Autocrine C‐peptide protects INS1 β cells against palmitic acid‐induced oxidative stress in peroxisomes by inducing catalase

Abstract Aims C‐peptide, produced by pancreatic β cells and co‐secreted in the bloodstream with insulin, has antioxidant properties in glucose‐ and hydrogen peroxide (H2O2)‐exposed INS1 β cells. Palmitic acid, the most physiologically abundant long‐chain free fatty acid in humans, is metabolized in peroxisomes of β cells accumulating H2O2 that can lead to oxidative stress. Here, we tested the hypothesis that C‐peptide protects β cells from palmitic acid‐induced stress by lowering peroxisomal H2O2. Materials and methods We exposed INS1 β cells to palmitic acid and C‐peptide in the setting of increasing glucose concentration and tested for changes in parameters of stress and death. To study the ability of C‐peptide to lower peroxisomal H2O2, we engineered an INS1 β cell line stably expressing the peroxisomal‐targeted H2O2 sensor HyPer, whose fluorescence increases with cellular H2O2. An INS1 β cell line stably expressing a live‐cell fluorescent catalase reporter was used to detect changes in catalase gene expression. Results C‐peptide protects INS1 β cells from the combined effect of palmitic acid and glucose by reducing peroxisomal H2O2 to baseline levels and increasing expression of catalase. Conclusions In conditions of glucolipotoxicity, C‐peptide increases catalase expression and reduces peroxisomal oxidative stress and death of INS1 β cells. Maintenance of C‐peptide secretion is a pro‐survival requisite for β cells in adverse conditions. Loss of C‐peptide secretion would render β cells more vulnerable to stress and death leading to secretory dysfunction and diabetes.


