Abstract
Aims:
Links between resistin, insulin resistance (IR), and resistin-stimulated cytokine signaling remain unknown in Mexican-Americans. A Mexican-American cohort was examined to determine (1) relationships between circulating resistin and IR, (2) resistin’s associations with cytokines and demographic and anthropometric variables, and (3) similar measurements with other adipokines.
Methods:
For cross sectional analyses, 953 adults (367 males and 586 females) in the Cameron County Hispanic Cohort (CCHC) were stratified into three groups: normal glucose tolerance, prediabetes, and diabetes mellitus. Differences in resistin and other adipokine levels were examined using linear regression via unadjusted model (Model 1), model adjusted for cytokines (Model 2), and model further adjusted for demographic and anthropometric variables (Model 3).
Results:
HOMA-IR increased with worsening glucose tolerance (p < 0.0001). In all models, resistin significantly decreased as glucose tolerance deteriorated. Model 3 resistin was positively associated with IL-1β (p = 0.0252) and IL-8 (p < 0.0001), inversely associated with TNF-α (p = 0.0352), but nonsignificantly associated with IL-6 (p = 0.8671). Model 3 leptin was significantly lower in diabetes mellitus compared to other groups (p < 0.005) and positively associated with female sex (p < 0.0001), age (p = 0.024), and BMI (p < 0.0001), without significant cytokine associations. Adiponectin displayed no significant associations with glucose tolerance, but was significantly associated with sex, BMI, and lipids (Model 3).
Conclusions:
Resistin unexpectedly decreased as IR increased while supporting evidence of a resistin-stimulated cytokine pathway in this Mexican-American cohort. Leptin fell with elevated IR after adjusting for cytokines, demographic and anthropometric variables. Adiponectin nonsignificantly decreased as IR increased while showing significant associations with sex, BMI, and lipids.
Keywords: Resistin, Insulin resistance, Leptin, Adiponectin, Mexican-American
1. Introduction
Diabetes mellitus and its complications are widespread. Among Hispanics living in the United States, there is an approximately 50% increased lifetime risk of developing diabetes mellitus, compared to non-Hispanic Whites [1]. Type 2 diabetes mellitus (T2DM) develops via a transition from normal glucose tolerance to prediabetes to T2DM. This transition accompanies an increase in insulin resistance (IR), which is a key factor in the development of diabetic microvascular complications and atherosclerotic cardiovascular disease [2,3].
Resistin is a key adipokine implicated in the development of increased IR, as shown in many murine studies and an increasing number of human studies. Some studies show positive associations between increased resistin level and increased IR [4–7] but other studies reveal no such differences [8–10]. Notably resistin levels vary across different ethnic groups, but Mexican-Americans are underrepresented in these studies [11].
Of particular interest is a proposed mechanism of resistin-mediated stimulation of four proinflammatory cytokines (IL-1β, IL-6, TNF-α, and IL-8) through a collaborative pathway, recently observed in Italian cohorts of healthy and insulin-resistant subjects [12]. Poorly controlled T2DM manifests as a chronic low grade inflammatory state [13], so the presence of this relationship is not surprising. What is less known is the role that resistin plays on the surrounding milieu of cytokines during the progression from normal glucose tolerance (NGT) to prediabetes to diabetes mellitus - along with concurrent IR deterioration - amongst Mexican-Americans.
Other than resistin, the adipokines leptin and adiponectin are implicated in the progression of IR. In non-obese insulin-resistant humans, leptin administration ameliorates IR and lowers blood glucose [14]. However in obese insulin-resistant humans, leptin levels are elevated; this represents a concurrent state of leptin resistance [15]. Data are conflicting regarding the association of circulating leptin levels with IR in Mexican-Americans [13,16] and other ethnic groups [17–19]. Adiponectin is a mediator of insulin sensitivity; lower levels are reported in insulin-resistant states in some Mexican-Americans [20,21], though this may not be true of all ethnicities [22].
The Cameron County Hispanic Cohort, which is a Mexican-American population living along the south Texas-Mexico border [23], provides a unique opportunity to study relationships between these adipokines and IR. Our prior studies show that this population exhibits unusually high insulin resistance compared to other ethnic groups [24], which is highly associated with the percentage of Amerindian ancestry markers [25]. T2DM in this population is associated with multiple elevated inflammatory markers [13].
The primary goal of this study is to define relationships between circulating resistin levels and IR in a homogeneous Mexican-American cohort. We concurrently delineate the effects of cytokines (IL-1β, IL-6, TNF-α, and IL-8) and demographic and anthropometric variables on resistin levels. The same assessments are repeated to comparatively explore associative patterns between other adipokines (leptin and adiponectin) and IR.
2. Methods
The study was approved by the Committee for the Protection of Human Subjects at The University of Texas Health Science Center at Houston. All work was carried out in accordance with the Declaration of Helsinki.
2.1. Study Participants
This cross-sectional analysis uses baseline data from the Cameron County Hispanic Cohort (CCHC), an ongoing randomly ascertained community cohort study recruited from households in a homogeneous Mexican-American population living along the southern Texas border. Participants enroll in the cohort by signing a written informed consent. This is followed by the completion of questionnaires (addressing anthropometric and socio-demographic characteristics), clinical examinations, and collection of laboratory specimens as previously described [23]. At the time of this analysis, the cohort consisted of 4,258 subjects recruited between 2004 and 2018 in Brownsville TX. Of these, 3,929 subjects were adults (18 years or older).
Only adult subjects with measurements of plasma resistin, leptin, and adiponectin were included in the analysis; these measurements were available in 1,834 of the 3,929 adult subjects. Participants were categorized into normal glucose tolerance, pre-diabetes, or diabetes mellitus based on American Diabetes Association criteria using fasting blood glucose (FBG) and or hemoglobin A1c (HbA1c) levels. Further information includes age, gender, blood pressure, and body mass index (BMI).
