WO2011068956A2 - Follistatin-like 3 (fstl3) levels in gestational diabetes - Google Patents

Abstract

Described are methods of determining a pregnant subject's risk of developing glucose intolerance and/or gestational diabetes mellitus (GDM), based on levels of FSTL3, also known as Follistatin Related Gene (FLRG).

Description

Follistatin-Like 3 (TSTL3 Levels in Gestational Diabetes

CLAIM OF PRIORITY

This application claims priority under 35 USC § 119(e) to U.S. Patent Application Serial No. 61/265,987, filed on December 2, 2009, the entire contents of which are hereby incorporated by reference.

TECHNICAL FIELD

This invention relates to methods of determining a pregnant subject's risk of developing glucose intolerance and/or gestational diabetes mellitus (GDM), based on levels of FSTL3, also known as Follistatin Related Gene (FLRG).

BACKGROUND

Gestational diabetes mellitus (GDM) afflicts 4% of pregnancies in the United States (Berg, C.J., et al, Obstet Gynecol, 2009. 113(5): p. 1075-81) and is associated with unfavorable perinatal outcomes including fetal macrosomia, shoulder dystocia, cesarean- section, and neonatal hypoglycemia (ACOG Practice Bulletin. Number 30, September 2001 (replaces Technical Bulletin Number 200, December 1994). Gestational diabetes. Obstet Gynecol, 2001. 98(3): p. 525-38) and importantly treatment of GDM has been shown to improve pregnancy outcomes (Crowther C et al, NEJM, 2005, 352:2477-2486). Following an episode of GDM, the future risk of subsequent type 2 diabetes mellitus in the mother approaches 70% (Kim, C, K.M. Newton, and R.H. Knopp, Diabetes Care, 2002. 25(10): p. 1862-8). GDM is characterized by glucose intolerance, beta cell dysfunction, and insulin resistance greater than the physiologic insulin resistance of pregnancy (Kim, C, K.M.

Newton, and R.H. Knopp, Diabetes Care, 2002. 25(10): p. 1862-8; Riskin-Mashiah, S., et al, Diabetes Care, 2009; Tsuchida, K., et al, J Biol Chem, 2000. 275(52): p. 40788-96; Sidis, Y, et al, Endocrinology, 2006. 147(7): p. 3586-97). However, the mechanisms responsible for the pathogenesis of GDM are not well elucidated. The glucose intolerance of GDM pregnancy resolves, at least temporarily, with the delivery of the infant and placenta, thus circulating factors released by the placenta may play a part in the pathogenesis of GDM.

In the United States, GDM is routinely detected at 24-28 weeks gestation by universal screening procedures. A recent study showed that higher first trimester levels of fasting blood glucose, even within the normal range, were linearly associated with an increased risk of GDM, cesarean-section, and macrosomia (Riskin-Mashiah, S., et al, Diabetes Care, 2009), suggesting that the pathophysiologic process which leads to the clinical diagnosis GDM is well underway weeks to months prior to its diagnosis. Thus, it is possible that factors directly or indirectly linked with the pathogenesis of this condition may be released by the placenta and present in blood samples well before the clinical diagnosis of GDM. Earlier detection of GDM through measurement of these proteins in the first trimester of pregnancy may permit more time for intervention, and subsequently improve both fetal and maternal outcomes.

SUMMARY

The present invention is based, at least in part, on the discovery that low FSTL3 levels, e.g., in the first trimester, are associated with the development of glucose intolerance and gestational diabetes mellitus later in pregnancy.

In general, the invention features methods of determining a woman's risk of developing glucose intolerance and/or gestational diabetes during pregnancy, by obtaining a biological sample, such as blood or serum, from a pregnant woman during the first or second trimester of pregnancy (e.g., at 5 weeks after conception, or in the first trimester (6 to 12 weeks), the second trimester (13 to 27 weeks), or at 8, 10, 12, 14, 16, 18, 20, or 24 weeks after conception); and measuring the level of FSTL3 in the sample; wherein the level of FSTL3 in the sample indicates the level of risk of developing gestational diabetes. In these methods, the sample can be a fasting or non-fasting sample. In some embodiments, FSTL3 is detected in a urine sample (rather than or in addition to in a blood, serum, or plasma sample).

In certain embodiments, the methods include measuring the level of FSTL3 in a biological sample, e.g., serum, blood, or urine, obtained from the pregnant subject; measuring the level of FSTL3 in the biological sample; comparing the FSTL3 level in the sample obtained from the pregnant subject with a reference FSTL3 level. In some embodiments, the reference FSTL3 level represents a level in a subject having a normal pregnancy; a low level of FSTL3 present in the sample obtained from the pregnant subject, as compared to the levels present in the reference representing a normal pregnancy, indicates that the pregnant subject has, or is predisposed to having, GDM. In some embodiments, the reference FSTL3 level represents a threshold level, and a level in the pregnant subject that is equal to or below the reference FSTL3 level indicates an increased risk of developing GDM. In some

embodiments, the reference FSTL3 level is a median level or a cutoff point for a tertile or quartile.

In some embodiments, the methods further include determining risk based on additional risk factors including, but not limited to, 1) the pregnant subject's race (African American, Hispanic, or Asian subjects have a higher risk than Caucasian subjects); 2) the pregnant subject's age (increasing age means increasing risk); 3) the pregnant subject's parity (multiparity increases risk); 4) the pregnant subject's body mass index (BMI over 25 or over 29 increases risk); the presence of polycystic ovary disease (PCOD) (increases risk); family history of diabetes (increases risk); or a previous delivery of a fetus > 9 lbs (increases risk).

In another aspect, the method further includes measuring the level of FSTL3 and at least one additional biomarker in the subject biological sample, e.g., urine, blood, or serum sample, and generating a subject profile comprising a value or plurality of values, each value representing a level of FSTL3 or an additional biomarker, and comparing the subject profile with a reference profile, wherein the reference profile comprises a value or plurality of values, each value representing a level of FSTL3 or the corresponding biomarker in a reference urine sample obtained from a reference subject. In some embodiments, the additional biomarker is activin, myostatin, a specific cytokine, Sex hormone-binding globulin (SHBG), sFLTl, and/or P1GF. The cytokine can be, e.g., an immune/hematopoietin, an interferon, a tumor necrosis factor (TNF)-related molecule or a chemokine. Examples include interleukin (IL)-6, IL-8, IL-lbeta, monocyte chemoattractant protein (MCP)-l or TNF-alpha, or any combination thereof. A reference profile can be generated from a sample obtained from any source containing, or believed to contain, a cytokine. Reference levels or thresholds of activin, myostatin, SHBG, P1GF, sFLTl, cytokines and/or growth factors can be used to generate reference profiles. For example, the reference profile can be obtained from the urine, serum, plasma, or blood of a reference subject. A reference subject can be a pregnant individual who has or later develops a gestational disorder or a pregnant individual having a normal pregnancy.

