Context-dependent effects of CCN2 on β-cell mass expansion and indicators of cell stress in the setting of acute and chronic stress

Abstract

Stimulation of functional β-cell mass expansion can be beneficial for the treatment of type 2 diabetes. Our group has previously demonstrated that the matricellular protein CCN2 can induce β-cell mass expansion during embryogenesis, and postnatally during pregnancy and after 50% β-cell injury. The mechanism by which CCN2 stimulates β-cell mass expansion is unknown. However, CCN2 does not induce β-cell proliferation in the setting of euglycemic and optimal functional β-cell mass. We thus hypothesized that β-cell stress is required for responsiveness to CCN2 treatment. In this study, a doxycycline-inducible β-cell-specific CCN2 transgenic mouse model was utilized to evaluate the effects of CCN2 on β-cell stress in the setting of acute (thapsigargin treatment ex vivo) or chronic [high-fat diet or leptin receptor haploinsufficiency (db/+) in vivo] cellular stress. CCN2 induction during 1 wk or 10 wk of high-fat diet or in db/+ mice had no effect on markers of β-cell stress. However, CCN2 induction did result in a significant increase in β-cell mass over high-fat diet alone when animals were fed high-fat diet for 10 wk, a duration known to induce insulin resistance. CCN2 induction in isolated islets treated with thapsigargin ex vivo resulted in upregulation of the gene encoding the Nrf2 transcription factor, a master regulator of antioxidant genes, suggesting that CCN2 further activates this pathway in the presence of cell stress. These studies indicate that the potential of CCN2 to induce β-cell mass expansion is context-dependent and that the presence of β-cell stress does not ensure β-cell proliferation in response to CCN2.

NEW & NOTEWORTHY CCN2 promotes β-cell mass expansion in settings of suboptimal β-cell mass. Here, we demonstrate that the ability of CCN2 to induce β-cell mass expansion in the setting of β-cell stress is context-dependent. Our results suggest that β-cell stress is necessary but insufficient for CCN2 to increase β-cell proliferation and mass. Furthermore, we found that CCN2 promotes upregulation of a key antioxidant transcription factor, suggesting that modulation of β-cell oxidative stress contributes to the actions of CCN2.

INTRODUCTION

Dysfunction or loss of functional β-cell mass leads to persistent hyperglycemia and eventual development of type 2 diabetes (T2D). In situations of increased metabolic demand, such as obesity and pregnancy, the failure of β cells to compensate and secrete enough insulin to maintain euglycemia results in the prolonged hyperglycemia that is a hallmark of T2D (1). The primary mechanism by which adult β-cell mass increases is replication of existing β cells (2), and it is known that β-cell replicative capacity decreases with age (3–5). Thus, identification of signaling pathways and factors that stimulate β-cell replication to increase functional β-cell mass is a goal for treating T2D.

One factor that has been identified to promote β-cell proliferation is cellular communication network factor 2 (CCN2; formerly known as CTGF). CCN2 is a member of the CCN family of proteins, a family of extracellular matrix-localized proteins that modulate signaling in various ways depending on the cellular microenvironment (6). Due to the modular structure of the protein, CCN2 is capable of interacting with a variety of molecules, leading to modulation of multiple cellular processes such as adhesion, migration, proliferation, and angiogenesis (6). In the developing pancreas, CCN2 is expressed in the vascular endothelium, ductal endothelium, and embryonic insulin-producing cells. However, expression of CCN2 in β cells is silenced soon after birth (7). Our group has demonstrated that CCN2 is involved in embryonic β-cell proliferation: homozygous CCN2 gene inactivation in mice results in decreased β-cell number at birth due to decreased β-cell proliferation at late gestation (7). Conversely, transgenic overexpression of CCN2 in embryonic insulin-producing cells results in increased β-cell proliferation and mass at birth (8). Global CCN2 haploinsufficiency during pregnancy impairs maternal β-cell proliferation in mice, demonstrating that normal levels of CCN2 are required for maternal β-cell compensation in a situation where metabolic demand for insulin is increased (9). Finally, induction of CCN2 in adult β cells following partial β-cell ablation promotes β-cell regeneration and restoration of functional β-cell mass (10).

Although CCN2 has the capacity to stimulate β-cell proliferation in these various metabolic settings, induction of CCN2 in adult β cells under euglycemic, unstressed conditions fails to increase β-cell proliferation (11). This suggests that CCN2 is capable of inducing β-cell proliferation only in settings of suboptimal β-cell mass, which can be considered a β-cell stressor, as it requires increased insulin production in remaining β cells to meet demand and maintain euglycemia. To date, the mechanisms by which CCN2 induces β-cell proliferation have not been elucidated; a receptor for CCN2 has not been identified in any context. We hypothesize that β cells can only respond to CCN2 during situations of suboptimal β-cell mass, which includes situations of β-cell stress.

