WO2015109116A1

WO2015109116A1 - Metabolic screening for gestational diabetes

Info

Publication number: WO2015109116A1
Application number: PCT/US2015/011645
Authority: WO
Inventors: Brian J. KOOS
Original assignee: The Regents Of The University Of California
Priority date: 2014-01-15
Filing date: 2015-01-15
Publication date: 2015-07-23
Also published as: EP3094968B1; US20160341739A1; EP3094968A1; EP3094968A4

Abstract

The invention provides a method for screening for and detection of diabetes mellitus in a subject that comprises assaying a test sample of urine from the subject for a metabolic marker of diabetes. An elevated or reduced amount of marker present in the test sample compared to a control sample is indicative of diabetes. The method can be used to screen for gestational diabetes early in pregnancy, or to detect diabetes or susceptibility to diabetes, in pregnant or non-pregnant subjects.

Description

METABOLIC SCREENING FOR GESTATIONAL DIABETES

[0001] This application claims the benefit of United States provisional patent application number 61/927,657, filed January 15, 2014, the entire contents of which are incorporated herein by reference. TECHNICAL FIELD OF THE INVENTION

[0002] The present invention relates generally to detection, diagnosis, and monitoring of diabetes mellitus, including gestational diabetes. The invention more specifically pertains to use of metabolic markers that can be detected in urine to screen for diabetes.

BACKGROUND OF THE INVENTION

[0003] The identification of gestational diabetes mellitus (GDM) currently involves either (1 ) universal screening by a 2-h, 75-g oral glucose tolerance test (OGTT) or a 1-h, 50-g oral glucose challenge test (OGCT) followed by a 3-h, 100-g oral glucose tolerance test (OGTT) for those with a positive OGCT, or (2) selective screening of high-risk groups based on age, race/ethnicity, body mass index (BMI), and family history. Venapuncture is required for both the OGCT and OGTT. The National Institutes of Health Consensus Development

Conference Statement provides the current standards for diagnosing GDM (Obstet Gynecol 122:358-370, 2013). The use of glucose challenges to screen and diagnose GDM is limited by a lack of reproducibility and accuracy, excessive time involved, late gestational age (24- 28 weeks of gestation) of testing, nausea, and the discomfort of venapunture (Brody et al. Obstet Gynecol 2003;101 :380-392; Blayo. Diabetes Metab 2004;30:575-580; Sacks et al. Am J Obstet Gynecol 1989;161 :642-645; Hanna, Diabet Med 2002:19:351-358).

[0004] The usefulness of serum/plasma markers for GDM have been examined, such as fasting glucose and insulin, C-reactive protein, sex hormone-binding globulin, adiponectin, and protein-associated plasma protein A (Smirnakis, et al. Am J Obstet Gynecol

2007;196:410e1 -410e7; Georgiou et al., Acta Diabetol 2008;45:157-165; Kulaksizglu S. et al. Gynecol Endocrinol 2013;29:137-140). While significant differences have been found in those with GDM, these measurements lack sensitivity and specificity compared to the standard OGCT and OGTT methods.

[0005] There is a need to identify improved markers for gestational diabetes. There is also a need for methods of detecting susceptibility to diabetes.

SUMMARY OF THE INVENTION

[0006] The invention provides a method of screening for susceptibility to diabetes in a subject. In a typical embodiment, the method comprises measuring the amount of a metabolic marker present in a test sample obtained from the subject, and comparing the amount of the metabolic marker present in the test sample to a control sample. The method further comprises identifying a subject as susceptible to diabetes if the amount of marker present in the test sample is increased or decreased relative to the control sample.

Metabolic markers to be measured are selected from the group consisting of the markers listed in Table 1. In one embodiment, the measuring comprises chromatography or spectrometry. The chromatography can be gas or liquid chromatography. The spectrometry can be, for example, mass spectrometry.

[0007] In one embodiment, the subject is pregnant, and the screening is for gestational diabetes mellitus (GDM). The method is typically performed when the subject is at least 6 weeks pregnant. In some embodiments, the method is performed when the subject is 6-38 weeks pregnant, typically when the subject is 6-14 weeks pregnant.

[0008] Testing can also be performed later in pregnancy, such as at about 22-30 weeks gestation, at term, or postpartum. The method can also be used in other populations as a general method of screening for diabetes and susceptibility to diabetes, such as type 2 diabetes mellitus.

