Patent Application
20110081675
This invention provides methods and biomarkers for identifying cardiotoxic effects of pharmaceuticals, biologics, and other chemical compounds and environmental agents. The invention specifically provides methods for identifying low molecular weight metabolites secreted by cardiomyocytes in response to in vitro exposure to cardiotoxic compounds. Metabolomic methods are provided for identifying candidate biomarkers predictive of cardiotoxicity by measuring low molecular weight metabolites produced and secreted by cardiomyocytes contacted with a chemical compound, pharmaceutical, biologic or environmental agent. Predictive biomarkers for cardiotoxic effects are also identified and provided herein.
Cardiotoxicity has become one of the leading causes of pharmaceutical lead compound attrition and subsequent withdrawal of FDA-approved drugs from the market. The development of screening methods that provide specificity and accuracy for predicting cardiotoxicity are needed to better enable safe drug development and to help reduce soaring financial losses associated with preclinical drug failure.
Currently, cardiotoxicity can only be inferred, predominantly by measuring in vitro alterations to the action potential duration (APD) in cardiomyocytes using patch-clamp procedures. Despite the invaluable knowledge generated by electrophysiology assays, patch clamp procedures are extremely time consuming and low throughput. Briefly, the APD response to pharmaceutical compounds is measured a single cell at a time, and even so-called “high throughput” systems, such as PatchExpress®, only permit recordings of dozens of cells per assay. Most importantly, however, the mechanism of pharmacological cardiotoxicity is not uniform across drugs; thus electrophysiology recordings are limited in their ability to predict the cardiotoxicity of multiple compounds. While certain compounds exert their toxicity primarily by interfering with proper function of cardiac ion channels (which translate into changes to the APD and thus can be detected using conventional assays), others are known disrupters of cardiomyocyte metabolism that are not currently assayed. The primary toxicity of chemotherapies and kinase inhibitors used for cancer therapy, for example, results in significant changes to metabolic indicators in cardiomyocytes. Independent of the mechanism, cardiotoxicity would ultimately produce changes to the comprehensive collection of low molecular weight molecules from cardiomyocytes.
Dysregulation of metabolite synthesis, processing and abundance has been associated with cardiotoxicity. Chemotherapeutic and anti-tumor regimens are accompanied by marked changes to mitochondrial function, including interference with oxidative phosphorylation and inhibition of ATP synthesis, myofibrillar structure, and other aspects of energy metabolism. (Takemura & Fugiwara, 2007, Progress in Cardiovascular Diseases 49(5): 330-352). Other metabolic processes that have been implicated in the cardiotoxicity of cancer drugs include lipid peroxidation, oxidation of proteins and DNA, and depletion of glutathione and pyridine nucleotide reducing equivalents. Cardiotoxic side-effects are not limited to pharmaceutical compounds, as cardiotoxicity has been observed with monoclonal antibody therapies and biologics. Therapeutic antibodies such as HER2/ERBB2 monoclonal antibodies and trastuzumab in association with paclitaxel treatment regimen have been shown to have a synergistic negative impact on adult cardiomyocytes. (Pentassuglia et al., 2007, Experimental Cell Research, 313: 1588-1601). Detrimental effects of biologics on cardiac safety are prevalent independent of combined therapies: for example, eleven percent of patients on trastuzumab develop cardiac toxicity (Guarneri et al., 2006, Journal of Clinical Oncology, 24: 4107-4115).
There remains a need in this art for in vitro methods for reliably determining cardiotoxicity of pharmaceuticals, biologics, and other chemical compounds and environmental agents.
The present invention provides reagents and methods for identifying a plurality of low molecular weight molecules, preferably secreted by cardiomyocytes or hESC-derived or human iPS-derived cardiac-specific cells, in response to pharmaceuticals, biologics, and other chemical compounds or environmental agents. In addition, the invention provides reagents and methods for identifying, in certain embodiments, particular metabolites produced by cardiomyocytes in response to a pharmaceutical, biologic, other chemical compound or environmental agent, as well as, in other embodiments, pluralities of cellular metabolites produced by cardiomyocytes in response to a pharmaceutical, biologic, other chemical compound or environmental agent, thereby also providing metabolic profiles of specific metabolites produced, for example, as the result of cardiotoxicity and that are secreted in response to exposure to particular pharmaceuticals, biologics, and other chemical compounds and environmental agents. The present invention thus provides reagents and methods for predicting cardiotoxic effects of pharmaceuticals, biologics, and other chemical compounds and environmental agents using profiles of low molecular weight metabolites identified via metabolomic analysis of human cardiomyocytes contacted with such agents in vitro.