Here we used complementary in vivo and ex vivo model systems to examine whether CCN2 induction during situations that are known to elevate β-cell stress can stimulate β-cell proliferation and whether CCN2 modulates markers of β-cell stress under these conditions. Thapsigargin (TG) was used to induce acute endoplasmic reticulum (ER) stress in ex vivo islet assays. In vivo chronic models of β-cell stress included db/+ mice (which we have shown express hallmarks of β-cell oxidative stress despite remaining euglycemic; 12) and prolonged high-fat diet (HFD) feeding (a well-established model of glucolipotoxicity-induced β-cell stress; 13). We find that the ability of β cells to respond to CCN2 is context dependent and the presence of β-cell stress does not ensure a proliferative response to CCN2. We also find that CCN2 has no effect on most markers of ER and oxidative stress, with the exception of increased expression of Nfe2l2 (the gene encoding Nrf2). Thus, CCN2 may act in part by enhancing the β-cell response to increased oxidative stress.

MATERIALS AND METHODS

Animals

Generation of mouse models.

RIP-rtTA;TetO-CTGF. Generation of mice harboring the rat insulin promoter driving expression of the reverse tetracycline transactivator protein (RIP-rtTA), and the Tet-On operator upstream of the CCN2 transgene (TetO-CTGF) was described previously (7, 14). RIP-rtTA;TetO-CTGF mice were maintained on a mixed C57BL6/J background. Genotyping primers are available on request. To induce CCN2 expression, mice were administered doxycycline (DOX) in the drinking water at 2 mg/mL with 2% Splenda. Only male mice were used for the HFD studies, whereas islets from both sexes were used for ex vivo assays. Mice were housed in a 12-h light/dark cycle with ad libitum access to food (11% kcal from fat; 5LJ5, Purina, St. Louis, MO) and water.

db/+;RIP-rtTA;TetO-CTGF. In-house bred Lepr^db/+ (db/+) on the C57BL/KsJ genetic background (Jax No. 000697) were bred with RIP-rtTA;TetO-CTGF mice to obtain experimental db/+;RIP-rtTA;TetO-CTGF mice and littermate controls. Mice were housed in a 12-h light/dark cycle and had ad libitum access to food (11% kcal from fat; 5LJ5, Purina, St. Louis, MO) and water. Four- or six-week-old male mice were administered DOX-containing water (2 mg/mL with 2% Splenda) and euthanized 2 wk later at 6 or 8 wk of age.

All mice were housed in the Vanderbilt University Medical Center (Nashville, TN) animal care facility. This facility is accredited by the American Association for Accreditation of Laboratory Animal Care. All mouse experiments obtained approval by the Vanderbilt University Medical Center Institutional Animal Care and Use Committee.

High-fat diet studies.

Male RIP-rtTA;TetO-CTGF and monogenic control (RIP-rtTA and TetO-CTGF) mice 8–10 wk of age were weighed and randomly assigned to one of two groups: 1) control diet (CD; BioServ F4031, Flemington, NJ) and 2) HFD (60% kcal from fat; BioServ F3282, Flemington, NJ). Mice were administered either CD or HFD for 1 wk or 10 wk. All mice were administered DOX-containing water (2 mg/mL with 2% Splenda) for the duration of the studies. Mice were housed in a controlled-temperature environment on a 12-h light/dark cycle with ad libitum access to food and water. Euthanasia was performed at time of pancreatic dissection using isoflurane until mice were unresponsive, followed by cervical dislocation.

Immunolabeling.

