MSACL 2024 Abstract

Introduction
Proinsulin is processed by a series of endopeptidase cleavages to yield intact insulin and C-peptide and typically circulates at low levels. However, perturbation of this metabolic pathway can indicate multiple metabolic, endocrine, or oncological pathologies. For example, recent studies suggest that elevations in the proinsulin/insulin ratio may be associated with risk of developing type 2 diabetes mellitus (T2DM). An increased proinsulin/insulin ratio may also reflect beta cell stress as demand on limited insulin secretory capacity is increased. Insulin and C-peptide serum measurements are primarily used for diagnosing glycemic disorders in diabetic and pre-diabetic patients and assessing insulin resistant syndromes as well as distinguishing insulin-secreting tumors from exogenous insulin administration as a cause for hypoglycemia.

Currently, proinsulin is commonly measured by immunoassay, which is limited by cross-reactivity with proinsulin intermediates des31,32- and des64,65-proinsulin (also known as “split proinsulins”). No commercially available assays can simultaneously and selectively quantify all proinsulin processing pathway intermediates. Here, we describe the development and characterization of an assay that quantifies intact proinsulin, the split proinsulins, insulin, C-peptide, and enzymatically missed-cleavage products of C-peptide with 1 to 2 additional N-terminal arginine residues (+R, +RR) resulting from incomplete proprotein convertase processing. Simultaneous measurement of these analytes provides the means for accurately characterizing proinsulin processing pathway dysfunction by identifying and quantifying the major metabolic intermediates or pathological products that may not distinguished by immunoassays.

Methods
Pre-analytic sample preparation was performed entirely using a TECAN Freedom Evo automated liquid handler. Matrix-matched calibrators, quality controls, and patient serum was delipidated with Cleanascite® after addition of internal standards (bovine insulin and stable isotope-labelled C-peptide, +R, and +RR). Target analytes were then immunoprecipitated with anti-insulin beta-chain monoclonal IgG and anti-C-peptide monoclonal IgG conjugated to magnetic beads. After sequential washing with sodium chloride, PBS, and water, enriched analytes were eluted from the beads with 0.1:30:70 (vol/vol/vol) formic acid:acetonitrile:water. Samples were then analyzed by on-line SPE-LC-MS/MS utilizing 4 multiplexed LC-streams coupled to a Shimadzu 8060NX triple quadrupole mass spectrometer. Analyte trapping was achieved with a Waters HLB on-line extraction cartridge and separation performed with a Phenomenex Kinetex C18 50x2.1 mm, 2.6 µm column. Analytes were eluted and detected within a 2.25 minute window of a 9-minute LC-MS/MS method, enabling 100% duty cycle utilization. Three MRM transitions were monitored for each analyte.

Results
Analytical measurement ranges (AMR) were established for insulin (3-320 uIU/mL), C-peptide and its +R and +RR missed-cleavage products (0.11-27.2 ng/mL for all 3 peptides) using calibration standards established using secondary reference materials with purity and peptide content confirmed by quantitative amino acid analysis (AAA). The AMR for proinsulin (3-500 pM) was established using primary NIBSC Reference Standard 09/296 for calibration. Because des31,32-proinsulin (AMR: 3-500 pM) and des64,65-proinsulin (AMR: 3-200 pM) reference standard materials were not commercially available, calibration standards were generated by digestion of proinsulin by trypsin and carboxypeptidase-B, with concentration assigned by LC-UV-Vis, traceable to NIBSC 09/296. Total imprecision for all analytes was less than 10%. The assay was free from serum index interferences except in cases of gross hemolysis and the presence of certain insulin analogue medications. No significant carryover was observed. A comparison using deidentified patent specimen discards (4 – 124 nM; n=81) between the described LC-MS/MS method and a proinsulin ELISA yielded a Deming Regression slope of 0.4207 and intercept of 2.645 pM with Spearman r of 0.6967 for proinsulin. Interestingly, the Spearman r improved (r=0.9034) when summing proinsulin, des31,32-proinsulin, and des64,65-proinsulin, indicating cross-reactivity of the ELISA antibodies with proinsulin metabolic intermediates.

Discussion
Improving standardization and traceability is paramount to high quality clinical diagnostics by ensuring results and clinical decision points are comparable from laboratory-to-laboratory. We have developed a multiplexed assay for the determination of proinsulin and its downstream metabolic products with performance characteristics in line with our pre-existing lab developed test for intact insulin/C-peptide by immunoprecipitation coupled with LC-MS/MS. Measurement of proinsulin by ELISA presents a challenge due to cross-reactivity of antibodies with intermediate forms. We leveraged this cross-reactivity, utilizing anti-insulin and anti-C-Peptide antibodies to enrich proinsulin and its metabolites from serum followed by LC-MS/MS, which offers greater analytical specificity for the discrimination of the targeted analytes. The bias between the proinsulin ELISA and LC-MS/MS assay is likely the result of differences in standardization. The proinsulin LC-MS/MS assay is calibrated directly using the NIBSC 09/296 proinsulin reference standard, which was validated across 17 independent laboratories with assigned quantities established with a ±10% error. The proinsulin ELISA described in our comparison uses a secondary reference standard traceable to NIBSC proinsulin standard 84/611, which was assigned an “approximate” mass based on measurements in 3 labs with no established range of uncertainty. Our assay enables the traceable quantification of key proinsulin pathway intermediates and suspected processing variants and a deeper understanding of their involvement in various pathologies.