Cisplatin Exposure Dysregulates Insulin Secretion in Male and Female Mice

Numerous studies have shown that cancer survivors, irrespective of age, treatment, and type of cancer, have an increased risk of developing new-onset type 2 diabetes (T2D) compared with the general population several months after the end of treatment (1–3). Cancer survivors also develop T2D earlier than control populations (4,5). Importantly, Sylow et al. (6) reported that out of 28,000 cancer survivors, those with new-onset T2D had a 21% higher all-cause mortality compared to cancer survivors without T2D.

Cisplatin is a platinum-based chemotherapeutic agent used to treat a wide variety of cancers (7). Cisplatin can enter cells through passive diffusion or the copper transporter 1 protein (8), which has not been well characterized in islets. Once cisplatin enters the cell, either one or both chlorine groups are replaced by water, effectively becoming aquated and biologically active (9). Aquated cisplatin is highly electrophilic and binds to nucleophilic centers on purine residues in nuclear and mitochondrial DNA, causing the cross-linking of DNA (9). Cisplatin also increases the production of reactive oxygen species (ROS) and leads to oxidative stress and mitochondrial deterioration (9,10); this, in turn, leads to increased cell senescence and cell death (9).

Unfortunately, the cytotoxic effects of cisplatin are not limited to cancerous cells. Cisplatin treatment has been linked to acute toxicities and severe off-target effects, including nephrotoxicity and neurotoxicity (11). Its effects on pancreatic islets have not been well characterized, despite compelling evidence linking cisplatin treatment with metabolic complications. Patients with testicular cancer treated with cisplatin have increased odds of developing metabolic syndrome compared with patients treated with surgery or radiotherapy and the general population, despite no major differences in BMI (12). In a group of 219 patients who received cisplatin-based chemotherapy for head and neck cancer, 5% developed diabetes during their treatment period (13), which was almost twice the global rate of diabetes prevalence at the time of the study (14).

The potential off-target effects of cisplatin on pancreatic islets are currently unknown. Given the established roles of DNA damage, mitochondrial dysfunction, and oxidative stress in β-cell dysfunction during T2D pathogenesis (15,16), we hypothesized that β-cells would be particularly susceptible to the off-target effects of cisplatin. Moreover, we predicted that the limited regenerative capacity of β-cells (17) would prevent recovery of the β-cell population following injury, contributing to an increased risk of T2D in cancer survivors who receive cisplatin treatment. In our study, we assessed the effects of cisplatin on islet function systemically by exposing male and female mice to cisplatin in vivo and directly by treating isolated male and female mouse islets, as well as female human donor islets, with cisplatin in vitro. We found that in vivo cisplatin exposure decreased plasma insulin levels in both male and female mice and in vitro cisplatin exposure profoundly dysregulated insulin secretion in both isolated mouse and human donor islets.

Male and female C57BL/6N mice (Charles River Laboratories), ∼16 weeks old, were maintained on a 12-h light/dark cycle with ad libitum access to a standard rodent chow diet (Teklad Diet #2018; Harlan Laboratories) and water. All experiments were approved by the Carleton University Animal Care Committee and performed in accordance with Canadian Council on Animal Care guidelines. Prior to starting experimental protocols, animals were tracked for 2 weeks before being randomly assigned to treatment groups, ensuring body weight and fasting blood glucose levels were consistent between groups.

At ∼18 weeks of age, mice received i.p. injections of 0.9% saline (vehicle control) or 2 mg/kg cisplatin (#232120-50MG; Sigma-Aldrich) every other day over the course of 14 days, for a total of seven injections and a cumulative dose of 14 mg/kg. A clinical dose of 50 mg/m² cisplatin in patients is equivalent to ∼14 mg/kg cisplatin in mice based on an animal equivalent dosing calculation that accounts for metabolic differences between species (18). Similar dosing protocols in rodents have shown cisplatin-induced toxicity with <20% body weight loss (19,20). Mice were allowed to recover for 1 week after the end of the exposure period before in vivo metabolic assessments were conducted. Mice were euthanized 2 weeks after cisplatin or vehicle exposure.

To determine whether the effects of cisplatin were reproducible in different strains of mice, male SCID-beige and DBA/2 mice were also exposed to cisplatin in vivo as described above, and metabolic assessments were performed. Details of the study protocol are described in the Supplementary Methods.

All metabolic analyses were performed in conscious, restrained mice, as previously described (21). Briefly, body weight and blood glucose were measured following a 4-h morning fast twice a week throughout the study. An insulin tolerance test (ITT) was conducted at 1-week postexposure and an i.p. glucose tolerance test (ipGTT) was performed the following week. Blood samples were collected at 0, 15, 30, and 60 min during the ipGTT for measuring plasma insulin levels by ultrasensitive rodent insulin ELISA (#80-INSMSU-E10; ALPCO). All blood glucose measurements were conducted using a handheld glucometer (Medisure).