Pancreata were dissected, fixed, and processed as described previously (15). Primary antibodies used were as follows: guinea pig anti-insulin (Dako, Cat. No. 10564, Carpinteria, CA), rabbit anti-Ki67 (1:400; Abcam, Cat. No. ab15580, Cambridge, MA), mouse anti-8-hydroxy-3-deoxyguanosine (8-OHdG) (1:100; Abcam, Cat. No. ab62623, Cambridge, MA), rabbit anti-Grp78 (also known as binding immunoglobulin protein or BiP) (1:500; Abcam, Cat. No. ab21685, Cambridge, MA), goat anti-Pdx1 (1:1000 Beta Cell Biology Consortium; Cat. No. AB2027), and rabbit anti-phospho-Nrf2 (1:200; Abcam, Cat. No. ab76026, Cambridge, MA). Primary antibodies were detected with the appropriate species-specific secondary antibodies (Jackson ImmunoResearch Laboratories, West Grove, PA): Cy2-conjugated anti-guinea pig IgG (1:400; Cat. No. GP 706-225-148), Cy3-conjugated anti-rabbit IgG (1:400; Cat. No. Rb 711-165-152), Cy3-conjugated anti-mouse IgG (1:400; Cat. No. Ms 715-165-150), and Cy3-conjugated anti-goat IgG (1:400; Cat. No. 705-165-003). Nuclei were visualized using 4′,6′-diamidino-2-phenylindole (DAPI, 1 µg/mL; Molecular Probes D1306, Eugene, OR). For Insulin/8-OHdG, Insulin/phospho-Nrf2, and Insulin/Pdx1, three pancreatic sections (at least 250 μm apart) were subject to sodium citrate (pH 6) antigen retrieval, and antibodies were diluted in 0.2% Triton X-100 in PBS with 5% normal donkey serum and 1% BSA. For all other immunolabeling, pancreatic sections were subject to sodium citrate (pH 6) antigen retrieval, and antibodies were diluted in 0.2% Triton X-100 in PBS with 5% normal donkey serum. A ScanScope FL scanner (Aperio Technologies, Inc.) was used for imaging Insulin/Ki67, Insulin/BiP, and Insulin/Pdx1 immunolabeling, and Stochastic Optical Reconstruction Microscopy (STORM; Nikon) was used for imaging Insulin/8-OHdG and Insulin/phospho-Nrf2 immunolabeling. Images were quantified using ImageScope software (Aperio Technologies, Inc.), ImageJ software (16), and NIS-Elements Imaging Software (Nikon).

β-cell mass.

At least five slides per animal spaced at least 250 μm apart (1%–2% of the entire pancreas) were immunolabeled for insulin and visualized using a DAB peroxidase substrate kit (Vector Laboratories, Inc), then counterstained with eosin. Quantification of β-cell mass was performed as described in Golson et al. (17).

β-cell proliferation.

Three slides (at least 250 μm apart) per animal were immunolabeled for insulin and Ki67. A minimum of 3,000 cells were counted using ImageScope software (Aperio Technologies, Inc). β-cell proliferation was determined by quantifying the number of dual-positive insulin-Ki67 cells using a macro generated with the CytoNuclear FL algorithm in eSlide Manager (Aperio Technologies, Inc.).

β-cell size.

Three slides (at least 250 μm apart) per animal were immunolabeled for insulin. Insulin-positive area was determined using a macro generated with the CytoNuclear FL algorithm in eSlide Manager (Aperio Technologies, Inc). β-cell size was determined by dividing insulin-positive area by the number of nuclei. A minimum of 3,000 β cells per animal were counted to calculate β-cell size.

Assessing oxidative and endoplasmic reticulum stress.

To quantify β-cell 8-OHdG, three slides (at least 250 μm apart) per animal were immunolabeled for 8-OHdG and insulin and imaged by STORM (Nikon). A macro generated using NIS-Elements Imaging Software (Nikon) was utilized to quantify the percentage of β cells with nuclear localization of 8-OHdG. To quantify BiP levels, three slides (at least 250 μm apart) per animal were immunolabeled for BiP and insulin, and imaged by a ScanScope FL scanner (Aperio Technologies, Inc.). Individual islets from each section were imaged using ImageScope software (Aperio Technologies, Inc.), and mean gray value was calculated using ImageJ software (16). At least 20 islets per animal were analyzed. To examine nuclear localization of phospho-Nrf2, three slides (at least 250 μm apart) per animal were immunolabeled for phospho-Nrf2 and insulin and imaged by STORM (Nikon). A macro generated using NIS-Elements Imaging Software (Nikon) was utilized to quantify the percentage of β cells with nuclear localization of phospho-Nrf2.

Assessing β-cell identity.

To quantify β-cell Pdx1 expression, three slides (at least 250 μm apart) were immunolabeled for Pdx1 and insulin and imaged using a ScanScope FL scanner (Aperio Technologies, Inc.). Individual islets from each section were imaged using ImageScope software (Aperio Technologies, Inc.). At least 1,000 β cells were manually counted using ImageJ software (16), and percentage of β cells labeled for Pdx1 was calculated by dividing the number of Pdx1+ β cells by the total number of β cells counted. At least 20 islets per animal were analyzed.

Islet isolations.

After euthanasia, pancreata were perfused with 0.5 mg/mL type IV collagenase that was dissolved in Hanks balanced salt solution as described previously (18). Islets isolated from wild-type C57BL/6J were handpicked and cultured overnight in 11 mM glucose RPMI-1640 with 10% horse serum and 1% penicillin/streptomycin (P/S) in preparation for ex vivo stress assays and RNA isolation. Isolated islets from RIP-rtTA;TetO-CTGF, RIP-rtTA, and TetO-CTGF animals were handpicked and cultured overnight in 11 mM glucose RPMI-1640 with 10% horse serum and 1% P/S supplemented with 2 μg/mL doxycycline to maintain CCN2 expression in preparation for ex vivo stress assays and RNA isolation.

