Neuroactive Kynurenine Metabolite Alterations Unveil Novel Association to Locomotor Deterioration in Diabetic Neuropathy

Abstract

Background:

The effect of prolonged hyperglycemia on the sensory pathway of the nervous system has been the focus of numerous diabetes studies that aim at understanding the pathophysiology of the underlying inflammatory condition and neuropathy. In this study, we investigate the effects of prolonged hyperglycemia on the motoneurons of the ventral horn of the spinal cord, a lesser-studied area of the nervous system, with a focus on alterations in the Kynurenine Pathway (KP) as potential factors contributing to the induction, progression, and/or chronicity of diabetic neuropathy.

Methods:

KP metabolites were identified and assessed by immunohistochemistry in cross-sections of the lumbar spinal cord of type 2 diabetes (T2D) streptozotocin-induced (STZ) adult Sprague-Dawley rats.

Results:

Neuropathy, hyperglycemia, and gait alterations were associated to myelin loss in the spinal cord. KP metabolites were identified in glia, motoneuron, and non-motoneuron. The KP induction, as evidenced by enhanced L-kynurenine (L-KYN) fluorescence, appears to be associated with increased levels of interferon-gamma (IFN-γ) and tumor necrosis factor-alpha (TNF-α). Notable differences in fluorescence merging of L-KYN with IFN-γ and TNF-α, of Quinolinic acid (QUIN) with 3-Hydroxykynurenine (3-HK), and of QUIN with advanced glycation end products (AGEs) were observed in the T2D group, contrasting with the control (P < .05). Additionally, in ventral horn cells, AGEs emerged as an added pro-inflammatory factor.

Conclusions:

The KP is activated during diabetic neuropathy, and it displays divergent metabolic profiles in glia, motoneuron, and non-motoneuron, which differ from the controls. Their presence also evolves with time, indicating the dynamic nature of the process.

Graphical Abstract.

Introduction

About 537 million individuals worldwide suffer from diabetes, a disease defined by ongoing hyperglycemia, which has caused approximately 6.7 million fatalities in 2021, but worse, it was predicted to be 643 million and 783 million diabetics by 2030 and 2045.^1,2 Type 2 diabetes (T2D) is a multifactorial disease^3,4 characterized by chronic hyperglycemia, which leads to metabolic alterations and initiates inflammatory signaling cascades, causing complications such as diabetic neuropathy, which involves nerve fiber damage in medium and small-diameter fibers due to prolonged exposure to hyperglycemia. ⁵

Neuropathy associated with diabetes has been reported as a multifactorial process that includes gate control dysfunction in the spinal dorsal horn, dorsal root ganglion, and supraspinal descending modulation. ⁶ Diabetic neuropathy patients face motor dysfunction, with balance loss and increased fall risk, which correlate with heightened physical disability. ⁷ Different contributing factors include peripheral axon myelin loss, ⁸ proximal axon hypertrophy, ⁹ and impaired motoneuron axon transport. ¹⁰ The mechanisms associated with axonal transport deficiencies include inflammation, excitotoxicity, oxidative stress, axonal cytoskeletal protein glycosylation, synaptic vesicle reduction, mitochondrial dysfunction, and distal axon mitochondria degeneration. ¹¹ Affected motoneurons show increased expression of the receptor for advanced glycation end products (RAGE) and poly ADP-ribose polymerase, both linked to cellular stress. ¹² As in T2D, in type 1 diabetes (T1D), synaptic alterations and motoneuron microenvironment changes have also been observed. ¹³

Chronic hyperglycemia exerts damage through biochemical mechanisms such as the polyol pathway and the advanced glycation end products (AGEs).^5,7 The kynurenine metabolic pathway (KP) could play a deleterious part.^14-26

A diabetogenic role of Xanthurenic acid (XA) has been reported since 1975,^14,15 indicating that XA and Kynurenic acid (KYNA) can inhibit pro-insulin synthesis in rat pancreatic islets. ¹⁶ Elevated L-kynurenine (L-KYN), 3-Hydroxykynurenine (3-HK), and Anthranilic acid (ANA) levels were found in diabetic patients with cataracts,^17,27,28 additionally, a positive correlation between the L-KYN/Tryptophan (Trp) ratio and pain intensity in T1D patients ¹⁸ ; thus, KP metabolites have been associated to diabetes development and even as potential biomarkers.^19,20 It remains unclear whether the different alterations play beneficial or deleterious roles. ²¹

