WO2012042020A2

WO2012042020A2 - Glyscosylation markers for the diagnosis of maturity onset diabetes of the young

Info

Publication number: WO2012042020A2
Application number: PCT/EP2011/067112
Authority: WO
Inventors: Harry Campbell; Igor Rudan; Alan F. Wright; Jayesh Jagunnathan Kattla; Pauline Rudd; Gordan Lauc
Original assignee: The University Court Of The University Of Edinburgh; National Institute for Bioprocessing Research and Training Limited; Genos Ltd
Priority date: 2010-09-30
Filing date: 2011-09-30
Publication date: 2012-04-05
Also published as: WO2012042020A3; GB201016464D0

Abstract

A method for diagnosing Maturity-Onset Diabetes of the Young (MODY) in a subject is provided. The method comprises providing a glycan- containing test sample from the subject; determining one or more levels of fucosylation of the glycans in the test sample; comparing the one or more levels of fucosylation with one or more reference values; and providing a diagnosis based on said comparison. This method may be used to distinguish HNF1A-MODY or HNF4A-MODY from other types of diabetes such that the correct treatment may be prescribed for the subject.

Description

GLYCOSYLATION MARKERS FOR THE DIAGNOSIS OF MATURITY- ONSET DIABETES OF THE YOUNG

Field of the invention

The present invention relates to a method for diagnosing disease based on the analysis of glycosylation. In particular, the present invention relates to a method for diagnosing Maturity-Onset Diabetes of the Young MODY) based on the analysis of glycosylation. Background to the Invention

Most human proteins are post-translationally modified in a process referred to as glycosylation by the addition of complex oligosaccharide structures (glycans). This has implications for protein folding, degradation, cell signalling, secretion, immune function and transcription. Mis- regulation of glycosylation results in a wide range of diseases including cancer, diabetes, cardiovascular, congenital, immunological and infectious disorders. Despite the impact on protein structure and function, the clinical consequences of changes in the human glycome remain largely unexplored, primarily because reliable analytical techniques have only recently been developed. Recent technological advances have allowed reliable, high-throughput quantification of N-glycans, which now permits investigation into the genetic regulation and biological roles of glycan structures and brings glycomics into line with genomics, proteomics and metabolomics.

Diabetes mellitus is a heterogeneous group of metabolic disorders wherein pancreatic cell dysfunction and insulin resistance of target tissues play a central role. This group includes type 2 diabetes, formerly known as non-insulin-dependent diabetes mellitus (NIDDM) or adult-onset diabetes, which is characterized by high blood glucose levels caused by insulin resistance and relative insulin deficiency. Type 2 diabetes is polygenic, that is, multiple genetic factors contribute to the development of type 2 diabetes. However, environmental factors such as physical inactivity, over-nutrition and obesity also play a role in the development of type 2 diabetes. Some other forms of diabetes result from mutations in a single gene (monogenic). Monogenic forms of diabetes account for about 2 to 5 percent of all cases of diabetes. Mutations disrupting function of the hepatocyte nuclear factor 1 -alpha (HNF1A) gene are responsible for the most common subtype of monogenic diabetes, HNF1 A-Maturity-Onset Diabetes of the Young (HNF1 A-MODY or MODY3). A second type of MODY is glucokinase-MODY (GCK-MODY or MODY2), which occurs when the glucokinase gene malfunctions. HNF4A-MODY (MODY1 ) is a less common form that is often diagnosed at a later stage. HNF1 B-MODY (MODY5) is associated with renal cysts. PDX1 or IPF1 (MODY4) and NeuroDI (MODY6) are very rare forms.

Like other forms of MODY, HNF1 A-MODY is characterized by autosomal dominant inheritance and pancreatic beta-cell dysfunction. This typically leads to diabetes diagnosed in early adulthood in the absence of beta-cell autoimmunity and insulin resistance. In clinical practice, diagnostic differentiation between HNF1 A-MODY and other causes of early-onset diabetes (including type 1 and type 2 diabetes and other forms of MODY) is complicated by the overlap of phenotypic features. In most countries, only a minority of HNF1 A-MODY cases are referred for definitive molecular testing (i.e. HNF1A sequencing). Even in centres where individuals with typical MODY characteristics (presentation before age 25 with parental diabetes) are offered diagnostic sequencing, the variability in clinical phenotype means that half of all cases are missed. Failures or delays in accurate molecular diagnosis have clinical repercussions, since, unlike other forms of diabetes, the optimal treatment for HNF1 A-MODY is low-dose sulfonylureas. Patients with undiagnosed HNF1A-MODY may spend many years receiving inappropriate treatment (including exogenous insulin) and enduring suboptimal glycemic control. Recent efforts to improve diagnostic performance by identifying

biochemical markers specific for MODY subtypes have met with varying success. It has recently been demonstrated that individuals with HNF1A- MODY have low levels of C-reactive protein (CRP) and that high- sensitivity (hs) CRP assays can discriminate well between HNF1A-MODY and both type 2 diabetes and HNF4A-MODY. However, CRP is an acute inflammatory marker and diagnostic discrimination can be disturbed by intercurrent infection. In addition, hsCRP assays do not discriminate well between HNF1A-MODY and type 1 diabetes. Therefore, a need remains to diagnose HNF1 A-MODY in asymptomatic subjects and further to separate HNF1 A-MODY cases from type 1 and type 2 diabetes cases and from other types of MODY. This would allow appropriate treatment of HNF1A-MODY.

Summary of the invention

The inventors of the present application have identified specific

glycosylation markers for diagnosing HNF1 A-Maturity-Onset Diabetes of the Young (HNF1 A-MODY) or HNF4A-Maturity-Onset Diabetes of the Young (HNF4A-MODY) and discriminating these from other types of diabetes.

Accordingly, according to a general aspect of the present invention there is provided a method for diagnosing or screening for HNF1 A-MODY or HNF4A-MODY in a subject in need thereof, comprising:

- providing a glycan-containing test sample from the subject; - determining one or more levels of fucosylation of the glycans in the test sample;

- comparing the one or more levels of fucosylation of the glycans in the test sample with one or more reference values; and

- providing a diagnosis based on said comparison.

In certain embodiments, the method is a method for distinguishing

HNF1 A-MODY or HNF4A-MODY from at least one of type 2 diabetes, type

1 diabetes and other forms of MODY such as glucokinase-MODY.

The main advantage of diagnosing HNF1 A-MODY lies in the potential for individualized therapy with low-dose sulfonylureas. The method of the present invention has the ability to differentiate HNF1 A-MODY and HNF4A-MODY from other types of diabetes, including type 1 and type 2 diabetes and other forms of MODY such as glucokinase-MODY. The method of the present invention may be used as a diagnostic adjunct particularly in those with diabetes diagnosed in the second to fourth decades, in whom the proportion of HNF1 A-MODY is highest (up to 5%) and clinical discrimination most difficult.

The method of the present invention offers stability over time as

fucosylation is thought to be less prone to spurious elevation from intercurrent infection than, for example, hsCRP levels. Furthermore, from a clinical perspective, it has been observed that adjustment for covariates (such as age, gender, sample type) does not alter the magnitude or significance of differences between diabetic subgroups, thus facilitating incorporation of the method into simple diagnostic protocols.

The method of the present invention includes screening the subject for HNF1 A-MODY or HNF4A-MODY. A diagnosis of HNF1 A-MODY or HNF4A-MODY therefore includes a prediction that the subject is likely to develop diabetes in the future where the subject does not already have diabetes. At-risk family members can therefore be screened for HNF1 A- MODY or HNF4A-MODY and offered diagnostic or predictive genetic testing.

In certain embodiments, the method is a method for diagnosing or screening for HNF1A-MODY. In certain embodiments, the method is a method for diagnosing or screening for HNF4A-MODY

The glycan-containing test sample may be any suitable sample obtained from the subject in need of the diagnosis which comprises N-linked glycans. The test sample may be obtained by any suitable technique known in the art. The test sample may comprise glycans of a sample of body fluid or body tissue. Further examples include, but are not limited to, test samples comprising glycans of whole serum, blood plasma, blood, dried blood spots, urine, sputum, seminal fluid, seminal plasma, pleural fluid, ascites, nipple aspirate, faeces or saliva. In particular, the test sample may comprise glycans of plasma proteins from the subject.

