Metabolomic signature of type 1 diabetes-induced sensory loss and nerve damage in diabetic neuropathy


Diabetic-induced peripheral neuropathy (DPN) is a highly complex and frequent diabetic late complication, which is manifested by prolonged hyperglycemia. However, the molecular mechanisms underlying the pathophysiology of nerve damage and sensory loss remain largely unclear. Recently, alteration in metabolic flux has gained attention as a basis for organ damage in diabetes; however, peripheral sensory neurons have not been adequately analyzed with respect to metabolic dysfunction. In the present study, we attempted to delineate the sequence of event occurring in alteration of metabolic pathways in relation to nerve damage and sensory loss. C57Bl6/j wild-type mice were analyzed longitudinally up to 22 weeks in the streptozotocin (STZ) model of type 1 diabetes. The progression of DPN was investigated by behavioral measurements of sensitivity to thermal and mechanical stimuli and quantitative morphological assessment of intraepidermal nerve fiber density. We employed a mass spectrometry-based screen to address alterations in levels of metabolites in peripheral sciatic nerve and amino acids in serum over several months post-STZ administration to elucidate metabolic dysfunction longitudinally in relation to sensory dysfunction. Although hyperglycemia and body weight changes occurred early, sensory loss and reduced intraepithelial branching of nociceptive nerves were only evident at 22 weeks post-STZ. The longitudinal metabolites screen in peripheral nerves demonstrated that compared with buffer-injected age-matched control mice, mice at 12 and 22 weeks post-STZ showed an early impairment the tricaoboxylic acid (TCA cycle), which is the main pathway of carbohydrate metabolism leading to energy generation. We found that levels of citric acid, ketoglutaric acid (2 KG), succinic acid, fumaric acid, and malic acid were observed to be significantly reduced in sciatic nerve at 22 weeks post-STZ. In addition, we also found the increase in levels of sorbitol and L-lactate in peripheral nerve from 12 weeks post-STZ injection. Amino acid screen in serum showed that the amino acids valine (Val), isoleucine (Ile), and leucine (Leu), grouped together as BCAA, increased more than twofold from 12 weeks post-STZ. Similarly, the levels of tyrosine (Tyr), asparagine (Asn), serine (Ser), histidine (His), alanine (Ala), and proline (Pro) showed progressive increase with progression of diabetes. Our results indicate that the impaired TCA cycle metabolites in peripheral nerve are the primary cause of shunting metabolic substrate to compensatory pathways, which leads to sensory nerve fiber loss in skin and contribute to onset and progression of peripheral neuropathy.


Diabetes is a chronic metabolic disease marked by hyperglycemia as a result of dysfunction in insulin secretion and/or action. Data from the World Health Organization shows that global prevalence of diabetes among adult is 8.5% [1]. Diabetes-induced peripheral neuropathy (DPN) is a highly complex and prevalent diabetic complication observed in 50% of diabetic patients. DPN is characterized by progressive loss of peripheral nerve axons resulting in pain, loss of sensation, and eventually a leading cause of lower extremity amputation [2, 3]. Despite a long history of research on delineating the pathophysiology of the disease, we are now beginning to understand molecular mechanisms underlying DPN.

Hyperglycemia-induced oxidative stress is postulated to be a primary driver in diabetic complication and organ dysfunction [4]. Several studies showed that hyperglycemia causes mitochondrial dysfunction leading to overproduction of superoxide and free radicals [5]. Elevated ROS levels trigger alterations in transcriptional factors function and release of inflammatory cytokines and chemokines. Other cellular consequences include the release of cytochrome c, activation of caspase 3, activation of an endonuclease-G (Endo-G), altered biogenesis, and cell death [6]. Hyperglycemia-induced ROS generation is linked to both enzymatic and nonenzymatic pathways. The enzymatic pathways include nicotinamide adenine dinucleotide phosphate oxidase (NADPH oxidase) and uncoupling of nitric oxide synthase (NOS). The nonenzymatic pathways include aberrant functionality of complex 1 and complex III of Electron Transport Chain (ETS). The metabolic inflexibility of cell to switch between nutrient utilization, in particular between fatty acid and glucose oxidation, is postulated to be the primary reason for ROS formation. Emerging evidence suggests that alterations in glycolytic flux, signaling molecules, and mitochondrial enzymes contribute to tissue-specific adaptations in fuel utilization that are associated with tissue dysfunction [7].

