Chain splitting of insulin: an underlying mechanism of insulin resistance?

Production of HI A-chain

HI A-chain was prepared by chain splitting of HI. HI (final concentration 150 µM) was incubated with GSH (1 mM) and GSSG (0.04 mM) for 16 h at 40 °C in 100 mM phosphate buffer, pH = 8. The resulting A-chain was purified using a Jupiter C4 RP-HPLC column (5 µ, 300 Å, 250 × 10 mm, Phenomenex, Torrance, CA, USA) with a linear gradient of acetonitrile in 0.1% TFA at a flow rate of 4 ml/min delivered by Gilson purification system (Middleton, WI, USA) and freeze dried.

In vitro plasma stability and quantification of in vivo samples by liquid chromatography with MS (LC-MS)

Human plasma was obtained from two healthy volunteers under an informed consent. The study and the informed consent were approved by the Danish Scientific Ethical Committee (VEK journal no H-D-2007-0055, approval date 09-NOV-2022). In vitro plasma stability of HI was assessed by incubation of 1 µM HI in 80% EDTA-stabilized human plasma and 20% PBS buffer (pH 7.4) at 37 °C with shaking. At selected time points (1.75, 4.5, 8, 25.5, 49.5, 73.5, 97.5, 169.5 h) one volume of the incubation mixture was subjected to protein precipitation using three volumes of 50% methanol/50% acetonitrile, followed by centrifugation and dilution of one volume of the supernatant with one volume of water before LC-MS analysis.

For quantification of the in vivo rat samples, selected plasma standard was prepared by spiking blank rat plasma with either intact HI (range: 0.45–8300 nM) or the A-chain reference standard (range: 0.05–200 nM) consisting of 2 out of 3 isomers¹. Prior to LC-MS analysis, the plasma standards, blank plasma and study samples were prepared by protein precipitation. For A-chain quantifications one volume of plasma sample was precipitated using three volumes of 50% methanol/50% acetonitrile, followed by centrifugation and dilution of one volume of the supernatant with four volumes of water containing 1% formic acid (FA). For quantification of intact HI and relative B-chain appearance, one volume of plasma sample was precipitated with two volumes of 20% methanol/80% acetonitrile, followed by centrifugation and dilution of one volume of the supernatant with one volume of water.

The LC-MS analysis was carried out on two different LC-MS systems: a TLX-2 TurboFlow high-performance LC system coupled to a Q Exactive Orbitrap mass spectrometer (Thermo Fisher Scientific, Bremen, Germany) for analysis of in vitro plasma stability samples and in vivo A-chain quantifications, and an Acquity I-Class LC system coupled to a Synapt G2-S MS (Waters Company, U.K.) for analysis of in vivo samples for intact HI and B-chain.

On the TLX-2 LC system, the mobile phases consisted of solvent A: Milli-Q water with 5% organic solvent (50% methanol/50% acetonitrile) and 1% FA; and solvent B: Milli-Q water with 95% organic solvent (50% methanol/50% acetonitrile) and 1% FA. For the in vivo A-chain quantification a TurboFlow Cyclone 0.5 × 1000 mm column from Thermo Fisher Scientific (Bremen, Germany) was used for online extraction before elution to the analytical column, whereas the samples from the in vitro plasma stability were loaded directly onto the analytical column. The analytical column used was a XBridge Protein BEH C4, 300 Å, 3.5 µm, 2.1 × 50 mm from Waters (Wilmslow, U.K.) controlled at 60 °C, and elution was performed using a flow rate of 400 µL/min and a linear 40% gradient of mobile phase solvent B from 40% to 80% over 2.5 min for the in vivo A-chain quantification, and for the analysis of the in vitro plasma stability samples a linear gradient of 10–90% B over 5 min was used. The Orbitrap MS was operating in positive ionization mode with a spray voltage of 4.0 kV, and a resolution of 35 K using a single ion monitoring scan (m/z 1187.5–1192.5) for the in vivo A-chain quantifications and a full scan (m/z 300–3500) mode for the in vitro plasma stability samples. The Orbitap LC-MS data were processed using the Quan Browser in the Xcalibur software from Thermo Fisher Scientific (Bremen, Germany) with a lower limit of quantification of the A-chain in the in vivo samples in the range of 0.2–0.5 nM.