Alternatively, the test sample may comprise serum or dried blood spots.

In certain embodiments, the test sample comprises a component of the sample initially obtained from the subject (e.g., serum, a serum fraction, a cell or tissue lysate, a glycan pool, an isolated protein, etc). The sample can be treated as necessary prior to determining the one or more levels of fucosylation, for example, cells and/or tissues are optionally lysed for detection of intracellular glycoprotein markers.

The test sample may comprise the total glycans in the sample initially obtained from the subject, for example, the total glycans in serum from the subject or the total glycans released from total plasma proteins of the subject. The one or more levels of fucosylation can thus be detected without purification of particular proteins from the sample. Alternatively, one or more glycoproteins (e.g., one or more plasma or acute phase proteins) are isolated from the sample from the subject prior to determining the one or more levels of fucosylation. Thus, the test sample may comprise a sample which has undergone one or more purification steps. For example, in certain embodiments, the test sample comprises glycans of one or more particular glycoproteins which have been purified from other glycoproteins present in the initial sample taken from the subject. The one or more particular glycoproteins may comprise, for example, one or more acute phase proteins (e.g., serum amyloid A, haptoglobin, a1 -acid glycoprotein, a1 -antitrypsin, a1 -antichymotrypsin, fibrinogen, transferrin, complement C3, a2-macroglobulin, prothrombin, factor VIII, von Willebrand factor or plasminogen) or other plasma protein(s) of interest. Affinity purification of proteins may be used to isolate them prior to determining the level of fucosylation. In certain embodiments, the test sample comprises one or more groups of glycans. The one or more groups of glycans may comprise one or more classes of glycans, such as, glycans with antennary fucose and/or core fucose, biantennary glycans, triantennary glycans and/or tetrantennary glycans.

In certain embodiments, the test sample comprises glycoproteins.

Optionally the test sample may be subjected to one or more steps to release glycans from the glycoproteins prior to determining the one or more levels of fucosylation. In certain embodiments, the sample comprises glycans released from glycoproteins. Glycans may be released from the glycoproteins contained in the test sample using enzymatic and or chemical means known to the person skilled in the art.

In certain embodiments, determining the one or more levels of fucosylation of the glycans in the sample comprises determining the level of antennary fucosylation (FUC-A), in particular determining the proportion of glycans with or without antennary fucose. In certain embodiments, the proportion of biantennary, triantennary and/or tetraantennary glycans which is fucosylated is determined. Determining the proportion of triantennary glycans which is fucosylated has been shown to be particularly effective in diagnosing HNF1A-MODY or HNF4A-MODY and distinguishing these from other types of diabetes. In addition, determining the proportion of biantennary fucosylated glycans has been shown to be useful in diagnosing HNF1A-MODY or HNF4A-MODY and, particularly, in distinguishing HNF1A-MODY from type 2 diabetes in plasma samples. Accordingly, in certain embodiments where the test sample is a plasma sample, determining the one or more levels of fucosylation of the glycans in the sample comprises determining the proportion of biantennary glycans which is fucosylated.

In certain embodiments, determining the one or more levels of fucosylation of the glycans in the sample comprises determining the level of core fucose (FUC-C). In certain embodiments, determining the one or more levels of fucosylation of the glycans in the sample comprises determining a ratio of antennary fucose of the glycans to core fucose of the glycans.

In certain embodiments, comparing the one or more levels of fucosylation of the test sample with the one or more reference values comprises comparing the one or more levels of fucoslyation of the test sample with one or more levels of the same type of fucosylation (e.g. core or antennary; biantennary or triantennary) in one or more patients with a type of diabetes other than HNF1 A-MODY or HNF4A-MODY, such as type 1 or type 2 diabetes or other forms of MODY. Accordingly, in certain embodiments, the one or more reference values are one or more corresponding fucosylation levels in a reference sample from one or more subjects diagnosed with a form of diabetes other than HNF1 A-MODY or HNF4A-MODY, such as, type 1 or type 2 diabetes or other forms of MODY. "Corresponding fucosylation levels" refer to the same type of fucosylation levels as that being determined in the test sample (e.g. core or antennary; biantennary or triantennary). The method of the present invention may therefore be used to distinguish subjects with HNF1A- MODY or HNF4A-MODY from subjects with other types of diabetes, such as type 1 or type 2 diabetes or other forms of MODY, such that the correct treatment may be prescribed for the subject. The reference sample can originate from a single individual or be a sample pooled from more than one individual. In certain embodiments, comparing the one or more levels of fucosylation of the test sample with the one or more reference values comprises comparing the one or more levels of fucoslyation of the test sample with one or more levels of the same type of fucosylation (e.g. core or antennary; biantennary or triantennary) in one or more subjects who do not have any form of diabetes. Accordingly, in certain embodiments, the one or more reference values are one or more corresponding fucosylation levels in a control sample from one or more subjects who do not have diabetes. The control sample can originate from a single individual or be a sample pooled from more than one individual. Typically, the person skilled in the art would be aware that any statistically significant difference between the one or more reference values and the one or more levels of fucosylation of the test sample would be determinant of a diagnosis. The degree of difference between the one or more levels of fucosylation in the test sample and the one or more reference values which would be determinant of a diagnosis would be well within the skill of the ordinary person to determine.

In certain embodiments, providing the diagnosis comprises determining whether the one or more levels of fucosylation in the test sample are above or below the one or more reference values, that is, the one or more reference values represent one or more threshold values. In certain embodiments, the one or more reference values have been previously determined and examples of threshold values for different types of fucosylation as determined by the present inventors are provided below. Alternatively, the one or more reference values may be determined by determining one or more fucosylation levels in a reference sample from one or more subjects with a type of diabetes other than HNF1 A-MODY or HNF4A-MODY or in a control sample from one or more subjects without diabetes. In certain embodiments, the one or more reference values are determined for each analytical method used and/or for each population. Determining the one or more reference values for each population allows for ethnic diversity which might require different cut-off values for each population. Thus, optimal values which will differentiate between healthy individuals and HNF1A-MODY/HNF4A-MODY are determined for each population. Typically, a level of fucosylation in the test sample which is below the indicated threshold value is indicative of a diagnosis of HNF1 A- MODY or HNF4A-MODY. Specifically, a reduced fucosylation level in the test sample may indicate a diagnosis of HNF1 A-MODY or HNF4A-MODY and distinguish these from type 1 and type 2 diabetes and from other forms of MODY as changes in glycosylation in total plasma or individual glycoproteins in other diseases generally involve an increase in fucose.

In certain embodiments, a fucosylation level of any individual protein or of a class of glycans which is decreased below 73%, typically 72%, more typically 71 %, even more typically 70%, even more typically 69% and even more typically 68% of the reference value of the control or reference sample is indicative of a diagnosis of HNF1 A-MODY or HNF4A-MODY. The class of glycans may be selected from at least one of the group consisting of antennary glycans, biantennary glycans, triantennary glycans and tetrantennary glycans. The individual protein may be a plasma protein. In certain embodiments, a reduction in the one or more

fucosylation levels of 30% or more with respect to the one or more reference values obtained from a control sample or a reference sample indicates a diagnosis of HNF1 A-MODY or HNF4A-MODY.

In certain embodiments, a reduction in the proportion of antennary glycans which is fucosylated indicates a diagnosis of HNF1 A-MODY or HNF4A- MODY. In certain embodiments, the presence of antennary fucose on less than 1 .8%, typically less than 1 .6%, of total glycans, for example in serum, is indicative of HNF1 A-MODY. Accordingly, when the one or more levels of fucosylation determined in the test sample comprises the proportion of total glycans which is fucosylated (that is, the amount of antennary glycans which are fucosylated calculated as a percentage of the total amount of antennary glycans including fucosylated and non- fucosylated antennary glycans), the reference value to which the determined level is compared is 1 .8%, and typically 1 .6%. A determined percentage of fucosylation in the test sample of less than the reference value indicates a diagnosis of HNF1 A-MODY. In certain embodiments, a reduction in the ratio of antennary to core fucose indicates a diagnosis of HNF1A-MODY or HNF4A-MODY. In certain embodiments, a ratio of less than 0.07, typically less than 0.065, in serum or blood plasma is indicative of HNF1A-MODY. The ratio may be calculated as disclosed in Knezevic et al, Glycobiology 19:1547-53. The normal range is between 0.08 and 0.14. Accordingly, when the one or more levels of fucosylation determined in the test sample comprises the ratio of antennary to core fucose, for example in serum or blood plasma, the reference value to which the determined level is compared is 0.07 and typically 0.065. A determined ratio in the test sample of less than the reference value indicates a diagnosis of HNF1A-MODY.