| INTRODUC TI ON
Serum conditions associated with diabetes, such as elevation of glucose, saturated free fatty acids (FFAs) and inflammatory cytokines, elicit intracellular production of reactive oxygen species (ROS) generating oxidative stress, which is a leading factor triggering pancreatic β cell degeneration in diabetes. As a consequence, type 1 and type 2 diabetes (T1D and T2D) subjects suffer from variable degrees of loss of β cells and impaired β cell secretion of both insulin and C-peptide. [1][2][3][4][5][6] C-peptide is the 31 amino acid peptide generated in the secretory granules of pancreatic β cells as part of normal insulin biosynthesis. 7 After its cleavage from proinsulin, C-peptide is stored in the β cell secretory granules and co-secreted in equimolar amount with insulin in the bloodstream of healthy individuals in response to ever-changing glycaemia. However, C-peptide does not undergo as much hepatic retention as insulin and circulates at a concentration approximately tenfold higher than that of insulin, with a biological half-life of more than 30 minutes in healthy adult humans, compared to 3-4 minutes for insulin. 8,9 Although for several decades C-peptide has been thought to have no biological activity of its own, more recent evidence point to a role of C-peptide as a 'sensor-effector' of cellular stress able to directly reduce ROS generation by inhibiting glucose-activated nicotinamide adenine dinucleotide phosphate (NADPH) oxidase at the plasma membrane 10,11 and restoring normal electron transport chain activity at mitochondria of endothelial cells. 12,13 In so doing, C-peptide inhibits downstream deleterious effects associated with ROS accumulation and inhibits pro-apoptosis enzymes caspase-3 and transglutaminase-2, while stimulating expression of survival protein Bcl-2 in a variety of peripheral target cells. 10,11,[14][15][16] Our laboratory has demonstrated a novel C-peptide mechanism, in which its beneficial activity expands to the same pancreatic β cells that synthesize and secrete C-peptide, in an autocrine fashion. 17 Thus, C-peptide appears to be more than a coincidental 'bystander' and could be directly acting on β cells over time to maintain a healthy secretory status. Stressful conditions that compromise C-peptide secretion might therefore put β cells at risk for further stress and death contributing to diabetes. 17,18 Palmitic acid (C16:0) is the most physiologically abundant longchain (LC) saturated FFA in the body and is elevated by industrialized diets. 19 High circulating levels of FFAs are a common feature of T2D, particularly in overweight subjects, 20,21 and are also found in T1D patients. 22 Prolonged exposure of β cells to LC (C10-C16) and very LC (C17-C24) FFAs inhibits glucose-induced insulin secretion, increases intracellular ROS production, and triggers β cell death, in a phenomenon termed lipotoxicity. 19,[23][24][25] The toxic actions of FFAs on β cells are dramatically increased in the setting of high glucose (glucolipotoxicity). 26 An early event leading to β cell oxidative stress after palmitic acid exposure is an increase in H 2 O 2 levels 27 from β-oxidation mainly at two subcellular sites, the mitochondria and the peroxisomes. However, excess of palmitic acid and very LC-FFAs which are highly toxic to β cells, critically depend on peroxisomal detoxification for β cell survival. 28 Peroxisomes are single membrane-bound, complex, and highly dynamic organelles present virtually in every eukaryotic cell. 29 Peroxisomes contain oxidase enzymes for FFA β-oxidation, generating H 2 O 2 as a by-product in addition to shortened acyl-CoA, and the enzyme catalase for H 2 O 2 detoxification. While the fatty acyl-CoA metabolites generated by β-oxidation are imported into mitochondria for further oxidation, which generates ATP, the reactive H 2 O 2 remains in the peroxisomes and if left to accumulate becomes toxic to cells. The detoxifying enzyme catalase protects the cells against the oxidative damage by decomposing H 2 O 2 into water and molecular oxygen (O 2 ). Peroxisomes are therefore specialized sites in the cell where LC-and very LC-FFAs-induced H 2 O 2 is both generated and scavenged. Experiments in which catalase was either overexpressed in the peroxisome or in the mitochondria, showed that only peroxisomal catalase provided protection against LC-FFAs-induced lipotoxic death of β cells, while mitochondrial catalase was not protective, thus demonstrating that peroxisomally generated H 2 O 2 mediates lipotoxicity in β cells. 28 Catalase is present in peroxisomes of most tissue, but at a low basal level in peroxisomes of pancreatic β cells, which renders these cells exquisitely sensitive to abnormally high H 2 O 2 levels. [30][31][32] Based on these observations, and that the major toxic product of palmitic acid metabolism is H 2 O 2 in the peroxisomes, we hypothesized that C-peptide can protect β cells against palmitic acid-induced stress by stimulating cellular pathways with antioxidant effectors in peroxisomes. While this hypothesis has remained untested generally for all cell types, in these studies we chose to test it specifically for β cells, which are the only cells of the body that make and secrete C-peptide. A protective autocrine action of C-peptide against palmitic acid-induced β cell stress would provide amplified protection to other peripheral target cells by protecting the very β cells that make and secrete the protecting agent, C-peptide, thereby providing more C-peptide in a positive feedback loop.

| Preparation of palmitic acid solution
Palmitic acid (Cayman Chemical Co.) was applied as a conjugate with BSA by dilution from 5 mmol/L palmitic acid/3.75% fatty acid-free BSA stock, prepared as follows. A 20 mmol/L solution of palmitic acid was made in 8 mL of 0.01 mmol/L NaOH in double distilled water and heated to 70°C for 30 minutes to form a palmitic acid soap. The palmitic acid soap was added to 24 mL of 50 mg/L fatty acid-free BSA (Sigma) in Phosphate Buffer Saline (PBS; Gibco) to form a stock solution of 5 mmol/L palmitic acid conjugated with 3.75% BSA. The 5 mmol/L palmitic acid/3.75% BSA stock was aliquoted and frozen at −20°C. For each experiment, one aliquot was thawed and diluted in the appropriate cell culture medium or PBS to the indicated palmitic acid concentration. For the 0 palmitate controls, the same procedure was used except that the stock was 0 mmol/L palmitic acid/3.75% fatty acid-free BSA, which was diluted to the equivalent BSA concentration used in each experiment. After each use, thawed stock solutions were discarded.