L-KYN, the key metabolite of KP derived from Trp catabolism, undergoes conversion by indoleamine 2,3-dioxygenase (IDO), which is induced by cytokines such as IFN-gamma, IL-1β, TNF-alpha, growth factors, infectious agents, and stress.^22-25,29 L-KYN can further be transformed into either KYNA, 3-HK, or ANA. The latter 2 give rise enzymatically to Picolinic acid (PA) and nonenzymatically to Quinolinic acid (QUIN). ²⁵ These compounds are subsequently metabolized into nicotinamide adenine dinucleotide (NAD⁺), serving as the final product of the KP pathway. The neuroprotective or neurotoxic effects of KP metabolites have been a focal point of research, considering their potential role in various diseases.^26,30-32

The roles involving KP neuroactive metabolites have been elucidated in various pro-inflammatory models. KYNA exerts an antagonist to the α7-nicotinic receptor ³³ and inhibits the opening of N-type Ca⁺² channels. ³⁴ QUIN acts as a weak agonist of channels such as NMDA and AMPA. ³⁵ QUIN’s role in excitotoxicity mechanisms emerges due to its ability to stimulate glutamate release from synaptosomes and to inhibit glutamate uptake by astrocytes. ³⁶ All these reports support the relevant physiological and pathological roles played by KP in Diabetes, however, the information on their effect on the spinal cord cells as a main sensory-motor processing center is limited.

This study aims to examine whether alterations in KP metabolites, that is, L-KYN, KYNA, 3-HK, and QUIN, are associated with motor function changes in a type 2 diabetes experimental model during the onset, progression, and/or chronicity of diabetic neuropathy.

Materials and Methods

Bioethics

All the procedures for animal management in this work were designed and put into practice, adhering to the Manual of Production and Ethical Use of Laboratory Animals, ³⁷ The Organic Law for the Protection of Wildlife in Captivity, ³⁸ and the institutional bioethics committee (Approval CDBBUC 012-2018/BASÑJKG).

STZ Induction of T2D

Male Sprague-Dawley rats, initially weighing 300 to 350 g, were maintained in a controlled environment of 27°C ± 1°C, with 60% relative humidity and a 12:12 light-dark cycle. Water containing 40 g.L⁻¹ sucrose was available ad libitum during the experimental time. A commercial diet containing 21% of crude protein, 9% saturated fat, 5% crude fiber, 8% ash, 12% moisture, 1.4% calcium, 1% phosphorus and enriched with pork saturated fats was provided for 5 weeks to 20 rats, which were then separated randomly into 10 Control and 10 T2D rat groups. From week 6 onwards, a normal diet and sucrose-supplemented water were supplied. T2D was induced in the group at the beginning of the 5th week using 35 mg.kg⁻¹ of streptozotocin i.p. (STZ, SIGMA™) in 0.9% NaCl.^39,40 The control animals received only 1 mL i.p. of the 0.9% NaCl.

The follow-up of both animal groups was made weekly by capillary blood glucose evaluation from week 4th until the end of the protocol (week 14th), a blood drop sample from the tail in glucose-reacting strips (FreeStyle™) based on the Glucose Dehydrogenase-NAD⁺ enzyme reaction ⁴¹ measured with an automatic digital strip glucometer (FreeStyle™).

Gait Pattern Protocol

The weekly gait pattern of each animal was recorded by the open field test using a 55 × 55 × 25 cm cage with rice husk floor and a 12.2 Mpx CCD digital camera (Fuji Film^®), which captured zenithal sequences at 15-second intervals for 6 minutes. Sequences were semi-automatically analyzed using Tracker^® v4.9 to evaluate the total distance traveled by each animal ⁴² obtaining the X-Y coordinates of the trajectories and the traveled distance.

Lumbar Spinal Cord (LSC) and Sciatic Nerve (SN) Extractions

Rats in the 12th and 14th weeks after STZ or vehicle administration were anesthetized for lumbar laminectomy and LSC extraction, using sodium thiopental (Pentothal^® sodium, Lab. Abbott) at 60 mg.kg⁻¹ i.p. under stable heart, respiratory, and body temperature conditions. ⁴¹ Laminectomy was performed as described elsewhere, ⁴³ the LSC ⁴⁴ and left sciatic nerves from the same animals were extracted and preserved in 4% paraformaldehyde for processing within the week.