Subjects were excluded for history of ischemic heart disease, stroke, malignancy, hyperthyroidism, or hypothyroidism from the analysis. As nicotine influences adipokine levels [26,27], current smokers (at time of visit) were excluded. Subjects were also excluded for use of medications which alter adipokine levels: statins [26,28], NSAIDs [28], thiazolidinediones [26,28,29], glucagon-like peptide-1 receptor agonists (GLP-1RAs) [30,31], dipeptidyl peptidase-4 inhibitors (DPP-4 inhibitors) [32], sodium-glucose cotransporter-2 (SGLT-2) inhibitors [33], angiotensin-converting-enzyme inhibitors (ACE-Is) and angiotensin II receptor blockers (ARBs) [26,28]. Among the 1,834 adult subjects with available adipokine data, application of these criteria led to the exclusion of 881 subjects. A final number of 953 study subjects qualified for the study.
2.2. Definitions
Participants were divided into one of three groups, representative of glucose tolerance: normal glucose tolerance (NGT), prediabetes, and diabetes mellitus. The 2010 American Diabetes Association Standards of Care were used to define diabetes mellitus was as HbA1c ≥ 6.5% or fasting blood glucose (FBG) ≥ 126 mg/dL [34]. Oral glucose tolerance tests are not conducted on the cohort participants. Participants who were on hypoglycemic medications, or were diagnosed with diabetes by a health care provider, were also defined as diabetes mellitus (i.e., as described above, subjects on some hypoglycemic medicines were ultimately excluded). Prediabetes was defined as HbA1c ranging from 5.7% to 6.4% or FBG ranging from 100 to 125 mg/dL, without either HbA1c or FBG meeting the definition of diabetes mellitus [34]. NGT was defined as both an HbA1c < 5.7% and FBG < 100 mg/dL.
2.3. Laboratory measurements
Fasting blood specimens were promptly aliquoted and stored at −80°C until analyses. Measurements were completed using methods described previously [13]. Fasting blood glucose (colorimetric assay), glycated hemoglobin (HbA1c, High Performance Liquid Chromatography), and lipid panel (colorimetric assay) were assayed in a local CLIA approved hospital laboratory. Plasma insulin levels were measured using ELISA assays (Mercodia, Uppsala, Sweden). HOMA-IR was calculated as fasting plasma insulin (FPI) × FBG (mmol/L)/22.5 and HOMA B (HOMA %Beta) was calculated as (20xFPI) / (FPG-3.5). For the purpose of these calculations, FPI was measured in mIU/L and FPG was measured in mmol/L [35].
Three key adipokines (resistin, leptin, and total adiponectin) and four relevant inflammatory markers (interleukin-1 beta (IL-1β), interleukin-6 (IL-6), interleukin-8 (IL-8), and tumor necrosis factor alpha (TNF-α)) were measured in house using the multiplex ELISA (Milliplex Map, Millipore, CA) bead technique, using two panels as previously described [13]. Results are read using the Luminex 200 System (Luminex Corp, Austin, TX).
2.4. Statistical analysis
We conducted univariable comparisons of demographic, anthropometric, and metabolic characteristics for the study participants grouped by glucose tolerance status (NGT vs. prediabetes vs. DM) using Chi-squared-tests for categorical variables, Analysis of Variance (ANOVA) for continuous variables, or their nonparametric equivalents (e.g. Kruskal-Wallis test) as appropriate. As the distribution of adipokines resistin, leptin, and total adiponectin were skewed, the data were transformed using the natural logarithm (ln) to produce approximately normal distributions for the main statistical analysis. We examined the relationship between those natural log-transformed adipokines and the glucose tolerance status using a univariable linear regression model (Model 1) and two multivariable linear regression models (Models 2 and 3). Model 1 examines the unadjusted relationship of adipokines and glucose tolerance status. Model 2 adjusts for markers of inflammation (IL-1β, IL-6, TNF-α, and IL-8), while Model 3 adjusts for the same markers of inflammation as well as demographic and anthropometric characteristics (sex, blood pressure, age, BMI, triglycerides, and HDL-C). Other potential confounding and interaction effects were also explored while developing final multivariable models. Statistical significance was set at p < 0.05. All analyses were performed on SAS® v. 9.4 (SAS Institute Inc., Cary, NC).
3. Results
3.1. Descriptive Demographic, Anthropometric, and metabolic characteristics by glucose tolerance status
These baseline analyses were performed on 953 Mexican-American subjects (61.49% female, average age = 41.36 ± 14.74 years, average BMI = 30.31 ± 6.34 kg/m2). Means and standard deviations are reported for normally distributed data; medians and interquartile range are reported for non-normally distributed data in Table 1.
Table 1.
Demographic, anthropometric and metabolic characteristics by glucose tolerance status.