Mice were euthanized by isofluorane overdose followed by cervical dislocation, and pancreata were inflated via common bile duct injection with collagenase (1,000 IU/mL, #C7657; Sigma-Aldrich) dissolved in Hanks’ balanced salt solution (HBSS) (137 mmol/L NaCl, 5.4 mmol/L KCl, 4.2 mmol/L NaH₂PO₄, 4.1 mmol/L KH₂PO₄, 10 mmol/L HEPES, 1 mmol/L MgCl₂, 5 mmol/L anhydrous dextrose, pH 7.2). Inflated pancreatic tissues were excised and incubated at 37°C for 10.5 min then vigorously agitated, after which the collagenase reaction was stopped by adding cold HBSS with 1 mmol/L CaCl₂. The digested pancreatic tissues were washed three times in HBSS with CaCl₂, and islets were separated from exocrine tissue via a Histopaque gradient (#10771; Sigma-Aldrich). Islets were then filtered through a 70-μm cell strainer, resuspended in RPMI medium (#11875093 [Gibco] or #350-000-CL [Wisent Bioproducts]) supplemented with 10% (v/v) FBS (#F1051-500ML; Sigma-Aldrich) and 1% (v/v) penicillin-streptomycin (#30-002-Cl [Corning] or #15140-122-100 [Gibco]) and handpicked under a dissecting scope to >95% purity.

Islets from three female human donors (ages 52, 57, and 69 years) were isolated at the Alberta Diabetes Institute IsletCore. Islet isolation was approved by the Human Research Ethics Board at the University of Alberta (Pro00013094). All donors’ families gave informed consent for the use of pancreatic tissue in research. Islets were shipped overnight in Connaught Medical Research Laboratories 1066 medium (#15110CV; Corning) supplemented with 0.5% (v/v) BSA (#BAL62; Equitech Bio Inc.), 1% (v/v) insulin-transferrin-selenium (#25800CR; Corning), 0.5% (v/v) penicillin-streptomycin (#09-757F; Lonza), and 2% (v/v) Gibco GlutaMAX supplement (#35050061; Thermo Fisher Scientific). Upon arrival, islets were handpicked into DMEM (#11-885-084; Thermo Fisher Scientific) supplemented with 10% FBS (Sigma-Aldrich) and 1% penicillin-streptomycin (Gibco) and allowed to rest overnight at 37°C with 5% CO₂ before cisplatin or vehicle exposure. See Supplementary Table 1 for donor characteristics.

The peak intact cisplatin concentration in patients after receiving one dose of 50–100 mg/m² cisplatin is ∼10 μmol/L (22). This, coupled with a review of cisplatin concentrations typically used in cell line studies (23,24), prompted us to use 10 μmol/L for our in vitro studies with isolated islets.

For all in vitro experiments, islets were isolated from male and female C57BL/6 mice on a mixed J/N background, as described above. For all end points except patch-clamp recordings, intact mouse and human islets were incubated in complete RPMI medium or DMEM, respectively, overnight at 37°C with 5% CO₂ before being handpicked for in vitro cisplatin or vehicle exposure. One-half of the islets from each biological replicate was transferred to complete RPMI medium or DMEM containing either 10 μmol/L cisplatin (Sigma-Aldrich) or 0.9% saline (vehicle control) and cultured between 6 and 48 h. Media were refreshed after 24 h. For patch-clamp recordings, intact mouse islets were dispersed with enzyme-free Hanks’-based cell dissociation buffer (#13150-016; Gibco) immediately after isolation, plated in a tissue culture–treated dish, and then incubated overnight at 37°C with 5% CO₂ in complete RPMI medium prior to a 4-h exposure to 10 μmol/L cisplatin or vehicle.

To assess static glucose-stimulated insulin secretion (GSIS), 25 islets/replicate were handpicked and underwent sequential 1-h incubations in Krebs-ringer bicarbonate HEPES buffer (KRBH) containing 2.8 mmol/L glucose (low glucose [LG]), KRBH with 16.7 mmol/L glucose (high glucose [HG]), and KRBH with 30 mmol/L KCl as previously described (21).

To assess GSIS dynamically, 70 islets/replicate were washed with prewarmed (37°C) PBS (#D8662; Sigma-Aldrich) and then loaded in Perspex microcolumns between two layers of acrylamide-based microbeads (#PERI-BEADS-20; Biorep Technologies). Islets were perifused for 40 min with prewarmed LG KRBH to equilibrate the islets. The islets were then perifused with LG KRBH for 15 min, HG KRBH for 45 min, LG KRBH for 25 min, KCl KRBH for 35 min, and LG KRBH for 25 min. Samples were collected every 5 min at a flow rate of 40 μL/min except for the first 15 min after perifusion with HG and KCl KRBH, during which samples were collected every 2.5 min at a rate of 80 μL/min. Islets and solutions were maintained at 37°C throughout the perifusion, while the collection plate was kept at 4°C using a built-in tray cooling system. Samples were stored at −80°C until analysis. Insulin concentrations of all static GSIS and perifusion samples collected from mouse islets were measured by rodent insulin chemiluminescence ELISA (#80-INSMR-CH10; ALPCO).

