Effects of human insulin and insulin aspart preparations on levels of IGF-I, IGFBPs and IGF bioactivity in patients with type 1 diabetes

Abstract

Background: Insulin aspart (IAsp) and its biphasic preparations BIAsp50 and BIAsp70 (containing 50% and 70% IAsp, respectively) have distinct glucose-lowering properties as compared to human insulin (HI). We investigated whether this affected the circulating IGF-system which depends on the hepatic insulin exposure.

Methods: In a randomized, four-period crossover study, 19 patients with type 1 diabetes received identical doses (0.2 U/kg sc) of IAsp, BIAsp70, BIAsp50 and HI together with a standardized meal. Serum total IGF-I and IGFBP-1 to -3 were measured by immunoassays for nine hours post-prandially. Bioactive IGF was determined by an in-house, cell-based IGF-I receptor kinase activation (KIRA) assay.

Results: Despite marked differences in peripheral insulin concentrations and plasma glucose, the four insulin preparations resulted in parallel decreases in IGFBP-1 levels during the first 3 hours, and parallel increases during the last part of the study (3-9 hours). Thus, only minor significances were seen. Insulin aspart and human insulin resulted in a lower area under the curve (AUC) during the first 3 hours as compared to BIAsp70 (p = 0.009), and overall, human insulin resulted in a lower IGFBP-1 AUC than BIAsp70 (p = 0.025). Nevertheless, responses and AUCs of bioactive IGF were similar for all four insulin preparations. Changes in levels of bioactive IGF were inversely correlated to those of IGFBP-1, increasing during the first 3 hours, whereafter levels declined (-0.83 ≤ r ≤ -0.30; all p-values <0.05).Total IGF-I and IGFBP-3 remained stable during the 9 hours, whereas IGFBP-2 changed opposite of IGFBP-1, increasing after 3-4 hours whereafter levels gradually declined. The four insulin preparations resulted in similar profiles and AUCs of total IGF-I, IGFBP-2 and IGFBP-3.

Conclusions: Despite distinct glucose-lowering properties, the tested insulin preparations had similar effects on IGF-I concentration and IGF bioactivity, IGFBP-2 and IGFBP-3 as compared to HI; only small differences in IGFBP-1 were seen and they did not affect bioactive IGF. Thus, insulin aspart containing preparation behaves as HI in regards to the circulating IGF-system. However, bioactive IGF appeared to be more sensitive to insulin exposure than total IGF-I. The physiological significance of this finding remains to be determined.

Trial registration: NCT00888732.

Figures

Figure 1

Insulin and glucose profiles. Serum insulin concentration (a) and plasma glucose concentration (b) during 9-hour treatment with insulin aspart (open circles), BIAsp70 (black rectangles), BIAsp50 (open rectangles) and human insulin (black circles). The arrow indicates human insulin injection time, and the vertical dotted line indicates the time of meal. Serum insulin and plasma glucose profiles are illustrated as mean levels (reprinted with permission from Diabetes Technology & Therapeutics).

Figure 2

IGF parameters profiles. (a) Serum total IGF-I concentration, (b) bioactive IGF concentration, (c) IGFBP-1 concentration, *P < 0.01 as BIAsp50 compared with insulin aspart and human insulin for AUC0–3; **P < 0.05, as BIAsp50 compared with human insulin for AUC3–6; ***P < 0.05, as BIAsp70 compared with human insulin for AUC0–9.(d) IGFBP-2 levels and (e) IGFBP-3 levels during 9-hour treatment with insulin aspart (open circles), BIAsp70 (black rectangles), BIAsp50 (open rectangles) and human insulin (black circles). For clarity, serum concentrations are illustrated as mean levels only.

Figure 3

Changes in IGF bioactivity and IGFBP-1. IGF bioactivity (a) and IGFBP-1(b) levels during 0–3 hours and 6–9 hours. Asterisks indicate that significant change was found in comparison with previous level (p < 0.05).

Figure 4

Linear regression between the changes of IGF bioactivity and IGFBP-1 level. 0–3 hours (a) and 6–9 hours (b). The symbols represent patients with insulin aspart (black circles), BIAsp70 (open circles), BIAsp 50 (black triangles) and human insulin (open triangles). For all four insulin preparations a significant inverse correlation with IGFBP-1 was observed, being r = - 0.30 for insulin aspart, r = - 0.47 for BIAsp70, r = - 0.40 for BIAsp50, and r = - 0.52 for human insulin group during 0–3 hours (a), while r = - 0.83 for insulin aspart, r = - 0.57 for BIAsp70, r = - 0.40 for BIAsp50, and r = - 0.58 for human insulin group during 6–9 hours (b); all p-values <0.05. For clarity the reference line illustrates the trend of data in four insulin groups.