| Variable | ALL n = 953 | Normal glucose tolerance n = 419 (43.97%) | Prediabetes n = 338 (35.47%) | Diabetes Mellitus n = 196 (20.57%) | p-value* |
|---|---|---|---|---|---|
| Demographic Characteristics | |||||
| Sex, Female N (%) | 586 (61.5) | 253 (60.4) | 204 (60.4) | 129 (65.8) | 0.3771 |
| Age (years), mean (SD) | 41.4 (14.7) | 36.7 (13.2) | 43.3 (14.5) | 47.9 (15.1) | <0.001 |
| BMI (kg/m2), mean (SD) | 30.4 (6.3) | 29.1 (6.4) | 31.2 (6.2) | 32.0 (6.0) | <0.001 |
| Waist circumference (cm), mean (SD) | 101.1 (14.9) | 97.5 (14.3) | 102.8 (14.3) | 105.8 (15.2) | <0.0001 |
| Hypertension, N (%) | 119 (12.5) | 25 (6.0) | 50 (14.8) | 44 (22.5) | <0.0001 |
| Anthropometric and Metabolic Characteristics | |||||
| HbA1c (%), median (IQR) | 5.4 (4.7,6.0) | 4.8 (4.0, 5.3) | 5.7 (5.1, 6.0) | 7.2 (6.5, 9.1) | <0.0001 |
| HbA1c (mmol/mol), median (IQR) | 36 (28, 52) | 29 (20, 34) | 39 (32, 42) | 55 (48, 76) | <0.0001 |
| Fasting blood glucose (mg/dL), median (IQR) | 96.0 (90.0, 105.0) | 91.0 (87.0, 95.0) | 102.0 (95.0, 106.0) | 118.5 (99.0, 170.5) | <0.0001 |
| Insulin (pmol/L), median (IQR) | 12.1 (7.8, 19.0) | 10.7 (7.0, 15.4) | 13.7 (9.6, 21.4) | 13.8 (8.2, 20.4) | <0.0001 |
| HOMA-IR, median (IQR) | 3.0 (1.8, 4.8) | 2.4 (1.5, 3.5) | 3.3 (2.3, 5.3) | 4.5 (2.4, 7.5) | <0.0001 |
| HOMA B, median (IQR) | 130.2 (81.5, 206.7) | 143.7 (94.0, 218.0) | 138.0 (92.5, 213.4) | 81.4 (46.7, 150.4) | <0.0001 |
| Resistin (ng/mL), median (IQR) | 19.2 (8.4, 30.4) | 20.7 (14.0, 31.0) | 19.6 (5.0, 34.0) | 13.1 (1.3, 24.0) | <0.001 |
| Leptin (ng/mL), median (IQR) | 17.1 (7.6, 28.7) | 14.3 (6.7, 25.2) | 20.0 (8.9, 33.1) | 18.7 (7.7, 28.6) | 0.0006 |
| Adiponectin (μg/mL), median (IQR) | 15.0 (10.0, 22.0) | 15.6 (11.1, 22.2) | 14.5 (9.7, 22.8) | 13.7 (9.5, 20.7) | 0.0403 |
| Triglycerides (mg/dL), median (IQR) | 121.0 (82.0, 171.0) | 104.5 (72.0, 142.0) | 132.0 (95.0, 179.0) | 149.0 (100.0, 205.0) | <0.0001 |
| Total Cholesterol (mg/dL), mean (SD) | 180.7 (40.1) | 177.6 (38.9) | 184.7 (42.4) | 180.5 (38.1) | 0.0501 |
| HDL-C (mg/dL), mean (SD) | 46.7 (12.0) | 47.6 (11.2) | 46.6 (13.1) | 44.6 (11.2) | 0.0049 |
| LDL-C (mg/dL), mean (SD) | 107.8 (34.6) | 106.8 (31.7) | 109.9 (35.9) | 106.1 (38.2) | 0.2592 |
| IL-1β (pg/mL), median (IQR) | 0.6 (0.5, 1.1) | 0.6 (0.4, 1.2) | 0.6 (0.6, 1.2) | 0.6 (0.6, 0.9) | 0.0183 |
| IL-6 (pg/mL), median (IQR) | 1.9 (0.9, 4.0) | 1.5 (0.7, 3.1) | 2.4 (1.0, 4.4) | 2.3 (1.0, 4.8) | <0.0001 |
| TNF-α (pg/mL), median (IQR) | 1.9 (0.9, 4.0) | 1.5 (0.7, 3.1) | 2.4 (1.0, 4.4) | 2.3 (1.0, 4.8) | 0.0782 |
| IL-8 (pg/mL), median (IQR) | 4.4 (3.2, 6.4) | 4.4 (3.1, 6.0) | 4.7 (3.3, 7.0) | 4.1 (3.1, 6.0) | 0.1095 |
p-values from Kruskal-Wallis test for non-normal data (summary statistics reported as median and IQR), ANOVA for normal data (summary statistics reported as mean and SD).
The progression of glucose intolerance (from normal glucose tolerance to prediabetes to diabetes mellitus) associated with a significant increase in HOMA-IR (p < 0.0001). As expected, this was also associated with increased age (p < 0.001), BMI (p < 0.001), waist circumference (p < 0.0001), percent of subjects with hypertension (p < 0.0001), HbA1c (p < 0.0001), fasting blood glucose (p < 0.0001), and insulin (p < 0.0001). Corresponding median HOMA-B levels decreased (p < 0.0001). Median triglyceride levels increased (p < 0.0001) and HDL-C levels decreased (p = 0.0049) with worsening glucose tolerance. Decreasing median levels of resistin (p < 0.001) and adiponectin (p = 0.0403) were associated with worsening glucose tolerance. Median leptin levels varied between groups in a nonlinear trend, with the lowest level seen in the normal glucose tolerance group (p = 0.0006). Significant differences across groups were observed in distribution of the inflammatory markers IL-1β (p = 0.0183) and IL-6 (p < 0.0001) (Table 1).
3.2. Univariable and multivariable associations between glucose tolerance status with adipokines
Both univariable and multivariable associations (after controlling for cytokines (IL-1β, IL-6, IL-8, and TNF-α), sex, hypertension, age, BMI, triglycerides, and HDL-C) between glucose tolerance groups (NGT, prediabetes, and diabetes mellitus) and log-transformed values of the above adipokines (resistin, leptin, and adiponectin) were assessed. The unadjusted (Model 1) and adjusted (Models 2 and 3) mean differences in log-transformed resistin (Table 2), leptin (Table 3), and adiponectin (Table 4) between glucose tolerance groups are presented. Model 2 adjusted for cytokines only. Model 3 adjusted for cytokines, sex, hypertension, age, BMI, triglycerides, and HDL-C. These variables were selected to reflect similar models from Menzaghi et al., who described the resistin-cytokine pathway in humans [12]. There were no statistically significant interaction effects observed, and there were no considerable confounders observed other than the variables included in the multivariable models.
Table 2.
