Insulin Resistance Is Associated With Increased Serum Concentration of IGF-Binding Protein–Related Protein 1 (IGFBP-rP1/MAC25)

A total of 113 nondiabetic healthy men, consecutively enrolled in a prospective study dealing with insulin sensitivity in northern Spain, were included in the current study. None of these subjects had evidence of metabolic disease other than nonmorbid obesity. Indeed, type 2 diabetes was ruled out by an oral glucose tolerance test. Exclusion criteria for this group were: 1) BMI ≥40 kg/m² and 2) concurrence of any systemic disease or medication use.

We also studied 43 type 2 diabetic patients, defined according to the criteria of the American Diabetes Association and prospectively recruited from diabetes outpatient clinics at the Hospital of Girona. All type 2 diabetic patients had stable metabolic control in the previous 6 months. Pharmacological therapy for these subjects included insulin (18%), oral hypoglycemic agents (45%), statins (25%), fibrates (10%), blood pressure–lowering agents (28%), aspirin (13%), and allopurinol (10%). Exclusion criteria for this group were 1) clinically significant hepatic, renal, neurological, endocrinologic, or other systemic disease, including malignancy; 2) current clinical evidence of hemochromatosis; and 3) an acute major cardiovascular event in the previous 6 months. A total of 13 patients exhibited one or more chronic diabetic complications, including macroangiopathy (3 cases), retinopathy (7 cases), and nephropathy (7 cases).

Both nondiabetic and type 2 diabetic subjects were of Caucasian origin, reported that their body weight had been stable for at least 3 months before the study, and were free from intercurrent illnesses or chronic infections at time of the study. Informed written consent was obtained after the purpose, nature, and potential risks of the study were explained to the subjects. The experimental protocol was approved by the ethics committee of the Hospital of Girona.

Each subject was studied in the postabsorptive state. BMI was calculated as the weight (in kilograms) divided by height (in meters) squared. Subjects’ waists were measured with a soft tape midway between the lowest rib and the iliac crest, hip circumference was measured at the widest part of the gluteal region, and waist-to-hip ratio was calculated accordingly. Blood pressure was measured in the supine position on the right arm after a 10-min rest; a standard sphygmomanometer of appropriate cuff size was used, and the first and fifth phases were recorded. Values used in the analysis are the average of three readings taken at 5-min intervals.

Insulin sensitivity was measured by frequently sampled intravenous glucose tolerance tests in nondiabetic subjects and by insulin tolerance tests in type 2 diabetic subjects. In brief, the experimental protocol for intravenous glucose tolerance test studies started between 8:00 and 8:30 a.m. after an overnight fast. A butterfly needle was inserted into an antecubital vein, and patency was maintained with a slow saline drip. Basal blood samples were drawn at −30, −10, and −5 min, after which glucose (300 mg/kg body wt) was injected over 1 min starting at time 0, and insulin (Actrapid, 0.03 units/kg; Novo Nordisk, Bagsværd, Denmark) was administered at time 20 min. Additional samples were obtained from a contralateral antecubital vein up to 180 min, as previously described (14).

For insulin tolerance test studies, the same preparation measures were followed, after which a bolus of insulin (Actrapid, 0.1 unit/kg; Novo Nordisk) was administered and blood was sampled from the dorsum of the same hand. To arterialize the venous blood, the hand was placed on a hot box at a constant temperature of 40°C for 20 min before the start of the study and kept there until the end of the test. Sampling was carried out every minute until 15 min after the injection of insulin. Insulin sensitivity was indicated by the first-order rate constant for disappearance of glucose (K_ITT) estimated from the slope of the regression line of the logarithm of blood glucose against time during the first 3–15 min of an insulin tolerance test. Medications, including insulin and oral hypoglycemic agents, were withheld within 12 h of the insulin tolerance test.

HDL cholesterol was quantified after precipitation with polyethylene glycol at room temperature. Total serum triglycerides were measured through the reaction of glycerol-phosphate-oxidase and peroxidase. Serum glucose concentrations were measured in duplicate by the glucose oxidase method, using a Beckman Glucose Analyzer II (Beckman Instruments, Brea, CA). The coefficient of variation (CV) was 1.9%.

