Patent Application
20160305928
The present invention relates to small molecules or metabolites that are found to have significantly different abundances or intensities between clinically diagnosed ovarian cancer-positive patients and normal disease-free subjects. The present invention also relates to methods for diagnosing ovarian cancer, or the risk of developing ovarian cancer.
Ovarian cancer is the fifth leading cause of cancer death among women (1). It has been estimated that over 22,000 new cases of ovarian cancer will be diagnosed this year, with 16,210 deaths predicted in the United States alone (2). Ovarian cancer is typically not identified until the patient has reached stage III or IV, which is associated with a poor prognosis; the five-year survival rate is estimated at around 25-30% (3). The current screening procedures for ovarian cancer involve the combination of bimanual pelvic examination, transvaginal ultrasonography, and serum screening for elevated cancer antigen-125 (CA125), a protein cancer antigen (2). The efficacy of CA125 screening for ovarian cancer is currently of unknown benefit, as there is a lack of evidence that the screen reduces mortality rates, and it is under scrutiny due to the risks associated with false positive results (1, 4). According to the American Cancer Society, CA125 measurement and transvaginal ultrasonography are not reliable screening or diagnostic tests for ovarian cancer, and that the only current method available to make a definite diagnosis is by surgery.
CA125 is a high molecular weight mucin that has been found to be elevated in most ovarian cancer cells as compared to normal cells (2). A CA125 test result that is higher than 30-35 U/ml is typically accepted as being at an elevated level (2). There have been difficulties in establishing the accuracy, sensitivity, and specificity of the CA125 screen for ovarian cancer due to the different thresholds used to define elevated CA125, varying sizes of patient groups tested, and broad ranges in the age and ethnicity of patients (1). According to the Johns Hopkins University pathology website, the CA125 test only returns a true positive result for ovarian cancer in roughly 50% of stage I patients and about 80% in stage II, III and IV patients. Endometriosis, benign ovarian cysts, pelvic inflammatory disease, and even the first trimester of a pregnancy have all been reported to increase the serum levels of CA125 (4). The National Institute of Health's website states that CA125 is not an effective general screening test for ovarian cancer. They report that only about three out of 100 healthy women with elevated CA125 levels are actually found to have ovarian cancer, and about 20% of ovarian cancer diagnosed patients actually have elevated CA125 levels.
It is clear that there is a need for improving ovarian cancer detection. A test that is able to detect risk for, or the presence of, ovarian cancer or that can predict aggressive disease with high specificity and sensitivity would be very beneficial and would impact ovarian cancer morbidity.
The present invention relates to small molecules or metabolites that are found to have significantly different abundances between persons with ovarian cancer, and normal subjects.
The present invention provides a method for identifying, validating, and implementing a high-throughput screening (HTS) assay for the diagnosis of a health-state indicative of ovarian cancer or at risk of developing ovarian cancer. In a particular example, the method encompasses the analysis of ovarian cancer-positive and normal biological samples using non-targeted Fourier transform ion cyclotron mass spectrometry (FTMS) technology to identify all statistically significant metabolite features that differ between normal and ovarian cancer-positive biological samples, followed by the selection of the optimal feature subset using multivariate statistics, and characterization of the feature set using methods including, but not limited to, chromatographic separation, mass spectrometry (MS/MS), and nuclear magnetic resonance (NMR), for the purposes of:
The present invention further provides a method for the diagnosis of ovarian cancer or the risk of developing ovarian cancer in humans by measuring the levels of specific small molecules present in a sample and comparing them to “normal” reference levels. The methods measure the intensities of specific small molecules, also referred to as metabolites, in the sample from the patient, and compare these intensities to the intensities observed in a population of healthy individuals. The sample obtained from the human may be a blood sample.
The present invention may significantly improve the ability to detect ovarian cancer or the risk of developing ovarian cancer, and may therefore save lives. The statistical performance of a test based on these samples suggests that the test will outperform the CA125 test, the only other serum-based diagnostic test for ovarian cancer. Alternatively, a combination of the test described herein and the CA125 test may improve the overall diagnostic performance of each test. The methods of the present invention, including development of HTS assays, can be used for the following, wherein the specific “health-state” refers to, but is not limited to, ovarian cancer:
1. Identifying small-molecule metabolite biomarkers which can discriminate between ovarian cancer-positive and ovarian cancer-negative individuals using any biological sample taken from the individual;
2. Specifically diagnosing ovarian cancer using metabolites identified in a sample such as serum, plasma, whole blood, and/or other tissue biopsy as described herein;
3. Selecting a number of metabolite features from a larger subset required for optimal diagnostic assay performance statistics using various statistical methods such as those mentioned herein;
4. Identifying structural characteristics of biomarker metabolites selected from non-targeted metabolomic analysis using LC-MS/MS, MSn, and NMR;
5. Developing a high-throughput tandem MS method for assaying selected metabolite levels in a sample;
6. Diagnosing ovarian cancer, or the risk of developing ovarian cancer, by determining the levels of any combination of metabolite features disclosed from the FTMS analysis of patient sample, using any method including, but not limited to, mass spectrometry, NMR, UV detection, ELISA (enzyme-linked immunosorbant assay), chemical reaction, image analysis, or other;
7. Monitoring any therapeutic treatment of ovarian cancer, including drug (chemotherapy), radiation therapy, surgery, dietary, lifestyle effects, or other;
8. Longitudinal monitoring or screening of the general population for ovarian cancer using any single or combination of features disclosed in the method;
9. Determining or predicting the effect of treatment, including surgery, chemotherapy, radiotherapy, biological therapy, or other.
10. Determining or predicting tumor subtype, including disease stage and aggressiveness.
In one embodiment of the present invention there is provided a panel of metabolites that differ between the normal and the ovarian cancer-positive samples (p<0.05). Four hundred and twenty four metabolites met this criterion, as shown in Table 1. These metabolites differ statistically between the two populations and therefore have potential diagnostic utility. Therefore, one embodiment of the present invention is directed to the 424 metabolites, or a subpopulation thereof. A further embodiment of the present invention is directed to the use of the 424 metabolites, or a subpopulation thereof for diagnosing ovarian cancer, or the risk of developing ovarian cancer.
In a further embodiment of the present invention there is provided a number of metabolites that have statistically significant different abundances or intensities between ovarian cancer-positive and normal samples. Of the metabolite masses identified, any subpopulation thereof could be used to differentiate between ovarian cancer-positive and normal states. An example is provided in the present invention whereby a panel of 37 metabolite masses is further selected and shown to discriminate between ovarian cancer and control samples.
In this embodiment of the present invention, there is provided a panel of 37 metabolite masses that can be used as a diagnostic indicator of disease presence in serum samples. The 37 metabolites can include those with masses (measured in Daltons) 440.3532, 446.3413, 448.3565, 450.3735, 464.3531, 466.3659, 468.3848, 474.3736, 478.405, 484.3793, 490.3678, 492.3841, 494.3973, 502.4055, 504.4195, 510.3943, 512.4083, 518.3974, 520.4131, 522.4323, 530.437, 532.4507, 534.3913, 538.427, 540.4393, 548.4442, 550.4609, 558.4653, 566.4554, 574.4597, 576.4762, 578.493, 590.4597, 592.4728, 594.4857, 596.5015, 598.5121, where a +/−5 ppm difference would indicate the same metabolite. This embodiment of the present invention also includes the use of the 37 metabolites, or a subpopulation thereof for diagnosing ovarian cancer or the risk of developing ovarian cancer.
In a further embodiment of the present invention, there is provided a panel of 31 metabolite masses that can be used as a diagnostic indicator of disease presence in serum samples. The 31 metabolites can include those with masses (measured in Daltons) 446.3413, 448.3565, 450.3735, 468.3848, 474.3872, 476.5, 478.405, 484.3793, 490.3678, 492.3841, 494.3973, 496.4157, 502.4055, 504.4195, 512.4083, 518.3974, 520.4131, 522.4323, 530.437, 532.4507, 538.427, 540.4393, 550.4609, 558.4653, 574.4597, 576.4757, 578.4848, 592.357, 594.4848, 596.5012, 598.5121, where a +/−5 ppm difference would indicate the same metabolite. This embodiment of the present invention also includes the use of the 31 metabolites, or a subpopulation thereof for diagnosing ovarian cancer or the risk of developing ovarian cancer.