Sciatic Nerve Myelin Assessment

Paraffin-included main trunk of the sciatic nerves was transversally cut into 5 µm sections. Luxol Fast Blue (LFB) staining protocol^45,46 was used for myelin staining. A Nikon Eclipse E600 microscope with a CCD Spot Flex camera was used for image acquisition from selected areas of identical sizes at magnifications 10× and 20×. ImageJ software (NIH, USA) and arbitrary spectral densitometry were used for myelin LFB affinity quantification.

Spinal Cord KP Immunohistochemistry

Transverse sections (5 µm) of LSC were analyzed for KP metabolite identification. The primary antibodies were L-KYN Abcam (ab119836, Rabbit), KYNA Abcam (ab37105, Rabbit), 3-HK Abcam (ab111013, Rabbit), and QUIN Abcam (ab85518, Mouse) for the KP and other analytes of interest such as IFNγ and AGES, TNF-α Abcam (ab6671, Rabbit) and AGES Abcam (ab23722, Rabbit) were used. The secondary antibodies used were anti-mouse conjugated with Rhodamine (Rd) and Anti-rabbit conjugated with Fluorescein Isothiocyanate (FITC) with emission wavelengths of 625 and 525 nm, respectively (NIH, USA).

Variations in the fluorescence intensity of the KP metabolites were recorded at weeks 12th (5 rats per group) and 14th (5 rats per group) by cell type (glia, motoneuron, and non-motoneuron). Three consecutive sections of the LSC were used to evaluate a pair of different analytes per slide, being the first section of the slide the negative control. Regions of interest (ROIs) were selected from each imaging, that is, the periphery of the nucleus, the periphery of the glial cell membrane, motor neurons, non-motor neurons, and finally the background.^47,48 Twenty-five ROIs of glia, 25 ROIs of motor neurons, 25 ROIs of non-motor neurons, and 10 ROIs of background were measured. The medians shown correspond, for example, to the evaluation of the L-kyn-kinase pair: 5 × 9 × 30 × 25 = 33,750 measurements for the diabetic group, and so on for each metabolite, thus, a total of 2500 images were processed from the 12th and 14th evaluation weeks for control and diabetes groups. Images were acquired using specific wavelengths for DAPI, Rd, and FITC. Background, neuronal shape, size, and fluorescence intensity were taken into account to differentiate among glia, motoneurons, and non-motoneurons, including nucleus and cell membrane peripheries measures in glial cells.^46,49,50

Antibody validation and negative controls: Autofluorescence quenching: 0.5% NH₃ was added to a Coplin jar containing 50 mL of ethanol, and the slides were incubated for 1 hour at room temperature. Antigen retrieval: The slides were placed in a Coplin jar containing 500 mL of HIER reagent (66 mM sodium citrate and HCl at pH = 6) and incubated for 10 minutes at boiling point. Membrane permeabilization: 0.1% Triton X100 in PBS was applied for 10 minutes at room temperature. Blocking: the blocking buffer solution was prepared with 3% bovine serum albumin (BSA) in PBS, plus 0.1% Triton X100 in PBS, and incubated at room temperature for 1 hour. To avoid false-positive results, the incubation with antibodies was then performed, except for the negative control preparation, which did not receive the primary antibody.

Statistical Analysis

Data from the open field evaluations, blood glucose measurements, and LFB staining were normalized. The results were compared using non-parametric, Mann-Whitney (bivariate analysis), and the Kruskal-Wallis (multivariate analysis) tests. The variance of the traveled distances recorded by the open field test was estimated with the coefficient of variation (CV), and between groups compared by the Fligner-Killeen test. Normalized fluorescence quantification results for KP metabolites, TNF-α, IFN-γ, and AGES were compared across cell types and weeks for both T2D and control animals. In addition, the multivariate analysis of variance (MANOVA) and Spearman’s correlation test were conducted. The analyses were performed in the R environment, employing the Nonpartest function, ⁵¹ and the cor function, ⁵² respectively. Potential variable associations were tested by principal component analysis (PCA) and cluster analysis. The Factoshiny, Factoextra, and FactoMineR packages from R environment version 4.0.2 ⁵³ were applied.