| Detection of cell death and apoptosis
INS1 β cells (10 000/well) were seeded in 96-well plates in regular medium for 48 hours at 37°C, 5% CO 2 . Then, cells were serumstarved overnight in minimum medium at 37°C, 5% CO 2 . The next day, medium was replaced with minimum medium with either 5.5, or 11, or 22 mmol/L glucose with or without 5 or 10 nmol/L C-peptide (Phoenix Pharmaceuticals) for 20 minutes at 37°C, 5% CO 2 before adding 100 or 200 μmol/L palmitic acid or equimolar BSA (Gibco) as control. Cells were then incubated for 24 hours at 37°C, 5% CO 2 . To detect cell death, 1 μg/mL of propidium iodide (PI; Sigma) was added and incubated in the dark on a rocking plate at room temperature for 15 minutes. PI is a red-fluorescent nuclear and chromatin counterstain that is not permeant to live cells, therefore commonly used to detect dead cells in a population. PI fluorescence excitation was 535 nm and emission was read at 617 nm using a computer-controlled BioTek Synergy H1 Hybrid Multi-Mode plate reader (Winooski, VT).

Each condition was tested in a total of 24 independent experiments.
Cell death is shown as arbitrary units of PI fluorescence at 617 nm expressed as mean ± standard error of the mean (SEM).
To detect INS1 β cell apoptosis, cells (10 000/well) were seeded in 48-well plates in regular medium for 72 hours at 37°C, 5% CO 2 . Then, cells were serum-starved overnight in minimum medium at 37°C, 5% CO 2 . The next day, medium was replaced with minimum medium with 11 mmol/L glucose with or without 5 or 10 nmol/L C-peptide (Phoenix Pharmaceuticals) for 20 minutes at 37°C, 5% CO 2 before adding 200 μmol/L palmitic acid or equimolar BSA (Gibco) as control.
Cells were then incubated for 24 hours at 37°C, 5% CO 2 . Apoptosis was detected using the Cell Death Detection ELISA PLUS kit (Roche  After washing three times with BSA solution, sections were incubated for 1 hour with a goat anti-rabbit Alexa 488 secondary antibodies (Invitrogen, 1:500) before being washed three times with BSA solution, then once with Hoechst dye (1 g/100 mL) for 30 seconds to stain nuclear DNA. The sections were washed and placed on coverslips with Gelvatol, a water-soluble mounting medium. Cells were evaluated using an Olympus Fluoview 1000 confocal microscope.
Cells exposed to equimolar amounts of BSA were used as a control.

| Live-cell catalase expression assays using CAT minigene-mNeonGreen reporter
To test for changes in catalase gene expression, we used a rat catalase minigene with DNA sequences encoding mNeonGreen fused to the C-terminal sequences encoding the catalase protein. The reporter includes 1874 bases of the catalase gene promoter including its PPARγ enhancer and the native 5′UTR and the entire catalase protein coding sequences followed by the mNeonGreen sequences.
Stable transfectants of the reporter in the INS1 β cells were isolated using a vector puromycin resistance marker, as described for the mCherry-Hyper INS1 β cell line above.
Once generated, INS1 β cells stably expressing the catalase reporter were seeded in 24-well plates (50 000/well) in regular medium and kept at 37°C, 5% CO 2 for 48 hours. Medium was removed and replaced with PBS with 11 mmol/L glucose in the absence or presence of 5, 10 nmol/L or 10 nmol/L heat-inactivated C-peptide (Phoenix Pharmaceuticals) for 20 minutes at 37°C, 5% CO 2 before addition of 200 μmol/L palmitic acid at 37°C, 5% CO 2 for 24 hours.
C-peptide was heat-inactivated by boiling it for 1 hour and quickly chilled in ice immediately before addition to cells in culture. 35 Cells exposed to equimolar amounts of BSA were used as a 0 μmol/L palmitic acid control. Each condition was tested in a total of 20 independent experiments performed over a two-month period. In each experiment changes in normalized green fluorescence were measured and normalized to Hoechst to account for any cell number variability from well to well by using the BioTek Synergy H1 Hybrid Multi-Mode plate reader, and expressed as live-cell fluorescence arbitrary units mean ± SEM

