Identification of Hybrid Insulin Peptides (HIPs) in Mouse and Human Islets by Mass Spectrometry

Abstract

We recently discovered hybrid insulin peptides (HIPs) as a novel class of post-translationally modified peptides in murine-derived beta cell tumors, and we demonstrated that these molecules are autoantigens in type 1 diabetes (T1D). A HIP consists of an insulin fragment linked to another secretory granule peptide via a peptide bond. We verified that autoreactive CD4 T cells in both mouse and human autoimmune diabetes recognize these modified peptides. Here, we use mass spectrometric analyses to confirm the presence of HIPs in both mouse and human pancreatic islets. We also present criteria for the confident identification of these peptides. This work supports the hypothesis that HIPs are autoantigens in human T1D and provides a foundation for future efforts to interrogate this previously unknown component of the beta cell proteome.

Keywords: beta cell proteome; hybrid insulin peptide (HIP); mass spectrometry; pancreatic islets; type 1 diabetes (T1D).

Scheme 1.

Hybrid Insulin Peptide (HIP) Formation^{a a}Representation of HIP formation, in which a C-terminally truncated C-peptide fragment (left peptide) combines with a cleavage product of another beta cell protein (right peptide) via a peptide bond. The mechanism of HIP formation is unknown.

Scheme 2.

In Silico Generation of a Hybrid Insulin Peptide Sequence Database^{a a}Database containing hypothetical HIP sequences was constructed by generating the list of possible truncated C-peptide sequences and combining each of these one at a time with the sequences of natural cleavage products of various beta cell proteins.

Figure 1.

Summary of peptides identified in NOD mouse islets and human islets by mass spectromery. Islets from (a) NOD mice or (b) non-diabetic human donors were processed and proteins were digested with either AspN or GluC, respectively. The resultant peptide mixtures were analyzed by LC-MS/MS. Data were searched against a standard proteome database then against a custom HIP database. Proteins were ranked by the number of distinct peptides identified for the protein and the top eight proteins are reported. HIPs constituted less than 2% of the total number of distinct peptides identified. MS/MS spectra matched to HIPs were later subjected to further analysis to assess the validity of each match. Numbers reported are combined totals across three independent experiments. For each NOD islet experiment, islets were harvested and pooled from a separate group of NOD mice, and each human islet sample was from a different non-diabetic donor.

Figure 2.

Putative HIPs identified by mass spectrometry in islets from NOD mice validate with synthetic peptide standards. Following identification of potential HIPs in NOD islets, the putative HIP sequences were synthesized and the synthetic peptides were analyzed by tandem MS. Mirror plots comparing the MS/MS spectra for the endogenous islet peptide (positive y-axis) and the synthetic peptide standard (reflected y-axis) for two representative HIPs are shown. Band y-ions (singly and doubly charged) are displayed in blue and red, respectively. A mass tolerance of ±20 ppm was used for labeling of band y-ions. Percent scored peak intensity (%SPI), score, and Pearson Correlation Coefficient (PCC) are indicated for each comparison. The difference in chromatographic retention time (ΔRT) for the peptide identified in islets as the sequence DLQTLAL-EVE and the synthetic validation peptide was not determined (n.d.).

Figure 3.

Putative HIPs identified by mass spectrometry in human islets validate with synthetic peptide standards. Following identification of potential HIPs in human islets, the putative HIP sequences were synthesized and the synthetic peptides were analyzed by tandem MS. Mirror plots comparing the MS/MS spectra for the endogenous islet peptide (positive y-axis) and the synthetic peptide standard (reflected y-axis) for two representative HIPs are shown. b- and y-ions (singly and doubly charged) are displayed in blue and red, respectively. A mass tolerance of ±20 ppm was used for labeling of b- and y-ions. The two peptides identified here represent different GluC cleavage products of the same HIP, with the second containing a missed cleavage. Percent scored peak intensity (%SPI), score, and Pearson Correlation Coefficient (PCC) are indicated for each comparison. For the shorter peptide, a retention time match between the endogenous and synthetic peptide was not confirmed.

Figure 4.

Pearson correlation coefficients (PCC) obtained by comparison of mass spectra of endogenous peptides and synthetic HIPs. Fragmentation spectra for putative HIPs identified in NOD and human islets were compared to spectra obtained by analysis of synthetic versions of the peptides. For each comparison, peaks with an abundance greater than three standard deviations above the average in either spectrum were included. Error bars indicate 95% confidence intervals. Values that surpassed the PCC validation threshold of 0.9 are shown in green and those that did not are shown in red.