Insulin concentrations of perifusion samples collected from islets of donor R542 were measured by human insulin chemiluminescence ELISA (#80-INSHU-CH10; ALPCO), while those from donors R551 and R552 were measured by radioimmunoassay (#HI-14K; Millipore). Human insulin concentrations are reported as fold change over basal (5–15 min) insulin secretion.

Islet respiration was quantified using the Seahorse XFe24 Analyzer (Agilent Technologies). A total of 70 mouse islets/replicate were handpicked and washed with Seahorse XF RPMI medium (103576, pH 7.4; Agilent Technologies) supplemented with 2 mmol/L sodium pyruvate, 2 mmol/L l-glutamine, and 1% (v/v) FBS containing either 2.8 or 16.7 mmol/L glucose. The islets were plated in a poly-d-lysine (#P7280; Sigma-Aldrich)–coated 24-well islet capture plate (#103518-100; Agilent Technologies) and incubated at 37°C without CO₂ for 1.5 h. Media were refreshed before loading the plate into the XFe24 Analyzer. The basal oxygen consumption rate was measured for ∼35 min. For islets pre-incubated in RPMI medium with 16.7 mmol/L glucose, wells were exposed to sequential injections of 2.5 μmol/L oligomycin for eight cycles, 3 μmol/L carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP) for five cycles, and a combination of 3 μmol/L rotenone and 3 μmol/L antimycin A for six cycles. For islets pre-incubated in RPMI medium with 2.8 mmol/L glucose, wells were exposed to 16.7 mmol/L glucose for six cycles prior to the same series of injections described above. With each cycle, the solutions were mixed for 3 min, samples were allowed to rest for 2 min, and oxygen consumption was measured for 3 min. Mitochondrial function parameters were calculated as indicated in Supplementary Table 2.

After 6, 24, and 48 h of cisplatin or vehicle exposure, 25 mouse islets/replicate were handpicked, dispersed, and stained with 0.5 μmol/L Hoescht (#62249; Thermo Fisher Scientific), 1.25 μmol/L calcein (#L3224; Invitrogen), and 0.5 μg/mL propidium iodide (PI) (#P21493; Invitrogen) dyes to assess cell viability as previously described (21). Cells were imaged with an Axio Observer 7 microscope.

Exocytotic capacity of dispersed mouse β-cells was measured following a 4-h incubation with cisplatin or vehicle in vitro. Cells were washed with 2.8 mmol/L glucose RPMI medium, preincubated with 2.8 mmol/L glucose RPMI medium for 1 h, and then patch-clamped in bath solution (118 mmol/L NaCl, 20 mmol/L tetraethylammonium chloride, 5.6 mmol/L KCl, 1.2 MgCl₂, 2.6 mmol/L CaCl₂, 5 mmol/L HEPES, pH 7.4) containing either 2.8 or 16.7 mmol/L glucose. The patch pipettes were filled with an internal pipette solution (125 mmol/L CsOH, 125 mmol/L glutamate, 10 mmol/L CsCl, 10 mmol/L NaCl, 1 mmol/L MgCl₂, 0.05 mmol/L EGTA, 5 mmol/L HEPES, 3 mmol/L MgATP, 0.1 mmol/L cAMP, pH 7.1). Whole-cell capacitance was recorded using a sine + DC lock-in function of an EPC10 amplifier and PatchMaster software (HEKA Electronics) ∼1 min after obtaining a whole-cell configuration. Capacitance was measured in response to 10 depolarizations from −70 to 0 mV, each 500 ms long. Changes in capacitance were normalized to initial cell size.

After 6, 24, and 48 h of in vitro cisplatin or vehicle exposure, 70 mouse islets/replicate were handpicked and stored in buffer RLT with 1% β-mercaptoethanol. Similarly, 70 human islets/replicate were handpicked after 48 h of in vitro cisplatin or vehicle exposure and stored in buffer RLT with 1% β-mercaptoethanol. RNA was extracted using the RNeasy Micro Kit (#74004; QIAGEN) per the manufacturer’s instructions. DNase treatment was performed prior to cDNA synthesis with iScript gDNA Clear cDNA Synthesis Kit (#1725035; Bio-Rad). Following this, quantitative PCR (qPCR) was performed using SsoAdvanced Universal SYBR Green Supermix (#1725271; Bio-Rad) and run on a CFX384 (Bio-Rad). All targets were run alongside no reverse transcription and no cDNA template controls. Ppia/PPIA was used as the reference gene given its stable expression under both control and treatment conditions. Data were analyzed using the 2^−ΔΔCt method. Primer sequences are listed in Supplementary Table 3.

Following 48-h in vitro cisplatin or vehicle exposure, RNA libraries were prepared from RNA lysate isolated from 70 mouse islets/replicate and processed with the TempO-Seq Mouse Whole Transcriptome (version 1.1) Assay panel (BioSpyder Technologies) according to the manufacturer’s instructions. Individual libraries were pooled and purified using the NucleoSpin Gel and PCR Clean Up Kit (#740609; Macherey-Nagel) according to modified BioSpyder instructions. The pooled library was sequenced using the Illumina NextSeq 2000 with a P1 Reagent Kit and 100-cycle high-throughput flow cell (Illumina).