In a further embodiment of the present invention, there is provided a panel of 30 metabolite masses that can be used as a diagnostic indicator of disease presence in serum samples. The 30 metabolites can include those with masses (measured in Daltons) substantially equivalent to 446.3396, 448.3553, 450.3709, 468.3814, 474.3736, 478.4022, 484.3764, 490.3658, 492.3815, 494.3971, 496.4128, 502.4022, 504.4179, 512.4077, 518.3971, 520.4128, 522.8284, 530.43351, 532.44916, 538.4233, 540.4389, 550.4597, 558.4648, 574.4597, 576.4754, 578.4910, 592.47029, 594.4859, 596.5016, and 598.5172, where a +/−5 ppm difference would indicate the same metabolite. This embodiment of the present invention also includes the use of the 30 metabolites, or a subpopulation thereof for diagnosing ovarian cancer or the risk of developing ovarian cancer. In this embodiment of the invention the molecular formulas of the metabolites with these masses are C28H46O4, C28H48O4, C28H50O4, C28H52O5, C30H50O4, C30H54O4, C28H52O6, C30H50O5, C30H52O5, C30H54O5, C30H56O5, C32H54O4, C32H56O4, C30H56O6, C32H54O5, C32H56O5, C32H60O5, C34H58O4, C34H60O4, C32H58O6, C32H60O6, C34H62O5, C36H62O4, C36H62O5, C36H64O5, C36H66O5, C36H64O6, C36H66O6, C36H68O6, and C36H70O6, respectively and the proposed structures are as shown below:
respectively.
In a further embodiment of the present invention, there is provided a panel of six C28 carbon molecules (neutral masses 450 (C28H50O4), 446 (C28H46O4), 468 (C28H52O5), 448 (C28H48O4), 464 (C28H48O5) and 466 (C28H50O5)) that were found to be significantly lower in serum of the ovarian patients as compared to controls.
In one embodiment of the present invention there is provided a method for identifying metabolites to diagnose ovarian cancer comprising the steps of: introducing a sample from a patient presenting said disease state, with said sample containing a plurality of unidentified metabolites, into a high resolution mass spectrometer, for example, a Fourier Transform Ion Cyclotron Resonance Mass Spectrometer (FTMS); obtaining, identifying, and quantifying data for the metabolites; creating a database of said data; comparing said data from the sample with corresponding data from samples of a control population; identifying one or more metabolites that differ; and selecting the minimal number of metabolite markers needed for optimal diagnosis.
In a further embodiment of the present invention there is provided a method for identifying ovarian cancer-specific metabolic markers comprising the steps of: introducing a sample from a patient diagnosed for ovarian cancer, with said sample containing a plurality of unidentified metabolites, into a Fourier Transform Ion Cyclotron Resonance Mass Spectrometer (FTMS); obtaining, identifying, and quantifying data for the metabolites; creating a database of said data; comparing said data from the sample with corresponding data from a control sample; identifying one or more metabolites that differ, wherein the metabolites are selected from the group consisting of metabolites with accurate masses (measured in Daltons) of, or substantially equivalent to, 440.3532, 446.3413, 448.3565, 450.3735, 464.3531, 466.3659, 468.3848, 474.3736, 478.405, 484.3793, 490.3678, 492.3841, 494.3973, 502.4055, 504.4195, 510.3943, 512.4083, 518.3974, 520.4131, 522.4323, 530.437, 532.4507, 534.3913, 538.427, 540.4393, 548.4442, 550.4609, 558.4653, 566.4554, 574.4597, 576.4762, 578.493, 590.4597, 592.4728, 594.4857, 596.5015, 598.5121, where a +/−5 ppm difference would indicate the same metabolite.
In a further embodiment of the present invention there is provided a method for identifying ovarian cancer-specific metabolic markers comprising the steps of: introducing a sample from a patient diagnosed for ovarian cancer, with said sample containing a plurality of unidentified metabolites, into a Fourier Transform Ion Cyclotron Resonance Mass Spectrometer (FTMS); obtaining, identifying, and quantifying data for the metabolites; creating a database of said data; comparing said data from the sample with corresponding data from a control sample; identifying one or more metabolites that differ, wherein the metabolites are selected from the group consisting of metabolites with accurate masses (measured in Daltons) of, or substantially equivalent to, 446.3413, 448.3565, 450.3735, 468.3848, 474.3872, 476.5, 478.405, 484.3793, 490.3678, 492.3841, 494.3973, 496.4157, 502.4055, 504.4195, 512.4083, 518.3974, 520.4131, 522.4323, 530.437, 532.4507, 538.427, 540.4393, 550.4609, 558.4653, 574.4597, 576.4757, 578.4848, 592.357, 594.4848, 596.5012, 598.5121, where a +/−5 ppm difference would indicate the same metabolite.
In a further embodiment of the present invention there is provided a method for identifying ovarian cancer-specific metabolic markers comprising the steps of: introducing a sample from a patient diagnosed for ovarian cancer, with said sample containing a plurality of unidentified metabolites, into a Fourier Transform Ion Cyclotron Resonance Mass Spectrometer (FTMS); obtaining, identifying, and quantifying data for the metabolites; creating a database of said data; comparing said data from the sample with corresponding data from a control sample; identifying one or more metabolites that differ, wherein the metabolites are selected from the group consisting of metabolites with accurate masses (measured in Daltons) of, or substantially equivalent to 446.3396, 448.3553, 450.3709, 468.3814, 474.3736, 478.4022, 484.3764, 490.3658, 492.3815, 494.3971, 496.4128, 502.4022, 504.4179, 512.4077, 518.3971, 520.4128, 522.8284, 530.43351, 532.44916, 538.4233, 540.4389, 550.4597, 558.4648, 574.4597, 576.4754, 578.4910, 592.47029, 594.4859, 596.5016, and 598.5172, where a +/−5 ppm difference would indicate the same metabolite In this embodiment of the invention the molecular formulas of the metabolites with these masses are C28H46O4, C28H48O4, C28H50O4, C28H52O5, C30H50O4, C30H54O4, C28H52O6, C30H50O5, C30H52O5, C30H54O5, C30H56O5, C32H54O4, C32H56O4, C30H56O6, C32H54O5, C32H56O5, C32H60O5, C34H58O4, C34H60O4, C32H58O6, C32H60O6, C34H62O5, C36H62O4, C36H62O5, C36H64O5, C36H66O5, C36H64O6, C36H66O6, C36H68O6, and C36H70O6, respectively and the proposed structures are as shown below:
respectively.
In a further embodiment of the present invention there is provided a method for identifying ovarian cancer-specific metabolic markers comprising the steps of: introducing a sample from a patient diagnosed for ovarian cancer, with said sample containing a plurality of unidentified metabolites, into a Fourier Transform Ion Cyclotron Resonance Mass Spectrometer (FTMS); obtaining, identifying, and quantifying data for the metabolites; creating a database of said data; comparing said data from the sample with corresponding data from a control sample; identifying one or more metabolites that differ, wherein the metabolites are selected from the group consisting of metabolites with accurate masses of, or substantially equivalent to six C28 carbon molecules (neutral masses 450 (C28H50O4), 446 (C28H46O4), 468 (C28H52O5), 448 (C28H48O4), 464 (C28H48O5) and 466 (C28H50O5)).
In yet a further embodiment of the present invention there is provided an ovarian cancer-specific metabolic marker selected from the group consisting of metabolites with an accurate mass (measured in Daltons) of, or substantially equivalent to, 440.3532, 446.3413, 448.3565, 450.3735, 464.3531, 466.3659, 468.3848, 474.3736, 478.405, 484.3793, 490.3678, 492.3841, 494.3973, 502.4055, 504.4195, 510.3943, 512.4083, 518.3974, 520.4131, 522.4323, 530.437, 532.4507, 534.3913, 538.427, 540.4393, 548.4442, 550.4609, 558.4653, 566.4554, 574.4597, 576.4762, 578.493, 590.4597, 592.4728, 594.4857, 596.5015, 598.5121, where a +/−5 ppm difference would indicate the same metabolite.
In yet a further embodiment of the present invention there is provided an ovarian cancer-specific metabolic marker selected from the group consisting of metabolites with an accurate mass (measured in Daltons) of, or substantially equivalent to, 446.3413, 448.3565, 450.3735, 468.3848, 474.3872, 476.5, 478.405, 484.3793, 490.3678, 492.3841, 494.3973, 496.4157, 502.4055, 504.4195, 512.4083, 518.3974, 520.4131, 522.4323, 530.437, 532.4507, 538.427, 540.4393, 550.4609, 558.4653, 574.4597, 576.4757, 578.4848, 592.357, 594.4848, 596.5012, 598.5121, where a +/−5 ppm difference would indicate the same metabolite.