The methods of the invention can be accomplished by contacting a sample obtained from a pregnant subject with a biomolecule specific for FSTL3, e.g., an immobilized biomolecule specific for FSTL3, and detecting a modification of the biomolecule. In some embodiments, the modification is indicative of the level of FSTL3 in a sample and can include stable or transient binding of the biomolecule to FSTL3. The subject FSTL3 levels can be compared to reference levels as described herein. Reference levels can further be used to generate a reference profile from one or more reference subjects. In some embodiments, the biomolecules are antibodies, such as monoclonal antibodies, or antigen-binding fragments thereof. In yet another aspect, the biomolecules are receptors.

In another aspect, the invention features arrays for detecting a gestational disorder. These arrays include a substrate (or substrates) having a plurality of addresses, each address having disposed thereon a set of one or more biomolecules, and each biomolecule in a set specifically detecting the same molecule; wherein at least one set of one or more

biomolecules specifically detects FSTL3. The arrays can further include biomolecules that specifically detect markers such as, for example, activin, myostatin, SHBG, soluble fms-like tyrosine kinase-1 receptor (sFlt-1), P1GF, interleukin (IL)-6, IL-8, IL-lbeta, monocyte chemoattractant protein (MCP)-l or TNF-alpha. In one aspect, an array of the invention further includes at least two addresses having disposed thereon an immobilized biomolecule that specifically detects at least one growth factor, such as, for example, vascular endothelial growth factor (VEGF), or fibroblast growth factor (FGF)-2. In some embodiments, the array is on beads rather than a flat substrate.

The invention also features a pre-packaged diagnostic kit for detecting a gestational disorder. The kit can include, e.g., antibodies or antigen-binding fragments thereof that bind specifically to and detect FSTL3 and optionally an additional marker, e.g., activin, myostatin, SHBG, sFlt-1, P1GF, interleukin (IL)-6, IL-8, IL-lbeta, monocyte chemoattractant protein (MCP)-l or TNF-alpha, as described herein and instructions for using the kit in a method described herein, e.g., to test a biological sample, e.g., a urine, blood, or serum sample, to determine a subject's risk of developing GDM. In some embodiments, the kit includes antibodies that detect FSTL3 and SHBG, or FSTL3 and one or both of activin and myostatin.

The invention also includes methods to determine the efficacy of a therapy administered to treat a gestational disorder. These methods include contacting the array with a sample obtained from a pregnant patient undergoing therapy for a gestational disorder. The level of FSTL3 can be determined and compared to a level of FSTL3 detected in a sample obtained from the patient prior to, or subsequent to, the administration of the therapy.

Subsequently, a caregiver can be provided with the comparison information for further assessment. An increase in FSTL3 levels would indicate a successful therapy.

Further, a subject profile can be entered into a computer system that contains, or has access to, a database that includes a plurality of digitally-encoded reference profiles. Each profile of the plurality has a plurality of values, each value representing a level of FSTL3 of a pregnant individual having, or predisposed to having, a gestational disorder. In this manner, a single subject profile can be used to identify a subject at risk for developing a gestational disorder based upon reference values.

Thus, in other aspects, the invention also features computer-readable media that contain a database including one or more digitally-encoded reference profiles, wherein a first reference profile represents a level of FSTL3 in one or more samples from one or more pregnant individuals having a gestational disorder.

The invention also features computer systems for determining whether a pregnant subject has, or is predisposed to having, a gestational disorder. These systems include a database that has one or more digitally-encoded reference profiles, wherein a reference profile comprises a value that represents a level of FSTL3 in one or more samples from one or more pregnant individuals having a gestational disorder; and a server that includes a computer-executable code for causing the computer to: i) receive a profile of a pregnant subject comprising a level of FSTL3 detected in a sample from the subject; ii) identify from the database a matching reference profile that is diagnostically relevant to the pregnant subject profile; and iii) generate an indication of whether of the subject has, or is predisposed to having, a gestational disorder.

In addition to their use to identify women who are at risk, the new methods can be used as a routine screen or "pre-screen" for all pregnant women to identify those women who are not at risk for gestational complications, thus avoiding the need for additional testing later during pregnancy in those women.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Although methods and materials similar or equivalent to those described herein can be used in the practice or testing of the present invention, suitable methods and materials are described below. All publications, patent applications, patents, and other references

(including database entries such as genbank entries) mentioned herein are incorporated by reference in their entirety. In case of conflict, the present specification, including definitions, will control. In addition, the materials, methods, and examples are illustrative only and not intended to be limiting.

Other features and advantages of the invention will be apparent from the following detailed description, and from the claims.

DESCRIPTION OF DRAWINGS

FIG. 1 is a bar graph showing risk of developing GDM according to first trimester FSTL3 levels, by quartiles.

FIG. 2 is a bar graph showing the risk (odds ratio) of developing GDM based on first tertile cutpoint of < 12,000 pg/ml. FIG. 3 is a box-whisker/scatter plot showing first trimester FSTL3 levels in women who did (GDM) and did not (NO GDM) develop GDM during pregnancy. FSTL3 levels were measured in serum collected from GDM and cases and controls at the first prenatal obstetric visit. FSTL3 levels were significantly lower in women who developed GDM (p<0.001). Box plots depict the median (horizontal line in each box), the 25th percentile (bottom of each box), and the 75th percentile (top of each box). Box whiskers extend to the highest/lowest non-outlier value. Outliers were defined as lying greater than three interquartile ranges outside the 25th or 75th percentile. Scatter plot overlay depicts levels of FSTL3 in individual subjects.

FIG. 4 is a box plot showing the odds of developing GDM by first trimester FSTL3 tertile. Subjects were divided into tertiles based on first trimester FSTL3 level. Univariate and multivariate logistic regression analyses were used to determine odds ratios for the development of GDM in each tertile. Multivariate logistic regression model includes adjustment for age, gestational age, diastolic blood pressure, BMI, nonwhite race, and multiparity. * Significantly different from the reference tertile at the p<0.05 level.