Low molecular weight metabolites can be sensitively detected in even low quantities by methods and technologies known in the art, including most particularly variations of liquid chromatography high resolution mass spectrometry (LC-MS) and/or electrospray ionization time of flight mass spectrometry (ESI-TOF). As disclosed herein the sensitivity of applying such methods to detecting metabolites produced by cardiomyocytes in response to pharmaceuticals, biologics, and other chemical compounds and environmental agents. provides improved outcomes for detecting cardiotoxicity compared with less robust methods known in the art. Advantages of the inventive methods disclosed herein include that they provide direct products of the cardiotoxic response—metabolites produced by cardiomyocytes in response to insults from pharmaceuticals, environmental agents, chemical compounds and/or biologic therapies. The invention disclosed herein also advantageously provides metabolite profiles produced by contacting cardiomyocytes in vitro with specific pharmaceuticals, biologics, and other chemical compounds and environmental agents. These profiles are comprised of non-limiting collections of candidate biomarkers, providing a biochemical metabolic signature indicative of cardiotoxicity.
In particular embodiments, the invention provides reagents and methods for in vitro screening using cardiomyocytes to detect metabolites associated with cardiotoxicity of specific pharmaceuticals, biologics, and other chemical compounds and environmental agents. The patterns and collections of metabolite biomarkers establish that such cardiomyocytes have a characteristic metabolic response to cardiotoxicity produced by contact with specific pharmaceuticals, biologics, and other chemical compounds and environmental agents.
Practice of the provided methods illustrates the cardiomyocyte metabolome includes potential human biomarkers for disease and cardiotoxic response. These biomarkers are identified by contacting cardiomyocytes with specific pharmaceuticals, environmental agents, chemical compounds and biologic therapies. The results set forth herein demonstrate that exposure of cardiomyocytes to known cardiotoxic drugs induced significant changes in different metabolic pathways, consistent with known activity as cardiotoxins, and further providing an exemplar for the practice of the inventive methods with uncharacterized pharmaceuticals, biologics, and other chemical compounds and environmental agents to determine the extent of any cardiotoxicity exhibited by these compounds.
Specific embodiments of this invention will become evident from the following more detailed description of certain preferred embodiments and the claims.
These and other objects and features of this invention will be better understood from the following detailed description taken in conjunction with the drawings wherein:
This invention is more particularly described below and the Examples set forth herein are intended as illustrative only, since numerous modifications and variations therein will be apparent to those skilled in the art.
As used in the description herein and throughout the claims that follow, the meaning of “a”, “an”, and “the” includes plural reference unless the context clearly dictates otherwise. The terms used in the specification generally have their ordinary meanings in the art, within the context of the invention, and in the specific context where each term is used. Some terms have been more specifically defined below to provide additional guidance to the practitioner regarding the description of the invention. In particular, the term “cell” as used herein can be singular or plural, but in a preferred embodiment is plural.
In one embodiment, the invention includes reagents and methods for determining the cellular and/or biochemical effects of exposure to cardiotoxic compounds. The term “cellular metabolite” or the plural form, “cellular metabolites,” as used herein refers to a low molecular weight molecule secreted by a cell. In general the size of the metabolites is in the range of about 55 Daltons to about 1500 Daltons. A cellular metabolite may include but is not limited to the following types of low molecular weight molecules: acids, bases, lipids, sugars, glycosides, amines, organic acids, lipids, amino acids, oximes, esters, dipeptides, tripeptides, fatty acids, cholesterols, oxysterols, glycerols, steroids, and/or hormones. In an alternative embodiment, the cellular metabolite is secreted from cardiomyocytes, human embryonic stem cell (hESC)-derived cardiomyocytes or human induced pluripotent stem cell (iPS)-derived cardiomyocytes. In a preferred embodiment, the cellular metabolites include but are not limited to the following low molecular weight molecules: Triethylamine; NN-Diethylamine; Hexylamine; p-Glucosyloxymandelonitrile; (s)-4-Hydroxymandelonitrilebeta-D-glucoside; 13,14-dihydro PGE1 (Prostaglandin E1); 7-Ketocholesterol; 1,25-Dihydroxyvitamin D3-26,23-lactone; Formononetin 7-O-glucoside-6″-O-malonate; Isochlorogenic acid b; 13-Dicaffeoylquinic acid; 3-Hexaprenyl-4-hydroxy-5-methoxybenzoic acid; 2-Phenylglycine; (E)-4-Hydroxyphenylacetaldehyde-oxime; (Z)-4-Hydroxyphenylacetaldehyde-oxime; Betaine; 2-Ethylhexyl-4-hydroxybenzoate; Glycerophosphocholine; N-Acetylgalactosamine; CGP52608; Biotin; DL-Homocystine; Ethenodeoxyadenosine; Queuine; N-Acetylaspartylglutamic acid; Tetrahydrocortisone; Cyclic Phosphatidic acid; 2-Methoxyestrone3-glucuronide; Diacylglycerol; Quercetin3-(2G-xylosylrutinoside); Niacinamide; Aspartic Acid; Iminodiacetate; Erythritol; D-Threitol; N-Acetylserine; L-Glutamic acid; L-4-Hydroxyglutamate semialdehyde; 2-Oxo-4-hydroxy-5-aminovalerate; O-Acetylserine; DL-Glutamate; DL-Glutaminic acid; 2-Aminoglutaric acid; Glutamate; D-Glutamic acid; 3-Pyridinebutanoic acid; Norsalsolinol; D-Phenylalanine; D-alpha-Amino-beta-phenylpropionic acid; L-Phenylalanine; 3-Methylhistidine; 1-Methylhistidine; (R)—N-Methylsalsolinol; (S)—N-Methylsalsolinol; Symmetric dimethylarginine; or Asymmetric dimethylarginine.