Serum insulin concentrations were measured in duplicate by a monoclonal immunoradiometric assay (Medgenix Diagnostics, Fleunes, Belgium). Intra- and interassay CVs were <7%. Serum adiponectin concentrations were measured by radioimmunoassay (Linco Research, St. Charles, MO). Samples were diluted 500 times before the assay. Sensitivity of the method is 2 mg/l. The intra- and interassay CVs were <5%. Serum C-reactive protein (CRP) was determined by an immunoturbidimetric assay (Beckman, Fullerton, CA) with intra- and interassay CVs <4% and a sensitivity of 1.0 mg/l. Serum soluble tumor necrosis factor (TNF) receptor 1 (sTNFR1) and sTNFR2 levels were analyzed, using commercially available solid-phase enzyme-amplified sensitivity immunoassays: sTNFR1 and sTNFR2 enzyme-amplified sensitivity immunoassay (Biosource Technologies, Fleunes, Belgium). The intra- and interassay CVs were <7 and <9%, respectively. sTNFR1 enzyme-amplified sensitivity immunoassay has no cross-reactivity with sTNFR2, and TNF-α does not interfere with the assay. Serum IGF-I, free IGF-I, and IGFBP-1 were measured by immunoradiometric assay (Diagnostic Systems Laboratories, Webster, TX). Sensitivities of these assays are 0.80, 0.03, and 0.33 μg/l, respectively. Intra- and interassay CVs were <9, <11, and <7%, respectively. Serum IGFBP-rP1 was measured by a recently reported enzyme-linked immunosorbent assay with a detection limit of 0.7 μg/l and intra- and interassay CVs <7% (7). Fasting C-peptide was measured in type 2 diabetic subjects by means of a fluorometric immunoassay (EG&G Wallac, Turku, Finland) with intra- and interassay CVs <6%.

Statistical analyses were performed, using SPSS 12.0 software. Unless otherwise stated, descriptive results of continuous variables are expressed as the means ± SD for Gaussian variables and as median and interquartile range for non-Gaussian variables. Parameters that did not fulfill normal distribution were mathematically transformed to achieve normality for subsequent analyses. The relation between variables was analyzed by simple correlation (Pearson’s test) and multiple regression in a stepwise manner. Unpaired Student’s t test was used to seek differences between nondiabetic and type 2 diabetic subjects. One-way ANOVA followed by ANCOVA, using general linear models (to correct for the effect of age and BMI among groups), were used to seek differences in insulin sensitivity and inflammatory markers among quartiles of IGFBP-rP1 in nondiabetic subjects. Levels of statistical significance were set at P < 0.05.

For a given value of P = 0.05, the study had an 80% power to detect a significant difference in insulin sensitivity index (S_i) of at least 0.7 SD between subjects in the highest and lowest quartile of serum IGFBP-rP1. The study was also powered to detect a difference of at least 0.6 SD in serum IGFBP-rP1 between nondiabetic and type 2 diabetic men.

Clinical and biochemical variables of the study subjects are summarized in Table 1.

To study whether IGFBP-rP1 could be associated with metabolic parameters and components of the IGF system in humans, nondiabetic subjects were stratified according to IGFBP-rP1 quartiles. Subjects in the highest quartile for serum IGFBP-rP1 were more insulin resistant (P < 0.01) and showed decreased serum adiponectin (P < 0.01) and increased CRP and sTNFR2 levels (both P < 0.05) compared with subjects in the lowest quartile (Table 2 and Fig. 1). Decreased serum-free IGF-I was also documented in subjects with the highest serum IGFBP-rP1; however, this association did not quite reach statistical significance after adjusting for age and BMI (P = 0.07).

The relationship between IGFBP-rP1 and S_i was also tested by comparing insulin-resistant with insulin-sensitive subjects (defined as having S_i values below and above the median for this group, respectively). Insulin-resistant subjects had higher serum IGFBP-rP1 levels than insulin-sensitive subjects (mean [95% CI] of 29 μg/ml [27–31] vs. 25 [24–27], respectively; P = 0.002 nonadjusted, and P = 0.02 adjusted for age and BMI).

