Acute hyperglycemia induces podocyte apoptosis by monocyte TNF-α release, a process attenuated by vitamin D and GLP-1 receptor agonists

Abstract

Targeting optimal glycemic control based on hemoglobin A1c (A1c) values reduces but does not abolish the onset of diabetic kidney disease and its progression to chronic kidney disease (CKD). This suggests that factors other than the average glucose contribute to the residual risk. Vitamin D deficiency and frequent episodes of acute hyperglycemia (AH) are associated with the onset of albuminuria and CKD progression in diabetes. This study aimed to determine if moderate levels of AH harm podocytes directly or promote a pro-inflammatory monocyte/macrophage phenotype that leads to podocyte apoptosis, and whether vitamin D deficiency accelerates these processes. We found that AH (16.7 mM D-glucose) didn’t induce podocyte apoptosis directly, but it did promote a pro-inflammatory response in human monocytes and macrophages, resulting in an increased TNF-α secretion causing podocyte apoptosis. The AH-induced monocyte TNF-α secretion was inversely correlated with healthy donors’ serum 25(OH)D levels. AH induced monocyte TNF-α release by increasing oxidative and ER stress, which in turn increased ADAM17 (A Disintegrin And Metalloprotease 17) and iRhom2 (inactive Rhomboid protein 2) expression, both essential for TNF-α secretion. Additionally, monocyte activation of glucagon-like peptide-1 receptor (GLP-1R), using a GLP-1R agonist, downregulated ADAM17/iRhom2 expression, decreasing TNF-α release and reducing podocyte apoptosis. These results show that a normal vitamin D status may attenuate a mechanism by which AH contributes to podocyte apoptosis and CKD progression and might enhance a novel anti-inflammatory role of GLP-1 to prevent AH-driven CKD progression in diabetes.

1. Introduction

Diabetes is the leading cause of end-stage renal disease (ESRD) in the United States and is a significant predictor of premature death [1,2]. Nearly half of patients with type 2 diabetes mellitus (T2DM) and 30 % of those with type 1 diabetes (T1DM) develop chronic kidney disease (CKD) [3,4]. Hyperglycemia is a major risk factor for CKD, and intensive glucose control reduces CKD risk. However, the traditional paradigm for glycemic control targeting hemoglobin A1c (A1c) levels has been reevaluated, particularly in patients with CKD [5,6]. Multiple studies indicate that oscillations of glucose during the day, known as glycemic variability, are associated with the development of microalbuminuria [7] and the decline in estimated glomerular filtration rates independently of A1C [8]. Therefore, it is essential to understand the mechanisms by which episodes of acute hyperglycemia contribute to CKD progression.

Diabetes can affect numerous cell types in the kidney [9]. Reduced podocyte numbers correlate strongly with albuminuria, and their loss represents an irreversible event leading to a decline in glomerular barrier function [10]. Hyperglycemia can be toxic to podocytes; in animal models of T2DM or T1DM, severe and prolonged hyperglycemia causes podocyte apoptosis [11,12]. In vitro studies demonstrate that severe and prolonged hyperglycemia (540 mg/dl) induced reactive oxygen species (ROS) generation, causing murine podocyte apoptosis [13]. Severe hyperglycemia increases podocyte apoptosis within 3 h, but lower glucose levels (360 mg/dl) require 18 h to induce caspase activity, suggesting that other glucose-related mechanisms, independent of direct glucose-induced podocyte apoptosis, are present at low glucose levels.

In human and mouse models of diabetes, chronic hyperglycemia contributes to the progression of renal impairment by increasing M1 macrophage glomeruli infiltration, causing podocyte apoptosis [14–16]. In mouse models of T1 or T2 diabetes, depletion of macrophage by clodronate liposomes or diphtheria on CD11DTR mice ameliorates chronic hyperglycemia-induced podocyte injury [17,18]. Interestingly, shorter periods of hyperglycemia also activate innate immunity [19]. In vitro, studies suggested that 12 h intermittent glucose fluctuations promote macrophage polarization into the inflammatory M1 phenotype with increased CD11c, IL-1β, TNF-α, IL-6, and monocyte chemo-attractant protein-1 expression, and induced TNF-α secretion to promote podocyte apoptosis [20,21]. In normal subjects, moderate hyperglycemia (195 mg/dl) for 6 h during a hyperglycemic clamp increased urinary excretion of inflammatory cytokines/chemokines, known to contribute to CKD [22]. In a rat model of intermittent hyperglycemia, short periods of post-prandial hyperglycemia increased monocyte adhesion to the endothelium vessel wall, which was worse than stable hyperglycemia, suggesting that acute hyperglycemia promotes the recruitment of proinflammatory monocytes secreting inflammatory cytokines causing podocyte apoptosis [23]. Therefore, it will be critical to prevent intermittent hyperglycemia or find factors promoting immunomodulation in the circumstance of acute hyperglycemia.