The phrase “identifying one or a plurality of cellular metabolites . . . differentially produced” as used herein includes but is not limited to comparisons of cells exposed to a test compound to untreated (i.e., control) cells. Detection or measurement of variations in low molecular weight molecule populations secreted by a cell, between experimental and control cells are included in this definition. As used herein, the terms “secrete,” “secreting,” and “secretion” are intended to encompass any cellular process by which a cellular metabolite produced by a cell is translocated outside the cell. Metabolites or small molecules, particularly those species secreted, excreted or consumed by the cells, or those metabolites that are fluxed through the cells, that participate in functional mechanisms of cellular response to pathological or chemical insult. Metabolites may also be produced as a result of apoptosis or necrosis.
In a preferred embodiment, alterations in cells or cell activity are measured by determining a profile of changes in low molecular weight molecules in treated versus untreated cells. Also included are comparisons between cells treated with different amounts, types or concentrations, durations or intensities of cardiotoxic or potential cardiotoxic compounds.
Alterations in cellular metabolites such as sugars, organic acids, amino acids, fatty acids, and low molecular weight compounds are measured and used to assess the effects of specific pharmaceuticals, environmental agents, chemical compounds and biologic therapies on biochemical pathways in cardiomyocytes. The screened low molecular weight compounds (i.e., metabolites) are secreted in response to a variety of biological activities, including, but not limited to inflammation, anti-inflammation, vasodilation, neuroprotection, fatty acid metabolism, collagen matrix degradation, oxidative stress, antioxidant activity, DNA replication and cell cycle control, methylation, biosynthesis of nucleotides, carbohydrates, amino acids and lipids, among others. Secreted low molecular weight molecules are precursors, intermediates and/or end products of in vivo biochemical reactions. Alterations in specific subsets of molecules correspond to a particular biochemical pathway and thus reveal the biochemical effects of cardiotoxicity.
The term “cardiomyocyte” or “cardiomyocyte cell(s)” as described herein refers to primary cardiomyocytes, cardiomyocyte precursor cells, clonal cardiomyocytes derived from adult human heart, immortalized cardiomyocytes, human embryonic stem cell (hESC)-derived cardiomyocytes, human induced pluripotent stem cell (iPS)-derived cardiomyocytes, or any cell displaying cardiomyocyte-specific markers such that a pathologist, scientist, or laboratory technician would recognize the cell to be cardiomyocyte-specific or cardiomyocyte derived.
The term “cardiotoxic” as described herein refers to a substance or treatment, particularly pharmaceuticals, biologics, and other chemical compounds and environmental agents, that induce cardiomyopathy, heart disease, and/or abnormal heart pathology and physiology. Examples of cardiotoxicities encompassed by the definition of the term as used herein include heart abnormalities that would be recognized by a physician, cardiologist, or medical researcher, which could be attributed to or a potential result of a drug-treatment regimen.
In a preferred embodiment the term “compound” or “test compound” includes but is not limited to pharmaceuticals, environmental agents, chemical compounds and biologic therapies, including antibody-based treatments, vaccines, or recombinant proteins and enzymes. In a particularly preferred embodiment, cardiotoxic compounds include tamoxifen, doxorubicin, and paclitaxel. In a further embodiment, potentially cardiotoxic compounds are screened for metabolite similarities to already known cardiotoxic compounds.
The term “cardiomyopathy” refers to heart disease, including but not limited to inflammation of the heart muscle and reduction of heart function. Cardiomyopathy can be classified as primary or secondary and may further include dilated, hypertrophic and restrictive cardiomyopathies. The heart cavity can be enlarged and stretched (e.g., cardiac dilation), and may not pump normally. Abnormal heart rhythms called arrhythmias and disturbances in the heart's electrical conduction also can occur. In this condition, the muscle mass of the left ventricle enlarges or “hypertrophies.”
