2026 Volume 51 Issue 3 Pages 183-199
Neurotoxicity of acrylamide has been demonstrated both in humans and animals, while the mechanisms remain largely unknown. We recently reported TNF-α deletion suppressed acrylamide-induced neurotoxicity in mice at low dose. Here we further investigated expression of antioxidant and proinflammatory cytokines to explore roles of TNF-α. Wild type and TNF-α KO mice were exposed to acrylamide at 0/12.5/25 mg/kg bw for 28 days. The results showed that acrylamide significantly decreased body weight at 12.5 and 25 mg/kg bw, but decreased brain weight only at 25 mg/kg bw. TNF-α deletion didn’t alleviate the above effects. Also, TNF-α deletion didn’t alleviate decrease of grip strength at 25 mg/kg bw. Immunohistochemical results showed that TNF-α deletion alleviated noradrenergic axon degeneration in cortex S1FL and S1HL regions at 12.5 mg/kg bw, but not 25 mg/kg bw. Moreover, TNF-α deletion suppressed acrylamide-induced upregulation of TGF-β and NF-κB, but didn’t suppress upregulation of IL-6, suggesting possible roles of IL-6 in acrylamide-induced neurotoxicity, particularly at high concentration. Moreover, this study showed a dipolar effect of TNF-α deletion on oxidative stress pathways, i.e., at 25 mg/kg bw, TNF-α deletion suppressed upregulation of oxidative stress (Keap1/HO-1/Gclc/Gclm/Sod/Cat/Gstm/MT-1); however, at 12.5 mg/kg bw, TNF-α deletion accelerated upregulation of Nqo1/Gclm/Sod1/Cat/Gsr. Taken together, genetic TNF-α ablation, at least partially, alleviated acrylamide neurotoxicity at low concentration. Limited alleviation effects at high concentration generated a hypothesis that this may be due to IL-6 signaling and dipolar regulating effects of TNF-α deletion on oxidative stress pathway. This study provided new insights into acrylamide neurotoxicity and TNF-α-targeting strategy.
Acrylamide (CH2=CHCONH2; CAS Number 79-06-1), a low-molecular-weight organic compound, is broadly used in industry as a monomer for producing polyacrylamide, which is widely applied in water and wastewater treatment, enhanced oil recovery, papermaking, textile processing, and the manufacture of dyes and cosmetics (Izumi et al., 2022). In addition to its industrial uses, acrylamide is a process contaminant in foods, first reported in 2002 across a range of heat-processed products (Tareke et al., 2002). It forms during high-temperature cooking of carbohydrate-rich foods via the Maillard reaction between reducing sugars and amino acids, most notably asparagine (Semla et al., 2017), and therefore fried and baked starchy foods such as French fries, potato chips, bread, and coffee often contain measurable levels of acrylamide (Friedman, 2003). Human exposure to acrylamide thus occurs via both occupational and dietary routes, raising concern that acrylamide may act as an environmental toxicant affecting diverse populations.
Robust evidence from both human cases and experimental animals indicates that acrylamide is neurotoxic and can induce neurodegeneration. Recent epidemiological studies reported that dietary acrylamide was associated with mild cognitive decline in non-smoking elderly men population in a prospective cohort study (Liu et al., 2017). Occupational exposure case series reported gait ataxia, tremor, and cerebellar dysfunction among chronically acrylamide-exposed workers in construction, mining, flocculant manufacturing, and tunneling (Pennisi et al., 2013). Human cases involving intoxication of acrylamide-contaminated drinking water have additionally reported severe truncal ataxia, disorientation, memory impairment, and confusion (Igisu et al., 1975). Pathological studies with animal models demonstrated that acrylamide induces multifocal swellings and degeneration of long myelinated axons in both central and peripheral nervous systems (Lehning et al., 2002; Tan et al., 2019). Previous studies of our group also reported that exposure to acrylamide induced noradrenergic axon degeneration in various areas of the murine brain (Davuljigari et al., 2021; Ekuban et al., 2021; Fergany et al., 2023). Nevertheless, the molecular mechanisms underlying acrylamide-induced neurotoxicity remain largely unknown.
