CANDIDATE

Cross Reference to Related Applications

[0001] The present application claims priority to U.S. Provisional Patent Application No.

61/440,597, filed February 8, 201 1 , entitled "A METHOD TO CUSTOMIZE PROTEIN BINDING DETERMINATION OF A DRUG," the disclosure of which is hereby incorporated by reference in its entirety.

Technical Field

[0002] Embodiments herein relate to the field of drug discovery, and, more specifically, to systems and methods for predicting drug efficacy using Differential Scanning Calorimetry.

Background

[0003] Pre-clinical drug discovery efforts aim to define new drug candidates that are capable of successfully navigating the multi-year, billion dollar, pharmaceutical drug

development pipeline. For a drug candidate to pass preclinical screening requirements, the chemical must be both potent and bioavailable. Potency is determined by the strength and specificity of binding. Bioavailability is defined using well known criteria: absorption, distribution, metabolism, excretion, and toxicity (ADMET). In addition to potency, ADMET properties may determine the efficacy of a drug in preclinical studies.

[0004] Approaches to pre-clinical drug discovery rely on combinatorial chemistry and genomics-based target generation approaches to predict which drug candidates should be given top priority for further development. To increase processing capacities, high throughput methods capable of screening large libraries of drug candidates have also become an implicit goal for emerging analytical techniques. Physical characterization of drug candidates also suggests those most likely for further preclinical and clinical studies. Part of the drug candidate assessment process may involve determining which plasma proteins bind to the drug candidate.

[0005] Many unsuccessful attempts have been made in the pharmaceutical industries to find alternative ways of obtaining early information about the ADMET properties of new drug candidate compounds. To date, there are very few accurate or reproducible models and systems for new drug discovery. Brief Description of the Drawings

[0006] Embodiments will be readily understood by the following detailed description in conjunction with the accompanying drawings. Embodiments are illustrated by way of example and not by way of limitation in the figures of the accompanying drawings.

[0007] Figures 1A and 1 B show the results of Differential Scanning Calorimetry (DSC) and fluorescence melting curves of Human Serum Albumin (HSA) in phosphate buffer (Figure 1A) and phosphate buffer with 5% DM50 (Figure 1 B), in accordance with various

embodiments;

[0008] Figure 2 shows example measurements of ligand binding (bromocresol green) to HSA by DSC, fluorescence melting, and Isothermal Titration Calorimetry (ITC), in

accordance with various embodiments;

[0009] Figure 3 shows example measurements of ligand binding (naproxen green) to

HSA by DSC, fluorescence melting, and ITC, in accordance with various embodiments;

[0010] Figure 4 shows example measurements of ligand binding (salicylate) to HSA by

DSC, fFluorescence melting, and ITC, in accordance with various embodiments;

[0011] Figure 5 shows example measurements of ligand binding (bromophenol blue) to

HSA by DSC, fluorescence melting, and ITC, in accordance with various embodiments;

[0012] Figures 6A and 6B show example measurements of ligand binding (phenol red) to HSA by DSC, fluorescence melting, and ITC, in accordance with various embodiments;

[0013] Figure 7 shows example measurements of ligand binding (bromosulfalein) to

HSA by DSC, fluorescence melting, and ITC, in accordance with various embodiments;

[0014] Figure 8 shows example measurements of ligand binding (ibuprofen) to HSA by

DSC, fluorescence melting, and ITC, in accordance with various embodiments;

[0015] Figure 9 shows example measurements of ligand binding (imipramine) to HSA by DSC, fluorescence melting, and ITC, in accordance with various embodiments;

[0016] FigurelO shows example measurements of ligand binding (chlorpromazine) to

HSA by DSC, fluorescence melting, and ITC, in accordance with various embodiments;

[0017] Figure 11 shows example measurements of ligand binding (oxacillin) to HSA by

DSC, fluorescence melting, and ITC, in accordance with various embodiments;

[0018] Figure 12 shows example measurements of ligand binding (penicillin G) to HSA by DSC, fluorescence melting, and ITC, in accordance with various embodiments;

[0019] Figure 13 shows example measurements of ligand binding (Evan's blue) to HSA by DSC, fluorescence melting, and ITC, in accordance with various embodiments;

[0020] Figure 14 shows example measurements of ligand binding (methyl orange) to

USA by DSC, fluorescence melting, and ITC, in accordance with various embodiments;

[0021] Figure15 shows example measurements of ligand binding (oetanoate) to HSA by DSC, fluorescence melting, and ITC, in accordance with various embodiments; [0022] Figure 16 shows example measurements of ligand binding (sodium oleate) to

HSA by DSC, fluorescence melting, and ITC, in accordance with various embodiments;

[0023] Figure 17 shows a data summary table of DSC and fluorescence melting of HSA in the presence of 18 different ligands, wherein the explicit data as shown in Figures 2-16 is not shown, in accordance with various embodiments;

[0024] Figure 18 shows a data summary table of DSC and fluorescence melting with

HSA-ligand (1 :10) in potassium phosphate buffer + DM50 in the presence of an additional 15 different ligands, wherein the explicit experimental data as shown in Figures 2-16 is not shown, in accordance with various embodiments;

[0025] Figure 19 shows a table having DSC of HSA melting in the presence of different binding ligands, wherein the table summarizes results of DSC melting of HSA in the presence of different binding ligands grouped according to binding site and observed ligand-dependent changes, in accordance with various embodiments;

[0026] Figure 20 shows a summary table of ITC saturation binding data, wherein the fitting parameters obtained from ITC saturation binding for the ligands examined are

summarized, and some binding curves are fit with a single site model, while most are best fit by a two site model, in accordance with various embodiments;

[0027] Figure 21 shows a table of ITC excess site data, wherein ITC excess binding is summarized, and most cases show two binding sites, in accordance with various embodiments;

[0028] Figure 22 illustrates complementarities of Surface Plasmon Resonance (SPR) and DSC, in accordance with various embodiments;

[0029] Figure 23 illustrates interactions with individual proteins, in accordance with various embodiments;

[0030] Figure 24 illustrates a strategy to screen a library for HSA binding, in

accordance with various embodiments;

[0031] Figure 25 illustrates a screen for binding to plasma proteins, in accordance with various embodiments;

[0032] Figure 26 illustrates drug interactions in situ, in accordance with various embodiments;

[0033] Figure 27 shows DSC thermograms of HSA in the presence of increasing amounts of penicillin, in accordance with various embodiments;

[0034] Figure 28 shows DSC thermograms of HSA in the presence of increasing amounts of bilirubin, in accordance with various embodiments;

[0035] Figure 29 shows calculations of Kn » Kd, wherein the simple equilibrium binding model indicates the observed behavior for penicillin is consistent with single-site binding to the native state of HSA, in accordance with various embodiments; [0036] Figure 30 shows calculations of Kn = Kd, wherein the simple equilibrium binding model indicates that the observed behavior for bilirubin is consistent with binding with equal strength to the native and denatured states of HSA, in accordance with various embodiments;

[0037] Figure 31 shows thermograms of HSA, HWP and HDP with no added ligand, in accordance with various embodiments;

[0038] Figure 32 shows thermograms of HSA in the presence of increasing concentrations of ligand, in accordance with various embodiments;

[0039] Figure 33 shows thermograms of HWP in the presence of increasing concentrations of ligand, in accordance with various embodiments;

[0040] Figure 34 shows thermograms of HDP in the presence of increasing concentrations of ligand, in accordance with various embodiments;

[0041] Figure 35 shows DSC thermograms of complement C3 at 10 μηη in the absence and presence of added ligand at 0.1 , 1 .0 and 10 μηι, in accordance with various embodiments;

[0042] Figure 36 shows DSC thermograms of complement C4 at 10 μηη in the absence and presence of added ligand at 0.1 , 1 .0 and 10 μηι, in accordance with various embodiments;

[0043] Figure 37 shows DSC thermograms of ceruloplasmin at 10 μηη in the absence and presence of added ligand at 0.1 , 1 .0 and 10 μηι, in accordance with various embodiments; and

[0044] Figure 38 shows DSC thermograms of transferrin at 10 μηη in the absence and presence of added ligand at 0.1 , 1 .0 and 10 μηη, in accordance with various embodiments.

Detailed Description of Disclosed Embodiments

[0045] In the following detailed description, reference is made to the accompanying drawings which form a part hereof, and in which are shown by way of illustration embodiments that may be practiced. It is to be understood that other embodiments may be utilized and structural or logical changes may be made without departing from the scope. Therefore, the following detailed description is not to be taken in a limiting sense, and the scope of

embodiments is defined by the appended claims and their equivalents.

[0046] Various operations may be described as multiple discrete operations in turn, in a manner that may be helpful in understanding embodiments; however, the order of description should not be construed to imply that these operations are order dependent.