Mass spectrometry-based platforms have been proposed as a means to select peptides and proteins, but not small-molecule metabolites, as candidate biomarkers of cardiotoxicity. For example, brain natriuretic peptide (BNP) and N-terminal proBNP (NTproBNP) are clinical biomarkers of heart failure. BNP hormone and the inactive NTproBNP are predominantly secreted in the ventricles of the heart in response to pressure overload and, consequently, are being investigated as markers of drug-induced cardiac hypertrophy in rat. (See Berna et al., 2008, Anal Chem 80: 561-566). In addition, myosin light chain 1 (Myl3), a 23-kDa isoform of one of the subunits of myosin and troponin have been proposed as biomarkers of cardiac necrosis to predict drug-induced cardiotoxicity (See Adamcova et al., 2005, Expert Opinion on Drug Safety 4(3): 457-472). Such peptides and proteins have been recognized in the art as products of the degenerative changes in heart muscle associated with cardiomyopathies.
Certain of the compounds used herein to demonstrate the usefulness of a metabolomics approach for identifying candidate biomarkers for cardiotoxicity in cardiomyocytes are known cardiotoxic compounds. These compounds are thus illustrative of the reagents and methods for detecting metabolomic markers for cardiotoxicity, and include doxorubicin, paclitaxel and tamoxifen. The assessment of low molecular weight molecule metabolic products secreted by cardiomyocytes in response to exposure to multiple drug-treatment regimens thus provides novel profiles of candidate biomarkers of cardiotoxicity that can be rationalized with these clinical indicia.
The term “control cell(s)” as used herein refers in general to non-cardiac derived cell types. In a preferred embodiment, control cells include human fibroblasts.
The term “control cardiomyocytes” as used herein refers to cardiomyocyte or cardiomyocyte-derived cells that are exposed to control conditions.
The term “control sets” as used herein refers to the exposure of a particular cell type to a condition that one of skill in the art would recognize as a control treatment. In a preferred embodiment this includes but is not limited to the following experimental conditions: the exposure of cardiac cells to non-toxic compounds, or the exposure of non-cardiac cells to cardiotoxic compounds. Conversely, as used herein, an “experimental set” includes cardiac-specific cells exposed to a compound of interest (e.g., test compound), such as specific pharmaceuticals, biologics, and other chemical compounds and environmental agents.
The term “subtracting” as used herein refers to the identification of common cellular metabolites secreted by experimental cells and control cells followed by the selective removal of those metabolites in common from a metabolic signature or biomarker profile of specific cardiotoxic response.
When identifying low molecular weight metabolites that are secreted by cardiomyocytes, a skilled technician or scientist would understand that such metabolites can be measured, for example, those metabolites secreted and/or released into cellular supernatant and/or present in cellular extracts, as well as a variety of other methods available for the assessment of secreted molecules. Identified metabolites may also be waste products excreted by cells.
The phrase “exposure to test compound” may refer to cell samples exposed to an individual compound separately or a plurality of compounds sequentially and/or collectively. In one embodiment, cells are exposed to an individual test compound. In a further embodiment, cells are exposed to multiple compounds. In an alternative embodiment, cells are not exposed to any compound (i.e., control). Cells may be cultured in the presence or absence of test compounds.
The phrase “selecting those with commonality” as used herein refers to secreted metabolites produced in commonality across more than one set of cells. Thus, for example, the metabolites in various cell sets are identified, compared, and those in common may be further selected for commonality.
The term “physical separation method” as used herein refers to any method known to those with skill in the art sufficient to produce a profile of changes and differences in low molecular weight molecules produced by cells exposed to pharmaceuticals, environmental agents, chemical compounds and biologic therapies according to the methods of this invention. In certain embodiments, physical separation methods permit detection of low molecular weight molecules including but not limited to acids, bases, lipids, sugars, glycosides, amines, organic acids, lipids, amino acids, oximes, esters, dipeptides, tripeptides, fatty acids, cholesterols, oxysterols, glycerols, steroids, and/or hormones. In particular embodiments, this analysis is performed by liquid chromatography high resolution mass spectrometry (LC-MS) and/or liquid chromatography/electrospray ionization time of flight mass spectrometry (LC-ESI-TOF-MS), however it will be understood that low molecular weight compounds as set forth herein can be detected using alternative spectrometry methods or other methods known in the art. For example, nuclear magnetic resonance (NMR) is another method that can identify low molecular weight compounds of the invention. Similar analyses have been applied to other biological systems in the art (Want et al., 2005, Chem Bio Chem 6:1941-51), providing biomarkers of disease or toxic responses that can be detected in biological fluids (Sabatine et al., 2005, Circulation 112: 3868-875). It is understood that different instruments may detect different low molecular weight compounds. Thus, for example, the profile developed by LC-MS and/or LC-ESI-TOF-MS may be the same as or different than the profile developed by NMR.
