USE OF THE ADIPOQ GENE EXPRESSION LEVEL TO CLASSIFY A SUBJECT INTO RISK GROUPS IN THE PREDICTION OR DIAGNOSIS OF TYPE II DIABETES MELLITUS
Patent Information
- Application Number
- DE502018015327
- Authority / Receiving Office
- DE · DE
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2017-11-24
- Filing Date
- 2018-11-19
- Publication Date
- 2024-11-28
- Estimated Expiration
- 2038-11-19
AI Technical Summary
Current methods for diagnosing and subclassifying type II diabetes mellitus are inadequate, as they rely on BMI and blood parameters that do not accurately predict insulin resistance or provide personalized treatment options, lacking a reliable method for early detection and differentiation of disease subclasses.
A method involving the determination of ADIPOQ gene expression levels in conjunction with HMGA2 and other markers to classify individuals into risk groups, allowing for more precise prognosis and diagnosis of type II diabetes, enabling tailored therapeutic approaches.
This approach enables further differentiation within the group of type II diabetes patients, facilitating appropriate therapies by identifying specific subtypes of the disease, potentially improving treatment outcomes and early detection of insulin resistance.
Description
[0001] It further relates to a method for the prognosis and / or diagnosis of diabetes mellitus type II disease for classifying a subject into risk groups, wherein the gene expression level of the gene ADIPOQ is determined and the test subject is then classified into risk groups taking this gene expression level into account.
[0002] The prevalence of overweight and obesity, as well as associated diseases such as type 2 diabetes mellitus (T2D), is a growing global problem. According to the World Health Organization (WHO), overweight is defined as a body mass index (BMI) of ≥ 25 kg / m 2 < - < 30 kg / m 2 <, and obesity as a BMI of ≥ 30 kg / m 2 <. Obesity is further divided into grades: grade I obesity (BMI = 30 - 34.9 kg / m 2 <), grade II obesity (BMI = 35 - 39.9 kg / m 2 <), grade III obesity (BMI ≥ 40 kg / m 2 <), and super-obesity (BMI ≥ 50 kg / m 2 <). A high BMI is not always a necessary cause of T2D; rather, BMI can only serve as a very rough predictor. Even people of normal weight (BMI < 25 kg / m²) develop T2D, but the exact scientific background as to why people of normal weight also develop diseases more commonly associated with overweight and obesity remains unclear.
[0003] Tomiyama et al. (Tomiyama AJ, Hunger JM, Nguyen-Cuu J, Wells C (2016) Misclassification of cardiometabolic health when using body mass index categories in NHANES 2005-2012. Int J Obes (Lond) 40:883-6) showed, based on measured blood parameters in a study with 40,420 participants, that 50% of the overweight and 29% of the obese subjects examined were considered cardiometabolically healthy, which was not to be expected based on classification based solely on BMI. In addition, 30% of the normal-weight subjects examined had an unhealthy cardiometabolic state. However, the measured blood parameters only show the current state of the subjects and reflect only a short, past period of cardiometabolic maldevelopment in the subjects. The HbA1c test, for example, provides information about the blood glucose level of the last 4 - 12 weeks and if the HbA1c value is ≥ 6.5%, you are considered diabetic.
[0004] T2D is characterized by a relative insulin deficiency and / or insulin resistance in the body's tissues. During the pathogenesis of T2D, insulin resistance and production progressively increase over many years, until the insulin-producing β-cells of the pancreas become dysfunctional due to years of overuse and reduce insulin production. This point is considered the onset of T2D, but it also takes an average of several more years until the diabetic is diagnosed as such by a doctor (Pearson J, Powers MA (2006) Systematically initiating insulin: the staged diabetes management approach. Diabetes Educ 32 (Suppl):19S-28S).In contrast to type 1 diabetes (T1D), where the cause of diabetes lies in a dysfunctional pancreas, in T2D, in addition to the pancreas, the liver, muscles, and / or abnormalities in fatty tissue can also be causative factors for the pathogenesis of T2D. Determining the HbA1c level alone cannot achieve a more specific diagnosis and the resulting personalized treatment of T2D diabetics. Personalized treatment can only be achieved by dividing the T2D disease into subclasses. To date, there is no reliable diagnostic procedure that allows for the subclassification of diabetics. Therefore, there is a need for differentiated procedures, both in the prognosis and diagnosis of diabetics.
[0005] Against this background, the object of the present invention was to provide a new possibility by means of which a different or additionally differentiated prognosis and / or diagnosis is possible compared to the prior art.
[0006] According to the invention, this object is achieved by a method for prognosticating and / or diagnosing diabetes mellitus type II in a subject, comprising the steps: a) Determine the gene expression level of the gene ADIPOQ in a sample of the subject and b) classifying the subject taking into account the gene expression level of the gene ADIPOQ into one of at least three risk groups, with at least two of the possible risk groups consisting of individuals who have already developed type II diabetes mellitus, wherein for classifying the test person in step b) the gene expression level of the gene HMGA2 is also taken into account, and wherein as possible groups for classification in step b) the groups III) strongly increased relative gene expression level for ADIPOQ , reduced relative gene expression level for HMGA2 and at least one of the marker situations selected from the group consisting of elevated HbA1c blood levels, elevated cholesterol levels, elevated triglycerides, reduced HDL cholesterol, elevated non-HDL cholesterol, elevated LDL cholesterol, elevated CRP, elevated blood sugar, elevated fasting blood sugar, elevated preprandial blood sugar, elevated postprandial blood sugar and age ≥ 45, IV) increased relative gene expression level for ADIPOQ, reduced relative gene expression level for HMGA2and at least one of the marker situations selected from the group consisting of non-elevated HbA1c blood levels, non-elevated cholesterol levels, non-elevated triglycerides, elevated HDL cholesterol, non-elevated non-HDL cholesterol, non-elevated LDL cholesterol, non-elevated CRP, non-elevated blood sugar, non-elevated fasting blood sugar, non-elevated preprandial blood sugar, non-elevated postprandial blood sugar and age ≥ 45, and V) reduced relative gene expression level for ADIPOQ , reduced relative gene expression level for HMGA2and at least one of the marker situations selected from the group consisting of non-elevated HbA1c blood levels, non-elevated cholesterol levels, non-elevated triglycerides, elevated HDL cholesterol, non-elevated non-HDL cholesterol, non-elevated LDL cholesterol, non-elevated CRP, non-elevated blood sugar, non-elevated fasting blood sugar, non-elevated preprandial blood sugar, non-elevated postprandial blood sugar and age ≤ 45, are available.
[0007] One way to diagnose abnormal development with regard to insulin resistance and production early on lies in adipose tissue, which, as a long-term store of fats and glucose, can buffer malnutrition over a very long period. Dysfunctional adipose tissue is probably a key element in the development of T2D, both in overweight / obese individuals and those of normal weight. Various study results show that adipose tissue should increasingly be viewed as an endocrine organ that actively intervenes in or controls physiological processes. The substances secreted by adipose tissue, called adipokines, are associated with, among other things, insulin sensitivity and resistance, reproduction, inflammation, and bone growth, as well as with immunological processes and fatty acid metabolism.An adipokine associated with the predisposition to the development of T2D is adiponectin (Szmitko PE, Teoh H, Stewart DJ, Verma S (2007) Adiponectin and cardiovascular disease: state of the art? Am J Physiol Heart Circ Physiol. 292:H1655-63).
[0008] Human adiponectin is produced by APM1 / ACDC / ACRP30 / GBP28 / ADIPOQ gene(Accession ID: D45371), which is located in chromosomal band 3q27. It contains three exons located in a 17-kb region. Exons one and two are 76 and 222 bp in size, respectively, and are separated by intron one, which is 10.3 kb in size. Exon three comprises approximately 4.28 kb. Translation begins in exon two and ends in exon three, thus leaving exon one and parts of exon three untranslated. The 30-kDa adiponectin protein is produced and secreted primarily by adipocytes. Adiponectin consists of a carboxy-terminal globular domain and a collagenous domain in the amino-terminal end. In blood plasma, adiponectin occurs as a complete, 244-amino-acid protein and as a proteolytic cleavage product fragment, also called globular adiponectin. Isoforms of adiponectin arise from different linkages between the globular and collagen domains.The effect of adiponectin on body cells is mediated by the AdipoR1 and AdipoR2 receptors. AdipoR1 is primarily found in skeletal muscle, and AdipoR2 in liver cells. However, further studies suggest that AdipoR1 and AdipoR2 are also expressed in cardiomyocytes, osteoblasts, and pancreatic β-cells. Adiponectin plays a key role in lipid and glucose metabolism. It alters insulin sensitivity through activation of 5'adenosine monophosphate-activated protein kinase (AMPK) and improves insulin resistance by increasing fatty acid oxidation and suppressing gluconeogenesis in the liver. Adiponectin deficiency is associated with an increased risk of diabetes and diabetic angiopathy, as well as an increased risk of heart attacks and strokes.
