as the measured rate of complex formation and is a special case of "binding profile". In some cases, the apparent rate of complex formation is equal to the true rate of complex formation. One non-limiting example when the apparent rate of complex formation is equal to the true rate of complex formation could be when a measurement of fluorescently labeled antibody binding to targets expressed on an object is conducted. In other cases, the apparent rate of complex formation is deviating from the true rate of complex formation. One non-limiting possible reason for deviation is that a radiolabeled antibody is emitting more radioactivity than the radiation detector is capable of quantifying. Another non-limiting possible reason for deviation is that a disturbing element, such as a fluorescence quencher, is present in close proximity to a fluorescently labeled antibody, so that the measured apparent rate of complex formation constitutes the true rate of fluorescently labeled antibody binding minus the effect of quenching, in total resulting in a lower apparent rate of complex formation as compared to the true rate of formation. In some instances, apparent binding rate is used interchangeably with apparent rate of complex formation. The present invention also includes kits of targets or ligands or solid supports used for quality control by the method.

Generally, the method of the invention in its first aspect is based on the provision of 6 characteristics.

• Providing an object under study,

• Providing a first species in solution, labeled with a detectable label,

• Providing a second species in solution, labeled with a quencher,

• Making a time resolved measurement of the interaction between the first species and the object,

• Bringing object and first species together, and detecting the rate of formation of (first species) - (object) complexes, • Adding second species and detecting if the apparent rate of formation of (first species) - (object) complexes decreases.

The object under study is often of biological or chemical nature. The present invention will be exemplified in the context of cell-based assays, but it is understood that the method as such can be applied on widely different types of objects. Objects that are of biological nature includes (but are not limited to) cells, tissue sections, organelles, bacteria, virus, macromolecules and the similar. Objects that are of chemical nature include (but are not limited to) paint, other surface coatings, solid or semisolid polymers (plastic, silicone rubber, and the similar), nanoparticles (iron particles, graphene particles, and the similar).

The first species may be an antibody known to bind specifically to a target structure on said object. The antibody is typically labeled with a fluorescent dye such as FITC, but may also be inherently fluorescent. Other common types of first species include, but are not limited to, proteins, peptides, RNA, DNA and smaller chemical compounds. The second species may also be an antibody known to bind specifically to a different target structure than the first species on said object. The second species is labeled with a suitable quencher capable of quenching the function of the label used for the first species. It is also possible that the second species is inherently quenching. Second species may be the same molecule as first species, albeit labeled with a quencher instead of with a fluorophore.

A device suitable for time-resolved detection of how first species interact with the intended target on the object is set up so as to monitor the rate of formation of (first species) - (object) complexes. A suitable apparatus for performing the time resolved measurement is the device described in

WO2005080967, which is incorporated by reference herein. A non-limiting example is described. One possible setup for applying the method for proximity detection is outlined in Figure 1. An object under study 1 1 1 is attached to a solid support 1 10. First species 120 dissolved in liquid is brought into contact with the object as attached on the solid support, and are placed in a device 150 suitable for measuring the rate of formation of complexes between first species and object. A measurement is started, and the typical output 151 will be an increasing number of complexes being detected. When the apparent rate of complex formation has been determined, the second species 121 dissolved in liquid is added to the solid support where first species and object are already in contact, thereby contacting the object with both first and second species at the same time. The measurement in device 150 is then resumed and in case the second species is binding to object in close proximity to the first species, the apparent rate of complex formation will be decreased which is visible as a decrease of slope of the continuously detected number of complexes 152.

The measurement time required to produce sufficient data to determine the apparent rate of complex formation of first species binding to object depends on the system under investigation and on the concentration of the first species. When first species is an antibody, the typical time required to produce sufficient data to determine the apparent rate of complex formation ranges from 10 s (for high concentrations of a rapidly binding antibody) to 10 hours (for a low concentration of a slowly binding antibody). If is often possible to adjust system parameters so as to produce sufficient data to determine the apparent rate of complex formation within 1 minute and 10 hours or even 24 hours, or within 1 minute to 30 minutes, or within 10 s to 30 minutes. The concentration of second species required to produce change of apparent binding rate depends on depends on the system under investigation and on the target occupancy level of first species at the time of adding second species. Under the assumption that first and second species are binding their respective target with comparable affinity, the concentration of second species required to induce a detectable deviation of the apparent binding rate is, surprisingly, typically much smaller than the concentration of first species used. There are cases where the concentration of second species is less than 0.5% of the concentration of first species while still producing a detectable change in apparent binding rate. Other preferred concentrations of second species are less than 1%, 4%, 10%, 33%, 50%, or 100% of the first species concentration. In other cases, the concentration of second species may be selected to be higher than the first species concentration.

A preferred device 150 has been previously disclosed [WO2005080967, which is incorporated by reference herein] and is schematically described in figure 2. In brief, the method relies on a target 202 being immobilized to a defined area on a solid support 201 , denoted an "active area". On the same solid support, there is also a reference area (in this case opposite to the active area) . A liquid containing a dissolved ligand is in contact with the solid support to enable an interaction between target and ligand.

Furthermore, the solid support is inclined and slowly rotated using a motor 203. Over the elevated portion of the solid support, a detector capable of detecting the label attached to the species used is mounted 204. Said detector is typically not in contact with the solid support, but registers e.g. emitted radiation of radioactive nuclides or emitted light from fluorescent labels. When the active area passes the detector, an elevated signal will be registered in case the ligand has bound to the target. The rate of ligand- target formation can be followed by depicting the difference between the detected signal from active area and reference area over time.

The interaction of species with the targets on the object may be very complex, but a general understanding of the mechanisms behind the interaction can be understood by use of simple mathematical modeling. One common description of how protein interacts is the monovalent interaction model, as discussed in detail in WO2005029077 (which is incorporated by reference herein). This interaction model is described together with an example in Figure 3 , wherein protein A is interacting with protein B thereby forming the complex AB. The quantity of complexes as well as the time required to reach equilibrium is dictated by the association rate (k_on, sometimes denoted k_a) and dissociation rate (koff, sometimes denoted kd) . Since both k_on and k₀fr are assumed to be greater than zero, there will always be free A, free B and complex AB available. It is known that the rates k_on and koff can vary considerably as discussed in the report "Label-free kinetic binding data as a decisive element in drug discovery" by Karl Andersson, Robert Karlsson, Stefan Lofas, Gary Franklin and Markku D Hamalainen published in Expert Opinion in Drug Discovery l (5):440-446 (which is incorporated by reference herein) . The impact of different kon and k₀fr on interaction binding traces is schematically shown for curves 1 and 2.