| Statistical analysis
Results are presented in boxplots, with the box indicating the central 50% of the data in which a line indicates the median and the whiskers the range. ANOVA followed by the Tukey post hoc test was used to assess differences between different conditions using Prism 8 (GraphPad Software, Inc). Values of P < .05 were considered statistically significant.

| Glucose enhances palmitic acid-induced INS1 β cell death
Exposure of β cells to elevated levels of palmitic acid causes oxidative stress and death. The toxic effect of palmitic acid on β cells is enhanced by concomitant elevated glucose (glucolipotoxicity). We therefore sought to study the effect of increasing glucose on palmitic acid-induced INS1 β cell death as measured by PI fluorescence.
In Figure 1

| C-peptide protects INS1 β cells from palmitic acid-induced cell death
We next analysed the effect of C-peptide on palmitic acid-induced

| C-peptide protects INS1 β cells from palmitic acid-induced death by decreasing apoptosis
To further investigate the effect of C-peptide on palmitic acid-induced INS1 β cell death with respect to the mechanisms involved, we exposed the cells to 200 μmol/L palmitic acid and tested the ability of C-peptide to protect from apoptosis. Figure 3 shows that addition of 200 μmol/L palmitic acid for 24 hours increased apoptosis however this decrease was not significantly different than the protection achieved with 5 nmol/L C-peptide. These results indicate that in palmitic acid-exposed INS1 β cells as well, C-peptide displays F I G U R E 2 C-peptide decreases palmitic acid-induced INS1 β cell death. Cells were plated in 96-well plates in regular medium and seeded for 48 h at 37°C, 5% CO 2 . Then, they were serum-starved for 24 h in minimum medium at 37°C, 5% CO 2 . The next day, medium was replaced with minimum medium with either 5.5, or 11, or 22 mmol/L glucose in the absence or presence of 5 or 10 nmol/L C-peptide for 20 min at 37°C, 5% CO 2 before addition of 100 or 200 μmol/L palmitic acid, or equimolar Bovine Serum Albumin (BSA) as 0 palmitic acid control. Cells were then incubated for 24 h at 37°C, 5% CO 2 . To detect cell death, 1 μg/ml of propidium iodide (PI) was added to each well, and fluorescence assayed after 15 min incubation using a BioTek H1 Hybrid plate reader. INS1 β cell death is shown as arbitrary units of PI fluorescence at 617 nm. In panel A, 100 μmol/L palmitic acid-induced an increased cell death compared to BSA in 5.5 mmol/L (***P < .005), 11 mmol/L (****P < .0001), and 22 mmol/L glucose (****P < .0001). Addition of 5 nmol/L C-peptide decreased palmitic acid-induced cell death in the presence of 22 mmol/L glucose (***P < .005). A further reduction in cell death was observed with 10 nmol/L C-peptide in 11 mmol/L (****P < .0001) and 22 mmol/L glucose (****P < .0001), and this reduction was significant when compared to 5 nmol/L C-peptide in both 11 mmol/L (**P < .01) and 22 mmol/L glucose (*P < .05). While palmitic acid significantly increased cell death compared to BSA alone (***P < .005) in 5.5 mmol/L glucose, C-peptide addition did not show a significant reduction in cell death at this glucose concentration. The protective effect of C-peptide on palmitic acid-induced beta cell death extends to conditions of 200 μmol/L palmitate exposure, as shown in panel B. Palmitic acid-induced cell death increased in all glucose concentrations as compared to BSA control (P < .0001). Addition of 5 nmol/L C-peptide protected against palmitic acid-induced cell death in all glucose concentrations compared to palmitic acid alone (***P < .005). A further protection was achieved with 10 nmol/L C-peptide as compared to palmitic acid alone in all glucose concentrations (****P < .0001), and this protection was significantly better than the one achieved with 5 nmol/L C-peptide in 5.