Data were processed using the Omics data analysis framework for regulatory application (R-ODAF; https://github.com/R-ODAF/R-ODAF_Health_Canada) pipeline (25). Differentially expressed genes (DEGs) were subjected to overrepresentation analysis using clusterProfiler (version 4.12.0) (26), querying the gene ontology (GO) (27) and Kyoto Encyclopedia of Genes and Genomes (KEGG) (28) databases. Significance was set at an adjusted P < 0.01 and Q < 0.05. Full details on library preparation and data processing are provided in the Supplementary Methods.

Aside from TempO-Seq, all statistical analyses were conducted using GraphPad Prism 10.1.2 (GraphPad Software). Specific statistical tests and sample sizes are indicated in figure legends. For all analyses, P < 0.05 was considered statistically significant. Data are presented as mean ± SEM.

Sequencing data are available from the National Center for Biotechnology Information Gene Expression Omnibus database (accession no. GSE278504). All data that support the findings of this study are available from the corresponding author upon reasonable request.

We first assessed how cisplatin impacts systemic glucose homeostasis in male and female mice exposed to vehicle or cisplatin over 2 weeks (Fig. 1A). Cisplatin exposure did not affect body weight (Supplementary Fig. 1A and C), fasting blood glucose levels (Supplementary Fig. 1B and D), or insulin sensitivity at 1 week posttreatment (Fig. 1B and E) in either sex. During an ipGTT at 2 weeks postexposure, male mice exhibited no differences in glucose tolerance (Fig. 1C), but female cisplatin-exposed mice had a slightly lower peak glucose value than vehicle-exposed female mice (Fig. 1F). Both male and female cisplatin-exposed mice had reduced plasma insulin levels compared with controls during the ipGTT (Fig. 1D and G). The reduction of plasma insulin in cisplatin-exposed mice was reproduced in two separate cohorts using DBA/2 and immunocompromised SCID-beige male mice (Supplementary Fig. 2), which are relevant mouse models to metabolic research and patients with cancer.

Islets isolated 2 weeks postexposure showed no differences in basal insulin release, GSIS, stimulation index, or insulin content between treatment groups for either sex (Fig. 1H–J and L–N). However, islets from cisplatin-exposed male mice had reduced KCl-stimulated insulin secretion and reduced oxygen consumption in response to glucose stimulation compared with controls (Fig. 1H and K); this was not observed in islets from female mice (Fig. 1L and O). This suggests that low plasma insulin levels in cisplatin-exposed mice may originate from intrinsic impairments within the islets but are likely also influenced by defects in peripheral tissues.

To determine whether cisplatin directly impairs islet function, islets were isolated from male and female mice and exposed to 10 μmol/L cisplatin or vehicle for 48 h in vitro prior to functional analyses as outlined in Fig. 2A. Male and female mouse islets exposed directly to cisplatin showed profoundly impaired oxygen consumption (Fig. 2B and J) and insulin secretion (Fig. 2C and K) after 48 h. Specifically, cisplatin-exposed mouse islets had trending increases in basal insulin secretion at the beginning and end of the perifusion assay (Fig. 2C–F and K–N) and had pronounced reductions in both GSIS and KCl-stimulated insulin release (Fig. 2C, G–I, K, and O–Q) compared with vehicle-exposed islets. Interestingly, despite cisplatin-exposed islets having nearly abolished KCl-stimulated insulin secretion (Fig. 2C, I, K, and Q), there was a delayed insulin peak post-KCl perifusion in both sexes (Fig. 2C and K). Cisplatin appears to hyperstimulate insulin release under basal glucose conditions, reduce the sensitivity of islets to glucose, and impair the release of insulin independent of glucose metabolism. Notably, the effects of cisplatin on the in vitro metabolic function of mouse islets were not sex-specific.

Since the effects of cisplatin on β-cell function in vitro were similar between sexes, we focused on male mouse islets for additional analyses. We first determined whether our in vitro cisplatin dosing protocol affected islet cell viability to assess if the profound decreases in metabolic function and insulin secretion were attributed to cell death. Islets were treated with cisplatin or vehicle for 6, 24, or 48 h. At each timepoint, intact islets were imaged to visualize morphology and then dispersed into a single-cell suspension to measure cell viability via an image-based assay (Fig. 3A). Based on a qualitative assessment, islets from both treatment groups appeared to be healthy at all timepoints (Fig. 3B). Cisplatin exposure did not affect the percentage of PI⁺ (i.e., dead/dying) islet cells at any timepoint (Fig. 3C and D). These data indicate that exposure to 10 μmol/L cisplatin for 48 h was not overtly cytotoxic to mouse islets.

