A role and mechanism for redox sensing by SENP1 in β-cell responses to high fat feeding

Abstract

Pancreatic β-cells respond to metabolic stress by upregulating insulin secretion, however the underlying mechanisms remain unclear. Here we show, in β-cells from overweight humans without diabetes and mice fed a high-fat diet for 2 days, insulin exocytosis and secretion are enhanced without increased Ca²⁺ influx. RNA-seq of sorted β-cells suggests altered metabolic pathways early following high fat diet, where we find increased basal oxygen consumption and proton leak, but a more reduced cytosolic redox state. Increased β-cell exocytosis after 2-day high fat diet is dependent on this reduced intracellular redox state and requires the sentrin-specific SUMO-protease-1. Mice with either pancreas- or β-cell-specific deletion of this fail to up-regulate exocytosis and become rapidly glucose intolerant after 2-day high fat diet. Mechanistically, redox-sensing by the SUMO-protease requires a thiol group at C535 which together with Zn⁺-binding suppresses basal protease activity and unrestrained β-cell exocytosis, and increases enzyme sensitivity to regulation by redox signals.

Fig. 1. Higher insulin secretion and β-cell exocytosis from islets of overweight donors and 2-day HFD mice.

A Insulin secretion from islets of young (21–45 years) donors at 1 (n = 27, 47, 5), 10 (n = 27, 41, 4) and 16.7 mM (n = 27, 47, 5) glucose. Effect of glucose (^**p = 0.0010; ^***p = 1 × 10⁻¹¹) and vs BMI < 25 (^&p = 0.034). B Exocytosis from β-cells of young (21–45) donors at 1 (n = 55, 120, 22), 5 (n = 130, 236, 36) and 10 mM (n = 88, 151, 21) glucose. Effect of glucose (^**p = 0.0056; ^***p = 4.6 × 10⁻⁷), vs BMI < 25 (^&&p = 0.0039), vs T2D BMI > 25 (^$p = 0.046; ^#p = 0.028). C Insulin secretion from islets of older (>45 years) donors 1 (n = 61, 111, 21), 10 (n = 40, 88, 18) and 16.7 mM (n = 55, 110, 21) glucose. Effect of glucose (^**p = 0.0029; ^***p = 2.3 × 10⁻¹⁰, 2.8 × 10⁻⁷), vs BMI < 25 (^&&p = 0.0047), vs T2D BMI > 25 (^$p = 0.040). D Exocytosis from β-cells of older (>45 years) donors at 1 (n = 164, 285, 78), 5 (n = 310, 474, 112) and 10 mM (n = 173, 364, 112) glucose. Effect of glucose (^*p = 0.031; ^***p = 7.9 × 10⁻¹¹; 5.8 × 10⁻¹¹), vs BMI > 25 (^###p = 1.8 × 10⁻⁴), vs BMI > 25 (^$$$p = 5.3 × 10⁻⁵). Data in A–E compared by two-way ANOVA followed by Tukey post-test. See Supplementary Fig. 1 for breakdown by sex and donor. E IPGTT of mice after CD and 2-day HFD (n = 14, 14). F Plasma insulin during IPGTT (n = 7, 7; **p = 0.0077). G, H Insulin secretion (panel G; n = 10, 11; ^***p = 5.1 × 10⁻⁵; ^**p = 0.0015, ^*p = 0.041, 0.011) and content (panel H; n = 11, 12). AUC – area under the curve (^*p = 0.019). I Representative traces (left), and average total responses, of β-cell exocytosis at 2.8 and 10 mM glucose (n = 21, 26, 32, 33 cells, ^@p = 0.017, from 3 pairs of mice, *p = 0.015, 0.040). J Representative traces, and average Ca²⁺ currents, of β-cells at 2.8 and 10 mM glucose (n = 9, 22, 19, 19 cells, ^*p = 0.023, 0.026, from 3 pairs of mice). K Single cell [Ca²⁺]_i response (n = 16, 34 cells from 4, 5 mice). Data are mean ± SEM, compared with student unpaired Student’s t test (two-sided; panels G, H, K) or two-way ANOVA followed by Tukey post-test (panels F, G, I, J), uncorrected Fisher’s LSD (panel J) or RM two-way ANOVA with matching and Bonferroni post-test (panel I). Source data are provided as a Source Data file.