Results

Hyperglycemia

The T2D group shows higher glycemia values than the control group (Figure 1). The mean values were 5.63 ± 0.69 mmol.L⁻¹ and 7.24 ± 0.67 mmol.L⁻¹, respectively, that is, mean 15.2 % higher in the T2D group over the controls, with maximum individual values of 14.15 mmol.L⁻¹ for T2D and 6.94 mmol.L⁻¹ for the control group (Figure 1). The differences in blood glucose concentrations were persistently statistically significant from week 6th (Mann-Whitney; P < .05, P < .01, P < .001). Hyperglycemia of the T2D group was 38% [95% CI: 12.2%-73.8%] higher than the control group when comparing the area under the curve of both groups (Figure 1).

Figure 1.

Capillary glycemia of the T2D (black line) and Control (gray line) groups from week 4 to 14. Each point is the average per week, the whiskers correspond to the standard deviation, n = 10 for each group and *P < .05, **P < .01 ***P < .001.

Open Field Trajectories

The T2D group showed a lower traveled distance than the control group from the 3rd week after STZ treatment (Mann-Whitney test; P < .05, most weeks with P < .01; Figures 2 and 3). The area under the curve from the traveled distance versus time (weeks) shows that the T2D group traveled 21.3% less distance than the control group from the 2nd week after the SZT administration. Gait pattern was characterized by long traveled distances in the Control group and by short traveled distances and higher clustering standpoints in the T2D group (Figure 3). The T2D group also showed a progressive decrease in gait distance over time, as shown by the negative correlation between both variables (r = −.74; P < .05) with a coefficient of determination (r²) of 55% [95% CI: 26.2%-87.8%] (P < .05). The coefficient of variation of traveled distance from the T2D group (85.4% [95% CI: 75.8%-94.7%]) was statistically higher (z = 3.75; P < .0001) than that from the Control group (63.1% [95% CI: 56.6%-69.4%]). Figure 3 (upper and mid panels) shows individual and collective samples of open field X-Y trajectory differences between animals from each group, with simpler, low complexity, and shorter trajectories in the T2D rats than the controls. Additionally, the T2D group shows more standpoints than the Controls (Figure 3 Bottom).

Figure 2.

Total gait distance per week in the open field test. T2D (black line) and Control (gray line) and groups during the development of diabetic neuropathy. Each point represents the median per week from 0 to 14, and the error bars are the interquartile range (p25-p75), n = 10 in each group and (*P < .05, **P < .01).

Figure 3.

Gait trajectories (upper and mid panels) of rats from Control or T2D groups. Typical individual Control (upper left) or T2D (upper right) rat trajectories during 6-minute tests at 14th weeks. The overplot of all trajectories from all recording weeks are shown for the Control (n = 10; mid left) and T2D group (n = 10; mid right). All traveled distances are presented on a log scale at the bottom of the graph for the Control (n = 10; blue) and T2D Group (n = 10; red) to detect the more frequent standpoints for the T2D group.

Sciatic Nerve Myelin

The T2D group shows a loss LFB myelin affinity compared to the control (Mann-Whitney test; P < .001) at tested weeks 12th and 14th (Figure 4A and B) which represents 55.6% [95% CI: 14.7%-94.7%] and 42.4% [95% CI: 5.27%-85.3%] less affinity respectively when comparing the area under the curves (Figure 4C).

Figure 4.

Sciatic nerve cross-sections affinity to LFB staining. The groups: (A) T2D, (B) Control, and (C) comparison of the affinity for LFB staining of both groups at week 12 and 14 (arbitrary densitometry units in blue spectrum). The yellow and red arrows in A and B indicate the different affinities of the tissue for the stain. A and B magnification is 40×. C presents the staining average of 500 measurements per image (n = 2) of each group per week, the scatter bars correspond to the standard deviation, and ***P < .001.

Spinal Cord KP Metabolites

The highest fluorescence intensity was observed at the intracellular level for all cell types (P < .05). The evaluation of the metabolites was carried out in pairs in adjacent cross-sections of LSC at weeks 12th and 14th. The evaluated pairs were L-KYN/QUIN, 3-HK/QUIN, KYNA/QUIN, L-KYN-/IFN-γ, L-KYN/TNF-α, and QUIN/AGES.^8-13 Cytoplasmatic fluorescence confluence was more evident in all cell types of the T2D group in contrast to the control, as shown in Figure 5.