A “biological sample” includes but is not limited to cells cultured in vitro, a patient sample, or biopsied cells dispersed and cultured in vitro. A “patient” may be a human or animal. A “patient sample” includes but is not limited to blood, plasma, serum, lymph, urine, cerebrospinal fluid, saliva or any other biofluid or waste.
The term “biomarker” as used herein refers, inter alia to low molecular weight compounds as set forth herein that exhibit significant alterations between experimental cell sets and control cell sets, particularly with regard to exposure to cardiotoxic compounds. In certain embodiments, biomarkers are identified as set forth above, by methods including, for example, LC-MS and/or LC-ESI-TOF-MS. In certain embodiments, the following low molecular weight molecules are provided herein, taken alone or in any informative combination, as biomarkers of cardiotoxicity: Triethylamine; NN-Diethylamine; Hexylamine; p-Glucosyloxymandelonitrile; (s)-4-Hydroxymandelonitrilebeta-D-glucoside; 13,14-dihydro PGE1 (Prostaglandin E1); 7-Ketocholesterol; 1,25-Dihydroxyvitamin D3-26,23-lactone; Formononetin 7-O-glucoside-6″7-O-malonate; Isochlorogenic acid b; 13-Dicaffeoylquinic acid; 3-Hexaprenyl-4-hydroxy-5-methoxybenzoic acid; 2-Phenylglycine; (E)-4-Hydroxyphenylacetaldehyde-oxime; (Z)-4-Hydroxyphenylacetaldehyde-oxime; Betaine; 2-Ethylhexyl-4-hydroxybenzoate; Glycerophosphocholine; N-Acetylgalactosamine; CGP52608; Biotin; DL-Homocystine; Ethenodeoxyadenosine; Queuine; N-Acetylaspartylglutamic acid; Tetrahydrocortisone; Cyclic Phosphatidic acid; 2-Methoxyestrone3-glucuronide; Diacylglycerol; Quercetin3-(2G-xylosylrutinoside); Niacinamide; Aspartic Acid; Iminodiacetate; Erythritol; D-Threitol; N-Acetylserine; L-Glutamic acid; L-4-Hydroxyglutamate semialdehyde; 2-Oxo-4-hydroxy-5-aminovalerate; O-Acetylserine; DL-Glutamate; DL-Glutaminic acid; 2-Aminoglutaric acid; Glutamate; D-Glutamic acid; 3-Pyridinebutanoic acid; Norsalsolinol; D-Phenylalanine; D-alpha-Amino-beta-phenylpropionic acid; L-Phenylalanine; 3-Methylhistidine; 1-Methylhistidine; (R)—N-Methylsalsolinol; (S)—N-Methylsalsolinol; Symmetric dimethylarginine; or Asymmetric dimethylarginine. In a preferred embodiment, the low molecular weight molecules described herein in Tables 2A-2D taken alone or in any informative combination, are reliable biomarkers of cardiotoxicity. Many of the identified low molecular weight molecules are identified by unique mass feature size or neutral mass, however some molecules are further identified by compound name.
The terms “metabolic signature” and “metabolic profile” as used herein refer to one or a plurality of metabolites identified by the inventive methods. Metabolic signatures and profiles according to the invention can provide a molecular “fingerprint” of the effects of cardiotoxicity and identify low molecular weight compounds significantly altered following exposure to pharmaceuticals, environmental agents, chemical compounds and biologic therapies that are cardiotoxic. In certain embodiments, metabolic signatures or metabolic profiles can be used to predict cardiotoxicity of a compound. In an alternative embodiment, a metabolic signature or profile may diagnose cardiotoxic effects from drug treatment regimens, pharmaceuticals, environmental agents, chemical compounds or biologic therapies.