Associations Between Glucose Tolerance Status and Log-Transformed Resistin based on Univariable (Unadjusted) and Multivariable (Adjusted) linear regression models.
| Resistin | Model 1 | Model 2 | Model 3 | |||
|---|---|---|---|---|---|---|
| Unadjusted β (SE) | p-value | Adjusted* β (SE) | p-value | Adjusted** β (SE) | p-value | |
| Glucose Tolerance Group | <0.0001 | <0.0001 | <0.0001 | |||
| Prediabetes vs. NGT | −0.234 (0.097) | 0.0157 | −0.266 (0.093) | 0.0042 | −0.313 (0.096) | 0.0011 |
| Diabetes Mellitus vs. NGT | −0.829 (0.115) | <0.0001 | −0.795 (0.110) | <0.0001 | −0.881 (0.117) | <0.0001 |
| Diabetes Mellitus vs. Prediabetes | −0.594 (0.119) | <0.0001 | −0.529 (0.114) | <0.0001 | −0.569 (0.115) | <0.0001 |
| IL-1β (pg/mL) | 0.015 (0.007) | 0.0405 | 0.016 (0.007) | 0.0252 | ||
| IL-6 (pg/mL) | 0.0001 (0.0004) | 0.9339 | 0.0001 (0.0004) | 0.8671 | ||
| TNF-α (pg/mL) | −0.008 (0.005) | 0.1109 | −0.010 (0.005) | 0.0352 | ||
| IL-8 (pg/mL) | 0.060 (0.006) | <0.0001 | 0.059 (0.006) | <0.0001 | ||
| Sex: Female vs. Male | −0.169 (0.088) | 0.0553 | ||||
| Hypertension: yes vs. no | −0.200 (0.131) | 0.1255 | ||||
| Age (year) | 0.005 (0.003) | 0.0991 | ||||
| BMI | 0.016 (0.007) | 0.0145 | ||||
| Triglycerides (mg/dL) | −0.0001 (0.0003) | 0.6402 | ||||
| HDL-C (mg/dL) | −0.011 (0.004) | 0.0026 | ||||
adjustment for IL-1β, IL-6, TNF-α, and IL-8.
additional adjustment for sex, hypertension, age, BMI, triglycerides, and HDL-C.
Note: all models were constructed using linear regressions based on each log-transformed marker (i.e. ln(resistin)).
NGT = normal glucose tolerance.
SE = Standard Error.
r2: 0.051839, 0.131629, 0.156741, for Model 1, 2 and 3, respectively.
Table 3.
Associations Between Glucose Tolerance Status and Log-Transformed Leptin based on Univariable (Unadjusted) and Multivariable (Adjusted) linear regression models.
| Leptin | Model 1 | Model 2 | Model 3 | |||
|---|---|---|---|---|---|---|
| Unadjusted β (SE) | p-value | Adjusted* β (SE) | p-value | Adjusted** β (SE) | p-value | |
| Glucose Tolerance Group | 0.0017 | 0.0021 | 0.0004 | |||
| Prediabetes vs. NGT | 0.274 (0.078) | 0.0004 | 0.269 (0.077) | 0.0005 | 0.068 (0.053) | 0.1957 |
| Diabetes Mellitus vs. NGT | 0.169 (0.092) | 0.0649 | 0.172 (0.092) | 0.06 | −0.181 (0.064) | 0.005 |
| Diabetes Mellitus vs. Prediabetes | −0.105 (0.095) | 0.2718 | −0.097 (0.095) | 0.3105 | −0.249 (0.063) | <0.0001 |
| IL-1β (pg/mL) | −0.003 (0.006) | 0.5886 | −0.005 (0.004) | 0.2447 | ||
| IL-6 (pg/mL) | 0.0001 (0.0003) | 0.8817 | −0.0001 (0.0002) | 0.6312 | ||
| TNF-α (pg/mL) | 0.002 (0.004) | 0.5404 | 0.005 (0.003) | 0.086 | ||
| IL-8 (pg/mL) | 0.009 (0.005) | 0.1058 | 0.007 (0.004) | 0.0603 | ||
| Sex: Female vs. Male | 1.239 (0.048) | <0.0001 | ||||
| Hypertension: yes vs. no | 0.029 (0.072) | 0.6874 | ||||
| Age (year) | 0.004 (0.002) | 0.024 | ||||
| BMI | 0.082 (0.004) | <0.0001 | ||||
| Triglycerides (mg/dL) | 0.0002 (0.0002) | 0.3567 | ||||
| HDL-C (mg/dL) | 0.002 (0.002) | 0.3021 | ||||
adjustment for IL-1β, IL-6, TNF-α, and IL-8
additional adjustment for sex, hypertension, age, BMI, triglycerides, and HDL-C
Note: all models were constructed using linear regressions based on each log-transformed marker (i.e. ln(leptin)).
NGT = normal glucose tolerance
SE = Standard Error
r2: 0.013253, 0.017487, 0.584275, for Model 1, 2 and 3, respectively
Table 4.
Associations Between Glucose Tolerance Status and Log-Transformed Adiponectin based on Univariable (Unadjusted) and Multivariable (Adjusted) linear regression models.
| Adiponectin | Model 1 | Model 2 | Model 3 | |||
|---|---|---|---|---|---|---|
| Unadjusted β (SE) | p-value | Adjusted* β (SE) | p-value | Adjusted** β (SE) | p-value | |
| Glucose Tolerance Group | 0.2013 | 0.2713 | 0.5616 | |||
| Prediabetes vs. NGT | −0.010 (0.065) | 0.883 | −0.023 (0.064) | 0.7228 | 0.008 (0.063) | 0.9008 |
| Diabetes Mellitus vs. NGT | −0.131 (0.077) | 0.0873 | −0.121 (0.076) | 0.1108 | −0.070 (0.077) | 0.3634 |
| Diabetes Mellitus vs. Prediabetes | −0.122 (0.080) | 0.1265 | −0.098 (0.079) | 0.2127 | −0.078 (0.075) | 0.3031 |
| IL-1β (pg/mL) | 0.007 (0.005) | 0.1394 | 0.005 (0.005) | 0.2784 | ||
| IL-6 (pg/mL) | 0.0001 (0.0003) | 0.6396 | 0.0001 (0.0003) | 0.7261 | ||
| TNF-α (pg/mL) | 0.0001 (0.003) | 0.9903 | 0.002 (0.003) | 0.5897 | ||
| IL-8 (pg/mL) | 0.023 (0.004) | <0.0001 | 0.024 (0.004) | <0.0001 | ||
| Sex: Female vs. Male | 0.235 (0.058) | <0.0001 | ||||
| Hypertension: yes vs. no | 0.038 (0.086) | 0.6613 | ||||
| Age (year) | 0.006 (0.002) | 0.0024 | ||||
| BMI | −0.018 (0.004) | <0.0001 | ||||
| Triglycerides (mg/dL) | −0.001 (0.000) | 0.0004 | ||||
| HDL-C (mg/dL) | 0.013 (0.002) | <0.0001 | ||||
adjustment for IL-1β, IL-6, TNF-α, and IL-8.
additional adjustment for sex, hypertension, age, BMI, triglycerides, and HDL-C.