Figure 5.

Anatomical distribution of the KP metabolites, AGES, IFN-γ, and TNF-α in LSC in T2D and Control. A: negative control-NC focusing only gray matter, B: the evaluated pair of the T2D and control groups is identified from left to right (L-KYN/QUIN, 3-HK/QUIN, KYNA/QUIN, L-KYN-/IFN-γ, L-KYN/TNF-α, and AGES/QUIN). Rd secondary antibodies (red), for QUIN and IFN-γ for L-KYN, 3-HK, KYNA, AGES, and TNF-α FITC (green). Merge images show glia with red arrows, motoneurons with green arrows, non-motoneuron with white arrows, and confluence fluorescence of Rd, FITC, and DAPI with blue arrows. 60× magnification.

KP Metabolites Fluorescence in Weeks 12th and 14th

At week 12th, the motoneuron and non-motoneuron cells show a significantly higher fluorescence for AGES, L-KYN, KYNA, QUIN, IFN-γ, and TNF-α (P < .05 or better) than the Control. It is noteworthy that fluorescence was significantly higher in motoneurons than in non-motoneuron (P < .05; Figure 6; Table 1), while glia of the T2D group show the highest intracellular fluorescence of AGES, QUIN, IFN-γ, and TNF-α (P < .01 or better) in contrast to L-KYN of the control group that shows higher fluorescence than that observed in the T2D group (P < .01).

Figure 6.

L-KYN florescence intensity time course from week 12 to 14. Each point represents the median intracellular fluorescence and the corresponding interquartile range in glia, motoneuron and non-motoneuron. Groups: T2D (n = 5) (black) and Control (n = 4) (gray) and *P < .05.

Table 1.

Intracellular Fluorescence Intensity by Metabolite in Glia, Motoneuron, and Non-Motoneuron of the T2D and Control Groups at Week 12.

	Glia				Motoneuron			Non motoneurons
Metabolite	T2D	P inter-group	Control	P inter-group	T2D	P inter-group	Control	T2D	P inter-group	Control
AGES	6.373 ± 2.196	***	1.282 ± 0.781		10.348 ± 6.879	*	2.801 ± 6.981	9.893 ± 4.729	***	5.945 ± 4.655
L-KYN	10.237 ± 6.311		19.224 ± 11.789	**	26.453 ± 16.149	**	10.227 ± 15.649	15.654 ± 10.461	***	10.675 ± 9.052
KYNA	24.766 ± 0.001		20.791 ± 10.964		39.055 ± 0.010	*	17.839 ± 23.748	39.055 ± 0.001	*	2.742 ± 2.673
3-HK	10.908 ± 1.888		8.558 ± 7.668		12.541 ± 5.159		12.694 ± 8.638	13.850 ± 6.054		3.539 ± 5.043
QUIN	33.399 ± 17.387	***	13.436 ± 16.690		76.389 ± 30.596	***	17.053 ± 19.696	51.760 ± 27.300	***	13.562 ± 19.003
IFN-γ	36.464 ± 23.243	**	23.061 ± 22.897		86.504 ± 76.978	***	42.518 ± 34.489	52.381 ± 46.443	***	32.808 ± 27.723
TNF-α	26.762 ± 11.064	***	17.066 ± 12.324		38.021 ± 18.604	***	27.521 ± 29.466	43.015 ± 38.647	***	26.806 ± 14.741
	λ = 625 nm		λ = 525 nm
CN	2.996 ± 2.631		3.061 ± 3.466

The data corresponds to the median ± interquartile range; *P < .05; **P < .01; ***P < .001. The emission wavelengths Rd: 625 nm and FITC 525 nm.

At week 14th, the glial cells, motoneurons, and non-motoneurons of the T2D group show higher intracellular fluorescence for L-KYN, KYNA, and 3-HK (P < .05 or better; Figure 6; Table 2) than those recorded from the Control. T2D motoneurons showed higher AGES fluorescence (P < .05), while TNF-α and IFN-γ presented significant differences in motoneurons and non-motoneurons of T2D, compared to the control (P < .05; Figure 6; Table 2).