In certain embodiments, a reduction in the proportion of biantennary glycans which is fucosylated indicates a diagnosis of HNF1A-MODY or HNF4A-MODY. In certain embodiments, the presence of fucose on less than 6%, typically less than 5%, of total biantennary glycans, for example in serum or plasma, is indicative of HNF1A-MODY. Accordingly, when the one or more levels of fucosylation determined in the test sample comprises the proportion of total biantennary glycans which is fucosylated (that is, the amount of biantennary glycans which are fucosylated calculated as a percentage of the total amount of biantennary glycans including fucosylated and non-fucosylated biantennary glycans), the reference value to which the determined level is compared is 6% and typically 5%. A determined percentage of fucosylation in the test sample of less than the reference value indicates a diagnosis of HNF1A-MODY. In certain embodiments, determining the proportion of biantennary glycans which are fucosylated allows HNF1 A-MODY to be distinguished from type 2 diabetes. In certain embodiments, a reduction in the proportion of triantennary glycans which is fucosylated indicates a diagnosis of HNF1 A-MODY or HNF4A-MODY. This type of fucosylation is thought to be less prone to spurious elevation from intercurrent infection than, for example, CRP levels. In certain embodiments, the presence of fucose on less than 18%, typically less than 16%, more typically less than 14%, even more typically less than 12% and even more typically less than 10% or 9% of total triantennary glycans, for example, in serum of plasma, is indicative of HNF1 A-MODY. The presence of fucose on less than 13.8%, typically less than 12.8%, of total triantennary glycans, for example in serum or plasma, is indicative of HNF1 A-MODY. Accordingly, when the one or more levels of fucosylation determined in the test sample comprises the proportion of triantennary glycans which is fucosylated (that is, the amount of triantennary glycans which are fucosylated calculated as a percentage of the total amount of triantennary glycans including fucosylated and non- fucosylated triantennary glycans), the reference value to which the determined level is compared is one of the above values. A determined percentage of triantennary fucosylation in the test sample of less than the reference value indicates a diagnosis of HNF1 A-MODY, which includes a likelihood of the subject developing HNF1 A-MODY where the subject does not already have diabetes. A determined percentage of triantennary fucosylation in the test sample of 18% or higher indicates the subject does not have HNF1A-MODY and is unlikely to develop HNF1A-MODY. In certain embodiments, a proportion of triantennary glycan fucoslyation of between 15% and 20%, typically around 18% or 19% indicates a diagnosis of HNF4A-MODY.

Reference samples for subjects with type 1 diabetes, type 2 diabetes or gluckokinase-MODY showed levels of triantennary fucosylation of approximately 25%. Similar levels were observed for control subjects not having diabetes.

In certain embodiments, a reduction in the proportion of tetraantennary glycans which is fucosylated indicates a diagnosis of HNF1A-MODY or HNF4A-MODY. In certain embodiments, the presence of fucose on less than 21 .5%, typically less than 19.5%, of total tetraantennary glycans, for example in serum or plasma, is indicative of HNF1A-MODY. Accordingly, when the one or more levels of fucosylation determined in the test sample comprises the proportion of total tetraantennary glycans which is fuscosylated (that is, the amount of tetraantennary glycans which are fucosylated calculated as a percentage of the total amount of

tetraantennary glycans including fucosylated and non-fucosylated tetraantennary glycans), the reference value to which the determined level is compared is 21 .5% and typically 19.5%. A determined percentage of fucosylation in the test sample of less than the reference value indicates a diagnosis of HNF1A-MODY.

In certain embodiments, a difference in one or more HPLC peaks containing fucosylated glycans as determined when compared to the one or more reference values for those peaks in a control sample or a reference sample indicates a diagnosis of HNF1A-MODY or HNF4A- MODY. Each peak relates to different glycan groups, some of which contain antennary fucose and some of which do not. Accordingly, the peaks may be increased or decreased in HNF1A-MODY or HNF4A-MODY depending upon whether the peaks in question contain antennary fucose. Differences in one or more specific HPLC glycan peaks associated with antennary fucose (GP10, GP1 1 , GP13, GP14, GP15, GP16, DG7, DG8, DG9, DG10, DG1 1 , DG12 or DG13 defined as disclosed in Knezevic et al, Glycobiology 19: 1547-53) are indicative of HNF1A-MODY as presented in Table 2. In certain embodiments, determining the one or more

fucosylation levels in the test sample comprises determining one or more of these glycan peaks. In certain embodiments, additional differences in the one or more fucosylation levels in subjects with HNF1 A-MODY, subjects with type 2 diabetes and control subjects which may be used in the method of diagnosis of the present invention are set forth in Table 2. In certain embodiments, the one or more levels of fucosylation are determined in the test sample and compared with the one or more reference values for the corresponding type of fucosylation.

In certain embodiments, determining the one or more levels of fucosylation comprises using any suitable technology or combination of technologies, such as HPLC alone or in combination with mass spectrometry. Examples of techniques include chromatography (e.g., normal phase or weak anion exchange HPLC), mass spectrometry, gel electrophoresis (e.g., one or two dimensional gel electrophoresis), an immunoassay (e.g., immuno-PCR), ELISA, lectin ELISA, Western blot, lectin immunoassay, capillary electrophoresis and lectin chromatography. Individual proteins can be isolated and their fucosylation analyzed by HPLC, mass spectrometry, capillary electrophoresis, ELISA or some other immunochemical methods. ELISA with specific lectins or antibodies can be used to detect reduced fucose level on whole serum or plasma. Flow cytometry can be used to detect decreased fucosylation on intact cells.

In certain embodiments, the method further comprises a step of analyzing the test sample for the presence of one or more known biomarkers of HNF1 A-MODY or HNF4A-MODY, for example, one or more of C-reactive protein (CRP), 1 ,5-anhydroglucitol, pancreatic autoantibodies and C- peptide. The addition of analysis of the markers of the present invention to analysis of existing biomarkers, such as CRP, 1 ,5-anhydroglucitol, pancreatic autoantibodies and C-peptide, improves the capacity for clinical discrimination of all major diabetes subtypes, and supports more efficient use of molecular diagnostics and personalized management of diabetic individuals and their relatives.

The subject may be a mammal, in particular a human. In certain embodiments, the subject is suspected of having diabetes. The method of the present invention may be used to distinguish HNF1A-MODY or

HNF4A-MODY from other types of diabetes, such as type 2 diabetes, type 1 diabetes and types of MODY other than HNF1 A-MODY and HNF4A- MODY. In certain embodiments, the subject is suspected of being at risk of developing diabetes, for example, the subject has a family member who has diabetes.

In certain embodiments, the diagnosis may be confirmed by subsequent testing, for example, molecular genetic investigation such as sequencing HNF1A or HNF4A.

The present invention further extends to a diagnostic kit for carrying out the above described method. Accordingly, according to a further aspect of the present invention there is provided a diagnostic kit for diagnosing or screening for HNF1 A-MODY or HNF4A-MODY in a subject in need thereof, comprising means for determining one or more levels of fucosylation of the glycans in a test sample and details of one or more reference values to which the one or more levels of fucosylation of the glycans in the test sample can be compared in order to provide a diagnosis. In certain embodiments, the kit includes instructions outlining the one or more levels of fucosylation which may be determined and/or provides instructions on how a diagnosis can be reached.