Given that we were not able to recover islets from the perifusion system to assess insulin content, we measured bulk insulin secretion and insulin content in mouse islets using a static GSIS assay after 48 h of cisplatin exposure in vitro (Fig. 4A). Cisplatin-exposed islets showed significantly elevated basal insulin secretion under LG conditions (Fig. 4B), a small decrease in stimulation index (Fig. 4C), and no change in total insulin content (Fig. 4D). The robust increase in basal insulin secretion in cisplatin-exposed islets aligned with the perifusion data (Fig. 2C), but the lack of difference under HG conditions in the static assay was surprising. This may be attributed to the longer period for second-phase insulin secretion during a static GSIS, which was not as severely impacted by cisplatin as first-phase insulin secretion (Fig. 2C and H). The similarity in insulin content between treatment groups confirms that cisplatin does not cause pronounced β-cell loss or depletion of insulin stores.

The altered insulin secretion in cisplatin-exposed islets could be attributed to changes in exocytotic capacity of the β-cells. To explore this, we measured the capacitance responses of male mouse β-cells by whole-cell patch-clamp. Under LG conditions, there was a clear increase in both the average and cumulative exocytotic capacity of cisplatin-exposed β-cells compared with vehicle-exposed β-cells (Fig. 4E–G). However, exocytotic capacity of β-cells from both treatment groups was comparable under HG conditions (Fig. 4H–J). These results suggest that cisplatin exposure increases exocytosis of insulin granules from β-cells in LG environments but not in HG environments. Thus, additional mechanisms are driving defective GSIS in cisplatin-exposed islets.

To better understand how cisplatin disrupts GSIS (Fig. 2C and K), we assessed mitochondrial function in vehicle- and cisplatin-exposed male mouse islets in greater detail. In this experiment, islets were pre-incubated in LG Seahorse XF RPMI medium to allow for assessment of oxygen consumption in response to HG. Cisplatin-exposed islets had significantly impaired oxygen consumption compared with vehicle-exposed islets throughout the assay and showed a significant or trending reduction in all calculated parameters (Fig. 5A–H), consistent with our previous experiment (Fig. 2B and J). Importantly, cisplatin-exposed islets did not robustly increase oxygen consumption when stimulated with HG (Fig. 5B). These data indicate that in vitro cisplatin exposure disrupts mitochondrial function in mouse islets, which likely contributes to the impaired GSIS (Fig. 2C and K).

We measured expression of key genes related to islet function and cell stress at 6, 24, and 48 h following vehicle or cisplatin exposure to better understand the temporal effects of cisplatin (Fig. 6). We first assessed expression of genes in the Bcl-2 family, key players in the intrinsic apoptosis pathway. Cisplatin-exposed islets showed a sustained, approximately fivefold downregulation of Bcl2 and a progressively increasing upregulation of Bcl2l1, both anti-apoptotic genes, compared with vehicle-exposed islets between 6 and 48 h (Fig. 6A and B). The pro-apoptotic gene Bax was upregulated approximately twofold in cisplatin-exposed islets at both 24 and 48 h (Fig. 6C). Additionally, Cdkn1a, which encodes for p21, a marker of DNA damage and senescence (29), was highly upregulated in cisplatin-exposed islets as early as 6 h postexposure (Fig. 6D). Together, these data suggest that cisplatin induces the intrinsic apoptosis pathway, but the anti-apoptotic gene Bcl2l1 is being activated to promote a pro-survival phenotype.

Ppargc1a, a transcriptional coactivator that regulates mitochondrial function (30,31), was modestly reduced in cisplatin-exposed islets at 6 h but not 24 or 48 h (Fig. 6E). Cisplatin-exposed islets had an approximately twofold upregulation of Nrf2, a marker of oxidative stress, at all three time points (Fig. 6F) and increased expression of Nrf2 downstream targets Gpx1 at 24 and 48 h and Hmox1 at 6 h (Fig. 6G and H). These changes imply that cisplatin acutely activates oxidative stress responses in mouse islets.

Cisplatin did not affect expression of insulin (Ins1, Ins2) or proprotein convertase genes (Pcsk1, Pcsk2) at 6 h but reduced expression of these genes at 24 and 48 h compared with vehicle (Fig. 6I–L). The most pronounced effect was on Pcsk2 expression, which was reduced approximately fourfold at 24 h and approximately eightfold at 48 h in cisplatin-exposed islets (Fig. 6L). These results suggest that cisplatin exposure impairs proinsulin production and processing.