Rodent and human studies results showed the downregulation of expression and activity of electron chain transport proteins in diabetic heart [8, 9] and skeletal muscles of patients with type 2 diabetes [10, 11]. Similarly, downregulation of mitochondrial enzymes in DRG and nerves has been observed after established DPN in genetically modified diabetic mice models, which do not replicate the clinical observation associated with disease progression. [12]. Moreover, the loss of peripheral nerve fibers in the epidermal layer of skin is considered to be a clinical diagnostic tool for DPN detection. Although these studies showed alterations in activity and expression of metabolic pathway enzymes in DRG, peripheral nerves, and nerve fiber in skin, what remains unanswered is whether these events occur after establishing DPN or they are the reasons for onset and development of DPN. However, to date, no attempts were made to investigate the progressive change happening in metabolic pathways before and during the development of DPN. Branched-Chain Amino Acid (BCAA), glucogenic, and ketogenic amino acids have reported to have critical role in metabolic health. Cross-section studies in type 2 diabetes patients and in rodents have reported that elevated levels of BCAA in serum contribute to the development of insulin resistance. Lack of longitudinal studies limits the association of amino acid to the progression of DPN. We therefore performed a comprehensive and unbiased amino acid screen to identify the signature tracking of the progression of DPN in the plasma of STZ-treated mice as compared with control buffer-injected mice.

In the present study, we have systemically analyzed the change in the kinetics of tri-carboxylic cycle (TCA) in peripheral nerve and amino acids in serum by employing mass spectrum-based measurement of TCA metabolites and amino acids at different points of time in mouse model of type 1 diabetes. We attempt to establish the link between observed metabolic alterations to pathological symptoms of DPN by performing behavioral analysis and changes in peripheral nerve fiber density in epidermal layer of skin. We report a series of changes in amino acids in nerves of diabetic mice and show that inhibition of glycolytic enzymes is an early event occurring in response to hyperglycemia.

Material and methods

Animal experiments

Age-matched 7–8-week-old C57BL6/j male mice were bought from Janvier Labs, Europe. Animals were maintained in humidity and temperature-controlled environment. Mice were housed in socially stable and well-nested individually ventilated cage-rack system (Techniplast, Italy) and had free access to water and food. All experiments were done in accordance with ARRIVE guidelines and approved by Regierungspräsidium Karlsruhe, Germany.

Rodent model of type 1 diabetic model

We employed low dose streptozotocin (STZ)-induced type 1 diabetic model for all the experiments. Multiple low-dose STZ (60 mg/kg/days, for 5 consecutive days) intraperitoneal (i.p.) administration leads to selective destruction of pancreatic beta cells. It is evident by increased levels of blood glucose. Blood glucose levels were monitored weekly using glucometer (Accu-Chek, Roche Diagnostics) for the entire course of the experiment. Mice only with blood glucose levels > 350 mg/dl were considered to be diabetic and included in experiments. Mice were analyzed over a period of 8 to 22 weeks post-STZ injection. Post-STZ mice blood glucose levels were maintained in a range between 350 and 500 mg/dl by sub-cutaneous administration of insulin.

Tissue extraction

Blood and tissue samples were collected at pre-diabetic and at different time points post-STZ injection. Age-matched citrate buffer-injected, nondiabetic mice were used as controls for blood and tissue collection. Mice were anesthetized under 2% isoflurane. The blood samples were drawn from the venous sinus using the retro-orbital bleeding method. The blood samples were collected from pre-diabetic, 8, 12, and 22 weeks post-STZ injection in EDTA-treated tubes, and plasma was separated by centrifugation. Sciatic nerve tissue was extracted from pre-diabetic, 12 and 22 weeks post-STZ-injected diabetic mice. The separated plasma and extracted sciatic nerve were snap-freezed using liquid nitrogen and stored at − 80 °C until analysis.

For paw punches biopsies, the mice at different time points post-STZ were perfused with 4% paraformaldehyde (PFA). The punch biopsies of the plantar skin of the hind paws were prepared and post-fixed in 4% PFA for 24 h at 4 °C. The tissue was incubated overnight in 30% sucrose, and 16-μm cryosections were made for immunohistochemical analysis.


To evaluate the degeneration of nerve fibers in the epidermal layer of diabetic mice, we performed immunohistochemistry using anti-CGRP and anti-PGP 9.5 antibodies [13]. Immunostaining was performed on punch biopsies of the plantar skin samples extracted from diabetic and nondiabetic C57BL6/j mice prior to and different time points post-STZ as described before. Briefly, cryosections were permeabilized in 0.5% PBST, washed, and blocked with 7% horse serum (HS). Sections were incubated overnight at 4 °C with anti-CGRP (1:1000, Sigma) primary antibody in 7% HS in PBS. Subsequently, sections were washed and incubated with Alexa Fluor-594-conjugated secondary antibody. Sections were washed and mounted in Mowiol. Fluorescence images were obtained using a laser-scanning spectral confocal microscope, and maximal projections were created using Leica SP8 software (Leica TCS SP8 AOBS, Bensheim, Germany). The acquired images were analyzed to quantify the fluorescence intensity from the epidermal area using ImageJ software.

