Reduced Expression Level of Protein Phosphatase PPM1E Serves to Maintain Insulin Secretion in Type 2 Diabetes

The prevalence of type 2 diabetes (T2D) is increasing worldwide (1,2). Recent large-scale genetics analyses have indicated that dysfunction of the insulin-secreting β-cell is the main culprit in this development (reviewed by Meier and Bonadonna [3] and Zhang et al. [4]). Reversible phosphorylation is arguably one of the most important mechanisms for regulating the biological activity of many intracellular proteins. Enzymatic activities of protein kinases and protein phosphorylation patterns change after stimulation of β-cell insulin release by nutrients (5–7). As a consequence, phosphorylation by kinases involved in the insulin secretory process have been widely investigated. Compared with their kinase counterparts, much less is known about the role of protein phosphatases in biological systems. This can be partly explained by technical difficulties related to accurately measuring protein dephosphorylation but also by a widely held view that phosphatases simply act as “housekeeping” enzymes (5). We now know that phosphatases are not merely housekeeping enzymes but exert specific actions and can even play a major role in regulation of physiological processes.

Short-term treatment of β-cells with the protein phosphatase inhibitor okadaic acid promotes Ca²⁺ entry and insulin secretion (8). Glycolytic and Krebs cycle intermediates, whose concentrations are increased after glucose stimulation, inhibit the enzymatic activity of phosphoserine/threonine phosphatases in β-cells (9). These findings point to a crucial role for protein (de)phosphorylation in regulating stimulus-secretion coupling in β-cells.

In this study, we measured expression levels of the type 2C family of protein phosphatases (PP2C), the subfamily of metal-dependent protein phosphatase (PPM) in human pancreatic islets; and we observed significant up- and downregulation of subsets of PP2C genes in islets from T2D donors. We focused our functional analyses on the PP2C gene that was most extensively altered, PPM1E, and proceeded to examine its functional role in β-cells and also whether its downregulation was a cause or consequence of T2D.

Human islets were obtained from the Nordic Network for Islet Transplantation at Uppsala University, Sweden, via the Human Tissue Laboratory at Lund University Diabetes Centre. The clinical characteristics of the islet donors used for assessing mRNA expression or insulin secretion are shown in Supplementary Tables 1 and 2 and Supplementary Fig. 1. All procedures using human islets were approved by the ethical committees in Uppsala and Malmö/Lund, Sweden (approval no. 2019-00357). mRNA expression of the PP2C family phosphatases in human islets was quantified by RNA sequencing, using a TruSeq RNA sample preparation kit (Illumina) (Supplementary Research Design and Methods).

INS-1 832/13 cells were cultured in RPMI 1640 medium at 37°C with 5% CO₂. In additional experiments, INS-1 832/13 cells were cultured with medium containing 16.7 mmol/L glucose and 0.15 mmol/L palmitate for 24 h or 48 h with or without addition of metformin. Male Wistar rats (aged 12 weeks) were housed at a designated animal facility under controlled conditions (temperature 21 ± 1°C, humidity 55%, 12-h light-dark cycle) with ad libitum access to food and water. Rat islets were isolated as previously described (10) and cultured in RPMI 1640 with GlutaMax supplemented with 10% FCS (Gibco-BRL). The protocols for animal care and use were approved by the ethics committee in Malmö/Lund, Sweden (2931/20).

Total mRNA was extracted using an RNA purification kit and reversely transcribed using RevertAid First-Strand cDNA synthesis kit. qRT-PCR was performed using TaqMan gene expression assays for both PPM1E and the housekeeping gene, HPRT1. Gene expression was quantified by the comparative cycle threshold value, in which the amount of target is expressed as 2^−ΔΔCt using HPRT1 as a reference gene (Supplementary Research Design and Methods).

For silencing of PPM1E expression, INS-1 832/13 cells were transfected by Lipofectamine RNAiMAX Reagent and 10 nmol/L small interfering RNAs (siRNAs) specific for rat PPM1E mRNA or a scrambled siRNA (as the negative control). For human or rat islets, 200–300 islets were transfected by Lipofectamine RNAiMAX Reagent and 50 nmol/L human or rat PPM1E siRNA for 24 h. The day after the first transfection, the islets were transferred into new plates and retransfected for another 24 h. For PPM1E overexpression, INS-1 832/13 cells were transfected by Lipofectamine 3000 transfection Reagent and PPM1E expression plasmid. All functional experiments were performed 72 h after the transfection in cells or islets (Supplementary Research Design and Methods).