In yet a further embodiment of the present invention there is provided an ovarian cancer-specific metabolic marker selected from the group consisting of metabolites with an accurate mass (measured in Daltons) of, or substantially equivalent to 446.3396, 448.3553, 450.3709, 468.3814, 474.3736, 478.4022, 484.3764, 490.3658, 492.3815, 494.3971, 496.4128, 502.4022, 504.4179, 512.4077, 518.3971, 520.4128, 522.8284, 530.43351, 532.44916, 538.4233, 540.4389, 550.4597, 558.4648, 574.4597, 576.4754, 578.4910, 592.47029, 594.4859, 596.5016, and 598.5172, where a +/−5 ppm difference would indicate the same metabolite. In this embodiment of the invention the molecular formulas of the metabolites with these masses are C28H46O4, C28H48O4, C28H50O4, C28H52O5, C30H50O4, C30H54O4, C28H52O6, C30H50O5, C30H52O5, C30H54O5, C30H56O5, C32H54O4, C32H56O4, C30H56O6, C32H54O5, C32H56O5, C32H60O5, C34H58O4, C34H60O4, C32H58O6, C32H60O6, C34H62O5, C36H62O4, C36H62O5, C36H64O5, C36H66O5, C36H64O6, C36H66O6, C36H68O6, and C36H70O6, respectively and the proposed structures are as shown below:
respectively.
In a further embodiment of the present invention there is provided a method for diagnosing a patient for the presence of an ovarian cancer, or the risk of developing ovarian cancer, comprising the steps of: screening a sample from said patient for the presence or absence of one or more metabolic markers selected from the group consisting of metabolites with an accurate mass of, or substantially equivalent to, the masses in Table 1, where a +/−5 ppm difference would indicate the same metabolite; wherein the absence or significant reduction of one or more of said metabolic markers indicates the presence of an ovarian cancer, or the risk of developing ovarian cancer, and wherein the method is a FTMS based method.
In a further embodiment of the present invention there is provided a method for diagnosing a patient for the presence of an ovarian cancer, or the risk of developing ovarian cancer, comprising the steps of: screening a sample from said patient for the presence or absence of one or more metabolic markers selected from the group consisting of metabolites with an accurate mass (measured in Daltons) of, or substantially equivalent to, 440.3532, 446.3413, 448.3565, 450.3735, 464.3531, 466.3659, 468.3848, 474.3736, 478.405, 484.3793, 490.3678, 492.3841, 494.3973, 502.4055, 504.4195, 510.3943, 512.4083, 518.3974, 520.4131, 522.4323, 530.437, 532.4507, 534.3913, 538.427, 540.4393, 548.4442, 550.4609, 558.4653, 566.4554, 574.4597, 576.4762, 578.493, 590.4597, 592.4728, 594.4857, 596.5015, 598.5121, where a +/−5 ppm difference would indicate the same metabolite; wherein the absence or significant reduction of one or more of said metabolic markers indicates the presence of an ovarian cancer, or the risk of developing ovarian cancer.
In a further embodiment of the present invention there is provided a method for diagnosing a patient for the presence of an ovarian cancer, or the risk of developing ovarian cancer, comprising the steps of: screening a sample from said patient for the presence or absence of one or more metabolic markers selected from the group consisting of metabolites with an accurate mass (measured in Daltons) of, or substantially equivalent to, 446.3413, 448.3565, 450.3735, 468.3848, 474.3872, 476.5, 478.405, 484.3793, 490.3678, 492.3841, 494.3973, 496.4157, 502.4055, 504.4195, 512.4083, 518.3974, 520.4131, 522.4323, 530.437, 532.4507, 538.427, 540.4393, 550.4609, 558.4653, 574.4597, 576.4757, 578.4848, 592.357, 594.4848, 596.5012, 598.5121, where a +/−5 ppm difference would indicate the same metabolite; wherein the absence or significant reduction of one or more of said metabolic markers indicates the presence of an ovarian cancer, or the risk of developing ovarian cancer.
In a further embodiment of the present invention there is provided a method for diagnosing a patient for the presence of an ovarian cancer, or the risk of developing ovarian cancer, comprising the steps of: screening a sample from said patient for the presence or absence of one or more metabolic markers selected from the group consisting of metabolites with an accurate mass (measured in Daltons) of, or substantially equivalent to masses to 446.3396, 448.3553, 450.3709, 468.3814, 474.3736, 478.4022, 484.3764, 490.3658, 492.3815, 494.3971, 496.4128, 502.4022, 504.4179, 512.4077, 518.3971, 520.4128, 522.8284, 530.43351, 532.44916, 538.4233, 540.4389, 550.4597, 558.4648, 574.4597, 576.4754, 578.4910, 592.47029, 594.4859, 596.5016, and 598.5172, where a +/−5 ppm difference would indicate the same metabolite; wherein the absence or significant reduction of one or more of said metabolic markers indicates the presence of an ovarian cancer, or the risk of developing ovarian cancer. In this embodiment of the invention the molecular formulas of the metabolites with these masses are C28H46O4, C28H48O4, C28H50O4, C28H52O5, C30H50O4, C30H54O4, C28H52O6, C30H50O5, C30H52O5, C30H54O5, C30H56O5, C32H54O4, C32H56O4, C30H56O6, C32H54O5, C32H56O5, C32H60O5, C34H58O4, C34H60O4, C32H58O6, C32H60O6, C34H62O5, C36H62O4, C36H62O5, C36H64O5, C36H66O5, C36H64O6, C36H66O6, C36H68O6, and C36H70O6, respectively and the proposed structures are as shown below:
respectively.
In a further embodiment of the present invention there is provided a method for diagnosing a patient for the presence of an ovarian cancer, or the risk of developing ovarian cancer, comprising the steps of: screening a sample from said patient for the presence or absence of one or more metabolic markers selected from the group consisting of metabolites with an accurate mass of, or substantially equivalent to neutral masses 450 (C28H50O4), 446 (C28H46O4), 468 (C28H52O5), 448 (C28H48O4), 464 (C28H48O5) and 466 (C28H50O5) wherein the absence or significant reduction of one or more of said metabolic markers indicates the presence of an ovarian cancer, or the risk of developing ovarian cancer.
In a further embodiment of the present invention there is provided a method for diagnosing the presence or absence of ovarian cancer in a test subject of unknown ovarian cancer status, comprising: analyzing a blood sample from a test subject to obtain quantifying data on molecules selected from the group comprised of molecules identified by the neutral accurate masses shown in Table 1, where a +/−5 ppm difference would indicate the same metabolite, or molecules having masses substantially equal to these molecules or fragments of derivatives thereof; comparing the quantifying data obtained on said molecules in said test subject with quantifying data obtained from said molecules from a plurality of ovarian cancer-positive humans or quantifying data obtained from a plurality of ovarian cancer-negative humans; and wherein said comparison can be used to determine the probability that the test subject is ovarian cancer-positive or -negative.
In a further embodiment of the present invention there is provided a method for diagnosing the presence or absence of ovarian cancer in a test subject of unknown ovarian cancer status, comprising: analyzing a blood sample from a test subject to obtain quantifying data on molecules selected from the group comprised of molecules identified by the neutral accurate masses (measured in Daltons) 440.3532, 446.3413, 448.3565, 450.3735, 464.3531, 466.3659, 468.3848, 474.3736, 478.405, 484.3793, 490.3678, 492.3841, 494.3973, 502.4055, 504.4195, 510.3943, 512.4083, 518.3974, 520.4131, 522.4323, 530.437, 532.4507, 534.3913, 538.427, 540.4393, 548.4442, 550.4609, 558.4653, 566.4554, 574.4597, 576.4762, 578.493, 590.4597, 592.4728, 594.4857, 596.5015, 598.5121, where a +/−5 ppm difference would indicate the same metabolite, or molecules having masses substantially equal to these molecules or fragments of derivatives thereof; comparing the quantifying data obtained on said molecules in said test subject with quantifying data obtained from said molecules from a plurality of ovarian cancer-positive humans or quantifying data obtained from a plurality of ovarian cancer-negative humans; and wherein said comparison can be used to determine the probability that the test subject is ovarian cancer-positive or -negative.
In a further embodiment of the present invention there is provided a method for diagnosing the presence or absence of ovarian cancer in a test subject of unknown ovarian cancer status, comprising: analyzing a blood sample from a test subject to obtain quantifying data on molecules selected from the group comprised of molecules identified by the neutral accurate masses (measured in Daltons) 446.3413, 476.5, 448.3565, 450.3735, 468.3848, 474.3872, 478.405, 484.3793, 490.3678, 492.3841, 494.3973, 496.4157, 502.4055, 504.4195, 512.4083, 518.3974, 520.4131, 522.4323, 530.437, 532.4507, 538.427, 540.4393, 550.4609, 558.4653, 574.4597, 576.4757, 578.4848, 592.357, 594.4848, 596.5012, 598.5121, where a +/−5 ppm difference would indicate the same metabolite, or molecules having masses substantially equal to these molecules or fragments of derivatives thereof; comparing the quantifying data obtained on said molecules in said test subject with quantifying data obtained from said molecules from a plurality of ovarian cancer-positive humans or quantifying data obtained from a plurality of ovarian cancer-negative humans; and wherein said comparison can be used to determine the probability that the test subject is ovarian cancer-positive or -negative.