Table 2.

Intracellular Fluorescence Intensity by Metabolite in Glia, Motoneuron, and Non-Motoneuron of the T2D and Control Groups at Week 14.

	Glia				Motoneuron			Non motoneurons
Metabolite	T2D	P inter-group	Control	P inter-group	T2D	P inter-group	Control	T2D	P inter-group	Control
AGES	9.799 ± 3.143		18.538 ± 5.885	**	22.407 ± 9.169	*	17.435 ± 10.430	15.146 ± 5.707		19.205 ± 10.430
L-KYN	66.017 ± 0.010	**	13.359 ± 10.220		76.281 ± 34.025	***	10.997 ± 14.922	76.281 ± 34.026	***	14.364 ± 16.412
KYNA	49.531 ± 1.650	*	3.898 ± 3.978		79.827 ± 16.786	***	5.009 ± 3.818	76.958 ± 23.610	***	4.229 ± 6.560
3-HK	44.232 ± 6.437	***	2.400 ± 6.316		76.570 ± 29.479	***	6.638 ± 5.999	76.570 ± 29.479	***	6.198 ± 7.299
QUIN	23.939 ± 6.435	***	10.075 ± 11.679		16.946 ± 6.997		15.935 ± 20.441	38.921 ± 14.027	***	19.085 ± 23.743
IFN-γ	60.931 ± 16.561	***	18.172 ± 10.435		62.680 ± 0.010	*	20.040 ± 9.909	62.680 ± 0.010	*	30.718 ± 13.344
TNF-α	28.118 ± 12.670		21.913 ± 16.690		59.456 ± 53.293	*	34.477 ± 28.641	32.967 ± 18.277	*	23.661 ± 25.835
	λ = 625 nm		λ = 525 nm
CN	1.254 ± 0.747		0.414 ± 1.612

The data corresponds to the median ± interquartile range; *P < .05; **P < .01; ***P < .001. The emission wavelengths Rd: 625 nm and FITC 525 nm.

Differences in KP fluorescence patterns between weeks 12th and 14th show that the highest intracellular fluorescence at week 12th was recorded for IFN-γ in glia, motoneuron, and non-motoneuron in both of T2D and Control groups (P < .05), whereas at the 14th week was for L-KYN (P < .05) in the glia of the T2D group than that recorded in the Control group. The L-KYN fluorescence in the T2D group increases with the passing of the weeks (r = .7; P < .05). At week 14th, the fluorescence of L-KYN, 3-HK, and KYNA was significantly higher in the T2D group than in the Control group (Figure 6; Table 2).

Association Analysis Among KP Metabolites

Principal component analysis of glial KP metabolites fluorescence from T2D and Control groups at week 12th show 2 main components, a right, horizontally elongated pattern (light blue; Figure 7A; 12th) for the T2D group and a left predominant, highly dispersed almost equidistant to the centroid pattern from the Control group (light red; Figure 7A; week 12th) with an overlapping of the 2 components at the origin of the cartesian graph, more for values from the component 1. Fluorescence values for KYNA, AGES, QUIN, 3-HK, IFN-γ, and TNF-α of glial cells were associated with component 1, whereas L-KYN was associated with component 2 (Figure 7A). The two-component pattern persists at the 14th week, yet without overlapping. The pro-inflammatory metabolites AGES, QUIN, and TNF-α were related to each other and to component 2. The metabolites KYNA, 3-HK, IFN-γ, and L-KYN were grouped and related to component 1 (Figure 7A; week 14th).

Figure 7.

Principal component analysis of T2D and Control at 12 and 14 weeks. A: glia; B: motoneuron and, C: non-motoneuron. Component 1 and component 2 are represented in the diagrams. Each vector is a parameter of KP metabolites, AGES, IFN-γ, and TNF-α. CONTROL (pink area) and T2D (blue area).

Motoneurons at week 12th showed the two-component pattern with strong overlapping at the origin of the Cartesian graph (Figure 7B; week 12th). The L-KYN associated with KYNA, AGES with QUIN and IFN-γ, and, together with 3-HK, were related to component 1, whereas the TNF-α was associated with component 2.