INS-1 832/13 cells were preincubated in secretion assay buffer (SAB) containing 2.8 mmol/L glucose for 2 h. Cells were stimulated for 1 h with SAB containing different insulin secretagogues (16.7 mmol/L glucose, 35 mmol/L KCl, 1 µmol/L forskolin, 10 mmol/L pyruvate) at 37°C. Secreted insulin levels in the supernatant were measured using rat insulin ELISA and the values were normalized to the total protein content extracted using RIPA buffer and measured with BCA Protein Assay Kit.

Four replicates of eight human islets per culture condition and donor were preincubated in SAB supplemented with 1 mmol/L glucose for 1 h and then stimulated with 20 mmol/L glucose, 35 mmol/L KCl, 100 μmol/L isobutylmethylxanthine (IBMX) for 1 h. Secreted insulin levels in the supernatant and also total insulin content were measured by human insulin ELISA kit; values were normalized to the total insulin content and total protein content, respectively (Supplementary Research Design and Methods).

The Abcam cell-ELISA protocol was used to quantify protein expression of PPM1E, phospho-Ca²⁺/calmodulin-dependent protein kinase type 2 (pCaMKII), phospho-AMP–dependent protein kinase (pAMPKα), and phospho-acetyl-CoA carboxylase (pACC) in cultured cells (Supplementary Research Design and Methods).

Mitochondrial oxygen consumption rate (OCR) was determined with the Seahorse Extracellular Flux Analyzer XF24 (Seahorse Bioscience, Billerica, MA). INS-1 832/13 cells were kept at 2.8 mmol/L glucose for 2 h and OCR was measured every 3 min for 90 min in the presence of glucose (2.8 and 16.7 mmol/L), IBMX (100 μmol/L), and, subsequently, oligomycin (4 μmol/L), mitochondrial inner membrane ionophore (carbonyl cyanide p-trifluoro-methoxyphenyl hydrazone [FCCP]; 4 μmol/L) and rotenone/antimycin A (1 μmol/L). Data were analyzed using Seahorse wave software (Agilent, Santa Clara, CA).

Mitochondrial membrane potential (ΔΨm) was measured using the fluorescent dye tetramethylrhodamine, methyl ester (TMRM) and confocal microscopy in quench mode, whereby ΔΨm is decreased after a transient increase in the whole-cell or islet fluorescence (Supplementary Research Design and Methods).

The single-cell cytosolic ATP to ADP ratio was measured by the genetically encoded biosensor Perceval HR and pHRed. Cytoplasmic-free Ca²⁺ levels were measured using the fluorescent dye Fluo4 AM and confocal microscopy in the INS-1 832/13 cells and rat islets (11) (Supplementary Research Design and Methods).

Cell viability was measured using RealTime-Glo MT Cell Viability Assay kit (Promega, Catalog G9711) according to the manufacturer’s instructions.

Total glutathione (GSH) and oxidized glutathione (GSSG) concentrations in cells were assessed using a glutathione colorimetric detection kit. Total reactive oxygen species (ROS) levels were measured by a total ROS assay kit and analyzed using a CytoFLEX flow cytometer (Supplementary Research Design and Methods).

The data sets generated and/or analyzed during this study are available from the corresponding author upon reasonable request.

Analyses were conducted with GraphPad Prism software (version 9) and all values are presented as mean ± SEM. Student t test and one-way ANOVA followed by Tukey post hoc test were used for single or multiple comparisons, respectively. A significance level of two-sided P values < 0.05 was considered statistically significant.

We first investigated whether expression of the PP2C family phosphatases was altered in islets from T2D donors. As shown in Fig. 1, mRNA expression of PPM1E (P = 0.004), PPM1H (P = 0.007), PDP2 (P = 0.013), and PPM1K (P = 0.036) was significantly reduced in islets from T2D donors, whereas expression of PPM1G (P = 0.003) was significantly increased as compared with islets from donors without diabetes (ND). In addition, expression of PPM1L (P = 0.078) and PDP1 was marginally (P = 0.054) lower and expression of PPM1F was marginally (P = 0.058) higher in islets from T2D donors, with no difference in expression levels of PPM1A, PPM1B, PPM1D, PPM1M, PHLPP1, PHLPP2, PPTC7, TAB1, and ILKAP, comparing expression from ND and T2D islet donors. Expression of PPM1J was undetectable in islets.