[0047] The description may use perspective-based descriptions such as up/down, back/front, and top/bottom. Such descriptions are merely used to facilitate the discussion and are not intended to restrict the application of disclosed embodiments.

[0048] The terms "coupled" and "connected," along with their derivatives, may be used.

It should be understood that these terms are not intended as synonyms for each other. Rather, in particular embodiments, "connected" may be used to indicate that two or more elements are in direct physical or electrical contact with each other. "Coupled" may mean that two or more elements are in direct physical or electrical contact. However, "coupled" may also mean that two or more elements are not in direct contact with each other, but yet still cooperate or interact with each other.

[0049] For the purposes of the description, a phrase in the form "A/B" or in the form "A and/or B" means (A), (B), or (A and B). For the purposes of the description, a phrase in the form "at least one of A, B, and C" means (A), (B), (C), (A and B), (A and C), (B and C), or (A, B and C). For the purposes of the description, a phrase in the form "(A)B" means (B) or (AB) that is, A is an optional element.

[0050] The description may use the terms "embodiment" or "embodiments," which may each refer to one or more of the same or different embodiments. Furthermore, the terms "comprising," "including," "having," and the like, as used with respect to embodiments, are synonymous.

[0051] In various embodiments, methods, apparatuses, and systems for predicting one or more properties of a drug candidate are provided. In some embodiments, a computing system may be endowed with one or more components of the disclosed apparatuses and/or systems, and may be employed to perform one or more methods as disclosed herein.

[0052] Disclosed in various embodiments are methods of using differential scanning calorimetry (DSC) as a platform to screen and characterize binding of a drug candidate in a complex mixture of one or more plasma proteins. In various embodiments, the disclosed methods enable high-throughput methods for determining binding properties of an active ligand in a complex mixture of specific and non-specific protein targets. Various embodiments of the systems and methods provided herein are based upon the observation that the extent to which a drug binds non-specifically to one or more plasma or serum proteins is determinative of the drug's in vivo distribution, availability for target binding, and rate of elimination.

[0053] Generally speaking, in order for a drug candidate compound to be considered effective in a whole organism, it should be capable of being delivered effectively to a target tissue, target cell or target molecular receptor. In general, to reach a tissue, a drug candidate must travel through the bloodstream to the desired target. As such, binding interactions of a drug candidate with non-target biomolecules (e.g., proteins, lipids, nucleic acids, etc) in the bloodstream may reduce the fraction of total administered drug candidate that is available for specific interaction with an intended target. For example, in some embodiments, a drug that is capable of binding to human serum albumin (HSA) may reach a drug target after traveling in the bloodstream. However, the unbound drug concentration is more closely related to the activity of a drug than total plasma/serum concentration because generally only an unbound drug may pass through most cell membranes.

[0054] Similarly, non-specific protein binding is also a factor in the determination of in vivo hepatic clearance based on in vitro intrinsic clearance. Knowledge of the partitioning behavior of a therapeutic compound into red blood cells may be important to the interpretation and understanding of the compound's pharmacokinetic profile. For example, a high partitioning ratio may indicate accumulation of the compound in red blood cells, and potential

hematotoxicity. Thus, as a consequence of these non-specific in vivo interactions, a drug compound that is capable of binding to a specific target molecule effectively in vitro may show substantially less efficacy than contemplated when administered in vivo. Thus, in various embodiments, the extent of binding to plasma proteins may be an important determinant of drug distribution and elimination (e.g., the D and E in ADMET).

[0055] While the intrinsic biochemical effect of a drug candidate (e.g., its binding affinity for a particular target) may be determined by in vitro tests at a very early research stage, the ultimate concentration at the target site can usually only be studied through experiments on the whole organism, which typically occurs at the expensive preclinical and clinical stages. This means that the information may only become available late in the research and development process, and is generally unavailable during the initial compound screening and

pharmacological optimization cycles.

[0056] DSC is a thermoanalytical technique in which the difference in the amount of heat required to increase the temperature of a sample and reference is measured as a function of temperature. The basic principle is that when a sample undergoes a physical transformation such as a melting transition, more or less heat will flow to the sample than the reference to maintain both at the same temperature. By observing the difference in heat flow between the sample and reference, DSC measures the amount of heat absorbed or released during temperature induced transitions. This heat is essentially the enthalpy of the transition and, in various embodiments, may provide insights into the thermodynamic stability of the protein sample. In accordance with various embodiments, when a ligand binds to a protein, it may affect stability of the protein structure and thereby induce changes in the shape of the heat signature, or thermogram.

[0057] Prior to the present disclosure, equilibrium dialysis generally was used to detect ligand binding. However, this method suffers from several inherent limitations. Even a modest binding study at a single ligand concentration with a subset of the 16 major proteins in plasma requires substantial resources. Also, some "sticky" compounds tend to bind ubiquitously to anything with which they come into contact, including the dialysis membrane, which can lead to erroneous binding results. Provided the ligand is soluble in aqueous solution, the methods disclosed in various embodiments using DSC do not suffer from such complications.

[0058] Thus, in various embodiments, the methods and systems disclosed herein may provide robust and reproducible results based on thermal stability rather than molecular weight and charge, which form the basis for conventional techniques (e.g., gel electrophoresis and mass spectrometry). The disclosed embodiments also may require only very small samples and short processing times. In various embodiments, the disclosed DSC-based systems and methods have many advantages that were not available prior to the present disclosure. More specifically, the disclosed DSC-based systems and methods may provide dissection of binding free-energy, into enthalpic and entropic components, providing a complete thermodynamic profile (e.g., ΔΘ = ΔΗ - TAS). Additionally the DSC-based methods and systems disclosed herein may provide unique insights into the origin of binding (e.g., hydrophobic vs. ionic interactions, or binding vs. non-binding). For example, in various embodiments, DSC titration of HSA with a ligand of known affinity shows clear binding characteristics, which may be illustrated by clear shifts in a DSC thermogram.

[0059] In various embodiments, additional techniques may be implemented in order to verify and complement the DSC measurements. For example, in various embodiments, a fluorescence melting curve may be used as an independent measure of the shift in melting temperature, AT _m , that occurs due to binding. In other embodiments, isothermal titration calorimetry (ITC) may be used as an independent measure of the melting enthalpy, ΔΗ, which may be compared with that measured by DSC. In various embodiments, ITC measurements also may provide an evaluation of the binding constant and the number of binding sites. Thus, in various embodiments, the following strategy may be used: if DSC provides AT _m and altered ΔΗ in the presence of the binding entity, further characterization may be achieved through fluoresence shift and ITC measurements.

[0060] Thus, disclosed in various embodiments are systems and methods, including integrated DSC-based systems and methods, that may be used to determine one or more favorable characteristics of in vivo absorption, distribution, metabolism, and/or excretion (ADME) for a biomolecule or small molecule ligand. In various embodiments, collectively, the four ADME criteria may influence the drug levels and kinetics of drug exposure to tissues, and consequently, may influence the performance and pharmacological activity of the ligand. In some embodiments, data generated by the disclosed systems and methods may be utilized in conjunction with currently available in silico drug design software. In some embodiments, a differential scanning calorimeter may be integrated with a computer employing enthalpy of transition calculation software to produce a high-throughput system for measuring ADMET parameters to facilitate the early stage selection of drug candidates in a drug discovery process based upon the pharmacokinetic and pharmacodynamic profiles for a drug interaction with a biomolecule (such as, e.g., human serum albumin (HSA) or blood plasma).

[0061] In various embodiments, the systems and methods described herein may be carried out under computer control, and may be used for selecting a lead drug candidate from a plurality of potential drug candidates. In various embodiments, lead drug candidates that are identified using these systems and methods may have reduced blood protein binding characteristics as compared to other candidate compounds. In some embodiments, the disclosed systems and methods may provide a means for identifying two or more candidate compounds from a selection of compounds that are capable of binding to a specific target molecule and, thereby, altering the function of the target molecule through in vivo activation, deactivation, catalysis, or inhibition. In various embodiments, the disclosed methods may include the step of combining two or more compounds from the plurality of compounds with at least one blood protein to form a first combination, a second combination, or more

combinations from the plurality of compounds.

[0062] Some embodiments may employ a device, such as a DSC, for generating and storing a range of physical transformation data, such as thermodynamic parameters, for one or more combinations of compounds and/or biomolecules, in operative combination with a computer that employs enthalpy of transition calculation software, for comparing and ranking physical transformation data for a first combination of compounds and/or biomolecules with physical transformation data for a second combination of compounds and/or biomolecules to permit the determination of one or more ADMET parameters for each of the first and second combinations. For example, in various embodiments, by comparing and ranking a range of physical transformation data for a first combination and a second combination, the disclosed systems and methods may permit the identification of compounds that are less likely to bind to a protein, such as a plasma protein, as compared to other compounds that are more likely to bind to the protein, thus reducing their bioavailability.

