DECREASED INSULIN SECRETION

PRIORITY CLAIM

[0001] This application claims priority to U.S. Provisional Patent Application Serial No. 62/309,040, filed March 16, 2016, the entire contents of which is incorporated herein by reference.

TECHNICAL FIELD

[0002] The present disclosure relates to methods of treating impaired glucose tolerance and/or decreased insulin secretion in a subject.

BACKGROUND

[0003] Adiponectin is an adipocyte-secreted hormone that acts on multiple tissues to enhance insulin action, lipid utilization, and endothelial function. Adiponectin concentrations vary by more than 25-fold among healthy individuals. While low adiponectin strongly predicts diabetes, cardiovascular disease, and death, little is known about effects of very high adiponectin concentrations.

SUMMARY

[0004] The present disclosure provides methods of treating impaired glucose tolerance and/or decreased insulin secretion in a subject.

[0005] In one embodiment, the present disclosure provides a method of treating or preventing diabetes in a subject, the method comprising determining a baseline level of adiponectin in a sample of the subject; and initiating a therapeutic regimen for the subject if the baseline level of adiponectin is at least about 49 pg/mL.

[0006] In another embodiment, the present disclosure provides a method of preventing insulin resistance in a subject having a normal HOMA-IR level, the method comprising identifying the subject as having a normal HOMA-IR level; determining an adiponectin level in a sample of the subject; and initiating a therapeutic regimen for the subject if the adiponectin level is less than 10 pg/mL.

[0007] In some embodiments, the present disclosure provides a method of identifying a subject as being at elevated risk for developing diabetes, the method comprising determining a level of adiponectin of at least 49 mg/mL in a sample associated with the subject and identifying the subject as being at elevated risk for developing diabetes based at least in part on the level of adiponectin in the sample.

[0008] In some embodiments, the present disclosure provides a method of identifying a subject as being at elevated risk for developing coronary artery disease (CAD), the method comprising obtaining a level of adiponectin in a biological sample associated with the subject; obtaining an age of the subject; and identifying the subject as being at elevated risk for developing CAD if a mathematical transformation (e.g., a logarithmic transformation) of the adiponectin level falls below a predetermined threshold.

[0009] In some embodiments, the present disclosure provides a method of treating or preventing metabolic disease in a subject, the method comprising determining an elevated level of adiponectin associated with the subject and thereafter initiating a therapeutic regimen in the subject to affect a decrease in the adiponectin level associated with the subject.

BRIEF DESCRIPTION OF THE DRAWINGS

[0010] FIG. 1 shows a ROC contrast plot with comparative AUC values for several predictive diabetes models consistent with the present disclosure in comparison to several predictive diabetes models that do not consider adiponectin levels.

[0011] FIG. 2 illustrates that adiponectin levels and LP-IR scores are inversely correlated, suggesting that low adiponectin levels may be associated with a dyslipoproteinemia component of metabolic disease.

[0012] FIG. 3 is a dendogram showing divisive clustering of various biomarkers using principal components, with four specific correlations (ln[eAG] and ln[A1 c]; Infinsulin] and ln[HOMA-IR]; In [leptin-BMI] and ln[leptin]; and ln[QA] and ln[FFA]) showing very strong correlations (r>0.90).

[0013] FIG. 4 shows a heat map for absolute value of Pearson's Correlation between biomarkers and cluster component scores (N ~1 ,500); lighter shading indicates areas of high correlation values (approaching 1 ) while darker shading indicates areas of relatively low correlation (approaching 0).

[0014] FIG. 5 illustrates that testing the 415 subjects of Example 1 with HOMA-IR (Glucose (in mg/dl_) ^* lnsulin/405) alone would have resulted in a 33% misdiagnosis rate of subjects with signs of insulin resistance. In contrast, diagnosis using adiponectin accurately identified this subset of subjects with signs of insulin resistance but normal HOMA-IR values.

[0015] FIG. 6 shows results of an OGTT test in subjects with discordant adiponectin (e.g., low adiponectin) compared to HOMA-IR. The plot on the top shows glucose response curves, while the plot on the bottom shows insulin response curves, for subjects with discordant (low) adiponectin levels and normal HOMA-IR (dashed line) compared subjects with normal adiponectin levels and normal HOMA-IR (solid line).

