Methods
Sample preparation: Insulin lispro and bovine insulin (internal standard) were spiked into rat plasma. Serial dilution was performed to prepare calibration curve standard samples. The samples were protein precipitated, and the supernatants were collected and processed by mixed-mode SPE. The eluents from the SPE plate were diluted with water and injected directly into the LC-MS/MS for analysis.
LC-MS conditions: Samples were analyzed in triplicate by a SCIEX Triple Quad 7500 LC-MS/MS System – QTRAP Ready, coupled with an ExionLC™ system. The method details are summarized in Table 1 and 2. The same sample set was also injected into a SCIEX Triple Quad 6500+ LC-MS/MS System, coupled with the same HPLC system, to measure the sensitivity difference between the two mass spectrometers. All MRM parameters were optimized on both systems for accurate performance comparison.
Data processing: Calibration curves were processed using Analytics in SCIEX OS Software 2.0.
Insulin lispro quantification results
The SCIEX Triple Quad 7500 LC-MS/MS System – QTRAP Ready has multiple novel hardware features to improve instrument sensitivity. First is the OptiFlow Pro Ion Source with E Lens Technology. 1 The E Lens Technology creates a stronger field at the ESI probe that the droplets must traverse, leading to more efficient release of ions from the droplet, and deflection of more ions towards the orifice, improving sensitivity.
Next is the addition of the D Jet Ion Guide, which more efficiently captures and transmits the ions in the higher vacuum region behind the orifice plate. Its tapered dodecapole geometry, efficiently focuses the ions into the second stage QJet® Ion Guide.
Together these technologies result in a significant sensitivity improvement in the SCIEX 7500 System over previous generations. To characterize the sensitivity improvement, the insulin calibration curves were analyzed on both the SCIEX 7500 System and the SCIEX Triple Quad 6500+ System. A 4-fold increase in S/N was observed (Figure 2).
Insulin lispro shares the same molecular weight, and almost the identical amino acid sequence, with human insulin, with only the two C terminal amino acids in B chain switched in position (Figure 1). This makes the differentiation of insulin lispro from human insulin particularly difficult. The only signature fragment ion that differentiates insulin lispro from human insulin is the y2 ion of B chain. All other y ions have the same m/z as human insulin. Therefore, the y2 ion was used as the quantifier to make the acquisition method applicable to human matrices as well. As shown in Figures 3, 4 and 5, the intact insulin lispro was robustly quantified at 0.1 ng/mL in rat plasma, with tight %CV (<13%) and high accuracies (87-111%) across the entire linear dynamic range (0.1-100 ng/mL). 100 ng/mL was the highest concentration prepared in the sample set, therefore the assay ULOQ was not fully characterized.