Ex vivo islet assays.

Islets were treated with thapsigargin (Sigma-Aldrich T9003, Darmstadt, Germany) diluted in phosphate-buffered saline (PBS) at 10 nM concentration for 24 h. Islets were harvested and prepared for RNA extraction in 1 mL Trizol reagent (ThermoFisher, Waltham, MA) and stored at −80°C before RNA isolation.

Islet gene expression analysis.

RNA isolation was performed via phenol-chloroform extraction. RNA integrity and concentration were assessed with a ND-1000 spectrophotometer (NanoDrop). A total of ≥50 ng cDNA was prepared from mouse islets using the high-capacity cDNA reverse transcription kit (ThermoFisher, Cat. No. 4368814, Waltham, MA). Quantitative RT-PCR (qRT-PCR) was performed as described previously (19). Primer sequences are available on request. Data were normalized to Actb. Mouse islet data are represented as 2^−ΔΔCt compared with the respective control groups.

Statistics.

All data were plotted as means ± SE. Student’s t test, one-way ANOVA, and the Tukey’s post hoc test were performed where appropriate (and as indicated in each figure legend). Statistical significance was set at P < 0.05 (denoted by one symbol), P < 0.01 (denoted by two symbols), and P < 0.001 (denoted by three symbols). Statistical analysis was conducted using Prism 9.0 (GraphPad Software, Inc.).

RESULTS

CCN2 Induction Does Not Attenuate Acute Cellular Stress

Our previous studies suggested that CCN2 is only capable of promoting β-cell proliferation under situations of suboptimal β-cell mass and/or increased β-cell stress, but not under euglycemic, unstressed conditions (10, 11). One possibility is that CCN2-mediated β-cell proliferation is associated with modulation of β-cell stress. Thus, we assessed the acute effects of CCN2 on markers of cell stress in isolated islets ex vivo. Isolated adult mouse islets were treated for 24 h with 10 nM thapsigargin which acts to deplete ER calcium, leading to activation of ER stress pathways and the unfolded protein response (20). ER stress is closely linked to oxidative stress (21). Thus, qRT-PCR was used to examine expression of both ER [Grp78, Ddit3 (CHOP), sXbp1, tXbp1, Atf6, Eif2s1] and oxidative (Nfe2l2, Sod1, Txn1) (22, 23) stress genes. The selection of ER stress markers included genes from the three ER unfolded protein response (UPR) pathways. The oxidative stress genes Sod1 and Txn1 were selected based on previous studies demonstrating the importance of these genes in preserving β-cell function and aiding β-cell survival during diabetogenic conditions (24–26). Nfe2l2 (Nrf2) was selected because of its role as a major mediator of the cellular antioxidant response (27). Successful induction of stress markers by thapsigargin was achieved, as expression of all markers assessed was increased at least 1.5-fold compared with control (Supplemental Fig. S1).

To assess whether induction of CCN2 specifically in β cells alters thapsigargin-mediated upregulation of stress markers, islets were isolated from 8- to 10-wk-old male and female RIP-rtTA;TetO-CTGF mice that were administered DOX for 2 days before islet isolation to induce CCN2 in vivo. After isolation, islets were cultured in media supplemented with DOX throughout the duration of the assay to sustain CCN2 expression in culture. Islets from littermates (RIP-rtTA or TetO-CTGF) of both sexes were utilized as controls. In agreement with previous studies from our laboratory, CCN2 gene expression was efficiently upregulated in response to DOX treatment (10; Supplemental Fig. S2A). Twenty-four hours of thapsigargin treatment resulted in a similar increase in expression of stress-related genes in control islets and those in which CCN2 was induced in β cells (Fig. 1). Nfe2l2, the gene encoding the antioxidant gene transcription factor Nrf2 (27), was further elevated with the combination of CCN2 induction plus thapsigargin treatment compared with control islets treated with thapsigargin (Fig. 1E).

Figure 1.

CCN2 upregulates genes associated with resolution of oxidative stress. A–L: qRT-PCR was performed on RNA from islets isolated from mice with/without CCN2 induction and with/without 10 nM thapsigargin treatment for 24 h. CCN2 induction increased expression of the antioxidant gene Nfe2l2 after TG treatment. n = 3–4. All samples were run in duplicate for these assays. Data were analyzed using one-way ANOVA and Tukey’s post hoc analysis. *P < 0.05, **P < 0.01, and ***P < 0.001.

