Insulin-induced changes in skeletal muscle microvascular perfusion are dependent upon perivascular adipose tissue in women

A total of 15 lean (BMI 18–25 kg/m²) and 18 obese (BMI >30 kg/m²) female volunteers participated in this study. Because of the different adipose tissue distribution between men and women, only women participated in this study. Physical health was determined by medical history, physical examination and screening blood tests. Inclusion criteria were: female sex, 18–55 years old and white descent. Exclusion criteria consisted of current illness, history of cardiovascular disease, hypertension, diabetes mellitus or impaired kidney function, use of medication affecting endothelial function or glucose metabolism, physical exercise more than three times a week, recent changes in body weight, pregnancy, postmenopausal state, alcohol abuse and smoking. All participants were recruited through advertisements. The study protocol was approved by the medical ethics committee of the VU University Medical Center, and conducted in accordance with the Declaration of Helsinki. All volunteers provided written informed consent before enrolment in the study.

Participants visited a quiet, temperature controlled room at the Clinical Research Unit on three separate days within 2 months, the first time for a screening visit, the second time for the hyperinsulinaemic–euglycaemic clamp with contrast enhanced ultrasound (CEU) measurements and the third time for the skeletal muscle biopsy. Participants were fasted overnight for the screening and the clamp, and refrained from physical exercise on the day before the clamp, and before and 2 days after the skeletal muscle biopsy. In three instances, the skeletal muscle biopsy was taken before the clamp—CEU measurements were then performed in the contralateral leg to avoid residual effects of wound healing.

Anthropometry was performed at the screening and before the clamp and fat percentage was assessed by bioelectrical impedance analysis (BF906, Maltron, Rayleigh, UK) [5].

After arrival at the Clinical Research Unit, the participants acclimatised for 30 min. Insulin sensitivity was determined by a hyperinsulinaemic–euglycaemic clamp, as described previously and depicted in Fig. 1 [18]. Insulin (Actrapid, Novo Nordisk, Bagsvaerd, Denmark) was infused in a primed (240 mU/m²), continuous manner, at a rate of 40 mU m⁻² min⁻¹ for 120 min. Euglycaemia was maintained at 5 mmol/l, according to whole-blood venous samples (YSI 2300 STAT Plus Analyzer, Yellow Springs, OH, USA), by adjusting the administration rate of the 20% wt/vol. glucose solution, at 5 min intervals. Whole-body glucose uptake or M value, was determined from the glucose infusion rate during the last 60 min of the clamp, and expressed as mg (lean kg)⁻¹ min⁻¹.

CEU measurements were performed with a Siemens-Acuson Sequoia 512 (Siemens-Acuson, Mountain View, CA, USA), equipped with a 17L5 transducer as described, at the time-points indicated in Fig. 1 [5]. The vastus lateralis muscle was imaged 15 cm proximal of the knee. During the baseline measurement, the probe location was outlined, and landmark structures indicated on-screen. For both the baseline and hyperinsulinaemia measurements, freshly prepared microbubbles (SonoVue; Bracco, Milan, Italy) were infused as an undiluted solution, during constant agitation, at a constant rate of 2.5 ml/min for 4 min in both lean and obese participants. After steady-state microbubble concentration was achieved (2.5 min; electronic supplementary material [ESM] Fig. 1), three real-time inflow curves of 30 s were generated at a mechanical index of 0.28 with linear postprocessing, after destruction of the microbubbles at a mechanical index of 1.7 [5, 19]. Video-intensities (VIs) were analysed using the Image Processing toolbox in MATLAB, version R2011a (Mathworks, Natick, MA, USA). Mean VI during the first 0.5 s was subtracted to correct for large vessels and background noise. Real-time curves from the region of interest in skeletal muscle were fitted to the exponential function VI = MBV[1 − e^{−MFV(t − 0.5)}], where t represents the time (s) after microbubble destruction, MBV is microvascular blood volume, MFV is microvascular flow velocity and e is the natural logarithm (See ESM Fig. 1). CEU does not provide an absolute measure of volume flow (ml min⁻¹ [g tissue]⁻¹), but is a relative measure used as a paired measurement within participants, and we therefore report only the percentage change in MBV, to minimise the chance that depth or different composition of skeletal muscle (adiposity) affects the outcome [20]. CEU has been shown to correlate with other methods of estimating insulin effects on the microcirculation [5].