Of the PP2C family, PPM1E was the most extensively downregulated gene in the islets from T2D donors (Figs. 1D and 2A). Downregulation of PPM1E in the islets of patients with T2D has also been reported in other studies (12,13). PPM1E is a ser/thr protein phosphatase and is insensitive to inhibitors, such as okadaic acid, that inhibit the majority of phosphoprotein phosphatases (14). β-Cells express several types of protein phosphatases. Because reversible protein phosphorylation is a critical process in the regulation of insulin secretion, protein phosphatases are crucial for the control of the cellular phosphorylation state. Here, we chose PPM1E for detailed analysis, with the goal of determining its role in β-cell function.

To determine whether expression of islet PPM1E was associated with glycemic control and metabolic status, the correlation between islet PPM1E mRNA levels and donor HbA_1c (Fig. 2B) and BMI (Fig. 2C) was analyzed. There was a negative correlation between PPM1E mRNA levels, HbA_1c (P = 0.001), and BMI (P = 0.004), demonstrating that an impairment of glycemic and metabolic control is associated with a progressive decrease of PPM1E expression in β-cells.

To understand the role of PPM1E in β-cell function, we overexpressed or silenced PPM1E in INS-1 832/13 cells and human and rat islets. Treatment of cells (Supplementary Fig. 2A and B) and rat (Supplementary Fig. 2C) and human (Supplementary Fig. 2D) islets with siRNAs or a PPM1E-expressing plasmid (Supplementary Fig. 2E) for 72 h was sufficient to effectively reduce or increase the expression of PPM1E. Once these models were established, we studied the consequences for β-cell function.

To gain insight into the effect of PPM1E on insulin secretion, we evaluated insulin secretion in response to glucose, pyruvate, KCl, and forskolin in PPM1E knockdown INS-1 832/13 cells. Reduced PPM1E expression significantly increased pyruvate-, glucose-, KCl-, and forskolin-stimulated insulin secretion after 15 min (Supplementary Fig. 3A–H) or 60 min (Fig. 3A–D and Supplementary Figs. 4A–D and 5) of treatment with secretagogues. PPM1E knockdown, however, did not have any effect on insulin content in INS-1 832/13 cells (Supplementary Fig. 4E and F). Conversely, PPM1E overexpression decreased glucose-stimulated insulin secretion (GSIS) in INS-1 832/13 cells (Fig. 3H and Supplementary Fig. 4G). Together, these results demonstrate an inhibitory effect of PPM1E on insulin secretion.

Treatment of INS-1 832/13 cells with 16.7 mmol/L glucose and 0.15 mmol/L palmitate for 24 h or 48 h creates a glucolipotoxic milieu that mimics pathogenic processes reminiscent of what β-cells are exposed to in T2D. Under these conditions, we measured expression of PPM1E at the mRNA (Fig. 4A) and protein (Fig. 4B) levels as well as GSIS (Fig. 4C). Compared with untreated cells, exposure to high levels of glucose and palmitate for 24 h or 48 h led to an inhibition of GSIS in response to 16.7 mmol/L of glucose and a reduction in mRNA expression of PPM1E; protein expression of PPM1E decreased after 48 h, but not 24 h, of treatment with high levels of glucose and palmitate.

Because PPM1E overexpression resulted in an inhibition of insulin secretion, we examined whether silencing of PPM1E would ameliorate decreased insulin secretion when β-cells were treated with high glucose and palmitate levels for 48 h (Fig. 4D). Compared with high levels of glucose and palmitate treatment alone (control β-cells), PPM1E knockdown resulted in increased GSIS when cells were exposed to high glucose and palmitate levels. All these results could be interpreted so that reduced expression of PPM1E is a consequence of T2D, being a compensatory response of β-cells to increase insulin secretion. To functionally elucidate this, we silenced PPM1E in human islets from T2D donors. Interestingly, GSIS (Fig. 3E) and KCl-promoted (Fig. 3F) or IBMX-promoted (Fig. 3G) GSIS were increased in PPM1E knockdown human islets from T2D donors. Similarly, PPM1E knockdown did not have any effect on insulin content in human islets when compared with islets treated with scrambled siRNA (Supplementary Fig. 4H).