Vitamin D deficiency aggravates the risk of albuminuria progression in patients with T1 or T2DM [24]. Indeed, calcitriol treatment in T1DM and CKD subjects reduces proteinuria and systemic inflammation (TNF-α and IL-6 levels) independently of changes in glycemic control [24–26]. In vitro, vitamin D reduces human monocyte/macrophage TNF-α secretion upon stimulation with lipopolysaccharide (LPS) through several mechanisms, including direct inhibition of TNF-α and ADAM17 gene expression [27,28]. TNF-α secretion requires the cleavage of its precursor on the cell membrane by ADAM17 (A Disintegrin and Metalloprotease 17) [29], which in turn uses iRhom2 (inactive rhomboid protein 2) as the chaperon for ADAM17 transport to the cell surface [30]. In hemodialysis patients with higher serum TNF-α levels and monocyte ADAM17 expression, correction of vitamin D deficiency and paricalcitol treatment reduced both ADAM17 expression and serum TNF-α [31]. These studies suggest that vitamin D deficiency could aggravate acute hyperglycemia-induced monocyte activation, prompting podocyte apoptosis.

The use of glucagon-like peptide-1 receptor agonists (GLP-1 RA) has changed the treatment of diabetes and its complications. They are associated with a reduction in albuminuria of ~14 % in adult patients with T2DM compared with placebo or conventional therapy [32]. GLP1 RA not only protects against prolonged hyperglycemia-induced podocyte apoptosis by reducing hyperglycemia-induced podocyte ROS [33], but also by additional systemic beneficial effects, such as lowering glucose and blood pressure, and suppressing renin-angiotensin-aldosterone system (RAAS) activation and inflammation [34–38]. Increased GLP-1 levels by dipeptidyl peptidase 4 inhibitors significantly reduced glomerular macrophage infiltration, glomeruli TNF-α, expression, and NF-κβ activity in rodent models of T1DM, suggesting an anti-inflammatory effect [39,40]. In humans, the GLP-1 receptor is expressed in monocytes/macrophages, and the GLP-1 receptor agonist (GLP1-RA) drives human macrophages towards an anti-inflammatory M2 phenotype [41]. However, it is unclear if monocyte GLP-1 receptor activation by GLP-1 RA prevents AH-induced podocyte apoptosis independently of their systemic effects lowering glucose and blood pressure. In this study, we evaluate the mechanisms by which acute hyperglycemia-induced human monocyte proinflammatory phenotype causes podocyte apoptosis, and the potential preventative role of a normal vitamin D status and monocyte GLP-1 R activation.

2. Materials and methods

2.1. Study design

This is a cross-sectional study of healthy, nonpregnant subjects without a history of diabetes who were voluntarily recruited from Volunteers for Health or outpatient clinics at Washington University. Subjects were recruited under protocol 201905025, approved by the Washington University Human Research Protection Office. All subjects provided informed consent and study procedures were carried out at Washington University School of Medicine in St. Louis, MO. This study was conducted in accordance with the Declaration of Helsinki.