Neuroinflammation, defined as activation of microglia and astrocytes with subsequent cytokine release within the central nervous system (CNS), is increasingly recognized as a driver of neurodegeneration recently (Adamu et al., 2024; Cunningham, 2013). Among them, tumor necrosis factor-alpha (TNF-α), is a central cytokine in regulating the proinflammatory responses and has been reported to play important roles in neurodegenerative disorders (Cacquevel et al., 2004). In Alzheimer’s disease and related dementias, elevated TNF-α and glial activation are consistently observed, and epidemiologic data link chronic TNF-α blockade with reduced dementia risk (Plantone et al., 2023). In Parkinson's Disease (PD) models, genetic deletion of TNF-α receptors in mice markedly attenuates dopaminergic neuron loss, underscoring TNF-α’s contribution to neuronal dysfunction (Muzio et al., 2021). Previous studies of our and other groups showed upregulated TNF-α after exposure of mice and in vitro models to acrylamide (Davuljigari et al., 2021; Santhanasabapathy et al., 2015; Zhao et al., 2017; Zong et al., 2025). In wild type C57BL/6 mice, acrylamide exposure at 300 ppm by drinking water induced neurotoxicity accompanied by upregulation of TNF-α, but activation of Nrf2 by sulforaphane attenuated acrylamide-induced neuropathy (Davuljigari et al., 2021). In Swiss albino mice acrylamide exposure for 4 weeks at 20 mg/kg bw induced reactive gliosis and expression of inflammatory cytokines TNF-α, interleukin-1 beta (IL-1β), and inducible nitric oxide synthase (iNOS) in the cortex, hippocampus and striatum, and farnesol isolated from the essential oils of ambrette seeds and citronella ameliorated acrylamide toxicity (Santhanasabapathy et al., 2015). In primary astrocytes and microglia, after exposure to acrylamide, nuclear factor kappa B subunit 1 (NF-κB) pathways were activated followed by upregulation of related cytokines, including interleukin 6 (IL-6), TNF-α, granulocyte colony stimulating factor (G-CSF), and IL-1β (Zhao et al., 2017). Recently our previous study with in vitro and in vivo models showed acrylamide-induced noradrenergic axon degeneration is promoted via a non-cell autonomous mechanism, involving microglial TNF-α induced protein 2 (Tnfaip2)/TNF-α and oxidative stress pathways, and targeting TNF-α pathway and antioxidant treatments showed potential alleviation effects (Zong et al., 2025). However, we also observed an unexpected result of limited effects of targeting TNF-α on acrylamide toxicity, especially at high exposure concentrations (Zong et al., 2025).
Therefore, in this study, using wild-type and TNF-α knockout (TNF-α KO) mice, we carried out further investigation to decipher the accurate roles of TNF-α in neurotoxicity of acrylamide. Following acrylamide exposure, we carried out series of comprehensive analyses, including selection of appropriate behavioral observation methodology, histopathological quantification of biomarkers for axonal degeneration, and molecular level investigation of various genes of related pathways (Fig. 1). Here, we show the unexpected results of comparing the neurotoxicity of acrylamide in wild-type and TNF-α depletion mice, and highlight the possible alternative role of IL-6 and dipolar effects of TNF-α deletion on oxidative stress pathway.
Schematic illustration of the study design. Wild type (WT) and TNF-α knockout (KO) mice (10-week-old, male, n=10) were acclimatized for 1 week and then grouped. The mice were exposed to acrylamide at 0, 12.5 or 25 mg/kg body weight (bw) by oral gavage, 7 days/week, for 4 weeks. After exposure, behavioral tests including open-field test and Y-maze test were performed. For pathological analysis, mice were perfused with 4% PFA and brains were dissected for cryosection. Immunohistochemical (IHC) staining for noradrenergic axon was performed by anti-noradrenaline transporter (NAT) antibody. After staining, axon length was calculated with Image J. For biochemical analysis, mice were sacrificed by decapitation. Brains were quickly dissected on ice and snap frozen at -80 degree deep-freezer. Gene expression was measured by real-time qPCR.