[0009] The invention also provides a method of screening for susceptibility to diabetes in a subject, wherein the method comprises measuring the amount of at least two metabolic markers present in a test sample obtained from the subject. The method further comprises comparing the amount of the metabolic markers present in the test sample to a control sample, and identifying a subject as susceptible to diabetes if the amount of the markers present in the test sample is increased or decreased relative to the control sample. Similarly, the invention provides a method of detecting diabetes in a subject. The method comprises measuring the amount of a metabolic marker present in a test sample obtained from the subject; and comparing the amount of the metabolic marker present in the test sample to a control sample. A subject is identified as having diabetes if the amount of marker present in the test sample is increased or decreased relative to the control sample. The metabolic markers are selected from the group consisting of the markers listed in Table 1. In some embodiments, the markers are selected from the group of markers listed in Table 2. In some embodiments, the markers are selected from the group of markers listed in Table 3.

[0010] In some embodiments of the invention, two markers selected from those described herein are measured. In other embodiments, three or more markers are measured.

Additional markers can be used to improve the screening and detection, including any combination of two, three, four, five, six, seven, or more of the metabolic markers described herein. In one embodiment, a combination of 8-1 1 markers is used together for screening. Representative combinations of the metabolic markers include:

Anserine and methylsuccinate;

Sucrose and 1-methylhistidine;

Xylonate and tiglyl carnitine,

Quinolinate and pro hydroxy pro,

N-acetylthreonine and carnitine,

Trigonelline and sucrose;

Trigonelline and alpha-CEHC glucuronide;

Pyroglutamine and 2-hydroxyisobutyrate;

Methylsuccinate, anserine and 1-methylhistidine;

Methylsuccinate, sucrose and 1-methylhistidine;

Sucrose, anserine and 1-methylhistidine;

Adipate, cystathionine, and cytidine;

Pyroglutamine, anserine, and 2-hydroxyisobutyrate;

Anserine, cytidine, and cystathionine;

Adipate, pyroglutamine, and cystidine; and

Methylsuccinate, sucrose, anserine, and 1-methylhistidine.

[0011] Other combinations of metabolic markers selected from the group consisting of: 1- methylhistidine, 2-hydroxyisobutyrate, 3-(3-hydroxyphenyl)proprionate Itaconate, 3- fucosyllactose, 3-hydroxy-3-methylglutarate, 5-methylthioadenosine, Acetylcarnitine, Adipate, Agmatine, Alpha-CEHC glucuronide, Anserine, Carnitine, Cystathionine, Cytidine, Dihydrobiopterin, Galactose, Gamma-CEHC glucuronide, Gluconate, Glutamate, Glutarate, Glycocholate, Homoserine, Hydroxyphenylacetate, Lactose, Leucine, Methylsuccinate, N- acetylarginine, N-acetylthreonine, Nicotinate, Pro-hydroxy-pro, Pyroglutamine, Quinolinate, Scyllo-inositol, Sorbose, Sucrose, Thymine, Tigyl carnitine, Trigonelline, and Xylonate are also contemplated.

DETAILED DESCRIPTION OF THE INVENTION

[0012] The invention described herein is based on the discovery that certain metabolic markers can be used to detect, diagnose and monitor diabetes, including gestational diabetes mellitus, as well as to guide in the monitoring and selection of treatment. Testing early in pregnancy is desirable, as the early initiation of counseling and therapeutic measures can be reasonably expected to have the most beneficial effects in reducing perinatal morbidity. The availability of an accurate test that can be used in early pregnancy provides an important advance in the screening and diagnosis of GDM. This noninvasive test has major advantages in that it is painless and the sample is easily collected by medical assistants. The gravida's urine can be collected randomly at a routine prenatal visit and can be sent to a commercial laboratory for analysis. The collection does not involve a highly trained, expensive nurse or phlebotomist, as is required for blood sampling. This additional advantage will allow GDM testing in clinics with restricted resources for patient care. Another advantage is that a single urine test could replace GDM screening by the 1-h, 50-g OGCT and offer sufficient accuracy to eliminate the need for the painful, time-consuming 2-h, 75-g OGTT or 3-h 100-g OGTT.

Definitions

[0013] All scientific and technical terms used in this application have meanings commonly used in the art unless otherwise specified. As used in this application, the following words or phrases have the meanings specified.

[0014] As used herein, a "sample" from a subject means a specimen obtained from the subject that contains urine, blood, serum, saliva, or other bodily fluid.

[0015] As used herein, the term "subject" includes any human or non-human animal. The term "non-human animal" includes all vertebrates, e.g., mammals and non-mammals, such as non-human primates, horses, sheep, dogs, cows, pigs, chickens, amphibians, reptiles, rodents etc.

[0016] As used herein, a "control" sample means a sample that is representative of normal measures of the respective analyte. The sample can be an actual sample used for testing, or a reference level or range, based on known normal measurements of the corresponding analyte.