FIG. 5 is a scatter plot showing the relationship between first trimester FSTL3 and glucose challenge test results. Glucose Challenge Test (GDM Screening) was performed at 24-28 weeks gestation. Fifty gram glucose load was administered orally and blood glucose was measured after one hour. R = - 0.3 P < 0.001 (Spearman correlation). Open circle markers represent GDM cases, filled circle markers represent controls.

DETAILED DESCRIPTION

Gestational diabetes mellitus (GDM), which complicates an estimated 3-7% of pregnancies, is associated with increased maternal and neonatal morbidity (Kim et al, Diabetes Care, 2002. 25(10): 1862-8; Solomon et al, JAMA, 1997. 278(13): 1078-83).

Identification of early risk markers may result in improved understanding of disease pathogenesis and identification of potential targets for intervention (Naylor et al, N Engl J Med, 1997. 337(22): 1591-6; Moses et al, Diabetes Care, 2009. 32(7): 1349-51.). There are no tests at the present time for the prediction of GDM. Currently, patients are diagnosed at 28 weeks of gestation based on an oral glucose loading and tolerance test. Early diagnosis i.e., before the onset of clinical symptoms, will allow clinicians to monitor at-risk patients closely and offer therapies such as insulin and metformin that have been shown to be beneficial in this population (HAPO Study Cooperative Group. N Engl J Med 2008, 358: 1991-02; Rowan et al, N Engl J Med, 2008, 358:2003-2015). Furthermore, novel therapies might allow the pregnancy to safely proceed, allowing valuable time for the growth of the baby and potentially limiting fetal as well as maternal morbidity and mortality (HAPO Study

Cooperative Group. N Engl J Med 2008, 358: 1991-02).

Current Methods of Testing for the Presence of GDM

Presently, there is no universally agreed-upon gold standard test for the detection of GDM. In the US, more than 90% of women are screened for GDM, in the category of "universal" screening (ACOG vs ADA recommendations). The usual screen is conducted at about 28 weeks of gestation (when insulin resistance is "maximum"), generally, the screen includes administering a 50 gm glucose challenge test (GCT) regardless of fasting status, and testing blood glucose 1 hour later. A level above 140 mg/dl is considered positive (although 130 mg/dl is used by some clinicians to capture more possible cases at the expense of specificity). Women with a positive test then generally take an oral glucose tolerance test (OGTT), during which a 100 gm glucose drink is administered to a subject after a 12 hour fast, and blood is collected before the glucose is administered and at intervals hereafter, e.g., at 1, 2, and 3 hours afterwards.

The following table shows exemplary normal glucose levels; levels above these can indicate the presence of GDM.

OGTT for gestational diabetes

j Fasting: j Less than 95 mg/dL or 5.2 mmol/L

j 1 -hour: j Less than 180 mg/dL or 10.0 mmol/L

j 2-hour: j Less than 155 mg/dL or 8.6 mmol/L

: 3-hour: j Less than 140 mg/dL or 7.7 mmol/L

These tests are not 100% reliable; over 20% of women with a positive GCT will test positive with an OGTT (false positive high), and performing a glucose loading test on two consecutive days can produce two entirely different results. The timing of screening is more or less arbitrary; earlier diagnosis would allow earlier intervention, but many women would be missed if only diagnosed using the GCT/OGTT paradigm; later diagnosis captures most women who really do have GDM, but the treatment may be too late to improve maternal or fetal outcomes. Left untreated, GDM can lead to delivery complications including shoulder dystocia, nerve palsy, and even fetal death, and an increased risk for diabetes in the baby. In addition, studies have shown that 50% of women who develop GDM go on to develop T2DM. Timely treatment can help reduce these complications, as well as reduce the frequency of labor induction, risk of preeclampsia (PE), and risk of Abruptio Placenta.

Present treatments generally include a restricted diet and increased exercise and/or the administration of insulin and/or oral hypoglycemic agents, e.g., glyburide or metformin.

FSTL3, Activin A, and Myostatin

Follistatin-Like-3 (FSTL3) is a structural and functional follistatin homolog which inhibits circulating members of the TGF subfamily of proteins (Tsuchida, K., et al, J Biol Chem, 2000. 275(52): p. 40788-96; Sidis, Y., et al, Endocrinology, 2006. 147(7): p. 3586-97; Schneyer, A., et al, Mol Cell Endocrinol, 2001. 180(1-2): p. 33-8). FSTL3 is expressed by the placenta (Ciarmela, P., et al., J Endocrinol Invest, 2003. 26(7): p. 641-5). FSTL3 and the proteins whose actions it antagonizes may play a major role in glucose homeostasis, as FSTL3 null mice are characterized by pancreatic beta cell hyperplasia and elevated insulin levels, increased glucose tolerance, depletion of liver glycogen stores, and up-regulation of hepatic gluconeogenesis (Mukherjee et al, Pro Natl Acad Sci USA, 2007, 104: 1348-53). Further evidence supporting a role for FSTL3 in glucose homeostasis includes the biologic activities of activin A and myostatin, both of which are antagonized by FSTL3. Activin A promotes differentiation and proliferation of beta-cells along with increased secretion of insulin (Florio, P., et al, J Endocrinol Invest, 2000. 23(4): p. 231-4; Park, M.K., et al., Transplantation, 2007. 83(7): p. 925-30). Myostatin is primarily known for its role as a negative regulator of muscle mass, but its absence has been shown to promote insulin sensitivity and protect against weight gain (Wolf, M., et al, J Clin Endocrinol Metab, 2002. 87(4): p. 1563-8; Gestational diabetes mellitus. Diabetes Care, 2004. 27 Suppl 1 : p. S88-90). Of note, levels of Activin A have been found to be elevated in pregnancies affected by GDM in two previous studies (Petraglia, F., et al, J Clin Endocrinol Metab, 1995. 80(2): p. 558-61; Gallinelli, A., et al, Eur J Endocrinol, 1996. 135(3): p. 340-4). While circulating levels of myostatin during pregnancy have not been previously described, myostatin is expressed by the human placenta and was shown to increase glucose uptake by placental explants

(Mitchell, M.D., et al, J Clin Endocrinol Metab, 2006. 91(4): p. 1434-7). FSTL3 is highly expressed by the placentas of infants with intrauterine growth restriction (Okamoto, A., et al, Placenta, 2006. 27(2-3): p. 317-21), and thus may also have a role in the regulation of placental nutrient partitioning.