Assume that protein A is in liquid phase and protein B is found in or on the object attached to the solid support. At time point Ti , liquid containing a known concentration of A is put in contact with the solid support.

Immediately complexes will be formed with velocity kon. After some time, there will be a balance between k_on and k₀ff resulting in equilibrium. Curves 1 and 2 have the same equilibrium level but different k_on and k₀fr, resulting in a short time T2 to equilibrium for curve 2 and a long time T3 to

equilibrium for curve 1. To apply this reasoning in the context of a quenching assay, assume that protein A is the first species labeled with a detectable label and protein B is labeled with a quencher, and further assuming that protein A and protein B interacts with an object in a manner that protein A and protein B are brought into close proximity, the effect of the quencher will be different depending on if a comparison is made before establishment of equilibrium of both interactions (e.g. at time point T2) or if a comparison is made at equilibrium of both interactions (e.g. at time point T3) . An end-point assay, such as all conventional quenching assays, will require extensive assay development to determine a suitable incubation time for which both the first species interaction and the second species interaction have reached equilibrium. Moreover, in cases where the object under study are cells, it is known that the interaction properties of a species in solution to a cell-surface receptor target can vary significantly, depending on which cell line is used, as described by Hanna Bjorkelund and co- authors in the report "Comparing the Epidermal Growth Factor Interaction with Four Different Cell Lines: Intriguing Effects Imply Strong Dependency of Cellular Context" as published in PLoS ONE 6(1): e 16536.

doi: 10.1371 /journal. pone.0016536, which is incorporated by reference herein. This means that in conventional quenching assays that rely on measurements being conducted at equilibrium, the required incubation time for reaching equilibrium has to be determined in each case, even if previous experience is at hand for a particular receptor as expressed on a different cell-line than the one under study. The generic properties of the preferred detection method shown in figure 2 are the following:

• a target is immobilized to a selected portion of a solid support,

• the target is exposed to a solution containing a ligand,

• a measurement is performed, capable of detecting an interaction

between said first and said second species, during which the amount of solution covering the selected portion of the support is reduced, and

• a reference measurement is performed on a portion of said support where no interaction takes place.

A conventional quenching assay typically relies on a plurality of end-point measurements conducted using different concentrations of quencher. An advantage of using a time-resolved assay for determining proximity is that the effect from quenching is evident as a discontinuity in the recorded binding trace. Such discontinuities are easily detected also when present only in small magnitude, meaning that the sensitivity of an assay where change in binding profile is used as assay output is high in comparison to conventional quenching assays. Furthermore, since a time-resolved assay is conducted using one and the same cell culture, the amount of lab-ware and manual work is typically smaller as compared to the conventional quenching assay. Yet another advantage of time-resolved quenching assays is that adjustments can be made during the course of the experiment. Should, for example, the concentration of added second species be too low, it is possible to add more second species and observe the effect. Such adjustments are typically not possible in conventional quenching assays. The time-resolved approach of the present invention further allows to detect molecules in proximity during a limited period of time, which could occur in for example molecular transportation in cells.

Molecular proximity is an important and common concept in biology. Cell surface receptors may for example form dimers, either homo-dimers

(dimerizing with another identical receptor) or heterodimers (two different receptors forming a complex). In other cases, the local receptor density may vary across a cell membrane, sometimes in lipid rafts or other forms of clusters, meaning that certain areas of a cell membrane have receptors of one type in close proximity of each other, whereas other areas of the same cell membrane have lower density of the same receptor. Another example is that certain biological functions require that two molecules are interacting with or binding to one receptor at the same time for a biological function to trigger, as exemplified by Greiner and co-authors in the report "Differential ligand- dependent protein-protein interactions between nuclear receptors and a neuronal-specific cofactor" as published in Proc Natl Acad Sci U S A. 2000 Jun 20; 97(13): 7160-7165, which is incorporated by reference herein. Molecular transport is another important field related to proximity. Still another example is that certain molecules are in motion inside a cell, such as translocating from the nucleus to the cytoplasm or vice versa, as discussed by Costa and co-authors in the report "Dynamic regulation of ERK2 nuclear translocation and mobility in living cells" as published in Journal of Cell Science 1 19, 4952-4963, 2006, which is incorporated by reference herein. Hence, in the field of biology, accurate proximity assays are important for elevating the understanding of the biology as such.

In some cases, the quantity, level or function of dimers vary depending on external factors, such as treatment of a cell with chemical compounds. One example where the quantity and function of dimers changed upon drug treatment has been disclosed by Bjorkelund an co-authors in "Gefitinib Induces Epidermal Growth Factor Receptor Dimers Which Alters the Interaction Characteristics with 125I-EGF." as published in PLoS ONE 6(9): e24739. doi: 10.1371 /journal.pone.0024739, later expanded into other tyrosine kinase inhibitors in the report by Bjorkelund and co-authors in "Resolving the EGF-EGFR interaction characteristics through a multiple- temperature, multiple-inhibitor, real-time interaction analysis approach" as published in Mol Clin Oncol. 2013 Mar; l (2):343-352. The method of the present invention can be used in such cases to provide a quantitative measure on dimer level, and hence provide a measure of how an external factor, such as a chemical compound, impacts dimer formation in a living cell. In cases when the first species is identical to the second species (with the exception of which label the species is conjugated to), binding of two species is typically similar or even identical. If binding of the species induces co- localization of the bound receptors, it is possible to model the rate of co- localization from the quenching effect. This can help deciphering mode of action of a drug, biology related to complement dependent cytotoxicity, lipid raft formation, and the similar.

In cases where bifunctional molecules are developed for the purpose of attracting two different targets, the present method can be used to confirm and quantify that the two targets are brought in close proximity by the bifunctional molecule. The targets may reside on one and the same cell, or alternatively one target each on two different cells. To illustrate how this can be done, the bifunctional molecule can be labeled with a fluorescent dye, and target binding proteins (typically antibodies) known to bind specifically to an intended target in a manner that does not obstruct dimerization.