In certain embodiments, cardiotoxicity of a test compound can be identified by cardiomyocyte secretion of a single known cardiotoxic biomarker. As an example, a single marker may include Betaine or Glycerophosphocholine. This may include metabolite(s) secreted in response to exposure to a single established cardiotoxic compound (e.g. doxorubicin). In other embodiments, cardiotoxicity is affirmed by detection of a metabolic signature (i.e., one or a plurality of low molecular weight metabolites) commonly produced by cardiomyocytes in response to two or more known cardiotoxic compounds (e.g., doxorubicin and paclitaxel, or doxorubicin, paclitaxel, and tamoxifen). In further embodiments, metabolic signatures of cardiotoxicity comprising one or a plurality of cellular metabolites provided in Tables 2A-2D, or described in the chromatograms of
Data for statistical analysis were extracted from chromatograms using the Agilent Mass Hunter software (Product No. G3297AA, Agilent Technologies, Inc., Santa Clara, Calif.); it will be understood that alternative statistical analysis methods can be used. Masses were binned together if they were within 10 ppm and eluted within a 2 minutes retention time window. A binned mass was considered to be the same molecule across different LC-ESI-TOF-MS analyses (referred to herein as an “exact mass,” which will be understood to be ±10 ppm). Binning of the data is required for statistical analysis and comparison of masses across the entire experiment. If multiple peaks with the same mass at the same retention time within a single sample were detected by Mass Hunter, they were averaged to assist data analysis. Masses lacking a natural isotopic distribution or with an absolute height of less than 1000 were removed from the data prior to analysis. It would be understood that the results from this assay provide relative values that are assessed according to annotated values within 10 ppm to provide an identity for the molecular weight detected. Thus, a mass shift within 10 ppm is considered consistent with determining the identity of a specific cellular metabolite previously annotated due to differences in ionization source and instrumentation, e.g. between different experiments or using different instruments.
As used herein, a mass was considered to be the same across LC/ESI-TOF-MS runs using a simple algorithm that first sorts the data by mass and retention time. After sorting, a compound was considered unique if it had an ordered retention time difference of less than or equal to 0.1 minutes and a mass difference less than or equal the weighted formula: consecutive masses did not differ by 10 ppm if under 175 Da, by 7 ppm over the range 175 to 300 Da, and by 5 ppm when greater 300 Da. If a series of measurements fit this definition it was considered to be from the same compound. If either the mass or the retention time varied by more than the limits listed above it was considered to be a different compound and given a new unique designation.
The data from the most reproducible mass features was log base 2 transformed and median centered prior to statistical analysis. Statistical analysis was performed using the open source statistical programming and analysis software R. Statistical significance of individual mass features were performed under the null hypothesis that no difference in abundance exists between control and drug treatment using a permutation based test statistic or a Welch T-test. To test the null hypothesis a one way permutation based t-test assuming a normal approximation of the conditional distribution was used and implement using the Conditional Inference Procedures in a Permutation Test Framework (Coin) library, a contributed package of programming code. Statistics tests were performed without replacement of missing values decrease the degrees of freedom rather than imputing missing values. This oneway test method is ideally suited for analysis of complex data sets where one may not be able to assume that every feature tested will have a normal distribution (Hothorn et al., 2006, Amer. Statistician, 60:257-263). False discovery rates (FDR) were controlled using the Q value estimator (Storey et al., 2003, Proc Natl Acad. Sci., 100:9440-5) and implemented using the qvalue library in R (Dabney et al., 2003, qvalue: Q-value estimation for false discovery rate control. R package version 1.10.0., www.CRAN.R-project.org).
In certain embodiments, a cardiotoxic biomarker may reference one or a collection of cellular metabolites produced by cardiomyocytes following exposure to known cardiotoxins. A cardiotoxic metabolic signature can comprise about 1, or about 6, or about 10, or about 20, or about 30 differentially secreted low molecular weight molecules, and while the cardiotoxic signature as disclosed herein comprises from about 1 to about 30 metabolites and includes the low molecular weight molecules set forth in Table 2A-2D herein, said cardiotoxic signature generally comprises a sufficient number of metabolites to independently identify an experimental test compound as being cardiotoxic. It will be understood by those with skill in the art that the differential fold change in metabolite secretion between untreated and treated cells can vary for each metabolite.
The Examples which follow are illustrative of specific embodiments of the invention, and various uses thereof. They set forth for explanatory purposes only, and are not to be taken as limiting the invention.
Human cardiomyocytes, clonal cardiomyocytes derived from adult human heart (Celprogen 36044-15at, San Pedro, Calif.) or cardiac precursor cells, were treated with varying doses of pharmacological compounds known to have cardiotoxic effects. Cardiomyocytes were treated with doxorubicin and paclitaxel, which are strong toxicants, as well as tamoxifen, a weak toxicant, for 24 or 48 hours. Some combinatorial treatments regimens appeared to exhibit synergistic cardiotoxic effects (e.g., for doxorubicin and trastuzumab combined therapies, see Pentassuglia et al., 2007, Experimental Cell Research 313: 1588-1601; for paclitaxel and doxorubicin combined therapies, see Robert, 2007, Cardiovasc Toxicol 7: 135-139)).