All animal experiment plans were approved by the Animal Experimentation Committee of Tokyo University of Science and followed the guidelines of Tokyo University of Science on animal experiments, in accordance with the Japanese act on welfare and management of animals. Homozygous TNF-α KO (-/-) mice with C57BL/6msSlc background were from Institute of Medical Science, the University of Tokyo (Taniguchi et al., 1997). Genotype was confirmed by polymerase chain reaction (PCR) using primers (sense: AGATGGAGAAGGGCAGTTAG, anti-sense1: ATACCAGGGTTTGAGCTCAG, anti-sense2: TACTTTGTTAAGAAGGGTGAGA). The PCR was conducted by a three-step cycle under conditions of 96°C for 2 min followed by 35 cycles of 96°C for 20 sec, 59°C for 30 sec and 72°C for 45 sec. The amplified DNA samples were then run on 2% agarose gel electrophoresis and visualized with a CCD camera (Fusion Solo S; Vilber Lourmat, Collegien, France). For wild type control, specific-pathogen-free (SPF) grade C57BL/6msSlc mice were purchased from SLC Japan (Tokyo) and allowed to acclimatize for one week before the start of the study. All mice were housed in a controlled environment of temperature (23–25°C), humidity (57–60%) and light (12 hr dark/12 hr light), with access to filtered drinking water and normal chow diet (Charles River Formular-1; 5LR1) ad libitum. Before the start of experiment, a weight-controlled randomization method was used for grouping of mice (n=10, male, 10-week-old). To minimize order- and time-of-day biases, a counterbalanced zig-zag order across treatment groups was used for all animal operations (exposure, behavioral testing, euthanasia/dissection).
ExposureAcrylamide (A9099, purity >99%, Sigma) was freshly prepared by dissolving in drinking water filtered through a G-10 ion exchange cartridge (Organo, Tokyo, Japan). Wild-type or TNF-α KO mice were administered via oral gavage with acrylamide at 12.5 and 25 mg/kg bw or vehicle (control group; filtered drinking water), 7 days per week, for 4 weeks. The exposure was carried out every day at 9-10 am. Body weight of mice was measured and recorded daily before administration of acrylamide.
Neurobehavioral testsOpen-field test
Open-field test was conducted in a dark and quiet environment at 9 pm of one day prior to dissection day. A square maze box with white acrylic walls (50 × 50 × 50 cm) was used. The central area of the maze box was defined as the middle 25 × 25 cm area of the field. The animals were accommodated to the environment for at least one hour before the behavior test. Video of the animal was collected with an infrared web camera, and a total of 20-min period was recorded after placing the animal into the center of the maze box. The maze box was cleaned with 70% ethanol and dried with clean tissue paper before next animal was put inside. Obtained data were analyzed with ImageJ software. Total distance of mice travelled (cm), distance of mice travelled in the center area (cm), and time spent in the center area (s) were calculated.
Y-maze testY-maze test was conducted in a dark and quiet environment, using a Y-shaped maze with three white-colored, opaque arms (arm A, B, C) orientated at 120-degree angles from each other. Y-maze test was performed after at least a 3 hr interval following open-field test. Video of the animal was collected with an infrared web camera. The mouse was introduced at a particular position of A-arm on the maze and allowed to explore the arms freely over a 10-min period. The Y-maze box was cleaned with 70% ethanol and dried with clean tissue paper before next animal was put inside. For calculation, an entry was defined when all four limbs of the mouse are within an arm. An alternation is defined as consecutive entries into all three arms. The percentage of spontaneous alternation (%) was calculated using the formula: number of spontaneous alternation/(total number of arm entries - 2) × 100%.
Functional Observational Battery (FOB)Functional Observational Battery (FOB) consists of a series of tests that assess the presence and severity of behavioral and/or neurologic dysfunction. In this study, grip strength test was carried out in accordance with the United States Environmental Protection Agency's (USEPA) recommended protocol as described in detail previously (Edwards and Parker, 1977).
Grip strength test was carried out at least 6 hr interval after Y-maze test and before dissection. For grip strength test, a grip strength meter (Melquest Ltd., Toyama, Japan) was used to measure the grip strength of forelimb and all-limbs, following the manufacturer's instructions. For each animal the test was repeated five times and the mean after removing the highest and lowest value was calculated and used as the representative value for the individual mouse for statistical analysis. The equipment was cleaned with 70% ethanol and dried with clean tissue paper before starting measurement for next animal.