2.1.1. Human monocyte culture

After an 8 h fasting, monocytes were isolated from 40 ml of peripheral blood collected in lithium heparin tubes by a standard Ficoll gradient technique, followed by CD14 microbead positive selection, as previously described (Miltenyi Biotec, Auburn, CA). Monocytes were then seeded at a concentration of 3 × 10⁵ cells per well in 12 well plates coated with type 1 collagen (Corning), recovered for 1 h in 5 mM D-glucose DMEM (Gibco), and exposed to 5 mM or 16.7 mM D-glucose DMEM supplemented with 1 % P/S with 0–1 % FBS for 6 h, with or without simultaneous inhibition of glycolysis with 10 mM 2-deoxy-D glucose (2DG), ER stress with (2 mM 4-phenylbutyric acid (PBA)), or oxidative stress with (500 μM N-acetylcysteine (NAC)). Conditioned media was collected, and cells were washed and collected in TrIzol (Thermo Fisher). For each subject, duplicates or triplicates of each experimental condition were conducted according to cell count. TNF-α release was quantified by ELISA (R+D). Q-PCR measured monocyte mRNA expression of IRHOM2 (primer forward 5’-GCGCCTCTCTCCTGGGC-3’, reverse 3’- CTCTGGCCGTCCGCTTG-5’), ADAM17 (primer forward 5’-ACAGCGACTGCACGTTGAAGG-3’, reverse 3’-CTGTGCAGTAGGACACGCCTTT5’)), vitamin D receptor (VDR) (primer forward 5’-CGCATCATTGCCATACTGCTGG, reverse 3’-CCACCATCATTCACACGAACTGG), TNF-α (primer forward 5’-CTCTTCTGCCT GCTGCACTTTG, reverse 3’-ATGGGCTACAGGCTTGTCACTC), cyp27b1 (primer forward 5’-CTCCACTCAGAGATCACAGCTG, reverse 3’-GGACACGAGAATTTCCAGGTACC), and normalized to the housekeeping gene L32.

2.1.2. THP-1 macrophage culture

The THP-1 monocytic cell line (ATCC) was grown in RPMI 1640 (Sigma 8578) with 10 % FBS and 1 % P/S and used before passage 10. THP-1 cells, seeded at a density of 1 × 10⁶ per well in 6 well plates (TPP), were differentiated to macrophages by exposure to 160 nM phorbol 12-myristate 13-acetate (PMA, Sigma) for 48 h in DMEM 5 mM D-glucose media (Gibco) and then treated with either DMEM 5 mM D-glucose, DMEM 16.7 mM glucose, or DMEM 16.7 mM glucose + 10 nM Glp-1 (a generous gift from Dr. Burton Wice, was custom synthesized by Bachem, Torrance, CA) for 24 h [42]. Media was collected and cells were washed with PBS and lysed in RIPA buffer containing protease inhibitors. Triplicates were conducted for each experimental condition in at least three independent experiments.

2.2. Podocyte experiments

Human immortalized podocytes (CIHP-1 line), originated from Dr Moin Saleem [43] were a gift from Dr. Maggie Chen (Renal Division, Washington University). Podocytes were grown in type 1 collagen (Sigma) coated T75 flasks (TPP) at 33°C, 5 % CO2 in RPMI-1640 (Sigma 8578) with 10 % FBS, 1 % P/S, 1 % Insulin-transferrin-selenium (Sigma), 1 % nonessential amino acids (Sigma) and used before passage 20. For differentiated podocytes, cells were transferred to 37 C 5 % CO2 and used on days 14–16. Differentiated podocytes were trypsinized and seeded at a density of either 300,000 cells/well in type 1 collagen-coated 12-well plates (TPP) to assess changes in protein content or 30,000 cells per well into type 1 collagen-coated 96 well plates (Corning) to measure apoptosis. These podocytes were exposed for 24 h to: a) DMEM 5 mM or 16.7 mM D-glucose; b) conditioned media from either human monocytes or THP-1 macrophages as described above; c) conditioned media from either human monocytes or THP-1 macrophages exposed to DMEM 5 mM or 16.7 mM glucose + /− 10 nM GLP-1 for 2 h for human monocytes and for 24 h in THP-1 macrophages; d) DMEM 5 mM glucose with increasing concentrations of recombinant TNF-α (50–500 pg/ml). For TNF-α neutralizing experiments, conditioned media from THP-1 macrophages exposed to DMEM 16.7 mM D-glucose for 24 h or conditioned media from monocytes exposed to DMEM 16.7 mM D-glucose for 6 h were incubated with 30 ng/ml TNF-α neutralizing antibody (Cell Signaling, D1B4) for 5 h at 37°C with at least triplicate measurements per condition from at least three independent experiments. Podocytes were then exposed to this conditioned media for 24 h, and apoptosis was measured by caspase 3/7 (Promega) following the manufacturer’s instructions. Baseline podocyte caspase activity after a 24 h exposure to 5 mM glucose DMEM with and without 30 ng/ml of the TNF-α antibody served as a control for a potential adverse impact of the anti TNF-α antibody on podocyte viability.