The sequence of the human FSTL3 is available at GenBank Acc. Nos. NM_005860.2 (nucleic acid) and P_005851.1 (protein). The mouse (mus musculus) sequence is available at GenBank Acc. Nos. NM_031380.2 (nucleic acid) and NP_113557.1 (protein). Activin A is a homodimer of the inhibin beta A subunit. The sequence of the human inhibin beta A is available at GenBank Acc. Nos. NM_002192.2 (nucleic acid) and

NP_002183.1 (protein). The mouse (mus musculus) sequence is available at GenBank Acc. Nos. NM_008380.1 (nucleic acid) and NP_032406.1 (protein).

The sequence of the human myostatin is available at GenBank Acc. Nos.

NM_005259.2 (nucleic acid) and NP 005250.1 (protein). The mouse (mus musculus) sequence is available at GenBank Acc. Nos. NM_010834.2 (nucleic acid) and NP_034964.1 (protein).

Assays

The methods described herein can include determining the level of FSTL3 in a sample, i.e., a biological sample, from a subject. The sample can be, e.g., blood, serum, plasma, or urine, but is preferably serum. In some embodiments, the sample is obtained from a pregnant woman during the first or second trimester of pregnancy (e.g., at 5 weeks after conception, or in the first trimester (6 to 12 weeks), the second trimester (13 to 27 weeks), or at 8, 10, 12, 14, 16, 18, 20, or 24 weeks after conception, or of gestation); in preferred embodiments, the sample is obtained between 6 and 14 weeks, between 6 and 12 weeks, between 8 and 14 weeks, between 8 and 12 weeks, between 10 and 14 weeks, between 10 and 12 weeks, before 12 weeks, before 14 weeks, or before 15 weeks after conception (or of gestation). References should generally be relevant to the stage of pregnancy. In some embodiments, the sample is obtained from a subject who is not fasting.

The methods can optionally include determining levels of one or more additional biomarkers for GDM as known in the art and/or described herein, e.g., SHBG, activin, and/or myostatin; a specific cytokine, SHBG, sFLTl, and/or P1GF. The cytokine can be, e.g., an immune/hematopoietin, an interferon, a tumor necrosis factor (TNF)-related molecule or a chemokine. Examples include interleukin (IL)-6, IL-8, IL-lbeta, monocyte chemoattractant protein (MCP)-l or TNF-alpha, or any combination thereof. In some embodiments, the methods include determining levels of FSTL3 and SHBG.

Methods for determining a level of a biomarker, e.g., FSTL3, in a sample are known in the art, and can include detecting levels of the biomarker, e.g., FSTL3, protein or mRNA using a biomolecule that is specific for the biomarker, e.g., FSTL3. For example, an antibody or antigen-binding fragment thereof can be used.

Monoclonal antibodies and methods of making them are known in the art. Antigen- binding fragments and methods of making them are also known in the art and include an Fab, F(ab')2, Fv, or a single chain Fv (scFv) where the Fvs of an H chain and L chain are linked by a suitable linker (Huston, J.S. el al, Proc. Natl. Acad. Sci. U.S.A., (1998) 85, 5879-5883).

Antibodies specific for FSTL3 are known in the art and are commercially available, e.g., from Abeam (Cambridge, MA); Developmental Studies Hybridoma Bank (Iowa City, IA); Novus Biologicals (Littleton, CO); Santa Cruz Biotechnology, Inc. (Santa Cruz, CA); R&D Systems (Minneapolis, MN); and Sigma-Aldrich (St. Louis, MO).

Antibodies specific for activin are known in the art and are commercially available, e.g., from Abeam (Cambridge, MA); Thermo Scientific (Rockford, IL.); AbD Serotec (Raleigh, NC); Gen Way Biotech, Inc. (San Diego, CA); Novus Biologicals (Littleton, CO); and Santa Cruz Biotechnology, Inc. (Santa Cruz, CA).

Antibodies specific for myostatin are known in the art and are commercially available, e.g., from Abeam (Cambridge, MA); Alpco Diagnostics (Salem, NH); BioVendor Laboratory Medicine, Inc. (Modrice, Czech Republic);Immundiagnostik (Bensheim, Germany); Novus Biologicals (Littleton, CO); LifeSpan Biosciences (Seattle, WA); and R&D Systems (Minneapolis, MN).

Antibodies specific for SHBG are known in the art and are commercially available, e.g., from Abeam (Cambridge, MA); AbD Serotec (Raleigh, NC); Gen Way Biotech, Inc. (San Diego, CA); LifeSpan Biosciences (Seattle, WA); Novus Biologicals (Littleton, CO); R&D Systems (Minneapolis, MN); and Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). For methods relating to SHBG, see U.S. Pat. No. 7,344,892, and Am J Thadhani et al, Obstet Gynecol. 189: 171-176 (2003), which are incorporated herein by reference.

In some embodiments, the antibodies or antigen-binding fragments thereof are immobilized, e.g., on a substrate, e.g., on a surface (e.g., of an array) or a bead, e.g., latex or polystyrene/latex beads, or magnetic beads. Methods of making such immobilized antibodies are known in the art. The antibodies or antigen-binding fragments thereof can also be modified, e.g., genetically or chemically, e.g., by PEGylation or attachment of a detectable moiety. The detectable moiety can facilitate quantification of the biomarker, e.g., FSTL3, in the sample.

The methods can thus include providing a sample, e.g., a sample of serum, and contacting the sample with an antibody under conditions that allow the formation of binding complexes comprising the antibody or antigen-binding fragment thereof and the antigen, i.e., FSTL3, and detecting the formation of those complexes. The methods will generally also include quantitation of the complexes, to determine the quantity of FSTL3 protein present in the original sample. The level of a protein can be determined using methods known in the art, e.g., using quantitative immunoassay methods, including enzyme-linked immunosorbent assays (ELISA). See, e.g., Harlow and Lane, Using Antibodies: A Laboratory Manual (Cold Spring Harbor Laboratory Press, 1998). In some embodiments, high throughput methods, e.g., protein or gene chips as are known in the art (see, e.g., Ch. 12, Genomics, in Griffiths et al, Eds. Modern genetic Analysis, 1999,W. H. Freeman and Company; Ekins and Chu, Trends in Biotechnology, 1999, 17:217-218; MacBeath and Schreiber, Science 2000, 289(5485): 1760- 1763; Simpson, Proteins and Proteomics: A Laboratory Manual, Cold Spring Harbor Laboratory Press; 2002; Hardiman, Microarrays Methods and Applications: Nuts & Bolts, DNA Press, 2003), can be used to detect the presence and/or level of a biomarker, e.g., FSTL3.