Behavioral measurements

All behavioral experiments were approved by Regierungspräsidium Karlsruhe, Germany. Mice were acclimatized to the experimental setup twice a day for 3 days. The animals were randomized, and the experimenter was blinded for the identity of treatment given to mice. Age-matched citrate buffer-injected control mice were used for all behavioral experiments. Thermal sensitivity was measured using Hargreaves apparatus. Briefly, the paw withdrawal latency on application of infrared (IR) heat source on the plantar surface of hindpaw was recorded with a cutoff of 25 s. Two consecutive heat applications on the same mouse were separated with time interval of 5 min. Mechanical sensitivity was analyzed using Von Frey monofilaments. Mice were placed on an elevated grid, and Von Frey monofilaments with specific force were applied to the plantar surface of hindpaw. Von Frey filaments of 0.16, 0.4, 0.6, 1.0, 1.4, 2.0, and 4.0 g were tested to determine the mechanical sensitivity. The force range was chosen, such that the mice response from no response to 100% response. Each monofilament was applied for five times at time interval of 10 min on the plantar hindpaw. Forty percent response frequency was calculated as “thresholds” at basal, 12, and 22 weeks post-STZ injection.

Determination of amino acid levels

Amino acids from mice serum were quantified after specific labeling with the fluorescence dye AccQ-Tag™ (Waters) according to the manufacturer’s protocol. The resulting derivatives were separated by reversed phase chromatography on an Acquity BEH C18 column (150 mm × 2.1 mm, 1.7 μm, Waters) connected to an Acquity H-class UPLC system and quantified by fluorescence detection (Acquity FLR detector, Waters, excitation 250 nm, emission 395 nm) using ultrapure standards (Sigma). The column was heated to 42 °C and equilibrated with 5 column volumes of buffer A (140-mM sodium acetate pH 6.3, 7-mM triethanolamine) at a flow rate of 0.45 ml/min. Baseline separation of amino acid derivates was achieved by increasing the concentration of acetonitrile (B) in buffer A as follows: 1-min 8% B, 16-min 18% B, 23-min 40% B, 26.3-min 80% B, hold for 5 min, and return to 8% B in 3 min. Data acquisition and processing were performed with the Empower3 software suite (Waters).

Determination of metabolites by gas chromatography/mass spectrometry from sciatic nerve tissue


Frozen material was extracted in 180 μl of 100% MeOH for 15 min at 70 °C with vigorous shaking. Five-microliter ribitol (0.2 mg/ml) was added as internal standard to each sample. It was followed by addition of 100-μl chloroform to each sample. All samples were shaken at 37 °C for 5 min. Subsequently, 200 μl of water was added to each sample and centrifuged at 11,000 g for 10 min. Three hundred microliters of the polar (upper) phase was transferred to a fresh tube and dried in a speed-vac (Eppendorf vacuum concentrator) without heating and used for derivatization.

Derivatization (methoximation and silylation)

Pellets were re-dissolved in 20-μl methoximation reagent containing 20-mg/ml methoxyamine hydrochloride (Sigma 226904) in pyridine (Sigma 270970) and incubated for 2 h at 37 °C with shaking. For silylation, 32.2-μl N-methyl-N-(trimethylsilyl)-trifluoroacetamide (MSTFA; Sigma M7891) and 2.8-μl alkane standard mixture (50-mg/ml C10-C40; Fluka 68281) were added to each sample. After incubation for 30 min at 50 °C, samples were transferred to glass vials for gas chromatography/mass spectrometry (GC/MS) analysis.

GC/MS analysis

GC/MS-QP2010 Plus (Shimadzu®) fitted with a Zebron ZB 5 MS column (Phenomenex®; 30 m × 0.25 mm × 0.25 μm) was used for GC/MS analysis. The GC was operated at an injection temperature of 230 °C, and 1-μl sample was injected with split mode (1:10). The GC temperature program started with a 1-min hold at 40 °C followed by a 6-°C/min ramp to 210 °C, a 20-°C/min ramp to 330 °C, and a bake-out for 5 min at 330 °C using helium as the carrier gas with constant linear velocity. The MS was operated with ion source and interface temperatures of 250 °C, a solvent cut time of 8 min, and a scan range (m/z) of 40–700 with an event time of 0.2 s. The “GCMS solution” software (Shimadzu®) was used for data processing.