[0063] Other embodiments are methods for identifying a drug candidate having suitable

ADME characteristics. These embodiments include: (a) measuring a first range of physical transformation data in a first combination having one or more biomolecule(s) without a drug candidate; (b) measuring a second range of physical transformation data in a second combination with the biomolecule(s) in the presence of the drug candidate; (c) calculating a difference between an enthalpy of transition for the first combination and the second combination (ΔΔΗ) to obtain at least one binding interaction parameter for the binding of the drug candidate with the biomolecule(s); and (d) comparing the binding interaction parameter to at least one mathematical expression correlating binding interaction data measured for a known drug compound having a known pharmacokinetic parameter to determine at least one pharmacokinetic parameter of the drug candidate.

[0064] Some embodiments further include measuring a third range of physical transformation data in a third combination of the biomolecule(s) and a second drug candidate. In some embodiments, these methods employ a system, as described in further detail herein, that includes a DSC in operative combination with a computer that employs enthalpy of transition calculation software to permit the determination of one or more ADMET parameters.

[0065] In various embodiments, the pharmacokinetic parameter may be an absorption parameter, a distribution parameter, a metabolism parameter, or an excretion parameter. For example, in various embodiments, the pharmacokinetic parameter may be a volume of distribution, a total clearance, protein binding, tissue binding, metabolic clearance, renal clearance, hepatic clearance, biliary clearance, intestinal absorption, bioavailability, relative bioavailability, intrinsic clearance, mean residence time, maximum rate of metabolism, a Michaelis-Menten constant, a partitioning coefficient between a tissue and blood or plasma, a fraction excreted unchanged in urine, a fraction of drug systemically converted to a metabolite, an elimination rate constant, a half-life, or a secretion clearance. Similarly, in various embodiments, partitioning coefficients between tissues and blood or plasma may be partitioning coefficients associated with the blood-brain barrier, blood-placenta barrier, blood-human milk partitioning, blood-adipose tissue partitioning, or blood-muscle partitioning.

[0066] In other embodiments, the methods may further include: (a) estimating at least two pharmacokinetic parameters of a drug candidate and/or (b) estimating a solubility property of the drug candidate. In other embodiments, the mathematical expression correlated from binding interaction data associated with known drug compounds may be a function fitted to a plurality of data points plotted on a Cartesian coordinate system. In some embodiments, the method may include one or more biomolecules that are selected from isolated plasma proteins, liposomes, CYP 450 enzymes, metabolic enzymes, and transport proteins.

[0067] Thus, in various embodiments, the present disclosure overcomes the

shortcomings of the conventional drug-development process by providing high throughput methods and systems for determining thermodynamic parameters for the binding of an active agent in a complex mixture of specific and non-specific molecules. Moreover, in various embodiments, the systems may be employed in methods to determine the varying degrees of T _m shift and enthalpy changes for a particular binding site or active agent type. In various embodiments, the systems and methods disclosed herein may be utilized in vitro to decrease the failure rate of lead compounds due to adverse ADMET properties.

[0068] Any DSC may be used to carry out the disclosed methods, for instance a GE

MicroCal DSC, a TA Instruments DSC, a Perkin Elmer DSC, or the like. One of ordinary skill in the art will recognize that the computing system may use software that may enable database creation and database comparison tools, and substitutions of software that is functionally equivalent is considered to be within the spirit and scope of the disclosure. DSC is discussed at greater length below.

[0069] Differential Scanning Calorimetry (DSC)

[0070] DSC is a thermoanalytical technique that may be used to determine the difference in the amount of heat required to increase the temperature of a sample and a reference, measured as a function of temperature, and is described in U.S. Patent No.

3,263,484, which is incorporated by reference herein in its entirety. Briefly, the technique may include simultaneously applying heat to a sample material and a reference material. In various embodiments, as the sample material goes through various physical and chemical changes such as crystallization, melting, freezing, oxidation, etc., its temperature may be affected by the changes in internal energy. In various embodiments, the differences in temperature between the sample and reference may be recorded, and calculations may then be made for

determining the internal energy changes occurring in the sample. [0071] Generally speaking, when the sample undergoes a physical transformation such as a phase transition, more or less heat may need to flow to it than the reference in order to maintain both at the same temperature. Whether less or more heat must flow to the sample may depend on whether the process is exothermic or endothermic. For example, as a solid sample melts to a liquid, it may require more heat flowing to the sample to increase its temperature at the same rate as the reference. This is due to the absorption of heat by the sample as it undergoes the endothermic phase transition from solid to liquid. Likewise, as the sample undergoes an exothermic process (such as crystallization), less heat may be required to raise the sample temperature. In various embodiments, by detecting the difference in heat flow between the sample and the reference, differential scanning calorimeters may measure the amount of heat absorbed or released during such transitions. In some embodiments, DSC may also be used to observe subtler phase changes, such as glass transitions.

[0072] In various embodiments, both the sample and the reference may be maintained at nearly the same temperature throughout the procedure. Generally, the temperature program for a DSC analysis may be designed such that the sample holder temperature increases linearly as a function of time. In various embodiments, the reference sample may typically have a well-defined heat capacity over the range of temperatures to be scanned.

[0073] DSC Curves

[0074] In various embodiments, DSC may result in a curve of heat flux versus temperature or versus time. Generally speaking, there are two different conventions:

exothermic reactions in the sample are observed with a positive or negative peak, depending on the kind of technology used. In various embodiments, this curve may be used to calculate enthalpies of transitions. In various embodiments, integrating the peak corresponding to a given transition may complete such calculations. The enthalpy of transition may be expressed using the following equation: ΔΗ = KA, wherein ΔΗ is the enthalpy of transition, K is the calorimetric constant, and A is the area under the curve. In various embodiments, the calorimetric constant may vary from instrument to instrument, and may be determined by analyzing a well-characterized sample with known enthalpies of transition.

[0075] Many companies produce DSC machines or functional equivalents of DSC machines that may be used in the disclosed embodiments, such as GE MicroCal (22 Industrial Drive East, Northampton, MA 01060); TA Instruments (159 Lukens Drive, New Castle, DE 19720); and Perkin Elmer (710 Bridgeport Avenue, Shelton, CT 06484), among others.

Additionally, one of ordinary skill in the art will appreciate that other methods may be employed for measuring the heat capacity of a biological fluid at a given temperature. As such, devices, methods, or equivalents thereof that may be useful for measuring the heat capacity of a fluid over a range of temperatures are also contemplated, and fall within the spirit and scope of the present disclosure. [0076] In order to understand how small variations in a complex composition may be used in various embodiments as comparative reference data sets, it is helpful to first understand how the same parameters may be affected when a pure substance is contaminated with other substances. Generally, forming a composition by mixing one or more elements with a pure element may alter the melting temperature, glass transition temperature, or

crystallization temperature of the composition, as well as other chemical and physical characteristics. For example, alloys may have enhanced properties when compared to the pure substance (e.g., steel is stronger than iron). Unlike a pure substance, mixtures do not necessarily have a single melting point, but may have a melting range in which the composition is a blend of solid and liquid phases. This combination, in turn, may produce a unique melting point.

[0077] In various embodiments, a DSC plasma or serum thermogram may represent a composite melting curve of 3,000 or more proteins that make up the plasma proteome. Of these, only about 16 major blood proteins are present in a concentration sufficient for their melting curves to directly manifest in the DSC thermogram, in accordance with various embodiments. Thus, in various embodiments, the complicated plasma or serum mixture may be sensitive to interactions with a drug candidate. For example, in embodiments, a drug candidate may bind to or interact with one or more of the 16 major plasma proteins, which may alter one or more of the primary, secondary, tertiary, and/or quaternary structures. In various embodiments, the result may be a radical shift, in a mass weighted manner, in the DSC plasma thermogram of a sample containing a drug candidate that binds to a plasma protein, vs. a drug candidate that does not bind to a plasma protein. Thus, in various embodiments, DSC plasma thermograms may be sensitive to interactions between a drug candidate and any of the (approximately) 16 major plasma component proteins, some of which include HSA, transferrin, fibrinogen, and IGg.

[0078] In various embodiments, because HSA is the most abundant plasma protein,

DSC may be used on a HSA sample to determine whether a drug candidate binds to HSA. Likewise, in various embodiments, DSC may be used to detect binding of a drug candidate to transferrin, fibrinogen, iGg, or any other plasma protein.