In a further embodiment of the present invention there is provided a method for diagnosing the presence or absence of ovarian cancer in a test subject of unknown ovarian cancer status, comprising: analyzing a blood sample from a test subject to obtain quantifying data on molecules selected from the group comprised of molecules identified by the neutral accurate masses (measured in Daltons) 446.3396, 448.3553, 450.3709, 468.3814, 474.3736, 478.4022, 484.3764, 490.3658, 492.3815, 494.3971, 496.4128, 502.4022, 504.4179, 512.4077, 518.3971, 520.4128, 522.8284, 530.43351, 532.44916, 538.4233, 540.4389, 550.4597, 558.4648, 574.4597, 576.4754, 578.4910, 592.47029, 594.4859, 596.5016, and 598.5172, where a +/−5 ppm difference would indicate the same metabolite, or molecules having masses substantially equal to these molecules or fragments of derivatives thereof; comparing the quantifying data obtained on said molecules in said test subject with quantifying data obtained from said molecules from a plurality of ovarian cancer-positive humans or quantifying data obtained from a plurality of ovarian cancer-negative humans; and wherein said comparison can be used to determine the probability that the test subject is ovarian cancer-positive or -negative. In this embodiment of the invention the molecular formulas of the metabolites with these masses are C28H46O4, C28H48O4, C28H50O4, C28H52O5, C30H50O4, C30H54O4, C28H52O6, C30H50O5, C30H52O5, C30H54O5, C30H56O5, C32H54O4, C32H56O4, C30H56O6, C32H54O5, C32H56O5, C32H60O5, C34H58O4, C34H60O4, C32H58O6, C32H60O6, C34H62O5, C36H62O4, C36H62O5, C36H64O5, C36H66O5, C36H64O6, C36H66O6, C36H68O6, and C36H70O6, respectively and the proposed structures are as shown below:
respectively.
In a further embodiment of the present invention there is provided a method for diagnosing the presence or absence of ovarian cancer in a test subject of unknown ovarian cancer status, comprising: analyzing a blood sample from a test subject to obtain quantifying data on molecules selected from the group comprised of molecules identified by the neutral accurate masses 450 (C28H50O4), 446 (C28H46O4), 468 (C28H52O5), 448 (C28H48O4), 464 (C28H48O5) and 466 (C28H50O5) comparing the quantifying data obtained on said molecules in said test subject with quantifying data obtained from said molecules from a plurality of ovarian cancer-positive humans or quantifying data obtained from a plurality of ovarian cancer-negative humans; and wherein said comparison can be used to determine the probability that the test subject is ovarian cancer-positive or -negative.
The identification of ovarian cancer biomarkers with improved diagnostic accuracy in human serum, therefore, would be extremely beneficial, as the test would be non-invasive and could possibly be used to monitor individual susceptibility to disease prior to, or in combination with, conventional methods. A serum test is minimally invasive and would be accepted across the general population. The present invention relates to a method of diagnosing ovarian cancer, or the risk of developing ovarian cancer, by measuring the levels of specific small molecules present in human serum and comparing them to “normal” reference levels. The invention discloses several hundred metabolite masses which were found to have statistically significant differential abundances between ovarian cancer-positive serum and normal serum, of which in one embodiment of the present invention a subset of 37, and in a further embodiment a subset of 31 metabolite masses, a further subset of 30 metabolite masses and a further subset of 6 metabolite markers are used to illustrate the diagnostic utility by discriminating between disease-positive serum and control serum samples. In yet a further embodiment of the present invention, any one or combination of the metabolites identified in the present invention can be used to indicate the presence of ovarian cancer. A diagnostic assay based on small molecules, or metabolites, in serum fulfills the above criteria for an ideal screening test, as development of assays capable of detecting specific metabolites is relatively simple and cost effective per assay. Translation of the method into a clinical assay compatible with current clinical chemistry laboratory hardware would be commercially acceptable and effective, and would result in a rapid deployment worldwide. Furthermore, the requirement for highly trained personnel to perform and interpret the test would be eliminated.
The selected 31 metabolites, identified according to the present invention, were further characterized by molecular formulae and structure. This additional information for 30 of the metabolites is shown in Table 35.
The present invention also discloses the identification of vitamin E-like metabolites that are differentially expressed in the serum of OC-positive patients versus healthy controls. The differential expressions disclosed are specific to OC.
In one embodiment of the present invention, a serum test, developed using an optimal subset of metabolites selected from the group consisting of vitamin E-like metabolites, can be used to diagnose the presence of OC, or the risk of developing ovarian cancer, or the presence of an OC-promoting or inhibiting environment.
In another embodiment of the present invention, a serum test, developed using an optimal subset of metabolites selected from the group consisting of vitamin E-like metabolites, can be used to diagnose the OC health-state resulting from the effect of treatment of a patient diagnosed with OC. Treatment may include chemotherapy, surgery, radiation therapy, biological therapy, or other.
In another embodiment of the present invention, a serum test, developed using an optimal subset of metabolites selected from the group consisting of vitamin E-like metabolites, can be used to longitudinally monitor the OC status of a patient on a OC therapy to determine the appropriate dose or a specific therapy for the patient.
The present invention also discloses the identification of gamma-tocopherol/tocotrienol metabolites in which the aromatic ring structure has been reduced that are differentially expressed in the serum of OC-positive patients versus healthy controls. The differential expressions disclosed are specific to OC. Therefore, according to the present invention, the metabolites can be used to monitor irregularities or abnormalities in the biological pathways or systems associated with ovarian cancer.
The present invention discloses the presence of gamma-tocopherol/tocotrienol metabolites in which there exists —OC2H5, —OC4H9, or —OC8H17 moieties attached to the hydroxychroman-containing structure in human serum.
In a further embodiment of the present invention there is provided a method for identifying and diagnosing individuals who would benefit from anti-oxidant therapy comprising: analyzing a blood sample from a test subject to obtain quantifying data on all, or a subset of, gamma tocopherols, gamma tocotrienols, omega-carboxylated gamma tocopherol and gamma tocotrienol, vitamin E-related metabolites or metabolic derivatives of said metabolite classes; comparing the quantifying data obtained on said molecules in said test subject with reference data obtained from the analysis of a plurality of OC-negative humans; wherein said comparison can be used to determine the probability that the test subject would benefit from such therapy.
In a further embodiment of the present invention there is provided a method for determining the probability that a subject is at risk of developing OC comprising: analyzing a blood sample from an OC asymptomatic subject to obtain quantifying data on all, or a subset of, gamma tocopherols, gamma tocotrienols, or metabolic derivatives of said metabolite classes; comparing the quantifying data obtained on said molecules in said test subject with reference data obtained from the analysis of a plurality of OC-negative humans; wherein said comparison can be used to determine the probability that the test subject is at risk of developing OC.
In a further embodiment of the present invention there is provided a method for monitoring irregularities or abnormalities in the biological pathway or system associated with ovarian cancer comprising: analyzing a blood sample from an test subject of unknown ovarian cancer status to obtain quantifying data on all, or a subset of, gamma tocopherols, gamma tocotrienols, or metabolic derivatives of said metabolite classes; comparing the quantifying data obtained on said molecules in said test subject with reference data obtained from the analysis of a plurality of OC-negative humans;
wherein said comparison can be used to monitoring irregularities or abnormalities in the biological pathways or systems associated with ovarian cancer.
In a further embodiment of the present invention there is provided a method for identifying individuals who respond to a dietary, chemical, or biological therapeutic strategy designed to prevent, cure, or stabilize OC or improve symptoms associated with OC comprising: analyzing one or more blood samples from a test subject either from a single collection or from multiple collections over time to obtain quantifying data on all, or a subset of, gamma tocopherols, gamma tocotrienols, omega-carboxylated gamma tocopherol and gamma tocotrienol, vitamin E-like molecules, or metabolic derivatives of said metabolite classes; comparing the quantifying data obtained on said molecules in said test subject's samples with reference data obtained from said molecules from a plurality of OC-negative humans; wherein said comparison can be used to determine whether the metabolic state of said test subject has improved during said therapeutic strategy.