Assuming that the bifunctional antibody binds both Target A and Target B, a first experiment where fluorescently labeled bifunctional molecule (first species) is binding followed by addition of Target A specific binding protein labeled with quencher (second species) and after that addition of Target B specific binding protein labeled with quencher (another second species). Should the signal obtained in the first incubation step (first species only) be reduced upon addition of the Target A specific second species, and then be further reduced upon addition of the Target B specific second species, then there is strong evidence that the bifunctional molecule actually binds the intended targets. To confirm that the bifunctional molecule actually brings the targets together in close proximity, a separate measurement is required, for example where a first species specific for Target A is followed by a second species specific for Target B, all during presence of unlabeled bifunctional molecule. The bifunctional aspect is of exemplary nature, because the method works equally well on tri-functional and higher order functional molecules. In a similar manner, molecules that change the level of dimers or cluster formation can be investigated using the present invention.

The first species need not necessarily be added to the measurement system, but can also be produced the object under study. One non-limiting example is to use ribosomes for protein production, most often as part of a cell but possibly in a cell free system. In such a case, the first species would comprise a target protein and a fluorescent entity or fluorescent tag (such as the green fluorescent protein (GFP) to mention one non-limiting example). Such a fusion protein construct can be produced in a cell, such as for example a bacterial cell transfected with a plasmid which encodes the combined target protein and GFP. When expressed in a cell, the cell would produce the target protein - GFP which in turn is capable of producing a baseline level and baseline slope of fluorescent signal. Upon adding a second species labeled with a suitable quencher, the baseline signal level and the baseline signal slope from the available target protein - GFP construct will be changed, most often reduced, in cases where the second species interact with the first species (target protein - GFP) . With this method it becomes possible to measure, in a time resolved manner, the intracellular interaction of target protein - second species. Should the target protein construct be confined to a specific cellular compartment, such as mitochondria, nucleus, endosomes, and cytoplasm, to mention a few non-limiting examples, the method would in addition to proximity also provide information about subcellular localization of the second species. As an example, should the target protein construct reside in the cellular membrane, the method would allow for estimation of relative motion of target construct and second species in the cellular membrane of the cell. The further define subcellular compartments, the fusion protein could be extended into multiple domains being fused together, such as combining a target protein, a subcellular compartment anchoring protein and a fluorescent entity into a first species, to mention one non-limiting example. A measurement of loss of baseline fluorescence signal does not strictly require comparison between a target area and a background area when implemented in the preferred device 150 , even though it is likely beneficial to include such a comparison.

Even though the present invention is illustrated using data from a preferred device, the method may be implemented on other suitable devices. One such suitable device is a flow cytometer. In a typical flow cytometer, the

fluorescence intensity from a first species binding to a target on cells are registered, one cell at a time, in a time-stamped manner. By binning fluorescent readings for all registered cells during a predefined time period (for example every second or every minute) a time resolved fluorescence intensity curve can be created. During such a measurement, a second species labeled with quencher can be added and should the second species bind to a target in close proximity to the first species, the apparent rate of binding would be decreased. Another suitable device is surface plasmon resonance (SPR) devices. Even though SPR is typically perceived "label free", there are labels described for dedicated use in SPR devices, as evident in the disclosure "Label-enhanced surface plasmon resonance applied to label-free interaction analysis of small molecules and fragments." by Eng and coauthors as published in Anal Biochem. 2016 Oct l ;510:79-87. doi:

10.1016/j.ab.2016.06.008. Hence, the present invention is possible to implement also in an SPR device. Data from time resolved quenching assays can be subjected to non-linear regression to predefined interaction models. One suitable model is to assume that the first species (labeled with a fluorophore) binds to its target according to a monovalent Langmuir model (also known as the one-to-one model) and that the second species (labeled with a quencher) also binds to its target according to the same model, where the two interactions are completely independent of each other. The loss of signal due to quenching would then depend on (a) the number of second species target sites being available, (b) the number of first and second species target sites being in close proximity, (c) the quenching efficiency of the quencher (d) the average number of labels being attached to each molecule (both first and second species) , the concentrations of the two different species and the number of first species targets. Many of these factors are difficult to distinguish from each other and accumulate into a "Bmax" parameter, where Bmax is the highest achievable signal (for the fluorophore) or signal loss (for the quencher). In the form of a differential equation, this model looks like the following:

F = BmaxF - Fl ;

dFl = +kaF * concF * F - kdF * Fl ;

Q = BmaxQ - Q l ;

dQ l = +kaQ * concQ * Q - kdQ * Ql ; Boundary values are defined at time = 0: F1=0 and Q1=0.

BmaxF is the maximum signal from the first species (labeled with fluorescence). BmaxQ is the maximum signal loss, or quenching effect, from the second species (labeled with a quencher) in which most correction factors like labels per molecule and quenching efficiency are accumulated. kaF, kdF, kaQ, kdQ are kinetic interaction parameters (association rate = ka, dissociation rate = kd, F for first species and Q for second species), and concF and concQ are concentrations of the respective species.

Using numerical integration, the differentials dFl and dQl are used to construct the curves Fl (t) and Ql (t), for example using the numerical integration method Euler forward. Using pseudo code, the programmed method would look similar to the following:

(start at time 0)

t=0

(populate vector with time values)

X(0)=T

(populate vector with first species binding levels over time, starts at 0) F1 (0)=0

(populate vector with second species binding levels over time, starts at 0) Q 1 (0)=0

(populate measured signal over time, starts at 0)

Y(0)=0

(set step length in the integrator, must be a sufficiently small number) dt =0.00001

(Set N so that N*dt > end time for estimated curve)

For j = 1 to N

(update concentration value if appropriate)

If X(j - 1 ) = (time point when concF or concQ changes) then Change concF, concQ to appropriate value(s)

End if

(calculate first species differential at current time)

F = BmaxF - Fl ;

dFl = +kaF * concF * F - kdF * Fl ;

(calculate first species differential at current time)

Q = BmaxQ - Q l ;

dQ l = +kaQ * concQ * Q - kdQ * Ql ;

(amend time vector)

