Characterization of Cytokine Treatment on Human Pancreatic Islets by Top-Down Proteomics

Abstract

Type 1 diabetes (T1D) results from autoimmune-mediated destruction of insulin-producing β cells in the pancreatic islet. This process is modulated by pro-inflammatory cytokine signaling, which has been previously shown to alter protein expression in ex vivo islets. Herein, we applied top-down proteomics to globally evaluate proteoforms from human islets treated with proinflammatory cytokines (interferon-γ and interleukin-1β). We measured 1636 unique proteoforms across six donors and two time points (control and 24 h post-treatment) and observed consistent changes in abundance across the glicentin-related pancreatic polypeptide (GRPP) and major proglucagon fragment regions of glucagon, as well as the LF-19/catestatin and vasostatin-1/2 region of chromogranin-A. We also observe several proteoforms that increase after cytokine-treatment or are exclusively observed after cytokine-treatment, including forms of beta-2 microglobulin (B2M), high-mobility group N2 protein (HMGN2), and chemokine (C-X-C motif) ligands (CXCL). Together, our quantitative results provide a baseline proteoform profile for human islets and identify several proteoforms that may serve as interesting candidate markers for T1D progression or therapeutic intervention. SUMMARY: This work applies a top-down proteomics workflow for the characterization and label-free quantification of proteoforms from human islets in the context of inflammation. The workflow is optimized for challenges unique to the islet proteome including high disulfide-linkage content and frequent truncation events, resulting in many proteoforms < 5kDa. There are limited examples of top-down proteomics characterization of human islets, thus this study provides a baseline characterization of the proteoforms of major hormones including chromogranin-A (CHGA), chromogranin-B/ secretogranin-1 (CHGB/SCG1), chromogranin-C/ secretogranin-2 (CHGC/SCG2), islet amyloid polypeptide (amylin/IAPP), insulin (INS), glucagon (GCG), pancreatic polypeptide prohormone (PPY), somatostatin (SST), and neurosecretory protein VGF (VGF). The quantitative results of proteoform abundances before and after cytokine treatment, which mimics the proinflammatory environment during T1D progression, provides interesting insights on how prohormone processing is altered under a proinflammatory environment.

Keywords: glucagon; insulin; islet; prohormone processing; top‐down proteomics.

FIGURE 1

(A) Workflow for processing human islets for top‐down proteomic analysis. Created with BioRender.com. (B) Mean number of unique genes (gray) and proteoforms (blue) observed from each human subject (N = 6, t = 2). Error bars represent ± sd. (C) Total unique genes (gray) and proteoforms (blue) found in the entire study. (D) Percentile bins containing proteoforms with different degrees of data completeness, for example, the percentage of acquisitions in which a given proteoform is observed.

FIGURE 2

Summary of the top 30 most abundant insulin (INS) and glucagon (GCG) proteoforms. (A) Median spectral count abundance for INS proteoforms. (B) Plot of INS proteoform truncations and modifications (C) Median spectral count abundance for GCG proteoforms. (D) Plot of GCG proteoform truncations and modifications. Right panels map the first and last amino acid of a given proteoform (x‐axis), and color fill denotes identified PTMs. Dashed vertical lines annotate the region of a given gene. Proteoforms are sorted top to bottom by ascending C‐terminal amino acid ending position, followed by ascending N‐terminal amino acid starting position. Y‐axis labels denote the first and last amino acid of a given proteoform, and “*” is used to denote modified proteoforms.

FIGURE 3

Volcano plots comparing proteoform fold changes pre‐ and post‐cytokine treatment. (A) Volcano plot showing all proteoforms quantified. Proteoforms that are also significant after p value adjustment are annotated with gene names. Color fill denotes if a proteoform has a p value < 0.05 and a log2 fold‐change cutoff of >1 (green) or ←1 (red), else proteoforms are colored black. (B) Volcano plot showing quantified insulin (INS) proteoforms. Color fill denotes which region of INS a proteoform is derived from. “Other” denotes proteoforms that span multiple regions. (C) Volcano plot showing quantified glucagon (GCG) proteoforms. Color fill denotes which region of GCG a proteoform is derived from. (D) Volcano plot showing quantified chromogranin‐C (CHGA) proteoforms. Color fill denotes which region of CHGA a proteoform is derived from. Horizontal dotted line indicates p value cutoff (0.05), and vertical dotted lines indicate log2 fold‐change cutoff of 1 and −1. Point size is scaled to the number of missing values present (i.e., larger point size indicates fewer missing values for a given proteoform).

FIGURE 4

Proteoforms are unique to cytokine treatment condition. (A) Heatmap of proteoforms unique to the cytokine treatment condition. Data is plotted as arbitrary proteoform identifiers (Gene_##) versus treatment_patient identifier. Fill denotes the median normalized log2(Intensity) as determined by label‐free quantification. NA values are shown in gray. Asterisks denote proteoforms that are significant (probability < 0.01) using a hypergeometric probability distribution. (B) Summarizes all proteoforms plotted in panel (A), including the first and last amino acid (firstAA, lastAA) based on the listed UniProt Accession.