[0017] As used herein, "a" or "an" means at least one, unless clearly indicated otherwise. Methods

[0018] The invention provides a method of screening for susceptibility to diabetes in a subject. The method is particularly useful for identifying susceptibility to gestational diabetes mellitus (GDM) during early pregnancy, including during the first trimester of pregnancy as well as later in pregnancy and postpartum. The invention also provides a method of detecting diabetes in a subject. In a typical embodiment, the method comprises measuring the amount of one, two, three or more metabolic markers present in a test sample obtained from the subject, and comparing the amount of the metabolic marker present in the test sample to a control sample. The method further comprises identifying a subject as susceptible to, or as having, diabetes if the amount of marker present in the test sample is increased or decreased relative to the control sample. [0019] An amount of marker is considered increased or decreased if it differs by a statistically significant amount from the amount present in the control. In some embodiments, the difference is an increase or decrease of at least 10%; in other embodiments, the difference is at least 20%, 30%, 40%, 50% or more. In other embodiments, a reference range has been identified for the amount of the metabolic marker present in a normal, control sample, and a test sample having an amount of the marker that is outside the reference range for that marker is susceptible to diabetes. In a typical embodiment, the sample is a urine sample. Other bodily fluids can be used for the sample, including serum and saliva.

[0020] Metabolic markers to be measured are listed in Tables 1 , 2, and 3, and in some embodiments, are selected from the group consisting of: adipate, methylsuccinate, pyroglutamine, cytidine, 1-methylhistidine, 2-hydroxyisobutyrate, cystathionine, sucrose and anserine. Additional markers useful for the methods of the invention include: tiglyl carnitine, xylonate, quinolinate, pro hydroxy pro, N-acetylthreonine, trigonelline, and alpha-CEHC glucuronide. Other groups of markers from which to select are described in the examples below, and/or are based on the gestational stage of the subject, the normalization used, and/or the method of analysis to be employed. For example, markers for use in a method of detecting diabetes or susceptibility to diabetes during the first trimester of pregnancy can be selected from the group consisting of: adipate, methylsuccinate, pyroglutamine, cytidine, 1- methylhistidine, 2-hydroxyisobutyrate, cystathionine, sucrose, glutamate, galactose, 3- hydroxy-3-methylglutarate, and anserine. Representative markers useful for the methods of the invention relating to pregnant subjects at 24-38 weeks gestation include: tiglyl carnitine, xylonate, quinolinate, pro hydroxy pro, N-acetylarginine, itaconate, leucine, scyllo-inositol, thymine, carnitine, acetylcarnitine, agmatine, sorbose, N-acetylthreonine, and trigonelline, Representative markers useful for the methods of the invention relating to subjects who are 4-12 weeks postpartum include: quinolinate, pro hydroxy pro, N-acetylarginine, trigonelline, lactose, 4-hydroxyphenylacetate, gamma-CEHC glucuronide, glutarate, dihydrobiopterin, sucrose, sarcosine, glycocholate, homoserine, pyrogluatime, 5-methylthioadenosine, nicotinate, and alpha-CEHC glucuronide. These postpartum markers can also serve as markers for non-pregnant subjects.

[0021] Examples of markers for which an increase relative to control is indicative of diabetes are listed in Example 3 below; and examples of markers for which a decrease relative to control is indicative of diabetes are also listed in Example 3 below. As noted in Example 3, the markers that increase or decrease with GDM relative to normal are different at 12 weeks gestation, 24 weeks gestation, and postpartum. As shown in the examples below, more than one marker may be used in combination to identify the presence of GDM. Table 1

3-(3-hydroxyphenyl)proprionate

Acetylcarnitine Itaconate

Ad i pate Lactose

Agmatine 25 Leucine

Anserine 1 - methylhistidine

Carnitine Methylsuccinate

Cystathionine 5-methylthioadenosine Cytidine N-acetylarginine

Dihydrobiopterin 30 N-acetylthreonine

3-fucosyllactose Nicotinate

Galactose Pro-hydroxy-pro

Gamma-CEHC glucuronide Pyroglutamine

Gluconate Quinolinate

Glutamate 35 Scyllo-inositol

Glutarate Sorbose

Glycocholate Sucrose

Homoserine Thymine

2- hydroxyisobutyrate Tiglyl carnitine

3-hydroxy-3-methylglutarate 40 Trigonelline

hydroxyphenylacetate Xylonate

Table 2

Alpha-CEHC glucuronide 50 Pro hydroxy pro

Anserine 2- hydroxyisobutyrate

Methylsuccinate N-acetylthreonine

Xylonate Carnitine

Tigyl carnitine Trigonelline

Quinolinate

Table 3

3- fucosyllactose

3-(3-hydroxyphenyl)proprionate

Itaconate

1-methylhistidine

Quinolinate

Sucrose [0022] In one embodiment, the measuring comprises chromatography or spectrometry. The chromatography can be gas or liquid chromatography. The spectrometry can be mass spectrometry. Other known methods of metabolic marker detection are also contemplated, and may be selected based on the characteristics of the individual marker of interest.