On bivariate correlations, IGFBP-rP1 was significantly associated with age (r = 0.31, P = 0.001), serum adiponectin (r = −0.24, P = 0.01), sTNFR2 (r = 0.27, P = 0.004), free IGF-I (r = −0.27, P = 0.005), and S_i (r = −0.28, P = 0.004) and showed borderline correlations with BMI, diastolic blood pressure, serum triglycerides, CRP (all r = 0.18–0.20, P ≤ 0.05), and IGF-I (r = −0.19, P < 0.05). Except for BMI, diastolic blood pressure, and IGF-I, all of these associations persisted after adjusting for age in partial correlation analyses. Although IGFBP-rP1 was not associated with serum IGFBP-1 on bivariate analysis (r = 0.11, P = 0.27), a significant correlation was evident after adjusting for BMI (a known covariate for circulating IGFBP-1 in these subjects; r = 0.25, P = 0.027).

On stepwise multiple regression analyses, S_i, sTNFR2, and age, but not BMI or free IGF-I, were independent predictors of IGFBP-rP1, explaining 9, 6, and 3% of its variance, respectively (Table 3). Although adiponectin did not enter in the final model with S_i as independent variable, subjects in the highest quartile for serum IGFBP-rP1 had persistently lower circulating adiponectin after adjusting for age, BMI, and S_i (data not shown), indicating that the association between IGFBP-rP1 and adiponectin is independent of S_i. A multiple regression analysis was also constructed with S_i as a dependent variable. In this model, BMI (β = −0.43, P < 0.0001), adiponectin (β = 0.19, P = 0.017), IGFBP-rP1 (β = −0.21, P = 0.011), and IGFBP-1 (β = 0.19, P = 0.034) were independent predictors of S_i, explaining 32, 6, 3, and 2% of its variance, respectively.

Finally, to test whether IGFBP-rP1 could be independently associated with components of the IGF system, multiple regression models were constructed with IGF-I, free IGF-I, or IGFBP-1 as dependent variables. Although no associations were documented for either IGF-I (not shown) or free IGF-I (Table 4) in these analyses, a modest but independent positive association of IGFBP-rP1 with serum fasting IGFBP-1 was evident, with IGFBP-rP1 explaining 2% of IGFBP-1 variance (Table 4). The strength of this relationship was increased when both nondiabetic and type 2 diabetic subjects were pooled together in a single group (β = 0.22, P = 0.002; see below).

As a group, type 2 diabetic patients had IGFBP-rP1 serum concentrations similar to those in nondiabetic subjects (Table 1). However, circulating IGFBP-rP1 was found to be increased in previously undiagnosed type 2 diabetic patients (n = 10) compared with both known type 2 diabetic patients receiving pharmacological therapy (n = 33) and nondiabetic subjects (IGFBP-rP1 of 35 μg/l [29–41], 26 [21–32], and 26 [22–30] in newly diagnosed type 2 diabetic, known type 2 diabetic, and nondiabetic subjects, respectively; P = 0.001, nonadjusted; P = 0.01, adjusted for age and BMI). These changes were independent of the modifications in circulating IGF-I, free IGF-I, and IGFBP-1 from nondiabetic to new type 2 diabetes and to known type 2 diabetes because only the latter subjects had significantly decreased serum IGF-I and free IGF-I (P < 0.0001) and increased serum IGFBP-1 (P < 0.01) (Fig. 2).

IGFBP-rP1 was associated with BMI in type 2 diabetes (r = 0.44, P = 0.004) and showed borderline correlations with CRP and fasting C-peptide (both r = 0.34, P ≤ 0.05), but not with age, fasting insulin, adiponectin, sTNFR2, IGF-I, or free IGF-I. IGFBP-rP1 was also negatively correlated with insulin sensitivity (K_ITT) after excluding one type 2 diabetic patient with very high fasting circulating insulin (>50 mIU/l; r = −0.44, P = 0.018). Similar to nondiabetic subjects, IGFBP-rP1 was not associated with serum IGFBP-1 on bivariate analysis (r = 0.12, P = 0.52) but showed a significant correlation with this binding protein after adjusting for fasting C-peptide (a known covariate for circulating IGFBP-1 in these subjects; r = 0.42, P = 0.021).