There is a strong overlap of both components, more for values from component 1, at the origin of the Cartesian graph. In the 14th week, the 2 components show no overlapping, are separated and distant from the origin of the Cartesian axis. The pro-inflammatory metabolites AGES, QUIN, and TNF-α were related to each other and to component 2, while KYNA, 3-HK, IFN-γ, and L-KYN were grouped and related to component 1 (Figure 7B; week 14th).

Non-motoneurons at week 12th show a clustered low low-dispersed Control component at the left quadrants and a wide dispersed left T2D component (Figure 7C; week 12th) associated with all recorded metabolites. These 2 components, at the 14th week, were separated from the origin of the cartesian axis. QUIN, AGES, and TNF-α were associated with component 2, while KYNA, 3-HK, IFN-γ, and L-KYN, associated with component 1 (Figure 7C; week 14th).

Discussion

In the present study, the hyperglycemia induced by low dose of STZ, a saturated fat-rich diet, and the sucrose-enriched hydration in rats successfully mimicked the clinical progression of T2D as seen in humans. This approach supports the notion of the critical role of diet and environmental factors in the development of diabetes-related complications, aligning with recent research that emphasizes lifestyle influences in T2D progression.^39,40

Open field test results suggest a progressive decrease in the distance covered and the increased detentions points by the T2D animals mimic the clinical progression of the human diabetic neuropathy but enhancing the motor dysfunctions ⁵⁷ caused by the synergic and deleterious action of the diabetes at CNS but also at the legs muscles. ⁵⁷ MRI analysis in T2D patients showed structural alterations in the motor cortex but also the spinal cord.^58,59 α-Motoneurons originating sciatic branches have been reported to be reduced in STZ-induced diabetes models. ⁶⁰ The T2D motor impairment also includes the neuromuscular junction and the number of motor units in human biopsies from the foot muscle showing neurogenic degeneration. ⁵⁷

In experimental STZ-induced diabetes, demyelination in the motor axon and mitochondrial disruptions at its terminal have also been reported. ⁶¹

The observed gait changes in T2D rats, characterized by reduced movement over time, agree with findings in T1D models. These similarities between T1D and T2D in motor function deterioration highlight a possible common pathway in diabetes-induced neuropathy, despite the differences in their etiology, thus suggesting a broader, systemic effect of diabetes on motor functions.^13,57,62,63

Myelin loss, intra-myelin edema, and a decrease of small, myelinated fibers in cross sections of the sciatic nerve confirmed one of the symptoms well described in diabetic neuropathies, making these results comparable with those already reported.^64,65 Experimental and clinical diabetic neuropathy are associated with impaired nerve conduction and structural nerve changes, adding to the growing evidence that diabetes has a direct and detrimental effect on neural integrity.^56,66

The temporal course of fluorescence changes of KP metabolites like L-KYN, KYNA, 3-HK, and QUIN along the 14 weeks of STZ-induced diabetes in this work strongly suggest that these compounds are involved in the progression of diabetic neuropathy. Their potential roles in allodynia, hyperalgesia, and motor dysfunction in T2D give additional insight into the pathophysiological understanding of diabetic neuropathy. KP metabolites have been pinpointed in controlled T2D patients with subclinical CNS dysfunction, ⁶⁷ highlighting the need for a more nuanced exploration of their functions. KP metabolites are also linked to neurodegenerative diseases such as Alzheimer’s, neurological disorders such as depression, metabolic disorders like atherosclerosis, and autoimmune disorders such as multiple sclerosis, among other pathological conditions.^{30 ,31,68-71}

The proinflammatory cytokines IFN-γ and TNF-α in glia, motoneuron, and non-motoneuron, and their possible role to induce KP metabolites in diabetic neuropathy as reported here, contribute to the ‘inflammatory soup’ hypothesis of chronic pain,^72-74 which opens novel anti-inflammatory strategies in managing diabetic neuropathy and underscores the intricate interplay between metabolic and inflammatory pathways in diabetes.

The KP metabolites could be associated with a neuroprotective and/or proinflammatory role, depending on the specific metabolite and the cellular context. According to our results, the proinflammatory environment was associated with the synthesis of neurotoxic metabolites such as QUIN and 3-HK. Different KP metabolites have been reported in CNS cells, depending on physiological or pro-inflammatory conditions.^54,55,75,76 Future research should aim at dissecting these interactions further and exploring therapeutic interventions targeting these pathways.