To further explore how cisplatin exposure alters islet gene expression, we performed a TempO-Seq analysis on male mouse islets following 48 h of treatment (Fig. 7 and Supplementary Fig. 3). In total, 1,022 probes were significantly different between cisplatin- and vehicle-exposed islets (Supplementary Table 4). Pathway analysis of DEGs using both the GO biological process (GO-BP) and the KEGG pathway databases identified insulin secretion as the most significantly enriched gene set (Fig. 7A and Supplementary Fig. 3A). In fact, the 20 most significantly enriched gene sets in the GO-BP database were associated with protein/hormone transport and secretion (Fig. 7A). Most of the enriched genes within each of the 20 gene sets were downregulated, with >75% of genes downregulated in 18 out of 20 gene sets (Fig. 7B and Supplementary Table 5). Since insulin secretion was the most enriched term across both databases, we next used hierarchal clustering of DEGs to more closely examine the insulin secretion pathway (Fig. 7C and Supplementary Fig. 3B). Out of 41 DEGs (represented by 46 probes), cisplatin exposure upregulated 6 genes and downregulated 35 genes (Fig. 7B). Of the 35 downregulated DEGs, several were β-cell maturity and identity markers, including Ucn3, Nkx6.1, and Mafa (Fig. 7B). Genes involved in paracrine signaling and potentiation of insulin secretion, including Gcg, Glp1r, and Abcc8, were also downregulated (Fig. 7B and Supplementary Fig. 3B). When visualized, several aspects of the insulin secretion pathway, from initial stimuli reception to insulin granule release, were downregulated (Fig. 7D and Supplementary Fig. 3C). Overall, the transcriptomic analysis confirmed that insulin secretion is highly impacted by cisplatin exposure, supporting our previous findings.

To begin translating our results in mouse islets to human health, we performed a dynamic insulin secretion assay and measured the expression of key genes related to islet and β-cell function in human islets obtained from three female organ donors (Fig. 8). Both GSIS and KCl-stimulated insulin secretion were reduced in cisplatin-exposed human islets compared with vehicle-exposed islets (Fig. 8A). Cisplatin-exposed human islets also showed a profound downregulation of insulin (INS) and proprotein convertase genes (PCSK1 and PCSK2) after 48 h compared with vehicle-exposed islets (Fig. 8B–D). We also observed a consistent decrease in β-cell identity genes GCK, UCN3, and MAFA in cisplatin-exposed human islets (Fig. 8E–G). Interestingly, BCL2 expression was downregulated in cisplatin-exposed human islets from all three donors (Fig. 8H), while two of the donors showed increased expression of both BCL2L1 and CDKN1A in cisplatin-exposed islets (Fig. 8I–J), similar to what was observed in mouse islets (Fig. 6A, B, and D). Overall, the decrease in GSIS and KCl-stimulated insulin secretion and downregulation of insulin processing and β-cell identity gene expression are consistent with what was observed in mouse islets 48 h after cisplatin-exposure (Figs. 2 C and K, 6 I–L, and 7 C).

The increased incidence of T2D among cancer survivors after chemotherapy (2,5) prompted us to investigate whether cisplatin treatment elicits off-target effects on pancreatic islets. Our research shows that cisplatin exposure profoundly dysregulates insulin secretion and oxygen consumption in islets from both male and female mice and significantly alters the expression of genes critical to islet function. Importantly, the adverse effects of cisplatin on insulin secretion were seen after both systemic exposure in mice and direct exposure to islets in vitro. Moreover, the acute effects of cisplatin, which disrupt insulin secretion and reduce expression of genes related to insulin processing and β-cell identity, were replicated in human islets from three female organ donors. These data suggest that cisplatin-induced damage to pancreatic islets may contribute to the long-term risks of metabolic dysregulation in cancer survivors.

Mouse islets exposed to cisplatin in vitro showed increased insulin release under LG conditions but reduced insulin release following HG or KCl stimulus. As observed during a dynamic GSIS assay, both cisplatin-exposed mouse and human islets displayed a diminished first-phase insulin peak but were still capable of gradual second-phase insulin release when stimulated with HG. This likely explains why there was no difference in total insulin released by mouse islets after 1 h of HG stimulation in a static GSIS assay. Basal hyperinsulinemia and the loss of first-phase GSIS are predictive markers for T2D (32,33). Our mouse study confirmed that some of these β-cell defects translated in vivo as well. The significantly reduced plasma insulin levels in cisplatin-exposed mice during a GTT is a key marker of β-cell dysfunction (34). It was surprising to find that islets isolated from these mice only exhibited mild defects in GSIS ex vivo, but this could be related to the recovery period provided to mice after cisplatin administration in vivo. Future studies should assess how long defects in GSIS persist after a prolonged washout period following in vitro cisplatin exposure. Additionally, there may be an interplay of both intrinsic defects in the islets and extrinsic effects caused by disruption of peripheral tissues, influencing the changes in plasma insulin observed in vivo. Cisplatin-induced toxicity has been characterized in various organs, including the kidney and brain (11). However, the effects of cisplatin on glucoregulatory tissues have not been well characterized. Future studies should investigate liver and adipose phenotypes in parallel with β-cell function.