In some embodiments, levels of free FSTL3, i.e., FSTL3 that is not bound to any other proteins, are determined. In some embodiments, levels of bound FSTL3, e.g., levels of FSTL3 bound to a binding partner such as activin and/or myostatin, are determined. In some embodiments, levels of total FSTL3 are determined, e.g., free plus bound FSTL3. In some embodiments, the methods further include calculating a ratio of total FSTL3 to bound FSTL3, or total FSTL3 to free FSTL3, and comparing the ratio with a reference or threshold ratio that is indicative of whether the subject has an increased risk of developing GDM.

Additional Risk Factors and Biomarkers

A number of additional risk factors for the development of GDM are known in the art, and thus the methods described herein can also include the determination of risk based on the presence of an additional risk factor. A factor that increases a subject's risk of developing GDM can include any one or more of: marked obesity; diabetes in a first-degree relative; personal history of glucose intolerance; prior delivery of macrosomic infant; or current glycosuria. The presence of all of the following factors would indicate a lower risk of developing GDM: age less than 25 years; low-risk ethnicity (i.e., other than Hispanic African, Native American, South or East Asian, Pacific Islander, or indigenous Australian); no diabetes in first-degree relatives; normal pre-pregnancy weight and normal pregnancy weight gain; no personal history of abnormal glucose levels; no prior poor obstetrical outcomes. See the recommendations of the American Diabetes Assn, Naylor et al, N. Engl. J. Med.

337: 1591-1597 (1997).

In some embodiments, the methods include selecting a subject for use of the present methods based on the presence of one or more factors that increase risk for GDM. The methods can also include determining levels of one or more clinically relevant biomarkers of risk for developing GDM. For example, levels of one or more of SHBG, sFLTl, and/or P1GF, interleukin (IL)-6, IL-8, IL-lbeta, monocyte chemoattractant protein (MCP)-l or TNF-alpha can be determined, and the subject's risk can be determined by comparison with a reference levels of the biomarker. See, e.g., U.S. Pat. No. 7,344,892, which is incorporated herein by reference in its entirety. For example, a number of biomarkers, and optionally one or more risk factors, and/or a glucose or insulin test such as homeostasis model assessment-estimated insulin resistance (HOMA-IR), see, e.g., Smirnakis et al, Diabetes Care, 28(5): 1207-1208 (2005), can be used to calculate a composite risk value using methods known in the art.

Reference Levels

The methods described herein include determining a subject's risk of developing GDM based on a comparison of a level of FSTL3 with a reference. In some embodiments, the reference FSTL3 level represents a level in a subject having a normal pregnancy; a low level of FSTL3 present in the sample obtained from the pregnant subject (i.e., statistically significantly lower), as compared to the levels present in the reference representing a normal pregnancy, indicates that the pregnant subject has, or is predisposed to having, GDM. For example, a reference level that represents a level of FSTL3 in a normal pregnancy (i.e., a level at or above which the subject is considered to have a low or normal risk of developing GDM) can be determined based on a clinical study cohort using methods known in the art.

Alternatively, the reference FSTL3 level can represent a level in a subject who has an increased (above normal) risk of developing GDM, and the presence of levels of FSTL3 in the subject that are less than or equal to the level of FSTL3 can indicate an increased risk of developing GDM. Again, an appropriate reference level can be determined based on a clinical study cohort using methods known in the art.

In some embodiments, the reference FSTL3 level represents a threshold level, and a level in the pregnant subject that is below, or equal to or below, the reference FSTL3 level indicates an increased risk of developing GDM. In some embodiments, the reference FSTL3 level is a median level or a cutoff point for a tertile or quartile (e.g., the lowest tertile or quartile), and a level in the pregnant subject that is below, or equal to or below, the reference (e.g., median or cutoff) FSTL3 level indicates an increased risk of developing GDM. Such references can be determined based on a clinical study cohort using methods known in the art. In some embodiments, the reference is a threshold level that is set at a percentage of a level in a cohort of subjects who have normal pregnancies, e.g., a level that is 25%, 30%, 40%, 45%, 50%, 55%, or 60% of a level in a normal cohort, e.g., a median level in a normal cohort. In some embodiments, e.g., wherein the assay is an ELISA, the reference level is 4, 6, 8, 10, 12, 14, 16, 18, or 20 ng/ml, or the equivalent thereof using a different assay, and the presence of a level that is at or below that level indicates that the subject has an increased risk of developing GDM.

Methods of Treatment

In addition to diagnosing or confirming the presence of or risk for GDM, the methods and compositions described herein also provide a means of optimizing the treatment of a subject having such a disorder. For example, the methods can include identifying (and selecting) a subject who is at increased risk of developing GDM based on a method described herein that includes determining levels of FSTL3 in the subject; and administering a treatment (e.g., a treatment that reduces the risk or clinical effects of GDM) to the subject, e.g., as known in the art or described herein, e.g., modifying diet and/or exercise, or administering insulin or an oral hypoglycemic agent such as glyburide. Optionally, the methods can include determining whether that treatment has an effect on levels of FSTL3 in the subject. The treatment can be continued or modified until a level of FSTL3 is attained that indicates a lower risk of developing GDM. The methods can be repeated to monitor changes in FSTL3 levels over time.

Theranostics

The methods described herein can be used to identify women at high risk for GDM such that a theranostic approach can be taken to test such individuals to determine the effectiveness of a particular therapeutic intervention (pharmaceutical or non-pharmaceutical) and to alter the intervention to 1) reduce the risk of developing adverse outcomes and 2) enhance the effectiveness of the intervention. Thus, the methods described herein can be used to provide a theranostic approach to treating and preventing GDM by integrating diagnostics and therapeutics to improve the real-time treatment of a subject having or at risk for developing GDM. Practically, this means using the present methods to identify which patients are most suited to a particular therapy, and providing feedback on how well a drug is working to optimize treatment regimens. In the area of GDM, theranostics can flexibly monitor changes in multiple important parameters over time. For example, theranostic multiparameter immunoassays specific for a series of diagnostically relevant molecules (e.g., FSTL3 and one or more of the clinically relevant biomarkers described herein, such as MCP- 1, CRP, IL-6, TNF-alpha, sFLTl and/or PIGF) can be used to follow the progress of a subject undergoing treatment for GDM. The markers provided herein are particularly adaptable for use in diagnosis and treatment because they are available in easily obtained body fluids such as urine, blood or serum.

Within the clinical trial setting, a theranostic method or composition of the invention can provide key information to optimize trial design, monitor efficacy, and enhance drug safety. For instance, "trial design" theranostics can be used for patient stratification, determination of patient eligibility (inclusion/exclusion), creation of homogeneous treatment groups, and selection of patient samples that are representative of the general population. Such theranostic tests can therefore provide the means for patient efficacy enrichment, thereby minimizing the number of individuals needed for trial recruitment. "Efficacy" theranostics are useful for monitoring therapy and assessing efficacy criteria. Finally, "safety" theranostics can be used to prevent adverse drug reactions or avoid medication error.