Examples of other assays that can be employed include immunoassay and electrochemical detection. Measures of test samples can be compared directly to controls, such as comparing the concentration of the metabolic marker present in the test sample to the concentration of the metabolic marker in a control sample or to a known normal

concentration of the metabolic marker. Alternatively, in some embodiments, the metabolite concentration can be normalized with respect to a selected normalizing marker, such as creatinine or osmolality.

[0023] In one embodiment, the subject is pregnant, and the screening is for gestational diabetes mellitus (GDM). The method is typically performed when subject is at least 6 weeks pregnant. In some embodiments, the method is performed when the subject is 6-38 weeks pregnant, typically when the subject is 6-14 weeks pregnant.

[0024] Testing can also be performed later in pregnancy, such as at about 22-30 weeks gestation, or at term. In some embodiments, testing is performed for patients contemplating pregnancy. It can also be used in nonpregnant populations as a general method of screening for diabetes and susceptibility to diabetes, such as type 2 diabetes mellitus.

Kits

[0025] For use in the diagnostic applications described herein, kits are also within the scope of the invention. Such kits can comprise a carrier, package or container that is compartmentalized to receive one or more containers such as vials, tubes, and the like, each of the container(s) comprising one of the separate elements to be used in the method. The probes, antibodies and other reagents of the kit may be provided in any suitable form, including frozen, lyophilized, or in a pharmaceutically acceptable buffer such as TBS or PBS. The kit may also include other reagents required for utilization of the reagents in vitro or in vivo such as buffers (i.e., TBS, PBS), blocking agents (solutions including nonfat dry milk, normal sera, Tween-20 Detergent, BSA, or casein), and / or detection reagents (i.e., goat anti-mouse IgG biotin, streptavidin-HRP conjugates, allophycocyanin, B-phycoerythrin, R- phycoerythrin, peroxidase, fluors (i.e.,

DyLight, Cy3, Cy5, FITC, HiLyte Fluor 555, HiLyte Fluor 647), and / or staining kits (i.e., ABC Staining Kit, Pierce)). The kits may also include other reagents and / or instructions for using antibodies, probes, and other reagents in commonly utilized assays described above such as, for example, liquid or gas chromatography, spectrometry, electrochemical assay, flow cytometric analysis, ELISA, immunoblotting (i.e., western blot), immunocytochemistry, immunohistochemistry.

[0026] In one embodiment, the kit provides the reagent in purified form. In another

embodiment, the reagents are immunoreagents that are provided in biotinylated form either alone or along with an avidin-conjugated detection reagent (i.e., antibody). In another embodiment, the kit includes a fluorescently labeled immunoreagent which may be used to directly detect antigen. Buffers and the like required for using any of these systems are well- known in the art and may be prepared by the end-user or provided as a component of the kit. The kit may also include a solid support containing positive- and negative-control protein and / or tissue samples. For example, kits for performing spotting or western blot-type assays may include control cell or tissue lysates for use in SDS-PAGE or nylon or other membranes containing pre-fixed control samples with additional space for experimental samples.

[0027] The kit of the invention will typically comprise the container described above and one or more other containers comprising materials desirable from a commercial and user standpoint, including buffers, diluents, filters, needles, syringes, and package inserts with instructions for use. In addition, a label can be provided on the container to indicate that the composition is used for a specific application, and can also indicate directions for use, such as those described above. Directions and or other information can also be included on an insert which is included with the kit.

[0028] In one embodiment, the kit comprises a set of assay reagents suitable for detecting and/or measuring alpha-CEHC glucuronide, anserine, methylsuccinate, xylonate, tigyl carnitine, quinolinate, pro hydroxy pro, 2-hydroxyisobutyrate, N-acetylthreonine, carnitine, and trigonelline, or any other combination of markers described herein.

EXAMPLES

[0029] The following examples are presented to illustrate the present invention and to assist one of ordinary skill in making and using the same. The examples are not intended in any way to otherwise limit the scope of the invention.

Example 1 : Metabolomic Screening for Gestational Diabetes

[0030] This example demonstrates that urinary metabolite concentrations can be used to identify pregnant women who will develop gestational diabetes mellitus (GDM).

Methods

[0031] Urine was collected from pregnant women from 6-38 weeks of gestation. Gestational diabetes was determined by standard clinical screening at 24-28 weeks of gestation: 50-g 1- h oral glucose challenge test (OGCT) followed by a 100-g 3-h oral glucose challenge test (OGTT) for those with an OGCT plasma glucose concentration of≥140 mg/dl. The results of the OGCT and OGTT separated the gravidas into two groups: normal gravidas (NG) and those with GDM.