Surgical skeletal muscle biopsies were taken in the non-fasting state from the vastus lateralis muscle at the same location as the CEU measurement. Local anaesthesia was achieved with lidocaine 2% wt/vol. before the open surgical muscle biopsy of approximately 7 mm × 7 mm × 7 mm. The tissue was immediately stored in ice-cold (0–5°C) MOPS-buffer (in mmol/l: 145 NaCl, 4.7 KCl, 3.3 CaCl₂, 2.0 MgSO₄, 1.4 NaH₂PO₄, 2.5 pyruvate, 0.02 EDTA, 3.0 MOPS [3-(N-morpholino) propanesulfonic acid], 5.6 glucose) and quickly transferred to the laboratory to harvest microvessels for testing.

To investigate direct effects of PVAT on insulin-induced vasoreactivity, microvessels were isolated on ice from one half of the skeletal muscle biopsy, and separated from the surrounding PVAT. Microvessels were then mounted on glass cannulae, and randomly assigned to incubation without, or with PVAT. PVAT from alongside its own microvessel was then fastened to one of the cannulae. Ex vivo vasoreactivity of isolated microvessels was studied in the pressure myograph at 80 mmHg and 37°C in K-MOPS (in mmol/l: 125 NaCl, 26 KCl, 3.3 CaCl₂, 2.0 MgSO₄, 1.4 NaH₂PO₄, 2.5 pyruvate, 0.02 EDTA, 3.0 MOPS, 5.6 glucose and 0.1% wt/vol. BSA) [11]. Microvessels were preconstricted by the potassium in the K-MOPS, and inner diameters recorded to determine baseline diameter. Diameter changes induced by four cumulative concentrations of insulin (0.02, 0.2, 2.0 and 20 nmol/l) were examined for 30 min each. To ascertain having isolated an arteriole or resistance artery, and to check endothelial integrity, acetylcholine (ACh) 1 × 10⁻⁷ mol/l and ACh 1 × 10⁻⁶ mol/l was tested at the end of each experiment. At least 10% vasodilation to ACh 1 × 10⁻⁶ mol/l had to be achieved; otherwise, it was excluded from analysis entirely. Maximum diameter was assessed after administration of papaverine (0.1 mmol/l). Vascular tone was expressed as the percentage of the maximum diameter. The insulin-induced vasoreactivity was expressed as the percentage change from baseline after preconstriction.

The second half of the skeletal muscle biopsy was used for histology. This half was stored in buffered formaldehyde (4% wt/vol.) and paraffin-embedded the next morning, 5 μm slices were stained with haematoxylin and eosin. Adipocyte cross-sectional areas were analysed with Image J in a blinded fashion [21]. Only adipocytes at a distance no greater than three adipocytes from the microvessel were included. Macrophage count in PVAT was quantified after CD68 staining and presented as the fraction of the number of adipocytes. For CD68 immunohistochemical analysis, slices were incubated in methanol/H₂O₂ (0.3% vol./vol.) to block endogenous peroxidases. Antigens were retrieved by heat inactivation in citrate buffer, followed by incubation with mouse anti-human CD68 (1:400, DakoCytomation, Glostrup, Denmark). Sections were incubated with Envision (undiluted, anti-mouse and anti-rabbit, DakoCytomation). Staining was visualised using 3,3'-diaminobenzidine (DAB 0.1 mg/ml, 0.02% H₂O₂ vol./vol.) and counterstained with haematoxylin.

Data were analysed with paired (within group) and unpaired (between groups) t tests. Normally distributed data are reported as mean ± SD. Non-normally distributed data were log-transformed, or reported as median and interquartile range and analysed with the Wilcoxon signed-rank test for paired data and the Mann–Whitney test for unpaired data. Pressure myography experiments were analysed using a two-way ANOVA with Bonferroni post hoc test. A p value smaller than 0.05 was considered statistically significant. Linear regression analyses were performed to examine the relations between two variables, controlling for age, and standardised β coefficients are reported. Bias corrected bootstrapping according to the mediation model by Preacher and Hayes was used to assess mediation effects [22]. In short, multiple regression analyses are performed, and the proposed mediator is important if the effect of the primary factor on the dependent decreases substantially when the proposed mediator is entered. The significance of this mediation is estimated through bootstrapping, where multiple random subsets of the dataset are run to estimate the significance of the change in β coefficient; this significance is indicated with a CI. Analyses were performed using IBM SPSS Statistics version 21 (Armonk, NY, USA) and Graphpad Prism 5.01 (La Jolla, CA, USA).