On the Acquity LC system, the mobile phases consisted of solvent A: Milli-Q water with 1% FA; and solvent B: Milli-Q water with 90% acetonitrile and 1% FA. The column used was an Acquity Peptide BEH C4 column, 300 Å, 1.7 µm, 2.1 × 150 mm from Waters (Wilmslow, U.K.) controlled at 50 °C, and elution was performed using a flow rate of 400 µL/min and a linear 25% gradient of mobile phase solvent B from 20% to 45% over 30 min. The Synapt G2-S MS was operated in positive ionization mode with a spray voltage of 3.1 kV, using the Resolution mode and recorded using mass intervals of m/z 1155–1170 and m/z 850–865 with target enhancements of m/z 1162 and m/z 858 and 0.25 s scan times, for intact HI and the B-chain, respectively (the mass spectrum of the B-chain contained characteristic multiply charged ions for the B-chain structure with monoisotopic masses of m/z 686.1396 and 857.4118). The Synapt G2-S LC-MS data were processed using the QuanLynx application in the MassLynx software from Waters Company (Wilmslow, U.K.) with a lower limit of quantification of HI in the in vivo samples of 4.3 nM. A mass interval of m/z 500–1500 with 1 s scan time was used for collection of full scan spectrum for qualitative analysis of the rat plasma following 180 min of infusion with HI.

In vivo experiments

Animal procedures were approved by the Danish National Animal Experiments Inspectorate and the Novo Nordisk Ethical Review Council.

In general, the rats had free access to fresh tap water and food (Altromin 1324, Brogaarden, DK) except during fasting periods, and followed a 12:12 h light-dark cycle with light on at 6 am and were allowed to acclimatize for approximately two weeks before any procedures were initiated.

Hyperinsulinaemic euglycaemic clamp study with HI

Male Sprague Dawley rats (Janvier Labs, France) weighing ~350 g at arrival were used. Rats were instrumented with permanent catheters in the Carotid artery (sampling) and Jugular vein (insulin and glucose infusions) 9–11 days before subjected to the clamp experiment which was performed as previously described³². Briefly, rats were fasted overnight, and human insulin (n = 2) was infused i.v. at 2 nmol/kg/min (a supramaximal dose to maximize the chance of detecting free A-chain in plasma) for 180 min while plasma glucose was measured at 10 min intervals and clamped at euglycaemia by adjusting the glucose infusion rates (GIR). Blood samples were taken at 10, 20, 30, 60, 120 and 180 min for determination of insulin and free A- and B-chain concentrations.

I.v. bolus pharmacokinetic study of the HI A-chain

Male Sprague Dawley rats were instrumented with permanent catheters and fasted overnight as described above. The A-chain (1, 7, and 50 nmol/kg) (n = 2–3) was administered i.v. at time zero and blood samples were taken at 2, 4, 6, 20, 60 and 120 min post dosing. Blood samples were immediately centrifuged, and plasma stored at −20 °C until assayed.

Data analysis

Data from the PK experiment with A-chain were analyzed by nonlinear least squares regression analysis of the dose normalized exposure values from all three doses using GraphPad Prism 9.0.1. Data could be fitted to a one-phase decay and the two parameters (intercept and rate constant) for the three different doses were not significantly different, so all the data were fitted to a one-phase decay constraining the parameters to be shared for the three different doses. The rate of chain splitting was then estimated by multiplying the A-chain exposure 180 min into the clamp experiment (2.823 nmol/L) with the A-chain clearance parameter from the PK experiment (0.14 L/kg/min). This calculation assumes that all chain splitting occurs in a single compartment (where sampling was performed) and that the A-chain concentration is in steady state at the end of the clamp experiment.