X(j)=X(j- l) + dt

(integrate to get first species binding level)

Fl (j)=Fl (j- l) + dt * dFl

(integrate to get second species binding level)

Q l (j) = Q l (j- 1) + dt * dFl

(combine Fl and Ql to get signal)

Y(j) = Fl (j) + Fl (j) / BmaxF * (- Q l (j))

End for

The calculated output signal Yl can then be compared to the measured signal Y to form a sum-of-squared residuals ( sum([Yl-Y]^A2) ) as measure of the closeness of the calculated curve. Input parameters (kaF, kdF, BmaxF, kaQ, kdD, BmaxQ) can then be iteratively adjusted in a manner that decreases the sum-of-squared residuals until a minimum value of sum-of- squared residuals has been found (which corresponds to the best fit and hence the most appropriate values of input parameters). Such a non-linear fitting procedure is readily known in the art, as evident in the report

"Experimental design for kinetic analysis of protein-protein interactions with surface plasmon resonance biosensors." by Karlsson and Fait as published in J Immunol Methods. 1997 Jan 15;200(l-2): 121-33. In typical cases, the Euler method for numerical integration discussed above is replaced with a faster but more complicated method such as a Runge-Kutta integrator. Nonlinear fitting is typically achieved using one or more of Levenberg-Marquard algorithms, Simplex optimization, quasi-Newton methods, gradient search methods, genetic algorithms or any other optimization algorithm readily known in the art. A critical element in the quenching interaction model is the correction for first species target occupancy, visible as the factor Fl (j) / BmaxF in the equation used for calculating signal level (i.e. Y(j) = Fl (j) - Fl (j) / BmaxF * Q 1 (j) in the pseudo code above). Also, the term describing quenching effect is visible in that the contribution of the Q 1 parameter is negative because quenching reduces signal.

Many other interaction models may be suitable for detailed kinetic analysis of data produced with the method of the present invention. One non-limiting example is first species binding two separate targets (a so called one to two model) where only one of the targets is located in close proximity to the second species target and second species bind to its target in a one to one manner. Another non-limiting example is if both first and second species bind to two different targets each, and only two of the four targets are close in proximity. Yet another non-limiting example is if first species is binding to one target in a two-step manner (a first short-lived interaction phase which gradually transforms into a second, typically stronger phase, also known as a transition state model) while the second species is binding its target in a on to one manner. These and similar combinations of common interaction models (one-to-one, one-to-two, bivalent) potentially with correction for technical properties (diffusion correction or mass transfer limitation correction for flow based setups, and depletion correction for fixed volume based methods) are all plausible for evaluation of data produced with the method of the present invention. Use of interaction models to describe acquired data is beneficial, in particular when comparing different molecules. The interaction model does, when it is describing the interaction in a sufficiently accurate way, provide an objective characteristic which is comparable between experiments and which is less sensitive to small variations of target properties,

concentrations used, and times used. Data evaluation using numerical integration and/ or non-linear fitting algorithms according to the present invention may be implemented by special-purpose software (or firmware) run on one or more general-purpose or special-purpose computing devices. Such a computing device may include one or more processing units, e.g. a CPU ("Central Processing Unit"), a DSP ("Digital Signal Processor"), an ASIC ("Application-Specific Integrated Circuit"), discrete analogue and/ or digital components, or some other programmable logical device, such as an FPGA ("Field Programmable Gate Array"). In this context, it is to be understood that each "component" of the numerical integration and / or non-linear fitting algorithm refers to a conceptual equivalent of a method step; there is not always a one-to-one correspondence between components and particular pieces of hardware or software routines. One piece of hardware sometimes comprises different components. For example, the processing unit may serve as one component when executing one instruction, but serve as another component when executing another instruction. In addition, one component may be

implemented by one instruction in some cases, but by a plurality of instructions in some other cases. The computing device may further include a system memory and a system bus that couples various system

components including the system memory to the processing unit. The system bus may be any of several types of bus structures including a memory bus or memory controller, a peripheral bus, and a local bus using any of a variety of bus architectures. The system memory may include computer storage media in the form of volatile and / or non-volatile memory such as read only memory (ROM), random access memory (RAM) and flash memory. The special-purpose software may be stored in the system memory, or on other removable/ non-removable volatile/ non-volatile computer storage media which is included in or accessible to the computing device, such as magnetic media, optical media, flash memory cards, digital tape, solid state RAM, solid state ROM, etc. The computing device may include one or more communication interfaces, such as a serial interface, a parallel interface, a USB interface, a wireless interface, a network adapter, etc. One or more I/ O devices may be connected to the computing device, via a communication interface, including e.g. a keyboard, a mouse, a touch screen, a display, a printer, a disk drive, etc. The special-purpose software may be provided to the computing device on any suitable computer-readable medium, including a record medium, a read-only memory, or an electrical carrier signal.

The following non-limiting examples of the invention will illustrate the principle behind it.

EXAMPLE 1

The method described above was tested with an object comprising a human cancer cell-line, SKOV3 cells, grown on one quarter of a 10 cm circular cell- dish. SKOV3 cells are known to have high levels of HER2 receptors on their surface. The antibody pertuzumab (sold under the trademark Omnitarg, Roche, CH) was selected as first species and was labeled with FITC (Sigma Aldrich) according to the manufacturer's instructions. The antibody trastuzumab (sold under the trademark Herceptin, Roche, CH) was selected as second species and was labeled with quencher ATTO 540 (ATTO-TEC GmbH Siegen, Germany) according to the manufacturer's instructions. It is known that trastuzumab binds HER2 on the C-terminal portion of domain IV and that pertuzumab binds HER2 at a completely different location near the center of domain II, as discussed in the report "Discovery of epitopes for targeting the human epidermal growth factor receptor 2 (HER2) with antibodies" by Johan Rockberg and co-authors as published in Molecular Oncology (Volume 3, Issue 3, June 2009, Pages 238-247) which is incorporated by reference herein. The dish was placed in a LigandTracer Green device (Ridgeview Instruments AB, Vange, Sweden) which operates according to the device disclosed in WO2005080967. The dish was rotated with a speed of approximately 1 round per minute. The recorded output is shown in Figure 4. After a baseline reading, 3 ml of liquid containing 4 nM of first species pertuzumab labeled with FITC was added (at time Ti) to the cell-dish. Since pertuzumab is known to bind to the cells, the measured fluorescence intensity will be higher when the area of the cell-dish holding target cells passes by the detector compared to when other areas of the cell dish pass by. Comparing the signal from cells with the signal from a reference area where no cells are attached makes it possible to determine the amount of pertuzumab binding to the cells. After 30 minutes of monitoring binding of pertuzumab to HER2 on SKOV3 cells, the second species trastuzumab labeled with ATTO Q540 was added (at time T₂) to the dish to achieve 4 nM concentration also of the second species. The