Statistical analysis

All the data were calculated and are presented as mean ± SEM. ANOVA for repeated measures and followed by Bonferroni’s test for multiple comparisons were employed to determine statistically significant differences. Changes with p ≤ 0.05 were considered to be significant.


Induction of STZ-induced DPN in mice

This study investigates alterations in amino acid profile in serum and the metabolic changes occurring in sciatic nerve over the course of progression of DPN. We employed the low-dose STZ model of type 1 diabetes, which does not involve direct neurotoxic effects. [14]. It is characterized by lymphocytic infiltration of pancreatic islets, leading to cell death, subsequently resulting in insulin deficiency and hyperglycemia [15]. We performed a long-term study in STZ-injected and age-matched citrate buffer-injected control mice to investigate the role of hyperglycemia-induced metabolic changes in the onset and development of DPN. Persistence of diabetes was examined by measuring blood glucose levels, and we had blood glucose levels to maintain all animals at uniform levels of hyperglycemia within the cohort. STZ-injected mice incessant showed blood glucose levels of > 400 mg/dl from 2 weeks onwards post-STZ injection over the entire course of the disease (Fig. 1a, p < 0.05, ANOVA). STZ-injected did not only suffer any weight loss over basal but also not gain any weight as age progressed, unlike buffer-injected age-matched controls at all time points (Fig. 1b, p < 0.05, ANOVA).

Fig. 1

Course and quantitative morphological analysis of DPN-associated pathology of peripheral nerves in STZ-injected C57Bl6/j mice and citrate buffer-injected age-matched control mice. a Blood glucose levels in C57BL6/j after STZ injection. b Course of body weight in mice after STZ-injection. c, d Summary of response thresholds to mechanical stimuli (defined as a force eliciting a response of paw withdrawal at least 40% response rate). e Identification of intra-epidermal nerve fibers via immunostaining for (i) CGRP or (ii) PGP 9.5 in the paw skin of STZ or buffer injected mice. The dotted white lines represent the epidermal layer in the skin. f Quantification of reduction in epidermal nerve fiber density post-STZ injection in STZ-injected and control mice. n = 8 mice per group. Data are represented as mean ± S.E.M. *p < 0.05 compared with baseline; # p < 0.05 as compared with the control group two-way ANOVA for repeated measurements with Bonferroni multiple comparison test. Scale bar represents 30 μm

We investigated STZ-injected diabetic and age-matched citrate buffer-injected mice in the behavioral model of evoked pain at basal, 12, and 22 weeks post-STZ injection to evaluate the course of DPN. Although behavioral changes have been reported [16], it was important to assess metabolic changes and sensory abnormalities in the same cohort of mice to derive meaningful relationships. The time point chosen for analyses is based on rational to address onset and development of DPN. STZ-treated mice showed no change in response threshold as compared with the control group on plantar application of Von Frey monofilaments at 12 weeks post-STZ injection (Fig. 1c). Similarly, in Hargreaves test, we did not find any significant difference between STZ-injected and control groups in withdrawal frequency at 12 weeks post-STZ injection (Fig. 1d). However, at 22 weeks post-STZ, we found that STZ-injected group of mice developed mechanical and thermal hypoalgesia, which demonstrate the pathophysiology of established DPN (Fig. 1c,d, *p < 0.05 two-way ANOVA). The control group of mice did not show any sign of neuropathy.

Assessment of Intra-Epidermal Nerve Fibers Density (IENFD) in skin biopsies is an important approach to appraise progression of the DPN, which is also used in the clinical diagnostic context [17]. We did not employ the nerve conduction velocity (NCV) measurements, which has been extensively used as one of the predictive indicator of diabetic neuropathy for diagnosis and assessment of disease progression in clinical trials. However, the major limitation of NCV studies is that it primarily measures only the large myelinated and motor fiber dysfunction. NCV testing failed to address any changes in small fiber functions, which are believed to be the earliest fibers damaged in diabetes [18, 19]. Sveen et al. [18] had shown that small fibers were more sensitive to metabolic changes than large fibers in long duration of type 1 diabetes. Although abnormal nerve parameter measured by NCV can associated with diabetic complications, but it cannot adequately evaluate the type and distribution of sensation loss and its severity. Furthermore, NCV is a complex, time-consuming procedure, and associated with numerous variable factors (e.g., nerve temperature at the time of recording), which are difficult to regulate in repetitive recordings in rodents. Considering the week attribute of NCV to diabetic complication, we focused on quantitative sensory testing and correlated the findings with nerve fiber density measurements and metabolic changes. Therefore, we attempt to correlate the IENFD with progression of diabetes in STZ-treated mice. We performed immunostaining using Calcitonin Gene-Related Peptide (CGRP) (Fig. 1e(i)) or labeling with Protein Gene Product 9.5 (PGP 9.5) (Fig. 1e(ii)) on skin sections to label nociceptive nerve endings, which are the main class of afferents in the epidermal zone. We found no significant changes in IENFD at 12 weeks post-STZ, but strikingly at 22 weeks post-STZ, IENFD was significantly reduced (Fig. 1e,f, *p < 0.05, ANOVA). The reduced IENFD and behavioral changes at 22 weeks post-STZ treatment fitted the sensory loss observed previously in STZ-treated mice [20].

