MON-271 Identification of IGF-1 Variants in a Clinical Study of 307,000 Specimens: Discrepancies with General Population and Novel Discoveries

High resolution accurate mass (HRAM) mass spectrometry (MS) is a highly specific and robust method for identifying and quantifying intact IGF-1. However, protein variants with mass-to-charge ratios differing from wild type (WT) are not included in quantification. Previously, we developed a naming convention for isotopic peaks, called the Isotopic Peak Index (IP(i)) that helps identify variants by comparing their IP(i)s. This method cannot differentiate some variants; however, the relative retention time (rRT), which is defined as the difference between the chromatographic RT of a variant and that of the WT, may help in some cases. We report IGF-1 variants detected using rRT and IP(i); we also compare expected and observed frequency of common variants A67T and A70T. For IPi, 4 variant groups (VGs), each monitored at a single m/z ratio, were identified to make detection of 15 variants more efficient. Variants within each VG could be distinguished if they had different IP(i)s. To distinguish some variants in the same VG with the same IP(i), we used rRT. We also developed a tandem mass-spectrometry (MS/MS) method that uses specific fragment ions generated during fragmentation to distinguish between the most abundant variants, A67T and A70T. Of the 307,269 samples we analyzed, 1,266 (0.4%) variants were identified. IP(i) identified variants with different indices and the rRT approach distinguished between pairs of variants that IP(i) could not (A38V vs A67V; R55K vs R36Q). The following variants (and their count) were identified: R50W/T4M/A67T/A70T (1,210), A67V (23), A38V (9), P66A (6), R36Q/R50Q (5), V17M/V44M (5), A67S (4), R55K/R56K (3), T29I/T41I (1), S34N (1), S33P (1), Y31H (1). We observed a different relative frequency of A67T and A70T variants than expected based on a general population ExAC DNA database. We expected to detect 2.1-fold more A67T than A70T counts. However, DNA sequencing of 74 samples in this study identified 1.6-fold more A70T occurrences (46) than A67T (28). Overrepresentation of A70T may suggest its clinical significance. The A38V variant was predicted to occur 5.1-times in the population studied, but it was detected 9 times. The A67V variant was not present in the ExAC database, but it was detected in 23 patients. We detected 2 variants that did not match WT or expected variants; DNA sequencing identified them as new variants (S33P, Y31H). DNA sequencing also identified 3 new variants (R50Q, R56K, T41I) at different positions than expected. Using a few test samples, the tandem MS technique has shown its ability to distinguish between A67T and A70T. It will be applied to this patient set. IP(i) and rRT can be used together to identify and differentiate IGF-1 variants. Such an approach may provide improved estimates of variant frequencies and identify new variants.