measurement was then continued. The slope Si shortly before addition of second species is higher than the slope S₂ shortly after addition of second species, indicating that the second species trastuzumab is binding to a location in close proximity of where the first species is binding. A similar control experiment (indicated with broken line (Curve 1) was made using the same first species (pertuzumab labeled with FITC) but after 30 minutes of monitoring binding of pertuzumab to HER2 on SKOV3 cells, 4 nM unlabeled trastuzumab was added at time T₂. The slope Si shortly before addition of unlabeled trastuzumab is similar to the slope S2 shortly after addition of unlabeled trastuzumab, indicating that the binding of trastuzumab alone does not interfere with the pertuzumab interaction in any significant manner. At time T3 the liquid containing trastuzumab and pertuzumab was replaced with liquid devoid of trastuzumab and pertuzumab, so as to follow the spontaneous release of bound trastuzumab and pertuzumab.

This example shows that in a container of cells, it is possible to detect the impact of the quenching second species in a rapid and sensitive manner. There is no need for use of concentration series of neither first nor second species. The measurement is further independent on the absolute number of cells seeded in the dish, because the slope of first species binding is compared to the slope after addition of second species.

Example 2

The experiment described in example 1 was repeated, but with the two antibodies labeled in opposite ways. This means that the antibody

trastuzumab was selected as first species and was labeled with FITC, and the antibody pertuzumab was selected as second species and was labeled with quencher ATTO 540. A measurement was conducted according to example 1 and the corresponding results are shown in Figure 5. The solid line 51 shows first species binding with supplement of unlabeled second species at 1.8 hours, and the dotted line 52 first species binding with supplement of quencher labeled second species at 1.8 hours (indicated as 50). The quencher clearly changes the pattern of the measured binding curve, indicating that first and second species are binding to a structure on the cells in close proximity. After a total of about 7 h in all cell dishes the medium was replaced with fresh cell medium to measure dissociation.

This example provides confirmation that the two antibodies trastuzumab and pertuzumab are binding to locations in close proximity.

Example 3

The experiments described in example 1 and example 2 were extended to include a blocking control. Three cell dishes with SKOV3 cells were prepared and placed in three different LigandTracer Green instruments. Recorded data is shown in Figure 6. One of the cell dishes was blocked with 30 nM of unlabeled Trastuzumab for 3 hours (solid black line 61), then 4nM FITC- Pertuzumab was added to all three cell dishes. After about one hour

(indicated as 60) of association 4nM of ATTO Q540-Trastuzumab was added to the cell dish which was blocked with 30nM unlabeled Trastuzumab (solid black line 61) and to one of the unblocked cell dishes (dotted black line 63). To the remaining cell dish 4nM of unlabeled Trastuzumab was added (dashed black line 62). After a total of about 6.5 h in all cell dishes the medium was replaced with fresh cell medium to measure dissociation.

When Trastuzumab labeled with the quencher was added to non-blocked cells a decrease in the slope of the FITC-Pertuzumab binding trace was observed, whereas no decrease was observed for cells that had been blocked with unlabeled Trastuzumab.

This example shows that when the antibody labeled with the quencher does not bind to the cells, since its target epitope is blocked, there is no effect on the binding signal of the first antibody.

Example 4

The experiment described in example 3 was repeated, but with the two antibodies labeled in opposite ways. Three cell dishes with SKOV3 cells were prepared and placed in three different LigandTracer Green instruments.

Recorded data is shown in Figure 7. One of the cell dishes was blocked with 30 nM of unlabeled Pertuzumab for 4 hours (solid black line 71), then 4nM FITC-Trastuzumab was added to all three cell dishes. After about 30 min of association (indicated as 70) 4nM ATTO Q540-Pertuzumab was added to the cell dish which was blocked with 30nM unlabeled Pertuzumab (solid black line 71) and to one of the unblocked cell dishes (dotted black line 72). To the remaining cell dish 4nM of unlabeled Trastuzumab was added (dashed black line 73). After a total of about 8 h in all cell dishes the medium was replaced with fresh cell medium to measure dissociation.

When Pertuzumab labeled with the quencher was added to non-blocked cells a clear decrease in the slope of the FITC-Trastuzumab binding trace was observed, whereas noticeably less change in the slope of the binding trace was observed for cells that had been blocked with unlabeled

Pertuzumab. In this configuration, the blocking was incomplete and ATTO Q540 -Pertuzumab could to some extent reduce signal also in the blocked dish (solid line 71). This shows that when the antibody labeled with the quencher does not bind to the same extent to the cells, since its target epitope is partially blocked, there is less effect on the binding signal or binding profile of the first antibody compared to an unblocked situation.

Example 5

The experiment described in example 1 was repeated using a different cell line: SKBR3 cells. SKBR3 cells also have high HER2 expression, but origin from a breast cancer, whereas SKOV3 cells are an ovarian cancer cell line. Figure 8 shows two LigandTracer measurements with a baseline

measurement for one hour, then 4 nM FITC-labeled Pertuzumab was added to the SKBR3 cells. After about 45 min measurement (indicated as 80) to one of the cell dishes 4 nM Trastuzumab labeled with the quencher ATTO Q540 is added (dotted line 82), whereas 4nM unlabeled Trastuzumab is added to the other (solid line 81). Adding Trastuzumab labeled with the quencher causes a change in the binding profile of the Pertuzumab-binding curve when compared to binding trace of the control experiment in which unlabeled Trastuzumab was added. This example shows that it is possible to detect proximity in a different cell line.