The present invention further extends to a method for prognosticating HNF1A-MODY or HNF4A-MODY in a subject in need thereof,

comprisingproviding a glycan-containing test sample from the subject, determining one or more levels of fucosylation of the glycans in the test sample and comparing the one or more levels of fucosylation of the glycans in the test sample with one or more reference values as described above, wherein a reduction in the one or more fucosylation levels in the test sample compared to the one or more reference values which is larger than the threshold values indicated above indicates that the prognosis is poor. The present invention further relates to a method for determining the relevance of a mutation in HNF1A, comprising:

- providing a glycan-containing test sample from a subject with the mutation;

- determining one or more levels of fucosylation of the glycans in the test sample;

- determining the functional relevance of the mutation based on said comparison.

In certain embodiments, the one or more fucosylation levels are one or more of those described above and a reduction in the one or more levels fucosylation as described above indicates that the mutation is functionally relevant, that is, it affects the function of HNF1 A. In that case, the mutation may play a role in HNF1A-MODY. Description of the Figures

The present invention will now be described with reference to the following example which is provided for the purpose of illustration and is not intended to be construed as being limiting on the present invention, and further with reference to the figures as described briefly below.

Figure 1 shows representations of the structures of the major glycans in DG8 and DG9 HPLC peaks. Both are triantennary glycans composed of two N-acetylglucosamine residues (black squares), three mannose residues (white circles), an additional three N-acetylglucosamine residues and three galactose residues (white diamonds). The only difference is the presence of a terminal fucose residue (diamond with a dot) attached to one of the antennae of the glycans in DG9. Alpha glycosidic linkages are represented by a dotted line, beta by a continuous line;

Figure 2 shows dot histograms illustrating the DG9-glycan index in different diabetes subtypes and non-diabetic control subjects. P values are calculated by Mann-Whitney tests in comparison with

HNF1A-MODY subjects. Median value is highlighted by a black dashed line;

Figure 3 shows ROC curves illustrating the performance of the DG9-glycan index to discriminate: (A) HNF1A-MODY and Type 2 diabetes; (B) HNF1A-MODY and Type 1 diabetes; (C) HNF1A- MODY and other diabetes subtypes combined; (D) HNF1 A-MODY and GCK-MODY; (E) HNF1A-MODY and HNF4A-MODY; and (F) HNF1 A-MODY and non-diabetic control subjects; and Figure 4 shows pedigrees of families with HNF1A mutations: A: R203H mutation. B: D526N mutation. Proband indicated by arrow in both panels. Squares, male subjects; circles, female subjects; black, diabetes; grey, impaired fasting glycemia; oblique line indicates that the respective subject has deceased; and number in diamonds indicates number of offspring. NM HNF1A mutation, and NN no HNF1A mutation found.

Detailed description of the invention

Genome-wide association studies are providing novel insights into the genetic architecture and biological basis of many diseases, but immediate translation into clinical practice has been limited. In a genome-wide association study of the human plasma N-glycome, evidence of association involving common variants near the hepatocyte nuclear factor 1 -alpha (HNF1A) gene established HNF1A as a master regulator of plasma protein fucosylation. HNF1A has been shown to promote both the de novo and salvage pathways for the synthesis of GDP-fucose, and to regulate fucosyltransferase VI. HNF1 A therefore regulates the expression of key fucosyltransferase and fucose biosynthesis genes and acts as a master regulator of plasma protein fucosylation. HNF1A thereby controls the outer-arm (antennary) fucosylation of proteins with N-linked glycans through effects both on the supply and incorporation of fucose. The inventors hypothesized that inactivating mutations in HNF1A, such as those found in HNF1A-MODY, would display altered patterns of N-linked glycan fucosylation, such as decreased antennary fucosylation of circulating proteins. The reduced fucosylation identified by the inventors differs from glycosylation changes in total plasma or on individual glycoproteins in other disease where fucose is generally increased.

Changes in fucosylation are therefore suitable for use as a marker to diagnose HNF1 A-MODY and differentiate this from other types of diabetes, such as type-2 diabetes, type-1 diabetes and other types of MODY such as glucokinase-MODY (GCK-MODY). The inventors have now shown that measurement of glycan profiles and antennary

fucosylation allows more effective diagnostic screening of patients with early-onset diabetes for HNF1A-MODY and HNF4A-MODY, and more efficient deployment of definitive molecular diagnostics.

Unless otherwise defined, all technical and scientific terms used herein have the meaning commonly understood by a person who is skilled in the art in the field of the present invention.

Throughout the specification, unless the context demands otherwise, the terms "comprise" or "include", or variations such as "comprises" or

"comprising", "includes" or " including" will be understood to imply the inclusion of a stated integer or group of integers, but not the exclusion of any other integer or group of integers.

The term N-linked glycans designates asparagines-linked glycans. Fucosylation is a type of glycosylation involving fucose. Fucose is a sugar that is evolutionary relatively novel. It is, in fact, deoxy-galactose, where galactose has lost one hydroxy-group (-OH), thus creating a new functional element in primates and other higher organisms. In general, by control sample or reference sample is meant a sample having a known composition or content of a particular integer against which a comparison to the test sample is made. The control sample or reference sample does not contain a level of the fucosylation being detected consistent with HNF1 A-MODY or HNF4A-MODY. As used herein, a control sample refers to a sample from a subject not having diabetes. A reference sample refers to a sample from a subject with a type of diabetes other than HNF1 A-MODY or HNF4A-MODY. Reference value may refer to a level of fucosylation in either a control sample or a reference sample.

EXAMPLE METHOD

Plasma protein N-glycans were profiled in 33 HNF1A-MODY and 41 type 2 diabetes subjects. The glycan index providing optimum discrimination between diabetes subtypes was further examined in subjects with HNF1A- MODY (n=189), Glucokinase-MODY (GCK-MODY) (n=1 17), HNF4A- MODY (n=40), type 1 diabetes (n=98), type 2 diabetes (n=167) and non- diabetic controls (n=98).

Subjects for initial and validation studies

Subjects carrying mutations in HNF1A (n=222), glucokinase {GCK, n=1 17) and hepatocyte nuclear factor 4-alpha (HNF4A, n=40) were recruited from five European centres. MODY cases had an established heterozygous loss-of-function mutation confirmed by sequencing in a certified diagnostic centre. All MODY mutations were considered pathogenic if they met one of these criteria: previously-published reports; presence of a truncating mutation; co-segregation of the mutation with a MODY phenotype within the family; absence of the variant in normal chromosomes. Additionally, 208 subjects were recruited with clinically-labelled type 2 diabetes diagnosed <45 years, 98 with clinically-labelled type 1 diabetes and 98 non-diabetic control subjects. Subjects were divided into an initial study and an independent validation dataset (see below). Full details of the subjects are available in Table 1 and are as follows. Subjects from Denmark

All MODY patients were identified in the molecular genetic diagnostic unit at Steno Diabetes Centre. The MODY samples comprise individuals with a mutation in either HNF1A (n=80 from 22 families), GCK (n=1 1 from 6 families) or HNF4A (n=23 from 6 families). All patients with type 2 diabetes and type 1 diabetes were identified from the outpatient clinic at Steno Diabetes Centre. The type 2 diabetes patients (n=101 ) were diagnosed with diabetes < 45 years of age and additional criteria are C- peptide positive, no requirement for permanent insulin treatment within 12 months of diagnosis and negative GAD antibodies. None of the type 2 diabetes patients met clinical criteria for MODY diagnostic testing. The type 1 diabetes patients (n=28) had HbA1 c >8%, GAD antibody positive and/or C-peptide <10 pmol/L, and diabetes for 3 years or more. All healthy control individuals (n=98) had previously participated in the population based sampled Inter99 study conducted at Research Centre for Prevention and Health, Lustrum University Hospital, Denmark. For this study all healthy control individuals were invited for re-examination at Steno Diabetes Centre. The healthy control individuals had no history of diabetes and had fasting plasma glucose <7mmol/L on the day of examination. Informed written consent was obtained from all subjects before participation. All subjects were of North European ethnicity. The Danish samples comprised serum samples and were all used in the validation study. The study protocol was approved by the regional Ethical Committees and was in accordance with the principles of the Declaration of Helsinki II.