Fig. 2. RNA sequencing of purified β-cells following 2-day HFD.

A β-cells from CD and 2-day HFD were isolated through fluorescence activated cell sorting (FACS) for RNA sequencing. 213 genes were identified differentially expressed (DE) genes after 2-day HFD (213 genes, n = 4 and 3 mice). No adjustments were made for multiple comparisons. B All DE genes were submitted to Metascape for functional enrichment analysis using the standard accumulative hypergeometric statistical test. Expression of DE genes on CD and HFD were normalized to z-scores by gene and colorized in circos heat map. The six most enriched pathways (PW) are shown. All DE genes were submitted to STRING database for protein-protein interaction. The related protein interactions for DE genes enriched in the six PW were highlighted as links in the circos plot. C Gene set enrichment analysis (GSEA) showed significant up- and down-regulated pathways after HFD with false discovery rate (FDR) less than 0.05. All nominal p-values were less than 0.01. Normalized enrichment score (NES) reflects the degree to which a gene set is overrepresented in a ranked list of genes. D Illustration of transcriptomic changes related to metabolism after 2-day HFD. Raw sequencing data available in the GEO repository (GSE249790).

Fig. 3. Reductive cytosolic redox signaling via SENP1 contributes to enhanced exocytosis during 2-day HFD.

A Oxygen consumption rate (OCR) measured by Seahorse assay (n = 6 pairs of mice, ^*p = 0.017, 0.030; ^**p = 0.0039), with relevant respiration parameters calculated at right as shown by italicized numbers (^**p = 0.0036, 0.0089). B Basal OCR measured by Fluorescence Lifetime Micro Oxygen Monitoring at 2.8 mM glucose (n = 4, 5 mice; ^*p = 0.018). C Representative image of pancreatic islets (white dashed circle) carrying cyto-roGFP2-Orp1 sensor (left). Scale bar = 100 µm. Redox ratio of individual islets (n = 65, 41 islets, ^@@@p = 1.3 × 10⁻¹², from 6 pairs of mice, ^*p = 0.041). D, E Exocytosis with infusion of 10 µM GSH (panel D; n = 27, 30, 31, 38 cells, ^@p = 0.046, 0.017; from 4 pairs of mice, ^*p = 0.026, ^**p = 0.0073) or 200 µM H₂O₂ (panel E; n = 43, 39, 56, 39 cells, ^@@@p = 7.1 × 10⁻⁶, 2.9 × 10⁻⁴; from 7 pairs of mice, ^*p = 0.015, 0.022) at 2.8 mM glucose after 2-day HFD. F Illustration of redox-control of insulin exocytosis. GRX1-glutaredoxin 1; GSH – reduced glutathione; GSSG- glutathione; SENP1- sentrin-specific SUMO-protease 1. G Senp1 expression by qRT-PCR in islets from CD or HFD fed mice (n = 7, 7, 8, 9, 4, 2 mice, ^**p = 0.0017). H Exocytosis with infusion of 4 μg/mL catalytic SENP1 (n = 21, 32, 31, 33 cells, ^@p = 0.032, 0.026, ^@@p = 0.0044; from 4 pairs of mice, ^*p = 0.18) at 2.8 mM glucose. In panels C, D, E, H data are shown as individual cells (gray) or cells averaged by animal (dark). Data are mean ± SEM, compared with students paired t-test (two-sided; panels A–C) or RM two-way ANOVA with matching and Bonferroni post-test (panel A), or two-way ANOVA followed by Tukey post-test (panels D, E, G, H). Levels of significance are indicated for analysis with cells as replicates (‘^@’) or with animals as replicates (*). ^@/*P < 0.05, ^@@/**P < 0.01, ^@@@/***P < 0.001 versus CD or as indicated. Source data are provided as a Source Data file.