Example 6

This experiment resembles the one described in example 1 , but here two different concentrations of second species (quencher labeled antibody) were applied on SKOV3 cells. Figure 9 shows two LigandTracer measurements, both with a baseline followed by FITC-Trastuzumab at 12 nM concentration.

After 2.5 hours of association (indicated as 90), 4 nM of ATTO Q540-

Pertuzumab was added to one (solid black line 91 ) cell dish and 12 nM of ATTO Q540-Pertuzumab to the other (dotted black line 92). After a total of about 4.5 h in all cell dishes the medium was replaced with fresh cell medium to measure dissociation. The change in slope was more pronounced when 12 nM quenching antibody was added compared to adding only 4 nM quenching antibody. For further analysis, curves were multiplied with - 1 and adjusted so that the signal level at time = 2.5 hours was 0, as shown in figure 10. Next, a one to one fit was applied for the quenched part of the curves starting at about 2.5 h and ending at about 4.5 hours. Fitting binding data to interaction models is well-known in the art, as evident in the report "Experimental design for kinetic analysis of protein-protein interactions with surface plasmon resonance biosensors." by Karlsson and Fait as published in J Immunol Methods. 1997 Jan 15;200(l-2): 121-33. The interaction parameters ka, kd and Bmax were set to "Global", because it is reasonable to expect that the quenching process behaves in a similar manner as an interaction process. The resulting fit shown in Figure 10 (solid line) is acceptable and indicates that the quenching process in this system follows a 1 : 1 kinetic model when the quenching antibody is applied at or close to equilibrium of the first antibody. This example shows that the quenching effect is concentration dependent, and that the quenching process is reasonably similar to a one to one molecular interaction.

Example 7

This experiment resembles the one described in example 2, but here different incubation times for the first species was evaluated. Three cell dishes with SKOV3 cell were prepared and placed in LigandTracer Green instruments. Results are shown in Figure 1 1. After an initial baseline measurement, 12 nM FITC-Trastuzumab was added and incubated for either 2h (black dashed line 1 1 1), 3h (black dotted line 1 12) or 4h (black solid line 1 13). After the respective incubation times 12nM ATTO Q540- Pertuzumab was added, and after approximately 3 hours of quenching the medium was replaced with fresh cell medium to measure dissociation.

The earlier the ATTO Q540 labeled antibody was added the more of the original fluorescent intensity was quenched 2 h after the addition of quenching antibody.

Example 8

This experiment resembles the experiment described in example 1 , but here low concentrations of the second species were evaluated. After about one hour of incubation (indicated as 125) of 12 nM of the first species two parallel experiments using dishes with SKOV3 cells, 50 pM of the second species were added to one of the dishes (dotted line 122) and 50pM of unlabeled trastuzumab (second species but unlabeled) were added to the other (solid line 121). After incubation for 1.5 hours the concentration of the second species was increased to 150 pM (indicated as 126) and then to 450 pM (indicated as 127) and 1.35 nM (indicated as 125) (dotted line 122). Results can be seen in Figure 12. This example shows that as low concentrations as 50 pM of the second species can give a visible deviation in the binding profile of the measured binding curve.

Example 9

This experiment resembles the experiment described in Example 2, but the antibodies were labeled with a different fluorophore, namely ATTO 633 (ATTO-TEC GmbH Siegen, Germany) and the dark quencher QC1 (LiCor, Lincoln, US) that quenches light in the range 500-800nm. Figure 13 shows binding of 4 nM ATT0633-labeled Trastuzumab to SKOV3 cells, after which (indicated as 130) 4nM Pertuzumab labeled with QC1 was added to one dish (dotted line 132) and 4 nM unlabeled Pertuzumab to the other (solid line 131).

Adding Pertuzumab labeled with the quencher caused a decrease in the slope of the Trastuzumab binding curve when compared to binding trace of the control experiment in which unlabeled Pertuzumab was added.

This example shows that dye and quencher are not limited to the ones used in the first set of examples for detecting proximity through altered binding profile.

Example 10

This experiment resembles the experiments described in Example 1 and Example 8, but the antibodies were labeled with a different fluorophore, namely ATTO 633 and the dark quencher QC1.

Figure 14 shows binding of 4 nM ATT0633-labeled Pertuzumab to SKOV3 cells, then (indicated as 140) 4nM Trastuzumab labeled with QC1 was added to one dish (dotted line 142) and 4 nM unlabeled Trastuzumab to the other (solid line 141). After a total of about 3.5 h in all cell dishes the medium was replaced with fresh cell medium to measure dissociation.

Adding Trastuzumab labeled with the quencher caused a decrease in the slope of the Pertuzumab binding curve when compared to binding trace of the control experiment in which unlabeled Trastuzumab was added.

This example shows that dye and quencher are not limited to the ones used in the first set of examples for detecting proximity. Example 1 1

An experiment was conducted using a completely different receptor system. A431 cells express large amounts of EGF receptors. Two different antibodies, both known to bind to EGFR, were used. The antibody abl 1400 (Abeam, Cambridge, MA) labeled with FITC and the therapeutic antibody cetuximab (trade name Erbitux®) labeled with quencher ATTO-540. Three cell dishes with A431 cells were prepared and placed in LigandTracer Green instruments. Resulting binding curves are shown in Figure 15. In all three measurements, baseline followed by incubation with 1.4 nM abl l400-FITC for 7 hours was conducted. Next (indicated as 150), one dish supplemented with 6nM unlabeled cetuximab (dotted black line 152), one dish was supplemented with 6 nM cetuximab-ATTO540 (dashed black line 153), and one dish was left without change (solid black line 151). The addition of quencher labeled cetuximab induced an immediate drop in signal, indicating that the two antibodies bind to structures located in close proximity.

This example shows that the method is applicable on a different biological model system than the one used in previous examples.

Example 12

An experiment was conducted using two different receptors that are known to form heterodimers. SKOV3 cells express moderate amounts of EGF receptor and high amounts of HER2 receptor. The EGF receptor is known to have a preference for dimerizing with HER2 compared to dimerization with itself. Two different antibodies were used: Cetuximab which is known to bind to EGFR and Trastuzumab which is known to bind to HER2.