[0016] FIG. 7 shows characteristics of subjects having very high adiponectin levels (at least 50 pg/mL) compared to subjects having adiponectin levels of less than 50 pg/mL. Generally, subjects with very high adiponectin levels (n=18) were older (e.g., about 20 years older, on average), had lower leptin levels, optimal apolipoprotein B, LDL-C, LDL-P and HDL-C levels, low insulin despite significant IGT, and much higher NT-proBNP levels compared to subjects with adiponectin levels of less than 50 pg/mL (n=408).

[0017] FIG. 8 shows that subjects with very high adiponectin levels (at least 50 pg/mL) are more glucose intolerant (top panel) but have lower fasting insulin levels and impaired insulin response to an oral glucose load (bottom panel) compared to subjects having adiponectin levels less than 50 pg/mL.

[0018] FIG. 9 shows ROC curves for subjects with very high adiponectin levels (at least 50 pg/mL) using (1 ) a base model including age, gender and BMI; (2) a HOMA-IR model; (3) the Japan IR model (log[lnsulin ^* Glucose/Adiponectin]); and (4) an OGTT Index model. Each of the HOMA-IR, Japan IR, and OGTT Index models perform significantly better than the base model for this subset of subjects with very high adiponectin levels.

[0019] FIG. 10 demonstrates that ROC curves derived from the Japan index (log[lnsulin ^* Glucose/Adiponectin]) and the HOMA-IR model (Glucose (in mg/dl_) ^* lnsulin/405) were not significantly different in predicting IGT in the subjects studied in Example 1 .

[0020] FIG. 1 1 shows probability of coronary artery disease (CAD) based on a bivariate effect of age and log[Adiponectin] in 18 test subjects (light circles) and 16 control subjects (dark circles).

DETAILED DESCRIPTION

[0021] Low adiponectin levels have been associated with type 2 diabetes mellitus (T2DM), cardiovascular disease (CVD), atherogenic lipid profiles, and essential hypertension. Studies have demonstrated that adiponectin has diagnostic and prognostic values as a biomarker of obesity, insulin sensitivity, and CVD.

[0022] Adiponectin is produced almost exclusively by adipose tissue. Circulating monomers assemble to form trimers, hexamers, high-molecular weight (HMW) oligomers, and other multimeric forms. Adiponectin exerts insulin-sensitizing, anti-inflammatory and anti-apoptotic actions on a variety of cell types. Metabolically unfavorable conditions tend to downregulate adiponectin secretion, which generally causes a reduction in its blood levels.

[0023] One study demonstrated that low adiponectin levels in Korean men and women were associated with a 1.7-fold and a 1 .8-fold increased risk, respectively, for developing T2DM within six years. Prediabetic men and women in the same study who had the lowest adiponectin levels were most likely to develop diabetes, compared to prediabetic men and women with higher adiponectin levels.

[0024] In another prospective study, an adiponectin level less than 3.9 pg/mL (men) or less than 6.01 pg/mL (women) was the strongest predictor of diabetes among subjects having stage 2 prediabetes based on impaired fasting glucose (IFG) levels. An adiponectin level of less than 6.01 g/mL also predicted diabetes among women having stage 1 or stage 2 prediabetes in the same study.

[0025] Another study demonstrated that high plasma adiponectin levels were associated with the lowest risk for T2DM and CVD in subjects without T2DM or CVD. More specifically, each doubling of adiponectin levels reduced risk of incident T2DM by 45% (p<0.001 ) and the risk of major adverse CV events after T2DM development by 66% (HR 0.34, p = 0.05).

[0026] Yet another study demonstrated that treatment of high-risk IGT subjects with pioglitazone resulted in a 3-fold increase in adiponectin, reversion to normal glucose tolerance, and protection from T2D development. Baseline adiponectin levels did not predict a subject's response to pioglitazone in this study, but subjects who reverted to normal glucose tolerance upon pioglitazone treatment had the highest final plasma adiponectin levels.

[0027] Another study examined the link between leptin:adiponectin ratios and sedentary time in men at high risk for developing T2DM. The study revealed, inter alia, that sedentary time was positively correlated with the leptin:adiponectin ratio in these subjects, independent of glycemic status and adiposity.