[0079] Statistical Analysis, Comparison, and Classification of DSC Thermograms

[0080] In accordance with various embodiments, melting curves of plasma from human blood or a component thereof measured by DSC thermograms may be used to detect binding to a drug candidate. In one specific, non-limiting embodiment, a general statistical methodology was developed to analyze DSC plasma thermogram data sets collected for human plasma and components thereof. In various embodiments, the statistical metric may provide estimates of the likelihood that a particular drug candidate binds to plasma or a component thereof, such as HSA. In some embodiments, analysis of an acquired DSC thernnogrann may involve comparison to a database of empirical reference sets of DSC thermograms. In various embodiments, two parameters, a distance metric (P) and correlation coefficient (r), may be used to produce a combined 'similarity metric', p, which can be used to classify drug candidates regarding their ability to bind plasma proteins.

[0081] In some embodiments, a DSC thermogram may be expressed as a p- dimensional vector x(T) = (xi , x ₂ , x _p ) with each entry x, corresponding to the measured heat capacity at each temperature T,. In some embodiments, data may be collected over temperatures ranging from about 25 to 1 10 °C, with 10 measurements for each degree (C), resulting in p ~ 750. In some embodiments, a reference set of thermograms may then include a collection of N vectors {Xj(T)| j = 1 ...N}, where each Xj(T) represents a single reference thermogram. In some embodiments, the median thermogram, x _ref (T), of a reference set may be computed from all curves within the set and serves as a template thermogram representing that category. In various embodiments, variance within a reference set may be quantified at each temperature by upper and lower quantile vectors, a _upP er(T) and ai _OW er(T), respectively. These vectors establish the 0.05 to 0.95 quantile boundaries, wherein 90% of the measured data lies. In the case that a more detailed knowledge of the distributions underlying the data is obtained, these parameters may be refined using distribution-dependent measures of central tendency and variance.

[0082] In various embodiments, the quality of a potential reference set of thermograms for a candidate drug may be assessed using descriptive statistical measures. For example, to determine the extent to which a given curve in the reference set aligns with the median thermogram, each reference curve may be compared to the median thermogram, and the linear correlation coefficient (rvalue) may be determined, resulting in a distribution of rvalues. In various embodiments, high levels (>0.8) of correlation may indicate that a given reference set is homogeneous in shape, and may be taken as an indication that the median thermogram x _ref (T) may be reliably used as a template curve. For simplicity, in some embodiments, the Pearson's correlation coefficient may be used here, however, more general and non-parametric correlations (such as Kendall's tau) may also be used when appropriate. In some embodiments, quantile box-plots may also be constructed to assess the variability or degree of homogeneity in a reference set.

[0083] Various methods of comparing thermograms may be used, in accordance with the present disclosure. In general, these methods may be aimed primarily at addressing the diagnostic need; e.g., determining to what degree a test curve, xtest(T), aligns with a given reference template curve, x _re f(T). In various embodiments, the degree of similarity between a test curve and a reference thermogram may be characterized by two factors: (1 ) closeness in space (standardized Euclidean distance) at each temperature point, and (2) similarity in shape (correlation). In general, two thermograms may be highly correlated but, due to vertical scaling, may be separated by a nontrivial distance in space. Conversely, two thermograms may be spatially close but poorly correlated due to fluctuations or noise in the data. For these reasons, in various embodiments, the metric employed to quantify similarities between test and reference curves may be a combination of both spatial distance and linear correlation.

[0084] In various embodiments, distance between two curves may be

determined using a similarity index, P(xtest, ref, CT| _OW er, CT _upP er) defined as follows. At each temperature T, the standardized distance between x _tes t and x _re f is calculated as, d(T) = abs[(x _test (T) - x _rei <T))/ σ _Γεί (Τ)], where o _ref {T) = x _rei {T) - Oi _ower (T) if x _test (T) < x _rei {T) and o _ref {T) = x _rei {T) + o _upper (T) if x _test (T) >

[0085] In various embodiments, the standardized distance d(T)may be

interpreted as the distance associated with a given reference data set that takes into account empirical distributions in the form of quantiles recorded for particular data sets. In various embodiments, a value of d(T) > 1 may indicate that, at the temperature T, the test curve is more distant from the median than 90% of the data in the reference set. In various embodiments, if the specific form of the distribution is known, then d(T) may be interpreted as a z-score, and the probability distribution function representing the reference data may be used to compute critical values at each temperature.

[0086] In various embodiments, the function p(T) may return high values (0.9) at temperatures for which the test curve falls within the quantile boundary, and may return low values (0.1 ) at temperatures for which the test curve falls outside the quantile boundary. Thus, in various embodiments, the function p(T) may represent a likelihood, based on quantile values, that the test curve is similar to the reference set at each temperature point. In various embodiments, no assumptions may be made about the distribution of the reference thermograms. As a result, in some embodiments, this choice of function may not be optimal for discrimination of test and template

thermograms. In the case of a known distribution of reference curves, more

appropriate forms of p(T), such as Gaussian or logistic functions mau be employed. [0087] In various embodiments, a scalar quantity representing similarity of the entire test curve to the reference set is then computed as the arithmetic mean of p(T) over all temperatures (T,, i=1 ,2, ... ,n):

P =∑p(T _; )/n

i= l

[0088] In various embodiments, the metric P may be interpreted as a probability determined by the standardized multivariate distance between the test curve and the reference template. In various embodiments, a value of P near unity may indicate the test curve is closer to the reference template than 90% of the reference data, while a value of P near zero indicates that the test curve is more distant than 90% of the data.

[0089] In various embodiments, similarity in shape between a test curve and a reference set may be quantified using a linear correlation, r, such as Pearson's or Kendall's tau correlation. In some embodiments, two thermograms that are linearly correlated may necessarily possess similar shapes, so the linear correlation may be an effective measure for discriminating between curves different shapes (assuming similar scaling of the data). In some embodiments, linear correlation may provide valuable information about the shape of test curves, and may help to support and strengthen conclusions about degrees of similarity between test and reference curves. In some embodiments, due to similarities in the overall protein composition of human blood plasma, any two thermograms may be highly correlated in certain temperature regions. In various embodiments, in very low (20-50°C) and very high (90-120°C) temperature regions, major differences in thermogram shape are seldom found. As a resultjn various embodiments, the value of a linear correlation coefficient, r, between a test curve and a reference median curve may, in practice, never be negative, and may seldom even be close to zero. On an absolute scale, in various embodiments, interpretation of r-values with regard to the strength of relationship on an absolute scale may be done with some amount of care. In practice, in various embodiments, initial characterization of the data may help to determine significant levels of rfor interpretive use. However, for the purposes of comparison of similar thermograms, in various embodiments, the relative scale of r may be more valuable, and may be established with training data and empirical calibration.

[0090] In various embodiments, a composite parameter, p, may then be introduced that combines both the standardized distance function, P, and the correlation coefficient, r, into a single metric. That is, (in the unlikely case that r≤ 0, p = 0). In embodiments, the range of p may be [0, 1 ], with values closer to zero indicating large differences in shape, and values approaching 1 indicating high similarity. In various embodiments, in order to produce a high value of p, high values of both P and r may be necessary, while a small value of either P or r alone may be sufficient to produce a low p value. In various embodiments, the absolute scale for p may depend on the particular reference set (or sets) employed. Instead, a relative scale may be empirically determined based on the training data, and the metric may be calibrated before application to unknown test curves. In various embodiments, the similarity metric, p, may incorporate both distance and correlation into a single quantitative statistic that may then be used for discrimination between test curves and reference templates.

[0091] Isothermal Titration Calorimetry

[0092] As described above, ITC may be used to confirm or refine a parameter detected by DSC. As used herein, the term "isothermal titration calorimetry" or "ITC" refers to a physical technique used to determine the thermodynamic parameters of interactions in solution. ITC is most often used to study the binding of small molecules (such as medicinal compounds) to larger macromolecules (proteins, DNA etc.) ITC is a quantitative technique that may directly measure the binding affinity (K _a ), enthalpy changes (ΔΗ), and binding stoichiometry (n) of the interaction between two or more molecules in solution. From these initial measurements, Gibbs energy changes (ΔΘ), and entropy changes (AS), may be determined using the relationship: AG = -RTInKa = AH -TAS, (where R is the gas constant and T is the absolute temperature).