Cetuximab was labeled with the fluorophore FITC and 2 nM was added to two separate dishes of SKOV3 cells at 1 h as shown in Figure 16. After 20 min of incubation (indicated as 160), 12 nM of Trastuzumab labeled with Q540 was added to one dish (black dotted line 162), whereas 12 nM of unlabeled Trastuzumab was added to the second dish (solid black line 161). The signal dropped upon addition of quencher labeled trastuzumab, indicating that HER2 and EGFR are located in close proximity. This experiment was repeated, but with two cell dishes being exposed to the tyrosin kinase inhibitor gefitinib at one micromolar concentration during 24 hours before start of measurement and throughout the full measurement and four dishes were left untreated. It is known that gefitinib increases the number of dimers as disclosed by Bjorkelund an co-authors in "Gefitinib Induces Epidermal Growth Factor Receptor Dimers Which Alters the

Interaction Characteristics with 125I-EGF." as published in PLoS ONE 6(9): e24739. doi: 10.1371 /journal. pone.0024739 and as disclosed by Anido and co-authors in "ZD 1839, a specific epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor, induces the formation of inactive EGFR/HER2 and EGFR/HER3 heterodimers and prevents heregulin signaling in HER2- overexpressing breast cancer cells." as published in Clin Cancer Res. 2003 Apr;9(4): 1274-83. All dishes were first subjected to 20 min of cetuximab- FITC, as above. After that, three of the dishes (one exposed to gefitinib and two untreated) were subjected to addition of 12 nM of unlabeled

Trastuzumab (three curves 166) and the remaining three were subjected to addition of 12 nM of Q540 labeled Trastuzumab (three curves curves 167). Gefitinib treated curve (dashed) 165 without treatment show a higher level of cetuximab signal, and also show a large difference to Q540 labeled curve 168. For untreated cells the curves connected to unlabeled trastuzumab (solid and dotted lines 166) give higher signal than the corresponding quencher related curves (solid and dotted lines 167), albeit smaller difference than for gefitinib treated cells. The resulting time-resolved data was subjected to non-linear regression to a monovalent, one-to-one, interaction model using the "solver" function in the program Microsoft Excel. The kinetic binding parameters ka and kd of Cetuximab were set as constants, derived from regular kinetic experiments with only cetuximab present during the measurement. The kinetic binding parameters ka and kd of Trastuzumab were similar for Gefitinib treated SKOV3 cells (ka = 7.2E4 1 / (M*s), standard error of the mean was

approximately 11%; kd = 2.5E-6 1 / (s), standard error of the mean was approximately 23%) and untreated SKOV3 cells (ka = 5.6E4 l / (M*s), standard error of the mean was approximately 12%; kd = 2.5E-6 1 / (s), standard error of the mean was approximately 23%). In the process of fitting kinetic parameters one also obtains a Bmax value that represents the theoretical maximum binding signal at an infinite ligand concentration. The Bmax value is dependent on how many labels the first and second species carry. When the same labeling batches are used in a set of experiments the ratio between the Bmax of the first species and the Bmax of the second species provides an estimation on how many molecules of first and second species are in proximity. When using the same labeling batches of FITC- Cetuximab and Q450-Trastuzumab it was possible to obtain a ratio of Bmax(second species) / Bmax(first species) of 0.20, 0.22, 0.25 and 0.26 in four independent experiments with untreated SKOV3 cells. When

performing the same experiment with SKOV3 cells pretreated with Gefitinib for 24 to 48 hours the estimated a Bmax ratio between first and second species of 0.27, 0.29, 0.31 and 0.37 in four independent experiments. The higher Bmax ratios for the treated cells reflect that gefitinib treated cells display more EGFR-Her2 heterodimers than untreated cells. This example shows that it is possible to detect if two different molecules (EGFR and HER2 in this case) are located in close proximity, putatively forming dimers. Results from a complementing experiment with a compound known to increase dimer levels, supports the hypothesis that dimers is cause of the observed proximity. The example also shows that it is possible to measure the effect of an external factor, such as the compound gefitinib, on the quantity of target molecules located in close proximity.

Example 13

An experiment with a completely different cell type and receptor system was performed. A type of B-cells, namely Daudi cells, that express high levels of a receptor called CD20 were used. The CD20 receptor is known to form clusters in lipid rafts on the cell surface upon binding of the therapeutic antibody Rituximab as disclosed by Janas and co-authors in "Rituxan (anti- CD20 antibody)-induced translocation of CD20 into lipid rafts is crucial for calcium influx and apoptosis." as published in Clin. Exp. Immunol. 139, 439-446 (2005). One aliquot of Rituximab was labeled with the fluorophore FITC and another was labeled with the Atto-Q540 quencher. Thus in this example, the same molecule acts as both first and second species. Three dishes with B-cells were prepared and to all of them 20 nM of FITC- Rituximab was added. Resulting binding curves are shown in Figure 17. At 1.5 hours (indicated as 170) 10 nM of unlabeled Rituximab was added to one of the dishes (dotted line 172), ΙΟηΜ of Q540-Rituximab was added to the second dish (dashed line 173) and an equivalent volume of PBS was added to the last dish (solid line 171). After about 30 min incubation

(indicated as 175) the concentrations of unlabeled Rituximab and Q540- Rituximab was increased to 40 nM.

The addition of unlabeled Rituximab to one of the dishes was necessary as the unlabeled molecule will compete with the labeled molecule and thus result in a decreased slope of the binding trace compared to the experiment were only labeled molecule is added. As can be seen in Fig 17 the addition of Q540 -Rituximab (dashed line 173) instead of unlabeled Rituximab (dotted line 172) results in a more pronounced change in the slope of the binding trace. This shows that despite competition there is an additional signal decrease which can be attributed to quenching.

This example shows that it is possible to use the same molecule as first and second species. This example also shows that CD20 molecules are located in close proximity on Daudi cells, which supports the previously published finding.