EXAMPLES

The invention is further described in the following examples, which do not limit the scope of the invention described in the claims.

Example 1

The experiments described in this example were performed to evaluate the hypothesis that circulating levels of FSTL3 might be altered in pregnancies complicated by GDM.

Design and Subjects

A nested case control study was conducted to determine if levels of FSTL3 in serum collected during the first trimester of pregnancy from women who went on to develop GDM were different from levels in the serum of women who did not go on to develop GDM. The parent study, MGH Obstetrical Maternal Study (MOMS), is a prospective cohort study of pregnant mothers which has been described previously (Wolf, M., et al, J Clin Endocrinol Metab, 2002. 87(4): p. 1563-8). In brief, the Massachusetts General Hospital obstetrics service provides community -based and high-risk obstetric care for women of varied socioeconomic and ethnic backgrounds from metropolitan Boston and throughout New England. MOMS subjects were recruited women receiving prenatal care at Massachusetts General Hospital and affiliated health centers between 1998 and 2005. Seventy percent of eligible women agreed to enroll in the cohort at their first prenatal visit at around ~10 weeks of gestation. Blood samples were collected at this time and stored at -80 degrees C for future analysis. Clinical information including age, height, race, last menstrual period, blood pressures and weights measured at prenatal visits, glucose challenge and glucose tolerance test results, mode of delivery, birth weight, and gestational age at delivery was collected from the obstetric electronic medical record. All subjects gave written informed consent.

The present study excluded subjects with a history of GDM in a previous pregnancy, preeclampsia (SBP>140 or DBP>100 and dipstick urine protein > 1+), small for gestational age infants (<10^th percentile in weight for gestational age at birth), multiple gestations, and subjects found to have glucose intolerance at less than 20 weeks gestation, as this may have represented pre-gestational diabetes. Standard of care at the MGH obstetric service is to administer a 50 gram, 1 hour, glucose loading test to all pregnant women at 24-28 weeks gestation to screen for GDM. Women whose blood glucose is greater than 140 mg/dL one hour after administration of the glucose challenge undergo subsequent glucose tolerance testing. Glucose tolerance testing consists of the administration of 100 grams of glucose after a 12 hour fast and the measurement of blood glucose levels 0, 1, 2, and 3 hours after glucose administration. GDM cases were defined as those women who had greater than 2 abnormal values on glucose tolerance test in accordance with American Diabetes Association guidelines (Gestational diabetes mellitus. Diabetes Care, 2004. 27 Suppl 1 : p. S88-90). Cases were also required to have an adequate first trimester blood sample available for analysis. Controls were chosen randomly from women who passed the glucose challenge test and/or glucose tolerance test. To the inventors' knowledge, circulating FSTL3 levels had not been measured in serum samples from pregnant women, thus precluding precise sample size and power calculations; this study proceeded with a 1 to 3 case to control ratio based on our previous studies of biomarkers in this population (Thadhani, R., et al, Am J Obstet Gynecol, 2003. 189(1): p. 171-6; Wolf, M., et al, Diabetes Care, 2004. 27(1): p. 21-7).

FSTL3 Measurement

ELISA testing of serum samples and cell culture supernatants was performed using human FLRG DuoSet ELISA (DY1288, R&D Systems, Minneapolis, M ). Assays were run in duplicate and the operator was blinded to the clinical data. A standard curve was generated using serial dilutions of recombinant FLRG (R&D Systems 1288-F3) at concentrations ranging from 10 ng/ml to 100 pg/ml in Reagent Diluent (1% BSA/Phosphate Buffered Saline, R&D Systems DY995). All reactions were carried out at room temperature. Plates were coated overnight with primary mouse anti-human FLRG antibody in PBS 4 μg/mL (100 μΐ/well). On the day of the assay each well was washed three times with wash buffer (0.05% Tween 20/Phosphate Buffered Saline, R&D Systems WA126) 400 μΐ/well using a Columbus Pro washer (Tecan, Inc) and then blocked with reagent Diluent 300 μΐ/well for 1 hour. Plates were then washed three times with wash buffer and 100 μΐ standard or sample was added to each well. Plates were incubated for 2 hours on a table top shaker and then washed three times as described above. Secondary biotinylated mouse anti-human FLRG antibody 400 ng/ml (100 μΐ/well) was added and plates were incubated for 2 hours on a table top shaker. Plates were washed again with Wash Buffer three times and Streptavidin conjugated to Horseradish-Peroxidase (100 μΐ/well) was added for 20 minutes. Plates were washed three times in wash buffer and color reagent containing H2O2 and tetramethylbenzidine (100 μΐ/well) was added (R&D Systems DY993). Plates were incubated for 20 minutes and Stop Solution 2N H₂S0₄ 50 μΐ (R&D Systems DY994) was added to each well. Optical density was measured using a Microplate Reader (BioRad, Model 660). Optical density at 550 nm was subtracted from optical density at 450 nm and serum concentrations were determined using a standard curve generated by a quadratic plot of the standard sample concentrations versus the measured optical density. Inter-assay coefficient of variation for this assay was approximately 15%.

Statistical Analysis

Subject characteristics and levels of FSTL3 in case and control subjects were compared using t-tests, Wilcoxon tests, or chi-squared tests as appropriate. Within the control group, one way analysis of variance was performed to determine whether FSTL3 levels differed in women who had failed the glucose challenge test, but passed the glucose tolerance test. Subjects were divided into tertiles based on FSTL3 levels, and odds ratios for the development of GDM were calculated for each tertile. Logistic regression analysis with appropriate indicator variables was performed to create univariate and multivariate models for the odds of developing GDM. Spearman's correlations between FSTL3 levels, blood glucose levels after 50g and lOOg glucose loads, and other subject characteristics were sought. Logarithmic transformation of FSTL3 levels was performed as appropriate. FSTL3 levels are reported as mean ± standard error of the mean unless otherwise noted. Results

Characteristics of the 52 subjects who developed gestational diabetes mellitus and the 295 control subjects are presented in Table 1. Subjects were similar in age, parity, and had similar blood pressures measured at the first prenatal visit. Mean gestational age at the first prenatal visit, the time of blood collection, was 10.9 weeks for all subjects and was similar in both groups. Subjects who developed gestational diabetes mellitus had slightly greater body mass index compared with those who did not, but this difference did not reach statistical significance. There was a greater percentage of Caucasians in the control group compared to the GDM group. Women with GDM delivered earlier in gestation than control women (38.8 ± 1.5 vs. 39.7 ± 1.1 weeks, p < 0.001), and GDM women gained less weight between the first visit and delivery. Infant birth weight and cesarean section rate did not differ between the groups

Table 1

FSTL3 levels (pg/m) in the subjects divided by category (e.g., presence of GDM, BMI, and pass/fail of GTT/GLT are shown in Table 2. Correlation of BMI with FSTL3 levels was R= -0.12, p=0.05. Table 2

The subjects were stratified into quartiles using univariate, multivariate (adjusted for age, BMI, and blood pressure), and OR-univariate model analysis. The results, shown in Fig. 1, demonstrate a clear increase for those subjects in the lowest quartile of FSTL3 levels.

