PKM2-mediated neuronal hyperglycolysis enhances the risk of Parkinson's disease in diabetic rats

Abstract

Epidemiological and animal studies indicate that pre-existing diabetes increases the risk of Parkinson's disease (PD). However, the mechanisms underlying this association remain unclear. In the present study, we found that high glucose (HG) levels in the cerebrospinal fluid (CSF) of diabetic rats might enhance the effect of a subthreshold dose of the neurotoxin 6-hydroxydopamine (6-OHDA) on the development of motor disorders, and the damage to the nigrostriatal dopaminergic neuronal pathway. In vitro, HG promoted the 6-OHDA-induced apoptosis in PC12 cells differentiated to neurons with nerve growth factor (NGF) (NGF-PC12). Metabolomics showed that HG promoted hyperglycolysis in neurons and impaired tricarboxylic acid cycle (TCA cycle) activity, which was closely related to abnormal mitochondrial fusion, thus resulting in mitochondrial loss. Interestingly, HG-induced upregulation of pyruvate kinase M2 (PKM2) combined with 6-OHDA exposure not only mediated glycolysis but also promoted abnormal mitochondrial fusion by upregulating the expression of MFN2 in NGF-PC12 cells. In addition, we found that PKM2 knockdown rescued the abnormal mitochondrial fusion and cell apoptosis induced by HG+6-OHDA. Furthermore, we found that shikonin (SK), an inhibitor of PKM2, restored the mitochondrial number, promoted TCA cycle activity, reversed hyperglycolysis, enhanced the tolerance of cultured neurons to 6-OHDA, and reduced the risk of PD in diabetic rats. Overall, our results indicate that diabetes promotes hyperglycolysis and abnormal mitochondrial fusion in neurons through the upregulation of PKM2, leading to an increase in the vulnerability of dopaminergic neurons to 6-OHDA. Thus, the inhibition of PKM2 and restoration of mitochondrial metabolic homeostasis/pathways may prevent the occurrence and development of diabetic PD.

Graphical abstract

Highlights

• •Increased risk of diabetic Parkinson's disease was related to high glucose of brain microenvironment.
• •PKM2 mediated high-glucose induced neuronal vulnerability to 6-OHDA
• •High glucose promoted abnormal mitochondrial fusion and apoptosis induced by 6-OHDA.
• •Inhibition of PKM2 blocked the development of diabetic Parkinson's disease.

1. Introduction

The incidence of metabolic and neurodegenerative diseases has been increasing worldwide for several decades [[1], [2], [3]]. Diabetes mellitus (DM) has reached epidemic proportions, and its incidence is predicted to double by 2030 compared to 2000 [[4], [5], [6], [7]]. Parkinson's disease (PD) is becoming increasingly prevalent in the elderly population, and the increased rate of PD is expected to be similar to that of DM [8]. Therefore, it is important to determine whether the prevalence of metabolic disorders increases the risk of developing PD.

Increasing evidence indicates that diabetes is a risk factor for an increased incidence of PD [[9], [10], [11]]. This association has been more clearly observed in epidemiological and experimental studies of PD cases, in which insulin resistance in patients with PD is associated with accelerated disease progression, increased severity of movement disorders, and an increased risk of PD [12,13]. In addition, many patients with DM but without PD exhibit pathologies related to subclinical striatal dopamine dysfunction [14]. DM and PD share a common pathogenetic mechanism, and the discovery of α-synuclein inclusions in pancreatic β cells of patients with diabetes further supports the link between these two diseases [[15], [16], [17]].

It is well documented that DM increases the risk of PD; however, few studies have shown that DM and PD risk have no correlation or are negatively correlated [[18], [19], [20]]. These discrepancies have been attributed to the heterogeneity of self-report or diagnostic criteria, the difference in size or design, the patient's genetic background, lifestyle, or other unhealthy confounding factors [21,22]. However, according to current studies, we believe that DM tends to increase the risk of PD. Based on this trend, an increasing number of studies have explored possible mechanisms underlying the association between DM and PD.

Hyperglycemia is considered to be a risk factor for neurodegeneration of the nigrostriatal pathway associated with motor symptoms in patients with PD [23]. Yang et al. [24] showed that Hoehn-YahR (H–Y) staging and hemoglobin A1c (HbA1c) ≥ 7% may be independent risk factors for cognitive impairment in PD, and abnormal blood glucose is significantly correlated with cognitive impairment in patients with PD. In addition, other studies have shown that hyperglycemia significantly increases glucose metabolism and leads to an increased risk of PD through the glycosylation pathway, oxidative stress, dopaminergic neuron degeneration, and high neuroinflammatory neurotoxin levels caused by hyperglycemia, all of which play a key role in the risk of patients with DM or PD [[25], [26], [27]].

Although recent studies, including large cohort studies, have supported the idea that hyperglycemia promotes PD development, the possible mechanisms remain unclear [28,29]. As a metabolic disease, diabetic hyperglycemia alters the metabolic characteristics of neurons [30,31]. Metabolomics, a technology developed in recent years, can be used to monitor global metabolite changes [32,33]. Cell metabolism determines the functional phenotype of the cells. We used metabolomics to explore the metabolic regulation of 6-hyroxydopamine (6-OHDA)-induced neurons by high glucose (HG), and whether this regulation leads to the vulnerability of neurons to 6-OHDA. Metabolomics revealed that HG levels further enhanced the 6-OHDA-induced neuronal glycolysis pathway. To verify this finding, we examined the levels of key enzymes in the glycolysis pathway using Western blotting and found that elevated pyruvate kinase M2 (PKM2) might mediate this process, while altering mitochondrial dynamic homeostasis, causing abnormal mitochondrial fusion, and decreasing mitochondrial membrane potential (MMP). In addition, as an inhibitor of PMK2, shikonin (SK) reversed excessive glycolysis and abnormal mitochondrial fusion to enhance neuronal tolerance to 6-OHDA and reduce the risk of PD in diabetic rats. This finding plays a guiding role in the clinical treatment strategy for patients with DM complicated with PD, and further exploration of the efficacy of SK in combination with other first-line clinical diabetes drugs is expected to determine the best treatment combination.

2. Materials and methods

2.1. Animals

Male Sprague-Dawley rats (SD rats) (n = 70) were obtained from Zhejiang Academy of Medical Sciences (Hangzhou, China) and treated with streptozotocin (STZ) to induce diabetes. All rats were 15- to 20-week-old and housed under a 12:12-h dark/light cycle at 22 ± 2 °C with food and water ad libitum. All animal studies were performed in accordance with the Institutional Animal Care and Use Committee of Zhejiang University and the Guide for Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised in 1996).

2.2. STZ-induced DM

STZ (50 mg/kg, i.p.; Sigma-Aldrich, St. Louis, MO, USA) was administered for seven consecutive days. Rats were considered to be diabetic when glucose levels monitored with a glucometer (Accu Chek Performa Nano, Roche, Basel, Switzerland), using tail blood after a 14-h fasting, were >250 mg/dL. Control mice were injected with the vehicle (10 mM sodium citrate, 0.9% NaCl; pH 4.5).

2.3. Microinjections of 6-OHDA

A subthreshold dose (2 μg/μL, 4 μL) of 6-OHDA (10 mM in 0.9% NaCl and 0.2% ascorbic acid solution; Sigma-Aldrich), corresponding to three-quarters the dose required to induce overt neurodegeneration of nigrostriatal axons, was used because it does not induce motor impairment in non-diabetic rats; it was injected unilaterally under 2% isoflurane anesthesia into the dorsal medial forebrain bundle (MFB) using a Hamilton syringe [34]. Stereotaxic coordinates were AP = −3.2, L = −1.5, and DV = −8.7. In the STZ-treated mice, 6-OHDA was injected at the onset of hyperglycemia. Rats were tested 2 weeks after 6-OHDA administration.

2.4. Drug administration

The rats were allowed to recover for 3 days post-surgery before drug injection. Insulin (4 U/kg) or vehicle (0.9% NaCl) was injected intraperitoneally for four weeks. For the SK treatment group, SK (5 mg/kg/day; Selleck, Shanghai, China) or an equal volume of vehicle (5% dimethyl sulfoxide (DMSO), 10% Tween-80, 30% PEG300, and 55% saline) was intragastrically administered for four weeks [35].