Example 14

In this example, data was produced using the same cells and antibodies as in Example 2. After an initial incubation during 10000 seconds (2.8 hours) of 12 riM FITC-trastuzumab, 24 nM of ATTO-Q540 - pertuzumab was added. Resulting binding curve is shown as circles in Figure 18. An interaction model where both trastuzumab and pertuzumab were assumed to bind according to independent one to one models were programmed using Visual Studio C++. The differential equation subjected to numerical integration was the following: void fevalOneOneQUE(double t, double Y[], double dY[], double kal , double kdl , double ka2, double kd2, double concl , double conc2, double BmaxO, double Bmaxl)

{

double BO;

double B 1 ;

BO = BmaxO - Y[0];

dY[0] = +kal * concl * BO - kdl * Y[0];

Bl = Bmaxl - Y[l];

dY[l] = +ka2 * conc2 * Bl - kd2 * Y[l];

}

Where kal , kdl and BmaxO relate to the first species binding

characteristics, ka2, kd2, Bmaxl relate to the second species binding characteristics, concl and conc2 represent concentration for first and second species respectively. Time is t, and binding level vector Y together with binding level differential vector dY are used for storing the integrated values throughout the numerical integration.

The estimated signal was constructed according to:

calcyValue = Y[0] - Y[0] / BmaxO * Y[l];

After numerical integration using a 5^th order Runge Kutta integrator, the model output was compared to the measured values, and the difference was minimized (difference as measured using the sum of squared residuals as penalty function) using the optimizer implemented in the TraceDrawer software package. After the fit had converged, the final estimated binding curve was plotted in Figure 18 (solid line). The following parameters were identified as the ones producing the best fit. ka(trastuzumab) = 5.2E4 (l / (M*s))

kd(trastuzumab) = 5.8E-6 (1 /s)

Bmax(trastuzumab) = 86 (signal units a.u.)

ka(pertuzumab) = 6.0E4 (l / (M*s))

kd(pertuzumab) = 4.3E-5 (1 /s)

Bmax(pertuzumab) = 32 (signal units a.u.)

This example shows that it is possible to fit an interaction model to acquired data so as to extract binding characteristics of both the first and the second species.

Example 15

In this example, a biological system is first subjected to characterization with respect to a defined biological proximity situation (such as CD20 clustering or EGFR-HER2 dimerization to mention two non-limiting examples) . The characterization may comprise one or multiple assays.

Next, at least one of the assays used in characterization step is repeated but with a test compound present and/ or test condition in place before or during the assay. A test compound can for example be a small chemical compound (such as gefitinib to mention a non-limiting example), a macromolecule such as protein molecule (including but not limited to an antibody), or a nucleic acid construct (for example aptamer) . Test conditions include, but is not limited to, environmental conditions (for example temperature, lighting conditions, and humidity) and culturing conditions (for example choice of cell culturing medium) . When results from the assay subjected to a test compound and/ or a test condition is available, the propensity of the test condition to alter the biological proximity situation as compared to the initial characterization conducted at regular conditions is calculated, reported and used as a quantification for the effect of the test condition on the biological proximity situation. With such a quantification available, it becomes possible to select test conditions that alters the biological proximity situation in a favourable manner.

Although the invention has been described with regard to its preferred embodiment, which constitutes the best mode currently known to the inventor, it should be understood that various changes and modifications as would be obvious to one having ordinary skill in this art may be made without departing from the scope of the invention as set forth in the claims appended hereto.

Claims

CLAIMS:

1. A method for determining if a first species and a second species bound to the same object are located in close proximity to each other, the method comprising:

immobilizing an object under study that comprises target structures being capable of interacting with the first and second species to form species-object complexes on a solid support;

providing a solution of said first species;

bringing said solution in contact with the object immobilized on the support,

determining an initial binding profile of first species-object complexes;

supplementing said solution with said second species;

determining a second binding profile of first species- object complexes in presence of second species;

comparing said initial binding profile to said second binding profile so as to determine if said second binding profile results in apparent loss of signal; wherein said first species is labelled with a detectable label; said second species is conjugated to a quencher capable of quenching the label attached to said first species; and wherein loss of apparent signal in the second binding profile compared to the initial binding profile is indicative of said first species being in close proximity of said second species.

2. Method as claimed in claim 1 , wherein said detection is performed without bringing the detector in contact with said solid support.

3. Method as claimed in claim 1 , wherein said detector is a

fluorescence detector, the first species is labelled with a fluorescent dye, and the second species is labelled with a quencher that quenches the fluorescent dye of said first species.

4. The method as claimed in claim 1 or 2 or 3, further comprising: immobilizing the object on a selected portion of the solid support; reducing the amount of solution covering the selected portion of the support prior to performing said measurement;

performing a reference measurement on a portion of said support where no interaction takes place.

5. The method as claimed in claim 4, wherein a difference between detection and reference measurements is calculated.

6. The method as claimed in any of the previous claims, wherein the solid support is an essentially flat dish capable of holding a solution confined within its boundaries.

7. The method as claimed in claim 4, 5 or 6, wherein the reduction of the amount of solution is achieved by orienting the support at an angle that deviates from the horizontal to provide an elevated part and a lower part of said support, such that the elevated part will be covered by less solution than the lower part, and wherein the support is rotated at a predetermined speed of rotation.

8. The method as claimed in any of the previous claims, wherein said object is an object of biological or chemical origin.

9. The method as claimed in any of the previous claims, wherein either of said species are inherently fluorescent, inherently radioactive or inherently quenching.

10. The method as claimed in any of the previous claims, wherein the time used for detecting the presence of interaction between the first species and the object is between 1 minute and 10 hours.

1 1. The method as claimed in any of the previous claims, wherein the quantity of first and second species being in close proximity is estimated through fitting a molecular interaction model [which includes a term for quenching effect] to obtained data.

12. Method according to claim 1 1 , wherein said step fitting a molecular interaction model to obtained data is computer-implemented.

13. A method for determining if a first species is bound to a second species, the method comprising:

providing an object under study that produces said first species; immobilizing said object on a solid support;

detecting a baseline level and baseline slope of signal from first species in said object;

providing a solution of said second species;

bringing said solution in contact with the object immobilized on the support,

detecting a post-addition level and a slope of signal of first species in presence of second species;

comparing said initial level and slope of signal from first species to said post-addition level and slope of signal from first species so as to determine if said initial level and slope is different than said post-addition level and slope; wherein said first species is a fusion molecule comprising a target protein of interest and a fluorescent entity; and said second species is conjugated to a quencher capable of quenching the fluorescent entity fused to said first species; and wherein a smaller second level and/ or a change in slope compared to the initial level and slope is indicative of said first species being in close proximity of said second species.