Alteration in amino acid level with onset and progression of DPN

We performed a longitudinal study to identify the amino acid signature in serum to track the progression of DPN in STZ-treated mice. Citrate buffer-injected control mice do not showed any significant changes in the levels of BCAA, glucogenic, and ketogenic amino acids at 12 and 22 weeks after injection. We found that the amino acids valine (Val), isoleucine (Ile) and leucine (Leu), grouped together as BCAA [21], increased more than twofold from 12 weeks post-STZ (Fig. 2a, *p < 0.05 as compared with basal, ANOVA). Leucine showed nearly threefold increase in STZ-treated mice over age-matched citrate buffer-injected control mice at 22 weeks. Several gluconeogenic and ketogenic amino acids were unchanged at 12 or 22 weeks in STZ-treated mice as compared with the controls. For example, aspartate (Asp), glutamate (Glu), glutamine (Gln), glycine (Gly), methionine (Met), threonine (Thr), and phenylalanine (Phe) remained unchanged at 12 and 22 weeks post-STZ injection, whereas tyrosine (Tyr), asparagine (Asn), serine (Ser), histidine (His), alanine (Ala), and proline (Pro) (Fig. 2b,c,d, *p < 0.05, ANOVA) showed progressive increase in plasma at 12 weeks post-STZ injection. Our results thus indicate diverse changes in plasma levels of BCAA and some gluconeogenic and ketogenic amino acids building up over time as diabetic complications set in.

Fig. 2

Quantitative analysis of amino acid levels in serum of STZ-injected and control mice. a Branched amino acids. b Gluconeogenic and ketogenic. c Gluconeogenic amino acids. d Other amino acids levels in serum at basal, 8, 12, and 22 weeks post-STZ and citrate buffer-injected mice. Fold changes of amino acid in serum of STZ-injected mice were calculated with age-matched citrate buffer-injected controls. N = 5 mice in each group. *p < 0.05 compared with baseline; ANOVA for repeated measurements with Bonferroni multiple comparison tests

Reduced TCA cycle intermediates in sciatic nerve of diabetic mice

The progressive impact of STZ-induced hyperglycemia on the kinetics of glycolytic and TCA cycle has not been described in nerves so far. Therefore, we performed targeted LC/MS-MS-based metabolomics screen. We measured the levels of glycolytic and TCA cycle intermediates in sciatic nerve of mice at basal state, 12, and at 22 weeks post-STZ injection. The levels of citric acid, ketoglutaric acid (2 KG), succinic acid, fumaric acid, and malic acid were observed to be significantly reduced in sciatic nerve at 22 weeks post-STZ as compared with age-matched citrate buffer-injected control mice (Fig. 3a, *p < 0.05, ANOVA). There was a trend for reduction of some of the TCA metabolites at 12 weeks, but changes only became significantly evident at 22 weeks, i.e., at the stage when sensory loss and nerve damage are clearly present. Progressive reduction in TCA metabolites results in metabolic shift to alternative pathways to meet the energy requirement. We found a continuous increase in levels of sorbitol in the sciatic nerve biopsies (Fig. 3b, *p < 0.05, ANOVA) at 12 and 22 weeks in STZ-treated mice. Gradual increase in levels of sorbitol and diminishing kinetic of TCA cycle with prolonged diabetic state indicate that there is a shift of glucose metabolism from glycolytic pathway to polyol pathway, which is linked to tissue dysfunction. Further, we found the elevated levels of L-lactate (Fig. 3c, *p < 0.05, ANOVA) at 22 weeks post-STZ injection.