Metabolon (Durham, NC) analyzed the urine samples on gas chromatography/mass spectrometry and liquid chromatography/mass spectrometry platforms for 441 low-molecular weight metabolites. Metabolite concentrations and metabolite levels normalized to urinary creatinine or osmolality were analyzed. The top 37 metabolites by random forest analysis that were found to have the most highly significantly different levels in GDM compared to NG were the following:

Creatinine Normalized Osmolality Normalized

6-17 weeks' gestation

Ad i pate 2-hydroxyisobutyrate

Methylsuccinate Galactose

Pyroglutamine 1 -methylhistidine

Cytidine Sucrose

1- methylhistidine 3-hydroxy-3-methylglutarate

2- hydroxyisobutyrate Gluconate

Glutamate Anserine

Cystathionine Methylsuccinate

Anserine

24-38 weeks' gestation

N-acetylarginine N-acetylthreonine

Xylonate Scyllo-inositol

Itaconate Thymine

Tiglyl carnitine Carnitine

Leucine Acetylcarnitine

Agmatine Sorbose

Sorbose

4-12 weeks postpartum

Quinolinate Trigonelline

Pro-hydroxy-pro Glycocholate

Lactose Homoserine

4-hydroxyphenylacetate Pyroglutamine

Gamma-CEHC glucuronide Sucrose Glutarate 5-methylthioadenosine

N-acetylarginine Alpha-CEHC glucuronide

Dihydrobiopterin Nicotinate

Trigonelline

Sucrose

Sarcosine

[0032] Further statistical scrutiny of these 37 candidate metabolites to distinguish GDM from NG by simultaneous classification tree analysis revealed the following optimal sensitivity, accuracy, and ROC area under the curve (AUC) using 1-2 metabolites:

Creatinine normalization

Osmolality normalization

GA, gestational age; PP, postpartum

Interpretation

[0033] Eleven metabolites have been identified that, in various combinations, have a high degree of accuracy in identifying gravidas who will develop gestational diabetes, as diagnosed by current clinical practice. In this study, the urinary analysis had up to 100% sensitivity, specificity and accuracy in early pregnancy (6-17 weeks). The detection of GDM in early pregnancy (-0.25 term) enables the clinical management of these patients (e.g., counseling, life-style changes, and glucose monitoring) long before GDM is normally diagnosed by current screening methods at about 26-30 weeks of gestation (-0.75 term). Commencing GDM surveillance and management at 10-12 weeks of gestation will likely reduce maternal and fetal and maternal morbidity related to hyperglycemia of GDM. [0034] Urinary metabolites were also useful in later gestation (-0.75 term) in that it had 92% sensitivity for indentifying GDM. Thus, a positive urinary metabolite screen would potentially eliminate the need for further GDM testing.

[0035] In postpartum (nonpregnant) women, the urinary metabolite screen identified all women who had had GDM (sensitivity and specificity up to 100%). These results suggest that such urinary metabolite testing would be useful for women who are considering pregnancy. A positive preconception screen would enable appropriate counseling and dietary alterations before pregnancy.

[0036] The ability of urinary metabolites to identify postpartum women (~6 weeks after delivery) who had GDM suggests that urinary metabolite screening would also be useful in identifying those in the general population who are either disposed to or actually have type 2 diabetes mellitus.

[0037] Metabolomic analysis have been performed on urine in pregnant women. Urine has particular advantages over blood tests because it is readily available, collected without discomfort to the patient, and does not require expensive skilled personnel for collection. But other body fluids can also be sampled for metabolomic analysis:

[0038] a) Blood. Pregnant women routinely have blood tests as part of the first prenatal visit. Therefore, metabolomic analysis of plasma samples should be done to establish whether plasma levels of these 1 1 or more (up to 37 total) metabolites are useful for identifying those who will develop GDM. Nonpregnant subjects commonly have screening blood tests (e.g., complete blood count, hemoglobin A1 C, glucose) as part of routine health care. Thus, metabolomic analysis of plasma levels of these 1 1 or more (up to 37 total) metabolites should be done to determine whether these blood metabolites are useful in identifying those who are either disposed to type 2 diabetes mellitus or have the disorder.

[0039] b) Saliva. Another body fluid that is easily sampled is saliva. Thus, metabolomic analysis of saliva should be performed to determine whether these 1 1 or more (up to 37 total) metabolites in saliva can predict gravidas who will develop GDM as well as nonpregnant subjects disposed to type 2 diabetes mellitus.

Example 2: Statistical Analysis of Metabolites Associated With Gestational Diabetes

[0040] In this example, the metabolites measured at -12 weeks (6-17 weeks gestation), -27 weeks (24-32 weeks gestation) and -6 weeks post partum were explored using both creatinine normalized and osmolality normalized data. The results show that the optimal metabolites for separating GDM from normals in this study differ by group. Thus, for each week and normalization method, a different subset of 8 to 1 1 metabolites from the 441 metabolites were used based on random forest analysis. There are 6 analyses (3 collection times with 2 normalization methods) using up to 1 1 metabolites in any one analysis.