Note: all models were constructed using linear regressions based on each log-transformed marker (i.e. ln(adiponectin)).
NGT = normal glucose tolerance.
SE = Standard Error.
r2: 0.003359, 0.032558, 0.152272, for Model 1, 2 and 3, respectively.
Resistin (Table 2) showed significant differences in Model 1 (unadjusted) and Models 2 and 3 (adjusted) between all three group pair-wise comparisons (prediabetes vs. NGT, diabetes mellitus vs. NGT, and diabetes mellitus vs. prediabetes), highlighting a significant association between lower resistin levels and worsening glucose tolerance (overall p< 0.0001 for each model). Increased IL-1β and IL-8 levels were significantly associated with higher levels of resistin (p = 0.0405 and p < 0.0001 for IL-1β and IL-8 levels respectively) based on Model 2, and these relationships remain intact after further adjusting for demographic and anthropometric characteristics (p = 0.0252 and p < 0.0001 for IL-1β and IL-8 levels respectively). Interestingly in Model 3, TNF-α also appears to decrease with increased resistin levels (p = 0.0352), but the significance was attenuated in Model 2 (p = 0.1109). IL-6 levels showed no significant association with resistin levels in Model 2 (p = 0.9339) and Model 3 (p = 0.8671). Based on Model 3, resistin levels were positively associated with BMI (p = 0.0145) and inversely associated with HDL-C level (p = 0.0026).
We also found significant differences in leptin levels (Table 3) across glucose tolerance groups based on both univariable and multivariable models (overall p = 0.0017 for Model 1, p = 0.0021 for Model 2 and p = 0.0004 for Model 3). Models 1 and 2 showed significantly higher leptin levels in the prediabetes compared to NGT group (p = 0.0004 and p = 0.0005, respectively) and no significant differences between diabetes mellitus vs. NGT nor diabetes mellitus vs. prediabetes. Model 3, however, after further controlling for demographic and anthropometric variables showed significantly lower leptin levels in diabetes mellitus compared to NGT (p = 0.005) and diabetes mellitus with prediabetes (p< 0.0001). Cytokines were not significantly associated with leptin levels based on Models 2 and 3. Higher leptin levels were seen in females (p < 0.0001) and positively associated with age (p = 0.024) and BMI (p < 0.0001).
There were no significant differences in adiponectin levels (Table 4) across glucose tolerance groups based on univariable or multivariable analyses (overall p = 0.2013 for Model 1, p = 0.2713 for Model 2, and p = 0.5616 for Model 3). Model 2 showed that increased IL-8 levels were significantly associated with higher levels of adiponectin (p < 0.0001) and this association remained significant with the further adjustments of Model 3 (p < 0.0001). In Model 3 adiponectin levels were increased in female sex (p < 0.0001), positively associated with age (p = 0.0024) and HDL-C level (p < 0.0001), and inversely associated with BMI (p <0.0001) and triglyceride level (p = 0.0004).
4. Discussion
Plasma resistin levels surprisingly decreased as IR increased in this Mexican-American cohort. When adjusting for metabolic comorbidities and cytokines previously implicated in resistin signaling, three of these four cytokines (IL-1β, TNF-α, and IL-8; IL-6 excluded) associated with resistin levels. The same models showed that IR, cytokines, demographic and anthropometric variables were associated with leptin and adiponectin levels in very different ways.
The novel resistin findings in Mexican-Americans contrast prior studies showing positive correlations between resistin and IR, seen in a Mexican population (i.e., a different cohort with a smaller sample size and excluding diabetes mellitus) [4] as well as a Japanese population [5]. Consistent with our BMI findings, some studies show that obesity is associated with higher serum resistin levels (when compared to non-obese states) in subjects from the United States and Japan [6,7]. However, other studies from Europe and the United States showed no relationship between these groups [8–10]. This inconsistent result in the literature may be explained by different population ethnicities, study sample sizes, tissues analyzed, disease status, and variations in resistin assays [11]. Compared to these studies, our study benefits from a homogenous study group, a larger sample size, and well-defined disease states. The roles of sex differences in adiposity status [10] and other hormonal factors in aging [11] may also affect findings. Notably in our study, higher BMI was associated with higher resistin levels, though sex and age showed no significant effect (Table 2). One may also reasonably speculate that genetic variance led to our findings [27,28]. Another study in a Mexican-Mestizo population demonstrated that that single nucleotide polymorphism (SNP) was associated with lower resistin levels [36]. In other (Korean and Italian) populations, polymorphisms in the RETN gene are implicated in significant variation in serum resistin levels [37,38]. Furthermore, four tag SNPs in the RETN gene were strongly associated with changing resistin levels in participants from the Framingham Offspring Study [39].