Subjects from Norway

The MODY subjects were recruited through the Norwegian MODY

Registry, comprising MODY patients from all over Norway. The MODY subjects have mutations (confirmed by a certified Norwegian diagnostic centre) in HNF1A (n=48 from 36 families), GCK (n=29 from 18 families) or HNF4A (n=10 from 4 families). Type 2 diabetes patients (n=16) were diagnosed <45 years and recruited at the diabetes outpatient clinic at Haukeland University Hospital. All subjects were of North European ethnicity. The Norwegian samples comprised serum samples and were all used in the validation study. The study was approved by the regional ethics committee of Western Norway. All participants gave written informed consent. Subjects from Slovakia

Subjects fulfilling the MODY clinical diagnostic criteria were actively searched in diabetes outpatient clinics throughout Slovakia. Patients with fasting hyperglycemia were tested for GCK mutations first and if negative, then for HNF1A and HNF4A mutations; patients with higher HbA1 c and requiring insulin treatment were tested for HNF1A mutations first and then if negative, were tested for HNF4A and GCK mutations. All patients from Slovakia carrying HNF1A (n=46), HNF4A (n=2) or GCK (n=77) mutations with relevant biochemical and clinical data were included in this study. All subjects were white Europeans. The Slovakian samples comprised EDTA plasma samples and were all used in the validation study. This study was approved by the Ethics Committees in the Children Faculty Hospital in Bratislava and the National Institute of Endocrinology and Diabetes in Lubochna. Subjects from Oxford, UK

Subjects were ascertained from the South of England. The MODY samples comprise subjects with a mutation (confirmed by sequencing in a certified UK diagnostic centre) in either HNF1A (n=34 from 17 families) or HNF4A (n=5 from 3 families). Type 2 diabetes subjects were selected from the Young Diabetes in Oxford (YDX) study, comprising subjects diagnosed with diabetes <45 years of age. Criteria for diagnosis were: C- peptide positive, no requirement for permanent insulin within 3 months of diagnosis and negative GAD antibodies. Type 1 diabetes subjects were selected from the YDX study and criteria for diagnosis were permanent insulin therapy from diagnosis and C-peptide<0.1 nmol/l and/or positive glutamic acid decarboxylase antibodies. Subjects with either a clinical label of type 1 diabetes or type 2 diabetes did not meet clinical criteria for MODY diagnostic testing or had been tested and were negative for mutations in HNF1A, HNF4A or GCK.

The Oxford samples were all plasma samples. 19 of the HNF1A-MODY subjects and 41 of the type 2 diabetes subjects were used in the initial study and the remainder were used in the validation study. One HNF1 A- MODY individual, 6 type 2 diabetes patients and 3 type 1 diabetes patients were of non-European ethnicity (4 Asian and 6 Black). The study was approved by the Oxfordshire Local Research Ethics Committee and all subjects gave informed consent.

Subjects from Edinburgh, UK

The MODY samples comprise subjects with a mutation (confirmed by sequencing in a certified UK diagnostic centre) in HNF1A (n=14). The Edinburgh samples were all plasma samples and used in the initial study. The study was approved by the Lothian Local Research Ethics Committee and all subjects gave written informed consent.

Subjects from Vis and Korcula, Croatia

The CROATIA-Vis study includes 1008 Croatians, aged 18-93 years, who were recruited from the villages of Vis and Komiza on the Dalmatian island of Vis during 2003 and 2004 within a larger genetic epidemiology program. The CROATIA-Korcula study includes 969 Croatians between the ages of 18 and 98. The field work was performed in 2007 and 2008 in the eastern part of the island, targeting healthy volunteers from the town of Korcula and the villages of Lumbarda, Zrnovo and Racisce. Both studies recruited adult individuals within a community irrespective of any specific

phenotype. Fasting blood samples were collected, biochemical and physiological measurements taken and questionnaire data for medical history as well as lifestyle and environmental exposures were collected following similar protocols. Plasma sample were available for glycan analysis. All studies conformed to the ethical guidelines of the 1975 Declaration of Helsinki and were approved by appropriate ethics boards with all participants signing informed consent prior to participation.

Diabetic status was determined by report of diabetes recorded by a physician in the medical history or report of treatment with an anti-diabetic medication regardless of glycemic status or record of HbA1 c measurement greater or equal to than 6.5%.

Subjects from ORCADES

The Orkney Complex Disease Study (ORCADES) was performed in the Scottish archipelago of Orkney and collected data between 2005 and 201 1 . Data from plasma samples for 889 participants aged 18 to 100 years from a subgroup of ten islands, was available for this study.

ORCADES conformed to the ethical guidelines of the 1975 Declaration of Helsinki and was approved by appropriate ethics boards with all participants signing informed consent prior to participation. Diabetic status was determined by any one of: a self-report of diabetes, treatment with an anti-diabetic medication regardless of glycemic status or record of HbA1 c measurement greater than or equal to than 6.5%.

Table 1. Clinical characteristics of subjects included in the initial and validation studies

Subjects with diabetes only * (i.e. excluded non-diabetic mutation carriers)

Normally distributed variables reported as mean ± SD, all others reported as median (25^th-75 ^h centiles)

# Unavailable for subjects from Edinburgh

Glycan release, labelling and analysis

Glycan release, labelling and analysis using hydrophilic interaction high performance liquid chromatography and sialidase digestion was performed as previously reported. In particular, plasma proteins were immobilized in a block of SDS-polyacrylamide gel and N-glycans were released by digestion with recombinant N-glycosidase F. This was done in a 96-well microtitre plate to achieve the best throughput of sample preparation. After extraction, glycans were fluorescently labelled with 2- aminobenzamide. Aliquots of the 2AB-labeled glycan pool were dried down in 96-well PCR plates. To these, the following was added: 1 μΙ of 500 mM sodium acetate incubation buffer (pH 5.5), 1 μΙ (0.005 units) of ABS, Arthrobacter ureafaciens sialidase (releases a2-3,6,8 sialic acid, Prozyme) and H₂0 to make up to 10 μΙ. This was incubated overnight (16-18 h) at 37 °C and then passed through AcroPrep™ 96 Filter Plates, 350 μΙ_ well, 10K (Pall Corporation, Port Washington, NY, USA) before applying to the HPLC.

The glycan quantification was performed using high precision liquid chromatography (HPLC) and a GlycoBase software as disclosed in Royle, L., et al., HPLC-based analysis of serum N-glycans on a 96-well plate platform with dedicated database software. Analytical Biochemistry, 2008. 376(1 ): p. 1 -12. Released glycans were subjected to hydrophilic interaction high performance liquid chromatography (HILIC) on a 250 x 4.6 mm i.d. 5μΓη particle packed TSKgel Amide 80 column (Tosoh Bioscience, Stuttgart, Germany) at 30°C with 50 mM formic acid adjusted to pH 4.4 with ammonia solution as solvent A and acetonitrile as solvent B. 60 minute runs were performed with fluorescence detector set with excitation and emission wavelengths of 330 and 420 nm, respectively. The system was calibrated using an external standard of hydrolyzed and 2-AB-labeled glucose oligomers from which the retention times for the individual glycans were converted to glucose units (GU).

Chromatograms from fluorescently-labelled plasma glycans were separated into 16 glycan groups (GP-series) and 13 desialylated glycan groups (DG-series), giving a total of 29 features (Table 2). The levels of glycans present in each peak were expressed as the percentage of the total plasma glycome. Levels of glycans sharing the same structural features were approximated by adding structures having the same characteristic. Individual glycan structures present in each glycan group were reported previously and are shown in Table 3. Glycan features were defined as shown in Table 4. Glycan analysis was performed in two centres, the National Institute for Bioprocessing Research and Training (NIBRT; Dublin, Ireland) and the Glycobiology laboratory of Genos Ltd (Zagreb, Croatia). Both laboratories used the same columns and separation conditions, and have previously demonstrated reproducibility of analytic results within and between laboratories. Table 2. Glycan measures in the initial study