Fig. 3

Mass spectrometry-based quantitative analysis of citric acid metabolites in sciatic nerve of STZ-injected mice. a Citric acid metabolites. b Sorbitol levels. c L-Lactate levels in sciatic nerve isolated from STZ- and citrate buffer-injected mice at basal, 12, and 22 weeks post-injection. The graph represents fold change in metabolite levels over age-matched control mice. N = 5 in mice in each group. *p < 0.05 compared with baseline; ANOVA for repeated measurements with Bonferroni multiple comparison test


DPN is the major reason for morbidity among diabetic patients [22]. The pathogenic triggers leading to onset and progression of DPN remain unidentified. As the development of DPN is not a static process, there is a continuous need to identify the markers, which then track specific and temporal alteration in onset and progression of DPN. We attempt to identify the metabolic changes occurring in the peripheral nerve tissue and in serum prior to the development of DPN. The main findings revealed by this study are as follows: (i) the switching of metabolic flux from oxidative pathway to alternative pathways in peripheral nerve is an early event occurring when no detectable symptoms of nerve dysfunction and sensory loss are present and (ii) the alteration in metabolic pathways in sciatic nerve is accompanied by increased levels of BCAA in serum, which may be a potential prognostic marker to detect diabetic complications.

Several studies in diabetic rat and in genetically modified-mice repeatedly reported metabolic pathway dysfunction in both type 1 and 2 diabetes models [12, 23, 24]. Oxidative stress, lipid and protein oxidations, inhibition of metabolic enzymes, increased BCAA [25], downregulation of mitochondrial enzymes in DRG, and the nerve have been observed after established DPN [12]. In this study, we attempt to understand the link between neuropathic symptoms with metabolic dysfunction. We demonstrated that at 12 weeks post-STZ injection, there was no detectable nociceptive dysfunction. It is further validated by no detectable changes in nerve fiber density in the epidermal layer of skin. In contrast, at similar state, we detect reduced levels of citric acid and a trend towards impairment of TCA cycle. Aconitase enzyme, which catalyzes the conversion of citrate to isocitrate, is the most sensitive TCA enzyme for ROS inhibition [26]. Hyperglycemia-induced aconitase inhibition is previously reported in a rodent model with established DPN [26].

Impairment of the TCA cycle deviates metabolites to compensatory metabolic pathways for energy generation. In nerve tissue, it is evident by elevated levels of sorbitol and L-lactate at a similar stage. Increased levels of sorbitol indicate deviation of glucose from glycolysis/TCA pathway to polyol pathway. Elevated L-lactate levels are an indicator of mitochondrial dysfunction and shift from oxidative phosphorylation to anaerobic metabolic pathways. Our results demonstrate that the reduced kinetic of TCA cycle leads to metabolic shift to alternative pathway, which with prolonged state results in development of DPN. It is conceivable that these changes lead to further dysfunction of energy metabolism and contribute to the onset and development of DPN. At late stage of 22 weeks post-STZ, DPN was evident by developed mechanical and thermal hypoalgesia, reduced nerve fiber density, and complete sink in the levels of TCA intermediates. Previous studies in sural nerve demonstrated that DPN is also marked by reduced expression of glycolytic enzymes and compromised axonal transport [27]. Our results indicate that these hyperglycemia-induced site-specific molecular changes in metabolic pathways are the first line of events, which may cause increased distal oxidative stress and contribute to onset and development of DPN.

In an attempt to develop a noninvasive prognostic marker, we correlated the metabolic changes in peripheral nerve tissue to changes in levels of amino acid in serum. In diabetic patient, it was shown that insulin resistance switches the metabolism of cell to utilize glucogenic and ketogenic amino acid as a source of energy. It results in reduction of these amino acids over time. In contrast, the levels of BCAAs were reported to be increased in patients with established diabetic complication. Our longitudinal studies demonstrate a trend for increasing levels of BCAAs as early as 5 weeks post-STZ, which were significantly elevated at 12 and 22 weeks post-STZ injection. The levels of tyrosine were also found to be increased from 12 weeks post-STZ, which has been recently marked a potential marker for obesity-induced modulation of insulin signaling [28]. Asn, His, Ala, and Pro showed significant increase at 12 weeks post-STZ, indicating the development of diabetic complications.


In summary, we propose that inhibition of glycolytic enzymes is an initiation event in sciatic nerve occurring at a stage when there are no evident symptoms of DPN. Inhibition of TCA cycle shifts glucose metabolism to alternative pathways, which are marked by early increase in the levels of sorbitol and L-lactate in peripheral nerves post-diabetes induction. These early events are paralleled by changes in BCAA serum levels, which come about very early and may hold some prognostic value. These observations from longitudinal analyses provide insights into delineating the temporal sequence of anomalies and may be supportive in designing and testing novel therapeutic measures to prevent or reverse the progression of DPN.



reactive oxygen species




diabetic peripheral neuropathy


sciatic nerve


sensory neuron-specific


mass spectrometry


  1. 1.

    Mathers CD, Loncar D (2006) Projections of global mortality and burden of disease from 2002 to 2030. PLoS Med 3(11):e442

    PubMed  PubMed Central  Article  Google Scholar 

  2. 2.