Significantly lower leptin levels in diabetes mellitus group were found compared to normal glucose tolerance group significantly when controlling for cytokines, sex, age, and BMI; there were no significant associations found between cytokines and leptin (Table 3). These results (now with a larger sample size) coincide with our prior findings in this cohort, which showed lower leptin levels in individuals with diabetes, compared to no diabetes [13]. However, they contrast those seen in Mexican-Americans from the San Antonio Heart Study, which showed no differences in leptin levels between subjects with and without diabetes mellitus [16]. The subjects from that study were slightly older than our subjects, but interestingly displayed the same associations of leptin with female sex, age, and BMI seen in our study (Table 3). Furthermore, among Mexican populations, multiple studies illustrated the association of increased obesity (a state of increased IR) with increased leptin levels [20,21]. These studies excluded diabetes mellitus and enrolled younger subjects as well as a larger percentage of female subjects. The results are consistent with our findings of positive relationships between leptin and BMI, though they contrast the inverse relationships observed between leptin and IR (Model 3). In spite of differences in ethnicity, subjects from Europe and the Middle East also displayed lower leptin levels in patients with diabetes, when compared non-diabetic control subjects [17–19]. Our cohort population contains an unusually high percentage of Amerindian genes (associated with IR), which highlights the importance of exploring the role of genetic variability [25]. Notably, some genetic studies show that SNPs of LEP and LEPR genes may affect leptin levels in Mexicans [40,41].
There were no significant differences found in adiponectin levels between IR groups, which contrasts prior findings seen in different Mexican-American populations [21,42,43]. However, adiponectin levels were higher in female sex, showed positive associations with age and HDL-C level, and inverse associations with BMI and triglyceride levels (Table 3); this is consistent with prior findings in Mexican-Americans [20]. A positive association between adiponectin and IL-8 level coincides with evidence that adiponectin increases IL-8 in some inflammatory diseases [44]. As Mexican-Americans reflect a diverse multiethnic group [25,42], it is imperative to study these individuals from a genetic perspective. Notably, SNPs in the ADIPOQ gene of Mexican-Americans are associated with IR [43].
There are several limitations to this study. Due to its cross-sectional nature, we cannot infer causal relationships, or specific temporal mechanisms for our findings. The study population is homogenously Mexican-American, so it is difficult to generalize the findings to all Hispanics. Furthermore, findings cannot be generalized to other ethnicities. Differences in genetic variability may play a role in our findings, given the high prevalence of Amerindian genes in this cohort [25].
5. Conclusions
Resistin levels decreased with increasing IR while supporting evidence of a previously-described cytokine pathway in this Mexican-American cohort. Leptin levels fell with elevated IR and were associated with sex, age, and BMI. Adiponectin levels showed no significant association with IR status and appeared to be associated with sex, BMI, and lipid levels. These findings are unique when compared to other ethnicities, including other Hispanics. Further studies should examine the effects of polymorphisms in RETN, LEP, LEPR, and ADIPOQ genes on adipokine levels and IR. As obesity largely influences adipokine levels [6,7,20], this should also be studied separately from the overall metabolic syndrome. One may also explore the mechanisms of these markers in IR by examining the effects of therapeutic agents known to reduce IR-thiazolidinediones and incretin mimetics - in this population.
Acknowledgements
We thank Rocio Uribe and her cohort team, who recruited and documented the participants and obtained the specimens. We also thank Marcela Morris and other laboratory staff for their contributions, and Christina Villarreal and Gloria de Luna for administrative support. We thank Valley Baptist Medical Center, Brownsville, Texas for providing us space for our Center for Clinical and Translational Science Clinical Research Unit. We also thank the communities of Brownsville, Laredo and Harlingen and the participants who so willingly gave their time for this study in their cities.
Funding
This work was supported by P20 MD000170, funded from the National Institute on Minority Health and Health Disparities, and the Center for Clinical and Translational Science Award 1U54RR023417-01 from the National Center for Research Resources. The sponsors support studies in the Cameron County Hispanic Cohort. They had no role in the study design, collection, analysis and interpretation of data, the writing of the manuscript, nor the decision to submit for publication.
Footnotes
CRediT authorship contribution statement
Absalon D. Gutierrez: Conceptualization, Methodology, Investigation, Data curation, Writing – original draft, Supervision. Carlos A. Flores: Conceptualization, Investigation, Writing – review & editing. Sapna Naik: Conceptualization, Investigation, Writing – review & editing. MinJae Lee: Methodology, Software, Validation, Formal analysis, Data curation, Writing – review & editing. Parisa Asgarisabet: Methodology, Validation, Formal analysis, Data curation, Writing – review & editing. Masha Resman: Conceptualization, Investigation, Writing – review & editing. Miryoung Lee: Investigation, Data curation, Writing – review & editing. Joseph B. McCormick: Conceptualization, Methodology, Resources, Writing – review & editing, Funding acquisition. Susan P. Fisher-Hoch: Conceptualization, Methodology, Resources, Writing – review & editing, Funding acquisition.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References