The cardiac origin of these cells was confirmed by immunohistochemistry using antibodies against cardiac alpha-actin protein (
In order to identify low molecular weight metabolites secreted by cardiomyocytes or cardiac precursors following exposure to cardiotoxic compounds, cells as described above in Example 1 were treated with doxorubicin, paclitaxel, and tamoxifen, for 24 or 48 hours.
The extracellular media from treated and untreated cells was processed as described in Cezar et al., (2007, Stem Cells Development 16: 869-882, this publication is incorporated by reference), for extraction of low molecular weight molecules (<3 kD) for metabolomics analysis. Extracellular low molecular weight molecule preparations were separated by liquid chromatography followed by electrospray ionization time of flight (LC-ESI-TOF-MS) mass spectrometry for ionization and detection of the full spectra of low molecular weight molecules present in each sample. More specifically, the samples were separated using the ESI_Luna_HILIC—95_t—06OACN—16 min method (HILIC chromatography). Statistical differences were inferred by subsequent bioinformatics and in silico mapping of deisotoped ESI-TOF-MS mass features as described below (also provided in Cezar et al. (2007, id.)).
Briefly, ionization (100 m/z-1500 m/z) was acquired on an Agilent 6520 Accurate-Mass Q-TOF in extended dynamic range and positive mode. Mass features were generated using two independent methods. First MassHunter Qualitative Analysis was used to generate mass features using the Molecular Feature Extraction algorithm (MFE). Features generated by MFE were binned in R and analyzed for differential accumulation in response to the drug treatments. The Agilent data files were also converted to mzData file format using Agilent's MassHunter Qualitative Analysis Workstation. The mzData files were analyzed in R using the software library XCMS to find mass feature bins differentially present in the presence of drug. MHD files created by MFE were converted to text files using MassHunterMFE version 44. The MHD text files were loaded into R and meta data corresponding to the file name, cell line (Celprogen Cardiomyocytes or solvent), plate (0, 1, 2 or 3), well (solvent, A, B, or C), experiment replication, cells (supernatant, uncultured media or solvent), cell culture passage number, drug treatment (15 uM tamoxifen, 15 μM paclitaxel, control, 26 μM doxorubicin), feature retention time group, retention time, feature neutral mass, mass feature mass standard deviation, abundance, saturation, height, number of ions in feature, min charge, max charge, charge number, width, and group feature count were added to each file.
In order to identify metabolites secreted by cardiomyocytes in response to cytotoxic drug treatment, metabolomic analysis was performed on cardiomyocytes from similar cell passages. Statistically-significant features that were common between the cytotoxic drug treatments were identified. Mass features that were present in at least 25% of LC-MS samples of control and drug treated cardiomyocytes were selected. The statistical significance of individual mass features was determined under the null hypothesis that no difference in abundance existed between control and drug treatment using a permutation-based test statistic like Students t-test. A one-way test assuming a normal approximation of the conditional distribution (see Horthon et al., 2006a, The American Statistician 60(3): 257-263) was used to test the null hypothesis and was implemented using the Conditional Inference Procedures in a Permutation Test Framework (Coin) library in R (see Horthon et al., 2006b, Conditional Inference Procedures in a Permutation Test Framework, R package version 0.4-5, CRAN.R-project.org). Statistics tests were performed on log base two transformed, median normalized abundance values without replacement of missing values reducing the degrees of freedom when a missing value was present. False discovery rates (FDR) were controlled using the Q value estimator (Storey et al., 2003, Proc. Natl. Acad. Sci. USA 100: 9440-45) with a lambda of 0 and implemented using the q value library in R (Dabney et al., 2003, qvalue: Q-value estimation for false discovery rate control. R package versions 1.10; CRAN.R-project.org; R Development Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing: 2007. ISBN 3-900051-07-0; www.R-project.org). After performing statistics, a universe of statistically-significant mass features was created from the comparisons of control to each drug treatment based on FDR-adjusted p values. Boolean logic was utilized to find the statistically significant features in common between the different drug treatments. An intersection of mass features that exhibited statistically significant differences in the drugs affecting cell viability (DOX, PAC), but that were not statistically significant in (TAM) was selected. This intersection represented common mass features that were associated with cardiotoxicity because they exhibited a statistically-significant change in cytotoxic treatments, but no statistically-significant change in non-cytotoxic treatments.