2.3. Western blot

To assess for changes in podocyte proteins, PBS-washed cells were lysed in RIPA buffer containing protease inhibitors. BCA assay (Thermo Fisher) quantified protein content. 15 ug of protein from each sample in 4X SDS loading buffer was separated by 10 % SDS-PAGE (Biorad), transferred onto incubated with the respective secondary antibodies linked to horse-radish peroxidase (HRP) (1:2500 anti-rabbit IgG-HRP, Santa Cruz Biotechnology) for 2 h at room temperature before protein detection with a chemiluminescent HRP substrate (Millipore Sigma) on a Licor Odyssey Fc Imaging system. To detect β-actin, membranes were exposed for 15 minutes to stripping buffer (Licor), washed three times with 1X TBST, and incubated with primary antibody (1:5000 mouse anti-β-actin) for 1 h, washed three times with 1X TBST, and incubated with secondary antibody (1:25000 anti-mouse IgG-HRP, Cell Signaling Technology) for 1 h before imaging as above. Protein expression was quantified by Image J 2.0 and normalized to beta-actin expression.

2.4. Statistical analysis

Experiments were carried out in duplicate or triplicate. Data are expressed as mean ± SEM and analyzed by paired or unpaired 2-tailed t-tests. Data with more than two groups were analyzed by ANOVA and Tukey’s post-test. Statistical analysis was conducted with GraphPad Prism version 8.4.3. Values are considered statistically significant if p < 0.05.

3. Results

3.1. Acute hyperglycemia induces podocyte apoptosis indirectly through pro-inflammatory immune cell activation

To assess for direct effects of hyperglycemia on podocyte viability and apoptosis, podocytes were exposed to euglycemic (5 mM D-glucose DMEM) or hyperglycemic (16.7 mM D-glucose DMEM) conditions for 24 h. Hyperglycemia had no effect either on the expression of the structural podocyte protein podocalyxin or on caspase 3/7 activity (Fig. 1a-b). Next, to evaluate the contribution of AH-induced monocyte activation to cause podocyte apoptosis, we exposed podocyte for 24 h to conditioned media from peripheral blood monocytes from healthy individuals, subjected to euglycemic or hyperglycemic conditions for 6 h. Podocytes exposed to monocyte hyperglycemic conditioned media had reduced podocalyxin protein expression and a 32 % higher apoptotic rate compared to those in podocytes exposed to monocyte euglycemic conditioned media (Fig. 1c-d). These findings suggest that short periods of hyperglycemia are sufficient to activate human monocytes to induce podocyte injury and initiate apoptosis. Interestingly, in THP-1 macrophages, a prolonged exposure to hyperglycemic conditions of at least 24 h was required to significantly reduce podocalyxin expression and increase podocyte apoptosis by 16 % (Fig. 1e-f) compared to the effects of conditioned media from THP-1 macrophages exposed to euglycemic conditions. The lower CD14 expression in THP-1 cells compared to circulating monocytes could account, in part, for their lower response to acute hyperglycemia, as previously demonstrated upon LPS stimulation [44,45].

Fig. 1.

Acute hyperglycemia does not injure podocytes directly but induces a monocyte/macrophage secretome that drives podocyte damage/apoptosis. Human immortalized podocytes were exposed to: (A, B) euglycemic (5 mM D-glucose, green dots) or hyperglycemic (HG: 16.7 mM D-glucose, blue dots) conditions for 24 h; (C, D) conditioned media from human normal monocytes exposed to 5 or 16.7 mM glucose for 6 h; (E, G) conditioned media from THP1 cells (macrophage phenotype) exposed to 5 or 16.7 mM glucose for 24 h. (A, C, D) Podocyte injury (Normalized podocalyxin:β-actin protein expression in euglycemia (green dots), and hyperglycemia-induced fold reduction in podocalyxin:β-actin ratio (blue dots) from at least 3 independent experiments; (B, D, F) Podocyte apoptosis (Normalized basal caspase 3/7 activity in euglycemia (green dots) and hyperglycemia-induced fold-increases in caspase 3/7 activity (blue dots). Each dot represent the average of quadruplicate measurements; the number of dots indicate the number of individuals or independent experiments. *p < 0.05.