Glycan HNF1A- Type 2 diabetes (n=41) P value measure MODY

GP1 0.11 (0.07- 0.08 (0.07-0.11) 0.057

GP2 2.65 (1.99- 2.25 (1.93-2.82) 0.042

GP3 1.69 (1.36- 1.50 (1.35-1.69) 0.101

GP4 4.72 (3.90- 3.96 (3.31-4.53) 0.005

GP5 2.80 (2.59- 2.81 (2.65-3.09) 0.644

GP6 4.25 (3.83- 3.90 (3.36-4.35) 0.062

GP7 13.24 (12.68- 13.56 (12.83-14.96) 0.267

GP8 8.75 (7.88- 8.11 (7.51-8.84) 0.058

GP9 34.08 (32.97- 34.42 (33.40-36.34) 0.561

GP10 5.18 (4.58- 6.63 (5.88-7.79) 1.3 x 10^"B

GP11 3.03 (2.71- 2.49 (2.20-3.08) 0.003

GP12 3.89 (3.57- 4.18 (3.64-4.52) 0.832

GP13 6.51 (5.23- 4.51 (3.89-6.08) 1.7 x 10^"b

GP14 5.46 (5.14- 7.92 (6.94-8.94) 4.2 x 10^"1U

GP15 1.03 (0.88- 0.71 (0.60-0.90) 1.3 x 10^"B

GP16 1.39 (1.28- 1.55 (1.28-1.79) 0.274

DG1 0.13 (0.09- 0.11 (0.09-0.14) 0.192

DG2 2.75 (2.34- 2.74 (2.18-3.10) 0.186

DG3 2.46 (2.10- 2.29 (1.94-2.47) 0.009

DG4 5.50 (4.93- 4.93 (4.13-5.76) 0.017

DG5 50.44 (49.25- 50.95 (49.76-53.96) 0.185

DG6 13.47 (12.07- 13.25 (12.07-14.85) 0.695

DG7 0.65 (0.45- 1.05 (0.91-1.19) 1.5 x 10^_

DG8 15.16 (13.43- 13.40 (12.10-15.85) 0.007

DG9 1.32 (0.95- 3.74 (2.95-4.61) 1.8 x 10^"B

DG10 0.65 (0.60- 0.89 (0.78-1.01) 1.8x 10^_

DG11 4.25 (3.68- 3.04 (2.37-3.45) 1.4 x 10^"B

DG12 0.37 (0.24- 0.91 (0.66-1.15) 7.6 x 10^"B DG13 1 .14 (0.92- 1 .19 (0.99-1 .31 ) 0.605

DG9/(DG8 0.08 (0.05- 0.29 (0.22-0.37) 1 .1 x 10^"'

DG7/(DG6 0.05 (0.03- 0.08 (0.06-0.09) 9.0 x 10^"B

Data presented as median (25 -75 centiles)

P value calculated using Mann-Whitney U tests

Table 3. Glycan structures present in different HPLC peaks

Table 4. Glycan features

Abbreviations

Structure abbreviations: all N-glycans have two core GlcNAcs; F at the start of the abbreviation indicates a core fucose a1-6 linked to the inner GlcNAc; Mx, number (x) of mannose on core GlcNAcs; D1 indicates that the a1-2 mannose is on the Mana1-6Mana1-6 arm, D2 on the Manal- 3Mana1-6 arm, D3 on the Mana1-3 arm of M6 and on the Manal- 2Mana1-3 arm of M7 and M8; Ax, number of antenna (GlcNAc) on trimannosyl core; A2, biantennary with both GlcNAcs as β1-2 linked; A3, triantennary with a GlcNAc linked β1-2 to both mannose and the third GlcNAc linked β1-4 to the a1-3 linked mannose; A4, GlcNAcs linked as A3 with additional GlcNAc β1-6 linked to a1-6 mannose; B, bisecting GlcNAc linked β1 -4 to β1 -3 mannose; Gx, number (x) of β1 -4 linked galactose on antenna; [3]G1 and [6]G1 indicates that the galactose is on the antenna of the a1 -3 or l -6 mannose; F(x), number (x) of fucose linked a1 -3 to antenna GlcNAc; Lac(x), number (x) of lactosamine (Gai i -4GlcNAc) extensions; Sx, number (x) of sialic acids linked to galactose; the numbers 3 or 6 or in parentheses after S indicate whether the sialic acid is in an a2- 3 or oc2-6 linkage. If there is no linkage number, the exact link is unknown.

Study Design and Analysis

All glycan traits were compared in an initial study of 33 HNF1 A-MODY and 41 type 2 diabetes subjects using a Mann-Whitney test. Based on these results (Table 2), the glycan ratio DG9/(DG8+DG9), hereafter referred to as the "DG9-glycan index", was chosen for an independent validation study.

The validation study was performed using 189 HNF1 A-MODY, 1 17 GCK- MODY, 40 HNF4A-MODY, 167 type 2 diabetes and 98 type 1 diabetes cases plus 98 non-diabetic controls. There was no overlap between the initial and validation studies. Evidence for important covariates was sought through analysis of parameters including age, gender, body mass index (BMI), high density lipoprotein, triglycerides, sample origin, processing laboratory and sample type (plasma vs. serum). An adjusted model incorporated significant covariates (age, BMI, sample origin, processing laboratory), as well as those already-known to affect specific glycan traits (gender). An additional analysis was performed without covariates. The use of DG9-glycan index as a discriminator of diabetes subtypes was analyzed using Receiver Operator Characteristic (ROC) curves from which the C-statistic was obtained. Performance of the DG9- glycan index as a clinical tool was evaluated by calculating sensitivity and specificity and other measures for the detection of HNF1 A-MODY at various thresholds. In particular, post-test diagnostic probabilities were estimated based on data from an etiological investigation of young adults with diabetes from the UK that indicated pre-test probabilities of 4% for HNF1A-MODY in early-onset type 2 diabetes and 1 % in type 1 diabetes. All analyses were performed using SPSS v17.0.

HNF1A Sequencing

In the subsequent case-finding study, the value of the DG9-glycan index as a screen for identifying HNF1 A-MODY was evaluated in a set of diabetic individuals who had not previously been suspected of having an HNF1A mutation. These were selected from general population cohorts from Croatia and Scotland (total n=2705, 183 with diabetes), with previously measured glycan profiles, as well as diabetic subjects in the initial or validation studies who had not been diagnosed with HNF1 A- MODY. Forty diabetic subjects with a DG9-glycan index <0.16 were selected, with current clinical labels of type 1 diabetes (n=5), type 2 diabetes (n=28, including 25 from the population cohorts) and GCK-MODY (n=7). Ages of diagnosis ranged from 5 to 79 years. The l O exons of HNF1A were amplified by PCR and bidirectional sequencing performed using M13 primers and a Big Dye Terminator Cycler Sequencing kit v1 .1 (Applied Biosystems, Warrington, UK). Reactions were analyzed on an ABI 3730 capillary sequencer (Applied Biosystems), and results compared with the reference sequence (NM_000545.3) using Mutation Surveyor v3.24 (SoftGenetics, Cambridge, UK). Mutation testing was undertaken in family members where available to establish co-segregation.

The study was performed according to the latest version of the Helsinki Declaration. Funding agencies had no influence on study design, collection of samples, analysis and interpretation of data. RESULTS

An index measuring the proportion of triantennary fucosylated glycans (glycan peaks DG9/[DG9+DG8]) was substantially lower in HNF1A-MODY than other diabetes subtypes in both initial and validation groups (P<5x10^" ⁷for all pairwise comparisons). C-statistic measures of discriminative accuracy were 0.94 for HNF1 A-MODY vs type 1 diabetes and 0.90 for HNF1 A-MODY vs type 2 diabetes. In the clinical setting of early-onset diabetes, this was estimated to translate into 88% sensitivity and 83% specificity to discriminate HNF1 A-MODY from other causes of diabetes. HNF1A sequencing in 40 diabetic subjects (both early- and late-onset) who displayed glycan profiles similar to those of known HNF1 A-MODY cases revealed two individuals with previously-unsuspected HNF1A mutations. HNF1 A-MODY and measures of antennary fucosylation

In the initial study of 33 subjects with HNF1 A-MODY and 41 with early- onset type 2 diabetes, marked differences were found in plasma glycome profiles. In all, 15 of 29 glycan measures differed significantly (P<0.05) between the two groups (Table 2). Patterns were consistent with the known effects of HNF1 A on fucosylation, in that subjects with loss-of- function mutations in HNF1A were characterized by an increase in the proportion of glycans without antennary fucose.

