Characterizing the effects of Dechlorane Plus on β-cells: a comparative study across models and species

Abstract

Epidemiological studies consistently link environmental toxicant exposure with increased Type 2 diabetes risk. Our study investigated the diabetogenic effects of a widely used flame retardant, Dechlorane Plus (DP), on pancreatic β-cells using rodent and human model systems. We first examined pancreas tissues from male mice exposed daily to oral gavage of either vehicle (corn oil) or DP (10, 100, or 1000 μg/kg per day) and fed chow or high fat diet for 28-days in vivo. DP exposure did not affect islet size or endocrine cell composition in either diet group. Next, we assessed the effect of 48-hour exposure to vehicle (DMSO) or DP (1, 10, or 100 nM) in vitro using immortalized rat β-cells (INS-1 832/3), primary mouse and human islets, and human stem-cell derived islet-like cells (SC-islets). In INS-1 832/3 cells, DP did not impact glucose-stimulated insulin secretion (GSIS) but significantly decreased intracellular insulin content. DP had no effect on GSIS in mouse islets or SC-islets but had variable effects on GSIS in human islets depending on the donor. DP alone did not affect insulin content in mouse islets, human islets, or SC-islets, but mouse islets co-exposed to DP and glucolipotoxic (GLT) stress conditions (28.7 mM glucose + 0.5 mM palmitate) had reduced insulin content compared to control conditions. Co-exposure of mouse islets to DP + GLT amplified the upregulation of Slc30a8 compared to GLT alone. Our study highlights the importance and challenges of using different in vitro models for studying chemical toxicity.

Keywords: Dechlorane Plus; diabetes; islets; persistent organic pollutants; stem cells; β-cells.

Figure 1.

The metabolic effects of insystemicvivo DP exposure in male mice are not driven by changes in islet morphology or endocrine cell composition. Male mice were treated daily with vehicle or DP (10, 100, 1000 µg/kg per day) and fed chow or high fat diet (HFD) for 28 days; paraffin-embedded pancreas sections from these mice were analyzed by immunofluorescence staining to quantify islet morphology and endocrine cell composition. (a) Average islet size, (b) average % insulin (Ins)⁺ area, (c) average % glucagon (Gcg)⁺ area, and (d) average % proinsulin⁺ area per islet. (e) Representative images showing immunofluorescence staining of islets for insulin and glucagon, or insulin and proinsulin. Scale bars = 100 µm. All data are presented as mean ± SEM. Individual data points represent biological replicates (n = 3–8 biological replicates per condition). *p-value < .05 (two-way ANOVA with Tukey post-hoc).

Figure 2.

DP exposure does not affect insulin secretion but reduces insulin content in INS-1 832/3 cells. (a) Representative brightfield images of INS-1 832/3 cells treated with vehicle (DMSO) or DP (1, 10, 100 nM) for 48-hours in vitro. (b) Insulin secretion was quantified after a 1-hour static incubation in KRBB containing low glucose (LG; 2.8 mM) followed by 1-hour in high glucose (HG; 16.7 mM). (c) Stimulation index (HG:LG ratio). (d) Insulin content of lysed INS-1 832/3 cells. n = 6 technical replicates per biological replicate and n = 4 biological replicates per condition. Data represent mean ± SEM. Individual datapoints represent biological replicates. **p-value < .01 (one-way ANOVA with Tukey post-hoc; insulin content = non-parametric one-way ANOVA). Scale bars = 100 $\mathit{\mu }$m.

Figure 3.

DP has no effects on insulin secretion or insulin content in human SC-islets. (a) Representative images showing SC-islet morphology following 48-hour exposure exposed to vehicle (DMSO) or DP (1, 10 nM). SC-islets were derived from INS-2A-EGFP stem cell line so GFP+ cells represent insulin+ cells. Insulin secretion in response to (b) low glucose (LG; 2.8 mM), (c) high glucose (HG; 16.7 mM), and (d) 30 mM KCl was assessed 48-hours post-exposure. (e) Total insulin content was measured in lysed cells. Data are normalized to the vehicle control condition for each endpoint (n = 6 biological replicates, i.e. differentiations per condition). Data represent mean ± SEM. Individual datapoints represent biological replicates. *p-value < .05 (two-way ANOVA with Tukey post-hoc). Scale bars = 750 µm.

Figure 4.

Effects of DP on insulin secretion from human islets varies between donors. Human islets from donors with no diabetes (ND; n = 3) or Type 2 diabetes (T2D; n = 2) were exposed to vehicle (DMSO) or DP (1, 10 nM) for 48-hours ex vivo. See Supplemental Table S1 for human donor characteristics. (a, d, g, j, m) Insulin secretion was quantified following a 1-hour static incubation in KRBB containing low glucose (LG; 2.8 mM) followed by high glucose (HG; 16.7 mM). (b, e, h, k, n) Stimulation index (HG:LG ratio). (c, f, i, l, o) Insulin content of islet lysates. All data represent the mean ± SEM. Individual datapoints represent technical replicates from a single donor. *p-value < .05, **p < .01 (GSIS: two-way ANOVA with Tukey post-hoc; stimulation index and insulin content: one-way ANOVA with Tukey post-hoc).

Figure 5.

DP alone does not alter insulin secretion or insulin content in mouse islets, but DP co-exposure with GLT reducesstress decreases insulin content. (a-e) Mouse islets were exposed to vehicle (DMSO) or DP (1 nM) for 48-hours and dynamic insulin secretion was assessed in response to LG (2.8 mM), HG (16.7 mM), and KCl (30 mM). (f-j) Mouse islets were exposed to vehicle (DMSO), glucolipotoxic conditions (GLT: 28.7 mM glucose + 0.5 mM palmitate), or GLT conditions + 1 nM DP for 48-hours. (a, f) Dynamic insulin secretion curves from the perifusion assay and (b-d,g-i) area under the curve (AUC) for different periods of the perifusion. (e, j) Insulin content was measured in islets that underwent the same treatment conditions. All data represent mean ± SEM. Individual datapoints represent biological replicates (i.e. islets from different mice; n = 4-6 per group). *p < .05 (Perifusion curves: two-way ANOVA with Tukey post-hoc; stimulation index, AUCs: One-way ANOVA with Tukey post-hoc, Insulin content: non-parametric one-way ANOVA).

Figure 6.

DP amplifies the effect of GLT on Slc30a8. Mouse islets were exposed to vehicle (DMSO), 1 nM DP alone, glucolipotoxic conditions alone (GLT: 28.7 mM glucose +0.5 mM palmitate), or combined GLT +1 nM DP for 48-hours. Gene expression for markers of β-cell maturity and function were analyzed by qPCR. All data represents mean ± SEM. Individual datapoints represent biological replicates (i.e. islets from different mice; n = 4-6 per group). *p < .05, **p < .01, ***p < .001 (two-way ANOVA with Tukey post-hoc).