Mitochondria play a primary role in β-cell energy metabolism and, thereby, insulin secretion (15). In addition to the transfer of electrons, the mitochondrial electron transport chain generates an electrochemical gradient across the mitochondrial inner membrane (i.e., the ΔΨm) by accumulating H⁺ in the intermembrane space; this is integral to the generation of ATP (15). To gain more insight into how PPM1E inhibits insulin secretion, we examined the effects of PPM1E knockdown on mitochondrial function. We monitored the relative changes in ΔΨm with TMRM in quench mode in INS-1 832/13 cells and rat islets (Fig. 5A and B and Supplementary Figs. 6 and 7). An increase in glucose concentration from 2.8 to 16.7 mmol/L hyperpolarized the inner mitochondrial membrane of INS-1 832/13 cells and rat islets; this response increased after silencing of PPM1E. The hyperpolarization was further enhanced in response to the ATP synthase inhibitor oligomycin but decreased upon treatment with the mitochondrial uncoupling agent FCCP. It therefore appears plausible that increased insulin secretion after silencing of PPM1E was caused, at least in part, by the exaggerated hyperpolarization of the inner mitochondrial membrane after glucose stimulation.

The translocation of protons and the electrochemical gradient over the inner mitochondrial membrane are coupled to the phosphorylation of ADP to generate ATP through ATP synthase (15). Thus, an increased ΔΨm, elicited by PPM1E knockdown, could potentially lead to an increase in mitochondrial ATP production. We therefore measured the cytosolic ATP to ADP ratio, using the genetically encoded PercevalHR probe in PPM1E knockdown INS-1 832/13 cells (Fig. 5C and Supplementary Fig. 8). We found that PPM1E knockdown significantly increased the elevation in the cytosolic ATP to ADP ratio upon stimulation with 16.7 mmol/L glucose in β-cells.

Because ATP in β-cells is largely generated through mitochondrial oxidative phosphorylation (OXPHOS) (16), increased ΔΨm and cytosolic ATP to ADP ratio in PPM1E knockdown β-cells could be due to the increased mitochondrial OXPHOS. To further address this, we examined OCR using an extracellular flux analyzer (Fig. 5D and Supplementary Fig. 9). In scramble control cells, we observed an increase in OCR in response to glucose, whereas OCR response decreased in response to respiratory complex inhibitors (oligomycin and antimycin plus rotenone) and was provoked to maximal capacity by adding FCCP. Surprisingly, despite the increased ΔΨm and cytosolic ATP to ADP ratio after PPM1E knockdown, basal mitochondrial OCR or maximal respiratory capacity was not affected by PPM1E knockdown. Thus, the effects of PPM1E on the ΔΨm and cytosolic ATP to ADP ratio are uncoupled from OXPHOS.

Insulin secretion is associated with transient changes in cytoplasmic free Ca²⁺ levels (16). Therefore, we examined whether PPM1E also influences cytoplasmic free Ca²⁺ levels in β-cells and rat islets (Fig. 5E and F and Supplementary Figs. 10 and 11). An increase in glucose concentration from 2.8 to 16.7 mmol/L increased cytoplasmic free Ca²⁺ levels in control cells and rat islets. PPM1E knockdown cells or islets showed enhanced cytoplasmic free Ca²⁺ levels in response to glucose. However, because this could be the result of higher ATP to ADP ratios rather than a direct effect of PPM1E on cytoplasmic Ca²⁺ dynamics, the cells and islets were then depolarized by addition of 30 mmol/L KCl. Surprisingly, silencing of PPM1E led to an increase in the maximal cytoplasmic Ca²⁺ elevation upon 30 mmol/L KCl stimulation in INS-1 832/13 cells (Fig. 5E) or rat islets (Fig. 5F). We also measured plasma membrane potential using a FLIPR Membrane Potential Assay Kit in single PPM1E knockdown β-cells. Although there was a trend toward an increase in plasma membrane potential in PPM1E knockdown INS-1 832/13 cells, the increase did not reach statistical significance (data not shown), demonstrating that increased cytosolic Ca²⁺ level in PPM1E knockdown cells is independent of Ca²⁺ entry through voltage-dependent Ca²⁺ channels. However, additional studies using the patch-clamp technique should be used to confirm Ca²⁺ channel conductance.