Fig. 4. βSENP1-KO mice develop glucose intolerance following 2-day HFD.

A, B IPGTT of male pSENP1-KO and pSENP1-WT mice fed with CD (n = 9, 8 mice) and 2-HFD (n = 10, 11 mice, ^*p = 0.023, 0.027). C, D OGTT of male pSENP1-KO and pSENP1-WT mice fed with CD (n = 6, 6 mice) and 2-HFD (n = 10, 8 mice, ^*p = 0.019, ^**p = 0.0017). E, F Exocytosis from β-cells of βSENP1-WT and βSENP1-KO mice after 2-day (n = 34, 32, 24, 34, 56, 56, 18, 21 cells, ^@@p = 0.0017, 0.0018, 0.0032, 0.0030; from 4, 3, 6, 2 mice, ^*p = 0.020, ^**p = 0.0029, 0.0036, 0.0092, 0.0097) and 4-week HFD (n = 46, 45, 32, 31, 48, 56, 68, 63 cells, ^@p = 0.031, 0.021, ^@@p = 0.0011, 0.0017, 0.0044; from 4, 3, 4, 6 mice, ^*p = 0.027, 0.044). G, H IPGTT of male βSENP1-KO and βSENP1-WT mice fed with CD (n = 6, 5 mice) and 2-day HFD (n = 10, 12 mice, ^*p = 0.044, 0.025). I Plasma insulin levels during IPGTT after CD or 2-day HFD (n = 8, 8 mice, ^**p = 0.0049). J, K OGTT of male βSENP1-KO and βSENP1-WT mice fed with CD (n = 4, 4 mice) and 4-week HFD (n = 13, 11 mice, ^*p = 0.021, 0.040, 0.050, 0.023). L Plasma insulin levels during OGTT after CD or 4-week HFD (n = 8, 8 mice, ^*p = 0.017). In panels E and F data are shown as individual cells (gray) or cells averaged by animal (dark). Data are mean ± SEM, compared with student two-sided unpaired Student’s t test (AUCs) or two-way ANOVA followed either by Bonferroni (two-sided; panels A–D, G–L) or Tukey (two-sided; panels E, F) post-test. Source data are provided as a Source Data file.

Fig. 5. Redox regulation of SENP1 activity requires C535.

A Multiple amino acid comparison among different isoforms of SENP across species. B SENP1 activity after indicated cysteine-to-serine substitution (n = 3 experiments, ^**p = 0.0034, 0.0020, ^***p = 0.00040, 0.00079). C SENP1 activity after cysteine-to-serine mutation and subsequent mutation of histidine 533 or aspartic acid 550 to serine (n = 5 experiments, ^***p = <1 × 10⁻¹⁵, <1 × 10⁻¹⁵, <1 × 10⁻¹⁵, 6.1 × 10⁻⁷). Control here was C608S that does not affect activity, but used to balance thiol groups. D SENP1 activity of control (C608S) SENP1 compared with serine mutants in the presence of H₂O₂ (100 μM), excess histidine (1 mM), or both (n = 4 experiments, Control: ^***p = 2.6 × 10⁻⁹, <1 × 10⁻¹⁵, <1 × 10⁻¹⁵, 3.8 × 10⁻⁸, 9.8 × 10⁻⁸; C535S: ^**p = 0.0013, ^***p = 4.6 × 10⁻⁹, <1 × 10⁻¹⁵, <1 × 10⁻¹⁵ 5.9 × 10⁻⁷; C535S/D550S: ^*p = 0.012, ^***p = 8.3 × 10⁻⁴, 1.3 × 10⁻⁸, 4.9 × 10⁻⁶). E Activity of SENP1 C603S, WT, and C535S upon full activation by DTT (10 mM), and subsequent inhibition by 5 mM H₂O₂ (n = 3 experiments, ^**p = 0.0024, ^***p = 0.9.6 × 10⁻⁹). F In β-cells from pSENP1-KO (red) mice the effect of 4 μg/mL glutathione-S-transferase (GST) peptide (Control), SENP1 WT, or C535S were infused with/without 10 µM H₂O₂ on exocytosis at 5 mM glucose compared with pSENP-WT β-cells (black) (n = 30, 21, 24, 26, 21, 29 cells, ^@p = 0.015, ^@@p = 0.0062, 0.0025, ^@@@p = 5.4 × 10⁻⁵, 9.6 × 10⁻⁴, 1.1 × 10⁻⁵; from 3, 3 mice, *p = 0.040, 0.033, 0.011, **p = 0.0031, 0.0089, 0.0026). G Activity of SENP1 C603S, WT, and C535S in the presence of GSH (0.1 mM) alone or with GRX1 (10 μg/ml). DTT (10 mM) was used to fully activate the enzymes (n = 3 experiments, ^***p = 3.3 × 10⁻⁴, 5.8 × 10⁻⁴, 7.5 × 10⁻⁵, 1.3 × 10⁻⁴). In panel F data are shown as individual cells (gray) or cells averaged by animal (dark). Data are mean ± SEM, compared with RM one-way ANOVA followed by Tukey post-test (panels B, F) or two-way ANOVA followed by Tukey (panels C, D, G) or Bonferroni (panel E) post-test within groups. Source data are provided as a Source Data file.