[0093] Drug Candidates

[0094] The disclosed system and methods may be used in various embodiments to detect binding of a drug candidate belonging to any class of drugs to a plasma or serum protein. As used herein, the term "drug class" refers to a classification method used to group a chemical or biological composition to the type of the active ingredient or by the way it is used to treat a particular condition, wherein each drug candidate may be capable of fitting into one or more drug classes. As used herein, the term "drug candidate" denotes one or more known or unknown chemical or biological compositions that may be classified by the chemical type of the active ingredient or by the way the drug candidate composition is proposed to treat a particular condition, wherein each "drug candidate composition" may be classified into one or more drug groups that having one or more of the following drug categories or drug classes including, but not limited to: ace inhibitors, addiction aids, aldosterone antagonists, alpha blockers, ALS agents, Alzheimer's disease drugs, aminoglycosides, anesthetics/sedatives, angiotensin II inhibitors, antacids, anti-arrhythmics, antibiotics, anti-cholinergics, anti-convulsants, anti- depressants, anti-diarrheals, anti-emetics, anti-fungals, anti-hepatitis agents, anti-herpetic agents, antihistamines, anti-hypertensive combinations, anti-hypertensives, anti-influenza agents, anti-malarials, anti-platelet agents, anti-psychotics, anti-spasmotics,

antitussives/expectorants, benzodiazepines, beta blockers, bile acid sequestrants,

bisphosphonates, burn preparations, calcium channel blockers, calcium supplements, cephalosporins, colony stimulating factors, corticosteroids, decongestants, antidiabetic agents, direct thrombin inhibitors, disease modifying agents, diuretics, erectile dysfunction drugs, fever inducing agents, fibrates, fluoroquinolones, H2 blockers, hypertension drugs, anti-HIV agents, hormone replacement drugs, interferons, immunizations, insulin, laxatives, low molecular weight heparins, macrolides, magnesium supplements, mouth & lip preparations, multiple sclerosis drugs, muscle relaxants, nasal preparations, neuromuscular blocking agents, nitrates, NSAIDs, ophthalmic preparations, otic preparations, opiates, anti-Parkinson's agents, analgesics, penicillins, phosphate supplements, potassium supplements, proton pump inhibitors, respiratory medications, statins, stimulants, tetracyclines, thrombolytics, thyroid hormones, anti-tuberculosis agents, topical (antibacterials), vaginal preparations, vasodilators, vasopressors and inotropes, and vitamin D analogs.

[0095] In various embodiments, a goal of a drug discovery process is to identify and characterize a chemical compound or ligand, e.g., binder or biomolecule (e.g., receptor) that affects the function of one or more other biomolecules (e.g., a drug "target") in an organism, usually a biopolymer, via a potential molecular interaction or combination. As used herein, the term "biopolymer" refers to a macromolecule that comprises one or more of a protein, nucleic acid (DNA or RNA), peptide or nucleotide sequence or any portions or fragments thereof. As used herein, the term "biomolecule" refers to a chemical entity that comprises one or more of a biopolymer, carbohydrate, hormone, or other molecule or chemical compound, either inorganic or organic, including, but not limited to, synthetic, medicinal, drug-like, or natural compounds, or any portions or fragments thereof. In various embodiments, the target molecule may be a disease-related target protein or nucleic acid for which it is desired to affect a change in function, structure, and/or chemical activity in order to aid in the treatment of a patient disease or other disorder. In other embodiments, the target may be a biomolecule found in a disease- causing organism, such as a virus, bacteria, or parasite, that when affected by the drug will affect the survival or activity of the infectious organism. In other embodiments, the target may be a biomolecule of a defective or harmful cell such as a cancer cell. In yet other

embodiments, the target may be an antigen or other environmental chemical agent that may induce an allergic reaction or other undesired immunological or biological response.

[0096] In some embodiments, the ligand may be a small molecule drug or chemical compound with desired drug-like properties in terms of potency, low toxicity, membrane permeability, solubility, chemical/metabolic stability, etc. In other embodiments, the ligand may be biologic, such as a protein-based or peptide-based drug. In yet other embodiments, the ligand may be a chemical substrate of a target enzyme. In some embodiments, the ligand may even be covalently bound to the target, or may be a portion of the protein, e.g., protein secondary structure component, protein domain containing or near an active site, protein sub- unit of an appropriate protein quaternary structure, etc.

[0097] Throughout the remainder of the detailed description, unless otherwise specifically differentiated, a (potential) molecular combination will feature at least one ligand and at least one target, the ligand and target are usually separate chemical entities, and the ligand will be assumed to be a chemical or biological compound while the target is typically a biological protein (mutant or wild type). As used herein, the term "molecular complex" refers to a bound state between target and ligand when interacting with one another in the midst of a suitable (often aqueous) environment. As used herein, a "potential" molecular complex refers to a bound state that may occur albeit with low probability and therefore may or may not actually form under normal conditions. As used herein, the term "associated" means constituents are bound to one another either covalently or non-covalently, the latter as a result of hydrogen bonding or other intermolecular forces. The constituents may be present in ionic, non-ionic, hydrated or other forms.

[0098] ADME characteristics that may be assessed by the methods disclosed in various embodiments include absorption/administration, distribution, metabolism, and

excretion/elimination. Absorption/administration refers to the process by which a compound reaches a target tissue following administration. This absorption often occurs on mucous surfaces like the digestive tract (intestinal absorption) - after being taken up by the target cells. This can be a serious problem at some natural barriers like the blood-brain barrier. Factors such as poor compound solubility, gastric emptying time, intestinal transit time, chemical instability in the stomach, and inability to permeate the intestinal wall can all reduce the extent to which a drug is absorbed after oral administration. Absorption critically determines the compound's bioavailability. Drugs that absorb poorly when taken orally must be administered in some less desirable way, like intravenously or by inhalation (e.g., zanamivir).

[0099] For distribution, a compound must be carried to its effective site, most often via the bloodstream. From there, the compound may distribute into tissues and organs, usually to differing extents. After entry into the systemic circulation, either by intravascular injection or by absorption from any of the various extracellular sites, the drug is subjected to numerous distribution processes that tend to lower its plasma concentration. Distribution is defined as the reversible transfer of a drug between one compartment to another. Some factors affecting drug distribution include regional blood flow rates, molecular size, polarity and binding to serum proteins, forming a complex.

[00100] Through the process of metabolism, a compound, such as a small-molecule drug, is converted from its administered form to an active form, typically in the liver by redox enzymes known as cytochrome P450 enzymes. As metabolism occurs, the initial (parent) compound is converted to new compounds called metabolites. When metabolites are pharmacologically inert, metabolism deactivates the administered dose of parent drug and this usually reduces the effects on the body. Metabolites may also be pharmacologically active, sometimes more so than the parent drug.

[00101] Compounds and their metabolites are removed from the body via excretion, usually through the kidneys or in the feces. Unless excretion is complete, accumulation of foreign substances can adversely affect normal metabolism. In general, there are at least three main sites where drug excretion occurs. The kidney is where products are excreted through urine. Biliary excretion or fecal excretion is the process that initiates in the liver and passes through to the gut until the products are finally excreted along with waste products or feces. The last method of excretion is through the lungs e.g. anesthetic gases.

[00102] Excretion of drugs by the kidney involves 3 main mechanisms: glomerular filtration of unbound drug; active secretion of (free & protein-bound) drug by transporters e.g., anions such as urate, penicillin, glucuronide, sulfate conjugates or cations such as choline, histamine; and filtrate, wherein a 100-fold concentrated substance in tubules form a favorable concentration gradient so that it may be reabsorbed by passive diffusion and passed out through the urine.

[00103] One of ordinary skill in the art will understand the term "binding mode" refers to the 3-D molecular structure of a potential molecular complex in a bound state at or near a minimum of the binding energy (e.g., maximum of the binding affinity), where the term "binding energy" (sometimes interchanged with "binding free energy" or with its conceptually antipodal counterpart "binding affinity") refers to the change in free energy of a molecular system upon formation of a potential molecular complex, e.g., the transition from an unbound to a (potential) bound state for the ligand and target. The term "system pose" is also sometimes used to refer to the binding mode. Here the term free energy generally refers to both enthalpic and entropic effects as the result of physical interactions between the constituent atoms and bonds of the molecules between themselves (e.g., both intermolecular and intramolecular interactions) and with their surrounding environment, meaning the physical and chemical surroundings of the site of reaction between one or more molecules. Examples of the free energy are the Gibbs free energy encountered in the canonical or grand canonical ensembles of equilibrium statistical mechanics.

[00104] In general, the optimal binding free energy of a given target-ligand pair directly correlates to the likelihood of combination or formation of a potential molecular complex between the two molecules in chemical equilibrium. The binding free energy describes an ensemble of (putative) complex structures and not one single binding mode. However, in computational modeling it is usually assumed that the change in free energy is dominated by a single structure corresponding to a minimal energy. This is generally true for tight binders (pK ~ 0.1 to 10 nanomolar) but questionable for weak ones (pK ~ 10 to 100 micromolar). The dominating structure is usually taken to be the binding mode. In some cases, it may be necessary to consider more than one alternative-binding mode when the associated system states are nearly degenerate in terms of energy.