| Palmitic acid and C-peptide stimulate catalase expression in INS1 β cells
To further evaluate the mechanism by which C-peptide lowers palmitic acid-induced peroxisomal H 2 O 2 , we tested whether C-peptide and palmitic acid increase peroxisomal catalase expression. We used a rat catalase minigene-mNeonGreen reporter to measure the effects of C-peptide and palmitic acid on catalase expression. Figure 5 shows fluorescence of the rat catalase-mNeonGreen reporter in INS1 β cells exposed to 11 mmol/L glucose in the presence of and were significantly lower than in 0 μmol/L palmitic acid with 10 nmol/L C-peptide (10 129 ± 557.4), indicating that a functional C-peptide is responsible for the increase in catalase expression.
These results indicate C-peptide and palmitic acid increase catalase expression and that C-peptide has autocrine protective capabilities through a catalase mechanism which is enhanced with palmitic acid.

| D ISCUSS I ON
The tective activity of C-peptide in palmitic acid-exposed INS1 β cells involves attenuation of apoptosis. [14][15][16][17][18] Glucolipotoxicity is characterized by abnormally high levels of intracellular ROS leading to β cell apoptosis. 30 Live-cell catalase expression assays using CAT minigene-mNeonGreen reporter. Cells were seeded in 24-well plates in regular medium for 24 h at 37°C and 5% CO 2 and then exposed to 200 μmol/L palmitic acid and 0, 5, 10 nmol/L C-peptide or 10 nmol/L inactivated C-peptide for 24 hr and assessed for catalase expression through fluorescence of the CAT minigene-mNeonGreen reporter using a BioTek H1 Hybrid plate reader. Fluorescence was obtained with an excitation and emission wavelength of 490 and 520 nm, respectively. 200 μmol/L palmitic acid and 10 nmol/L C-peptide both increased catalase expression (arbitrary units) compared to the control of no treatment (****P < .0001). Catalase expression increased with 200 μmol/L palmitic acid and 5 nmol/L C-peptide (****P < .0001) and 10 nmol/L C-peptide (****P < .0001) compared with 0 nmol/L C-peptide. No significant difference was observed between 0 nmol/L C-peptide and 10 nmol/L inactivated C-peptide, but 10 nmol/L inactivated C-peptide was significantly lower than 10 nmol/L C-peptide (****P < .0001). Results are presented in boxplots, with the box indicating the central 50% of the data in which a line indicates the median and the whiskers the range   [30][31][32] There is evidence that expression of these enzymes can be induced under conditions of hyperglycaemia and by exposure to insulinotropic agents in wild-type murine islets and cell lines. 36,37 A second important finding of the current study is that C-peptide increases catalase expression and that the presence of both palmitic acid and C-peptide have an additive effect relative to either alone.
The highest expression detected was with 200 μmol/L palmitic acid and 5 nmol/L or with 200 μmol/L palmitic acid and 10 nmol/L C-peptide, compared with just 10 nmol/L C-peptide alone or 200 μmol/L palmitic acid alone. These results provide evidence that C-peptide has autocrine protective capabilities in peroxisomes against excess palmitic acid-induced ROS through a catalase mechanism. Thus, C-peptide antioxidant capability is evidently not limited to inhibiting ROS generation by only acting on plasma membrane-associated NADPH-dependent oxidase or by improving mitochondrial respiration, 10-13 but extends to the stimulation of pathways that decrease toxic levels of H 2 O 2 in peroxisomes where ROS does the most damage in β cells. 28 Our results provide evidence for a model featuring novel organellar mechanisms involving activation of pathways that increase catalase expression levels. Taken together, our results provide further evidence suggesting that C-peptide functions beyond its role for proper folding and disulphide bond formation of the A and the B chains of insulin within the β cell. These findings support models in which C-peptide has acquired physiological roles that occur after its secretion from β cells and functions to provide antioxidant actions. 17,18,38 Here, we consider the nature and components of the pathways resulting in secreted C-peptide protecting β cells.