β-Cell CCN2 Induction Has No Effect on Stress Markers in the Setting of Chronic In Vivo β-Cell Stress

To determine whether β-cell induction of CCN2 modulates cell stress in a chronic in vivo setting, we used two mouse models of β-cell stress: mice heterozygous for the Lepr^db mutation (db/+) and HFD feeding. Mice carrying the Lepr^db mutation in the homozygous state are commonly used as a model of T2D as they become obese and develop overt diabetes (28). In the current study, db/+ animals were interbred with RIP-rtTA;TetO-CTGF animals. Littermates (both db/+;RIP-rtTA and db/+;TetO-CTGF) were utilized as controls. CCN2 gene expression was efficiently upregulated in response to DOX treatment, similar to previously published studies from our laboratory (10) (Supplemental Fig. S2B). Assessment of insulin secretion by dynamic perifusion assay demonstrated that β-cell-specific induction of CCN2 does not alter the insulin secretory response to various secretagogues (Supplemental Fig. S3).

Previous studies from our group have revealed that β cells from db/+ animals exhibit mild oxidative stress (12) as evidenced by increased nuclear localization of the Nrf2 transcription factor and 8-OHdG, a marker of oxidative DNA damage (29). The current study provides additional support for this. Analysis of pancreatic sections from control (RIP-rtTA or TetO-CTGF) mice and mice carrying the leptin receptor mutation (db/+;RIP-rtTA or db/+;TetO-CTGF) reveals that islets from mice carrying the leptin receptor mutation showed a trend toward higher levels of BiP (Supplemental Fig. S4), an ER lumen chaperone protein increased with ER stress (30), and significantly more 8-OHdG levels compared with controls (Supplemental Fig. S4). Taken together these results support a higher level of cellular stress in β cells from db/+ animals.

To generate a model of chronic mild in vivo β-cell stress in which CCN2 could be induced in a β-cell-specific manner, male mice on the db/+ background were treated with DOX for 2 wk during two different time windows: 4–6 wk and 6–8 wk of age. These time points were chosen based on a previous study showing that a compensatory increase in β-cell proliferation occurs at 4 wk of age in db/db mice, but by 6 wk of age β-cell proliferation is comparable with wild-type controls (31). Two weeks of β-cell-specific CCN2 induction in mice carrying the leptin receptor mutation had no effect on glucose tolerance at either time point (Supplemental Fig. S5). In addition, CCN2 induction did not result in any changes in 8-OHdG (Fig. 2, A–C) or BiP (Fig. 2, D–F) in β cells. A subset of acinar cells showed increased levels of BiP in all animals regardless of genotype or CCN2 induction (Fig. 2D).

Figure 2.

CCN2 induction does not alter β-cell oxidative or ER stress in db/+ animals. A: representative images of 8-OHdG immunolabeling in pancreata from db/+ animals with CCN2 induction from 4 to 6 wk of age. 8-OHdG—red, Ins—green, DAPI—blue. Quantification of percent of β cells expressing 8-OHdG in the nucleus after CCN2 induction from 4 to 6 wk of age (B) and from 6 to 8 wk of age (C). D: representative images of BiP immunolabeling in paraffin-embedded pancreata from db/+ animals with CCN2 induction from 6 to 8 wk of age. BiP—red, Ins—green, DAPI—blue. BiP quantification in β cells after CCN2 induction from 4 to 6 wk of age (E) and from 6 to 8 wk of age (F). n = 5–6 animals/group. Control groups for 4- to 6-wk 8-OHdG: 3 RIP-rtTA, 2 TetO-CTGF. Control groups for 4- to 6-wk BiP: 3 db/+;RIP-rtTA, 2 db/+;TetO-CTGF. Control groups for 6- to 8-wk 8-OHdG: 2 RIP-rtTA, 2 TetO-CTGF. Control groups for 6- to 8-wk BiP: 2 RIP-rtTA, 3 TetO-CTGF. 4 db/+;RIP-rtTA and 2 db/+;TetO-CTGF. Control groups for 6- to 8-wk studies: 2 db/+;RIP-rtTA and 3 db/+;TetO-CTGF. Data were analyzed using a Student’s t test (scale bar, 50 μm). ER, endoplasmic reticulum.

HFD feeding was used as a second model of chronic β-cell stress. Male RIP-rtTA;TetO-CTGF and littermate controls (both RIP-rtTA or TetO-CTGF) were placed either on control diet (CD) or HFD for 1 wk or 10 wk beginning at 8 wk of age. These time points were chosen based on a previous study by our group demonstrating a significant increase in β-cell proliferation in the absence (1 wk HFD) or presence (9 wk or more HFD) of insulin resistance (32). Previous studies by our group demonstrated that β-cell-specific induction of CCN2 during euglycemic, unstressed conditions has no effect on β-cell proliferation or mass (10, 11). Thus, we did not include a CD + CCN2 group in the current study. However, the effect of HFD feeding on CCN2 gene expression in islets had never been assessed. qRT-PCR on islets from CD, HFD, and HFD with β-cell-specific CCN2 induction confirmed previous studies showing that CCN2 is not normally expressed in adult islets (7; Supplemental Fig. S2C). HFD administration did not induce CCN2 expression in islets (Supplemental Fig. S2C). CCN2 expression was efficiently upregulated after 1 wk of DOX treatment and HFD mirroring that of previously published studies from our laboratory (10; Supplemental Fig. S2C).