[00105] Binding affinity is of direct interest to drug discovery and rational drug design because the interaction of two molecules, such as a protein that is part of a biological process or pathway and a drug candidate sought for targeting a modification of the biological process or pathway, often helps indicate how well the drug candidate will serve its purpose. Furthermore, where the binding mode is determinable, the action of the drug on the target can be better understood. Such understanding may be useful when, for example, it is desirable to further modify one or more characteristics of the ligand to improve its potency (with respect to the target), binding specificity (with respect to other target biopolymers), or other chemical and metabolic properties. In various embodiments, the interaction of two known molecules may be significantly altered with the introduction of one or more unknown molecules into the two- molecule system. Various embodiments may allow a user to analyze one or more drug candidates together with one or more biological targets, and the resulting data may be useful for further drug modification or new drug selection.

[00106] Other embodiments involve separating reference and test samples into relative concentrations of the about 16 major component proteins of a plasma or tissue sample, which include HSA, transferrin, fibrinogen, and IGg, wherein a drug candidate interaction with various relative concentrations may yield predictive results for ADMET parameters, In various embodiments, the global signature yielded by thermograms in the presence and absence of a drug candidate may reflect the status of the whole dynamic system. As such, in various embodiments, thermograms may show changes in the "forest rather than in a single tree," or the single tree depending on the procedure.

[00107] The following examples are provided to further illustrate several specific embodiments and the manner in which they may be carried out. It will be understood by one of ordinary skill in the art that the specific details given in the examples have been chosen for purposes of illustration, and are not intended to be construed as limiting.

[00108] EXAMPLES

[00109] EXAMPLE 1: NEW DRUG DEVELOPMENT - ADME SCREENING

[00110] Example 1 illustrates the use of DSC as a screening tool for determining parameters for absorption, distribution, metabolism, elimination (ADME) or toxicity of candidate drug compositions. Figures 1A and 1 B illustrate the results of DSC and fluorescence melting curves of HSA at a concentration of 25 μΜ in phosphate buffer (left) and in the same phosphate buffer plus 5% DMSO (right). The DSC melting curve is shown on the far left of both panels. The DSC curve had a single peak with low temperature and high temperature linear baselines preceding and succeeding the melting transition. The integrated area under the baseline corrected DSC melting curve provides a measure of the melting transition enthalpy, ΔΗ. Parameters obtained in the DSC melting examples, in the two solvents, are summarized in the boxes directly above the displayed DSC melting curves. All values in phosphate buffer without DMSO (1002.69 ± 5.68 kJ/mol) and with DMSO (1000.52 ± 2.17 kJ/mol) were identical.

Likewise, the T _max values measured in the two solvents (62.94 ± 0.14 and 61.27 ± 0.23) were in good agreement.

[00111] DMSO is used as an additive to improve solubility of added binding ligands. The observation that there were no differences in the melting curve parameters with and without DMSO indicates that the effect of DMSO is confined to its intended purpose of only enhancing ligand solubility, while not affecting the stability of HSA.

[00112] The curves on the right of Figures 1A and 1 B are melting curves (10 in the left panel, 12 in the right panel) under the boxes labeled "96 well fluorescence melt" are HSA melting curves measured in a 96 well plate reader of the fluorescence intensities of HSA as a function of temperature in the presence of the fluorescent dye, cypro orange. Florescence of cypro orange linearly decreases with increasing temperature, and cypro orange binds preferentially to the denatured (melted) protein and produces an increased florescence directly proportional to the extent of melting measured as function of temperature.

[00113] The curves below the 96 well fluorescence melting curves, labeled HSA 25 μΜ and HSA 25 μ (5% DM50) are the negative derivatives dF/dT, where F is the fluorescence signal and T is temperature. Derivatives were taken to more precisely determine the transition temperature, T _m . Although at a lower resolution, the fluorescence melting curve measurement provides an independent assessment of the T _m (referred to as T _max , in DSC). As seen, the T _m values measured by fluorescence melting were within 3°C of those measured by DSC, indicating that the two techniques were essentially in agreement for the measured stability of HSA in the two solvents. Although highly amenable to high throughput, fluorescence melting does provide direct measurements of the transition enthalpy, AH, and therefore cannot provide any quantitative insight into the nature of the melting transition other than the relative overall thermodynamic stability reflected in the value of T _m .

[00114] When the example illustrated in Figures 1A and 1 B was repeated in the presence of a ligand, such as bromocresol green, it was found that the composition binds HSA. Figure 2 shows an example of measurements of ligand binding to HSA using three different technologies: (1 ) DSC, (2) fluorescence melting, and (3) ITC. On the right side of Figure 2 is the DSC melting curve (top) of the ligand bromocresol green that binds to HSA. The ratio of HSA to ligand concentration was 1 :10.

[00115] The thermodynamic parameters determined by DSC were increased when compared to the values in Figures 1A and 1 B. More specifically, in the presence of the ligand, ΔΗ increased from 1000.52 to 1021.25 kJ/mol. The DSC T _max increased by 9.26° C and the fluorescent melting temperature, T _m , increased by 10.7°C. These measurements indicate that the ligand, in this case bromocresol green, binds to HSA. The increase in AH and T _m values due to ligand binding is directly related the strength of the "collective" ligand binding constant. However, these measurements generally do not provide further direct insight into the type of binding between bromocresol green and HSA.

[00116] In order to gain insight into the type of binding, ITC experiments may be performed. In an ITC procedure, the ligand is added at a constant temperature in a titrating fashion to the HSA, and the enthalpy and entropy of binding are determined as a function of relative concentration of ligand to HSA. The heat response (enthalpy) measured by ITC upon addition of the ligand at increasing molar ratios of the ligand to HSA is shown for a series of injections in the top of Figure 2 on the left.

[00117] Directly below the ITC binding curve is the enthalpy of binding. This value was determined from the ITC curve and was plotted versus the molar ratio of ligand to HSA. This "saturation binding" curve showed that the ligand was added to excess molar ratios to saturate all available sites on HSA. This binding data was then fit with a two site binding model varying the site occupation numbers, N1 and N2, the binding enthalpies to both sites, DH1 and DH2, binding entropies for both sites, DS1 and DS2, and the binding constants for both sites, K1 and K2, respectively. Results shown in Figure 2 in the box on the left below the plot were obtained using Origin software. Fitting parameters were obtained from a "Mathematica fit" in the box below were obtained from fitting the binding curve using Mathematica and obtaining the listed parameters.

[00118] A comparison of the values obtained using DSC and the other two methods indicates that all three methods provide consistent results, and indicate that at least two sites on HSA are bound by the ligand, in this case, bromocresol green. Although ITC may provide quantitative insight into the type of binding that occurs, it is not easily amenable to high throughput. In the screening strategy, if the DSC and/or fluorescence melting curves reveal a shift due to the ligand, then ITC may be enlisted to more deeply characterize the binding mechanism where required.

[00119] EXAMPLE 2: BINDING OF HSA BY NAPROXEN

[00120] Naproxen sodium is a nonsteroidal anti-inflammatory drug (NSAID) commonly used for the reduction of pain, fever, and inflammation. The diagrams shown in Figure 3 are similar to those found in Figure 2, with the exception that HSA is bound by naproxen.

Moreover, ITC binding curves are shown on the left of Figure 3, along with fitted parameters obtained for the two-site binding model, derived from the two software routines described above. The DSC and fluorescence melting curves are shown on the right, along with the melting curve parameters that were measured. The increased ΔΗ, T _max , and T _m values of HSA in the presence of the ligand verify the ligand binds HSA. The results of two independent ITC measurements and fits of the saturated binding curves indicate the binding mechanism is complicated, can be fit with the two site model, but from the differences in their ITC binding curves, binding behavior of naproxen to HSA is clearly different than binding of bromocresol green in Figure 2.

[00121] EXAMPLE 3: BINDING OF HSA BY SALICYLA TE

[00122] The diagrams shown in Figure 4 are similar to those found in Figure 3, with the exception that HSA is bound by salicylate. Replicate ITC binding curves are shown on the left of Figure 4, along with fitted parameters obtained using a single site model (as opposed to the multi-site models that were fitted in Figures 2 and 3. (Generally, the two-site model is invoked if the data cannot be fit satisfactorily with a one site model). The much smaller values of ΔΗ, T _max , and T _m for HSA in the presence of the ligand verify that the ligand binds HSA, but in a much weaker fashion than do the ligands shown in Figures 2 and 3. The simple linear binding curves were measured by ITC. The observation of relatively weak binding to HSA by salicylate indicates that it has properties desired for a better drug candidate compared to those of naproxen (Figure 3) and bromocresol green (Figure 2), which bind HSA much more strongly.

