Twenty-four male Wistar rats (8 - 10 weeks old, 250 - 300 g) were obtained from the breeding facility of the Faculty of Science and Technology, Marien Ngouabi University (Congo). Animals were housed under controlled conditions (22˚C ± 2˚C; 55% ± 10% humidity; 12 h light/dark cycle) with free access to standard chow and water.

Diabetes was induced by a single intraperitoneal injection of streptozotocin (STZ, 50 mg/kg body weight), a protocol that predominantly causes direct β-cell destruction and insulin-deficient hyperglycemia rather than the peripheral insulin-resistant phenotype that defines type 2 diabetes; no insulin-resistance testing (e.g., insulin tolerance test, HOMA-IR) was performed. Animals are therefore referred to throughout this manuscript as STZ-diabetic rats, and not as a type 2 diabetes model. Persistent fasting hyperglycemia > 250 mg/dL was used as the diagnostic criterion for successful induction.

After anesthesia with isoflurane (5% induction, 2% maintenance), animals were sacrificed by decapitation. Pancreata were rapidly removed and digested with collagenase P (0.5 mg/mL, 15 - 20 min, 37˚C). Islets were purified by differential centrifugation on a Ficoll gradient, washed in HBSS, and cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum, penicillin (100 U/mL), and streptomycin (100 µg/mL).

For each rat, approximately 10 islets were isolated. Twenty-four rats were allocated in three equal cohorts of eight to the three sub-experiments described in Section 2.3 (glucose dose-response, Sections 3.1 - 3.3; inhibitor dose-ranging, Section 3.4; and matched-control comparison, Section 3.5). Within each sub-experiment, islets from a given rat were distributed across the treatment conditions of that sub-experiment, rather than assigned in full to a single condition; the rat, and not the individual islet, was therefore treated as the independent biological unit for statistical analysis (Section 2.6), with islet-level measurements averaged within each rat/condition prior to group-level comparison. Islets were incubated for 24 h at 37˚C (5% CO₂, saturated humidity) to allow recovery before experimental stimulation.

Islets were exposed to different glucose concentrations representing normoglycemia (5.5 mM), moderate hyperglycemia (15 mM), and severe hyperglycemia (25 mM). To assess the effect of pharmacological NOX2 perturbation, islets were pretreated for 30 min with apocynin or diphenyleneiodonium (DPI) at 5, 15, or 25 µM (final DMSO ≤ 0.1% v/v). A vehicle-only control (DMSO ≤ 0.1% v/v, no apocynin/DPI) was included at each glucose concentration. Eight rats were allocated to this sub-experiment; islets pooled from these animals were processed into three independent biological islet preparations per condition, consistent with the biological triplicate design described in Section 2.6.

A second, independent experiment was carried out to test protection against glucotoxicity directly, rather than dose-response alone. Islets were exposed to normoglycemic (5.5 mM), moderately hyperglycemic (15 mM), or severely hyperglycemic (25 mM) medium and randomly assigned to one of four arms: an untreated, glucose-matched, vehicle-only (DMSO ≤ 0.1% v/v) control, apocynin (15 µM in vehicle), DPI (15 µM in vehicle), or apocynin combined with DPI (15 µM each in vehicle), with a 30 min inhibitor pretreatment as in the paragraph above. Islets were then cultured under these conditions for 48 h, and NOX2 and caspase-3 were measured in islet lysates (protein-normalized), and insulin in the culture supernatant, all collected at that endpoint (48 h); reported insulin values represent accumulated secretion over the 48 h culture period rather than an acute glucose-stimulated insulin secretion (GSIS) challenge. Including an inhibitor-free control at every glucose concentration made it possible to evaluate pharmacological NOX2 perturbation against untreated glucotoxicity, rather than only across increasing inhibitor doses. Eight rats were allocated to this sub-experiment, with islets pooled from all eight animals and processed into three independent biological islet preparations per condition (Section 2.6).

2.4.1. NOX2 Quantification

NOX2 levels were measured using a specific sandwich ELISA kit (MyBiosource, San Diego, CA, USA), according to the manufacturer’s protocol. Results were expressed as ng/mg total protein.

2.4.2. Caspase-3 Quantification

Caspase-3 levels were quantified with a commercial ELISA kit (SunLong Biotech, Hangzhou, China; ref. SL0152Ra), in the same islet lysates as NOX2. Concentrations were calculated from a standard curve and corrected for protein content (Bradford assay), and expressed as ng/mg protein.