- [1].Narayan KM, Boyle JP, Thompson TJ, Sorensen SW, Williamson DF, Lifetime risk for diabetes mellitus in the united states, JAMA 290 (14) (2003) 1884–1890. [DOI] [PubMed] [Google Scholar]
- [2].Wu S, McCormick JB, Curran JE, Fisher-Hoch SP, Transition from pre-diabetes to diabetes and predictors of risk in mexican-americans, Diabetes Metab. Syndr. Obes 10 (2017) 491–503. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [3].DeFronzo RA, Banting lecture. from the triumvirate to the ominous octet: A new paradigm for the treatment of type 2 diabetes mellitus, Diabetes 58 (4) (2009) 773–795. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Nieva-Vazquez A, Pérez-Fuentes R, Torres-Rasgado E, Lopez-López JG, Romero JR, Serum resistin levels are associated with adiposity and insulin sensitivity in obese hispanic subjects, Metab. Syndr. Relat. Disord 12 (2) (2014) 143–148. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [5].Hasegawa G, Ohta M, Ichida Y, Obayashi H, Shigeta M, Yamasaki M, Fukui M, Yoshikawa T, Nakamura N, Increased serum resistin levels in patients with type 2 diabetes are not linked with markers of insulin resistance and adiposity, Acta Diabetol. 42 (2) (2005) 104–109. [DOI] [PubMed] [Google Scholar]
- [6].Degawa-Yamauchi M, Bovenkerk JE, Juliar BE, Watson W, Kerr K, Jones R, et al. Serum resistin (FIZZ3) protein is increased in obese humans. J Clin Endocrinol Metab 2003;88(11):5452–5. [DOI] [PubMed] [Google Scholar]
- [7].Azuma K, Katsukawa F, Oguchi S, Murata M, Yamazaki H, Shimada A, et al. Correlation between serum resistin level and adiposity in obese individuals. Obes Res 2003;11(8):997–1001. [DOI] [PubMed] [Google Scholar]
- [8].Lee JH, Chan JL, Yiannakouris N, Kontogianni M, Estrada E, Seip R, Orlova C, Mantzoros CS, Circulating resistin levels are not associated with obesity or insulin resistance in humans and are not regulated by fasting or leptin administration: Cross-sectional and interventional studies in normal, insulin-resistant, and diabetic subjects, J. Clin. Endocrinol. Metab 88 (10) (2003) 4848–4856. [DOI] [PubMed] [Google Scholar]
- [9].Heilbronn LK, Rood J, Janderova L, Albu JB, Kelley DE, Ravussin E, et al. Relationship between serum resistin concentrations and insulin resistance in nonobese, obese, and obese diabetic subjects. J Clin Endocrinol Metab 2004;89(4): 1844–8. [DOI] [PubMed] [Google Scholar]
- [10].de Luis DA, Sagrado MG, Conde R, Aller R, Izaola O, Perez Castrillon JL, Dueñas A, Relation of resistin levels with cardiovascular risk factors and insulin resistance in non-diabetes obese patients, Diabetes Res. Clin. Pract 84 (2) (2009) 174–178. [DOI] [PubMed] [Google Scholar]
- [11].Codoñer-Franch P, Alonso-Iglesias E, Resistin: Insulin resistance to malignancy, Clin. Chim. Acta 438 (2015) 46–54. [DOI] [PubMed] [Google Scholar]
- [12].Menzaghi C, Marucci A, Antonucci A, De Bonis C, Ortega Moreno L, Salvemini L, Copetti M, Trischitta V, Di Paola R, Suggestive evidence of a multi-cytokine resistin pathway in humans and its role on cardiovascular events in high-risk individuals, Sci. Rep 7 (1) (2017), 10.1038/srep44337. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [13].Mirza S, Hossain M, Mathews C, Martinez P, Pino P, Gay JL, Rentfro A, McCormick JB, Fisher-Hoch SP, Type 2-diabetes is associated with elevated levels of TNF-alpha, IL-6 and adiponectin and low levels of leptin in a population of mexican americans: A cross-sectional study, Cytokine 57 (1) (2012) 136–142. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].Meek TH, Morton GJ, The role of leptin in diabetes: Metabolic effects, Diabetologia 59 (5) (2016) 928–932. [DOI] [PubMed] [Google Scholar]
- [15].Kershaw EE, Flier JS, Adipose tissue as an endocrine organ, J. Clin. Endocrinol. Metab 89 (6) (2004) 2548–2556. [DOI] [PubMed] [Google Scholar]
- [16].Haffner SM, Stern MP, Miettinen H, Wei M, Gingerich RL, Leptin concentrations in diabetic and nondiabetic mexican-americans, Diabetes 45 (6) (1996) 822–824. [DOI] [PubMed] [Google Scholar]
- [17].Tatti P, Masselli L, Buonanno A, Di Mauro P, Strollo F, Leptin levels in diabetic and nondiabetic subjects, Endocrine 15 (3) (2001) 305–308. [DOI] [PubMed] [Google Scholar]
- [18].Mohammadzadeh G, Zarghami N, Serum leptin level is reduced in non-obese subjects with type 2 diabetes, Int. J. Endocrinol. Metab 11 (1) (2013) 3–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [19].Abdelgadir M, Elbagir M, Eltom M, Berne C, Ahr B [eacute]n, Reduced leptin concentrations in subjects with type 2 diabetes mellitus in sudan, Metabolism 51 (3) (2002) 304–306. [DOI] [PubMed] [Google Scholar]
- [20].Black MH, Shu Y-H, Wu J, Koebnick C, MacKay A, Watanabe RM, Buchanan TA, Xiang AH, Longitudinal increases in adiposity contribute to worsening adipokine profile over time in mexican americans, Obesity (Silver Spring) 26 (4) (2018) 703–712. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [21].Xiang AH, Black MH, Shu Y-H, Wu J, MacKay A, Koebnick C, Watanabe RM, Buchanan TA, Gualillo O, Association of weight gain and fifteen adipokines with declining beta-cell function in mexican americans, PLoS ONE 13 (8) (2018) e0201568. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [22].Hulver MW, Saleh O, MacDonald KG, Pories WJ, Barakat HA, Ethnic differences in adiponectin levels, Metabolism 53 (1) (2004) 1–3. [DOI] [PubMed] [Google Scholar]