Statistical methods

[0041] The statistical methods are the same as described in Example 1 above.

[0042] Univariate - A nonparametric ROC was carried out on each metabolite separately to identify the best threshold, sensitivity, specificity and unweighted accuracy for each.

Unweighted accuracy is defined as

Unweighted accuracy = (sensitivity + specificity)/2

[0043] Means and standard deviations by group are reported on the log (base e) scale since the distribution of the metabolites is much closer to a normal (Gaussian) distribution on the log scale. Thus the original scale GDM/normal mean ratio is computed as the antilog of the log scale mean differences and is also the ratio of geometric means, NOT the ratio of arithmetic means. The geometric mean is also the theoretical median when the log scale data follow a normal distribution.

(log (GDM/normal)) = log(GDM) - log(normal) , GDM/normal=exp(log(GDM) - log(normal))

[0044] To be conservative, p values were computed using the nonparametric Wilcoxon rank sum test.

[0045] Multivariate - The nonparametric classification tree model was used to simultaneously evaluate all 8 metabolites and the corresponding sensitivity, specificity and accuracy is reported.

Results:

[0046] Univariate - The tables below show the univariate log scale means, standard deviations, thresholds, sensitivity, specificity and unweighted accuracy by normalization method and time (12 weeks, 24 weeks, post partum). The original scale GDM/normal geometric mean ratio is also reported. For example, at 12 weeks using creatinine normalization, methylsuccinate provides the highest unweighted nominal accuracy of 90.6% and has a GDM/normal mean ratio of 3.1 on the original scale. Using osmo normalized data at 12 weeks, 2 hydroxyisobutyrate provides the highest unweighted nominal accuracy of 85.8% and has a mean ratio of 1.47. Univariate metabolite comparison: GDM vs normal, log e scale nominal specificity (spec) and sensitivity (sens). P values from Wilcoxon rank sum test

* original scale geometric mean ratio

Creatinine normalized, 12 wks, 9 GDM, 16 normals

! 0.35 \ 0.024 0.65

Osmolality normalized, 12 wks, 9GDM, 16 normals

Creatinine normalized, 24 wks, 12 GDM, 16 normals

Osmolality normalized, 24 wks, 12 GDM, 16 normals

Creatinine normalized, post partum, 12 GDM, 16 normals

[0047] Multivariate tree - The multivariate tree results were analyzed and reviewed by normalization method and time. The results including sensitivity, specificity and unweighted accuracy are summarized in the tables below.

[0048] At week 12, using the creatinine normalized data, the tree uses only anserine and methylsuccinate and creates four classification groups (terminal tree nodes). The nominal accuracy is 100%. Using the osmo-normalized data at week 12, the tree only uses 2- hydroxyisobutyrate (same as in the univariate results) and creates two classification groups. This tree correctly classifies 15 of 16 normals (94%) and 7 of 9 GDM (78%) for a nominal unweighted accuracy of (94% + 78%)/2 = 86%.

Creatine normalization - with xenobiotics

Osmo normalization - with xenobiotics

not including root level

Best metabolites in Wulff report - 12 - with xenobiotics

Creatine normalization - without xenobiotics

Osmolality normalization - without xenobiotics

^* not including root level

Best metabolites in Wulff re ort - 12 - without xenobiotics

Example 3: Notable Metabolites by Random Forest Analysis

[0049] This example provides a list of metabolite markers of interest, as determined from the random forest analysis. The arrows indicate whether the levels were higher or lower than in normals. This listing is followed by a series of tables that provide the tree with cutoffs and key metabolites.