The endoplasmic reticulum (ER) is the main internal Ca²⁺ store. Therefore, the observed elevated, cytoplasmic free Ca²⁺ levels could be a result of the release from ER. To understand the contribution of the ER in increased cytoplasmic free Ca²⁺ levels after PPM1E knockdown, we quantified Ca²⁺_ER indirectly in control and PPM1E knockdown β-cells by measuring cytoplasmic free Ca²⁺ levels in response to 30 mmol/L KCl in the cells incubated with 5 μmol/L of the sarco-endoplasmic Ca²⁺ ATPase inhibitor thapsigargin 10 min before running the experiment. Subtraction of the Ca²⁺ signal obtained from thapsigargin-pretreated cells from the Ca²⁺ signal of untreated cells revealed modulation of Ca²⁺_ER homeostasis by PPM1E (Fig. 5G–I). The observation of elevated cytoplasmic free Ca²⁺ levels after PPM1E knockdown is in accordance with previous studies reporting that phosphatase inhibition may raise intracellular Ca²⁺ levels through enhancing the phosphorylation level of proteins involved in Ca²⁺ transport, such as ryanodine receptor, phospholamban, and plasma membrane Ca²⁺ channels (17,18).

PPM1E was first identified as a CaMK phosphatase; it dephosphorylates phospho-threonine residues involved in the activation of CaMKI, CaMKII, and CaMKIV (19). CaMKs are involved in regulation of insulin secretion and β-cell mass (20,21). Inhibition of CaMKII in β-cells decreases GSIS and glucose-stimulated Ca²⁺ entry via voltage-dependent Ca²⁺ channels and suppresses glucose-stimulated action-potential firing frequency (20). Later, PPM1E also was shown to be a primary AMPK phosphatase (22,23). Dysregulation of AMPK, the master regulator of energy metabolism in cells, is associated with metabolic disorders, insulin resistance, and T2D (24). Because activation of CaMKII and AMPK is implicated in T2D, and also because PPM1E has an inhibitory effect on these kinases, increased insulin secretion after PPM1E knockdown could be related to increased phosphorylation and activation of these kinases. Thus, we measured phosphorylation levels of CaMKII (Fig. 6A), AMPK (Fig. 6B), and acetyl-CoA carboxylase (ACC) (Fig. 6C), a bona fide substrate of AMPK and a marker of AMPK activation (25). Phosphorylation levels of CaMKII, AMPK, and ACC were largely enhanced in the PPM1E knockdown β-cells. Thus, downregulation of PPM1E in T2D is predicted to increase activity of CaMKII and AMPK and in promoting insulin secretion.

A growing body of data shows that activation of AMPK is the mechanism by which metformin and phenformin exert their beneficial effects in treatment of T2D. Metformin-induced inhibition of inflammation (26), lipid accumulation (27), and hepatic gluconeogenesis (28), as well as increased muscle glucose uptake (29), are mediated through activation of AMPK. Given activation of AMPK by metformin and the potential inhibitory effect of PPM1E on AMPK, we hypothesized that the increased activation of AMPK upon metformin treatment is a consequence of downregulation of PPM1E by metformin in β-cells. To elucidate this, expression of pAMPK (Fig. 7A), pACC (Fig. 7B), and PPM1E (Fig. 7C) was measured in INS-1 832/13 cells treated with 16.7 mmol/L glucose and 0.15 mmol/L palmitate for 48 h in the presence of metformin. Treatment with high glucose and palmitate levels decreased expression of PPM1E, pAMPK, and pACC in β-cells. Metformin treatment for 48 h (Fig. 7A–C), but not 24 h (Supplementary Fig. 12), reduced protein expression of PPM1E (Fig. 7C), whereas levels of pAMPK (Fig. 7A) and pACC (Fig. 7B) increased in β-cells treated with high levels of glucose and palmitate. This finding consistent with those of another study reporting that phenformin, through inhibition of PPM1E activity, increases AMPK phosphorylation in HEK293 cells (22). It should be noted that metformin also acts through AMPK-independent pathways; it reduces palmitate-induced elevation of ROS and ER stress markers in an AMPK-independent manner in β-cells (30). Because PPM1E also influenced Ca²⁺_ER homeostasis, it may also be involved in AMPK-independent effects of metformin.