[0009] Accordingly, "adiponectin" within the meaning of this text is the adiponectin protein or its gene or the associated mRNA and / or parts of this protein or gene (or its mRNA), preferably with at least one amino acid chain of ≥ 7 amino acids, more preferably ≥ 15 amino acids, and particularly preferably ≥ 20 amino acids, or a nucleic acid chain of ≥ 20 nucleic acids, more preferably ≥ 40 nucleic acids, and particularly preferably ≥ 55 nucleic acids, if applicable per strand. Adiponectin is an important adipokine involved in the control of lipid metabolism and insulin sensitivity and has direct anti-diabetic, anti-atherogenic, and anti-inflammatory effects. It stimulates AMPK phosphorylation and activation in the liver and skeletal muscle, thereby increasing glucose utilization and fatty acid combustion.The human adiponectin gene is located in the chromosomal region 3q27 and consists of three exons spanning a region ≥17 kb in length. It encodes, among other things, a 244-amino-acid complete adiponectin protein with a molecular mass of 30 kDa. The adiponectin protein is characterized by a carboxy-terminal globular domain and a collagenous domain at the amino-terminal end. Adiponectin occurs in plasma as a complete protein (244 amino acids) and as a proteolytic cleavage product fragment, also called globular adiponectin. The adiponectin isoforms arise from different linkages between the globular and collagenous domains. Three main complexes circulate in plasma: a low molecular weight trimer (LMW), a medium molecular weight hexamer (MMW) and a high molecular weight complex (HMW).
[0010] The value for the relative gene expression level within the meaning of the present invention can be determined in any manner known to the person skilled in the art. The gene expression level is preferably determined at the mRNA level or at the protein level. The mRNA level is preferred.
[0011] In the present invention, relative gene expression levels are determined. If reference is made to "gene expression levels" below, this always refers to relative gene expression levels, unless otherwise stated. The relative gene expression levels are preferably determined by determining the gene expression level of the gene to be investigated in relation to the expression level of a housekeeping gene, preferably selected from the group consisting of HPRT, 18S rRNA, GAPDH, GUSB, PBGD, B2M, ABL, RPLP0, where especially HPRT is preferred.
[0012] The term "prognosis" as used in this text means a prediction of an increased probability of the development or occurrence of a clinical condition or disease.
[0013] The term "diagnosis" of a disease, as used in this text, means that a disease that already shows clinical symptoms is identified and / or confirmed.
[0014] Diabetes mellitus type II is also abbreviated as T2D in the following text.
[0015] "Test subject" in the sense of this application means humans and animals, with preference being given to humans.
[0016] "Risk groups" as defined in this text are those groups that can be separated from one another by appropriate distinguishing characteristics and each share a common increased or non-increased risk with regard to the development or presence of a disease, particularly T2D. Furthermore, risk groups may also differ from one another by other physiological differences, which may have therapeutic or prophylactic relevance.
[0017] Surprisingly, it has been found that the method according to the invention makes it possible to further differentiate within the group of individuals (test subjects) suffering from type II diabetes mellitus. This differentiation helps in identifying suitable therapies for the respective individual.
[0018] One such therapeutic approach could be, for example, administering medications that influence adiponectin levels depending on ADIPOQ expression within T2D patients. This could, for example, make it possible to reduce the adiponectin levels in groups with relatively low ADIPOQ Expression to increase adiponectin levels. A suitable drug for this could be metformin.
[0019] Before the method according to the invention, a corresponding sample from a test subject is of course provided.
[0020] As already described, by including the relative gene expression level of the gene ADIPOQ in one of at least three risk groups, the possibility of establishing a further differentiation within the already affected diabetes mellitus type II subjects, which in turn can be used for therapeutic purposes.
[0021] According to the invention, in the present invention, when classifying the subjects in step b), the gene expression level of the gene HMGA2 (and possibly IL-6) is taken into account.
[0022] "HMGA2" in the context of this text refers to the High Mobility Group AT-Hook Protein 2 (HMGA2) or its gene or the associated mRNA and / or parts of this protein or gene (or its mRNA), preferably with at least one amino acid chain of ≥ 7 amino acids, more preferably ≥ 15 amino acids, and particularly preferably ≥ 20 amino acids, or a nucleic acid chain of ≥ 20 nucleic acids, more preferably ≥ 40 nucleic acids, and particularly preferably ≥ 55 nucleic acids, possibly per strand. HMGA2 is a transcription factor that influences the regulation of gene expression and belongs to the group of High Mobility Group A proteins (HMGA proteins). HMGA proteins are chromatin-associated, acid-soluble non-histone proteins that bind to sequence-independent, specific DNA motifs. As architectural transcription factors, they increase or inhibit the binding capacity of other transcription factors through structural changes in chromatin organization. The human HMGA2The gene is located in the chromosomal region 12q14∼15 and consists of five exons spanning a region of ≥160 kb. It encodes a 109 amino acid protein with a molecular mass of 12 kDa. The HMGA2 protein is characterized by three highly conserved DNA-binding domains, the so-called AT hooks, and an acidic, negatively charged C-terminal domain.
[0023] "PPAR-gamma" in the sense of this text is the peroxisome proliferator-activated receptor gamma (PPAR-gamma) or its gene or the associated mRNA and / or parts of this protein or gene (or its mRNA preferably at least one amino acid chain ≥ 7 amino acids, more preferably ≥ 15 amino acids and particularly preferably ≥ 20 amino acids or a nucleic acid chain of ≥ 20 nucleic acids, more preferably ≥ 40 nucleic acids and particularly preferably ≥ 55 nucleic acids, if applicable per strand). PPAR-gamma is a ligand-binding nuclear transcription factor of the PPAR subfamily, which belongs to the group of nuclear hormone receptors. Via heterodimerization with the retinoid X receptor α (RXRα), PPAR-gamma activates the transcription of various genes. The human PPAR-gamma gene is located in the chromosomal band 3p25 and consists of 11 exons. The human PPAR-gamma -Gene encodes 2 isoforms, each representing a 477 and a 505 amino acid long protein.
[0024] "IL-6" as used herein refers to interleukin-6 or its gene or the associated mRNA and / or parts of this protein or gene (or its mRNA), preferably with at least one amino acid chain of ≥ 7 amino acids, more preferably ≥ 15 amino acids, and particularly preferably ≥ 20 amino acids, or a nucleic acid chain of ≥ 20 nucleic acids, more preferably ≥ 40 nucleic acids, and particularly preferably ≥ 55 nucleic acids, if applicable, per strand. Interleukin-6 is a cytokine that plays a role in both inflammatory reactions and the maturation of B lymphocytes. Furthermore, it has been demonstrated to be an endogenous, inflammatory substance, a so-called pyrogen, which can trigger high fever in the presence of autoimmune diseases or infections. The protein is predominantly produced at sites of acute or chronic inflammation, from where it is secreted into the serum and triggers an inflammatory response via the interleukin-6 receptor alpha.Interleukin-6 is involved in various inflammation-associated disease states, including a predisposition to diabetes mellitus and systemic juvenile idiopathic arthritis (Still syndrome). The human . IL-6 The IL-6 gene is located in chromosomal band 7p15.3 and consists of six exons. The IL-6 precursor protein consists of 212 amino acids. After cleavage of a 28-amino acid signal peptide, the mature interleukin-6 is 184 amino acids long (Hirano T, Yasukawa K, Harada H, Taga T, Watanabe Y, Matsuda T, Kashiwamura S, Nakajima K, Koyama K, Iwamatsu A, et al., 1986. Complementary DNA for a novel human interleukin (BSF-2) that induces B lymphocytes to produce immunoglobulin. Nature 324: 73-76).
[0025] It has been found that in the use according to the invention or in the method according to the invention, the combination of the data for ADIPOQwith those of the aforementioned preferred genes lead to particularly reliable group distinctions within the risk groups.