6-17 Weeks Pregnant

[0050] Osmolality Normalized GDM/Normal Mean Ratio

2- hydroxyisobutyrate †

Galactose I

1-methylhistidine †

Sucrose I

3- hydroxy-3-methylglutarate †

Gluconate †

Anserine †

Methylsuccinate †

[0051] Creatinine Normalized

Adipate †

Methylsuccinate †

Pyroglutamine j

Cytidine †

1-methylhistidine †

2-hydroxyisobutyrate †

Glutamate †

Cystathionine †

Anserine † 24-38 Weeks Pregnant

[0052] Osmolality Normalized GDM/Normal Mean Ratio

N-acetylthreonine Ϊ

Scyllo-inositol i

Thymine i

Carnitine i

Acetylcarnitine i

Sorbose † [0053] Creatinine Normalized

N-acetylarginine †

Xylonate †

Itaconate †

Tiglyl carnitine †

Leucine †

Agmatine I

Sorbose † POSTPARTUM

[0054] Osmolality Normalized GDM/Normal Mean Ratio

Trigonelline I

Glycocholate I

Homoserine I

Pyroglutamine I

Sucrose I

5-methylthioadenosine I

Alpha-CEHC glucuronide †

Nicotinate I

[0055] Creatinine Normalized

Quinolinate †

Pro-hydroxy-pro †

Lactose †

4-hydroxyphenylacetate †

Gamma-CEHC glucuronide †

Glutarate †

N-acetylarginine †

Dihydrobiopterin †

Trigonelline I

Sucrose I

Sarcosine †

Cross Tables From Tree Analyses

creatinine normalization, 12 weeks

methylsuccinate Normal GDM GDM

>=0.9294

methylsuccinate Normal normal GDM

<0.9294

[0056] Anserine is higher in GDM compared to normals. Methylsuccinate is higher in GDM compared to normals in those with anserine between 0.405 and 14.32. creatinine normalization, 24 weeks

[0057] Both xylonite and tiglyl carnitine are higher in GDM compared to normal. Normals are lower on both Xylonate and tigly carnitine. creatinine normalization, post partum

[0058] If Quinolinate >= 0.9202, pro hydroxyl pro is higher in GDM than normals. If pro hydoxy pro >=1.0385, quinolinate is higher in GDM than normals. Those with GDM are higher on both Quinolinate and pro hydroxy pro. osmo normalization, 12 weeks

2. hydroxyisobutyrate is higher in GDM compared to normals.

osmo normalization, 24 week s

[0059] N.acethythreonine is lower in GDM compared to normals in those with Carnitine >=0.81 13.

[0060] Normals are higher on both N.acethythreonine and Carnitine. osmo normalization, post partum

trig = trigonelline N methylnicotinate

alpha CEHC = alpha CEHC glucuronide

[0061] Those with GDM have both lower trig and higher alpha CEHC compared to normals.

Example 4: Metabolic Signature of Gestational Diabetes in First Trimester [0062] This example describes the distinctive metabolic profile that can be observed in early pregnancy for gravidas who develop GDM.

[0063] Study Design: Urine samples were collected at <14 weeks' gestation from 370 healthy gravidas with singletons. Abnormal OGCTs (24-28 wks) were followed by 100-g OGTT. Nine GDM subjects were matched with 16 normal gravidas (NG) . Urine samples were analyzed on GC/MS and LC/MS/MS platforms by Metabolon for 441 low molecular weight metabolites. Scaled biochemical levels were normalized separately to creatinine or osmolality. Random forests (RF) analysis identified 8 biochemicals that differed most significantly between GDM and NG. A nonparametric ROC identified the optimal threshold, sensitivity and specificity for each metabolite in predicting GDM. A non-parametric classification tree model (CART) was used to evaluate simultaneously the 8 metabolites and compute the corresponding sensitivity and specificity. P-values were calculated by unpaired t-test and Fisher's exact test where appropriate. Results are expressed as means ± SE.

[0064] Results: NG and GDM had similar age (NG: 33 ± 1.2; GDM: 34 ± 1.1 years), BMI (NG: 25.6 ± 1.4; GDM: 28.4 ± 2.3 kg/m2), ethnicity, parity, and GA at urine collection (NG:8.8 ± 0.7; GDM: 10.1 ± 0.9 weeks). CART analysis for the top 8 creatinine-normalized metabolites that differed between NG and GDM showed that 100% sensitivity and specificity (no misclassification) was obtained using a decision rule based on anserine and

methylsuccinate. CART analysis of these same metabolites referenced to osmolality revealed a 100% sensitivity and 88% specificity (2 normals misclassified) using a decision rule based on anserine and sucrose.

[0065] Conclusion: RF and CART were surprisingly accurate in identifying critical metabolites that separate NG and GDM in early pregnancy. Example 5: Metabolic Signature of Gestational Diabetes Throughout Pregnancy

[0066] This example describes the distinctive metabolic profile that can be observed in different stages of pregnancy for gravidas who develop GDM.

[0067] Methods: Fed-state urine samples were collected at 6-14 weeks' and 22-32 weeks' gestation and 6 weeks postpartum from 375 healthy gravidas with singletons. Abnormal 50-g glucose challenge tests (GCT) at 24-28 weeks' gestation were followed by 100-g GTT. Nine GDM subjects were matched with 16 normal gravidas (NG). Urine samples were analyzed on GC/MS and LC/MS/MS platforms for 441 metabolites (Metabolon). Scaled metabolite levels were normalized to creatinine. Random forest analysis identified 8 biochemicals that most distinguished GDM from NG. ROC analysis identified optimal threshold, sensitivity, and specificity of each metabolite for predicting GDM. Classification tree model (CART) was used to compute simultaneously the corresponding sensitivity and specificity for each metabolite. P-values were calculated by unpaired t-test and Fisher's exact test where appropriate.