In a further embodiment of the present invention, there is provided a method for identifying individuals who are deficient in the cellular uptake or transport of vitamin E and related metabolites by the analysis of serum or tissue using various strategies, including, but not limited to: radiolabeled tracer studies, gene expression or protein expression analysis of vitamin E transport proteins, analysis of genomic aberrations or mutations in vitamin E transport proteins, in vivo or ex vivo imaging of vitamin E transport protein levels, antibody-based detection (enzyme-linked immunosorbant assay, ELISA) of vitamin E transport proteins.
This summary of the invention does not necessarily describe all features of the invention.
These and other features of the invention will become more apparent from the following description in which reference is made to the appended drawings wherein:
The present invention relates to the diagnosis of ovarian cancer (OC), or the risk of developing OC. The present invention describes the relationship between endogenous small molecules and OC. Specifically, the present invention relates to the diagnosis of OC, or the risk of developing OC, through the measurement of vitamin E isoforms and related metabolites. More specifically, the present invention relates to the relationship between vitamin E-related metabolites in human serum and the implications thereof in OC.
The present invention discloses for the first time clear and unambiguous biochemical changes specifically associated with OC. These findings also imply that the measurement of these biomarkers may provide a universal means of measuring the effectiveness of OC therapies. This would dramatically decrease the cost of performing clinical trials as a simple biochemical test can be used to assess the viability of new therapeutics. Furthermore, one would not have to wait until the tumor progresses or until the patient dies to determine whether the therapy provided any benefit. The use of such a test would enable researchers to determine in months, rather than years, the effectiveness of dose, formulation, and chemical structure modifications of OC therapies.
The present invention relates to a method of diagnosing OC by measuring the levels of specific small molecules present in human serum and comparing them to “normal” reference levels. In one embodiment of the present application there is described a novel method for the early detection and diagnosis of OC and the monitoring the effects of OC therapy.
One method of the present invention uses accurate masses in an FTMS based method. The accurate masses that can be used according to this invention include the masses shown in Table 1, or a subset thereof.
A further method involves the use of a high-throughput screening (HTS) assay developed from a subset of metabolites selected from Table 1 for the diagnosis of one or more diseases or particular health-states. The utility of the claimed method is demonstrated and validated through the development of a HTS assay capable of diagnosing an OC-positive health-state.
The impact of such an assay on OC would be tremendous, as literally everyone could be screened longitudinally throughout their lifetime to assess risk and detect ovarian cancer early. Given that the performance characteristics of the test are representative for the general OC population, this test alone may be superior to any other currently available OC screening method, as it may have the potential to detect disease progression prior to that detectable by conventional methods. The early detection of OC is critical to positive treatment outcome.
The term “vitamin E” collectively refers to eight naturally occurring isoforms, four tocopherols (alpha, beta, gamma, and delta) and four tocotrienols (alpha, beta, gamma, and delta). The predominant form found in western diets is gamma-tocopherol whereas the predominant form found in human serum/plasma is alpha-tocopherol. Tocotrienols are also present in the diet, but are more concentrated in cereal grains and certain vegetable oils such as palm and rice bran oil. Interestingly, it is suggested that tocotrienols may be more potent than tocopherols in preventing cardiovascular disease and cancer (5). This may be attributable to the increased distribution of tocotrienols within lipid membranes, a greater ability to interact with radicals, and the ability to be quickly recycled more quickly than tocopherol counterparts (6). It has been demonstrated that in rat liver microsomes, the efficacy of alpha-tocotrienol to protect against iron-mediated lipid peroxidation was 40 times higher that that of alpha-tocopherol (6). However, measurements in human plasma indicate that trienols are either not detected or present only in minute concentrations (7), due possibly to the higher lipophilicity resulting in preferential bilary excretion (8).
A considerable amount of research related to the discrepancy between the distribution of alpha and gamma tocopherol has been performed on these isoforms. It has been known and reported as early as 1974 that gamma- and alpha-tocopherol have similar intestinal absorption but significantly different plasma concentrations (9). In the Bieri and Evarts study (9), rats were depleted of vitamin E for 10 days and then fed a diet containing an alpha:gamma ratio of 0.5 for 14 days. At day 14, the plasma alpha:gamma ratio was observed to be 5.5. The authors attributed this to a significantly higher turnover of gamma-tocopherol, however, the cause of this increased turnover was unknown. Plasma concentrations of the tocopherols are believed to be tightly regulated by the hepatic tocopherol binding protein. This protein has been shown to preferentially bind to alpha-tocopherol (10). Large increases in alpha-tocopherol consumption result in only small increases in plasma concentrations (11). Similar observations hold true for tocotrienols, where high dose supplementation has been shown to result in maximal plasma concentrations of approximately only 1 to 3 micromolar (12). More recently, Birringer et al (8) showed that although upwards of 50% of ingested gamma-tocopherol is metabolized by human hepatoma HepG2 cells by omega-oxidation to various alcohols and carboxylic acids, less than 3% of alpha-tocopherol is metabolized by this pathway. This system appears to be responsible for the increased turnover of gamma-tocopherol. In this paper, they showed that the creation of the omega COOH from gamma-tocopherol occured at a rate of >50× than the creation of the analogous omega COOH from alpha-tocopherol. Birringer also showed that the trienols are metabolized via a similar, but more complex omega carboxylation pathway requiring auxiliary enzymes (8).
It is likely that the existence of these two structurally selective processes has biological significance. Birringer et al (8) propose that the purpose of the gamma-tocopherol-specific P450 omega hydroxylase is the preferential elimination of gamma-tocopherol/trienol as 2,7,8-trimethyl-2-(beta-carboxy-3′-carboxyethyl)-6-hydroxychroman (gamma-CEHC). We argue, however, that if the biological purpose is simply to eliminate gamma-tocopherol/trienol, it would be far simpler and more energy efficient via selective hydroxylation and glucuronidation. The net biological effect of these two processes, which has not been commented on in the vitamin E literature, is that the two primary dietary vitamin E isoforms (alpha and gamma), upon entering the liver during first-pass metabolism, are shunted into two separate metabolic systems. System 1 quickly moves the most biologically active antioxidant isoform (alpha-tocopherol) into the blood stream to supply the tissues of the body with adequate levels of this essential vitamin. System 2 quickly converts gamma-tocopherol into the omega COOH. In the present invention it is disclosed that significant concentrations of multiple isoforms of gamma-tocopherol/tocotrienol omega COOH are present in normal human serum at all times. We were able to estimate that the concentration of each of these molecules in human serum is in the low micromolar range by measuring cholic acid, an organically soluble carboxylic acid-containing internal standard used in the triple-quadrupole method. This is within the previously reported plasma concentration range of 0.5 to 2 micromolar for γ-tocopherol (approximately 20 times lower than that of alpha-tocopherol) (13) The cumulative total, therefore, of all said novel γ-tocoenoic acids in serum is not trivial, and likely exceeds that of γ-tocopherol itself. None of the other shorter chain length gamma-tocopherol/trienol metabolites described by Birringer et al (8) were detected in the serum. Also, the alpha and gamma tocotrienols were also not detected in the serum of patients used in the studies reported in this work, suggesting that the primary purpose of the gamma-tocopherol/trineol-specific P450 omega hydroxylase is the formation of the omega COOH and not gamma-CEHC. Not to be bound by the correctness of the theory, it is therefore suggested that the various gamma-tocopherol/tocotrienol omega COOH metabolites disclosed in the present application are novel bioactive agents and that they perform specific and necessary biological functions for the maintenance of normal health and for the prevention of disease.
Of relevance is also the fact that it has been shown that mammals are able to convert trienols to tocopherols in vivo (14, 15). Since several of the novel vitamin E-like metabolites disclosed herein contain a semi-saturated phytyl side chain, the possibility of a tocotrienol precursor cannot be excluded.
Just as trienols have been reported to have biological activities separate from the tocopherols (16), gamma-tocopherol has been reported to have biological functions separate and distinct from alpha-tocopherol. For example, key differences between alpha tocopherol and alpha tocotrienol include the ability of alpha tocotrienol to specifically prevent neurodegeneration by regulating specific mediators of cell death (17), the ability of trienols to lower cholesterol (18), the ability to reduce oxidative protein damage and extend life span of C. elegans (19), and the ability to suppress the growth of breast cancer cells (20, 21). Key differences between the gamma and alpha forms of tocopherol include the ability of gamma to decrease proinflammatory eicosanoids in inflammation damage in rats (22) and inhibition of cyclooxygenase (COX-2) activity (23). In Jiang et al (23) it was reported that it took 8-24 hours for gamma-tocopherol to be effective and that arachadonic acid competitively inhibits the suppression activity of gamma-tocopherol. It is hypothesized that the omega COOH metabolites of gamma-tocopherol may be the primary bioactive species responsible for its anti-inflammation activity. The conversion of arachadonic acid into eicosanoids is a critical step in inflammation. It is more conceivable that omega COOH forms of gamma-tocopherol, due to their structural similarities to arachadonic acid, are more potent competitive inhibitors of this formation than native gamma-tocopherol.