2.5. Motor function

The rats were tested at 24-h intervals and killed within 24 h after the last test (15 days after 6-OHDA administration).

2.5.1. Rota-rod test

An accelerating rota-rod apparatus (Hugo Basile, Gemonio, Italy) was used, as previously described [36]. The rotation was accelerated from 5 to 20 rpm within 2 min, and the fall time was determined.

2.5.2. Forepaw adjusting steps (FAS) test

Referring to the methods in the literature [37], the experimenter fixed and lifted the rat's hindlimb skeleton with one hand and fixed its right forelimb with the other hand so that the left forelimb was placed on the table. The rat's body was positioned at an angle of approximately 45° from the table. The rats were then artificially adjusted to move at a constant speed of 90 cm/10 s along the edge of the table six times at an interval of 10 s. During this time, the number of steps was recorded in each test, and an average of six measurements was taken as the number of steps taken by the rats.

2.5.3. Cylinder test

The rats were individually placed in a transparent cylinder to assess asymmetry in the spontaneous use of the forelimbs [38]. The number of wall contacts made with the ipsilateral and contralateral forepaws (relative to the 6-OHDA injection side) in 3 min was recorded via video and scored.

2.5.4. Amphetamine (AP)-induced rotation test

Following administration of d-amphetamine sulfate (5 mg/kg, i.p.; Sigma-Aldrich), full-body turns toward the ipsilateral 6-OHDA injection side were scored using rotometer bowls [39].

2.6. Immunohistochemistry and lesion quantification

All animals used for histological assessment were tested for motor performance. Coronal free-floating sections (30 μm) from a slicing vibratome were incubated overnight with a tyrosine hydroxylase (TH) antibody for diaminobenzidine immunoperoxidase staining. Images were obtained using a bio-optical microscope (Nikon ECLIPSE50i, Nikon, Tokyo, Japan) and measured using ImageJ software.

2.7. Cell culture

The PC12 cell line was obtained from the American Type Culture Collection (ATCC; Manassas, VA, USA), cultured at 37 °C in a humidified atmosphere of 5% CO₂, and passaged twice a week. The experiments were performed on the cells between passages 4 and 15.

2.8. Cell differentiation

The differentiation medium contained RPMI-1640 with l-glutamine, gentamicin, amphotericin B, 1% dihydroxy-1-selenolane (DHS), and nerve growth factor (NGF) (Cat. No. N2513; Sigma-Aldrich ) at a concentration of 100 ng/mL. The culture medium was changed every 48 h [40].

2.9. Cell viability assay

Cell viability was measured using the sulforhodamine B (SRB) method, as described in a previous study [41]. Briefly, PC12 cells were seeded into 96-well plates and incubated for 24 h. After treatment with 100 ng/mL NGF, the cells were treated for three days. For glucose or SK, different concentrations of these reagents were added to the wells and incubated for a further 48 h. For the 6-OHDA treatment group, different concentrations of 6-OHDA were added to the wells for another 24 h of incubation. After treatment, cell viability was measured using the SRB assay (Sigma-Aldrich). The percentage of cell viability was calculated using the following formula: Cell viability (%) = cell number in the treatment group/cell number in the DMSO control group × 100%. The data were obtained from three independent experiments.

2.10. Metabolite extraction

After treatment, the cell medium was removed and the cells were treated with pre-cold methanol (−40 °C), scraped quickly in a dry ice-methanol bath, and then centrifuged at 20,000 g for 10 min at 4 °C to obtain the supernatant and precipitate. The supernatant was considered as the intracellular sample and dried in a freeze dryer for subsequent liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis, and the protein concentration of the precipitate was measured using the bicinchoninic acid (BCA) assay.

2.11. Metabolomics data acquisition

The dried samples were resuspended in normalized volume (according to the protein amount in each sample) of 50% acetonitrile in Millipore water (Manassas, VA, USA), vortexed for 30 s, and centrifuged at 20,000 g for 10 min at 4 °C to obtain the supernatant for LC-MS/MS analysis. Metabolites were separated on a Waters ACQUITY UPLC BEH amide column (100 mm × 2.1 mm, 1.7 μm; Manassas, VA, USA). The temperature was maintained at 25 °C. The gradient elution system contained mobile phase A (100% H₂O + 25 mM CH₃COONH₄ + 25 mM NH₄OH) and B (acetonitrile(ACN)). The elution program was set as follows: 95% mobile phase B from 0 to 1 min, decreased mobile phase B to 65% from 1 to 14 min, decreased mobile phase B to 40% from 14 to 16 min, and then kept till 18 min. The flow rate was 0.3 mL/min. MS data acquisition was performed using a Q Exactive Plus system (Thermo Fisher Scientific Inc., Rockford, IL, USA). In the full-scan MS/dd-MS² mode, the resolutions of the full-scan MS and dd-MS² were set at 70,000 and 17,000, respectively. The spray voltage was set as 3.5 and 3.2 kV for positive and negative modes, respectively. The capillary and auxiliary gas heater temperatures were set to 320 °C and 350 °C, respectively. The flow rates of the sheath and auxiliary gas were set to 35 and 15 (in arbitrary units), respectively. All samples were randomly injected during data acquisition.

2.12. Metabolomic data processing

Following LC-MS data acquisition, the raw data were further processed using Compound Discoverer 3.2 (Thermo Fisher Scientific Inc.) with the metabolite databases mzCloud, mzVault, Masslist, and Chemspider. Principal component analysis (PCA) and cluster analysis of the identified metabolites were performed using R software (version 3.6.3).

2.13. Measurement of lactate and glucose

Blood, cerebrospinal fluid, and cultured cells were homogenized in lysis buffer and sonicated at 300 W (3 s on and 7 s off) for 3 min on ice, followed by centrifugation at 12,000 g for 10 min at 4 °C. The supernatants were collected, and lactate and glucose levels were measured using a lactate colorimetric assay kit (A019-2-1, Nanjing Jiancheng Bioengineering Institute, Nanjing, China) and a glucose colorimetric assay kit (F006-1-1, Nanjing Jiancheng Bioengineering Institute), respectively, according to the manufacturer's instructions.

2.14. Mitochondria quantification

The number of mitochondria was measured using the MitoTracker® Green FM assay kit (40742ES50; Yeasen Biotechnology Co., Ltd., Shanghai, China) according to the manufacturer's instructions. Mito-Tracker is a bright green fluorescent probe (Ex = 490 nm, Em = 523 nm), which hardly fluoresces in aqueous solution and only fluoresces when concentrated in the lipid environment of the mitochondria.

2.15. Transmission electron microscopy (TEM)

TEM was performed at the Center of Cryo-Electron Microscopy (CCEM) at Zhejiang University. For TEM observations, the samples were fixed for 72 h with 2.5% glutaraldehyde at pH 7.3 and buffered with phosphate-buffered saline (PBS, 1 mM). After fixation with 1% osmic acid for 1 h, the cells were stained with 2% uranyl acetate for half an hour and dehydrated in a graded ethanol series. Then, the cells were penetrated in penetrating agent (embedding agent:pure acetone, 1:1, V/V) at 25 ℃ for 2 h, and ultra-thin (± 70 nm) sections on copper grids, finally imaged on a Tecnai T10 transmission electron microscope (Hillsboro, OR, USA).

2.16. JC-1 assay

The MMP of the cells was measured using JC-1 (40705ES03; Yeasen Biotechnology Co., Ltd.), a dual-emission membrane potential-sensitive probe that exists as a green fluorescent monomer at low MMP, and forms aggregates with red/orange fluorescence at high MMP. Cells cultured in 96-well plates were washed twice with PBS, and JC-1 (1 mg/mL) was added for 30 min at 37 °C. The fluorescence intensity at 488/530 nm (green) and 549/595 nm (red) was monitored using a microplate reader, and the ratio of red to green fluorescence intensity represented the level of MMP.

2.17. Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining

After drug treatment, the cells present in the supernatant were pooled with trypsinized adherent cells, centrifuged, and resuspended in PBS. Aliquots of this suspension were deposited on slides using a cytospin centrifuge. The cells were fixed with ice-cold 4% paraformaldehyde for 10 min and washed twice with PBS. They were then subjected to TUNEL (40308ES20; Yeasen Biotechnology Co., Ltd.) according to the manufacturer's instructions. Next, 4′,6-diamidino-2-phenylindole (DAPI) (1/1000 in PBS) was added to the cells for 5 min at 25 °C. Images were obtained using a fluorescence microscope (Nikon ECLIPSE 50i). The percentage of apoptotic cells was determined by counting the TUNEL⁺ cell fraction of at least 500 cells in random fields.