CCN2 induction in β cells had no effect on HFD-induced weight gain across genotypes (Supplemental Fig. S6). One week or 10 wk of HFD feeding resulted in impaired glucose tolerance compared with controls; there was no effect of CCN2 induction in β cells on HFD-induced impairments in glucose tolerance at either time point (Supplemental Fig. S7). Despite published studies showing that HFD upregulates markers of cell stress such as Atf6 and Eif2s1 in β cells (33), we did not observe a significant increase in 8-OHdG or BiP at either time point in response to HFD (Fig. 3). CCN2 induction in β cells had no effect on levels of 8-OHdG (Fig. 3, A–C) or BiP at either time point (Fig. 3, D–F). There was also no difference in nuclear localization of the active, phosphorylated form of Nrf2 at either time point (Fig. 3, G and H).

Figure 3.

CCN2 induction during HFD does not alter markers of oxidative or ER stress in β cells. A: representative images of 8-OHdG immunolabeling in pancreata after CCN2 induction during 1 wk of HFD. 8-OHdG—red, Ins—green, DAPI—blue. Quantification of percent of β cells expressing 8-OHdG in the nucleus after CCN2 induction during 1 wk of HFD (B) and 10 wk of HFD reveals no significant differences (C). D: representative images of BiP immunolabeling in paraffin-embedded pancreata after CCN2 induction during 10 wk of HFD. BiP—red, Ins—green, DAPI—blue. BiP quantification in β cells after 1 wk of HFD and CCN2 induction (E) and 10 wk of HFD and CCN2 induction reveals no significant differences (F). n = 4–5 animals/group. G: representative images of p-Nrf2 immunolabeling in pancreata after CCN2 induction during 1 wk of HFD. p-Nrf2—red, Ins—green, DAPI—blue. Quantification of nuclear phospho-Nrf2 in insulin+ cells after CCN2 induction during 1 wk of HFD (H) and 10 wk of HFD (I) reveals no significant differences. n = 3/group. Control groups for 1-wk 8-OHdG: CD: 1 RIP-rtTA, 3 TetO-CTGF; HFD: 3 RIP-rtTA; 1 TetO-CTGF. Control groups for 1-wk BiP: CD: 2 RIP-rtTA, 3 TetO-CTGF; HFD: 3 RIP-rtTA, 2 TetO-CTGF. Control groups for 10-wk 8-OHdG: CD: 3 RIP-rtTA, 1 TetO-CTGF; HFD: 2 RIP-rtTA, 2 TetO-CTGF. Control groups for 10-wk BiP: CD: 4 RIP-rTTA; 1 TetO-CTGF; HFD: 3 RIP-rtTA, 2 TetO-CTGF. Control groups for 1-wk p-Nrf2: CD: 1 RIP-rtTA, 2 TetO-CTGF; HFD: 1 RIP-rtTA, 2 TetO-CTGF. Control groups for 10-wk p-Nrf2: CD: 2 RIP-rtTA, 1 TetO-CTGF; HFD: 2 RIP-rtTA, 1 TetO-CTGF. Data were analyzed using one-way ANOVA (scale bar, 50 μm). CD, control diet; ER, endoplasmic reticulum; HFD, high-fat diet.

β-Cell CCN2 Induction in In Vivo Models of Chronic β-Cell Stress Has Differential Effects on β-Cell Mass and Identity

Since previous studies from our laboratory suggested that CCN2-mediated induction of β-cell proliferation occurs only in situations of suboptimal β-cell mass and/or β-cell stress (9, 10), we examined whether CCN2 induction would increase β-cell proliferation and mass in db/+ mice (from 4 to 6 or 6 to 8 wk of age) or HFD-fed mice (1 wk or 10 wk). Two weeks of CCN2 induction had no effect on β-cell proliferation in db/+ mice at either time point (Fig. 4, A and B), and no effect on β-cell mass from 4 to 6 wk of age (Fig. 4C). β-cell mass was significantly decreased in db/+ animals when CCN2 was induced from 6 to 8 wk of age (Fig. 4D). Expression of several genes associated with cell death (Bax, Bad, Bcl-XL, and Casp3) was analyzed from isolated islets from controls (db/+;RIP-rtTA and db/+;TetO-CTGF) and db/+;RIP-rtTA;TetO-CTGF animals treated with DOX from 6 to 8 wk. No difference in expression of any of these genes was observed (Fig. 4E). There was also no change in β-cell size that could explain the decrease in β-cell mass (Fig. 4F). We also examined whether CCN2 induction in β cells affected expression of the critical β-cell identity transcription factor, Pdx1 protein. We found no difference in the percentage of insulin+ cells expressing nuclear Pdx1 when comparing db/+ animals with and without CCN2 induction (Fig. 4G).