    Vincent AM, Callaghan BC, Smith AL, Feldman EL (2011) Diabetic neuropathy: cellular mechanisms as therapeutic targets. Nat Rev Neurol 7(10):573–583

    CAS  PubMed  Article  Google Scholar 

  3. 3.

    Feldman EL, Nave KA, Jensen TS, Bennett DLH (2017) New horizons in diabetic neuropathy: mechanisms, bioenergetics, and pain. Neuron. 93(6):1296–1313

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  4. 4.

    Giacco F, Brownlee M (2010) Oxidative stress and diabetic complications. Circ Res 107(9):1058–1070

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  5. 5.

    Rolo AP, Palmeira CM (2006) Diabetes and mitochondrial function: role of hyperglycemia and oxidative stress. Toxicol Appl Pharmacol 212(2):167–178

    CAS  PubMed  Article  Google Scholar 

  6. 6.

    Brownlee M (2001) Biochemistry and molecular cell biology of diabetic complications. Nature 414(6865):813–820

    CAS  Article  PubMed  Google Scholar 

  7. 7.

    Obrosova IG, Pacher P, Szabó C, Zsengeller Z, Hirooka H, Stevens MJ, Yorek MA (2005) Aldose reductase inhibition counteracts oxidative-nitrosative stress and poly(ADP-ribose) polymerase activation in tissue sites for diabetes complications. Diabetes. 54(1):234–242

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  8. 8.

    Lashin OM, Szweda PA, Szweda LI, Romani AM (2006) Decreased complex II respiration and HNE-modified SDH subunit in diabetic heart. Free Radic Biol Med 40(5):886–896

    CAS  PubMed  Article  Google Scholar 

  9. 9.

    Yang JY, Yeh HY, Lin K, Wang PH (2009) Insulin stimulates Akt translocation to mitochondria: implications on dysregulation of mitochondrial oxidative phosphorylation in diabetic myocardium. J Mol Cell Cardiol 46(6):919–926

    CAS  PubMed  Article  Google Scholar 

  10. 10.

    Kelley DE, He J, Menshikova EV, Ritov VB (2002) Dysfunction of mitochondria in human skeletal muscle in type 2 diabetes. Diabetes 51(10):2944–2950

    CAS  PubMed  Article  Google Scholar 

  11. 11.

    Mogensen M, Sahlin K, Fernström M, Glintborg D, Vind BF, Beck-Nielsen H et al (2007) Mitochondrial respiration is decreased in skeletal muscle of patients with type 2 diabetes. Diabetes 56(6):1592–1599

    CAS  PubMed  Article  Google Scholar 

  12. 12.

    Akude E, Zherebitskaya E, Chowdhury SK, Smith DR, Dobrowsky RT, Fernyhough P (2011) Diminished superoxide generation is associated with respiratory chain dysfunction and changes in the mitochondrial proteome of sensory neurons from diabetic rats. Diabetes 60(1):288–297

    CAS  PubMed  Article  Google Scholar 

  13. 13.

    Hsieh YL, Lin CL, Chiang H, Fu YS, Lue JH, Hsieh ST (2012) Role of peptidergic nerve terminals in the skin: reversal of thermal sensation by calcitonin gene-related peptide in TRPV1-depleted neuropathy. PLoS One 7(11):e50805

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  14. 14.

    Koulmanda M, Qipo A, Chebrolu S, O'Neil J, Auchincloss H, Smith RN (2003) The effect of low versus high dose of streptozotocin in cynomolgus monkeys (Macaca fascilularis). Am J Transplant 3:267–272

    CAS  PubMed  Article  Google Scholar 

  15. 15.

    Leiter EH (1982) Multiple low-dose streptozotocin-induced hyperglycemia and insulitis in C57BL mice: influence of inbred background, sex, and thymus. Proc Natl Acad Sci U S A 79(2):630–634

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  16. 16.

    Rajashree R, Kholkute SD, Goudar SS (2011) Effects of duration of diabetes on behavioral and cognitive parameters in streptozotocin-induced juvenile diabetic rats. Malays J Med Sci 18(4):26–31

    PubMed  PubMed Central  Google Scholar 

  17. 17.

    Yagihashi S, Mizukami H, Sugimoto K (2011) Mechanism of diabetic neuropathy: where are we now and where to go? J Diabetes Investig 2(1):18–32

    CAS  PubMed  Article  Google Scholar 

  18. 18.

    Sveen KA, Karimé B, Jørum E, Mellgren SI, Fagerland MW, Monnier VM et al (2013) Small- and large-fiber neuropathy after 40 years of type 1 diabetes: associations with glycemic control and advanced protein glycation: the Oslo Study. Diabetes Care 36(11):3712–3717

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  19. 19.