Given that there was no change in insulin content from mouse islets isolated after in vivo cisplatin exposure and no change in total insulin content or percentage of PI⁺ islet cells following in vitro cisplatin exposure, we speculate that cisplatin-induced impairments in insulin secretion are not driven by β-cell loss but, rather, through intrinsic defects within islet endocrine cells. For example, cisplatin exposure significantly downregulated the expression of Abcc8, the gene encoding for sulfonylurea receptor 1 (SUR1), a key component of the K_ATP channel in β-cells. Loss-of-function in this gene has been linked to malfunction of the K_ATP channel, which leads to accumulation of potassium ions in the β-cell and depolarization without external nutrient stimulation (35). This could explain the increased basal insulin secretion and exocytotic capacity of β-cells under LG conditions in cisplatin-exposed mouse islets. While no differences were noted in the exocytotic capacity of β-cells under HG conditions, the heightened exocytotic capacity of cisplatin-exposed β-cells in LG conditions could reduce the availability of insulin granules in the readily releasable pool, causing a delayed release of insulin upon stimulation while the pool replenishes itself (36). Further investigation into the granule docking machinery of β-cells and calcium signaling is required to better understand how cisplatin alters exocytotic capacity of β-cells.

Cisplatin is known to inhibit the replication/transcription of mitochondrial DNA (9), so we assessed mitochondrial function in cisplatin-exposed islets. Islets isolated from cisplatin-exposed male mice had modest defects in glucose-stimulated oxygen consumption at 2 weeks postexposure. Both male and female mouse islets exposed to cisplatin in vitro had substantially decreased basal oxygen consumption and lacked appropriate changes in oxygen consumption rates in response to glucose or electron transport chain modulators. Because the basal oxygen consumption of cisplatin-exposed mouse islets is near maximal respiration, these islets are likely unable to accommodate changes in energetic demand as effectively as their vehicle-exposed counterparts. The reduced mitochondrial ATP production in cisplatin-exposed mouse islets aligns with the observed downregulation of Gck, which encodes for glucokinase, a critical protein in the transformation of glucose to pyruvate. Glucokinase is considered the glucose sensor of the β-cell; individuals with mutations in GCK have monogenic diabetes and, in turn, require greater nutrient stimulation to trigger insulin secretion (37). The downregulation of Gck and GCK in cisplatin-exposed mouse and human islets, respectively, may explain the reduced insulin secretion observed upon HG stimulation in cisplatin-exposed islets. This, in combination with the inhibited oxygen consumption, likely contributes to dysregulated GSIS in cisplatin-exposed mouse islets.

The rapid upregulation of Nrf2, a master regulator of oxidative stress response pathways, and its downstream targets suggests that cisplatin exposure increases oxidative stress in mouse islets, much like in other tissues (9). β-Cells are particularly sensitive to oxidative stress, as they express relatively low levels of antioxidant enzymes compared with other cells (16). Moreover, the downregulation of Ppargc1a in cisplatin-exposed mouse islets may inhibit ROS detoxification (31), further contributing to cisplatin-induced oxidative stress in islets. Additional investigation is required to understand whether ROS accumulation is involved in cisplatin-induced β-cell dysfunction and to determine whether interventions with antioxidants could protect islets from the adverse effects of cisplatin.

Members of the Bcl-2 family are key regulators of β-cell fate (38). The robust decline in Bcl2 expression, along with the stark upregulation of Bax expression, suggests that cisplatin induces the intrinsic apoptosis pathway in mouse islets. However, we did not observe increased PI⁺ cells or reduced insulin content in cisplatin-exposed mouse islets, so we speculate that the upregulation of Bcl2l1 (and likely other anti-apoptotic factors) contributes to protecting cisplatin-exposed islets from cell death. Fiebig et al. (39) found that Bcl-x_L, the protein product of Bcl2l1, is more effective at preventing cell death than Bcl-2 in cells treated with etoposide, another chemotherapeutic agent. Other studies have shown that while upregulation of Bcl-x_L prevents cell death, it also impairs mitochondrial function, oxygen consumption, and insulin secretion (40,41); thus, the upregulation of Bcl2l1 may contribute to the dysregulation in oxygen consumption and insulin release in cisplatin-exposed mouse islets. Interestingly, decreased Bcl2 expression has also been correlated with elevated levels of basal insulin secretion (42), which aligns with our findings. Taken together, our data suggest that cisplatin exposure may promote a prosurvival phenotype in islets, which could shift cell fate toward senescence rather than apoptosis. This notion is further supported by the stark upregulation of Cdkn1a observed in mouse islets as early as 6 h postexposure, a characteristic noted in islets from donors with T2D (43). In human donor islets, we observed similar trends in the expression of in BCL2, BCL2l1, and CDKN1A following cisplatin exposure. However, the limited sample size of human donors prevented us from fully characterizing the responses to cisplatin. A much higher number of human donors is required to fully delve into the effects of cisplatin on human islets to account for the biological variability between donors. Studies using senolytics and small molecule agonists/antagonists of the Bcl-2 family will provide additional insight into the role of senescence in cisplatin-induced β-cell dysfunction.