2.18. PKM2 knockdown

Differentiated PC12 cells were transfected with PKM2 (Norway rat) siRNA-468 (5′-CCGCAGAGGUGGAGCUGAATT-3′), PKM2 (Norway rat) siRNA-867 (5′-GCAAGAACAUCAAGAUCAUTT-3′), and PKM2 (Norway rat) siRNA-640 (5′-GGAGAAAGGUGCUGACUACTT-3′) (Obio Technology (Shanghai) Corp., Ltd., Shanghai, China) using UltraFection 3.0 (FXP135-020; Beijing 4A Biotech Co., Ltd., Beijing, China) for 24 h, and then analyzed by Western blotting (Fig. S1, see Fig. S2 for full immunoblots).

2.19. Western blotting

After cell treatment or animal administration, cells or brain tissues were lysed using RIPA lysis buffer (Cell Signaling Technologies, Beverly, MA, USA) according to the manufacturer's instructions, and the protein concentrations were determined using the Bradford reagent (Bio-Rad, Hercules, CA, USA).

For the rat brain tissues, the whole brain was weighed, washed with PBS, and then homogenized with 1 mL of cold PBS solution using a Tissue Ruptor instrument (Shanghai Jingxin Industrial Development Co., Ltd., Shanghai, China). The homogenized brain tissues were lysed using RIPA buffer, and the protein concentration was measured. An equal amount of protein (50 μg/well) was loaded onto sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gels for electrophoresis. The gel-separated proteins were then transferred onto polyvinylidene fluoride membranes, which were blocked with 5% non-fat dried milk for 1 h. The membranes were incubated with primary antibodies such as TH (25859-1-AP; Proteintech, Wuhan, China), PFKFB3 (13763-1-AP; Proteintech), PKM1 (15821-1-AP; Proteintech), PKM2 (15822-1-AP; Proteintech), LDHA (19987-1-AP; Proteintech), PDHB (14744-1-AP; Proteintech), MFN2 (12186-1-AP; Proteintech), FIS1 (10956-1-AP; Proteintech), and β-actin (20536-1-AP; Proteintech) overnight at 4 °C, followed by incubation with horseradish peroxidase-conjugated secondary antibodies for 1 h. Finally, visualization of protein bands was performed using the Ultra ECL Western Blotting Detection Reagent (36208ES76; Yeasen Biotechnology (Shanghai) Co., Ltd.) and detected by the ChemiDoc MP Imaging System (Bio-Rad). Band intensities were quantified using ImageJ software (ImageJ 1.46r; National Institutes of Health, Bethesda, MD, USA), and the expression of the proteins of interest relative to that of β-actin was calculated.

2.20. Statistical analyses

Data were analyzed using one-way analysis of variance (ANOVA), followed by Dunnett's or Bonferroni's tests, using R software (version 3.6.3). Data are presented as the mean ± standard error of mean (SEM) unless otherwise indicated. Significance was set at P < 0.05 and expressed as ^∗P < 0.05, ^∗∗P < 0.01, and ^∗∗∗P < 0.001.

3. Results

3.1. DM caused motor dysfunction following injection of rats with a subthreshold dose of 6-OHDA

To verify the effect of diabetic hyperglycemia on the occurrence and development of PD, STZ-induced type 1 diabetes mellitus (T1DM) rats were used (Fig. 1A). Glucose levels in the rat's blood and cerebrospinal fluid were measured. As shown in Fig. S3, the glucose levels of the STZ-induced diabetic rats with or without a subthreshold dose of 6-OHDA were significantly increased, but there was no significant difference between these two groups. However, the changes observed in the STZ-induced diabetic rats did not translate into significant changes in motor function (Figs. 1B–E). Therefore, we investigated whether these changes were associated with increased sensitivity to neurodegeneration in the substantia nigra-striatum pathway. As expected, non-diabetic rats injected at the unilateral medial forebrain bundle with the subthreshold dose of 6-OHDA showed no motor injury (Figs. 1B–E). In contrast, STZ-treated animals showed decreased latency to fall on the rota-rod (Fig. 1B), a reduced number of lifts of the left forelimb in FAS (Fig. 1C), decreased use of the opposite paw on the injection side in the cylinder test (Fig. 1D), and increased ipsilateral rotation after amphetamine injection (Fig. 1E). The motor dysfunction of insulin-treated rats was significantly increased (Figs. 1B–E), which was consistent with the significant reduction in cerebrospinal fluid glucose levels. These results suggested that HG levels might increase the susceptibility of dopaminergic neurons to 6-OHDA, leading to motor dysfunction induced by subthreshold doses of 6-OHDA in diabetic rats.

Fig. 1.

Increased motor dysfunction and nigrostriatal neurodegenerative damage in T1DM rats after 6-OHDA administration. (A) Schematic diagram of the animal experimental design. (B–E) Data from non-diabetic or STZ-induced diabetic rats were evaluated on rota-rod test (B), FAS test (C), cylinder test (D) and AP rotation test (E) (n = 7). (F) Representative immunoblots of TH and β-actin of the striatum (Cropped blot images are shown; see Fig. S4 for full immunoblots). (G) Quantification of TH normalized to β-actin (n = 3). (H) Representative photomicrographs of coronal sections of the SNc and striatum of nondiabetic or STZ-induced diabetic rats immunostained for TH. (I and J) Histograms represent the proportional stained area of immunoreactivity (n = 3). TH⁺ fibers in striatum (I), and TH⁺ cells in SNc (J). Data are shown as mean ± standard error of mean (SEM). A one-way ANOVA was used for statistical analysis, followed by a Tukey's multiple comparisons test. ^∗∗∗P < 0.001 vs. Con; ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 vs. STZ+6-OHDA. SD-Rat: Sprague-Dawley rats; STZ: streptozotocin; T1DM: type 1 diabetes mellitus; 6-OHDA: 6-hydroxydopamine; Glc: glucose; Ins: insulin; AP: amphetamines; FAS: forepaw adjusting steps; SNc: substantia nigra pars compacta; TH: tyrosine hydroxylase; L/R: left/right.

3.2. A subthreshold dose of 6-OHDA caused nigrostriatal neurodegenerative damage in diabetic rats

To confirm that motor dysfunction is associated with dopaminergic fiber loss in the substantia nigra-striatum pathway, Western blotting and immunohistochemistry analyses of TH immunopositive fibers in the rat striatum were performed. Consistent with the Western blotting results (Figs. 1F and G, see Fig. S4 for full immunoblots), the loss of TH-immunopositive fibers in the striatum of STZ-diabetic rats induced by unilateral microinjection of the subthreshold dose of 6-OHDA was significantly greater than that in non-diabetic rats (Figs. 1H and I). To further determine whether TH loss in the striatum of 6-OHDA-treated diabetic rats was associated with midbrain dopamine neuron loss, we measured the number of TH-immunoreactive neurons in the substantia nigra pars compacta (SNc). In STZ-diabetic rats, 6-OHDA reduced the number of TH-positive neurons, whereas no significant neuronal loss was found in non-diabetic control rats (Figs. 1H and J), indicating that the loss of substantia nigra cells might contribute to motor deficits. Together, these results suggested that the changes observed at the molecular and functional levels were associated with a higher susceptibility of dopaminergic neurons in the substantia nigra-striatum pathway to neurodegeneration in diabetic rats than in non-diabetic controls.