A mass was considered to be the same across LC/ESI-MS runs using a simple algorithm that sorts the data by mass and retention time as performed by the software and methods described above. The criteria used for treated-cells were based on a sliding mass scale to compensate for detector efficiency. Because of flow rate, a mass was considered equivalent if it was within (0.00001×mass) when under 175 Da, (0.000007×mass) when 176 Da-300Da, and (0.000005×mass) when over 300 Da with a retention time difference of 1.5 min. If a series of measurements fit this definition, it was considered to be from the same compound within each experiment. If either the mass retention time varied by more than the limits listed above, the compound was considered to be a different one and given a different bin description. Specifically, 774,645 features were identified by the MassHunter software with an average of 6455 and a median of 5869 features per LC/MS run. The mass features were then sorted by mass and retention time groupings and feature ID bins were created for each set of mass and retention groupings that did not differ.
The neutral exact mass and/or empirical chemical formula of each compound, detected by LC-ESI-TOF-MS, was queried in public searchable databases, METLIN (metlin.crips.edu), The Human Metabolome Database (hmda.ca), Kyoto Encyclopedia of Genes and Genomes (genome.jp/keg), and the Biological Magnetic Resonance Bank (bmrb.wisc.edu/metabolomics) for candidate identities. LC-MS-measured mass signals matched small molecules present in the databases if their exact masses were within 10 parts per million (0.00001×mass). Exact mass measurements and chemical formulae are generally nonambiguous for small molecules up to a certain size. Analytical-grade chemical standards were purchased from Sigma for comparative LC-MS. Aliquots of conditioned medium used in experiments were spiked with 1 mM chemical standards followed by standard LC-ESI-TOF-MS, as described above. The neutral exact masses and retention times for standard compounds in spiked conditioned medium were used to re-extract peaks in experimental samples using Analyst software (Agilent).
The doses for each compound were based upon published standards, equivalent to therapeutic circulating levels whenever possible (see Table 1). Trypan Blue exclusion/cell death assays using the aforementioned concentrations have shown that these doses corroborate published findings, whereby doxorubicin and paclitaxel induced significantly higher cell death as described in Example 1. (
Metabolite trends observed in initial studies are shown in
Identified features are provided in Table 2A-2D. Specifically, Table 2A provides identified mass features with commonality between paclitaxel and doxorubicin treatments. Table 2B provides identified mass features with commonality between paclitaxel, doxorubicin, and tamoxifen treatments. Table 2D provides identified mass features secreted from cardiac precursor cells treated with doxorubicin and then paclitaxel.
Statistically-significant changes in metabolite secretion can be examined for novel or non-annotated low molecular weight molecules, using the approach reported previously (Cezar et al, 2007, Stem Cells and Development 16: 869-882). Initial experiments have shown that a subset of human metabolites are indeed statistically-significantly altered in response to pharmaceuticals that are strong inducers of cardiomyopathies, namely doxorubicin and paclitaxel, in comparison to weak/moderate inducers such as tamoxifen. (
Data were accrued from n=107 mass spectrometry injections following exposure of human cardiomyocytes (Celprogen 36044-15at, San Pedro, Calif.) to three experimental treatments with different degrees of cardiotoxicity: (1.) doxorubicin; (2.) pacitaxel; and (3.) tamoxifen (weak toxicant). Following statistical analysis, with False Discovery Rates (FDR 0.05) adjustments, 187 significant features (e.g., candidate biomarkers), were identified in response to doxorubicin, 185 significant features in response to paclitaxel and 148 significant features in response to tamoxifen. Seventy-three statistically significant features were found to be in common to the strong cardiotoxicants doxorubicin and paclitaxel as described in the Preferred Embodiments. (
The putative annotation of the exact neutral masses of such features in chemical databases revealed that several candidate biomarkers map onto energy metabolism pathways, such as NADPH2: oxygen oxidoreductase activity, UDPglucuronate beta-D-glucuronosyltransferase, glycolysis, gluconeogenesis as well as oxidative stress. These results are consistent with published reports on the mechanisms of cardiotoxicity for these particular compounds.
Strong robustness and high reproducibility of low molecular weight molecules identified following exposure of human cardiomyocytes to the three established cardiotoxins: paclitaxel, doxorubicin, and tamoxifen was observed. The identification of metabolites secreted by cardiomyocytes in response to two or three cardiotoxins permitted enrichment for candidate biomarkers and provided a metabolic signature of cardiotoxicity.
In addition, the claimed invention is not intended to be limited to the disclosed embodiments. It should be understood that the foregoing disclosure emphasizes certain specific embodiments of the invention and that all modifications of alternatives equivalent thereto are within the spirit and scope of the invention as set forth in the appended claims.
This application claims the priority benefit of U.S. provisional patent application Ser. No. 61/249,150 filed Oct. 6, 2009, the entirety of which is herein incorporated by reference.
| Number | Date | Country | |
|---|---|---|---|
| 61249150 | Oct 2009 | US |