In one aspect of this invention there is provided novel gamma-tocopherol/tocotrienol metabolites in human serum. These gamma-tocopherol/trienol metabolites have had the aromatic ring structure reduced. In this aspect of the invention, the gamma-tocopherol/tocotrienol metabolites comprise —OC2H5, —OC4H9, or —OC8H17 moieties attached to the hydroxychroman structure in human serum.
Not wishing to be bound by any particular theory, in the present invention it is hypothesized that the novel metabolites disclosed herein are indicators of vitamin E activity and that the decrease of such metabolites is indicative of one of the following situations:
In all of the aforementioned related epidemiological studies concerning vitamin E, there is little known about the correlation between gamma tocopherol and OC. At the time of this application, a PubMed search for “Ovarian Cancer” and “Gamma Tocopherol” returned only one publication reporting no change in plasma gamma tocopherol levels between OC patients and controls (24). More recent findings have eluded to a potential inverse association between alpha-tocopherol supplementation and ovarian cancer risk (25). Basic research has shown that alpha tocopherol can inhibit telomerase activity in ovarian cancer cells in vitro, suggesting a potential role in the control of ovarian cancer cell growth. No in vitro effects of gamma tocopherol on ovarian cancer cells has been reported.
Based on the discoveries disclosed in this application, it is contemplated that although dietary deficiencies or deficiencies in specific vitamin E metabolizing enzymes may increase the risk of OC incidence, it is also contemplated that the presence of OC may result in the decrease of vitamin E isoforms and related metabolites. These decreased levels are not likely to be the result of a simple dietary deficiency, as such a strong association would have been previously revealed in epidemiological studies, such as in the study performed by Helzlsouer et al (24).
Based on the discoveries disclosed in this application, it is also contemplated that the decreased levels of vitamin E-like metabolites are not the result of a simple dietary deficiency, but rather impairment in the colonic epithelial uptake of vitamin E and related molecules. This therefore represents a rate-limiting step for the sufficient provision of anti-oxidant capacity to epithelial cells under an oxidative stress load. In this model, the dietary effects of increased iron consumption through red meats, high saturated fat, and decreased fiber (resulting in a decreased iron chelation effect (26)) results in the previously mentioned Fenton-induced free radical propagation, of which sufficient scavenging is dependent upon adequate epithelial levels of vitamin E. Increases in epithelial free radical load, combined with a vitamin E-related transport deficiency, would therefore be reflected by a decrease in vitamin E-like metabolites as anti-oxidants, as well as decreases in the reduced carboxylated isoforms resulting from hepatic uptake and P450-mediated metabolism. It has recently been shown that the uptake of Vitamin E into CaCo-2 colonic epithelial cells is a saturable process, heavily dependent upon a protein-mediated event (27). Because protein transporters are in essence enzymes, and follow typical Michaelis-Menton kinetics, the rate at which vitamin E can be taken up into colonic epithelial cells would reach a maximal velocity (Vmax), which may not be capable of providing a sufficient anti-oxidant protective effect for the development of OC. At some point in time, therefore, increasing rates of oxidative stress above the rate at which vitamin E can be transported from the diet will deplete the endogenous pool.
Discovery and Identification of Differentially Expressed Metabolites in Ovarian Cancer-Positive Versus Normal Healthy Controls
Clinical Samples. In order to determine whether there are biochemical markers of a given health-state in a particular population, a group of patients representative of the health-state (i.e. a particular disease) and a group of “normal” counterparts are required. Biological samples taken from the patients in a particular health-state category can then be compared to equivalent samples taken from the normal population with the objective of identifying differences between the two groups, by extracting and analyzing the samples using various analytical platforms including, but not limited to, FTMS and LC-MS. The biological samples could originate from anywhere within the body, including, but not limited to, blood (serum/plasma), cerebrospinal fluid (CSF), urine, stool, breath, saliva, or biopsy of any solid tissue including tumor, adjacent normal, smooth and skeletal muscle, adipose tissue, liver, skin, hair, kidney, pancreas, lung, colon, stomach, or other.
For the ovarian cancer diagnostic assay described herein, serum samples were obtained from representative populations of healthy ovarian cancer-negative individuals and professionally diagnosed ovarian cancer-positive patients. Throughout this application, the term “serum” will be used, but it will be obvious to those skilled in the art that plasma or whole blood or a sub-fraction of whole blood may also be used in the method. The biochemical markers of ovarian cancer described in the invention were derived from the analysis of 20 serum samples from ovarian cancer positive patients and 25 serum samples from healthy controls. In subsequent validation tests, 539 control samples (not diagnosed with ovarian cancer; 289 subjects using the C28 HTS panel, and another 250 using the 31 molecule HTS panel) and 241 ovarian cancer samples were assessed. All samples were single time-point collections, while 289 ovarian cancer samples were taken either immediately prior to or immediately following surgical resection of a tumor (prior to chemotherapy or radiation therapy). The 250 ovarian subset (shown in
Non-Targeted Metabolomic Strategies.
Multiple non-targeted metabolomics strategies have been described in the scientific literature including NMR (28), GC-MS (29-31), LC-MS, and FTMS strategies (28, 32-34). The metabolic profiling strategy employed for the discovery of differentially expressed metabolites in this application was the non-targeted FTMS strategy invented by Phenomenome Discoveries Inc. (30, 34-37). Non-targeted analysis involves the measurement of as many molecules in a sample as possible, without any prior knowledge or selection of components prior to the analysis. Therefore, the potential for non-targeted analysis to discover novel metabolite biomarkers is high versus targeted methods, which detect a predefined list of molecules. The present invention uses a non-targeted method to identify metabolite components that differ between ovarian cancer-positive and healthy individuals, followed by the development of a high-throughput targeted assay for a subset of the metabolites identified from the non-targeted analysis. However, it would be obvious to anyone skilled in the art that other metabolite profiling strategies could potentially be used to discover some or all of the differentially regulated metabolites disclosed in this application, and that the metabolites described herein, however discovered or measured, represent unique chemical entities that are independent of the analytical technology that may be used to detect and measure them.
Sample Processing.
When a blood sample is drawn from a patient there are several ways in which the sample can be processed. The range of processing can be as little as none (i.e. frozen whole blood) or as complex as the isolation of a particular cell type. The most common and routine procedures involve the preparation of either serum or plasma from whole blood. All blood sample processing methods, including spotting of blood samples onto solid-phase supports, such as filter paper or other immobile materials, are also contemplated by the invention.
Sample Extraction.
The processed blood sample described above is then further processed to make it compatible with the analytical technique to be employed in the detection and measurement of the biochemicals contained within the processed blood sample (in our case, a serum sample). The types of processing can range from as little as no further processing to as complex as differential extraction and chemical derivatization. Extraction methods may include, but are not limited to, sonication, soxhlet extraction, microwave assisted extraction (MAE), supercritical fluid extraction (SFE), accelerated solvent extraction (ASE), pressurized liquid extraction (PLE), pressurized hot water extraction (PHWE), and/or surfactant assisted extraction (PHWE) in common solvents such as methanol, ethanol, mixtures of alcohols and water, or organic solvents such as ethyl acetate or hexane. The preferred method of extracting metabolites for FTMS non-targeted analysis is to perform a liquid/liquid extraction whereby non-polar metabolites dissolve in an organic solvent and polar metabolites dissolve in an aqueous solvent. The metabolites contained within the serum samples used in this application were separated into polar and non-polar extracts through sonication and vigorous mixing (vortex mixing).
Mass Spectrometry Analysis of Extracts.
Extracts of biological samples are amenable to analysis on essentially any mass spectrometry platform, either by direct injection or following chromatographic separation. Typical mass spectrometers are comprised of a source, which ionizes molecules within the sample, and a detector for detecting the ionized particles. Examples of common sources include electron impact, electrospray ionization (ESI), atmospheric pressure chemical ionization (APCI), matrix assisted laser desorption ionization (MALDI), surface enhanced laser desorption ionization (SELDI), and derivations thereof. Common ion detectors can include quadrupole-based systems, time-of-flight (TOF), magnetic sector, ion cyclotron, and derivations thereof.
The present invention will be further illustrated in the following examples.