3.3. High glucose promoted the damaging effect of 6-OHDA on neurons

To further explore whether HG levels increase neuronal susceptibility to 6-OHDA, NGF-PC12 cells were used. Consistent with the results from the animal experiments, the cells were first treated with HG (30 mM) (Fig. S5A) for 24 h, and then treated with a subthreshold dose of 6-OHDA (50 μM) (Fig. S5B) for 24 h; cell viability was examined using SRB (Fig. 2A). The results showed that HG stimulation alone reduced cell viability by approximately 9%, and 6-OHDA had no significant effect on cell viability (4%) (Fig. 2B). However, the combined intervention of HG and 6-OHDA resulted in a 46% decrease in cell viability (Fig. 2B). To further confirm the damaging effects of HG and 6-OHDA on neurons, we performed Western blotting to determine the expressions of apoptotic proteins, such as Bax and Bcl2. Figs. 2C–E show that HG or 6-OHDA alone significantly upregulated Bax level, while only 6-OHDA downregulated Bcl2 level (see Fig. S6 for full immunoblots). However, both HG and 6-OHDA upregulated the ratio of Bax/Bcl2, indicating that HG or 6-OHDA alone could cause cell apoptosis. Compared to 6-OHDA alone, combined intervention further upregulated Bax, downregulated Bcl2, and increased the Bax/Bcl2 ratio (Fig. 2 F). TUNEL staining also demonstrated that HG promoted 6-OHDA-induced apoptosis in NGF-PC12 cells (Figs. 2G and H). The results showed that HG increased the sensitivity of neurons to 6-OHDA, which further confirmed the key role of HG in the development of PD in vivo.

Fig. 2.

High glucose accelerated the damage effect of 6-OHDA on neurons in vitro. (A) Schematic diagram of the cell experimental design. (B) Cell viabilities were tested by sulforhodamine B (SRB) in NGF-PC12 cells cultured in the presence of HG (30 mM glucose) with or without 50 μM 6-OHDA. (C) Representative immunoblots of Bax, Bcl2 and β-actin of the NGF-PC12 cells cultured in the presence of HG (30 mM glucose) with or without 50 μM 6-OHDA (Cropped blot images are shown; see Fig. S6 for full immunoblots). (D, E) Quantification of Bax (D) and Bcl2 (E) normalized to β-actin. (F) The ratio of Bax/Bcl2 in NGF-PC12 cells cultured in the presence of HG (30 mM glucose) with or without 50 μM 6-OHDA. (G) Reprentative images of DAPI and TUNEL staining of NGF-PC12 cells cultured in the presence of HG (30 mM glucose) with or without 50 μM 6-OHDA. (H) Quantification of TUNEL⁺ apoptotic cells in NGF-PC12 cells cultured in the presence of HG (30 mM glucose) with or without 50 μM 6-OHDA. Data are shown as mean ± standard error of mean (SEM); n = 3, biologically independent samples. A one-way ANOVA was used for statistical analysis, followed by a Tukey's multiple comparisons test. ^∗∗P < 0.01, ^∗∗∗P < 0.001 vs. Con; ^###P < 0.001 vs. 6-OHDA; ND: no data. NGF: nerve growth factor; NGF-PC12 cells: differentiated PC12 cells by nerve growth factor; HG: high glucose; 6-OHDA: 6-hydroxydopamine; Bax: Bcl2 associated X protein; Bcl2: B-cell lymphoma-2; TUNEL: terminal deoxynucleotidyl transferase dUTP nick end labeling; DAPI: 4′,6-diamidino-2-phenylindole.

3.4. Metabolomic changes of the neurons treated with HG combined with subthreshold dose 6-OHDA

However, the mechanism underlying 6-OHDA-induced neuronal injury induced by HG levels is still unknown. On the one hand, HG can promote cellular glycolysis [42,43]. On the other hand, 6-OHDA competitively inhibits dopamine in brain tissue and strongly blocks the mitochondrial respiratory chain, leading to mitochondrial damage in dopaminergic neurons, a significant decrease in local antioxidant content and ATP production, and an increase in cell apoptosis or necrosis [[44], [45], [46]]. It is reasonable to hypothesize that low doses of 6-OHDA may lead to cellular stress and dependence on glycolysis for energy production by inhibiting the mitochondrial respiratory chain. At the same time, HG induces excessive glycolysis in cells and synergizes with 6-OHDA, resulting in cell metabolism imbalance and subsequent death. Therefore, we performed untargeted metabolomics to assess the influence of the combined intervention of HG and 6-OHDA on the metabolic characteristics of NGF-induced PC12 neurons. PCA and cluster analysis showed that the metabolic composition of 6-OHDA-induced cells was significantly different from that of the control group, and that HG treatment seemed to amplify this difference (Fig. 3A). Analysis of cell glycolysis and TCA metabolism intermediates revealed that HG upregulated cellular glycolysis intermediates (Fig. 3B), leading to a significant increase in the ratio of lactate to pyruvate (Lac/Pyr) and a decrease in the ratio of citrate to pyruvate (Cit/Pyr) (Fig. 3C). In addition, colorimetric measurements of extracellular lactate levels showed that HG promoted extracellular lactate release (Fig. 3D). Similar to HG, the subthreshold dose of 6-OHDA also activated glycolysis and promoted the release of extracellular lactate (Figs. 3B–D). The combined intervention of HG and 6-OHDA further enhanced the glycolysis metabolism of cells, especially the levels of lactate (Figs. 3C and D). To verify this, Western blotting was used to detect the levels of key enzymes involved in the intracellular glycolysis pathway. The results showed that HG or subthreshold 6-OHDA upregulated the levels of PFKFB3, PKM2, and LDHA and downregulated the level of PDHB (Figs. 3E–I, see Fig. S7 for full immunoblots), which was consistent with the results of glycolysis activation. Interestingly, HG increased the levels of PKM2 and LDHA, but not PFKFB3, and downregulated the level of PDHB in 6-OHDA-induced cells compared with 6-OHDA (Figs. 3E–I). Combined with the results of glycolysis metabolite levels, PKM2 might mainly mediate glycolysis in cells under the combined intervention of HG and 6-OHDA. In addition, our results showed that the PKM2 increase was accompanied by a decrease in PKM1 expression (Figs. 3E and J, see Fig. S7 for full immunoblots). It has been reported that inhibition of PKM2 can activate PKM1 in microglia of 5XFDA mice, thereby promoting the transformation of cellular metabolic characteristics from glycolysis to oxidative phosphorylation [47]. Combined with our results, 6-OHDA inhibited the expressions of PDHB and PKM1 (Figs. 3I and J), which was consistent with its inhibitory effect on the mitochondria. Thus, HG might aggravate 6-OHDA-induced glycolysis by increasing the level of PKM2 and decreasing the levels of PKM1 and PDHB to promote pyruvate conversion to lactate rather than Ac-CoA in NGF-PC12 cells.

Fig. 3.

High glucose accelerated 6-OHDA-induced hyperglycolysis. (A) Principal component analysis (PCA) of cell metabolism characteristics under positive and negative ion modes of metabolomics analysis. (B) Heatmap of glycolysis and tricarboxylic acid cycle (TCA) alterations was generated by comparison with their control with increased metabolite level in red and decreased metabolite level in white. The numbers represent the fold changes to Con. (C) The ratio of lactate/pyruvate and citrate/pyruvate in NGF-PC12 cells cultured in the presence of HG (30 mM glucose) with or without 50 μM 6-OHDA. (D) Colorimetric assay assessment of extracellular lactate in NGF-PC12 cells cultured in the presence of HG (30 mM glucose) with or without 50 μM 6-OHDA. (E) Representative immunoblots of PFKFB3, PKM1, PKM2, LDHA, PDHB and β-actin in NGF-PC12 cells cultured in the presence of HG (30 mM glucose) with or without 50 μM 6-OHDA (Cropped blot images are shown; see Fig. S7 for full immunoblots). (F–J) Quantification of PFKFB3 (F), PKM2 (G), LDHA (H), PDHB (I), and PKM1 (J) normalized to β-actin, respectively. Data are shown as mean ± standard error of mean (SEM); n = 3, biologically independent samples. A one-way ANOVA was used for statistical analysis, followed by a Tukey's multiple comparisons test. ^∗P < 0.05, ^∗∗P < 0.01, ^∗∗∗P < 0.001, vs. Con; ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 vs. 6-OHDA. NGF: nerve growth factor; NGF-PC12 cells: differentiated PC12 cells by nerve growth factor; HG: high glucose; 6-OHDA: 6-hydroxydopamine; Pyr: pyruvate; Lac: lactate; Cit: citrate; i-Cit: isocitrate; Suc: succinate; Mal: malate; FC: fold change; Lac/Pyr: the ratio of lactate to pyruvate; Cit/Pyr: the ratio of citrate to pyruvate; PFKFB3: 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 3; PKM1: pyruvate kinase 1; PKM2: pyruvate kinase 2; LDHA: lactate dehydrogenase A; PDHB: pyruvate dehydrogenase E1 subunit beta.