2.4.3. Insulin Secretion

Culture supernatants were collected at 1 h, 24 h, and 48 h post-stimulation. Insulin concentrations were determined by rat-specific ELISA (Mercodia, Uppsala, Sweden) and expressed as ng/mg protein. Unless otherwise stated, values reported in Section 3 correspond to the 48 h collection point and represent insulin accumulated in the medium over the culture period under sustained glucose exposure, rather than an acute glucose-stimulated insulin secretion (GSIS) response to a step change in glucose concentration.

All experimental procedures involving animals were conducted in accordance with the Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines and institutional regulations on animal care and use, and were approved by the Health Ethics Committee under approval number 053-25/MESRSIT/DGRST/CERSSA/-25.

All experiments were performed in biological triplicates with technical duplicates and analyzed using GraphPad Prism 5 software (GraphPad Software, Inc., CA, USA). Data are expressed as mean ± standard deviation (SD). Group comparisons were performed using Student’s t-test for pairwise comparisons, or one-way ANOVA followed by Tukey’s post-hoc test for multiple comparisons. Non-normally distributed data were analyzed with the Mann-Whitney or Kruskal-Wallis test, as appropriate. Correlations were assessed using Spearman’s coefficient. Differences were considered statistically significant at p < 0.05.

Sample size was determined from preliminary data and a power analysis (α = 0.05, power (1 − β) = 0.80, Cohen’s d ≈ 0.6), which justified the inclusion of 8 animals per group.

Because islets from a given rat were distributed across treatment conditions within a sub-experiment, islet-level measurements are not statistically independent within a rat. The rat was therefore treated as the biological experimental unit for all comparisons: islet-level values were averaged within each rat and condition prior to group-level analysis. The 24 rats were allocated in equal cohorts of eight to each of the three sub-experiments (Sections 3.1 - 3.3, 3.4, and 3.5). For the glucose dose-response experiment (Sections 3.1 - 3.3), the n reported in each figure legend refers directly to these 8 rats. For the two inhibitor sub-experiments (Sections 3.4 and 3.5), islets pooled from the corresponding 8-rat cohort were processed into three independent biological islet preparations per condition, consistent with the biological triplicate design noted above; this is why the figure legends for those sections report n = 3 biological islet preparations per condition, each preparation drawing on the shared 8-rat cohort.

ELISA analysis revealed a significant increase in caspase-3 concentration proportional to glucose concentration. Levels remained low under normoglycemic conditions (5.5 mM), whereas severe hyperglycemia (25 mM) induced a marked increase in protein expression, confirming a direct glucotoxic impact on pancreatic β-cells (Figure 1).

Figure 1. Caspase-3 concentration in pancreatic islets (n = 8 rats, 80 islets per condition, with islets from each rat distributed across the three glucose conditions). Results are expressed as mean ± standard deviation.

ELISA measurements showed that NOX2 concentration increased in parallel with rising glucose concentrations, peaking at 25 mM and reflecting the heightened oxidative stress experienced by islets under hyperglycemia (Figure 2).

Exposure of islets to hyperglycemia resulted in a marked decrease in insulin secretion. Secretion remained effective under physiological glucose conditions but progressively collapsed under severe glucotoxic stress, illustrating the functional decline of β-cells (Figure 3).

Figure 2. NOX2 concentration (ng/mg total protein) measured by ELISA in islet lysates. Means ± SD, n = 8 rats (80 islets per condition).

Figure 3. Effect of glucose variation on accumulated insulin secretion at 48 h. Means ± SD, n = 8 rats/condition (80 islets per condition).

To characterize the dose-response relationship of pharmacological NOX2 perturbation, islets under severe hyperglycemia (25 mM) were pretreated with apocynin or DPI across a 5 - 25 µM range. Both agents lowered NOX2 concentration in a clearly dose-dependent manner (Figure 4(c), Figure 4(f)). This reduction did not, however, translate into a parallel benefit downstream. Insulin secretion declined progressively as the concentration of either agent increased (Figure 4(b), Figure 4(e)), and caspase-3 concentration actually rose across the DPI series (Figure 4(a)), with only a modest decline for apocynin (Figure 4(d)). NOX2 was being suppressed, yet islet function was not improving, and for DPI the apoptotic

Figure 4. (a) Effect of DPI inhibitor on caspase-3 concentration. (b) Effect of DPI inhibitor on insulin secretion. (c) Effect of DPI inhibitor on NOX2 concentration. (d) Effect of apocynin inhibitor on caspase-3 concentration. (e) Effect of apocynin inhibitor on insulin secretion. (f) Effect of apocynin inhibitor on NOX2 concentration. Means ± SD, n = 3 biological islet preparations/condition, drawn from a cohort of 8 rats (Section 2.6). Note: this experiment compares inhibitor doses under fixed severe hyperglycemia (25 mM) and has no untreated, glucose-matched control; Figure 5 provides that comparison.