The enrichment of insulin secretion in both the GO-BP and KEGG databases, robust downregulation of genes in this pathway, and reduced expression of Ins1, Ins2, Pcsk1, and Pcks2 over time highlight the detrimental effects of cisplatin on β-cell function. This is further supported by the pronounced and consistent downregulation of INS, PCSK1, and PCSK2 in cisplatin-exposed human islets from all three donors. Interestingly, cisplatin-exposed mouse and human islets showed downregulation of Ucn3/UCN3 and Mafa/MAFA, genes critical to β-cell identity. The loss of these genes is linked to β-cell dedifferentiation and contributes to β-cell failure in T2D (44). Compromised β-cell identity ultimately leads to downstream effects in insulin processing and signaling (45). Future studies should prioritize the collection of intact fixed pancreatic tissue to characterize the effects of in vivo cisplatin exposure on key β-cell identity markers at the protein level. Furthermore, downregulation of Gcg and Glp1r in cisplatin-exposed mouse islets could point to defects in paracrine signaling between endocrine cells. Deficits in paracrine signaling within the islet contribute to impaired insulin secretion in islets from donors with T2D compared with donors without diabetes (46). Future studies exploring the effects of cisplatin on other pancreatic endocrine cells will shed light on how paracrine signaling could be contributing to cisplatin-induced defects in insulin secretion.

Incidences of new-onset metabolic syndrome have been reported in a wide age range of adult cancer survivors between 18 and 50 years old (47). In our experiments, mice were exposed to cisplatin between 10 and 18 weeks of age, which would be comparable to young adults (48). We chose to focus on relatively young and healthy mice to assess the effect of cisplatin on islet function without the confounding influence of age. However, it will be important to elucidate the effects of cisplatin on an aged cohort of mice to better understand how age may influence susceptibility to cisplatin-induced islet dysfunction.

The development of new-onset diabetes in cancer survivors is well documented (2,5), but the cause of this dysglycemia remains unclear. Our research strongly indicates that cisplatin exposure causes acute defects in islet function, highlighting an urgent need for further investigation. Our initial experiments in human donor islets suggest that our findings in mouse islets are applicable to human health. Longer-term mouse studies are needed to better understand the chronic impacts of this insulin dysregulation following chemotherapy and elucidate mechanisms by which cisplatin, as well as other chemotherapeutic drugs, affect islet function. While our initial findings in human donor islets are compelling, it is crucial to expand upon these findings with a larger sample size to account for the natural heterogeneity in islet function observed in humans. Additionally, clinical studies examining β-cell function in cancer survivors will be vital in translating our results to patient outcomes. Ultimately, this research will provide critical insight for designing targeted interventions to improve long-term metabolic health outcomes in cancer survivors and reduce their risk of developing T2D after treatment.

Acknowledgments. Human islets for research were provided by the Alberta Diabetes Institute IsletCore at the University of Alberta in Edmonton (https://www.bcell.org/adi-isletcore.html) with the assistance of the Human Organ Procurement and Exchange (HOPE) program, Trillium Gift of Life Network (TGLN), and other Canadian organ procurement organizations. The authors are extremely grateful to Dr. Bruce McKay (Carleton University) for mentorship and guidance for this project. The authors also sincerely thank Andrea Smith, Kayleigh Rick, and Emilia Poleo-Giordani, all from Carleton University, for their early contributions to this project.

Funding. This research was supported by Diabetes Canada grant OG-3-22-5610-JB (to J.E.B.), Canadian Institute of Health Research (CIHR) grant 148451 (to P.E.M.), and a research grant funded by CIHR, Breakthrough T1D Canada, and Diabetes Canada (5-SRA-2021-1149-S-B/TG 179092 to P.E.M. and J.E.B.). L.B. was supported by a CIHR CGS-D award and a Natural Sciences and Engineering Research Council (NSERC) CREATE award on behalf of the Canadian Islet Research Training Network (CIRTN-R2FIC). L.G. was supported by master’s and doctoral training scholarships from Fonds de recherche du Québec–Santé. M.E.A.C. was supported by NSERC CGS-M and NSERC CGS-D awards. J.P. was supported by the Guiding Interdisciplinary Research on Women’s and Girls’ Health and Wellbeing (GROWW) scholarship and an Ontario Graduate Scholarship (OGS). E.P.v.Z. was supported by an NSERC-CREATE PDF award on behalf of CIRTN-R2FIC and a CIHR fellowship. M.P.H. was supported by an OGS and CIHR CGS-D award. K.A.v.A. was supported by an NSERC-CREATE award on behalf of CIRTN-R2FIC. E.E.M. is supported by Diabetes Canada grant OG-3-21-5591-EM. C.L.Y. has support from Canadian Foundation for Innovation grant 233109 and Canada Research Chairs Program grant CRC-202-00060. J.A.M. was supported by NSERC discovery grant RGPIN-2024-04456. P.E.M. holds a Canada Research Chair in Islet Biology. J.E.B. was supported by an Ontario Early Researcher Award.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. L.B., L.G., M.E.A.C., J.D.H.S., A.A.H., J.P., E.P.v.Z., M.P.H., K.S.M., K.A.v.A., H.L., X.D., A.B., E.F., E.E.M., P.E.M., and J.E.B. were involved with data acquisition and analysis. L.B., L.G., and J.E.B. conceived the experimental design. L.B. and J.E.B. wrote the manuscript. C.L.Y. and J.A.M. were involved with data acquisition. All authors contributed to manuscript revisions and approved the final version of the article. J.E.B. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.