3.5. High glucose accelerated 6-OHDA-induced mitochondrial deletion

We demonstrated that HG promoted 6-OHDA-induced hyperglycolysis in NGF-PC12 cells. It is well known that all neurons are highly dependent on the mitochondria for survival. We hypothesized that HG may promote mitochondrial damage caused by 6-OHDA and then increase glycolytic pathway stress in NGF-PC12 cells. Therefore, we first examined the mitochondrial levels using the MitoTracker assay. Figs. 4A and B show that the subthreshold dose of 6-OHDA resulted in a great loss of neuronal mitochondria. Surprisingly, like 6-OHDA, HG also reduced cellular mitochondrial levels synergistically with 6-OHDA. In addition, we performed a JC-1 assay to evaluate mitochondrial function. The results showed that HG or 6-OHDA alone could decrease the MMP, and HG enhanced the effect of 6-OHDA on MMP in NGF-PC12 cells (Fig. 4C). TEM revealed that the size of mitochondria in cells treated with HG or 6-OHDA was larger than that in control cells, whereas the number of mitochondria in HG or 6-OHDA was less than that in control cells (Fig. 4D). Compared to 6-OHDA alone, treatment with the combination of HG and 6-OHDA resulted in larger and fewer mitochondria (Fig. 4D). Accordingly, we hypothesized that HG might disrupt the mitochondrial dynamics. Therefore, we examined the expressions of mitochondrial dynamics proteins. As is shown in Figs. 4E and F, HG or 6-OHDA alone promoted the expression of MFN2 and inhibited the expression of FIS1 (see Fig. S8 for full immunoblots). These results suggested that HG accelerated 6-OHDA-induced mitochondrial deletion by promoting normal mitochondrial fusion.

Fig. 4.

High glucose accelerated 6-OHDA-induced mitochondrial deletion. (A) Representative fluorescence images of mito-tracker (in green) in NGF-PC12 cells cultured in the presence of HG (30 mM glucose) with or without 50 μM 6-OHDA. (B) Mito-tracker fluorescence intensity was detected by microplate analyzer. (C) JC-1 assay was used to measure the mitochondrial membrane potential of NGF-PC12 cells cultured in the presence of HG (30 mM glucose) with or without 50 μM 6-OHDA. (D) Representative transmission electron microscopy images of mitochondrial in NGF-PC12 cells cultured in the presence of HG (30 mM glucose) with or without 50 μM 6-OHDA. Arrowheads denote mitochondrial. (E) Representative immunoblots of MFN2, FIS1 and β-actin in NGF-PC12 cells cultured in the presence of HG (30 mM glucose) with or without 50 μM 6-OHDA (Cropped blot images are shown; see Fig. S8 for full immunoblots). (F) Quantification of MFN2 and FIS1 normalized to β-actin, respectively. Data are shown as mean ± standard error of mean (SEM); n = 3, biologically independent samples. A one-way ANOVA was used for statistical analysis, followed by a Tukey's multiple comparisons test. ^∗∗∗P < 0.001, vs. Con; ^###P < 0.001, vs. 6-OHDA. NGF: nerve growth factor; NGF-PC12 cells: differentiated PC12 cells by nerve growth factor; HG: high glucose; 6-OHDA: 6-hydroxydopamine; MMP: mitochondrial membrane potential; MFN2: mitochondrial fusion protein 2; FIS1: mitochondrial fusion 1 protein.

3.6. Pharmacological inhibition of PKM2 could reduce mitochondrial loss and cell damage induced by 6-OHDA under HG treatment

To confirm the key role of PKM2 in HG-induced vulnerabiity of neurons to 6-OHDA, we investigated whether shikonin as a PKM2 inhibitor could improve HG-induced vulnerabiity of neurons to 6-OHDA [48]. We first investigated the levels of mitochondria and found that SK alone had no effect on the cell mitochondria (Figs. 5A and B). However, it significantly reduced the mitochondrial deletion induced by the combined treatment with HG and 6-OHDA (Figs. 5A and B). The JC-1 assay confirmed that SK prevented the HG+6-OHDA-induced decrease in MMP (Fig. 5C). Thus, we hypothesized that SK might regulate mitochondrial dynamics by inhibiting the mitochondrial fusion induced by HG+6-OHDA in NGF-PC12 cells. Consistent with our hypothesis, TEM showed mitochondrial fission in HG+6-OHDA-induced NGF-PC12 cells treated with SK (Fig. 5D). To further confirm this result, we performed Western blotting to examine changes in the levels of mitochondrial proteins. As shown in Figs. 5E and F, the levels of MFN2 decreased and the levels of FIS1 increased in HG+6-OHDA-induced cells after SK treatment (see Fig. S9 for full immunoblots). Previous studies have shown that HG promotes 6-OHDA-induced aberrant mitochondrial fusion, leading to apoptosis. SK may also inhibit HG+6-OHDA-induced apoptosis. The results shown in Figs. 5E and F show that the apoptotic protein Bax and the ratio of Bax/Bcl2 were significantly downregulated and the anti-apoptotic protein Bcl2 was significantly upregulated in HG+6-OHDA-induced NGF-PC12 cells after SK intervention. Consistently, TUNEL staining also confirmed this result (Figs. 5G and H). Finally, we examined cell viability using the SRB assay, and the results showed that SK ameliorated cell damage (Fig. 5I).

Fig. 5.

Inhibition of PKM2 can reduce mitochondrial loss and cell damage induced by 6-OHDA under HG treatment. (A) Representative fluorescence images of mito-tracker (in green) in NGF-PC12 cells cultured in the presence of HG (30 mM glucose) and 50 nM SK with or without 50 μM 6-OHDA. (B) Mito-tracker fluorescence intensity was detected by microplate analyzer. (C) JC-1 assay was used to measure the mitochondrial membrane potential of NGF-PC12 cells cultured in the presence of HG (30 mM glucose) and 50 nM SK with or without 50 μM 6-OHDA. (D) Representative transmission electron microscopy (TEM) images of mitochondria in NGF-PC12 cells cultured in the presence of HG (30 mM glucose) and 50 nM SK with or without 50 μM 6-OHDA. Arrowheads denote mitochondria. (E) Representative immunoblots of MFN2, FIS1, Bax, Bcl2 and β-actin of NGF-PC12 cells cultured in the presence of HG (30 mM glucose) and 50 nM SK with or without 50 μM 6-OHDA (Cropped blot images are shown; see Fig. S9 for full immunoblots). (F) Quantification of MFN2, FIS1, Bax and Bcl2 normalized to β-actin, respectively. (G) Reprentative images of DAPI and TUNEL staining of NGF-PC12 cells cultured in the presence of HG (30 mM glucose) and 50 nM SK with or without 50 μM 6-OHDA. (H) Quantification of TUNEL⁺ apoptotic cells in NGF-PC12 cells cultured in the presence of HG (30 mM glucose) and 50 nM SK with or without 50 μM 6-OHDA. (I) Cell viabilities were tested by sulforhodamine B (SRB) in NGF-PC12 cells cultured in the presence of HG (30 mM glucose) and 50 nM SK with or without 50 μM 6-OHDA. Data are shown as mean ± standard error of mean (SEM); n = 3, biologically independent samples. A one-way ANOVA was used for statistical analysis, followed by a Tukey's multiple comparisons test. ^∗∗∗P < 0.001, HG+6-OHDA vs. Con; ^###P < 0.001, HG+6-OHDA+SK vs. HG+6-OHDA. NGF: nerve growth factor; NGF-PC12 cells: differentiated PC12 cells by nerve growth factor; HG: high glucose; 6-OHDA: 6-hydroxydopamine; SK: shikonin; MMP: mitochondrial membrane potential; MFN2: mitochondrial fusion protein 2; FIS1: mitochondrial fusion 1 protein; Bax: Bcl2 associated X protein; Bcl2: B-cell lymphoma-2; TUNEL: terminal deoxynucleotidyl transferase dUTP nick end labeling; DAPI: 4′,6-diamidino-2-phenylindole.