The invention described herein involved the analysis of serum extracts from 45 individuals (20 with ovarian cancer, 25 healthy controls) by direct injection into a FTMS and ionization by either ESI or APCI in both positive and negative modes. The advantage of FTMS over other MS-based platforms is the high resolving capability that allows for the separation of metabolites differing by only hundredths of a Dalton, many which would be missed by lower resolution instruments. Sample extracts were diluted either three or six-fold in methanol:0.1% (v/v) ammonium hydroxide (50:50, v/v) for negative ionization modes, or in methanol:0.1% (v/v) formic acid (50:50, v/v) for positive ionization modes. For APCI, sample extracts were directly injected without diluting. All analyses were performed on a Bruker Daltonics APEX III FTMS equipped with a 7.0 T actively shielded superconducting magnet (Bruker Daltonics, Billerica, Mass.). Samples were directly injected using electrospray ionization (ESI) and atmospheric pressure chemical ionization (APCI) at a flow rate of 600 μL per hour. Ion transfer/detection parameters were optimized using a standard mix of serine, tetra-alanine, reserpine, Hewlett-Packard tuning mix, and the adrenocorticotrophic hormone fragment 4-10. In addition, the instrument conditions were tuned to optimize ion intensity and broad-band accumulation over the mass range of 100-1000 amu according to the instrument manufacturer's recommendations. A mixture of the abovementioned standards was used to internally calibrate each sample spectrum for mass accuracy over the acquisition range of 100-1000 amu.
In total six separate analyses comprising combinations of extracts and ionization modes were obtained for each sample:
Aqueous Extract
1. Positive ESI (analysis mode 1101)
2. Negative ESI (analysis mode 1102)
Organic Extract
3. Positive ESI (analysis mode 1201)
4. Negative ESI (analysis mode 1202)
5. Positive APCI (analysis mode 1203)
6. Negative APCI (analysis mode 1204)
Mass Spectrometry Data Processing. Using a linear least-squares regression line, mass axis values were calibrated such that each internal standard mass peak had a mass error of <1 ppm compared with its theoretical mass. Using XMASS software from Bruker Daltonics Inc., data file sizes of 1 megaword were acquired and zero-filled to 2 megawords. A sinm data transformation was performed prior to Fourier transform and magnitude calculations. The mass spectra from each analysis were integrated, creating a peak list that contained the accurate mass and absolute intensity of each peak. Compounds in the range of 100-2000 m/z were analyzed. In order to compare and summarize data across different ionization modes and polarities, all detected mass peaks were converted to their corresponding neutral masses assuming hydrogen adduct formation. A self-generated two-dimensional (mass vs. sample intensity) array was then created using DISCO VAmetrics™ software (Phenomenome Discoveries Inc., Saskatoon, SK, Canada). The data from multiple files were integrated and this combined file was then processed to determine all of the unique masses. The average of each unique mass was determined, representing the γ-axis. A column was created for each file that was originally selected to be analyzed, representing the x-axis. The intensity for each mass found in each of the files selected was then filled into its representative x,y coordinate. Coordinates that did not contain an intensity value were left blank. Once in the array, the data were further processed, visualized and interpreted, and putative chemical identities were assigned. Each of the spectra were then peak picked to obtain the mass and intensity of all metabolites detected. These data from all of the modes were then merged to create one data file per sample. The data from all 45 samples were then merged and aligned to create a two-dimensional metabolite array in which each sample is represented by a column and each unique metabolite is represented by a single row. In the cell corresponding to a given metabolite sample combination, the intensity of the metabolite in that sample is displayed. When the data is represented in this format, metabolites showing differences between groups of samples (i.e., normal and cancer) can be determined.
Advanced Data Interpretation.
A student's T-test was used to select for metabolites that differ between the normal and the ovarian cancer-positive samples (p<0.05). Four hundred and twenty four metabolites met this criterion (shown in Table 1). These are all features that differ statistically between the two populations and therefore have potential diagnostic utility. The features are described by their accurate mass and analysis mode (1204, organic extract and negative APCI), which together are sufficient to provide the putative molecular formulas and chemical characteristics (such as polarity and putative functional groups) of each metabolite. Table 1 also shows the average biomarker intensities and standard deviations of the intensities in the normal and ovarian samples. A log(2) ratio of the metabolite intensities (normal/ovarian) is shown in the far right column. By definition, since each of the metabolites in Table 1 shows a statistically significant difference (p<0.05) between the ovarian and control populations, each mass alone could be individually used to determine whether the health state of a person is “normal” or “ovarian” in nature. For example, this diagnosis could be performed by determining optimal cut-off points for each of the masses in Table 1, and by comparing the relative intensity of the biomarker in an unknown sample to the levels of the marker in the normal and ovarian population, a likelihood ratio for either being ovarian-positive or normal calculated for the unknown sample. This approach could be used individually for any or all of the masses listed in Table 1. Alternatively, this approach could be used on each mass, and then a combined average likelihood score based upon all the masses used.
Similar approaches to the above example would include any methods that use each or all of the masses to generate an averaged or standardized value representing all measure biomarker intensities for ovarian cancer. For example, the intensity of each mass would be measured, and then either used directly or following a normalization method (such as mean normalization, log normalization, Z-score transformation, min-max scaling, etc) to generate a summed or averaged score. Such sums or averages will differ significantly between the ovarian and normal populations, allowing cut-off scores to be used to predict the likelihood of ovarian cancer or normality in future unclassified samples. The cutoff scores themselves, whether for individual masses or for averages or standardized averages of all the masses in Table 1, can be selected using standard operator-receiver characteristic calculations.
A third example in which all masses listed in Table 1 could be used to provide a diagnostic output would be through the use of either a multivariate supervised or unsupervised classification or clustering algorithms. Similar to those listed below for optimal feature set selection, multivariate classification methods such as principal component analysis (PCA) and hierarchical clustering (HCA) (both unsupervised, ie, the algorithm does not know which samples belong to which disease variable), and supervised methods such as supervised PCA, partial least squared discriminant analysis (PLSDA), logistic regression, artificial neural networks (ANNs), support vector machine (SVMs), Bayesian methods and others (see 38 for review), perform optimally with more features. This is shown in the example in
However, the incorporation and development of 424 signals into a commercially useful assay is impractical, and therefore supervised methods such as those listed above are often employed to determine the fewest number of features required to maintain an acceptable level of diagnostic accuracy. In this application, no supervised training classifiers were used to narrow the list further; rather, the list was reduced to 37 (see Table 2) based on univariate analysis, 13C filtering, and mode selection. Any other subset from the 424 masses listed in Table 1 can be used according to the present invention to develop a assay for detecting ovarian cancer. A subset of 30 metabolite markers is listed in Table 35. Furthermore, a subset of 29 metabolite markers is listed in Table 3. Alternatively, several supervised methods also exist, of which any one could have been used to identify an alternative subset of masses, including artificial neural networks (ANNs), support vector machines (SVMs), partial least squares discriminant analysis (PLSDA), sub-linear association methods, Bayesian inference methods, supervised principal component analysis, shrunken centroids, or others (see (38) for review).
The identification of metabolites that can distinguish ovarian cancer patient serum from healthy control serum began with the generation of comprehensive metabolomic profiles of 20 ovarian cancer patients and 25 controls, as described in Example 1. The full dataset comprised 1,244 sample-specific masses, of which 424 showed p-values of less than 0.05 when the data was log(2) transformed and a student's t-test between the ovarian cancer samples and controls performed (Table 1). Each of these masses is statistically significant in discriminating between the ovarian cancer and control cohorts, and therefore has potential diagnostic utility. In addition any subset of the 424-metabolite markers has potential diagnostic utility. Table 1 shows these masses ordered according to the p-value (with the lowest p-values at the beginning of the table).