In multiple regression models, BMI stood as the only independent predictor of serum IGFBP-rP1 (β = 0.54, P = 0.005), explaining 26% of its variance. Multivariate models were also constructed with IGF-I, free IGF-I, and IGFBP-1 as dependent variables. Although no associations were evident between IGFBP-rP1 and either IGF-I or free IGF-I (not shown), IGFBP-rP1 was an independent predictor of IGFBP-1 (β = 0.39, P = 0.016), explaining 11% of its variance. Other predictive variables for IGFBP-1 were fasting C-peptide (β = −0.74, P < 0.0001) and adiponectin (β = 0.30, P = 0.033), which explained 30 and 8% of its variance, respectively.

To our knowledge, this is the first report of an association between insulin resistance and IGFBP-rP1 in humans. Although significant binding affinities (>10⁶ mol/l⁻¹) of IGFBP-rP1 for either insulin or IGFs have not been definitely proven (5,12,15), the fact that the protein can compete for insulin binding to its receptor and can inhibit insulin receptor and insulin receptor substrate-1 phosphorylation (12) suggests a possible contribution of IGFBP-rP1 to the insulin resistance syndrome in humans. This thesis is supported by the independent associations between IGFBP-rP1 and S_i in nondiabetic subjects and between IGFBP-rP1 and BMI in type 2 diabetic patients in the current study.

Second, in addition to its relationship with S_i, IGFBP-rP1 was significantly associated with both anti-inflammatory (adiponectin) and proinflammatory (sTNFR2 and CRP) molecules, all of which have been shown to play a role in the development and modulation of insulin resistance (14,16,17). Although it is uncertain whether IGFBP-rP1 may upregulate the inflammatory response and inhibit adiponectin expression and/or secretion, or whether IGFBP-rP1 is downregulated by adiponectin and stimulated by inflammation, it can be concluded that the association between IGFBP-rP1 and insulin resistance may also result from modulation of the chronic low-grade inflammatory response. In support of the latter association are recent findings showing upregulation of tissue IGFBP-rP1 expression (liver and muscle) after induction of inflammation by either thermal injury or intravenous injection of TNF-α in rats (18,19).

Third, IGFBP-rP1 serum concentrations were found to be increased in newly diagnosed type 2 diabetic patients, an observation that strengthens the association of IGFBP-rP1 with insulin resistance. Although circulating IGFBP-rP1 failed to be independently associated with insulin sensitivity in type 2 diabetes, it was strongly related to BMI, an important surrogate for insulin resistance in these subjects. An additional effect of increasing blood glucose on IGFBP-rP1 should be considered here because moderate hyperglycemia is associated with enhanced IGFBP-rP1 immunoreactivity in human cancer tissues (20). It is also tempting to speculate that pharmacological therapy can reverse the effect of IGFBP-rP1 on insulin resistance and inflammation in type 2 diabetic patients. The design of our study, however, precludes a detailed analysis of the interactions between insulin sensitivity, inflammation, and IGFBP-rP1 taking into consideration the effect of pharmacological therapy.

Fourth, our data indicate that IGFBP-rP1 is independently associated with higher serum IGFBP-1 concentration in both nondiabetic and type 2 diabetic subjects. Because IGFBP-1 can regulate the bioactivity of IGFs and because overexpression of IGFBP-1 results in hyperinsulinemia and glucose intolerance (21), we suggest that the association of IGFBP-rP1 with insulin resistance can be also explained, at least in part, by possible interactions of IGFBP-rP1 with the IGF system.

Finally, IGFBP-rP1 has been shown to be induced during senescence in several assays in vitro (4,22). The positive correlation between IGFBP-rP1 and age, independent of other covariates, appears to indicate that increased IGFBP-rP1 is also associated with the deterioration of insulin sensitivity observed in human aging.

In summary, IGFBP-rP1 is increased with insulin resistance and shows novel associations with the IGF system in humans.

This study was supported, in part, by grants PI041407 (to A.L.-B.) and G03/212 and G03/028 (to J.M.F.-R.) from the Fondo de Investigación Sanitaria, Health Institute Carlos III, Spain, and partly funded by grant BFU2004-03654/BFI (to J.M.F-R.) from the Ministry of Education and Science, Spain. A.L.-B. is a Research Investigator of the Fund for Scientific Research “Ramon y Cajal” (Ministry of Education and Science, Spain).