Next, we assessed if higher levels of TNF-α, a cytokine known to cause podocyte apoptosis, were present in the AH-induced monocyte/macrophage media. We exposed monocytes to hyperglycemia for 6 h or THP-1 macrophages for 24 h and found that hyperglycemia significantly increased monocyte TNF-α secretion compared to monocytes/macrophages incubated in euglycemic conditions (Fig. 2a,b). Moreover, AH-induced monocyte TNF-α release provoked podocyte apoptosis (Fig. 2c), which resembles the average caspase activity in resting podocytes after exposure to recombinant TNF-α at concentrations from 50–500 pg/ml (Supplemental Fig. 1a). This pro-apoptotic TNF-α effect is further supported by a significant reduction of apoptosis upon the addition of TNF-α neutralizing antibody to the conditioned media from human monocytes exposed to hyperglycemia for 6 h, or from THP-1 macrophages exposed to hyperglycemia for 24 h (Fig. 2d). No effect on podocyte apoptosis was induced by co-incubation of podocytes in euglycemic media with the TNF-α -neutralizing antibody (Supplemental Figure 1b). These findings suggest that in the setting of acute hyperglycemia, an early (6 h) TNF-α release by circulating monocytes contributes to podocyte apoptosis, and that THP-1 macrophages require a longer hyperglycemic exposure than peripheral blood monocytes to increase TNF-α secretion to cause podocyte apoptosis.

Fig. 2.

Acute hyperglycemia-induced monocyte TNF-α secretion causes podocyte apoptosis in a vitamin D-dependent manner. (A) Normalized TNF-α secretion from peripheral blood monocytes from 11 normal human subjects exposed ex vivo for 6 h to euglycemic (5 mM D-glucose, green dots) conditions, and individual fold increases in TNF-α secretion in response to hyperglycemia (HG, 16.7 mM D-glucose, blue dots); (B) Normalized TNF-α secretion in 24 h by THP-1 macrophages cultured under euglycemic conditions (green dots), and fold increases in TNF-α secretion in response to hyperglycemia (blue dots) (from n = 4 independent experiments); (C) Podocyte apoptotic rate (normalized caspase 3/7 activity) after 24 h exposure to conditioned media from peripheral blood monocytes from 6 normal individuals, each exposed to hyperglycemic conditions (dark blue dots) + /− TNF-α neutralizing antibody (light blue dots) for 6 h; (D) Podocyte apoptotic rate (normalized caspase 3/7 activity; n = 5) after 24 h exposure to conditioned media from THP-1 macrophages exposed to hyperglycemic conditions (dark blue dots) + /− TNF-α neutralizing antibody (light blue dots) for 24 h; (E) Fold differences in monocyte TNF-α secretion from human monocytes based on serum vitamin D levels (n = 10); (F) Correlation between serum vitamin D levels and monocyte TNF-α secretion (n = 10). Each dot represent the average of cuadruplicate measurements; the number of dots indicate the number of individuals or independent experiments. Bars and error bars represent mean+SEM. *p < 0.05 **p < 0.01, *** p < 0.001.

3.2. Vitamin D deficiency/insufficiency aggravates hyperglycemia-induced increases in immune cell TNF-α secretion causing podocyte apoptosis

Previous studies demonstrate that vitamin D suppresses the expression of the TNF-α gene [27,46]. Then, we examined whether vitamin D deficiency/insufficiency exacerbated monocyte release of TNF-α in response to transient hyperglycemia. We found that acute hyperglycemia of 6 h was sufficient to upregulate monocyte TNF-α and CYP27b gene expression and reduce VDR mRNA levels (Supplemental Fig 2). Moreover, we found that in individuals with normal 25(OH)D levels (>30 ng/ml), monocyte release of TNF-α in response to AH was significantly lower than that in monocytes from vitamin D deficient or insufficient subjects (Fig. 2e). Furthermore, there was a significant inverse correlation between serum vitamin D and the levels of TNF-α released to the conditioned media in response to acute hyperglycemia (6 h) (Fig. 2d). 25OHD suppresses monocyte ER stress and decreases their proinflammatory phenotype, adhesion to fibronectin, and migration [47]. However, the lack of measurements of monocyte CYP27b1 activity, VDR protein content, and nuclear localization precludes an assessment of the contribution of monocyte conversion of 25(OH)D to 1, 25(OH)2D3 to local VDR activation to suppress TNF-α expression and release. Taken together, these findings suggest that in the setting of acute hyperglycemia, an early (6 h) increase in TNF-α release by circulating monocytes contributes to podocyte apoptosis, an effect attenuated by a normal vitamin D status.