Fig. 2 shows the odds ratios for developing GDM based on a first tertile cutpoint of < 12,000 pg/ml. The OR based on the univariate model was 10.1 (95% CI 5.2-19.6), with a sensitivity of 58% and a PPV of 65%. The OR based on the multivariate model was 12.0 (95% CI 5.8-25.5), with a specificity of 86% and NPV of 83%.

Example 2

The experiments described in this example were performed to confirm that circulating levels of FSTL3 might be altered in pregnancies complicated by GDM. This study focused on those subjects for whom enough serum was available in the sample bank to allow the FSTL3 measurements to be performed in duplicate. They were part of the same cohort of subjects tested in Example 1, and used the same methodology. Results

Characteristics of the 37 subjects who developed gestational diabetes mellitus and the 127 control subjects are presented in Table 3. Subjects were similar in age, parity, and had similar blood pressures measured at the first prenatal visit. Mean gestational age at the first prenatal visit, the time of blood collection, was 10.9 weeks for all subjects and was similar in both groups. Subjects who developed gestational diabetes mellitus had slightly greater body mass index compared with those who did not, but this difference did not reach statistical significance. There was a greater percentage of Caucasians in the control group compared to the GDM group. Women with GDM delivered earlier in gestation than control women (39.0 ± 0.20 vs. 39.8 ± 0.10 weeks, p < 0.001), and GDM women gained less weight between the first visit and delivery. Infant birth weight and cesarean section rate did not differ between the groups.

Table 3. Characteristics of Subjects with and without GDM at Prenatal Visit and Delivery

GDM (n=37) Control (n=127)

p-value¹ mean ± SE mean ± SE

First Prenatal Visit

Gestational Age (weeks) 10.5 ± 0.32 11.0 ± 0.15 0.20

Age (years) 34.2 ± 0.86 34.0 ± 0.41 0.80

BMI (kg/m²) 29.2 ± 1.4 26.8 ± 0.49 0.20

SBP (mmHg) 114 ± 1.6 112 ± 1.05 0.32

DBP (mmHg) 72 ± 1.4 70 ± 0.73 0.09

% Nulliparous 27% 24% 0.75

% Caucasian 65% 81% 0.04*

Delivery

Gestational Age (weeks) 39.0 ± 0.20 39.8 ± 0.10 <0.001*

Weight Gain (lbs) 22.3 ± 1.9 27.5 ± 1.3 0.005*

Birth Weight (g) 3491 ± 91 3538 ± 43 0.62

C-Section (%) 30% 28% 0.80

¹ P-value reflects: T-test for normally distributed variables: Age, Gestational age, Birth Weight. Mann- Whitney- Wilcoxon for non -normally distributed variables: BMI, Blood Pressure, Weight Gain. Chi-squared for categorical variables: Race, Parity, Cesarean-Section. * Significant at the p<0.05 level.

First trimester FSTL3 levels were lower in subjects who developed GDM compared with control subjects (17,216 ± 3130 vs. 43,038 ± 3,564 pg/ml, pO.001, see Figure 3).

Average FSTL3 levels in subjects who failed the 50 gram glucose challenge screening test

(blood glucose at 1 hour > 140), but passed the glucose tolerance test (diagnostic test), was

38,861 ± 7632 pg/ml, lower but not significantly different from FSTL3 levels in women who passed the glucose challenge test (43,683 ± 3950 pg/ml, p=0.614) and significantly greater than those who went on to develop GDM (p<0.001). When subjects were divided into tertiles based on FSTL3 levels, women in the first tertile had an 11.2-fold risk of developing GDM as compared to women in the 3^rd tertile (Figure 4). This odds ratio increased to 15.0 after adjustment for gestational age, maternal age, race, BMI, blood pressure, and parity. In the multivariate logistic regression model, only FSTL3 tertile and nonwhite race, and not baseline body mass index, blood pressure, age, or smoking history, were significant independent predictors of GDM.

FSTL3 was not correlated with gestational age (p=0.44), BMI (p=0.29), systolic blood pressure (p=0.87), diastolic blood pressure (p=0.96) nonwhite race (p= 0.40), or age (p=0.71). Levels of FSTL3 in white and non-white women did not differ (p=0.90) and FSTL3 levels did not differ among women with different smoking histories (p=0.80). Glucose challenge test result (blood glucose 1 hour after a 50 gram glucose load) was inversely correlated with FSTL3 level (r= -0.30 pO.001, Figure 5) and positively correlated with BMI (r=0.18, p=0.03) and non-white race (r=0.23, p=0.003). In a multivariate linear regression model, only FSTL3 and non-white race independently predicted glucose challenge test result.

Conclusion

In this nested case-control study, low FSTL3 levels in the first trimester were shown to be associated with the development of glucose intolerance and gestational diabetes mellitus later in pregnancy. When subjects were divided into tertiles based on FSTL3 levels, women with the lowest levels (tertile 1) demonstrated a markedly elevated (11 -fold) risk compared to women with the highest FSTL3 levels. This estimate of risk was further increased to 15-fold after adjustment for age, body mass index, nonwhite race, blood pressure, nonwhite race, and multiparity, which are known risk factors for GDM (Solomon, C.G, et al, JAMA, 1997. 278(13): p. 1078-83). Levels of FSTL3 were inversely correlated with levels of blood glucose during a 50 gram glucose challenge test. These data are consistent with a role for FSTL3 in GDM pathophysiology. Furthermore, the association between first-trimester FSTL3 levels and subsequent gestational diabetes may allow FSTL3 to be tested as a candidate biomarker for estimation of GDM risk early in pregnancy.