[0028] These studies clearly teach that high adiponectin levels were not associated with diabetic conditions. Instead, these studies specifically state that high levels of adiponectin did not predict incident diabetes in the study subjects.

[0029] In contrast to the studies described above, the present technology is based on the surprising discovery that subjects with high adiponectin levels exhibited enhanced insulin sensitivity, but a greater glucose excursion and decreased insulin response during OGTT, compared to subjects with normal adiponectin levels.

[0030] In some embodiments, the present disclosure provides a method of treating or preventing diabetes in a subject, the method comprising determining a baseline level of adiponectin in a sample of the subject; and initiating a therapeutic regimen for the subject if the baseline level of adiponectin is at least about 49 pg/mL. In some embodiments, the therapeutic regimen comprises a change in diet. In some embodiments, the therapeutic regimen comprises an increase in exercise. In some embodiments, the therapeutic regimen comprises administering a composition comprising an omega-3 fatty acid to the subject. In some embodiments, the therapeutic regimen comprises administering conjugated linoleic acid to the subject. In some embodiments, the therapeutic regimen comprises administering an adiponectin agonist to the subject. In some embodiments, the method further comprises determining a baseline HOMA-IR level associated with the subject before the step of initiating the therapeutic regimen. In some embodiments, the baseline HOMA-IR level indicates the subject is normoglycemic. In some embodiments, the method further comprises determining a baseline leptin level in the sample of the subject before the step of initiating the therapeutic regimen. In some embodiments, the baseline leptin level is below a normal leptin level associated with a normoglycemic subject of the same gender as the subject. In some embodiments, the method further comprises determining a baseline apolipoprotein B level in the sample of the subject before the step of initiating the therapeutic regimen. In some embodiments, the baseline apolipoprotein B level is within a normal apolipoprotein B range associated with normoglycemic subjects of the same gender as the subject. In some embodiments, the method further comprises determining a baseline LDL-C level in the sample of the subject before the step of initiating the therapeutic regimen. In some embodiments, the baseline LDL-C level is within a normal LDL-C range associated with normoglycemic subjects of the same gender as the subject. In some embodiments, the method further comprises determining a baseline LDL- P level in the sample of the subject before the step of initiating the therapeutic regimen. In some embodiments, the baseline LDL-P level is within a normal LDL-P range associated with normoglycemic subjects of the same gender as the subject. In some embodiments, the method further comprises determining a baseline HDL-C level in the sample of the subject before the step of initiating the therapeutic regimen. In some embodiments, the baseline HDL-C level is within a normal HDL-C range associated with normoglycemic subjects of the same gender as the subject. In some embodiments, the method further comprises determining a baseline HbA1 c level in the sample of the subject before the step of initiating the therapeutic regimen. In some embodiments, the baseline HbA1 c level is within a normal HbA1 c range associated with normoglycemic subjects of the same gender as the subject. In some embodiments, the method further comprises determining a baseline fasting glucose level in the sample of the subject before the step of initiating the therapeutic regimen. In some embodiments, the baseline fasting glucose level is within a normal fasting glucose range associated with normoglycemic subjects of the same gender as the subject. In some embodiments, the method further comprises determining a baseline NT-proBNP level in the sample of the subject before the step of initiating the therapeutic regimen. In some embodiments, the baseline NT-proBNP level is greater than a normal NT-proBNP range associated with normoglycemic subjects of the same gender as the subject. In some embodiments, the method further comprises determining a second level of adiponectin in a sample of the subject obtained after the step of initiating the therapeutic regimen. In some embodiments, the method further comprises discontinuing the therapeutic regimen if the second level of adiponectin is less than 49 pg/mL. In some embodiments, the sample of the subject is a blood sample.