Figure 4.

CCN2 induction in db/+ β cells does not induce β-cell proliferation but significantly decreases β-cell mass. β-cell proliferation after 2 wk of CCN2 induction from 4 to 6 wk of age (A) and 6 to 8 wk of age (B). β-cell mass after 2 wk of CCN2 induction from 4 to 6 wk of age (C) and 6 to 8 wk of age (D). n = 5–6 animals/group. E: qRT-PCR analysis of expression of cell death genes after 2 wk of DOX in db/+ control and db/+ animals with CCN2 induction. F: β-cell size in 8-wk-old db/+ animals after 2 wk of CCN2 induction. G: percent Pdx1+/insulin+ cells in 8-wk-old db/+ animals treated with DOX for 2 wk. Controls for 4- to 6-wk proliferation and mass studies: 4 db/+;RIP-rtTA, 2 db/+;TetO-CTGF. Control groups for 6- to 8-wk proliferation and mass studies: 2 db/+;RIP-rtTA, 3 db/+;TetO-CTGF. Control groups for Pdx1 study (6–8 wk): 2 db/+;RIP-rtTA, 1 TetO-CTGF. Control groups for β-cell size (6–8 wk): 2 db/+;RIP-rtTA, 1 TetO-CTGF. Data were analyzed using a Student’s t test. *P < 0.05. DOX, doxycycline.

We observed no difference in β-cell proliferation with CCN2 induction after 1 wk or 10 wk of HFD (Fig. 5, A and B). β-cell mass was unchanged after CCN2 induction during 1 wk of HFD (Fig. 5C), which was not unexpected. However, β-cell mass was significantly increased after 10 wk of CCN2 induction during HFD (Fig. 5D). This could not be explained by an increase in β-cell size, which was unchanged (Fig. 5E). We found that after 1 wk HFD feeding there were no differences across genotypes in Pdx1 nuclear localization in insulin+ cells (Fig. 5F). After 10 wk of HFD, we observed a trend toward reduced numbers of insulin+ cells with Pdx1 in the nucleus. CCN2 induction seems to preserve nuclear Pdx1 in insulin+ cells (Fig. 5G).

Figure 5.

CCN2 induction during HFD significantly increases β-cell mass. β-cell proliferation after 1 wk of CCN2 induction during HFD (A) and 10 wk of CCN2 induction during HFD (B). β-cell mass after 1 wk of CCN2 induction during HFD (C) and 10 wk of CCN2 induction during HFD (D). n = 7–10/group. E: β-cell size after 10 wk of CCN2 induction during HFD. n = 3/group. Percent Pdx1+/insulin+ cells after 1 wk (F) or 10 wk (G) of HFD. Control groups for 1-wk time point proliferation and mass: CD: 2 RIP-rtTA, 8 TetO-CTGF; HFD: 3 RIP-rtTA, 6 TetO-CTGF. Control groups for 10-wk time point proliferation and mass: CD: 5 RIP-rtTA, 2 TetO-CTGF; HFD: 7 RIP-rtTA, 1 TetO-CTGF. Control groups for 1-wk time point Pdx1: CD: 3 TetO-CTGF; HFD: 1 RIP-rtTA, 2 TetO-CTGF. Control groups for 10-wk time point Pdx1: CD: 2 RIP-rtTA, 1 TetO-CTGF; HFD: 1 RIP-rtTA, 2 TetO-CTGF. Data were analyzed using one-way ANOVA and Tukey’s post hoc analysis. *P < 0.05. CD, control diet; HFD, high-fat diet.

DISCUSSION

Our group previously showed that CCN2 could induce adult mouse β-cell proliferation under euglycemic conditions in the setting of increased metabolic demand, including during pregnancy and after fifty percent β-cell ablation (9, 10). However, CCN2 induction in β cells in euglycemic adult mice in the absence of any physiological stressor or increased insulin demand fails to induce β-cell proliferation (10, 11). Thus, we concluded that β cells are only able to respond to CCN2 in settings of suboptimal β-cell mass or increased β-cell stress. In the current study, we investigated the potential of CCN2 to induce β-cell proliferation and modulate β-cell stress using in vivo and ex vivo models of β-cell stress. β-cell CCN2 induction had a specific effect on expression of Nfe2l2, a master regulator of genes that ameliorate oxidative stress, only in the setting of simultaneous thapsigargin treatment. β-cell CCN2 induction had no effect on markers of cell stress in mouse models of chronic β-cell stress in vivo (leptin receptor heterozygosity and HFD feeding), at least as assessed by BiP expression, nuclear 8-OHdG, and active Nrf2. CCN2 induction specifically in β cells stimulated β-cell mass expansion in the setting of prolonged, but not short term, HFD feeding (associated with insulin resistance), but failed to induce β-cell proliferation or increase β-cell mass in db/+ mice.