Perfusion, embedding, and cryosection of the brainFor histopathological study (n=4), after 28-day exposure, intracardiac perfusion was performed utilizing ascending aorta cannulation with ice-cold 4% paraformaldehyde (PFA, pH 7.4, Wako, Japan). To ensure sufficient perfusion of brain, a high pressure of 120 mmHg was used with a pump. After perfusion, the brain was post-fixed for 1 hr on ice, and then the brain was dissected out and further fixed in 4% PFA for an additional 24 hr at 4°C. After dehydration with a series of 10%, 20% and 30% sucrose solutions, brain tissues were embedded in optimum cutting temperature (OCT) medium (Sakura Finetek, Japan) and stored at -80°C. For cryosection, OCT-embedded brain tissues were serially sectioned in the coronal plane on a cryosectioning microtome (Leica CM3050S; Leica Microsystems, Wetzlar, Germany) at 40 μm thickness from bregma –0.34 according to Mouse Brain in Stereotaxic Coordinates (Paxinos and Franklin, 2019), which is representative of the full extent of somatosensory cortex in mice. The tissue sections were placed on MAS-coating adhesive glass slides (Matsunami Glass, Osaka, Japan), then air-dried, and stored at -30ºC until immunostaining.
Immunohistochemical stainingThe frozen sections were air-dried at room temperature, followed by antigen retrieval with 10 mM sodium citrate buffer (pH 6.0, Wako) in a 95°C water bath for 30 min. Endogenous peroxidase activity was blocked by incubating the sections for 20 min with Bloxal peroxidase blocking reagent (Vector Laboratories, Burlingame, CA, USA), and non-specific protein binding was blocked at 4°C overnight using protein blocking reagent containing 1% bovine serum albumin (Sigma Aldrich), 2.5% normal horse serum (Vector Laboratories), 0.3 M glycine (Wako) and 0.1% Tween-20 (Wako). Endogenous biotin was blocked by avidin/biotin blocking reagent (Vector Laboratories), as described by the manufacturer. The sections were then incubated for 2 hr at 37°C with mouse anti-noradrenaline transporter antibody (1:1000, Abcam, Cambridge, UK). After incubation with the primary antibody, the sections were washed three times in Tris buffered saline with 0.01% Tween-20 (TBST) and then incubated for 1 hr with horse anti-mouse biotinylated secondary antibody (Vector Laboratories) and further washed three times in TBST. Then the sections were incubated with the avidin-biotin peroxidase complex (Vector Laboratories) for 30 min and reacted with 3,3′-diaminobenzidine (DAB) reagent (Vector Laboratories), followed by mounting with an aqueous mounting medium (Vector Laboratories). The slides were then air-dried at room temperature until observation.
Morphometric analysis of noradrenergic axonsStained sections were observed with a microscope (BX 50, Olympus, Tokyo) equipped with a digital camera (FlexCam C1, Leica). Previously we evaluated the effects of exposure to acrylamide on noradrenergic axon degeneration in the secondary somatosensory cortex region (S2) of mice brain (Zong et al., 2025). In this study, brain sub-regions of the primary somatosensory cortex (S1), including forelimb (S1FL), hindlimb (S1HL) and barrel field (S1BF) at Bregma –0.34, were observed and photomicrographs were taken. Length of axons was quantified using the Segmented Line tool of ImageJ software (National Institute of Health, Bethesda, MD, USA).
Isolation of total RNA, synthesis of cDNA and real-time quantitative polymerase chain reaction (real-time qPCR)For biochemical study, following euthanasia of mice with decapitation, the brain was quickly dissected on ice, snap frozen with dry-ice and stored at -80ºC until analysis. Total RNA was isolated from the cerebral cortex (n=6 each group) using the ReliaPrep RNA Tissue Miniprep System (Promega, Madison, WI, USA), following the manufacturer’s instruction. The concentration of the extracted mRNA was measured using NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). The quality of mRNA was determined by confirming that A260/A280>2.0 and A260/A230>2.0. Complementary DNA (cDNA) was then synthesized using Superscript IV Vilo reverse transcriptase (Invitrogen, Carlsbad, CA). Real-time quantitative PCR was performed by using the THUNDERBIRD SYBR NEXT qPCR Mix (Toyobo, Osaka) and the AriaMx Real-Time PCR System (Agilent Technologies, Santa Clara, CA, USA). A standard curve constructed using serial diluted cDNA samples from a pooled sample was used to quantify the expression level of each gene. The mRNA expression was then calculated by standardization to the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Primer sequences are listed in Table 1.