3.3. Increased glycolysis, ER- and oxidative stress mediate hyperglycemia-induced monocyte TNF-α secretion

To determine whether increased monocyte glycolytic activity mediated hyperglycemia induction of TNF-α secretion, human monocytes were exposed to hyperglycemia with or without the glycolysis inhibitor, 2-deoxy-glucose (2DG). 2DG reduced monocyte TNF-α levels in the incubation media by 51 %, indicating that the activation of glycolysis is required for monocytes to increase TNF-α secretion in response to hyperglycemic conditions (Fig. 3). Several studies suggest that chronic hyperglycemia in patients with diabetes increases markers of ER stress and oxidative stress in circulating monocytes [47,48,49,50]. ER stress is also critical to enhance ROS production in many organs and cellular systems. To assess the role of hyperglycemia-induced monocyte ROS and ER stress in TNF-α secretion, we inhibited ROS production with N-acetyl cysteine (NAC) or ER stress with 4-phenyl-butyric acid (PBA) and found a 39 % or 63 % reduction of hyperglycemia-induced TNF-α release, respectively, by an exclusive hyperglycemic stimulus (Fig. 3). Thus, targeting monocyte ER or oxidative stress during hyperglycemic stimuli could attenuate inflammation and its complications in diabetes.

Fig. 3.

Inhibition of glycolysis, ER and oxidative stress attenuates acute hyperglycemia-induced monocyte TNF-α secretion. TNF-α secretion (N = 5) by 50,000 human monocytes cultured under hyperglycemic conditions (HG: 16.7 mM D-glucose, dark blue bar) with and without the glycolysis inhibitor 2-deoxyglucose (2DG red bar), the ER stress inhibitor 4-phenylbutyric acid (PBA, light blue bar), or the oxidative stress inhibitor N-acetylcysteine (NAC, yellow bar) for 6 h. Bars and error bars represent Mean + SEM, *p < 0.05.

3.4. Hyperglycemia induction of monocyte ADAM17/iRhom2 expression controls TNF-α release

The release of TNF-α from its transmembrane precursor requires an enzymatic cleavage by ADAM17 at the cell surface. In turn, iRhom2, its regulatory-adapter molecule, mediates ADAM17 translocation from the ER to the cell membrane, an essential step for TNF-α release [30]. We found that human monocytes exposed to 6 h of hyperglycemia had a 3 or a 10-fold higher monocyte iRhom2 or ADAM17 mRNA expression, respectively, compared to their expression levels in monocytes exposed to euglycemic conditions (Fig. 4a,b). Since ER stress is known to upregulate ADAM17 transcription in other tissues [51], we explored whether hyperglycemia-mediated activation of monocyte ROS and ER stress induced TNF-α release by regulating ADAM17/iRhom2 expression. We found that NAC inhibition of ROS production and PBA, an ER stress inhibitor, prevented hyperglycemia-induced ADAM17 and iRhom2 expression (Fig. 4c-f). These results suggest that induction of monocyte TNF-α release by hyperglycemia is partly due to ER stress/ROS-mediated activation of ADAM17/iRhom2.

Fig. 4.

ER and oxidative stress activation in human monocytes mediate acute hyperglycemia-induced iRhom2/Adam17 expression. (A, B) ADAM17 and its partner for TNF-α release iRhom2 mRNA expression in human monocytes cultured either under euglycemic (5 mM D-glucose, green bar) or hyperglycemic (HG: 16.7 mM D-glucose, dark blue bar) conditions for 6 h (n = 8); ADAM17 or iRhom2 mRNA expression in human monocytes cultured under (C,D) ER inhibitors (PBA n = 6) or (D, E) oxidative stress inhibitor (NAC; n = 6). Bars and error bars represent mean + SEM. *p < 0.05.

3.5. GLP1 RA controls hyperglycemia-induced monocyte TNF-α release

Postprandial hyperglycemia at 2 h is a recognized risk factor for cardiovascular disease in patients with T2DM [52]. In vivo, transient episodes of hyperglycemia promoted myelopoiesis and proinflammatory monocytes [53]. Since GLP-1 RA have been found to attenuate ER and oxidative stress and decrease proinflammatory cytokine production [32, 54,55,56], we examined whether GLP-1 RA is able to attenuate the inflammatory monocyte activation induced by 2 h of acute hyperglycemia, thus mimicking postprandial hyperglycemic episodes. We found that monocytes exposed to hyperglycemia for 2 h in the setting of GLP-1 RA had reduced TNF-α levels, ADAM17, and iRhom2 expression compared to the levels in monocytes exposed to hyperglycemia alone (Fig. 5a-c). Similarly, the conditioned media from THP-1 macrophages incubated under hyperglycemic conditions with GLP-1 RA had lower TNF-α levels and reduced podocyte apoptosis (Fig. 5d,e). These findings suggest that downregulation of TNF-α release by GLP-1 RA is a potential mechanism to attenuate podocyte apoptosis and reduce CKD progression in diabetes.