[0026] A method according to the invention is preferred, wherein for the classification into risk groups (in step c)) one or more characteristics of the subject selected from the group consisting of age, BMI, height, weight, gender, waist and hip circumference, body fat percentage, muscle mass, total body water (TBW), blood pressure, smoking status, high blood pressure, intake of antihypertensive medication are further taken into account.
[0027] Thus, by considering additional characteristics of the test subjects, it is possible to create additional certainty in the breakdown within the risk groups. Of course, it is also possible to further refine the breakdown of the risk groups with each additional marker (taking into account another characteristic).
[0028] It is known that blood tests can provide relevant information about the status of T2D.
[0029] By means of the inventive use of ADIPOQ In combination with the additional markers / blood parameters, further group differentiations can be created.
[0030] The following table provides information on typical values for blood parameters with a normal reference range according to the 2012 guideline of the German Diabetes Society and the 2016 ESC / EAS Guidelines for the Management of Dyslipidaemias: Table 1 parameter Standard reference range (mg / dL) In T2D group In non-T2D group Triglyceride < 150 Increased Non-elevated / low Cholesterol, total* <20 years: <170 Increased Non-elevated / low 20-30 years: <200 30-40 years: <220 >40 years: <240 HDL cholesterol Men > 40 Humiliated Increased Women > 48 non-HDL cholesterol + -Very high risk: < 100 Increased Non-elevated / low -high risk: <130 -Low to moderate risk: <145 LDL cholesterol + -Very high risk: <70 Increased Non-elevated / low -high risk: <100 -Low to moderate risk: <115 CRP < 5.0 mg / L Increased Non-elevated / low Fasting glucose (venous plasma) < 100 Increased Non-elevated / low 2h glucose value (venous plasma) <120 Increased Non-elevated / low Preprandial glucose (venous plasma) < 100 Increased Non-elevated / low HbA1c < 5,7% Increased Non-elevated / low
[0031] * The average total cholesterol level for people aged 35 to 65 in Germany is approximately 236 mg / dl, with a standard deviation of ±46 mg / dl.
[0032] + The grading is based on very high, high, and low to moderate risk for cardiovascular disease (see current guidelines from the European Society of Cardiology (ESC) and European Atherosclerosis Society (EAS) for 2016: ESC / EAS Guidelines for the Management of Dyslipidaemias). The risk can be calculated using a SCORE (Systematic Coronary Risk Estimation). The most common scoring systems are the ESC-SCORE and the PROCAM-SCORE (Prospective Cardiovascular Munster Study).
[0033] In this case, when differentiating the status of the respective blood values, a good combination with other markers can lead to additional splitting within the risk groups with appropriate evaluation (see also below).
[0034] A method according to the invention is preferred, wherein the sample was obtained from fatty tissue.
[0035] The "fat tissue sample" refers to the sample that provides the values for at least one gene expression level, preferably for all gene expression levels, used for step b). Of course, blood values cannot be obtained from fat tissue.
[0036] For the preferred method according to the invention, it is preferred that the sample of fatty tissue was obtained by puncturing the subcutaneous abdominal fatty tissue.
[0037] Fan-shaped punctures under suction are particularly effective for obtaining cells and cell clusters, which facilitate molecular genetic analysis. This fan-shaped puncture procedure reduces the risk of fat cells sticking to or clogging the needle tip, and also allows cells to be obtained from various regions of the affected fatty tissue, thus providing a representative cross-section of the distribution of different fatty tissue cell types.
[0038] Particularly preferred is a method according to the invention, wherein the sample mass for the samples from fatty tissue is ≤ 50 mg, preferably ≤ 20 mg and more preferably ≤ 5 mg.
[0039] Surprisingly, it has been shown that reliably differentiated results can be achieved even with very small sample volumes of adipose tissue. It is particularly preferred that the sample be obtained by puncture as a fine-needle aspirate.
[0040] From the inventors' point of view, determining the respective parameters from fatty tissue also has the following advantages in terms of prognosis: Firstly, after determining the parameters from fatty tissue, different risk groups of individuals who already suffer from type II diabetes mellitus can be identified without, for example, the need for further blood sample determinations. Secondly, determining the parameters from fatty tissue according to the invention makes it possible to identify individuals earlier who have a higher probability of developing type II diabetes mellitus than, for example, conventional HbA1c tests. These HbA1c tests only provide information about the blood glucose level over the last four to 12 weeks and are therefore less suitable for a long-term prognosis for the development of type II diabetes mellitus.
[0041] According to the invention, the test subject is preferably a human, since a differentiated prognosis and diagnosis in the case of T2D in humans is of particular importance from both an economic and health policy perspective.
[0042] As already indicated above, it is preferable to determine gene expression levels at the mRNA level. This allows reliable data to be obtained with extremely small sample sizes and using established methods.
[0043] A method according to the invention is preferred, wherein the classification into the risk groups (in step b)) is carried out using the multivariate model of the self-organizing maps according to Kohonen.
[0044] For the methodology of self-organizing maps, please refer to the methodology section below.
[0045] According to the invention, a method is preferred wherein the classification in step b) is carried out into one of at least 5 groups, wherein at least two of the groups consist of individuals who have an increased probability of developing diabetes mellitus type II.
[0046] It has been found that by taking into account the relative gene expression level of ADIPOQ (or, alternatively but not according to the invention, genes that are linearly statistically correlated with them), it is possible to develop differentiation even within the groups of individuals / subjects who have an increased probability of developing type II diabetes mellitus. This finding can also be used to derive therapeutic approaches or appropriate preventive measures.
[0047] Further preferred is a method according to the invention in which the gene expression level is measured at the mRNA level relative to the gene expression level of a housekeeping gene. This method is particularly suitable for obtaining reliable results from small sample quantities.
[0048] A method according to the invention is preferred, wherein as possible groups for classification in step b) the groups with the marker situation I) reduced relative gene expression level for ADIPOQ, greatly increased relative gene expression level for HMGA2and at least one of the marker situations selected from the group consisting of elevated HbA1c blood levels, elevated cholesterol levels, elevated triglycerides, reduced HDL cholesterol, elevated non-HDL cholesterol, elevated LDL cholesterol, elevated CRP, elevated blood sugar, elevated fasting blood sugar, elevated preprandial blood sugar, elevated postprandial blood sugar and age ≥ 45, and II) reduced relative gene expression level for ADIPOQ, reduced relative gene expression level for HMGA2 and at least one of the marker situations selected from the group consisting of elevated HbA1c blood levels, elevated cholesterol levels, elevated triglycerides, low HDL cholesterol, elevated non-HDL cholesterol, elevated LDL cholesterol, elevated CRP, elevated blood glucose, elevated fasting blood glucose, elevated preprandial blood glucose, elevated postprandial blood glucose and age ≥ 45.
[0049] These two preferred classification options reflect the additional potential of using data from ADIPOQ in a special way. With the respective combinations of marker values (always a combination of three values), the preferred classification options according to the invention can be applied – in addition to the classification options into further groups. This makes it possible to obtain reliable additional information regarding the distinguishability of subjects / individuals suffering from T2D.
[0050] In the method according to the invention, there are as possible groups for classification in step b) (at least) the groups with the marker situation III) strongly increased relative gene expression level for ADIPOQ, reduced relative gene expression level for HMGA2and at least one of the marker situations selected from the group consisting of elevated HbA1c blood levels, elevated cholesterol levels, elevated triglycerides, reduced HDL cholesterol, elevated non-HDL cholesterol, elevated LDL cholesterol, elevated CRP, elevated blood sugar, elevated fasting blood sugar, elevated preprandial blood sugar, elevated postprandial blood sugar and age ≥ 45, IV) increased relative gene expression level for ADIPOQ, reduced relative gene expression level for HMGA2and at least one of the marker situations selected from the group consisting of non-elevated HbA1c blood levels, non-elevated cholesterol levels, non-elevated triglycerides, elevated HDL cholesterol, non-elevated non-HDL cholesterol, non-elevated LDL cholesterol, non-elevated CRP, non-elevated blood sugar, non-elevated fasting blood sugar, non-elevated preprandial blood sugar, non-elevated postprandial blood sugar and age ≥ 45, and V) reduced non-elevated relative gene expression level for ADIPOQ , reduced relative gene expression level for HMGA2and at least one of the marker situations selected from the group consisting of non-elevated HbA1c blood levels, non-elevated cholesterol levels, non-elevated triglycerides, elevated HDL cholesterol, non-elevated non-HDL cholesterol, non-elevated LDL cholesterol, non-elevated CRP, non-elevated blood sugar, non-elevated fasting blood sugar, non-elevated preprandial blood sugar, non-elevated postprandial blood sugar and age ≤ 45, are available.