C-peptide binds to its receptor which could be GPCR146. 39 Signalling at the β cell surface membrane or along its endocytic itinerary, 40 the C-peptide/GPCR complex presumably activates one or more cytosolic factors that elevate peroxisome proliferator activator (PPAR)-γ levels in the cytosol, nucleus, or both cellular compartments, as shown in kidney or in lung epithelial cells. [41][42][43] PPAR-γ is a member of the nuclear receptor superfamily of ligand-activated transcription factors, that is known to transcriptionally regulate catalase expression. 44 The increase in PPAR-γ could be achieved by C-peptide signalling enhanced synthesis via transcription or translation or slowed degradation. The increase in PPAR-γ might also be accompanied by an increase in its phosphorylation, which could increase catalase gene transcription leading to increased catalase, and consequently the observed attenuation of peroxisomal H 2 O 2 levels in the setting of glucolipotoxicity. These underlying intracellular mechanisms of C-peptide antioxidant activity against glucolipotoxicity are under investigation.
The results indicating that C-peptide induces catalase expression and reduces peroxisomal peroxide levels in β cells do not in any way exclude other components of the peroxisomal antioxidant defence pathways being involved. In mammalian cells, several other ROS detoxifying enzymes are known to contribute to peroxisomal redox balance, which include superoxide dismutases, peroxiredoxins, glutathione S-transferases, and epoxide hydrolases. 29 It will be important to test whether C-peptide displays any effects on superoxide dismutase as well as other antioxidant enzymes in addition to catalase as part of the mechanisms of protecting β cells. The findings support models in which C-peptide stimulates antioxidant pathways, including those that functionally protect against the oxidative stress consequent to excess long-chain fatty acid catabolism in peroxisomes. These antioxidant pathways raise the question whether similar mechanisms of C-peptide protection in peroxisomes are active in other cell types known to be protected by C-peptide, such as those of the endothelium and smooth muscle.
Given that C-peptide is both made by and protects β cells, changes in C-peptide levels, up or down, will have effects amplified by its autocrine action or lack thereof. Less secreted C-peptide will result in less protection from oxidative stress which, depending on the oxidative load, will put β cells at risk of apoptosis, resulting in less secreted C-peptide, less protection, and so forth in a downward spiral. Accordingly, more C-peptide will result in more protection from oxidative stress and decreased apoptosis resulting in greater numbers of β cells surviving the oxidative threat in the setting of glucolipotoxicity. The autocrine implications of the model logically extends to therapeutic increases in C-peptide, which would be predicted to decrease β cell oxidative stress and death by apoptosis, as shown here, in the setting of not only diabetes but also, possibly, obesity in the absence of diabetes, and thereby slow the loss of functional β cell mass. These results and observations provide further support to test for and develop co-replacement C-peptide with insulin therapy clinically. Interestingly, there is a general positive association between C-peptide levels and the protection of β cell functional mass year after year of T1D. [2][3][4][5][6] Also, consistent with the present findings, there is a negative association between C-peptide levels and the rate of development of diabetic complications in T1D. [45][46][47][48][49][50] What is lacking in therapeutic interventions might be simply adding back more than part of what was lost in T1D, not only the insulin but the C-peptide as well.
Given that glucolipotoxicity is a shared setting of diabetes in general, these developments appear to apply to T2D as well as T1D subjects.

ACK N OWLED G EM ENTS
The authors would like to thank the Center for Biologic Imaging at the University of Pittsburgh School of Medicine for outstanding fluorescence instrumentation and analysis and Dr Alexander Sorkin for critical discussions.

CO N FLI C T O F I NTE R E S T
The authors declare that there is no conflict of interest associated with this manuscript.

E TH I C S S TATEM ENT
There is no ethics statement for this manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.