[0031] In some embodiments, the present disclosure provides a method of preventing insulin resistance in a subject having a normal HOMA-IR level, the method comprising identifying the subject as having a normal HOMA-IR level; determining an adiponectin level in a sample of the subject; and initiating a therapeutic regimen for the subject if the adiponectin level is less than 10 pg/mL. In some embodiments, the therapeutic regimen comprises a change in diet. In some embodiments, the therapeutic regimen comprises an increase in exercise. In some embodiments, the therapeutic regimen comprises administering a composition comprising an omega-3 fatty acid to the subject. In some embodiments, the therapeutic regimen comprises administering conjugated linoleic acid to the subject. In some embodiments, the therapeutic regimen comprises administering an adiponectin agonist to the subject. In some embodiments, the method further comprises determining an OGTT result associated with the subject before the step of initiating the therapeutic regimen. In some embodiments, the OGTT result indicates glucose intolerance and/or impaired insulin response. In some embodiments, the OGTT comprises administering a 75g oral glucose load to the subject. In some embodiments, the sample of the subject is a blood sample. [0032] In some embodiments, the present disclosure provides a method of identifying a subject as being at elevated risk for developing diabetes, the method comprising determining a level of adiponectin of at least 49 mg/ml_ in a sample associated with the subject; and identifying the subject as being at elevated risk for developing diabetes based at least in part on the level of adiponectin in the sample. In some embodiments, the step of identifying comprises identifying the subject as being at elevated risk for developing diabetes based on the level of adiponectin, a fasting insulin level, and a fasting glucose level in the sample, and an age, gender and BMI of the subject. In some embodiments, the step of identifying comprises determining a ratio of a multiple of the fasting insulin level with the fasting glucose level to the level of adiponectin in the sample. In some embodiments, the step of identifying comprises identifying the subject as being at elevated risk for developing diabetes based at least in part on a ratio of leptin to the level of adiponectin in the sample. In some embodiments, the step of identifying comprises identifying the subject as being at elevated risk for developing diabetes based on the level of adiponectin in the sample but not based on an age, gender or BMI of the subject. In some embodiments, the step of identifying comprises determining a logarithmic transformation of at least the level of adiponectin in the sample.

[0033] In some embodiments, the present disclosure provides a method of identifying a subject as being at elevated risk for developing coronary artery disease (CAD), the method comprising obtaining a level of adiponectin in a biological sample associated with the subject; obtaining an age of the subject; and identifying the subject as being at elevated risk for developing CAD if a mathematical transformation of the adiponectin level falls below a predetermined threshold. In some embodiments, the mathematical transformation comprises a logarithmic transformation. In some embodiments, the mathematical transformation does not include adjusting the adiponectin level based on an age of the subject. In some embodiments, the mathematical transformation does not include adjusting the adiponectin level based on a gender of the subject. In some embodiments, the mathematical transformation does not include adjusting the adiponectin level based on an HDL-C level of the subject. In some embodiments, the mathematical transformation does not include adjusting the adiponectin level based on level of small dense low density lipoprotein (sdLDL) of the subject. In some embodiments, the mathematical transformation does not include adjusting the adiponectin level based on an LDL level of the subject. In some embodiments, the mathematical transformation does not include adjusting the adiponectin level based on a cystatin C level of the subject. In some embodiments, the cystatin C level of the subject is a log[cystatin C] level. In some embodiments, the mathematical transformation does not include adjusting the adiponectin level based on a cholestanol level of the subject. In some embodiments, the mathematical transformation does not include adjusting the adiponectin level based on an HDL ApoE level of the subject. In some embodiments, the mathematical transformation does not include adjusting the adiponectin level based on an insulin level of the subject.

[0034] In some embodiments, the present disclosure provides a method of treating or preventing metabolic disease in a subject, the method comprising determining an elevated level of adiponectin associated with the subject; and thereafter initiating a therapeutic regimen in the subject to affect a decrease in the adiponectin level associated with the subject. In some embodiments, the elevated level of adiponectin associated with the subject is at least 50 pg/mL. In some embodiments, the therapeutic regimen comprises administering to the subject one or more n-3 polyunsaturated fatty acids (n-3 PUFAs). In some embodiments, the therapeutic regimen comprises administering to the subject conjugated linoleic acids (CLA).

EXAMPLES

Example 1 .

[0035] 426 patients at risk for diabetes underwent a 75g 2hr OGTT. Blood collected at baseline and during the OGTT was sent to a national reference laboratory (Health Diagnostic Laboratory, Inc., Richmond, VA) for analysis.