[0051] Groups IV and V include only individuals who have not developed the disease. Group III includes individuals who already show clinical symptoms of T2D.
[0052] Regarding the description "increased" / "decreased", the following applies in case of doubt: For the relative gene expression levels, based on the mean value of the respective patient collective, the following situation applies: + / - 5% unchanged + 5-10 % slightly increased - 5-10 % slightly lowered + 5-50% increased - 5-50 % humiliated + / > 50 % greatly increased - / > 50 % severely reduced
[0053] The grading applies accordingly to blood values, whereby the value given in Table 1 serves as the reference value and only the range that lies beyond the normal range is considered (there are therefore always only "increased" or "decreased" values of the respective gradations in addition to the "unchanged" values).
[0054] Of course, the collective patenting that forms the basis always includes individuals who already suffer from T2D.
[0055] For the division of the groups, please also refer to the examples.
[0056] Also of interest is a kit for a method according to the invention, comprising a) a primer pair that annealed to the cDNA of HMGA2 and b) a primer pair that binds to the cDNA of ADIPOQ binds and prefers c) a primer pair that binds to the cDNA of a housekeeping gene selected from the group consisting of HPRT,18S rRNA, GAPDH, GUSB, PBGD, B2M, ABL, RPLP0 binds.
[0057] With this kit it is possible to determine the gene expression levels of ADIPOQ and HMGA2 The primer pair combination mentioned is thus suitable for a preferred variant of the method according to the invention.
[0058] The invention is explained in more detail below using examples and taking into account the respective methodology. Evaluation procedure methodology
[0059] The aim of the methods applied was to create new classifications (clusters) of diabetics and non-diabetics based on various biomarkers such as HMGA2, ADIPOQ, IL-6 or PPAR-gamma, which will expand the existing classifications both diagnostically and therapeutically in the future.
[0060] For the formal description of the study data, standard methods of descriptive statistics were used. Absolute and relative frequencies were reported for nominal parameters, and the median was also reported for ordinal parameters. Mean and standard deviation were calculated for metric values. Normal distributions were tested using the Kolmogorov-Smirnov test (KS test). Non-parametric correlations between the biomarkers were calculated using Kendall's tau-b. The χ2 test was used for comparisons between categorical variables. Self-organizing maps (SGMs) were used to calculate clusters that were not known a priori. SOM)). SOMs (here Kohonen maps after Teuvo Kohonen, cf. Teuvo Kohonen: Self-Organizing Maps. Springer-Verlag, Berlin 1995, ISBN 3-540-58600-8) are types of artificial neural networks with an unsupervised learning process with the goal of creating a topographical feature map in the form of clusters of the input space (patient data). Patients within a cluster should be maximally homogeneous and between clusters maximally inhomogeneous. SOMs are used for clustering, visualizing complex relationships, prediction (evaluation), modeling, and data exploration. The network used here consists of 1000 neurons; correlations were automatically compensated, and missing values were taken into account. The SOM-WARD clustering method (2-stage hierarchical clustering algorithm) was used to generate the clusters. Color coding was performed using heat maps. The resulting clusters were compared descriptively.To describe the clusters using decision trees or facts and rules, various classification algorithms were used, such as C5.0, CART, and Exhausted Chaid. The performance measures evaluated for the various classifiers were classification accuracy, model compactness (e.g., size of a decision tree), model interpretability, efficiency, and robustness to noise and missing values.
[0061] To further validate the models and calculate the importance of the biomarkers for the various classification models, RBF networks (radial basis function networks) were created as a prediction model. RBF networks provide a suitable approximation of the cluster assignment of the SOMs. The input vectors were normalized (subtracting the mean and dividing by the range (x-min) / (max-min); normalized values range between 0 and 1). The softmax function σ is used as the activation function, a normalized radial basis function. Softmax σ maps a k-dimensional vector z to a k-dimensional vector σ(z).
[0062] The network performance (how "good" is the network) was checked using the following data: Model summary: Results including error, relative error, or percentage of incorrect predictions. Classification results: A classification table was provided for each dependent variable. ROC curves:. ROC curves (Receiver Operating Characteristic curves) indicate the sensitivity and specificity for each possible cutpoint of the input variables. The area under the curve (AUC) is a measure of the quality of the classification. Cumulative profit charts.
[0063] The following methods were used: 1. Self-organizing neural networks (-7 Kohonen maps) 2. Classification algorithms Entropy-based learning methods (C5.0) Exhausted Chaid and CART 3. Radial basis functions (special type of neural networks) 4. Descriptive and inductive statistics 1) Self-organizing neural networks
[0064] As self-organizing maps ( SOM)) refers to types of artificial neural networks with an unsupervised learning method aimed at achieving a topological representation of the input space (here, patient data). The most well-known SOMs are the topology-preserving Kohonen maps developed by Teuvo Kohonen. The learning algorithm independently generates classifiers, which it uses to divide the input patterns into (previously unknown) clusters. The goal is to achieve maximal homogeneity among patients within a cluster and maximal inhomogeneity among clusters.
[0065] Core idea (Topographic feature map): "Neighboring" input vectors (here patient data) should belong to neighboring neurons in the map so that the density and distribution of the neurons correspond to the probability model of the training set.
[0066] Advantages: Neighborhood relationships in the "confusing" input space can be read directly in the output layer.
[0067] Applications: SOMs are used for clustering, visualizing complex relationships, prediction (evaluation), modeling, and data exploration. The focus of the application for the questions at hand is clustering, visualization, and prediction.
[0068] (Tools: e.g. Self-Organizing Maps in R (R is a free programming language for statistical calculations and graphics. R is part of the GNU project, see also https: / / cran.r-project.org / web / packages / som / som.pdf and Figure 1 ) 1.1) Formal description of the Kohonen network model (algorithm)
[0069] A SOM consists of two layers of neurons (Input layer and output layer), see also Figure 2Each neuron in the input layer is connected to each neuron in the output layer (the neurons in the input layer are fully interconnected with those in the output layer). Each neuron in the input layer corresponds to a parameter of the data set. The number of input neurons is the dimension of the input layer. The output neurons are related to each other by a neighborhood function. The strength of the connection is expressed by a number (=weight) w[i][j] (w[i][j] is the weight indicating the strength of the connection between the i-th input neuron and the j-th output neuron). The vector w[j] represents all weights w[i][j] (i=1...n, n is the number of parameters acquired for each patient) to the j-th output neuron. Input vectors and weight vectors are normalized (length=1). Initialize the weight vectors.As a starting point, arbitrary values generated by a random generator are specified for the weights. Each patient p defines an input vector through its values. Input p = (xp [1], xp [2],..., X p [n]) with the components Input p [i] = xp [i]. These input vectors are first normalized to 1. For each patient p and each neuron j in the output layer, the Euclidean distance s p j = ‖ w j − input p ‖ = ∑ i w j i − input p i 2 between the weight vector w[j] and the input vector Input p. The output neuron that has the smallest distance to the input vector Input p is called Winner-takes-all neuron. The weight vector of the winning neuron is most similar to the input vector, ie, it has the "maximum excitation" under the input vector Input p . If two or more output neurons have the same minimum distance, a neuron is randomly selected. Winner neuronand its neighbors receive the "contribution," meaning they are allowed to represent the input. This defines a function f(input p ) that assigns each vector input p of the input space (pattern space, feature space) a location α in a representational layer (map) f: input p →f(input p ) .= arg min(llw[j] - input p ll). The minimum is calculated over all weight vectors w[j]. The function arg returns the index of the winning neuron. Determine the winning neuron for each patient (input pattern) according to the above rule. Neighborhood function and Weight adaptation: In the next step, the weights of the winning neuron and its surrounding neurons are adapted. The degree of proximity to the winning neuron plays a significant role. Let's assume the winning neuron has the index α. For an input[i] and an output neuron j, Dw[j][i] denotes the weight change within a learning rule. This is calculated as follows: Dw j i = η t ∗ input i − w i j * NbdWt α j where η(t) is the learning rate(η(t) with 0 < η(t) < 1) and NbdWt(α, j) the Neighborhood function (neighborhood weighting function). A neighborhood function calculates the degree of proximity of an output neuron to the winner neuron, which lies between 0 and 1. Convergence proofs against a statistically describable equilibrium state exist, for example, for η(t) := η*t -a< with 1 <a≤1. Nearer weight vectors proportional to the neighborhood function to each other The weight vectors w[j] of all neurons j according to the neighborhood function NbdWt(a, j) of neuron α are updated. The new weight w new [i][j] is calculated as follows: w new [j][i] := w[j][i] + Dw[j][i] The neighborhood function can be one of the following functions: NbdWt α j = 1 1 + d α j s 2 NbdWt α j = e − d α j s 2 with d( α , j):= ( α- j); s is a scalar. The width s = s(t) of the neighborhood function and the learning step size η(t) are reduced over time: the winning neuron and its neighbors are "activated" according to the function NbdWt(a, j). Normalize weight vectors to length 1, present the next data point. For various methods for color coding (heat maps) of maps, see the literature (e.g., http: / / citeseerx.ist.psu.edu / viewdoc / download?doi=10.1.1.100.500&rep=rep1&type=pdf, https: / / arxiv.org / pdf / 1306.3860.pdf, or https: / / www.visualcinnamon.com / 2013 / 07 / self-organizing-maps-creating-hexagonal.html ). 2) Classification algorithms
[0070] To evaluate the SOM models (description of the classes by facts and rules or decision trees), three different classification algorithms are used: (1) Entropy-based learning methods (C5.0), (2) Exhausted Chaid and (3) CART. 2.1) Definition.