Validation study

Whilst several alternative glycan measures showed promising

discrimination in the initial study, a subset for further validation was selected. In particular, DG9 and DG8 were focused on as measures of triantennary glycans with and without antennary fucose, respectively (Figure 1 ). The DG9-glycan index (DG9/[DG8+DG9]) therefore

summarizes the proportion of triantennary glycans which is fucosylated. As well as consistency with the existing data on HNF1A effects on fucosylation, and strong evidence from the initial study (Table 2), triantennary glycans are not affected by the removal of fibrinogen during coagulation allowing validation studies to include both serum and plasma samples.

The distributions of DG9-glycan index measures for the 709 individuals in the validation study (none of them included in the initial study) differed significantly between diagnostic categories (Figure 2; Table 5). Levels were substantially lower in subjects with HNF1A-MODY (median 0.09

[interquartile range 0.06-0.13]) than in those with early-onset type 2 diabetes (0.25 [0.18-0.33], P =1x10^"39 vs. HNF1A-MODY), type 1 diabetes (0.28 [0.20-0.34], P=6x10^"34 vs. HNF1 A-MODY) or GCK-MODY (0.25

[0.18-0.30], P=5x10^"31 vs. HNF1 A-MODY). DG9-glycan index levels in those with HNF1 A-MODY were also lower when compared against controls (0.24 [0.19-0.29], P=5x10^"32 vs. HNF1 A-MODY), and against all other diabetic patients combined (0.25 [0.18-0.31 ], P=5x10^"55 vs. HNF1A- MODY). Consistent with evidence that HNF4A also regulates fucosylation, HNF4A-MODY cases showed DG9-glycan index levels intermediate between those in HNF1 A-MODY and other forms of diabetes (0.18 [0.09- 0.24], P=4x10^"7 vs. HNF1A-MODY).

Table 5. Levels of DG9-glycan index and estimated marginal means for the adjusted DG9-glycan index across diabetes subtypes and healthy controls in the Validation study

*Estimated marginal means (95% CI) DG9-glycan index adjusted for covariates (age at sampling, gender, BMI, processing laboratory and sample origin)

P value calculated using Mann-Whitney U tests Adjustment for significant covariates had no appreciable impact on the magnitude or significance of differences in DG9-glycan index between groups (Table 5).

Receiver operating characteristic (ROC) curve analyses

Statistically convincing between-group differences do not necessarily translate into a clinically-valid screening test. C-statistic measures of discriminative accuracy derived from ROC curve analyses (Figure 3; Table 6) reached 0.94 for HNF1A-MODY against type 1 diabetes, and 0.90 for HNF1A-MODY against early-onset type 2 diabetes. Similar discrimination was observed for the comparison of HNF1A-MODY and GCK-MODY (C- statistic 0.89), but the DG9-glycan index performed less well in

differentiating HNF4A- and HNF1A-MODY (C-statistic 0.76). These measures were not affected by sample type (serum vs plasma: Table 6). Table 6. C-statistic for the DG9-glycan index and the DG7/(DG5+DG6) ratio to discriminate HNF1A-MODY subjects from those with other diabetes etiologies and healthy controls across sample types

All samples Plasma samples Serum samples only only

C- SE Sig. C- SE Sig. C- SE Sig

STAT STAT STAT

HNF1A DG 0.905 0.01 1x1 0.921 0.0 3x1 0.898 0.0 6x

-MODY 9- 6 Q-39 27 0-14 21 10^" vs. 27

giy

Other glvcan measures

As described earlier, primary validation analyses focused on the DG9- glycan index. However, availability of full glycome profiles for the validation samples allowed the relative performance of other measures highlighted in the initial study to also be explored. Other glycan ratios, such as DG7/(DG5+DG6) representing the proportion of biantennary fucosylated glycans, offered good discrimination between diabetes subtypes (Table 6). The DG7/(DG5+DG6) ratio provided near perfect discrimination between HNF1 A-MODY and type 2 diabetes (C-statistic >0.99) in plasma samples, but performed less well in analyses of serum samples (C-statistic 0.77).

Use of the DG9-glvcan index in clinical practice

To examine the performance of the DG9-glycan index as a diagnostic screen in clinical practice, analyses were based on UK data showing that unrecognized HNF1 A-MODY subjects account for approximately 4% of early-onset type 2 diabetes (diagnosed <45 years) and 1 % of type 1 diabetes. Based on the validation study, it was estimated that a diagnostic threshold for the DG9-glycan index of 0.16 confers 88% sensitivity and 81 % specificity for the discrimination of HNF1 A-MODY from early-onset type 2 diabetes, and 88% sensitivity and 88% specificity for equivalent comparisons with type 1 diabetes (Table 7). In contrast, an age of diagnosis of diabetes below 25 years, which is the most widely-used diagnostic feature for MODY, has lower sensitivity (64%) though higher specificity (99%) for the discrimination of HNF1 A-MODY from early-onset type 2 diabetes. It was calculated that a patient with diabetes diagnosed up to 45 years and who has an existing clinical label of type 2 diabetes who is found to have a DG9-glycan index <0.16 has a post-test probability of harbouring an underlying HNF1A mutation of 16%, whereas the same patient with a DG9-glycan index above 0.16 has a 1 % post-test probability of having unrecognized HNF1 A-MODY. Table 7 demonstrates the impact of different DG9-glycan index thresholds and pre-test probabilities on these estimates.

Table 7. Sensitivity and specificity figures at various thresholds of the DG9-glycan index

HNF1 A-MODY vs. Type 2 diabetes subjects

Threshold Sensit Sped Positive Post- Neg. Post-test for ivity ficity liketest likeprobability

HNF1A (%) (%) lihood probab lihood (%) sequencin ratio ility ratio

g (%)

Diabetes 64.0 98.8 53.33 69.0 0.36 1 .5 diagnose

d up to

25y

DG9- 95.8 61 .7 2.50 9.4 0.07 0.3 glycan

index

<0.22 DG9- 92.6 68.3 2.92 10.8 0.1 1 0.5 glycan

index

<0.20

DG9- 90.5 74.9 3.61 13.1 0.13 0.5 glycan

index

<0.18

DG9- 87.8 80.8 4.65 16.2 0.15 0.9 glycan

index

<0.16

DG9- 79.9 86.2 5.79 19.4 0.23 0.9 glycan

index

<0.14

DG9- 73.0 89.8 7.23 23.2 0.30 1 .2 glycan

index

<0.12

DG9- 59.3 92.8 8.24 25.6 0.44 1 .8 glycan

index

<0.10

Diabetes 89.4 88.6 7.84 24.6 0.12 0.5 diagnose

d up to

25y or

DG9- glycan index

<0.12

Diabetes 95.2 80.2 4.81 16.7 0.06 0.2 diagnose

d up to

25y or

DG9- glycan

index

<0.16

HNF1A-MODY vs. Type 1 diabetes subjects

DG9- 95.8 69.4 3.13 3.1 0.06 0.1 glycan

index

<0.22

DG9- 92.6 78.6 4.33 4.2 0.09 0.1 glycan

index

<0.20

DG9- 90.5 84.7 5.92 5.6 0.1 1 0.1 glycan

index

<0.18

DG9- 87.8 87.8 7.20 6.8 0.14 0.1 glycan

index

<0.16

DG9- 79.9 91 .8 9.74 9.0 0.22 0.2 glycan index

<0.14

DG9- 73.0 93.9 1 1 .97 10.8 0.29 0.3 glycan

index

<0.12

DG9- 59.3 98.0 29.65 23.0 0.42 0.4 glycan

index

<0.10

*Based on pre-test probability of 4% misdiagnosed HNF1A-MODY cases in clinically- labelled young adult-onset Type 2 diabetes and 1 % misdiagnosed HNF1A-MODY cases in clinically-labelled Type 1 diabetes. Positive and negative likelihood ratios are shown for carrying an HNF1A mutation in the context of either a positive or negative DG9-glycan index respectively.