Despite the detection of increased cellular stress in islets of db/+ mice (Ref. 12 and this study), induction of CCN2 in β cells of db/+ mice had no effect on β-cell proliferation and actually led to a significant decrease in β-cell mass when induced from 6 to 8 wk of age. The mechanism for this decrease in β-cell mass has yet to be determined as β-cell size was unchanged and assays for markers of β-cell death were negative (qRT-PCR) or inconclusive (TUNEL and cleaved caspase; data not shown). It is possible that the β-cell microenvironment in the setting of leptin receptor heterozygosity differs from that present in the setting of prolonged HFD, altering β-cell responsiveness to CCN2 induction. Alternatively, CCN2 induction in β cells under the strong insulin promoter may result in some β-cell stress itself, although we did not observe any increases in 8-OHdG or BiP levels in β cells of db/+ mice with CCN2 induction compared with control animals. However, in ex vivo studies we found that β-cell CCN2 induction leads to increased expression of Nfe2l2, the gene encoding the antioxidant master transcriptional regulator Nrf2 (25). Nrf2 is critical for β-cell proliferation in other settings: overexpression of Nrf2 induces β-cell proliferation in both mouse and human islets (34), and loss of Nrf2 causes a reduction in glucose-stimulated β-cell proliferation and HFD-induced β-cell mass expansion (35). Activation of Nfe2l2/Nrf2 by CCN2 induction suggests that Nrf2 plays a role in CCN2-induced β-cell proliferation. It is possible that if we had analyzed markers of cellular stress at a later time point, we would have detected a decrease in markers of oxidative stress due to the CCN2-mediated increase in Nrf2.

Previous studies utilizing wild-type C57BL/6J mice reported a significant increase in β-cell proliferation after 1 wk of HFD (32). However, we did not observe an effect of HFD on β-cell proliferation at 1 wk or 10 wk in the current study. The RIP-rtTA;TetO-CTGF mouse model is maintained on a mixed genetic background, which may alter the HFD-induced compensatory response. Furthermore, it is unknown if the presence of DOX alters β-cell responsiveness to HFD. Despite this, we did observe an increase in β-cell mass with CCN2 induction in the setting of prolonged HFD feeding. It is possible that CCN2 increased β-cell proliferation at an earlier time point that was not assessed in these studies. The current study demonstrates that although CCN2 is not induced in adult β cells in response to HFD feeding, it can further enhance HFD-mediated β-cell mass expansion when induced in β cells. However, the lack of increase in β-cell proliferation following CCN2 induction in β cells from db/+ mice indicates that increased β-cell stress alone is necessary but not sufficient to render β cells responsive to CCN2. Further studies will be required to determine how the islet microenvironment or β-cell phenotype alters CCN2 responsiveness allowing for increased β-cell proliferation to increase functional β-cell mass.

DATA AVAILABILITY

Data will be made available upon reasonable request.

SUPPLEMENTAL DATA

Supplemental Methods: https://figshare.com/s/333dc153fea14de34f53.

Supplemental Fig. S1: https://doi.org/10.6084/m9.figshare.23264354.

Supplemental Fig. S2: https://doi.org/10.6084/m9.figshare.23264435.

Supplemental Fig. S3: https://doi.org/10.6084/m9.figshare.23264651.

Supplemental Fig. S4: https://doi.org/10.6084/m9.figshare.23264660.

Supplemental Fig. S5: https://doi.org/10.6084/m9.figshare.23264762.

Supplemental Fig. S6: https://doi.org/10.6084/m9.figshare.23266952.

Supplemental Fig. S7: https://doi.org/10.6084/m9.figshare.23304143.

GRANTS

S.E.T. was supported in part by the Vanderbilt University Training Program in Molecular Endocrinology (5T32 DK7563-30). M.G. was supported by grants from the American Diabetes Association (1-16-IBS-100) and the National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health (R01 DK105689) and by VA Merit Awards (I01 BX003744-01 and I01 BX005399). The Islet Procurement and Analysis Core of the Vanderbilt Diabetes Research and Training Center was supported by the National Institutes of Health (Grant DK-20593).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

S.E.T. and M.G. conceived and designed research; S.E.T. and J.D.F. performed experiments; S.E.T., J.D.F., and M.G. analyzed data; S.E.T. and M.G. interpreted results of experiments; S.E.T. prepared figures; S.E.T. drafted manuscript; S.E.T. and M.G. edited and revised manuscript; S.E.T., J.D.F., and M.G. approved final version of manuscript.