[0071] Classification procedures are methods and criteria for classifying objects (here patients) into classes (here types and subtypes of healthy prediabetics or diabetics).
[0072] From a training set of examples with known class membership, a classification algorithm is used to Classifier in the form of decision trees or equivalent in the form of facts and "if-then" rules The classification differs from clustering (see also SOMs) in that the classification the classes are known a priori, while in Clustering the classes must first be searched for.
[0073] Decision trees are used for decision-making through a tree-like structure consisting of a root node (start node), nodes, edges and leaves (end nodes) ( Figure 2 ).
[0074] Formally, a tree is a finite graph with the properties: 1) There is exactly one node where no edge ends (the "Root").2) Exactly one edge ends in each node different from the root, the edges are directed 3) Each node can be reached from the root via exactly one path.
[0075] A Decision tree is a tree: 1) Each node tests an attribute 2) Each branch corresponds to an attribute value 3) Each leaf (node without outgoing edges) assigns a class.
[0076] The Figure 3 shows the formal components of a classification algorithm.
[0077] Problem: The classifier is initially optimized for the training data. It may deliver poorer results on the entire data set (underfitting or overfitting: hypothesis class is too weak or too complex). See Figure 4 .
[0078] A possible overfitting can be avoided by Pruning or boosting methodsbe reduced, whereby the expert selects the number of iteration steps required to improve group formation.
[0079] In general, the set of existing examples is divided into two subsets (train-and-test). Training amount: to learn the classifier (construction of the model) Test quantity: to evaluate the classifier
[0080] If this is not applicable because the set of objects with known class membership is small, the so-called m-fold Cross-validation used.
[0081] As Quality measures for classifiers the following criteria are taken into account: Classification accuracy, model compactness (e.g., decision tree size), model interpretability, efficiency, and robustness to noise and missing values. 2.2) Construction of decision trees
[0082] (See also Quinlan, J. Ross (1986): Induction of decision trees. In: Machine Learning 1 (1), pp. 81-106. Quinlan, J. Ross (1993): C4.5. Programs for machine learning. In: J. Ross Quinlan. San Mateo, Calif.: Morgan Kaufmann (The Morgan Kaufmann series in machine learning). Quinlan, J. Ross (1996): Improved use of continuous attributes in C4. 5. Journal of artificial intelligence research 4, pp. 77-90.) Basic algorithm
[0083] Ribbon: 1. Select the "best" decision attribute A for the next node. 2. For each value of A, create a new child node. 3. Assign the training data to the child nodes. 4. If the training data is correctly classified, STOP. Otherwise, iterate over child nodes (→ 1.). Designations:
[0084] Training dataset T, number of training data |T|, classes C i : The dataset of all training data in the class C i . | Ci list the number of elements in class C i . The following applies: Σ | C i | = |T| (i=1,..k). Attribute A = { a 1 , a 2 , ... , a m}. The attribute A divides the data set T into m subsets T 1 , T 2 , ... ,T m . |T i | is the number of the subset T i . Given training dataset T and attributes A; Output: Information gain(T, A) of attributes A for the training dataset T 3) Algorithm (calculating the information gain using entropy using ID3 as an example)
[0085] The empirical entropy for a lot T of training objects with the classes C i (i=1,...,k) is defined as entropie T = ∑ i = 1 k p i ⋅ log p i mit p i : = C i / T The attribute A has the partitioning T 1 , T 2 , ..., T m The empirically conditioned Entropy G(T,A) for a lot T and an attribute A is defined as G T A = ∑ i = 1 m T i T ⋅ entropie T i The Information gain Gain(T, A) by attribute A with respect to T is defined as Gain T A = entropie T − G T A = entropie T − ∑ i = 1 m T i T ⋅ entropie T i Input: Training dataset T with the classes Ci (i=1, ...,k), attributes A and threshold ε ; Output: Decision tree E; Algorithm:
[0086] 1. Create a node K; 2. If all sample data in T have an identical class C j or the number of data is less than threshold ε, then the only node with class C j back to node K as a leaf; 3. If A = 0̸, then the only node with the most frequent class in T is assigned to node K as a leaf; 4. Calculate the information gain Gain of A in T and determining the best attribute A g with maximum information gain based on Gain T A = entropie T − G T A = entropie T − ∑ i = 1 m T i T ⋅ entropie T i 5. Denote the node K by A g 6. Calculate for all attribute values A gj of A g the subset T gi of all examples from the training dataset with A gj 7. If T gi = 0̸, then a leaf with the most frequent class is appended to node K otherwise 8. Recursion of the branch of A g . Remark:
[0087] For a random event y, which occurs with probability P(y), the following applies: Information content: h y ≡ − log 2 P y
[0088] Entropy is the mean amount of information content of the random variable y H y ≡ ∑ y P h y = − ∑ y P y log 2 P y Classification algorithms used ID3
[0089] Discrete attributes, no missing attributes Information Gain as a quality measure. C4.5 / C5.0
[0090] Extension of ID3. Information Gain Ratio as a quality measure. Missing attributes. Numerical and real-valued attributes. Pruning of the decision tree.
[0091] The Information GainRatio is defined as GainRatio T A : = Gain T ,A SplitInformation T ,A SplitInformation T A = − ∑ i = 1 m T i T ⋅ log 2 T i T CART (Classification And Regression Trees)
[0092] Compare also Hastie, T., Tibshirani, R., Friedman, J.H. (2001). The elements of statistical learning: Data mining, inference, and prediction. New York: Springer Verlag.
[0093] Procedure analogous to ID3 or C5.0. The information level is determined by the Gini index The Gini index is minimized (instead of maximizing the Gini gain).
[0094] Gini gain Gini T A = ∑ i = 1 k T i T ⋅ Gini T i Mit Gini T = 1 − ∑ i = 1 k p i 2 P i is the relative frequency of class C i in T CHAID ( Chi -square A utomatic Interaction D etectors)
[0095] See also Sonquist, JA and Morgan, JN (1964): The Detection of Interaction Effects. Survey Research Center, Institute for Social Research, University of Michigan, Ann Arbor.
[0096] CHAID is another algorithm for constructing decision trees. The differences from C5.0 or CART are that the CHAID algorithm uses the chi-square independence test to select attributes, and the CHAID algorithm stops the tree from growing before it becomes too large. Thus, the tree is not allowed to grow indefinitely and then pruned again. 4) Radial basis function networks (RBF networks) See also Zell, A.: Simulation of neural networks. Oldenbourg 1994
[0097] RBF networks are used to further validate the models and calculate the importance of biomarkers for the various classification models. RBF networks create predictive models. They are particularly suitable for approximating functions.
[0098] The RBF network consists of an input layer with n neurons, a hidden layer with k neurons, and an output layer with m neurons. An n-dimensional pattern is thus mapped into an m-dimensional output space.
[0099] The input layer is a pure forwarding layer. Each neuron distributes its value to all hidden neurons. In the hidden layer, the distance between the input and the center c is calculated in each neuron using a Euclidean norm. Radial basis functions are used as the network input and activation functions (see Figure 5 ).