Fig. 5.

GLP-1 receptor activation ameliorates podocyte apoptosis by attenuating monocyte/macrophage TNF-α secretion through a reduction in iRhom2 and ADAM17 expression. (A-C) TNF-α secretion (n = 5), iRHOM2 (n = 6) and ADAM17 (n = 6) expression in human monocytes cultured under hyperglycemic (HG: 16.7 mM D-glucose) conditions for 2 h (dark blue dots or bars) + /− 10 nM GLP-1 RA (light blue dots or bars); (D) TNF-α secretion from THP-1 macrophages exposed to hyperglycemic (HG, 16.7 mM D-glucose) conditions (dark blue dots) or HG + GLP-1 RA (light blue dots) for 24 h (n = 4); (E) caspase 3/7 activity in podocytes exposed for 24 h to conditioned media from THP-1 macrophages cultured under hyperglycemic (HG, 16.7 mM D-glucose, dark blue dots) conditions or HG + GLP-1 RA (light blue dots) for 24 h (n = 4). Each dot represents the average of duplicate measurements per individual or experimental condition, and the number of dots indicates the total number of individuals or independent experiments. Bars and error bars indicate Mean ± SEM of 3 independent experiments per experimental condition. *p < 0.05..

4. Discussion

Chronic and severe hyperglycemia are recognized inducers of podocyte apoptosis [12,13]. However, we demonstrated that short periods of moderate hyperglycemia did not induce podocyte structural injury or apoptosis directly, suggesting that other mechanisms are involved in AH to cause CKD progression. In this study, we found that acute hyperglycemia stimulates monocyte glycolysis, causing increased ROS production and ER stress, which in turn results in upregulation of iRhom2 and Adam17 expression, thereby increasing monocyte TNF-α secretion causing podocyte apoptosis. This process in human monocytes was exacerbated by vitamin D deficiency and attenuated by GLP-1 receptor agonists.

In normal subjects, acute hyperglycemia for 4–6 h during hyperglycemic clamps increases circulating TNF-α, IL-6, and IL-18 levels, and these levels were higher and lasted longer in glucose-intolerant individuals [57]. In vitro studies indicate that monocytes could be a potential source of the hyperglycemia-induced increases in circulating cytokines. Monocytes exposed to severe and prolonged hyperglycemia for 18–24 h or 12 h of intermittent hyperglycemia have increased cytokine secretion, but it is unclear whether short bounds of hyperglycemia commonly present in glucose-intolerant individuals, or in diabetes with glycemic variability, can induce a monocyte proinflammatory phenotype capable of causing podocyte apoptosis [19,20,58,59]. Our study found that exposure of normal human monocytes to 6 h of moderate hyperglycemia increased TNF-α release to levels sufficient to cause podocyte apoptosis, as corroborated through the reduction in apoptotic rates by TNF-α neutralizing antibodies. This demonstrates a potential mechanism by which acute hyperglycemia during episodes of glucose variability may contribute to CKD progression in diabetes. Importantly, upon exposure of monocytes to hyperglycemic conditions in the presence of 2-deoxy-D-glucose (2-DG), which blocks glycolysis at the level of the hexokinase, there was no effect of hyperglycemia on TNF-α release. This suggests that an excessive glucose entrance to monocyte-macrophages promotes oxidative and ER stress, inducing ADAM17 and IRhom expression and increasing TNF-α secretion.

Previous studies indicate that prolonged hyperglycemia increased monocyte TNF-α transcription by ROS activation of transcription factors such as nuclear factor-κβ and activating protein-1 [58]. Chronic hyperglycemia also induced ROS in monocytes by activating protein kinase C (PKC) [60], as blocking both ROS and PKC inhibited monocyte activation [58]. TNF-α release from its precursor on the cell membrane requires the metalloproteinase ADAM17 [61]. The chaperone proteins iRhom1 and iRhom2 are critical for translocating ADAM17 from the ER through the Golgi to the cell surface [61,62] to ensure ADAM17 activity. ER stress induces ADAM17 activation in nonimmune cells [63]. In this study, we found that acute moderate hyperglycemia-induced monocyte TNF-α release by increasing ADAM17 and iRhom2 gene expression. Moreover, inhibition of ROS or downregulation of ER stress prevented acute hyperglycemia-induced monocyte TNF-α release, in part by suppressing ADAM17 and iRhom2 expression. Our data illustrates the multilevel regulation by which acute hyperglycemia promotes proinflammatory monocyte activation and highlights the novel role of the oxidative and ER stress/iRhom2/ADAM17 axis in regulating the TNF-α release induced by short periods of acute hyperglycemia.