- [23].Fisher-Hoch SP, Vatcheva KP, Rahbar MH, McCormick JB, Kirchmair R, Undiagnosed diabetes and pre-diabetes in health disparities, PLoS ONE 10 (7) (2015) e0133135. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [24].Qu H-Q, Li Q, Rentfro AR, Fisher-Hoch SP, McCormick JB, Vella A, The definition of insulin resistance using HOMA-IR for americans of mexican descent using machine learning, PLoS ONE 6 (6) (2011) e21041. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [25].Qu H-Q, Li Q, Lu Y, Hanis CL, Fisher-Hoch SP, McCormick JB, Ancestral effect on HOMA-IR levels quantitated in an american population of mexican origin, Diabetes Care 35 (12) (2012) 2591–2593. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [26].Koh KK, Park SM, Quon MJ, Leptin and cardiovascular disease: Response to therapeutic interventions, Circulation 117 (25) (2008) 3238–3249. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [27].Kotani K, Hazama A, Hagimoto A, Saika K, Shigeta M, Katanoda K, Nakamura M, Adiponectin and smoking status: A systematic review, J. Atheroscler. Thromb 19 (9) (2012) 787–794. [DOI] [PubMed] [Google Scholar]
- [28].Westerink J, LJ F. Visseren, Pharmacological and non-pharmacological interventions to influence adipose tissue function, Cardiovasc. Diabetol 10 (1) (2011) 13, 10.1186/1475-2840-10-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [29].Bajaj M, Suraamornkul S, Hardies LJ, Pratipanawatr T, DeFronzo RA, Plasma resistin concentration, hepatic fat content, and hepatic and peripheral insulin resistance in pioglitazone-treated type II diabetic patients, Int. J. Obes. Relat. Metab. Disord 28 (6) (2004) 783–789. [DOI] [PubMed] [Google Scholar]
- [30].Derosa G, Maffioli P, Salvadeo SAT, Ferrari I, Ragonesi PD, Querci F, Franzetti IG, Gadaleta G, Ciccarelli L, Piccinni MN, D’Angelo A, Cicero AFG, Exenatide versus glibenclamide in patients with diabetes, Diabet. Technol. Ther 12 (3) (2010) 233–240. [DOI] [PubMed] [Google Scholar]
- [31].Derosa G, Putignano P, Bossi AC, Bonaventura A, Querci F, Franzetti IG, Guazzini B, Testori G, Fogari E, Maffioli P, Exenatide or glimepiride added to metformin on metabolic control and on insulin resistance in type 2 diabetic patients, Eur. J. Pharmacol 666 (1–3) (2011) 251–256. [DOI] [PubMed] [Google Scholar]
- [32].Derosa G, Carbone A, D’Angelo A, Querci F, Fogari E, Cicero AFG, Maffioli P, Variations in inflammatory biomarkers following the addition of sitagliptin in patients with type 2 diabetes not controlled with metformin, Intern. Med 52 (19) (2013) 2179–2187. [DOI] [PubMed] [Google Scholar]
- [33].Ferrannini E, Baldi S, Frascerra S, Astiarraga B, Heise T, Bizzotto R, Mari A, Pieber TR, Muscelli E, Shift to fatty substrate utilization in response to sodium-glucose cotransporter 2 inhibition in subjects without diabetes and patients with type 2 diabetes, Diabetes 65 (5) (2016) 1190–1195. [DOI] [PubMed] [Google Scholar]
- [34].American diabetes association. standards of medical care in diabetes - 2010. 2010; 33(Jan):S11–S61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [35].Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC, Homeostasis model assessment: Insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man, Diabetologia 28 (7) (1985) 412–419. [DOI] [PubMed] [Google Scholar]
- [36].Chavarria-Avila E, Ruiz Quezada SL, Guzman-Ornelas MO, Castro-Albarran J, Aguilar Aldrete ME, Vasquez-Del Mercado M, et al. Association of resistin gene 3’UTR+62G>A polymorphism with insulin resistance, adiposity and the adiponectin-resistin index in mexican population. Nutr Hosp 2013;28(6):1867–76. [PubMed] [Google Scholar]
- [37].Menzaghi C, Coco A, Salvemini L, Thompson R, De Cosmo S, Doria A, et al. Heritability of serum resistin and its genetic correlation with insulin resistance-related features in nondiabetic caucasians. J Clin Endocrinol Metab 2006;91(7): 2792–5. [DOI] [PubMed] [Google Scholar]
- [38].Cho YM, Youn B-S, Chung SS, Kim KW, Lee HK, Yu K-Y, Park HJ, Shin HD, Park KS, Common genetic polymorphisms in the promoter of resistin gene are major determinants of plasma resistin concentrations in humans, Diabetologia 47 (3) (2004) 559–565. [DOI] [PubMed] [Google Scholar]
- [39].Hivert M-F, Manning AK, McAteer JB, Dupuis J, Fox CS, Cupples LA, Meigs JB, Florez JC, Association of variants in RETN with plasma resistin levels and diabetes-related traits in the framingham offspring study, Diabetes 58 (3) (2009) 750–756. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [40].López-Quintero A, García-Zapién AG, Flores-Martínez SE, Díaz-Burke Y, Gonźalez-Sandoval CE, Lopez-Roa RI, Medina-Díaz E, Muñoz-Almaguer ML, Sánchez-Corona J, Contribution of polymorphisms in the LEP, LEPR and RETN genes on serum leptin and resistin levels in young adults from mexico, Cell Mol. Biol. (Noisy-le-grand) 63 (8) (2017) 10, 10.14715/cmb/2017.63.8.3. [DOI] [PubMed] [Google Scholar]
- [41].Chavarria-Avila E, Vazquez-Del Mercado M, Gomez-Banuelos E, Ruiz-Quezada SL, Castro-Albarran J, Sanchez-Lopez L, et al. The impact of LEP G-2548A and LEPR Gln223Arg polymorphisms on adiposity, leptin, and leptin-receptor serum levels in a mexican mestizo population. Biomed Res Int 2015;2015:539408. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [42].Pereira RI, Low Wang CC, Wolfe P, Havranek EP, Long CS, Bessesen DH. Associations of adiponectin with adiposity, insulin sensitivity, and diet in young, healthy, mexican americans and non-latino white adults. Int J Environ Res Public Health 2015;13(1):ijerph13010054. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [43].Richardson DK, Schneider J, Fourcaudot MJ, Rodriguez LM, Arya R, Dyer TD, Almasy L, Blangero J, Stern MP, DeFronzo RA, Duggirala R, Jenkinson CP, Association between variants in the genes for adiponectin and its receptors with insulin resistance syndrome (IRS)-related phenotypes in mexican americans, Diabetologia 49 (10) (2006) 2317–2328. [DOI] [PubMed] [Google Scholar]
- [44].Gomez R, Scotece M, Conde J, Gomez-Reino JJ, Lago F, Gualillo O, Adiponectin and leptin increase IL-8 production in human chondrocytes, Ann. Rheum. Dis 70 (11) (2011) 2052–2054. [DOI] [PubMed] [Google Scholar]