signal appeared to worsen at the highest concentration tested. A plausible explanation is a biphasic, dose-dependent response, in which off-target or cytotoxic effects of these agents at higher concentrations offset or outweigh the benefit of NOX2 inhibition itself; this would be consistent with reports that DPI broadly inhibits other flavoproteins beyond NOX2, and that apocynin can act non-specifically above the low micromolar range [20][21]. It should be noted that this dose-ranging experiment did not include an untreated, glucose-matched control, so on their own these data cannot establish whether any of the concentrations tested conferred net protection relative to unopposed glucotoxicity. They do, however, define a dose range in which NOX2 suppression alone is a poor predictor of islet outcome, which is what motivated the controlled comparison described in Section 3.5.

To test directly whether NOX2 perturbation protects islets against glucotoxicity, we ran a second, independent experiment comparing apocynin (15 µM) and DPI (15 µM), alone or combined, with an untreated, glucose-matched, vehicle-only control at each glucose concentration (5.5, 15, and 25 mM; n = 3 biological islet preparations per condition, drawn from a cohort of 8 rats (Section 2.6)), analyzed by one-way ANOVA with Tukey’s post-hoc test at each glucose level, as described in Section 2.6. All three treatment arms showed lower NOX2 concentration and lower caspase-3 concentration than the untreated control at every glucose level, together with higher insulin secretion. Of the 27 pairwise comparisons against the untreated control (3 treatment arms × 3 outcome measures × 3 glucose concentrations), 26 reached statistical significance (Tukey-adjusted p < 0.05, and p < 0.001 for most); the sole exception was apocynin versus control for caspase-3 concentration at 15 mM glucose (p = 0.14). Combining apocynin and DPI gave numerically the strongest protection at each glucose concentration, although no formal interaction analysis was performed; this combined effect should therefore be described as more pronounced rather than as a formally demonstrated additive effect (Figures 5(a)-(c)). At 25 mM glucose, for instance, insulin secretion rose from 20.0 ± 0.2 ng/mg protein in untreated islets to 26.0 ± 0.2 ng/mg protein with combined treatment (Tukey-adjusted p < 0.001), while caspase-3 concentration fell from 6.1 ± 0.1 to 5.1 ± 0.1 ng/mg protein over the same comparison (p < 0.001).

Because this experiment includes the glucose-matched control missing from the dose-ranging study in Section 3.4, it provides more direct support for a genuine protective effect of NOX2 perturbation, at least at a defined, moderate

Figure 5. Effect of apocynin (15 µM), DPI (15 µM), and their combination on (a) NOX2 concentration, (b) insulin secretion, and (c) caspase-3 concentration, relative to untreated, glucose-matched, vehicle-only controls, at increasing glucose concentrations (5.5, 15, 25 mM). Means ± SD, n = 3 biological islet preparations/condition, drawn from a cohort of 8 rats (Section 2.6).

concentration (15 µM) of either agent. It also suggests that this protection does not necessarily keep improving as the dose is increased further.

Correlation analysis demonstrated a robust positive relationship between NOX2 concentration and caspase-3 concentration (r = 0.93; p = 0.0004), underscoring the pivotal role of oxidative stress in driving apoptotic signaling in β-cells (Figure 6(a)). This correlation was calculated using individual rat-level values (averaged per rat and condition, as described in Section 2.6) across the 8 rats and 3 glucose conditions of the dose-response experiment (n = 24 paired observations; Sections 3.1 - 3.3), rather than from only 3 condition-level means, which would be too few points to support the reported p-value. Even so, because all 8 rats contributed data at all 3 glucose levels, observations are not fully independent of one another, and the correlation should be interpreted as an association within this dose-response design rather than as a fully independent-sample effect-size estimate.

Conversely, caspase-3 concentration was strongly and inversely correlated with insulin secretion (r = −0.95; p = 0.0002) (calculated on the same n = 24 rat-level, per-condition observations described above), a pattern consistent with, but not proof of, a contribution of apoptosis to secretory loss. Because glucose exposure can independently drive both increased caspase-3 concentration and decreased insulin secretion, this association is best described as a treatment-associated relationship rather than as evidence that caspase-3-driven apoptosis is the principal cause of reduced insulin output (Figure 6(b)).

Figure 6. (a) Positive correlation between NOX2 concentration and caspase-3 concentration. (b) Inverse correlation between caspase-3 concentration and insulin secretion.