It is expected that the study sample is representative of the pregnant population treated at a large New England hospital, with the exception that no subject had a previous history of GDM. Twenty-three percent of subjects were ethnic/racial minorities. Subjects with and without GDM were similar with the exception of race, gestational age at delivery, and weight gain during pregnancy. Nonwhite race is a known risk factor for GDM (Solomon, C.G, et al, JAMA, 1997. 278(13): p. 1078-83); this likely explains the difference in racial makeup of the case and control group. Women with gestational diabetes mellitus are followed closely for fetal growth and induction of labor or cesarean section is targeted if there is concern for macrosomia (ACOG Practice Bulletin. Number 30, September 2001 (replaces Technical Bulletin Number 200, December 1994). Gestational diabetes. Obstet Gynecol, 2001. 98(3): p. 525-38). This likely accounts for the younger gestational age at delivery, lower weight gain, and similar birth weights in the GDM group. FSTL3 levels were not correlated with BMI, parity, maternal age, nonwhite race or smoking history, clinical factors known to be associated with GDM (Solomon, C.G, et al, JAMA, 1997. 278(13): p. 1078-83; Naylor, CD., et al, Toronto Trihospital Gestational Diabetes Project Investigators. N Engl J Med, 1997. 337(22): p. 1591-6; Weijers, R.N., D.J. Bekedam, and Y.M. Smulders, Diabetes Care, 2002. 25(1): p. 72-7). It is probable that this study was not powered to detect these relationships, but it is also possible that a low FSTL3 level conveys risk independent of these factors. In our sample, BMI, parity, maternal age, and smoking history were not significantly associated with a higher risk of GDM, implying that our study may have been underpowered to detect these known relationships. In contrast, FSTL3 level and nonwhite race predicted GDM independently, implying that FSTL3 levels in non-white women would not completely explain the reason for their known elevated risk for GDM.

REFERENCES

1. Fisher, P.M., H.W. Sutherland, and P.D. Bewsher, Diabetologia, 1980. 19(1): p. 15-20.

2. Van Assche, F.A., L. Aerts, and F. De Prins, Br J Obstet Gynaecol, 1978. 85(11): p. 818-20.

3. Yen, S.S., C.C. Tsai, and P. Vela, Am J Obstet Gynecol, 1971. 111(6): p. 792-800.

4. Botta, R.M., et al., Acta Diabetol Lat, 1988. 25(1): p. 81-8.

5. Yamaoka, T., et al., J Clin Invest, 1998. 102(2): p. 294-301.

6. Kirwan, J.P., et al., Diabetes, 2002. 51(7): p. 2207-13.

7. Ryan, E.A., M.J. O'Sullivan, and J.S. Skyler, Diabetes, 1985. 34(4): p. 380-9.

8. Fisher, P.M., H.W. Sutherland, and P.D. Bewsher, Diabetologia, 1980. 19(1): p. 10-4.

9. McLachlan, K.A., et al., Diabetes Metab Res Rev, 2006. 22(2): p. 131-8.

10. Homko, C, et al., J Clin Endocrinol Metab, 2001. 86(2): p. 568-73.

11. Catalano, P.M., et al., Am J Physiol, 1993. 264(1 Pt 1): p. E60-7.

12. Catalano, P.M., et al., Am J Obstet Gynecol, 1999. 180(4): p. 903-16.

13. Lapolla, A., et al., J Clin Endocrinol Metab, 2008. 93(3): p. 876-80.

14. Kautzky- Wilier, A., et al., Diabetes Care, 1997. 20(11): p. 1717-23.

15. Guo, T., et al., PLoS One, 2009. 4(3): p. e4937. 16. Wilkes, J.J., D.J. Lloyd, and N. Gekakis, Diabetes, 2009. 58(5): p. 1133-43.

17. Ostlund, I. and U. Hanson, Acta Obstet Gynecol Scand, 2003. 82(2): p. 103-8.

18. Berger, H. and M. Sermer, Diabetes Care, 2009. 32(7): p. 1352-4.

19. Moses, R.G. and N.W. Cheung, Diabetes Care, 2009. 32(7): p. 1349-51.

20. Phaloprakarn, C, S. Tangjitgamol, and S. Manusirivithaya, Eur J Obstet Gynecol Reprod Biol, 2009. 145(1): p. 71-5.

21. Pryor-Koishi et al., BJOG 114:1128-1137 (2007).

22. Day et al., Nat. Chem. Biol. 5(10):749-757 (2009)

23. Petraglia et al., "Ch. 9: Changes in Activin A Secretion in Gestational Diabetes," in Inhibin, Activin. and Follistatin. (Serono Sumposia, Norwell, MA) 1997, pp. 96-103.

OTHER EMBODIMENTS

It is to be understood that while the invention has been described in conjunction with the detailed description thereof, the foregoing description is intended to illustrate and not limit the scope of the invention, which is defined by the scope of the appended claims. Other aspects, advantages, and modifications are within the scope of the following claims.

Claims

WHAT IS CLAIMED IS:

1. A method of determining a pregnant subject's risk of developing gestational diabetes mellitus (GDM), the method comprising:

providing a biological sample from the subject;

determining a level of Follistatin-Like 3 (FSTL3) in the sample; and

comparing the level of FSTL3 in the sample with a reference level,

wherein the presence of a level of FSTL3 in the sample that is equal to or less than the reference level indicates that the subject has an increased risk of developing GDM.

2. The method of claim 1, wherein the biological sample was obtained from the subject before 15 weeks of gestation.

3. The method of claim 1, further comprising determining a level of sex hormone-binding globulin (SHBG) in the sample; comparing the level of the second marker to a reference level of SHBG, wherein the presence of levels of SHBG that are equal to or less than the reference levels indicate an increased risk of developing GDM.

4. The method of claim 1, wherein the sample comprises blood, serum, plasma, or urine.

5. The method of claim 1, wherein determining a level of FSTL3 in the sample comprises contacting the sample with an antibody or antigen-binding fragment thereof that binds specifically to FSTL3, allowing the formation of complexes between FSTL3 and the antibody or antigen-binding fragment thereof, and detecting the complexes.

6. The method of claim 5, wherein the antibody or antigen-binding fragment thereof is immobilized on a substrate.

7. The method of claim 6, wherein the substrate is a bead.

8. The method of claim 6, wherein the antibodies are monoclonal antibodies.

9. The method of claim 1, wherein the reference level is a percentage of a median level in a cohort of subjects who have normal pregnancies.

10. The method of claim 1, wherein the reference level is a threshold level that represents the level in a subject with a normal risk of developing gestational diabetes.