[0100] The activation function of each hidden neuron is a so-called radial function, i.e. a monotonically decreasing function f : IR 0 + → 0,1 mit f 0 = 1 und lim x → ∞ f x = 0 m
[0101] The input vectors are normalized (Subtraction of the mean and division by the range (x-Min) / (Max-Min); Normalized values are in the range between 0 and 1). The activation function is the Softmax function σ as a normalized radial basis function. Softmax σ maps a k-dimensional vector z to a k-dimensional vector σ(z). σ z j = e z j ∑ k = 1 K e z k , j = 1 , … , k
[0102] The number of neurons in the hidden layer is determined by the "Bayesian Information Criterion "(Compare Schwarz, Gideon E. (1978), "Estimating the dimension of a model", Annals of Statistics, 6 (2): 461-464, MR 468014, doi:10.1214 / aos / 1176344136) (BIC). The best number of hidden units is the one that results in the smallest BIC based on the training data.
[0103] For the output layer, we used the identity function as the activation function. The output units are thus simply weighted sums of the hidden units. The output of the network is therefore a linear combination of the radial basis functions of the inputs and the weights. Network performance
[0104] Network performance checks how "good" the network is. A range of results are provided for this purpose. Model summary. Results including error, relative error, or percentage of false predictions, and training time. Classification results. A classification table is provided for each categorical dependent variable. ROC curves. The ROC curves (receiver operating characteristic curves) indicate the sensitivity and specificity for each possible cutpoint of the input variable. The area under the curve (AUC) is a measure of classification quality. Cumulative gain plots. Examples
[0105] The following examples are based on two different patient groups. Patient group 1 was used as the basis for the following examples 1 and 3. It is characterized by a number of 7 Individuals in whom 4 a T2D was present and patent collective 2, which forms the basis for examples 2, 4 and 5, consisted of 61 Subjects of which 26T2D was present.
[0106] A sample with the following ΔC T served as calibrator for the relative gene expression values: Patient collective 1 & 2: HMGA2: 6,513 ADIPOQ: 5,484
[0107] The ΔCT values are the CT value of the target gene minus the CT value of the endogenous control (housekeeping gene), where the CT value is the value at which the signal for the respective cDNA exceeds the threshold for the first time during amplification.
[0108] The following mean values were obtained for the relative gene expression in patient group 2: HMGA2: 1,663117 ADIPOQ: 0,95365 Example 1 Detection of expression of ADIPOQ and HMGA2 in adipose tissue samples obtained by fine needle aspiration of subcutaneous abdominal adipose tissue Materials and methods Fine needle aspiration
[0109] Fine-needle aspirates were obtained by puncturing subcutaneous abdominal hWAT using a 20 ml syringe and a disposable injection cannula (diameter 0.90 x 40 mm). After disinfecting the puncture site, the cannula was inserted into the subcutaneous fat. Negative pressure was created with the syringe, and the cannula was moved back and forth in a fan-like motion within the tissue to aspirate fat cells. Immediately after puncture, the samples were placed in 1 ml of QIAzol Lysis Reagent (QIAGEN, Hilden, Germany), and the cannula was rinsed several times with QIAzol. The samples were then frozen at -80 °C. RNA isolation
[0110] Total RNA was isolated using the RNeasy Lipid Tissue Mini Kit (QIAGEN, Hilden, Germany) in a QIAcube (QIAGEN, Hilden, Germany) according to the manufacturer's instructions. Fine needle aspirates (5 mg) in 1 ml of QIAzol Lysis Reagent were homogenized in a Tissue Lyser II (QIAGEN, Hilden, Germany), and the homogenate was incubated at room temperature for 5 min. This was followed by the addition of 200 µl of chloroform, which was mixed with the sample by vigorous manual shaking for 15 s. The sample was incubated again for 2 min at room temperature and centrifuged at 12,000 xg for 15 min at 4 °C. The upper aqueous phase was then transferred to a new 2 ml cup, and total RNA was isolated using a Qiagen RNeasy Mini Spin column (QIAGEN, Hilden, Germany) in a QIAcube according to the manufacturer's instructions. cDNA synthesis
[0111] For cDNA synthesis, ≤ 250 ng of RNA was transcribed into cDNA using 200 U M-MLV reverse transcriptase, RNase Out (Thermo Fisher Scientific, Darmstadt, Germany) and 150 ng of random primers (Thermo Fisher Scientific, Darmstadt, Germany) according to the manufacturer's instructions. The RNA was denatured at 65°C for 5 min and then stored on ice for at least 1 min. After addition of the enzyme, the mix was incubated for 10 min at 25°C to anneal the random primers to the RNA. Subsequent reverse transcription was carried out at 37°C for 50 min, followed by inactivation of the reverse transcriptase at 70°C for 15 min. Pre-amplification of cDNA
[0112] 5 µl cDNA was prepared using RealTime ready cDNA Preamp Mastermix (Roche, Mannheim, Germany) using HMGA2 and HPRT(hypoxanthine phosphoribosyltransferase 1)-specific primers according to the manufacturer's instructions. The gene-specific primers bind appropriately to the respective cDNA, so that the generated amplicons contain the binding sites of the primers of the gene-specific assays used in quantitative real-time PCR. The cDNA preamplification was performed according to the following temperature profile: 95°C for 1 min, followed by 14 cycles at 95°C for 15 sec and at 60°C for 4 min. Quantitative real-time PCR (qRT-PCR)
[0113] Relative quantification of gene expression was performed using real-time PCR on the Applied Biosystems 7300 Real-Time PCR System. Commercially available gene expression assays (Life Technologies, Carlsbad, CA, USA) were used to quantify mRNA levels of HMGA2 (Assay ID Hs00171569_m1) and ADIPOQ(Assay ID Hs00605917_m1) was used. As an endogenous control, as described by Klemke et al. (Klemke M, Meyer A, Hashemi Nezhad M, Beige G, Bartnitzke S, Bullerdiek J (2010) Loss of let-7 binding sites resulting from truncations of the 3' untranslated region of HMGA2 mRNA in uterine leiomyomas. Cancer Genet Cytogenet 196:119-123), HPRTused. All measured samples were determined in triplicate. Quantification of gene expression was performed in 96-well plates using the pre-amplified cDNA to be analyzed, the respective gene-specific assay, and the FastStart Universal Probe Master (Rox) (Roche, Mannheim, Germany). The temperature profile of the real-time PCR followed the manufacturer's instructions: The template was denatured at 95°C for 10 minutes. This was followed by amplification in 50 cycles, beginning with denaturation for 15 seconds at 95°C and a combination of annealing / elongation for 60 seconds at 60°C. The obtained data were analyzed using the comparative delta-Ct method (ΔΔCT method). [(Livak KJ, Schmittgen TD (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2 -ΔΔC(T) Method. Methods 25: 402-408)]. Result
[0114] The gene expression of ADIPOQ and HMGA2was measured by qRT-PCR in seven samples from human patients ( Fig. 6 ). The results clearly show that ADIPOQ- and HMGA2 mRNAs can be reliably quantified from fine-needle aspirates containing even very small amounts of adipose tissue cells. Furthermore, the amount of RNA isolated from the fine-needle aspirates is sufficient to determine the expression of additional genes after pre-amplification.