[0068] Results: GDM and NG were similar in age, ethnicity, parity, and BMI. Results of CART analysis of the 8 most discriminating metabolites for predicting GDM are shown in the following table.

[0069] Conclusions: The most discriminating biochemicals in tree analysis differ according to gestational age with a similar high accuracy in predicting GDM regardless of the time of urine collection. [0070] Throughout this application various publications are referenced. The disclosures of these publications in their entireties are hereby incorporated by reference into this application in order to describe more fully the state of the art to which this invention pertains.

[0071] From the foregoing it will be appreciated that, although specific embodiments of the invention have been described herein for purposes of illustration, various modifications may be made without deviating from the spirit and scope of the invention. Accordingly, the invention is not limited except as by the appended claims.

Claims

What is claimed is:

1. A method of screening for susceptibility to diabetes in a subject, the method comprising:

(a) measuring the amount of a metabolic marker present in a test sample obtained from the subject;

(b) comparing the amount of the metabolic marker present in the test sample to a control sample;

(c) identifying a subject as susceptible to diabetes if the amount of marker present in the test sample is increased or decreased relative to the control sample; wherein the metabolic marker is selected from the group consisting of: 3- fucosyllactose, 3-(3-hydroxyphenyl)proprionate, Itaconate, 1 -methylhistidine, Quinolinate, and Sucrose.

2. The method of claim 1 , wherein the measuring comprises chromatography or spectrometry.

3. The method of claim 2, wherein the chromatography is gas or liquid chromatography.

4. The method of claim 2, wherein the spectrometry is mass spectrometry.

5. The method of claim 1 , wherein the subject is 6-38 weeks pregnant.

6. The method of claim 1 , wherein the subject is 6-14 weeks pregnant.

7. The method of claim 1 , wherein the subject is postpartum.

8. The method of claim 1 , wherein the metabolic marker is normalized to urinary creatinine or osmolality.

9. A method of screening for susceptibility to diabetes in a subject, the method comprising:

(a) measuring the amount of at least two metabolic markers present in a test sample obtained from the subject;

(c) identifying a subject as susceptible to diabetes if the amount of the markers

present in the test sample is increased or decreased relative to the control sample; wherein the metabolic markers are selected from the group consisting of: alpha- CEHC glucuronide, anserine, methylsuccinate, xylonate, tigyl carnitine, quinolinate, pro hydroxy pro, 2-hydroxyisobutyrate, N-acetylthreonine, carnitine, and trigonelline.

10. The method of claim 9, wherein the metabolic markers are anserine and methyl succinate.

1 1 . The method of claim 9, wherein the metabolic markers are pyroglutamine and 2- hydroxyisobutyrate.

12. The method of claim 9, wherein at least three markers are measured.

13. The method of claim 12, wherein the metabolic markers are adipate, cystathionine, and cytidine.

14. The method of claim 12, wherein the metabolic markers are pyroglutamine, anserine, and 2-hydroxyisobutyrate.

15. The method of claim 12, wherein the metabolic markers are anserine, cytidine, and cystathionine.

16. The method of claim 12, wherein the metabolic markers are adipate, pyroglutamine, and cystidine.

17. The method of any of the preceding claims wherein the diabetes is diabetes mellitus.

18. A method of detecting diabetes in a subject, the method comprising:

(c) identifying a subject as having diabetes if the amount of marker present in the test sample is increased or decreased relative to the control sample; wherein the metabolic marker is selected from the group consisting of: 1 - methylhistidine, 2-hydroxyisobutyrate, 3-(3-hydroxyphenyl)proprionate Itaconate, 3- fucosyllactose, 3-hydroxy-3-methylglutarate, 5-methylthioadenosine, Acetylcarnitine, Adipate, Agmatine, Alpha-CEHC glucuronide, Anserine, Carnitine, Cystathionine, Cytidine, Dihydrobiopterin, Galactose, Gamma-CEHC glucuronide, Gluconate, Glutamate, Glutarate, Glycocholate, Homoserine, Hydroxyphenylacetate, Lactose,

Leucine, Methylsuccinate, N-acetylarginine, N-acetylthreonine, Nicotinate, Pro- hydroxy-pro, Pyroglutamine, Quinolinate, Scyllo-inositol, Sorbose, Sucrose, Thymine, Tigyl carnitine, Trigonelline, and Xylonate.

19. The method of any of the preceding claims wherein the marker is any one, or any combination of two or more of the markers selected from the group consisting of: alpha- CEHC glucuronide, anserine, methylsuccinate, xylonate, tigyl carnitine, quinolinate, pro hydroxy pro, 2-hydroxyisobutyrate, N-acetylthreonine, carnitine, trigonelline, 3- fucosyllactose, 3-(3-hydroxyphenyl)proprionate, itaconate, 1-methylhistidine, and sucrose.

20. The method of any of the preceding claims wherein test sample is urine.