A statistical analysis technique called principal component analysis (PCA) was used to examine the variance within a multivariate dataset. This method is referred to as “unsupervised”, meaning that the method is unaware of which samples belong to which cohorts. The output of a PCA analysis is a two or three-dimensional plot that projects a single point for each sample on the plot according to its variance. The more closely together that points cluster, the lower the variance is between the samples, or the more similar the samples are to each other based on the data. In
However, the incorporation of all 424 masses with p<0.05 into a routine clinical screening method is not practical. As described above, any number of statistical methods, including both supervised and non-supervised methods, could be used to extract subsets of these 424 masses as optimal diagnostic markers, and various methods would yield slightly different results. A subset of 37 metabolites (see Table 2) was selected from the list of 424 as one potential panel of ovarian cancer screening markers. The 37 metabolites were selected by filtering the data for masses with p-values less than 0.0001, removing all 13C isotopes, and excluding metabolites not detected in mode 1204. The list of 37 metabolites are shown in Table 2, and include masses (measured in Daltons) 440.3532, 446.3413, 448.3565, 450.3735, 464.3531, 466.3659, 468.3848, 474.3736, 478.405, 484.3793, 490.3678, 492.3841, 494.3973, 502.4055, 504.4195, 510.3943, 512.4083, 518.3974, 520.4131, 522.4323, 530.437, 532.4507, 534.3913, 538.427, 540.4393, 548.4442, 550.4609, 558.4653, 566.4554, 574.4597, 576.4762, 578.493, 590.4597, 592.4728, 594.4857, 596.5015, 598.5121, where a +/−5 ppm difference would indicate the same metabolite. A PCA plot based solely on these masses, is shown in
The PCA plot does not adequately allow one to visualize the actual intensities of the metabolites responsible for the separation of the clusters. A second statistical method was therefore used, called hierarchical clustering (HCA), to arrange the patient samples into groups based on a Euclidean distance measurements using the said 37 metabolites, which themselves were clustered using a Pearson correlation distance measurement. The resulting metabolite array is shown in
The metabolites and their associations with the clinical variables described in Example 1 are further confirmed using an independent mass spectrometry system. Representative sample extracts from each variable group are re-analyzed by LC-MS using an HP 1050 high-performance liquid chromatography (HPLC), or equivalent, interfaced to an ABI Q-Star (Applied Biosystems Inc., Foster City, Calif.), or equivalent, mass spectrometer to obtain mass and intensity information for the purpose of identifying metabolites that differ in intensity between the clinical variables under investigation. This is also a non-targeted approach, which provides retention time indices (time it takes for metabolites to elute off the HPLC column), and allows for tandem MS structural investigation. In this case, to verify that the sample extracts from the ovarian cancer patients and the controls did indeed have differential abundances of said markers, selected extracts from each cohort were analyzed independently using said approach. Of the 37 said metabolites described previously, 29 were detected across a set of 10 ovarian cancer and 10 control samples. A PCA plot based on these 29 masses is shown in
The retention times of the 29 metabolites shown in
The following example describes the tandem mass spectrometry analysis of a subset of the ovarian markers. The general principle is based upon the selection and fragmentation of each of the parent ions into a pattern of daughter ions. The fragmentation occurs within the mass spectrometer through a process called collision-induced dissociation, wherein an inert gas (such as argon) is allowed to collide with the parent ion resulting in its fragmentation into smaller components. The charge will then travel with one of the corresponding fragments. The pattern of resulting fragment or “daughter ions” represents a specific “fingerprint” for each molecule. Differently structured molecules (including those with the same formulas) will produce different fragmentation patterns, and therefore represents a very specific way of identifying the molecule. By assigning accurate masses and formulas to the fragment ions, structural insights about the molecules can be determined.
In this example, MSMS analysis was carried out on a subset of 31 ovarian markers (from Tables 2 and 3). The resulting fragment ions for each of the selected parent ions are listed in Tables 4 through 34. The parent ion is listed at the top of each table (as its neutral mass), and the subsequent fragments shown as negatively charged ions [M-H]. The intensity (in counts and percent) is shown in the middle and right columns, respectively. The specific retention time (from the high performance liquid chromatography) is shown at the top of the middle column. The ovarian markers all had retention times under the chromatographic conditions used (see methods below) between 16 and 18 minutes.
Proposed structures based upon interpretation of the fragmentation patterns are summarized in Table 35. Subsequent Tables 36 through 65 list the fragment masses and proposed structures of each fragment for each parent molecule. The masses in the table are given as the nominal detected mass [M-H] and the proposed molecular formula is given for each fragment. In addition, the right-hand column indicates the predicted neutral fragment losses.
Interpretation of the MSMS data revealed that the metabolite markers are structurally related to the gamma-tocopherol form of vitamin E, in that they comprise a chroman ring-like moiety and phytyl side-chain. However, these molecules possess several important differences from gamma tocopherol:
a). omega-carboxylated phytyl sidechains (carboxylation at the terminal carbon position of the phytyl chain).
b). semi-saturated and open chroman ring-like systems
c). increased carbon number due to potential hydrocarbon chain addition to the ring system.
Based on the similarity to gamma-tocopherol and the presence of the omega-carboxyl moieties, the class of novel metabolites was named “gamma-tocoenoic acids.”
HPLC analysis were carried out with a high performance liquid chromatograph equipped with quaternary pump, automatic injector, degasser, and a
Hypersil ODS column (5 μm particle size silica, 4.6 i.d×200 mm) and semi-prep column (5 μm particle size silica, 9.1 i.d×200 mm), with an inline filter. Mobile phase: linear gradient H2O-MeOH to 100% MeOH in a 52 min period at a flow rate 1.0 ml/min.
Eluate from the HPLC was analyzed using an ABI QSTAR® XL mass spectrometer fitted with an atmospheric pressure chemical ionization (APCI) source in negative mode. The scan type in full scan mode was time-of-flight (TOF) with an accumulation time of 1.0000 seconds, mass range between 50 and 1500 Da, and duration time of 55 min. Source parameters were as follows: Ion source gas 1 (GS1) 80; Ion source gas 2 (GS2) 10; Curtain gas (CUR) 30; Nebulizer Current (NC)-3.0; Temperature 400° C.; Declustering Potential (DP)-60; Focusing Potential (FP)-265; Declustering Potential 2 (DP2)-15. In MS/MS mode, scan type was product ion, accumulation time was 1.0000 seconds, scan range between 50 and 650 Da and duration time 55 min. For MSMS analysis, all source parameters are the same as above, with collision energy (CE) of −35 V and collision gas (CAD, nitrogen) of 5 psi.
The following example describes the development of a high-throughput screening (HTS) assay based upon triple-quadrupole mass spectrometry for a subset of the ovarian markers. The preliminary method was initially established to determine the ratio of six of the ovarian 28-carbon containing metabolites to an internal standard molecule added during the extraction procedure. This is similar to the HTS method reported in applicant's co-pending CRC/Ovarian PCT application published on Mar. 22, 2007 (WO 2007/030928). The ability of this method to differentiate between ovarian cancer patients and subjects without ovarian cancer is shown in
Based upon completion of MSMS analysis of the remaining molecules, a new HTS triple-quadrupole method was developed to analyze a larger subset of the ovarian markers. This expanded triple-quadrupole method measures a comprehensive panel of the gamma Tocoenoic acids, and includes the metabolites listed in Table 67. The method measures the daughter fragment ion of each parent, as well an internal standard molecule (see methods below). The biomarker peak areas are then normalized by dividing by the internal standard peak areas.
The method was then used to validate the reduction of gamma tocoenoic acids in a subsequent independent population of controls and ovarian cancer positive subjects. The graph in
Serum samples are extracted as described for non-targeted FTMS analysis. The ethyl acetate organic fraction is used for the analysis of each sample. 15 uL of internal standard is added (1 ng/mL of (24-13C)-Cholic Acid in methanol) to each sample aliquot of 120 uL ethyl acetate fraction for a total volume of 135 uL. The autosampler injects 100 uL of the sample by flow-injection analysis into the 4000QTRAP. The carrier solvent is 90% methanol:10% ethyl acetate, with a flow rate of 360 uL/min into the APCI source.
The MS/MS HTS method was developed on a quadrupole linear ion trap ABI 4000QTrap mass spectrometer equipped with a TurboV™ source with an APCI probe. The source gas parameters were as follows: CUR: 10.0, CAD: 6, NC: −3.0, TEM: 400, GS1: 15, interface heater on. “Compound” settings were as follows: entrance potential (EP): −10, and collision cell exit potential (CXP): −20.0. The method is based on the multiple reaction monitoring (MRM) of one parent ion transition for each metabolite and a single transition for the internal standard. Each of the transitions is monitored for 250 ms for a total cycle time of 2.3 seconds. The total acquisition time per sample is approximately 1 min. The method is similar to that described in the PCT case referred to above (WO 2007/030928), but was expanded to include a larger subset of the molecules as shown in Table 67.
All citations are hereby incorporated by reference.
The present invention has been described with regard to one or more embodiments. However, it will be apparent to persons skilled in the art that a number of variations and modifications can be made without departing from the scope of the invention as defined in the claims.
Biol Chem 2003; 278: 43508-15.
Each table shows the collision energy in voltage, the HPLC retention time in minutes and the percent intensity of the fragment ion. Masses in the title of the table are neutral, while the masses listed under m/z (amu) are [M-H] and correspond to units in Daltons.
This application is a divisional of U.S. patent application Ser. No. 12/524,641, which is a national stage application under 35 U.S.C. 371 of PCT/CA2008/000270, filed Feb. 1, 2008, and claims the benefit of U.S. Provisional Patent Application Ser. No. 60/887,693, filed Feb. 1, 2007.
| Number | Date | Country | |
|---|---|---|---|
| 60887693 | Feb 2007 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 12524641 | Oct 2009 | US |
| Child | 14482158 | US |