[0036] Adiponectin groups were defined as low (<10 pg/rnl; n=198), normal (10-49 pg/ml; n=210), or high (>49 pg/rnl; n=18). The low cut-point has been reported previously, and the high cut-point was data-derived for best model fit. [0037] As expected from other studies, the low adiponectin group had higher BMI, fasting insulin, (peptide, CRP, LDL-P, and triglycerides, lower HDL-C and HDL-P, and increased glucose and insulin excursions during OGTT (all p<0.02). The latter were analyzed by repeated measures ANOVA.

[0038] Remarkably, the high adiponectin group exhibited enhanced insulin sensitivity but a greater glucose excursion and decreased insulin response during OGTT as compared to the normal group (p<0.05). Ferritin did not differ.

[0039] As shown in FIG. 1 and Table 1 , a sensitivity and specificity ROC plot was used to compare improvements to a base model including age, gender and BMI (ref. no. 10). Adding HOMA-IR to the base model (ref. no. 20) significantly increased the AUC by 0.0467 (p = 0.0408). Adding fasting insulin, fasting glucose and adiponectin to the base model increased the AUC by 0.0309 (ref. no. 30), but that increase was not significant.

[0040] Adding the HDL OGTT parameters (a combination of adiponectin, free fatty acids, a-hydroxybutyrate, ferritin, LGPC and C-peptide) to the base model was most predictive (ref. no. 40), increasing the AUC by 0.0923 (p = 0.0012). Removing adiponectin from the HDL OGTT parameters was insignificant (AUC 0.804 vs. 0.795). Use of adiponectin alone or leptin:adiponectin ratio alone were each less predictive (AUC values of 0.702 and 0.716, respectively) than HOMA-IR.

Table 1 . ROC Contrast Estimation and Testing

[0041] Interestingly, however, using adiponectin alone predicted IGT just as well as the base model including age, gender and BMI (AUC values of 0.702 vs. 0.7015). These data demonstrate that adiponectin alone may be used in place of demographic parameters such as age, gender and BMI in predicting IGT in some subjects.

[0042] FIG. 2 shows an inverse correlation between adiponectin and lipoprotein insulin resistance score (LP IR Score). The plot reveals a Pearson Correlation of r = -0.48 (p < 0.0001 ) for n=379 subjects. These data suggest that low adiponectin levels may be associated with the dyslipoproteinemia component of metabolic disease.

[0043] These data enabled the discovery that four pairs of parameters have very strong correlations (r > 0.90), which are shown circled in FIG. 3. Average glucose ("eAG") and hemoglobin A1 c ("A1 c") were very strongly correlated (r = 0.99), as were leptin/BMI and leptin (r = 0.98), insulin and HOMA-IR (r = 0.93), and oleic acid and free fatty acids (r = 0.90). These data suggest that the parameters in these pairs should not be considered independent if they are used in IGT predictive models. For example, an IGT predictive model that reveals abnormal levels of insulin and HOMA-IR in a subject should be considered a single abnormal indicator of insulin resistance rather than two separate (e.g. , independent) indicators of insulin resistance.

[0044] Referring now to FIG. 4, adiponectin and LP-IR Score may be viewed as forming their own dimension in a cluster analysis. This implies from a latent variable point of view that both adiponectin and LR-IR are manifested from the same underlying physiology.

[0045] To arrive at the heat map shown in FIG. 4, each disjoint cluster includes a cluster component score based on a linear combination of the weighted, standardized biomarker values contained within that cluster. Principal component (pc) analysis that maximized the amount of explained variability yielded linear combinations.; disjoint clusters are correlated because the principal components were rotated (i.e., not orthogonal). The number of clusters was determined by considering several factors, including eigenvalues, minimum R-squared value between a biomarker and its cluster component score, total variability explained in the data, and subject matter knowledge.

[0046] The diagonal of high correlation values (lighter shades) seen in FIG. 4 indicates those variables that are homogeneous. Darker shades indicate greater independence between clusters and biomarkers. A summary of the clusters appears Table 2 below.

Table 2. Cluster Summary

[0047] A summary of the biomarkers appears in Table 3 below.