Fig. 6. Zn²⁺ tunes SENP1 redox sensitivity and SENP1-dependent β-cell exocytosis.

A Activity of SENP1 WT and C535S in the presence or absence of Ca²⁺, Co²⁺, Ni²⁺ and Zn²⁺ (n = 3 experiments, ^**p = 0.0071, ^***p = 0.00029). Data were normalized to C535S activity, as this is more resistant to oxidation. B Dose-response curve of SENP1 inhibition by ZnCl₂ (n = 3 experiments, ^***p = 1.9 × 10⁻¹⁴). C SENP1 inhibition by ZnCl₂ could be reversed with the chelating reagent EDTA (n = 4 experiments, ^***p = 1.6 × 10⁻⁸). D–F SENP1 activity, measured using a native-PAGE assay, inhibited by Zn²⁺-carrying (MT1X, panel D, ^***p = <1 × 10⁻¹⁵, 1.5 × 10⁻⁶, 1.5 × 10⁻¹²) but not Zn²⁺-depleted (T1X, panel E) metallothionine or by Zn²⁺-carrying MT1X in the presence of EDTA (panel F, ^***p = all are 1.9 × 10⁻¹⁴) (n = 6 experiments). G Effect of Zn²⁺ on exocytosis from β-cells of βSENP1-WT and -KO mice at 5 mM glucose (n = 46, 35, 46, 39, 23, 23 cells, ^@p = 0.023, ^@@p = 0.0035, 0.0041, 0.0012, ^@@@p = 7.4 × 10⁻⁷; from 6, 6, 3 pairs of mice, ^*p = 0.030, ^**p = 0.030). H Concentration-response curve of SENP1 activity, with enzyme prepared with or without ZnCl₂ during refolding, to inhibition by H₂O₂. Activity of SENP1 C603S is shown for comparison (n = 6 experiments). I Activity of SENP1 C603S, WT, and C535S in the presence of 1 µM Zn²⁺ and subsequent activation with 5 mM GSH and 10 µg/ml GRX1. 2-mercaptoethanol (2-ME, 1 mM) was used to fully activate enzymes (n = 4 experiments, ^*p = 0.014, ^***p = 3.5 × 10⁻⁶, 5.1 × 10⁻¹⁰, 5.1 × 10⁻¹¹, 2.6 × 10⁻⁶, 6.5 × 10⁻⁵). In panel G data are shown as individual cells (gray) or cells averaged by animal (dark). Data are mean ± SEM, compared with RM one-way ANOVA followed by Tukey post-test (panels C–F) or RM two-way ANOVA followed by Bonferroni post-test within groups (panels A, I) or across conditions (panels B, G). Source data are provided as a Source Data file.