[0115] Fig.6 represents the relative expression of HMGA2 and ADIPOQ in seven samples. It shows that a reliable quantification of the HMGA2- and ADIPOQ expression even from very small amounts of RNA (concentration ≤ 25 ng / µl) obtained by fine needle aspiration. Fig.6 that the expression of genes varies between individual samples and thus inter-individual differences can be detected. Example 2 The expression of ADIPOQ in adipose tissue samples from type 2 diabetics and non-diabetics Materials and methods Tissue samples
[0116] Human subcutaneous abdominal adipose tissues were collected during surgery and stored in liquid nitrogen after surgery. The samples were then frozen at -80°C. All human adipose tissue samples used met the requirements of the Declaration of Helsinki. Written informed consent for the use of the tissue samples was provided by the patients (n = 61). RNA isolation
[0117] Total RNA was isolated using the RNeasy Lipid Tissue Mini Kit (QIAGEN, Hilden, Germany) in a QIAcube (QIAGEN, Hilden, Germany) according to the manufacturer's instructions. Adipose tissue samples (50–100 mg) in 1 ml of QIAzol Lysis Reagent were homogenized in a Tissue Lyser II (QIAGEN, Hilden, Germany), and the homogenate was incubated at room temperature for 5 min. This was followed by the addition of 200 µl of chloroform, which was mixed with the sample by vigorous manual shaking for 15 s. The sample was incubated again for 2 min at room temperature and centrifuged at 12,000 xg for 15 min at 4 °C. The upper aqueous phase was then transferred to a new 2 ml cup, and total RNA was isolated using a Qiagen RNeasy Mini Spin column (QIAGEN, Hilden, Germany) in a QIAcube according to the manufacturer's instructions. The RNA concentration was determined using a photometer and the samples were then stored at -80°C. cDNA synthesis
[0118] For cDNA synthesis, 250 ng of RNA was transcribed into cDNA using 200 U M-MLV reverse transcriptase, RNase Out (Thermo Fisher Scientific, Darmstadt, Germany) and 150 ng of random primers (Thermo Fisher Scientific, Darmstadt, Germany) according to the manufacturer's instructions. The RNA was denatured at 65°C for 5 min and then stored on ice for at least 1 min. After addition of the enzyme, the mix was incubated for 10 min at 25°C to anneal the random primers to the RNA. Subsequent reverse transcription was carried out at 37°C for 50 min, followed by inactivation of the reverse transcriptase at 70°C for 15 min. Quantitative real-time PCR (qRT-PCR)
[0119] The relative quantitative real-time PCR was performed as described in Example 1. Result
[0120] The differences in adiponectin levels in the blood plasma of type 2 diabetics and non-diabetics described in the literature may also be reflected at the level of gene expression of ADIPOQ in the fatty tissue of type 2 diabetics and non-diabetics ( Fig. 7 ). The Fig. 7 shows the significant difference (p < 0.05) in the relative gene expression of ADIPOQ in the adipose tissue of 35 non-diabetics and 26 type 2 diabetics. Example 3 Different gene expression of ADIPOQ in adipose tissue samples from type 2 diabetics Materials and methods Sample preparation
[0121] Sample collection was performed by fine-needle aspiration as described in Example 1. The subsequent steps of sample preparation, i.e., RNA isolation, cDNA synthesis, cDNA pre-amplification, and quantitative real-time PCR, were also performed as described in Example 1. Result
[0122] The Fig. 8 shows the expression of ADIPOQin adipose tissue samples from four patients. Patients 2 and 4 have an HbA1c value of 6.7% and 7.0%, respectively, and are therefore diabetics; patients 1 and 3 are non-diabetics (HbA1c values of 5.6% and 4.4%, respectively). The measured ADIPOQ gene expression values are 1.734 for patient 1; 0.609 for patient 2; 1.389 for patient 3; and 1.397 for patient 4. The measured gene expressions of ADIPOQ in samples from human patients ( Fig. 8 ) that surprisingly, type 2 diabetics like patient 4 also have a high ADIPOQ expression. This result was not expected based on the current literature. The different ADIPOQExpression in patients 2 and 4 would also mean a different form of therapy with regard to, for example, the administration of oral antidiabetics. The protein adiponectin, encoded by the ADIPOQ gene, suppresses hepatic gluconeogenesis via the receptor AdipoR1 and stimulates glucose uptake in skeletal muscle via the receptor AdipoR2. Therefore, at high ADIPOQ expression values, the administration of medications from the group of biguanides or biguanide derivatives (e.g. metformin) is not recommended, since this group of drugs inhibits hepatic gluconeogenesis and improves peripheral insulin action, thus the mode of action of biguanides or biguanide derivatives is similar to that of adiponectin. ADIPOQ -expression values, however, the administration of biguanide or biguanide derivatives could (possibly) be recommended in order to compensate for the relative lack of adiponectin. Example 4 The gene expression of ADIPOQ in adipose tissue samples from normal and overweight individuals Materials and methods Sample preparation
[0123] The adipose tissue samples were obtained during surgery as described in Example 2. The subsequent steps of sample preparation, i.e., RNA isolation, cDNA synthesis, and quantitative real-time PCR, were also performed as described in Example 2. Result
[0124] The Fig. 9 shows the expression of ADIPOQ in adipose tissue samples from 20 patients with normal weight (BMI < 25) and 41 patients with overweight (BMI ≥ 25). The gene expression of ADIPOQ in the normal weight group was 1.207 and was highly significantly higher (p < 0.001) than the ADIPOQ -expression in the overweight group with 0.834. The analysis of the correlation of ADIPOQ -expression and BMI in 61 patients showed a moderate (-0.507) and highly significant (p < 0.001) inverse association ( Fig. 10). This result is generally consistent with the inverse relationship between adiponectin levels in blood plasma and body weight described in the literature. However, comparing the gene expression of ADIPOQ of individuals with overweight or obesity or normal weight with the ADIPOQ -Expression of the entire patient collective (see Fig. 11 ), it is surprisingly evident that there are exceptions to the inverse relationship between adiponectin levels in blood plasma and body weight. Fig. 6 shows that the male subject 1 with a BMI of 37.5 and the female subject 2 (BMI = 29.8) have a ADIPOQ -expression of 1.519 and 1.364 respectively, which are more likely to ADIPOQ- Expression of the normal weight group with 1.207. In addition, the normal weight male subjects 3 (BMI=24.8) and 4 (BMI=24.1) show a ADIPOQ -expression of 0.612 and 0.659 respectively, which are in the range of ADIPOQexpression of the overweight is 0.834. These data suggest that the inverse correlation of ADIPOQ -expression and BMI corresponds to another factor. The patent DE10 2015 208 083 B3 discloses a connection between HMGA2- Expression level and BMI. An analysis of the ADIPOQ- and HMGA2- Expression level and BMI using a surface plot, as in Fig. 12 shown, shows that the correlation of ADIPOQ -Expression level and BMI depending on HMGA2- Expression level. This correlation was not to be expected based on the scientific data and the state of the art. Patent DE10 2015 208 083 B3 also discloses a correlation between high HMGA2 -expression level, low PPAR-gamma expression levels and a high BMI. Since the peroxisome proliferator-activated receptor gamma (PPAR-gamma) regulates adipokines, such as adiponectin, the gene expression of PPAR-gamma and ADIPOQ It was shown, as in Fig. 13 shown a strong (0.838) and highly significant (p < 0.001) correlation between the gene expression of PPAR-gamma and ADIPOQ. Example 5 Data analysis using self-organizing maps shows five different clusters within the patient collective Materials and methods Sample preparation
[0125] The adipose tissue samples were obtained during surgery as described in Example 2. The subsequent steps of sample preparation, i.e., RNA isolation, cDNA synthesis, and quantitative real-time PCR, were also performed as described in Example 2. Statistical analysis See above, Methodology Result
[0126] Various studies have shown in the past that low adiponectin levels in the blood are associated with an increased incidence of type 2 diabetes (T2D). Adiponectin is considered a "protective factor" against T2D and coronary heart disease, and high levels of the hormone are seen as an indicator of a lower risk of disease. However, as examples 3 and 4 show, high ADIPOQ-expression levels can also be found in patients with T2D and obesity. These findings were not expected based on the scientific data and therefore have potential implications for the treatment of these patients. In order to enable personalized treatment of, for example, patients with T2D, it is useful to identify the specific subtype of T2D that the patient suffers from. A specific T2D subtype could, for example, be associated with insulin resistance, while another subtype could be associated with problems in insulin secretion. To identify these T2D subtypes, the analysis of the relationship between the biomarkers can be used. ADIPOQ, HMGA2 and help the elderly using Kohonen's Self-Organizing Maps (SOM).
[0127] The Fig. 14 shows after analysis of the biomarkers ADIPOQ, HMGA2and age (as a further marker), SOM surprisingly resulted in a classification of the patient collective into five groups and not, as one would expect, for example, from an HbA1c test, into two groups, namely type 2 diabetics and non-diabetics. The corresponding data can be found in Table 2 and in the Figures 15-17 The division of the patient collective into five groups is also surprising, since the patent specification DE10 2015 208 083 B3 disclosed a division of the patients into four groups, namely two overweight and two normal weight groups.