Table 3. Biomarker Summary

[0048] As shown in FIG. 5, one-third (n=140) of the entire cohort had discordant low adiponectin (<10 mg/L) despite normal HOMA-IR. This suggests that estimating risk of insulin resistance in this cohort using HOMA-IR alone would have misdiagnosed nearly 34% of these subjects.

[0049] FIG. 6 shows glucose response and insulin response curves for the subset of subjects with normal HOMA-IR scores. In the top panel, the subset of subjects with discordant low adiponectin (<10 mg/L) show signs of glucose intolerance compared to the subset of subjects with normal adiponectin levels (at least 10 mg/L) after a 75g oral glucose load. Similarly, the bottom panel shows that the discordant subjects showed signs of impaired insulin response compared to the subset of subjects with normal adiponectin levels after the 75g oral glucose load.

[0050] As shown in FIG. 7, eighteen subjects in the cohort had very high adiponectin levels (at least 50 pg/mL; median 99.0 pg/mL). Those subjects were found to be older

(about 20 years older on average) and lower leptin levels, but optimal lipids and lipoproteins. Interestingly, traditional glycemic markers such as HbA1 c and fasting glucose were not significantly different compared to the subjects with adiponectin levels below 50 pg/mL.

[0051] FIG. 8 shows glucose response and insulin response curves for the subjects with very high adiponectin levels (at least 50 pg/mL; median 99.0 pg/mL). These subjects tended to be more glucose intolerant (top panel) but had lower fasting insulin levels and impaired insulin response (bottom panel) to an oral glucose load compared to the subjects with normal or low adiponectin levels.

[0052] ROC curves for the subjects with adiponectin levels below 50 pg/mL (i.e., excluding the subjects with very high adiponectin) are shown in FIG. 9. In this subset of the subjects, models including HOMA-IR (ref. no. 20); a combination of Insulin, Glucose and Adiponectin (ref. no. 30); or the combination of adiponectin, free fatty acids, a- hydroxybutyrate, ferritin, LGPC and C-peptide (i.e., "HDL OGTT-lndex"; ref. no. 40) to the base model (age, gender and BMI; ref. no. 10) each significantly improved IGT prediction. Specific increases for each improved model are shown in Table 4 below.

Table 4. ROC Contrast Estimation and Testing for Subjects with Adiponectin <50 pg/mL

[0053] FIG. 10 shows a comparison of the Japan IR Index model (Log[lnsulin ^* Glucose/Adiponectin]; ref. no. 30) with the HOMA-IR model (ref. no. 20) and a base model including age, gender and BMI (ref. no. 10). In this comparison, the Japan IR Index model accounted for 2% more variability than the HOMA-IR model in predicting IGT. However, that improvement was not statistically significant, as shown in Table 5 below. Table 5. Comparison of HOMA-IR and Japan IR Index Models in Subjects with Adiponectin <50 pg/mL

Example 2.

[0054] A logistic regression analysis of 56 biomarkers in 50 subjects (34 test; 16 control), of which 60% were male and the mean age was 57 years was performed to develop a model for predicting the probability of having coronary artery disease (CAD).

[0055] As shown in Table 6, three biomarkers— HDL-cholesterol (HDL-C), homocysteine (Log[Homocysteine]) and adiponectin (Log[Adiponectin])— were significantly different between subjects having CAD and those without CAD when modeled individually and adjusted for age and gender.

Table 6. Biomarker Odds Ratios for CAD

[0056] In addition to the above three markers, HDL-C, sdLDL/LDL, Log[Cystatin-C], cholestenol, HDL ApoE, and insulin were significantly different in subjects with CAD compared to those without CAD, but the significance of each of those biomarkers disappeared when adjusted for age and gender.

[0057] FIG. 1 1 graphically shows the probability of CAD in these subjects based on the bivariate effect of age (p = 0.019) and adiponectin (p = 0.032) averaged over gender (p = 0.056).

[0058] These data demonstrate that very high adiponectin levels are unexpectedly associated with impaired glucose tolerance and decreased glucose-stimulated insulin secretion in some individuals.

[0059] From the foregoing, it will be appreciated that specific embodiments of the invention have been described herein for purposes of illustration, but that various modifications may be made without deviating from the scope of the invention. Accordingly, the invention is not limited except as by the appended claims.