To exclude the possibility that the effects of SK are due to its non-PKM2 inhibitory pharmacological activity, we used siRNA to specifically silence PKM2. TUNEL staining showed that siRNA PKM2 also rescued apoptosis induced by HG+6-OHDA in NGF-PC12 cells (Figs. S10A and B). Next, we found that silencing of PKM2 reduced the level of Bax and the raito of Bax/Bcl2 and promoted the expression of Bcl2 (Figs. S10C and D, see Fig. S11 for full immunoblots). In addition, Western blotting showed that PKM2 silencing maintained mitochondrial homeostasis by inhibiting abnormal mitochondrial fusion and increaseing mitochondrial numbers and MMP (Figs. S10E–G). Thus, we preliminarily determined that PKM2 mediated neuronal vulnerability to HG+6-OHDA via the regulation of mitochondrial homeostasis and cell apoptosis.

3.7. Pharmacological inhibition of PKM2 contributed to the flow of glucose-derived carbon into the TCA cycle

To further explore whether the potential protective mechanism of SK was related to cellular metabolic reprogramming, we performed untargeted metabolomics to assess its effect on the metabolic characteristics of NGF-induced PC12 neurons under the combined intervention with HG and 6-OHDA. PCA and cluster analysis showed that SK treatment restored the metabolite composition of cells to that of the control group (Fig. 6A). Next, Western blotting was used to determine the levels of key glycolytic enzymes. Compared to the combined intervention with HG and 6-OHDA, SK significantly reduced PFKFB3, PKM2 and LDHA expressions and upregulated PKM1 expression (Figs. 6B–F, see Fig. S12A for full immunoblots). The metabolomic study showed that SK significantly reduced intracellular lactate and upregulated TCA intermediates such as citrate, isocitrate, succinate, and malate (Fig. 6G). Colorimetry showed that SK reduced the levels of extracellular lactate (Fig. 6H). The ratio of Lac/Pyr decreased significantly in the SK treatment group, whereas the ratio of Cit/Pyr increased significantly after SK administration (Fig. 6I). At the same time, PDHB levels were significantly upregulated in HG+6-OHDA induced NGF-PC12 cells after SK treatment (Figs. 6J and K, see Fig. S12B for full immunoblots). Furthermore, silencing PKM2 also inhibited the glycolytic enzymes PKM2 and LDHA and promoted the expressions of PKM1 and PDHB, which was comparable to the regulation of glycolytic metabolic enzymes by SK in HG+6-OHDA-induced NGF-PC12 cells (Fig. S13, see Fig. S14 for full immunoblots). In conclusion, these results demonstrated that SK inhibited PKM2 and promoted PKM1 and PDHB, changing cellular metabolism from glycolysis to oxidative phosphorylation, thus increasing neuronal tolerance to 6-OHDA in HG environments.

Fig. 6.

Inhibition of PKM2 promotes the tricarboxylic acid cycle (TCA) in HG+6-OHDA induced PC12-derived neurons. (A) Principal component analysis (PCA) analysis of cell metabolism characteristics under positive and negative ion modes of metabolomics analysis. (B) Representative immunoblots of PFKFB3, PKM1, PKM2, LDHA and β-actin of NGF-PC12 cells cultured in the presence of HG (30 mM glucose) and 50 nM SK with or without 50 μM 6-OHDA (Cropped blot images are shown; see Fig. S12A for full immunoblots). (C–F) Quantification of PFKFB3 (C), PKM2 (D), LDHA (E) and PKM1 (F) normalized to β-actin, respectively. (G) Heatmap of glycolysis and TCA alterations was generated by comparison with their control with increased metabolite level in red and decreased metabolite level in white. The numbers represent the fold changes to Con. (H) Colorimetric assay assessment of extracellular lactate of NGF-PC12 cells cultured in the presence of HG (30 mM glucose) and 50 nM SK with or without 50 μM 6-OHDA. (I) The ratio of lactate/pyruvate and citrate/pyruvate in NGF-PC12 cells cultured in the presence of HG (30 mM glucose) and 50 nM SK with or without 50 μM 6-OHDA (J) Representative immunoblots of PDHB and β-actin of NGF-PC12 cells cultured in the presence of HG (30 mM glucose) and 50 nM SK with or without 50 μM 6-OHDA (Cropped blot images are shown; see Fig. S12B for full immunoblots). (K) Quantification of PDHB normalized to β-actin. Data are shown as mean ± standard error of mean (SEM); n = 3, biologically independent samples. A one-way ANOVA was used for statistical analysis, followed by a Tukey's multiple comparisons test. ^∗∗∗P < 0.001, HG+6-OHDA vs. Con; ^###P < 0.001, HG+6-OHDA+SK vs. HG+6-OHDA. NGF: nerve growth factor; NGF-PC12 cells: differentiated PC12 cells by nerve growth factor; HG: high glucose; 6-OHDA: 6-hydroxydopamine; SK: shikonin; Pyr: pyruvate; Lac: lactate; Cit: citrate; i-Cit: isocitrate; Suc: succinate; Mal: malate; FC: fold change; Lac/Pyr: the ratio of lactate to pyruvate; Cit/Pyr: the ratio of citrate to pyruvate; PFKFB3: 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 3; PKM1: pyruvate kinase 1; PKM2: pyruvate kinase 2; LDHA: lactate dehydrogenase A; PDHB: pyruvate dehydrogenase E1 subunit beta.

3.8. Pharmacological inhibition of PKM2 rescued motor dysfunction and nigrostriatal neurodegenerative damage in diabetic rats after 6-OHDA administration

To examine whether SK could prevent the PD process in diabetic rats, we orally administered SK to 6-OHDA-treated diabetic rats every day for 4 weeks and then conducted behavioral tests (Fig. 7A). SK significantly improved motor dysfunction in 6-OHDA-treated diabetic rats (Figs. 7B–E). In the rota-rod, FAS, cylinder, and amphetamine rotation tests, the motor function performance of the SK-treated rats was significantly better than that of the 6-OHDA-treated diabetic rats (Figs. 7B–E). In addition, consistent with the Western blotting results (Figs. 7F and G, see Fig. S15 for full immunoblots), immunohistochemical results showed that the loss of TH immunopositive fibers in the striatum of SK-treated rats was significantly lower than that of STZ-diabetic rats induced by unilateral microinjection of the subthreshold dose of 6-OHDA (Figs. 7H and I). We then measured the number of TH-immunoreactive neuronal cell bodies in the SNc. In 6-OHDA-induced STZ-diabetic rats, SK increased the number of TH-positive neurons (Figs. 7H and J). In conclusion, both behavioral and pathological findings suggested that SK prevented PD in diabetic rats. However, unlike insulin treatment, SK did not reduce glucose levels in the peripheral blood and cerebrospinal fluid, which was consistent with the results of our cell experiments (Fig. S16). SK protected neurons from 6-OHDA injury by regulating cell metabolism and making cells more tolerant to the extracellular high-glucose environment rather than directly reducing the glucose levels through insulin-like effects.

Fig. 7.

Pharmacological inhibition of PKM2 reduces motor dysfunction and nigrostriatal neurodegenerative damage in debatic rats after 6-OHDA administration. (A) Schematic diagram of the animal experimental design. (B–E) Data from non-diabetic or STZ-diabetic rats evaluated on rota-rod test (B), FAS test (C), cylinder test (D) and AP rotation test (E) (n=7). (F) Representative immunoblots of TH and β-actin of the striatum (Cropped blot images are shown; see Fig. S15 for full immunoblots). (G) Quantification of TH normalized to β-actin (n=3). (H) Representative photomicrographs of coronal sections of the substantia nigra and striatum of nondiabetic or STZ-diabetic rats immunostained for TH. (I–J) Histograms represent the proportional stained area of immunoreactivity (n=3). TH⁺ fibers in striatum (I), and TH⁺ cells in SNc (J). Data are shown as mean ± standard error of mean (SEM). A one-way ANOVA was used for statistical analysis, followed by a Tukey's multiple comparisons test. ^∗∗∗P < 0.001 vs. Con; ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 vs. STZ+6-OHDA. SD-Rat: Sprague-Dawley rats; STZ: streptozotocin; T1DM: type 1 diabetes mellitus; Glc: glucose; SK: shikonin; AP: amphetamines; FAS: forepaw adjusting steps; SNc: substantia nigra pars compacta; TH: tyrosine hydroxylase; L/R: left/right.