[00123] EXAMPLE 4: BINDING OF HSA BY BROMOPHENOL BLUE

[00124] Bromophenol blue is commonly used as a color marker to monitor the process of agarose gel electrophoresis and polyacrylamide gel electrophoresis. Although bromophenol blue is not considered a drug, it may be used as an example of a drug candidate that carries a slight negative charge at moderate pH, wherein it can bind with proteins and give blue color. As such, the diagrams shown in Figure 5 are similar to those found in Figure 2, with the exception that HSA is bound by bromophenol blue. As shown in Figure 5, the binding of HSA by bromophenol blue is similar to the binding of bromocresol green in Figure 2. The increased values of ΔΗ, T _max and T _m of HSA in the presence of the ligand verify that the ligand binds HSA. The ITC measurements and resulting binding curve and suitable fits with the two site model indicate that bromophenol blue binds HSA relatively strongly, comparable to bromocresol green in Figure 2. Comparison of this ITC binding curve with the ITC binding curve in Figure 2 indicates that the curves are similar and both are best fit by a two site binding model, indicating they have similar types of (complicated) binding and binding strength.

[00125] EXAMPLE 5: BINDING OF HSA BY PHENOL RED

[00126] The binding characteristics of Phenol Red shown in Figures 6A and 6B are similar to those of salicylate shown in Figure 4. The ITC binding curves that are displayed are not simply replicates under the same conditions (as in previous figures). The binding curve on the left is similar to the ones in previous figures and is similar to saturation binding, when the molar ratio of ligand to HSA is high enough that the ligand saturates all binding sites. As shown, this binding curve is best fit by the single site model with the displayed parameters

(Figure 6A). The binding curve in Figure 6B is the "excess site binding" curve as was measured at molar ratios where the HSA is vast excess of ligand. These procedures were performed to precisely evaluate an enthalpy for the average binding of the ligand to HSA. The enthalpy evaluated in the excess site binding case is -2,399.8 cal/mol, which is about 65% larger than that evaluated from fitting the saturation binding curve with the single site model. This indicates the deficiency of the single site model, and shows that there may be more than one site.

[00127] EXAMPLE 6: BINDING OF USA BY BROMOSULFALEIN

[00128] The binding of HSA by bromosulfalein shown in Figure 7 and the ITC saturation binding curve (far left) indicates binding very similar to that of bromocreosol green in Figure 2. The ITC excess binding procedures revealed an average enthalpy of binding of - 4743.9 cal/mol, which was very close to the average of the two enthalpies determined from the two-site fit of the saturation binding curve, (-3984 + -6101 )/2 = 5043, indicating there are probably two sites on HSA bound by the ligand.

[00129] EXAMPLE 7: BINDING OF HSA BY IBUPROFEN

[00130] The binding of HSA by ibuprofen is shown in Figure 8. The ITC binding saturation-binding curve (left) is similar to that of phenol red in Figure 6 and salicylate in Figure 4. Parameters obtained from fitting the single site model are shown on the left. From two excess site binding curves (middle) different average binding enthalpies were evaluated indicting the presence of multiple sites.

[00131] EXAMPLE 8: BINDING OF HSA BY IMIPRAMINE

[00132] The binding of HSA by imipramine is shown in Figure 9. The small increases in ΔΗ, T _max and T _m of HSA in the presence of the ligand indicate the ligand binds very weakly, if at all, to HSA. This is further evidenced by the flat saturation (left) and excess site ITC binding curves (middle). This is an example of a ligand that does not appreciably bind HSA, which is a desirable characteristic for a therapeutic.

[00133] EXAMPLE 9: BENDING OF HSA BY CHLORPROMAZINE

[00134] The binding of HSA by chlorpromazine is shown in Figure 10, which is similar to that observed for ibuprofen (Figure 8), phenol red (Figure 6), and salicylate (Figure 4).

[00135] EXAMPLE 10: BINDING OF HSA BY OXACILLIN

[00136] The binding of HSA by oxacillin is shown in Figure 11 , which is similar to chlorpromazine (Figure 10), ibuprofen (Figure 8), and phenol red (Figure 6).

[00137] EXAMPLE 11: BINDING OF HSA BY PENICILLIN G

[00138] The binding of HSA by penicillin G (benzylpenicillin) is shown in Figure 12. The small increases in ΔΗ, T _max and T _m of HSA in the presence of the ligand indicate the ligand binds very weakly, if at all, to HSA. This is further evidenced by the fiat ITC binding curves shown on the left. Similar to what was observed for binding of imipramine in Figure 9, this is an example of a ligand that does not appreciably bind HSA, which is a favorable characteristic for a therapeutic.

[00139] EXAMPLE 12: BINDING OF HSA BY EVAN'S BLUE

[00140] The binding of HSA by Evan's blue is shown in Figure 13, and this is a complicated binding similar to bromosulfalein in Figure 7. [00141 ] EXAMPLE 13: BINDING OF HSA BY METHYL ORANGE

[00142] The binding of HSA by methyl orange is shown in Figure 14. This complicated binding is similar to Evan's blue in Figure 13 and bromosulfalein in Figure 7.

[00143] EXAMPLE 14: BINDING OF HSA BY OCTANOA TE

[00144] The binding of HSA by octanoate is shown in Figure 15. This pattern shows similar binding of HSA observed for naproxen in Figure 3.

[00145] EXAMPLE 15: BINDING OF HSA BY SODIUM OLEATE

[00146] The binding of HSA by sodium oleate is shown in Figure 16. These patterns axe similar to oxacillin (Figure 11 ), chlorpromazine (Figure 10), thupropin (Figure 8), and phenol red (Figure 6).

[00147] EXAMPLE 16: SUMMARY OF HSA BINDING BY 18 DIFFERENT LIGANDS

[00148] The results of DSC and fluorescence melting of HSA in the presence of 18 different Iigands is summarized in Figure 17. For some of the Iigands in Figure 17, the explicit experimental data as shown in Figures 2-16 is not shown.

[00149] EXAMPLE 17: SUMMARY OF HSA BINDING BY 15 DIFFERENT LIGANDS

[00150] A summary of DSC and fluorescence melt procedures is shown in Figure 18. Figure 18 shows similar results to Figure 17. For these Iigands the explicit experimental data as shown in Figures 2-16 is not shown.

[00151 ] EXAMPLE 18: SUMMARY OF HSA BINDING GROUPED BY BINDING SITE

[00152] DSC results of HSA melting in the presence of different binding Iigands are shown in Figure 19, grouped according to binding site and experimentally observed ligand dependent changes.

[00153] EXAMPLE 19: ITC SATURATION BINDING SUMMARY

[00154] The fitting parameters obtained from ITC saturation binding procedures for the

Iigands examined are summarized in Figure 20. Some binding curves were fit with a single site model, while most were best fit by a two-site model.

[00155] EXAMPLE 20: ITC EXCESS SITE BINDING SUMMARY

[00156] The results of ITC excess binding are summarized in Figure 21. In most cases, results of two independent procedures are shown. These procedures were performed in vast molar excess of HSS to ligand provide an evaluation of the average enthalpy of binding the ligand to HSA.

[00157] EXAMPLE 21: ITC EXCESS SITE BINDING SUMMARY

[00158] FIGURE 22 shows a comparison of information provided by surface plasmon resonance (SPR) and DSC. SPR provides indirect measurements of the free-energy, Δ G, for a binding reaction. By direct measurement, DSC provides values of the entropy, ΔΗ, and entropy (AS), that together comprise AG and provide not only an evaluation of the binding strength, but in addition to the binding strength also yield insights into the precise

thermodynamic nature and origins of interactions driving the binding reaction. For the binding reaction, SPR can only provide an indirect estimate of AG. DSC provides a direct measure of ΔΚ, AS, and AG.

[00159] EXAMPLE 22: INTERACTION OF HSA WITH THREE AGENTS

[00160] Interaction of HSA with three agents with known and different binding affinities for HSA at a concentration of about 10 ^"5 M is shown in Figure 23. For phenol red, a compound known to have a very low binding affinity for HSA, K _B < 10 ⁵ M, with increased concentrations of the binding ligand, very little change is observed on the measured DSC melting profiles. In contrast, for two ligands (bromocresol green and bromphenol blue) known to have a greater binding affnity for HSA, K _B > 10 ⁵ , clear shifts are observed in the DSC melting profiles.

[00161 ] EXAMPLE 23: STRA TEGY TO SCREEN A LIBRARY OF UNKNOWN IBA

LIGANDS

[00162] A strategy to screen a library of potential binding ligands for HSA is shown in

Figure 24. Ligands having relative binding strengths for HSA, relative shifts in the DSC melting profile are observed. The stronger the binding, the larger the shift.