We identified advanced glycation end products (AGES) in glial cells, motoneurons, and non-motoneurons of LSC cross sections, with notably higher fluorescence in the diabetic neuropathy group. This observation aligns with the understanding that diabetic neuropathy involves an activation of the AGES formation pathway. Furthermore, we noted a correlation between AGES, proinflammatory cytokines, and neurotoxic KP metabolites, as indicated by the principal component analysis (PCA). This is consistent with previous reports of AGES accumulation in advanced diabetes stages and their identification in the endoneurium, perineurium, and microvesicles of peripheral nerves in T2D patients.^5,77 The increased expression of AGES receptors in spinal cord motor cells ¹² and their co-presence with TNF-α in retinal glia ⁷⁸ further substantiates their role in diabetic neuropathy.

The distinct metabolic profiles of the Control and T2D groups, revealed through PCA across weeks 12 and 14, underscore the unique metabolic alterations associated with diabetic neuropathy. These findings demonstrate that the neuropathic process is dynamic, with changes in metabolite associations depending on the cell type and the progression stage. This highlights the evolving nature of diabetic neuropathy and its complex metabolic interplay.

Based on the results of the present report, we propose a model where hyperglycemia in diabetic neuropathy leads to increased blood-brain barrier (BBB) permeability (Figure 8). This alteration potentially allows a higher influx of metabolites like Trp, L-KYN, and 3-HK into the CNS. In a pro-inflammatory environment, these metabolites are processed by neurons and microglia to promote QUIN synthesis through KP. Conversely, astrocytes, typically synthesizing KYNA under normal conditions, may shift toward QUIN production from peripheral 3-HK in inflammation. The rise in IFN-γ and TNF-α further stimulates KP through IDO activation in the CNS. Such stimulation leads to increased glutamate release and calcium influx in dorsal and ventral horn neurons.

Figure 8.

Proposal about KP alterations in ventral horn cells in type 2 diabetic neuropathy.

The roles of KP metabolites such as L-KYN, 3-HK, QUIN, and KYNA in inflammation vary depending on the stage of T2D and the development of diabetic neuropathy.

Furthermore, our study suggests that inflammation in ventral horn neurons is mediated by AGES, TNF-α, and QUIN. The interaction between glial cells and neurons appears to foster KP activation and inflammation, potentially impacting motor function and exacerbating motor dysfunction. Our results in the diabetic neuropathy model indicate that neurotoxic metabolites like 3-HK and QUIN may trigger a signaling cascade that promotes inflammation and neurodegeneration, as illustrated in Figure 8.

In this study, we identified a connection between the activation of the kynurenine pathway in the glia, motoneuron, and non-motoneuron of the lumbar spinal cord (LSC) and the development of diabetic neuropathy. This finding broadens the array of factors implicated in diabetic neuropathy, opening avenues for future research. Such research endeavors could include investigating biomarkers, identifying therapeutic targets, developing experimental methods, and designing novel therapies focused on regulating the induction of the kynurenine pathway.

Limitations of the Study

The STZ+HFD model is widely employed to replicate key features of human type 2 diabetes (T2D), including insulin resistance, β-cell dysfunction, and metabolic alterations. ⁷⁹ However, several limitations must be considered for the present report. STZ induces rapid pancreatic β-cell destruction, with an abrupt decline in insulin production, which contrasts with the progressive β-cell dysfunction observed over years in human T2D.^79-81 In this model, β-cell death occurs primarily via DNA alkylation induced by STZ, diverging from the glucolipotoxicity and chronic inflammatory mechanisms implicated in human T2D pathophysiology, thus lacking the pre-diabetic phase sequence of the human T2D that is, obesity – insulin resistance – abnormal glucose tolerance.^81,82 Rodent high-fat diets exceeded typical human lipid intake that could promote metabolic and neuroinflammatory responses with potential confounding behavioral and cognitive assessments induced by STZ toxicity ⁸³ or acute hyperglycemia. ⁸⁴ Although the sample size per group was relatively small, it was consistently maintained across all experimental groups and throughout the 12-week study period, ensuring internal comparability. While Western blot and pre-adsorption controls were not included, which limits the confirmation of findings at the protein level, the data nonetheless provide valuable insights at the immunohistochemical level.