HNF1A-MODY case-finding in diabetic subjects

Diagnostic HNF1A sequencing was performed in 40 subjects with diabetes not known to have HNF1A-MODY and with a DG9-glycan index <0.16. Cases with a wide range of ages at diagnosis (5 to 79 years) who were considered to have a variety of causes for their diabetes were examined. Two individuals were found to have HNF1A mutations: in each case, those mutations had been previously-reported as causal for HNF1A-MODY. The first (Proband #1 : Figure 4A) was heterozygous for a c.608G>A

p.Arg203His missense mutation in exon 3 and had a clinical phenotype consistent with HNF1 A-MODY, including a 2-generation history of early- onset diabetes. Although previously assumed to have type 1 diabetes, this patient had residual endogenous insulin secretion 17 years after diagnosis (C-peptide, 0.27nmol/L). The second (Proband #2: Figure 4B) was heterozygous for a c.1576G>A p.Asp526Asn missense mutation in exon 8. The clinical phenotype in this case was more ambiguous: the proband has diet-treated diabetes diagnosed at 63 years and the mutation did not show strong co-segregation with disease. However, it is well-known that mutations in the 3' exons of HNF1A (which are not expressed in all isoforms) can result in a later age of onset. The pathogenicity of this mutation is supported by SIFT and PolyPHEN analyses. The values of DG9-glycan index in these two individuals were 0.05 (ranked 3^rd of those selected for HNF1A sequencing) and 0.08 (ranked 7^th), respectively.

CONCLUSION

The DG9-glycan index allowed clinically-useful discrimination between HNF1 A-MODY and other diabetes subtypes. Clinical deployment of this assay can improve case-finding through more precise targeting of subsequent molecular genetic investigation. Potential advantages of the DG9-glycan index in this context include stability over time and differentiation of HNF1 A-MODY from both common types of diabetes. Although there is some indication that glycan profiles are affected by acute inflammation, all four HNF1 A-MODY subjects in the present study with elevated hsCRP levels (>10mg/l) had DG9-glycan indices below 0.16. This suggests the DG9-glycan index is less prone to spurious elevation from intercurrent infection, although this will require confirmation in larger numbers. The ability to discriminate between HNF1 A-MODY and type 1 diabetes in subjects with recently-diagnosed diabetes is likely to be particularly important because incorrect diagnostic classification can lead to the unwarranted decision to recommend lifelong therapy with exogenous insulin. The addition of the DG9-glycan index to existing biomarkers such as hsCRP, 1 ,5-anhydroglucitol, pancreatic autoantibodies and C-peptide therefore improves the capacity for clinical discrimination of all major diabetes subtypes, and should support more efficient use of molecular diagnostics and personalized management of diabetic individuals and their relatives.

The greatest clinical impact of the DG9-glycan index is likely to be as a diagnostic adjunct in those with diabetes diagnosed in the second to fourth decades, in whom the proportion of HNF1A-MODY is highest (up to 5%) and clinical discrimination most difficult. From a clinical perspective, the observation that adjustment for covariates (such as age, gender, sample type) does not alter the magnitude or significance of differences between diabetic subgroups, facilitates incorporation into simple diagnostic protocols.

Although the invention has been described in connection with specific preferred embodiments, it should be understood that the invention as claimed should not be unduly limited to such specific embodiments.

Indeed, various modifications of the described modes of carrying out the invention which are obvious to those skilled in the art are intended to be covered by the present invention.

Claims

1 . A method for diagnosing or screening for HNF1 A-MODY or HNF4A- MODY in a subject in need thereof, comprising:

- providing a glycan-containing test sample from the subject;

- comparing the one or more levels of fucosylation of the glycans in the test sample with one or more reference values; and - providing a diagnosis based on said comparison.

2. The method as claimed in claim 1 wherein the test sample comprises glycans of plasma proteins.

3. The method as claimed in claim 1 wherein the test sample comprises glycans from whole serum.

4. The method as claimed in claim 1 wherein the test sample comprises glycans of one or more isolated glycoproteins.

5. The method as claimed in claim 4 wherein the one or more isolated glycoproteins are one or more plasma proteins.

6. The method as claimed in any preceding claim wherein the method further comprises a step of releasing glycans from glycoproteins in the test sample prior to determining the one or more levels of fucosylation of the glycans in the test sample.

7. The method as claimed in any preceding claim wherein the one or more levels of fucosylation comprises the proportion of glycans with antennary fucose.

8. The method as claimed in claim 7 wherein the glycans are selected from at least one of the group consisting of biantennary glycans, triantennary glycans and tetraantennary glycans.

9. The method as claimed in claim 8 wherein the glycans are triantennary glycans.

10. The method as claimed in claim 8 wherein the glycans are biantennary glycans.

1 1 . The methods as claimed in claim 10 wherein the test sample comprises glycans of plasma proteins.

12. The method as claimed in any one of claims 1 to 6 wherein the one or more levels of fucosylation comprises a ratio of antennary fucose to core fucose.

13. The method as claimed in any preceding claim wherein providing a diagnosis based on said comparison comprises determining whether the one or more levels of fucosylation of the glycans in the test sample are below the one or more reference values, wherein a finding that the one or more levels of fucosylation of the glycans in the test sample are below the one or more references values indicates a diagnosis of HNF1A-MODY or HNF4A-MODY.

14. The method as claimed in any preceding claim wherein the one or more reference values are one or more corresponding fucosylation levels in one or more subjects who do not have any form of diabetes.

15. The method as claimed in any one of claims 1 to 13 wherein the one or more reference values are one or more corresponding fucoslyation levels in one or more subjects who have a type of diabetes other than HNF1A-MODY or HNF4A-MODY.

16. The method as claimed in claim 14 or 15 wherein a finding that the one or more fucosylation levels of the glycans of one or more individual proteins in the test sample are below 70% of the one or more reference values indicates a diagnosis of HNF1 A-MODY or HNF4A-MODY.

17. The method as claimed in claim 14 or 15 wherein a finding that the one or more fucosylation levels of one or more classes of glycans in the test sample are below 70% of the one or more reference values indicates a diagnosis of HNF1 A-MODY or HNF4A-MODY.

18. The method as claimed in claim 17 wherein the one or more classes of glycans are selected from at least one of the group consisting of biantennary glycans, triantennary glycans and tetrantennary glycans.

19. The method as claimed in claim 13 when dependent on claim 7.

20. The method as claimed in claim 13 when dependent on claim 9.

21 . The method as claimed in claim 20 wherein the one or more reference values is 16% and a finding that the one or more reference values is below 16% indicates a diagnosis of HNF1A-MODY.

22. The method as claimed in claim 20 wherein the one or more reference values is 14% and a finding that the one or more reference values is below 14% indicates a diagnosis of HNF1A-MODY.

23. The method as claimed in claim 13 when dependent on claim 10.

24. The method as claimed in claim 23 wherein the one or more reference values is 6% and a finding that the one or more reference values is below 6% indicates a diagnosis of HNF1A-MODY.

25. The method as claimed in claim 13 when dependent on claim 7 wherein the glycans are tetraantennary glycans.

26. The method as claimed in claim 25 wherein the one or more reference values is 21 .5% and a finding that the one or more reference values is below 21 .5% indicates a diagnosis of HNF1 A-MODY.

27. The method as claimed in claim 13 when dependent on claim 12.

28. The method as claimed in claim 27 wherein the one or more reference values is 0.065 and a finding that the one or more reference values is below 0.065 indicates a diagnosis of HNF1 A-MODY.

29. The method as claimed in claim 7 when dependent on claim 3 wherein the one or more reference values is 1 .8% and a finding that the proportion of glycans with antennary fucose in the test sample is below 1 .8% indicates a diagnosis of HNF1 A-MODY.

30. The method as claimed in claim 9 wherein a finding that the proportion of triantennary glycans with antennary fucose in the test sample is between the one or more references values of 15% and 20% indicates a diagnosis of HNF4A-MODY.

31 . The method as claimed in claim 30 wherein a finding that the proportion of triantennary glycans with antennary fucose in the test sample is between 17% and 19% indicates a diagnosis of HNF4A-MODY.

32. A method for determining the relevance of a mutation in HNF1 A, comprising:

- providing a glycan-containing test sample from a subject with the mutation;

33. The method as claimed in claim 32 wherein determining the one or more levels of fucosylation of the glycans in the test sample comprises determining the proportion of glycans with antennary fucose.

34. The method as claimed in claim 33 wherein the glycans are triantennary glycans.

35. The method as claimed in any one of claims 32 to 34 wherein a finding that the one or more levels of fucosylation of the glycans in the test sample are below the one or more references values indicates the mutation is functionally relevant.