Given the inhibitory effect of PPM1E on insulin secretion and the observed attenuation of PPM1E expression after metformin treatment, we examined the effect of metformin on insulin secretion in the cells treated with 16.7 mmol/L glucose and 0.15 mmol/L palmitate. Chronic (24 h) (Fig. 7D), but not acute (1 h) (Supplementary Fig. 13A–C), metformin treatment slightly increased GSIS in β-cells treated with high levels of glucose and palmitate, whereas no significant effects on GSIS were observed in control β-cells. The effect of metformin on GSIS is controversial because stimulatory (31), no change (32), or even inhibitory (33) effects have been reported.

Another beneficial effect of metformin is its capacity to increase cell viability. Thus, in the next set of experiments, we examined the effect of metformin on cell viability in control or PPM1E knockdown cells treated with high levels of glucose and palmitate (Fig. 7E). Interestingly, metformin treatment for 24 h or 48 h decreased cell viability under normal cell culture conditions while it protected β-cells against cytotoxicity induced by high levels of glucose and palmitate. Indeed, metformin triggers autophagy by AMPK activation and subsequent inhibition of mTOR (34). Although moderate autophagy protects against cellular tumorigenesis and high-fat diet–induced cell death (34,35), its overactivation may cause excessive digestion of essential cytoplasmic constituents, leading to cell death (36). PPM1E knockdown did not have any effect on cell viability under normal cell culture conditions but protected β-cells against cytotoxicity induced by high levels of glucose and palmitate for 24 h but not 48 h. However, compared with the PPM1E knockdown β-cells treated with high glucose and palmitate levels, metformin treatment, did not significantly affect cell cytotoxicity, demonstrating that metformin-induced cell protection against cytotoxicity induced by high levels of glucose and palmitate, at least in part, is due to inhibition of PPM1E.

Oxidative stress plays a pivotal role in the impairment of β-cell function. Because of a low level of redox buffering capacity, compared with other cell types, β-cells are highly sensitive to oxidative stress (37). Under diabetic conditions, chronic hyperglycemia provokes ROS formation, which deteriorates β-cell function (38). Because β-cell function was improved upon PPM1E knockdown, we hypothesized that PPM1E knockdown can protect β-cells against oxidative stress. To this end, we treated control and PPM1E knockdown cells with 16.7 mmol/L glucose and 0.15 mmol/L palmitate for 24 h or 48 h and determined the GSH to GSSG ratio (Fig. 8A and B) as well as ROS levels (Fig. 8C and Supplementary Fig. 14).

Treatment with high levels of glucose and palmitate for 24 h (Fig. 8A) or 48 h (Fig. 8B) significantly reduced the GSH to GSSG ratio in INS-1 832/13 cells. PPM1E knockdown had no effect on the GSH to GSSG ratio under normal cell culture conditions but resulted in a marginal increase in the GSH to GSSG ratio after 24 h, but not 48 h, of treatment with 16.7 mmol/L glucose and 0.15 mmol/L palmitate in INS-1 832/13 cells. Flow cytometric analysis revealed a significant increase in ROS production after 48 h (Fig. 8C and Supplementary Fig. 14) but not 24 h (not shown) after β-cell treatment with high levels of glucose and palmitate. PPM1E knockdown, on the other hand, resulted in reduced ROS levels when cells were treated with high levels of glucose and palmitate (Fig. 8C and Supplementary Fig. 14).

Decreased oxidative stress can be attributed to activation of AMPK (p-AMPK/p-ACC) after silencing of PPM1E. Antioxidative effects of AMPK are mediated by increases in NADPH and prevention of ATP depletion (39). Indeed, AMPK blocks hyperglycemia-induced PKC activation, which is responsible for NADPH oxidase activation and, consequently, ROS formation (40).