Vitamin D controls TNF-α and ADAM17 gene expression in different cell types [27,28] and attenuates ER and oxidative stress [64,65,66]. 25OHD suppresses monocyte ER stress and decreases their proinflammatory phenotype, adhesion to fibronectin, and migration [64]. Herein, we confirm an inverse correlation between serum 25(OH)D and the release of TNF-α by normal monocytes exposed to 6 h of hyperglycemia, suggesting that while vitamin D deficiency/insufficiency worsens the acquisition of a pro-inflammatory phenotype in circulating monocytes, a normal vitamin D status could attenuate the AH-induced monocyte TNF-α secretion. Also, 1,25(OH)2D3 has a direct renoprotective mechanism through its ability to stimulate nephrin expression, which plays a critical role in the renal filtration barrier [67]. Moreover, 1,25(OH)2D3 and its analog paricalcitol were shown to help maintain the nephrin PI3k-Akt signal involved in podocyte viability, thus reducing pro-apoptotic signals in ex vivo models of glomeruli and in mouse models of type 1 and type 2 diabetes [68]. Therefore, vitamin D deficiency could also aggravate hyperglycemia-induced podocyte apoptosis by multiple mechanisms. In our study, it is possible that in individuals with a normal vitamin D status, a higher conversion of 25 (OH)D to 1,25(OH)2D3 by pro-inflammatory monocyte-macrophages could counteract more effectively hyperglycemia-induced podocyte apoptosis, not only through the attenuation of oxidative- and ER stress-mediated increases in TNF-α release but also through direct suppression of TNF-α -independent pro-apoptosis signals.

GLP-1 RA has been shown to delay diabetic kidney disease progression through a direct reduction of podocyte oxidative stress and podocyte apoptosis [33], and also indirectly by facilitating glycemic control, reducing RAAS activation, and glomerular and interstitial macrophage infiltration in humans and animal models of chronic hyperglycemia [32,34,37–39]. Previous studies suggest that GLP-1 RA inhibits monocyte ROS and promotes an anti-inflammatory M2 macrophage phenotype in human monocytes [41,69]. However, it is unclear whether GLP-1 RA effects on podocyte apoptosis during acute hyperglycemia are through decreasing monocyte proinflammatory phenotype. In this study, GLP-1 RA recapitulated the attenuation of the proinflammatory monocyte activation during acute hyperglycemia by inhibitors of ROS and ER stress. GLP-1 RA reduced monocyte TNF-α secretion by reducing iRHOM2 and ADAM17 expression and prevented TNF-α induced podocyte apoptosis. Previous studies indicate that GLP-1R and Vitamin D receptor activation attenuate endoplasmic reticulum stress, reduce macrophage migration, and decrease the expression of proinflammatory molecules in macrophages [41,70–73]. These results highlight GLP-1 RA and vitamin D as potential synergistic therapeutic modulators of the monocyte iRhom2/ADAM17 axis by regulating the TNF-α-induced podocyte apoptosis in human cells.

A limitation in this study is that we exclusively evaluated human podocyte injury as a mechanism of immune glomerular injury in kidney disease, but hyperglycemia-induced CKD onset and progression involves an interplay between multiple cell types, including T lymphocytes, dendritic, mesangial and glomerular endothelial cells [74,75,76]. Therefore, more studies must be conducted to assess the complex interaction between human innate immunity and glomerular cells to better understand pathogenic mechanisms leading to early glomerular injury in diabetes.

5. Conclusions

This study revealed a novel monocyte oxidative/ER stress/iRhom2/ADAM17 pathway induced by acute hyperglycemia that regulates TNF-α induced podocyte apoptosis and identified vitamin D and GLP1 RA as potential therapeutic immunomodulators for hyperglycemia-induced kidney disease.

Data availability

Data will be made available on request.