    Themistocleous AC, Ramirez JD, Serra J, Bennett DL (2014) The clinical approach to small fibre neuropathy and painful channelopathy. Pract Neurol 14(6):368–379

    PubMed  PubMed Central  Article  Google Scholar 

  20. 20.

    Beiswenger KK, Calcutt NA, Mizisin AP (2008) Epidermal nerve fiber quantification in the assessment of diabetic neuropathy. Acta Histochem 110(5):351–362

    PubMed  PubMed Central  Article  Google Scholar 

  21. 21.

    Wolfe RR (2017) Branched-chain amino acids and muscle protein synthesis in humans:myth or reality? J Int Soc Sports Nutr 22(14):30

    Article  CAS  Google Scholar 

  22. 22.

    Barrett AM, Lucero MA, Le T, Robinson RL, Dworkin RH, Chappell AS (2007) Epidemiology, public health burden, and treatment of diabetic peripheral neuropathic pain: a review. Pain Med 2:S50–S62

    Article  Google Scholar 

  23. 23.

    Hinder LM, Vivekanandan-Giri A, McLean LL, Pennathur S, Feldman EL (2013) Decreased glycolytic and tricarboxylic acid cycle intermediates coincide with peripheral nervous system oxidative stress in a murine model of type 2 diabetes. J Endocrinol 216(1):1–11

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  24. 24.

    Ola MS, Berkich DA, Xu Y, King MT, Gardner TW, Simpson I, (2006) Analysis of glucose metabolism in diabetic rat retinas. Am J Physiol Endocrinol Metab 290(6):E1057–E1067

    CAS  PubMed  Article  Google Scholar 

  25. 25.

    Piccolo BD, Graham JL, Stanhope KL, Fiehn O, Havel PJ, Adams SH (2016) Plasma amino acid and metabolite signatures tracking diabetes progression in the UCD-T2DM rat model. Am J Physiol Endocrinol Metab 310(11):E958–E969

    PubMed  PubMed Central  Article  Google Scholar 

  26. 26.

    Tretter L, Adam-Vizi V (2000) Inhibition of Krebs cycle enzymes by hydrogen peroxide: a key role of [alpha]-ketoglutarate dehydrogenase in limiting NADH production under oxidative stress. J Neurosci 20(24):8972–8979

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  27. 27.

    Figliomeni B, Bacci B, Panozzo C, Fogarolo F, Triban C, Fiori MG (1992) Experimental diabetic neuropathy. Effect of ganglioside treatment on axonal transport of cytoskeletal proteins. Diabetes 41(7):866–871

    CAS  PubMed  Article  Google Scholar 

  28. 28.

    Hellmuth C, Kirchberg FF, Lass N, Harder U, Peissner W, Koletzko B, Reinehr T (2016) Tyrosine is associated with insulin resistance in longitudinal metabolomic profiling of obese children. J Diabetes Res 2016:2108909

    PubMed  Article  CAS  Google Scholar 

Download references


The authors thank Rose LeFaucheur for the secretarial help, Karin Meyer, Dunja Baumgartl-Ahlert and Hans-Joseph Wrede for technical assistance, and Vijayan Gangadharan (V.G) for scientific discussion. We would like to thank the Metabolomics Core Technology Platform of the Excellence Cluster CellNetworks for support with amino acid and metabolite quantification.


This work was supported by a grant from the Deutsche Forschungsgemeinschaft (DFG) in the Collaborative Research Center 1118 (SFB1118 Project B06) to N.A. and R.K. and (SFB1158, Project A03) to V.G. and R.K.

Author information




NA and RK designed the study; DRR and NA performed and analyzed the experiments; NA and RK wrote the manuscript. All authors approved the final version.

Corresponding author

Correspondence to Nitin Agarwal.

Ethics declarations

Conflict of interest

The authors declare that there is no conflict of interest.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.


• Inhibition of glycolytic enzymes is an initiation event in sciatic nerve occurring at a stage when there are no evident symptoms of DPN

• Importantly, inhibition of TCA cycle shifts glucose metabolism to alternative pathways.

• Finally, the metabolic dysfunction observed in peripheral nerve is paralleled by changes in BCAA serum levels

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Rojas, D.R., Kuner, R. & Agarwal, N. Metabolomic signature of type 1 diabetes-induced sensory loss and nerve damage in diabetic neuropathy. J Mol Med 97, 845–854 (2019).

Download citation


  • Hyperglycemia
  • DPN
  • TCA
  • Streptozotocin
  • Sensory neurons