4. Discussion

STZ-induced diabetic rats are widely used as a T1DM. STZ selectively destroys pancreatic islets, leading to a lack of insulin and hyperglycemia [49,50]. In agreement with previous studies, STZ-treated rats exhibited increased vulnerability to nigrostriatal neurodegeneration [51]. Here, we showed that increased vulnerability to 6-OHDA in diabetic animals was associated with hyperglycemia-induced upregulation of PKM2 in neurons, which caused hyperglycolysis and abnormal mitochondrial fusion and apoptosis.

Growing evidences from epidemiological and animal studies indicate that pre-existing diabetes increases the risk of developing PD [[52], [53], [54]]. Recently, researchers have suggested mechanisms of such an association, in which DM promotes striatal oxidative stress, alters dopamine neurotransmission, and increases vulnerability to neurodegenerative damage leading to motor impairment [51]. However, the mechanism by which diabetes promotes oxidative stress damage in neurons remains unclear. Our study found that susceptibility to PD in diabetic rats might be related to high central glucose levels. As shown in Fig. 1, unilateral microinjections of a subthreshold dose of 6-OHDA significantly increased the levels of glucose in the cerebrospinal fluid and exacerbated movement disorder in STZ-induced diabetic rats compared to nondiabetic controls. In addition, diabetic rats showed significantly increased loss of TH-immunopositive fibers in the striatum and SNc following treatment with 6-OHDA, compared to control, 6-OHDA-treated non-diabetic rats. To further confirm the relationship between HG levels and motor dysfunction, we treated rats with 6-OHDA + STZ and insulin. Interestingly, insulin reduced glucose levels in the cerebrospinal fluid and improved motor dysfunction induced by 6-OHDA while reducing the loss of TH immunopositive neurons in STZ-diabetic rats. In conclusion, our study highlights that HG levels are a key factor leading to susceptibility to PD in diabetic rats.

However, the effect of hyperglycemia on neuronal injury remains unclear. Current studies mostly support two mechanisms to explain the increased risk of PD caused by hyperglycemia: protein glycosylation and oxidative stress [[55], [56], [57], [58], [59]]. In addition, some studies have suggested that it may be related to neuroinflammation and neurotoxins produced by hyperglycemia [60,61]. However, these studies are limited to cell damage phenotypes, and the exact drivers remain unknown. Cellular metabolism is often thought to be the engine that drives signaling pathways. Metabolomics can monitor metabolic changes in neurons globally [33], and we found that HG promoted the flow of a large number of glucose metabolism intermediates to lactic acid, but not to TCA. At the same time, 6-OHDA had an effect on neurons similar to that of high glucose., which upregulated the neuronal glycolysis pathway, promoted the conversion of pyruvate to lactic acid, and weakened TCA. This difference was further amplified in neurons treated with HG and 6-OHDA. Therefore, we hypothesize that 6-OHDA may act through further activation of HG-induced hyperglycolysis, leading to cell damage. Combined with metabolomic data analysis, we found that PKM2 was upregulated in neurons treated with HG or 6-OHDA, while PKM1 and PDHB were decreased. In addition, the changes in PKM1-PDHB and PKM2 were significantly enhanced in neurons treated with HG and 6-OHDA, which was consistent with the above results of cellular glucose metabolism (Fig. 3). It has been reported that PKM1 promotes the oxidative phosphorylation of citrate to produce ATP to adapt to the high energy demand of neurons, whereas PKM2 mainly mediates glycolysis to promote biosynthesis and meet the high synthesis demand of stem cells [[56], [57], [58]]. Therefore, we hypothesize that HG levels enhance the susceptibility of neurons to 6-OHDA through the upregulation of PKM2-mediated glycolysis.

Therefore, the activation of glycolysis by HG enhances the vulnerability of neurons to 6-OHDA. Combined with the literature reports and our current experimental results, we hypothesized that the key reason for excessive glycolysis caused by HG or 6-OHDA intervention alone may be closely related to mitochondrial homeostasis. Because neuronal mitochondria are highly active, when stimulated by HG or 6-OHDA, their function is impaired, leading to the passive activation of cellular glycolysis. Our results also confirmed that treatment with HG or 6-OHDA alone was sufficient to disturb cellular mitochondrial homeostasis, leading to abnormal mitochondrial fusion and reduced membrane potential, which was mediated by PKM2. This may explain why neurons treated with HG were more sensitive to 6-OHDA. This was because HG levels altered mitochondrial homeostasis and led to increased abnormal mitochondrial fusion. At this time, 6-OHDA intervention further aggravated mitochondrial dyshomeostasis, leading to a decline in mitochondrial membrane potential, and apoptosis. Therefore, the HG-induced upregulation of PKM2 and abnormal fusion of mitochondria are the root cause of 6-OHDA vulnerability.

According to the discussion above, our study explained the connection between DM and PD from the perspective of neuronal metabolic reprogramming and confirmed that diabetes-mediated HG levels in the brain are the key cause of progressive neuronal loss. To some extent, this study provides evidence for the possible association between glucose metabolism disorder and PD in patients with DM and also suggests possible mechanisms of this association.

SK, obtained from the roots of Lithospermum erythrorhizon (Boraginaceae), is a PKM2 inhibitor (IC₅₀ = 2.95 μM) that has anti-cancer, anti-inflammatory, anti-bacterial, and other pharmacological activities [48,62,63]. From the above results, we infer that SK may maintain the balance of mitochondrial dynamics and reprogram neuronal metabolism, leading to the transition from glycolysis to the TCA oxidative phosphorylation pathway via inhibition of PKM2, thus reducing HG-induced vulnerability of neurons to 6-OHDA. Consistent with this, our study found that SK inhibited PKM2 expression and alleviated abnormal mitochondrial fusion and apoptosis of NGF-PC12 cells following their treatment with HG+6-OHDA. A metabolomic study found that SK promoted intermediate metabolite levels in TCA rather than glycolysis, further confirming the key role of PKM2 in HG-induced vulnerability of neurons to 6-OHDA. To further clarify the pharmacological significance of PKM2 inhibition in PD development, animal experiments were conducted to evaluate its efficacy. The results showed that SK improved motor dysfunction induced by 6-OHDA and reduced the loss of TH-immunopositive neurons in STZ-induced diabetic rats. However, unlike insulin, SK did not reduce glucose levels in rat cerebrospinal fluid. As both insulin and SK can prevent the DM-mediated occurrence and development of PD, we can make sense from the perspective of their functional characteristics. Insulin blocks the occurrence and development of 6-OHDA-induced PD in rats by reducing extracellular levels of glucose, which also highlights the key role of HG as a risk factor for PD in diabetic rats. SK reshapes the metabolic characteristics of neurons, resulting in neurons with a TCA oxidative phosphorylation pathway sufficient to tolerate high extracellular glucose, reducing its effect on neuronal metabolism, and inhibiting HG-induced hypersensitivity of neurons to 6-OHDA.

In conclusion, metabolomics can effectively identify metabolic disorders in the body and identify their possible mechanisms. Starting from metabonomics, our study found that the increased risk for diabetic PD was related to HG-induced promotion of PKM2-mediated neuronal hyperglycolysis. By inhibiting PKM2, neuronal metabolic characteristics that were dependent on mitochondrial energy could be restored by SK, thereby preventing the occurrence and development of diabetic PD. Our study provides mechanistic support for the possible development of SK as a clinical drug to improve motor function in DM patients who develop PD and expands the range of clinical indications for SK, which is no longer limited to adjuvant therapy for hepatobiliary diseases. Furthermore, SK is expected to reduce a risk of PD in the population with DM.

CRediT author statement

Ya Zhao and Yanwei Wang: Project administration, Methodology, Validation, Investigation, Writing - Original draft preparation; Yuying Wu: Investigation, Writing - Original draft preparation; Cimin Tao and Rui Xu: Software, Validation, Data curation; Yong Chen and Linhui Qian: Writing - Reviewing & Editing; Tengfei Xu and Xiaoyaun Lian: Conceptualization, Supervision, Resources, Project administration.

Declaration of competing interest

The authors declare that there are no conflicts of interest.