[0128] The data analysis using self-organizing maps divides the test group taking into account the parameters HMGA2- and ADIPOQ expression in five groups. The non-diabetics are divided into two clusters (C1 and C4). The mean HMGA2 -expression in both clusters is compared to the mean HMGA2-expression of the entire collective was reduced. The mean ADIPOQ Expression in cluster C1 is increased compared to the overall collective, but decreased in cluster C4. Diabetics are divided into clusters C2, C3, and C5. The mean HMGA2 -expression is strongly increased in cluster C2, the ADIPOQ -expression is reduced. In cluster C3, the expression of both genes is reduced. The third diabetic cluster C5 shows a reduced HMGA2 -expression, which ADIPOQ -expression is strongly increased in this cluster. ADIPOQ Segmentation
[0129] Table 2 Cluster Designation mean HMGA2 Deviation [%] HMGA2 mean ADIPOQ Deviation [%] ADIPOQ Mean age C1 Non-diabetics ST1 1.21 -27.2 humiliated 1.12 17.4 increased 64,7 C2 T2D ST1 3.48 109.2 greatly increased 0.74 -22.4 humiliated 76,4 C3 T2D ST2 1.19 -28.4 humiliated 0.67 -29.7 humiliated 66,3 C4 Non-diabetics ST2 1.21 -27.2 humiliated 0.79 -17.2 humiliated 37,3 C5 T2D ST3 1.01 -39.3 humiliated 1.79 87.7 greatly increased 75,3 Without being bound to any theory, there are the following biological explanations for the clusters:
[0130] C2 (T2D subtype 1): dysfunctional adipose tissue due to many preadipocytes (high HMGA2 -value and relatively low ADIPOQ-value), more "insulin-resistant" T2D subtype. Differentiation of preadipocytes towards mature insulin-sensitive adipocytes would be desirable here; for example, drugs from the group of glitazones and metformin could be used. However, there is evidence that insulin and insulin-like growth factors promote the differentiation of preadipocytes towards mature insulin-sensitive adipocytes (Ayoubi et al., 1999; Klemm et al., 2001). Therefore, it would also be conceivable to use insulin and insulin-promoting drugs at high HMGA2 -values.
[0131] C5 (T2D subtype 3): functional adipose tissue due to many mature insulin-sensitive adipocytes (high ADIPOQ-values), a generally "insulin-sensitive T2D subtype" is expected. This T2D subtype presumably has a problem with insulin production / secretion in the β-cells of the pancreas, so treatment here could aim to improve / increase insulin production (e.g., sulfonylurea and glinides).
[0132] C3 (T2D subtype 2): dysfunctional adipose tissue probably due to impaired (or at least too low) adipokine production (medium HMGA2 -value (rather inconspicuous) and very low ADIPOQ value), probably more "insulin-resistant" T2D subtype. Since preadipocytes are probably not too many here, the administration of metformin could help to ADIPOQ -values (see Kubota et al. 2006 and Adamia et al. 2007).
[0133] C1 (non-diabetic subtype 1): reflects functional adipose tissue in terms of fat cell composition (ratio of preadipocytes to adipocytes presumably in the "normal / healthy" range) as well as function or adipokine production ("moderately high" ADIPOQ -values).
[0134] C4 (non-diabetic subtype 2): reflects more functional adipose tissue in terms of fat cell composition (preadipocyte to adipocyte ratio presumably in the "normal / healthy" range). The lower values compared to cluster C1 ADIPOQ -values could possibly be related to age (but the reason could also be the smaller number of subjects in the C4 cluster).
Claims
1. Method for prognosing and / or diagnosing a type II diabetes mellitus disease of a subject, comprising the steps of: a) determining the gene expression level of the ADIPOQ gene in a sample of the subject and b) classifying the subject into one of at least three risk groups, taking the gene expression level of the ADIPOQ gene into account, wherein the gene expression level of the HMGA2 gene is further taken into account for the classification of the subject in step b), and wherein the groups III) greatly increased relative gene expression level for ADIPOQ, lowered relative gene expression level for HMGA2 and also at least one of the marker situations selected from the group consisting of increased HbAlc blood values, increased cholesterol values, increased triglycerides, lowered HDL cholesterol, increased non-HDL cholesterol, increased LDL cholesterol, increased CRP, increased blood sugar, increased fasting blood sugar, increased preprandial blood sugar, increased postprandial blood sugar and age ≥ 45, IV) increased relative gene expression level for ADIPOQ, lowered relative gene expression level for HMGA2 and also at least one of the marker situations selected from the group consisting of nonincreased HbAlc blood values, nonincreased cholesterol values, nonincreased triglycerides, increased HDL cholesterol, nonincreased non-HDL cholesterol, nonincreased LDL cholesterol, nonincreased CRP, nonincreased blood sugar, nonincreased fasting blood sugar, nonincreased preprandial blood sugar, nonincreased postprandial blood sugar and age ≥ 45, and V) lowered relative gene expression level for ADIPOQ, lowered relative gene expression level for HMGA2 and also at least one of the marker situations selected from the group consisting of nonincreased HbAlc blood values, nonincreased cholesterol values, nonincreased triglycerides, increased HDL cholesterol, nonincreased non-HDL cholesterol, nonincreased LDL cholesterol, nonincreased CRP, nonincreased blood sugar, nonincreased fasting blood sugar, nonincreased preprandial blood sugar, nonincreased postprandial blood sugar and age ≤ 45 are available as possible groups for a classification in step b).
2. Method according to Claim 1, wherein one or more features of the subject selected from the group consisting of BMI, height, weight, sex, abdominal and hip circumference, body fat percentage, muscle mass, total body water (TBW), blood pressure, smoker status, high blood pressure and ingestion of antihypertensive drugs are further taken into account in the classification in step b).
3. Method according to any of Claims 1 to 2, wherein the sample was obtained from adipose tissue.
4. Method according to Claim 3, wherein the sample was obtained by puncture of subcutaneous abdominal adipose tissue.
5. Method according to any of Claims 1 to 4, wherein the subject is a person.
6. Method according to any of Claims 1 to 5, wherein the determination of the gene expression level is done at the mRNA level.
7. Method according to any of Claims 1 to 6, wherein the classification in step b) is done using the multivariate model of self-organizing maps by Kohonen.
8. Method according to any of Claims 1 to 7, wherein the classification in step b) is done into one of at least 5 groups, at least two of the groups consisting of individuals who have an increased probability of forming type II diabetes mellitus.
9. Method according to any of Claims 1 to 8, wherein the gene expression level is measured at the mRNA level relative to the gene expression level of a housekeeping gene.
10. Method according to any of Claims 1 to 9, wherein the groups I) lowered relative gene expression level for ADIPOQ, greatly increased relative gene expression level for HMGA2 and also at least one of the marker situations selected from the group consisting of increased HbAlc blood values, increased cholesterol values, increased triglycerides, lowered HDL cholesterol, increased non-HDL cholesterol, increased LDL cholesterol, increased CRP, increased blood sugar, increased fasting blood sugar, increased preprandial blood sugar, increased postprandial blood sugar and age ≥ 45, II) lowered relative gene expression level for ADIPOQ, lowered relative gene expression level for HMGA2 and also at least one of the marker situations selected from the group consisting of increased HbAlc blood values, increased cholesterol values, increased triglycerides, lowered HDL cholesterol, increased non-HDL cholesterol, increased LDL cholesterol, increased CRP, increased blood sugar, increased fasting blood sugar, increased preprandial blood sugar, increased postprandial blood sugar and age ≥ 45, III) greatly increased relative gene expression level for ADIPOQ, lowered relative gene expression level for HMGA2 and also at least one of the marker situations selected from the group consisting of increased HbAlc blood values, increased cholesterol values, increased triglycerides, lowered HDL cholesterol, increased non-HDL cholesterol, increased LDL cholesterol, increased CRP, increased blood sugar, increased fasting blood sugar, increased preprandial blood sugar, increased postprandial blood sugar and age ≥ 45, IV) increased relative gene expression level for ADIPOQ, lowered relative gene expression level for HMGA2 and also at least one of the marker situations selected from the group consisting of nonincreased HbAlc blood values, nonincreased cholesterol values, nonincreased triglycerides, increased HDL cholesterol, nonincreased non-HDL cholesterol, nonincreased LDL cholesterol, nonincreased CRP, nonincreased blood sugar, nonincreased fasting blood sugar, nonincreased preprandial blood sugar, nonincreased postprandial blood sugar and age ≥ 45, and V) lowered relative gene expression level for ADIPOQ, lowered relative gene expression level for HMGA2 and also at least one of the marker situations selected from the group consisting of nonincreased HbAlc blood values, nonincreased cholesterol values, nonincreased triglycerides, increased HDL cholesterol, nonincreased non-HDL cholesterol, nonincreased LDL cholesterol, nonincreased CRP, nonincreased blood sugar, nonincreased fasting blood sugar, nonincreased preprandial blood sugar, nonincreased postprandial blood sugar and age ≤ 45 are available as possible groups for a classification in step b).