[00163] EXAMPLE 24: DSC SCREEN FOR MAJOR PLASMA PROTEINS

[00164] Methods showing how DSC may be used to screen for ligands that bind to the major proteins present in plasma are shown in Figure 25. For a single binding ligand,

(bromcresol green) the relative binding strength to each of the six plasma proteins manifests in the relative shift of their DSC melting profile in the presence of the ligand. As shown, strong binding, as in this case exists for HSA, a significant shift in the melting curve is observed. For lower binding affinity, minimal shifts are observed and only more subtle changes occur in the

DSC melting profile.

[00165] EXAMPLE 25: DSC AND DRUG INTERACTIONS WITH WHOLE PLASMA

[00166] Drug interactions with whole plasma and with an individual plasma protein, HSA, are shown in Figure 26. The drug is bromocreosol green. This figure demonstrates how the individual components of plasma that bind the ligand may be identified.

[00167] EXAMPLE 26: BINDING OF PENICILLIN AND BILIRUBIN TO HSA

[00168] DSC thermograms were measured for HSA in the presence of increasing concentrations of each ligand. As shown in Figures 27 and 28, HSA thermograms were affected differently by the presence of penicillin and bilirubin. In Figure 27, for HSA at 10 μΜ, as penicillin is increased from 0.1 to 50 μΜ, the thermogram peak temperature (T _m ) shifted up and increased in height with increasing ligand

concentration. The clear shifts in Tm increasing ligand concentration indicate it binds in a likely specific manner. Thermodynamically, the enthalpy of melting was increasing, indicating as ligand concentration increases, the specific site(s) on HSA were bound by the ligand. This binding to the native structure of HSA stabilizes it against denaturation. The observed stabilization by site-binding is consistent with those inferred from ITC measurements, wherein penicillin was shown to bind HSA at a single site with a significant binding enthalpy.

[00169] In Figure 28, again for HSA at 10 μΜ, as bilirubin was increased from 0.1 to 50 μΜ, the behavior of the thermogram with increased ligand concentration differed from penicillin. As ligand concentration increased, the thermogram peak height increased, but the T _m did not significantly change. In independent ITC experiments, bilirubin was shown to bind HSA in a pseudo multiple-site manner. Clearly, DSC can not only detect ligand binding, but can also distinguish between the different types of binding to HSA by penicillin and bilirubin. In order to obtain insight into the origins of these differences in binding, a simple binding model analysis was performed.

[00170] An equilibrium model was formulated to describe the following binding reaction: NL \ *H N + L j D + L j DL , in which the native (N) protein melts to the denatured form (D) with equilibrium constant Ku; and the native and denatured forms can be bound by the ligand (L), with binding constants K _N and K _D , respectively resulting in NL and DL. Equilibrium constants are given by,

K _N = C _NL /C _N ^■ C _L ; K _U = C _DL /C _NL ; K _D = C _DL /C _D ^■ C _L . From conservation of mass considerations and the total protein concentration given by C _p ^Tot = C _N + C _NL + C _D + C _DL , the fraction of molecules in each state (e.g., N _L , N, D and D _L ) are determined from the relative concentrations and the respective enthalpies of each binding reaction, e.g., AH = f _N ^■ AH _NL + f _D - AH _ND + f _DL - AH _DL the quantity compared to experiments is given

[00171] Calculations of AC _P were made by inputting values of AH _NL , ΔΗ _Ν0 , ΔΗ ₀ ι_ and protein and ligand concentrations. Two scenarios were considered, and results are shown in Figures 29 and 30. In the first case, binding to only the native form in the reaction was considered, e.g., K _N » K _D , where binding of the ligand to the denatured form was not allowed. As Figure 29 clearly shows, a titratable increase in thermogram T _m and peak height were observed with increasing ligand concentration. Comparison with the data in Figure 27 for penicillin revealed that the curves and trends were in semi-quantitative agreement.

[00172] The calculated thermograms in Figure 30 were obtained with the model by assuming binding of the ligand to both the native and denatured forms is allowed and occurs with equal likelihood, K _N = K _D . As seen in Figure 30, under these conditions, a titratable increase in peak height, but no change in T _m was observed. This behavior was semi-quantitatively similar to that observed for the HSA thermograms in the presence of bilirubin in Figure 28. Thus, the model was able to produce experimental observations for two different classes of binding ligands.

[00173] EXAMPLE 27: DETECTION AND CHARACTERIZATION OF UNKNOWN

LIGAND BINDING

[00174] In this Example, binding an unknown ligand (an anonymous drug candidate for cancer) to HSA, whole plasma, and HSA-depleted plasma was investigated by DSC analysis. As in other binding experiments, total protein concentration was 10 μΜ. Thermograms were measured for protein solutions in the presence of 0.1 , 1 .0 and 10 μΜ ligand. Control thermograms for the three protein samples are shown in Figure 31. Thermograms of the different protein solutions with added ligand are shown in Figures 32-34. In Figure 32, effects of binding on HSA are shown. This behavior is qualitatively similar to that observed for HSA in the presence of bilirubin in Figure 29. However, in this case, the peak height decreased (in contrast to bilirubin) with increasing ligand concentration. Model analysis indicated this behavior can be predicted if the ligand binding to the denatured form is favored over the native form, KN « KD, indicating thqt binding to the native form of HSA is relatively weak. Figure 33 shows the effect of ligand binding to whole plasma.

[00175] Clear shifts in the plasma thermogram are observed with increasing ligand concentration. Evidently, some of the protein components in plasma are bound by the ligand. Owing to the results observed for HSA alone, observations on the whole plasma thermogram in the presence of ligand are due to binding of a protein component in plasma other than HSA. Results in Figure 34 are consistent with this assessment, and show the effect of ligand binding to plasma depleted of HSA. A clear change in thermogram shape and size is observed with increasing ligand concentration. This observation clearly indicates the ligand binds protein(s) in plasma other than HSA.

[00176] EXAMPLE 28: SPECIFICITY STUDIES

[00177] The aim of this Example was to determine if the compound interacts with four of the most abundant proteins in plasma. These are Complement C3, Complement C4,

Transferrin (TRF), and Ceruloplasmin. Changes in the protein thermograms with increased concentrations of added compound revealed the presence of binding to the protein. In each melting procedure, the individual protein was present at 10 μηη and the ligand was present at 0.1 , 1 .0 and 10 μηη, corresponding to molar ratios of protein to ligand of 100:1 , 10:1 , and 1 :1 , respectively.

[00178] Figure 35 shows DSC thermograms of complement C3 at 10 μηη in the absence and presence of added ligand at 0.1 , 1 .0 and 10 μηη. Compared to the C3 thermogram in the absence of ligand (solid line), at the lowest ligand concentration (0.1 μηη) the thermogram actually decreased in intensity with no change in transition temperature. At the higher concentration in the presence of ligand, the thermogram intensity increased significantly, but again with no change in the transition temperature. These observations indicate reasonable binding of the ligand to complement C3.

[00179] Figure 36 shows DSC thermograms of complement C4 at 10 μηη in the absence and presence of added ligand at 0.1 , 1 .0 and 10 μηη. Compared to the C4 thermogram in the absence of ligand (solid line) which displayed two peaks, in the presence of ligand the peaks coalesced to a single peak. At the lowest and highest ligand concentrations (0.1 and 10 μΜ) the thermogram peak increased in intensity with increased ligand concentration with little or no change in transition temperature. At a ligand concentration 1.0 μηη, the thermogram intensity decreased below that of the protein alone. These observations indicate a more complicated binding mechanism for C4, but are consistent with reasonable binding of the ligand to the protein.

[00180] Figure 37 shows DSC thermograms of ceruloplasmin at 10 μηη in the absence and presence of added ligand at 0.1 , 1 .0 and 10 μηη. Compared to the ceruloplasmin thermogram in the absence of ligand (solid line), as ligand concentration increased, the thermogram peak increased with minor changes in the transition temperature. These observations further indicate reasonable binding of the ligand to ceruloplasmin.

[00181] Figure 38 shows DSC thermograms of transferrin at 10 μηη in the absence and presence of added ligand at 0.1 , 1 .0 and 10 μηη. Compared to the transferrin thermogram in the absence of ligand (solid line), as ligand concentration increased, the thermogram peak intensity decreased with no changes in the transition temperatures. In contrast to what was observed for complement C3, complement C4 and ceruloplasmin, these observations indicate relatively weaker TRF binding to the ligand.

[00182] Although certain embodiments have been illustrated and described herein, it will be appreciated by those of ordinary skill in the art that a wide variety of alternate and/or equivalent embodiments or implementations calculated to achieve the same purposes may be substituted for the embodiments shown and described without departing from the scope.

Those with skill in the art will readily appreciate that embodiments may be implemented in a very wide variety of ways. This application is intended to cover any adaptations or variations of the embodiments discussed herein. Therefore, it is manifestly intended that embodiments be limited only by the claims and the equivalents thereof.