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Fig. S1

The siRNA PKM2 was screened by western blotting. (A) Representative immunoblots of PKM2 and β-actin of PC12 cells cultured in the presence of siRNA PKM2 or siRNA Con. (cropped blot images are shown, see Supplementary Fig. S2 for full immunoblots). SiRNA Con: siRNA negative control; siRNA PKM2^#1: PKM2 (Norway rat) siRNA-468 (5′-CCGCAGAGGUGGAGCUGAATT-3′); siRNA PKM2^#2: PKM2 (Norway rat) siRNA-867 (5′-GCAAGAACAUCAAGAUCAUTT-3′); siRNA PKM2^#3: PKM2 (Norway rat) siRNA-640 (5′-GGAGAAAGGUGCUGACUACTT-3′).

Fig. S2

Uncropped figures from western blots. Uncropped western blot images that correspond to Fig. S1. PKM2: pyruvate Kinase 2.

Fig. S3

The glucose levels of blood and cerebrospinal fluid. Colorimetric assay assessment of glucose of blood and cerebrospinal fluid. Data are shown as mean ± SEM; n = 7, biologically independent animals. A one-way ANOVA was used for statistical analysis followed by a Tukey's multiple comparisons test. ∗∗∗P < 0.001, vs Con; ^###P < 0.01, vs STZ+6-OHDA. STZ: streptozotocin; T1DM: type 1 diabetes mellitus; Ins: insulin.

Fig. S4

Uncropped figures from western blots. Uncropped western blot images that correspond to Fig. 1F. TH: Tyrosine Hydroxylase.

Fig. S5

Toxicity test of glucose and 6-OHDA onNGF-PC12 cells. Cell viability were tested by SRB assay. (A) The toxicity of Glc on NGF-PC12 cells. (B) The toxicity of 6-OHDA on NGF-PC12 cells. Data are shown as mean ± SEM; n = 5, biologically independent samples. A one-way ANOVA was used for statistical analysis followed by a Tukey's multiple comparisons test. ∗∗P < 0.01, ∗∗∗P < 0.001 vs Con. Glc, Glucose; NGF-PC12 cells: differentiated PC12 cells by nerve growth factor; 6-OHDA: 6-Hydroxydopamine; SRB: sulforhodamine B.

Fig. S6

Uncropped figures from western blots. Uncropped western blot images that correspond to Fig. 2C. Bax: Bcl-2AssaciatedXprotein; Bcl-2: B-cell lymphoma-2.

Fig. S7

Uncropped figures from western blots. Uncropped western blot images that correspond to Fig. 3E. PFKFB3: 6-Phosphofructo-2-Kinase/Fructose-2,6-Biphosphatase 3; PKM1: pyruvate Kinase 1; PKM2: pyruvate Kinase 2; LDHA: lactate dehydrogenase A; PDHB: pyruvate dehydrogenase E1 subunit beta.

Fig. S8

Uncropped figures from western blots. Uncropped western blot images that correspond to Fig. 4E. MFN2: mitochondrial fusion protein 2; FIS1: mitochondrial Fission 1 Protein.

Fig. S9

Uncropped figures from western blots. Uncropped western blot images that correspond to Fig. 5E (A). MFN2: mitochondrial fusion protein 2; FIS1: mitochondrial Fission 1 Protein; Bax: Bcl-2AssaciatedXprotein; Bcl-2: B-cell lymphoma-2.

Fig. S10

Silence of PKM2 can reduce mitochondrial loss and cell damage induced by 6-OHDA under HG treatment. (A) Reprentative images of DAPI and TUNEL staining of NGF-PC12 cells cultured in the presence of HG (30 mM Glc) with or without 50 μM 6-OHDA after silence PKM2. Scale bar, 50 μm. (B) Quantification of TUNEL⁺ apoptotic cell in NGF-PC12 cells cultured in the presence of HG (30 mM Glc) with or without 50 μM 6-OHDA after silence PKM2. (C) Representative immunoblots of Bax, Bcl2, MFN2, FIS1 and β-actin of NGF-PC12 cells cultured in the presence of HG (30 mM Glc) with or without 50 μM 6-OHDA after silence PKM2. (cropped blot images are shown, see Figure S9B for full immunoblots). (D-E) Quantification of Bax, Bcl2, Bax/Bcl2, MFN2 and FIS1 normalized to β-actin, respectively. (F) Mito-tracker fluorescence intensity was detected by microplate analyzer. (G) JC-1 assay was used to measured the mitochondrial membrane potential of NGF-PC12 cells cultured in the presence of HG (30 mM Glc) with or without 50 μM 6-OHDA after silence PKM2. Data are shown as mean ± SEM; n = 3, biologically independent samples. A one-way ANOVA was used for statistical analysis followed by a Tukey's multiple comparisons test. ∗∗∗P < 0.001, siRNA Con+HG+6-OHDA vs Con; ^###P < 0.001, siRNA PKM2+HG+6-OHDA vs HG+6-OHDA. NGF: nerve growth factor; NGF-PC12 cells: differentiated PC12 cells by nerve growth factor; Glc: glucose; HG: high glucose; 6-OHDA: 6-Hydroxydopamine; TUNEL: Terminal deoxynucleotidyl transferase dUTP nick end labeling; DAPI: 4′,6-diamidino-2-phenylindole; Bax: Bcl-2AssaciatedXprotein; Bcl-2: B-cell lymphoma-2; MFN2: mitochondrial fusion protein 2; FIS1: mitochondrial fission 1 Protein.

Fig. S11

Uncropped figures from western blots. Uncropped western blot images that correspond to Fig. S10C. MFN2: mitochondrial fusion protein 2; FIS1: mitochondrial Fission 1 Protein; Bax: Bcl-2AssaciatedXprotein; Bcl-2: B-cell lymphoma-2.

Fig. S12

Uncropped figures from western blots. Uncropped western blot images that correspond to Fig. 6B (A), Fig. 6J (B). PFKFB3: 6-Phosphofructo-2-Kinase/Fructose-2,6-Biphosphatase 3; PKM1: pyruvate Kinase 1; PKM2: pyruvate Kinase 2; LDHA: lactate dehydrogenase A; PDHB: pyruvate dehydrogenase E1 subunit beta.

Fig. S13

Silence of PKM2 inhibited the key glycolysis enzymes by 6-OHDA under HG treatment. (A) Representative immunoblots of PKM1, PKM2, LDHA, PDHB and β-actin of NGF-PC12 cells cultured in the presence of HG (30 mM Glc) with or without 50 μM 6-OHDA after silence PKM2. (cropped blot images are shown, see Figure S9C for full immunoblots). (B-E) Quantification of PKM1, PKM2, LDHA and PDHB normalized to β-actin, respectively. Data are shown as mean ± SEM; n = 3, biologically independent samples. A one-way ANOVA was used for statistical analysis followed by a Tukey's multiple comparisons test. ∗∗∗P < 0.001, siRNA Con+HG+6-OHDA vs Con; ^###P < 0.001, siRNA PKM2+HG+6-OHDA vs HG+6-OHDA. NGF: nerve growth factor; NGF-PC12 cells: differentiated PC12 cells by nerve growth factor; Glc: glucose; HG: high glucose; 6-OHDA: 6-Hydroxydopamine; PKM1: pyruvate Kinase 1; PKM2: pyruvate Kinase 2; LDHA: lactate dehydrogenase A; PFKFB3: 6-Phosphofructo-2-Kinase/Fructose-2,6-Biphosphatase 3; PKM1: pyruvate Kinase 1; PKM2: pyruvate Kinase 2; LDHA: lactate dehydrogenase A; PDHB: pyruvate dehydrogenase E1 subunit beta.

Fig. S14

Uncropped figures from western blots. Uncropped western blot images that correspond to Fig. S12. PKM1: pyruvate Kinase 1; PKM2: pyruvate Kinase 2; LDHA: lactate dehydrogenase A; PDHB: pyruvate dehydrogenase E1 subunit beta.

Fig. S15

Uncropped figures from western blots. Uncropped western blot images that correspond to Fig. 7F. TH: Tyrosine Hydroxylase.

Fig. S16

The glucose levels of blood and cerebrospinal fluid. Colorimetric assay assessment of glucose of blood and cerebrospinal fluid. Data are shown as mean ± SEM; n = 7, biologically independent animals. A one-way ANOVA was used for statistical analysis followed by a Tukey's multiple comparisons test. ∗∗∗P < 0.001, vs Con. STZ: streptozotocin; T1DM: type 1 diabetes mellitus; SK: shikonin.