This study reports the expression levels of PP2C phosphatases in islets of patients with T2D and suggests a functional role of PPM1E in β-cells, given that PPM1E is the most extensively downregulated PP2C phosphatase in T2D. We showed that PPM1E exerts an inhibitory effect on insulin secretion and is downregulated in human T2D islets, most likely as a compensatory response of β-cells to uphold insulin secretion under metabolic stress.

Although PPM1E knockdown increased hyperpolarization of the inner ΔΨm and cytoplasmic ATP to ADP ratio after glucose stimulation in β-cells, we did not observe an effect on mitochondrial OCR. It should be noted that a change in the cellular ATP to ADP ratio does not entirely reflect mitochondrial function. Cellular ATP levels can be altered in many different situations that affect the total adenine nucleotide pool size but not necessarily bioenergetic status (41). In this context, PPM1E knockdown increased phosphorylation and activation of AMPK, which is activated by cellular stress; once activated, the system stimulates catabolic pathways that increase ATP production, while switching off ATP-consuming processes (42). Thus, an increased cellular ATP to ADP ratio without any change in mitochondrial OCR can be attributed to increased AMPK activity and inhibition of ATP-consumption after PPM1E knockdown.

We found that PPM1E knockdown increased cytosolic Ca²⁺ levels in β-cells. Increases in cytosolic Ca²⁺ levels directly stimulate insulin granule exocytosis but also initiate a broad range of cellular responses, resulting in (de)phosphorylation events and activation of downstream transcription factors (43). Here, we observed increased phosphorylation of CaMKII in PPM1E knockdown β-cells; CaMKII is a Ca²⁺ sensor, playing a role in Ca²⁺-induced insulin secretion. CaMKII is also known to phosphorylate ryanodine receptor 2 (RyR2), which is located in the ER membrane and is responsible for the Ca²⁺ release from ER (18). Because PPM1E has an inhibitory effect on CaMKII, elevated ER Ca²⁺ release after PPM1E knockdown could be due to increased activation of CaMKII and phosphorylation of RyR2. In the β-cells, RyR2 maintains insulin content and secretion and regulates the proteome, most likely via DNA methylation (44). Hence, increased GSIS after PPM1E knockdown may, at least in part, be due to increased phosphorylation and activation of CaMKII. Inhibition of CaMKII after PPM1E activation has also been reported in mouse neuroblastoma Neuro2a cells and a wheat-embryo cell-free protein expression system (45,46).

Our data showed that overall expression of PP2C phosphatases is changed in T2D. Given that PPM1E was most extensively downregulated in T2D islets, we chose this phosphatase for further studies. Our data, although largely supportive of this notion, do not exclude a role for other PP2C phosphatases in β-cells and T2D. Reduced expression of PPM1E can be interpreted as a compensatory response of β-cells to increase insulin secretion when metabolically challenged. Establishing a causal relationship in T2D, however, would require further extensive studies involving regulation at the transcriptional and translational levels, as well as posttranslational control of PPM1E activity and counterregulatory processes, both in vitro and in vivo. Nevertheless, favoring a compensatory role of this phosphatase, reduced expression of PPM1E stimulated insulin secretion through elevated cytosolic Ca²⁺ levels and ATP to ADP ratio; increased hyperpolarization of the inner ΔΨm; enhanced phosphorylation of CaMKII, AMPK, and ACC; decreased cellular ROS formation; and increased cell viability. Given this capacity for functional compensation, modulation of phosphatase expression and activity in general, and of PPM1E in particular, may have potential therapeutic effects in the preservation and/or restoration of β-cell function in patients with T2D.

Funding. This study was supported by grants from the Swedish Research Council (2021-01777), the Novo Nordisk, Hjelt, Stiftelsen Lars Hiertas Minne, and Albert Påhlsson’s Foundations, and the Royal Physiographic Society of Lund.

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

Author Contributions. S.G. contributed to the study conceptualization, methodology, investigation, writing the original draft, review and editing of the manuscript, and visualization. L.R.C., A.H., M.H., and S.K. contributed to the study methodology and investigation. M.F. and H.M. contributed to the study conceptualization, methodology, resources, reviewing and editing the manuscript, supervision, and funding acquisition. S.G. 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.

Prior Presentation. Preliminary results of this study were presented at the virtual 57th Annual Meeting of the European Association for the Study of Diabetes, 27 September to October 2021.