;%0A%09%09%09newWindow.document.open();%0A%09%09%09newWindow.document.write('<img src=%22./Graphics/51_183-t1.jpg%22>');%0A%09%09%09newWindow.document.close();%0A%09%09)
Data were expressed as mean ± standard deviation (SD). Differences among groups in each genotype were analyzed by one-way analysis of variance (ANOVA) followed by Dunnett’s multiple comparison test. For comparison of effects with same concentration among genotypes, Student’s t-test was carried out. Simple regression was used to evaluate linear trend. Multiple regression was used to evaluate interaction between two factors. A probability (p) of <0.05 denoted the presence of a statistically significant difference. All statistical analyses were performed using JMP (Version 17, SAS Institute, Cary, NC, USA).
The reported oral median lethal dose (LD50) value of acrylamide ranges from 107 to 195 mg/kg bw in mice, guinea pigs, and rabbits (US ATSDR, 2012). Following oral administration, acrylamide is rapidly absorbed and widely distributed to various tissues, including liver, lung, muscle, and brain (Doerge et al., 2005). Compared with intravenous route, oral administration was found to attenuate acrylamide bioavailability to 23% from the diet and to 32-52% from aqueous gavage (Doerge et al., 2005). Previous studies reported that after oral administration of [1,3-14C]-labelled acrylamide to F344 rats at 30 mg/kg bw for 13 days, the tissue mean concentration of acrylamide in the brain was 53.52 μg/g (US ATSDR, 2012; Zong et al., 2025). In this study, to investigate the mechanism of acrylamide neurotoxicity at non-lethal doses, 12.5 and 25 mg/kg bw were selected as low and high exposure groups. These doses are also consistent with those commonly used in vivo doses in animal studies on acrylamide neurotoxicity (Ekuban et al., 2021; Davuljigari et al., 2021; Fergany et al., 2023).
No death of mice was observed in wild type group and TNF-α KO group during oral exposure to acrylamide at 0, 12.5 and 25 mg/kg bw for 28 days. Symptoms such as gait abnormality, ataxia, and paralysis, were observed both in wild type group and TNF-α KO group following the procedure of acrylamide exposure. The above typical symptoms of acrylamide toxicity are consistent with previous reports, indicating appropriate administration of acrylamide into mice and good reproducibility of reported acrylamide toxicity.
Exposure to acrylamide at 12.5 and 25 mg /kg bw for 28 days significantly decreased body weight gain, both in wild-type and in TNF-α KO mice (Fig. 2). In wild-type mice, compared to control group, 25 mg/kg bw exposed group showed decreased body weight at day 23, 25, 26, 28, while 12.5 mg/kg bw only at day 28 (Fig. 2A). In TNF-α KO mice, compared to control group, 25 mg/kg bw exposed group showed decreased body weight at day 15-23 and day 28, while 12.5 mg/kg bw only at day 16 and 28 (Fig. 2B). No significant differences were detected between 12.5 and 25 mg/kg bw groups (p>0.05, by Student’s test).
Changes of body weight in wild-type and TNF-α knockout mice after exposure to acrylamide for 28 days. The body weights of mice were measured every day before administration of acrylamide by oral gavage. Data are means of each group. *: p<0.05, 12.5 mg/kg vs relative control. #: p<0.05, 25mg/kg bw vs relative control.
Brain and cerebellum weight significantly decreased after exposure to acrylamide at 25 mg /kg bw for 28 days in wild-type mice and TNF-α KO mice, but was not changed after exposure to 12.5 mg/kg bw (Table 2). Spleen weight and testis weight significantly decreased after exposure to acrylamide at 25 mg /kg bw for 28 days in wild-type mice, but not in TNF-α KO mice. No significant changes were observed in weights of liver or kidney, neither in wild-type nor in TNF-α KO mice (Table 2).
;%0A%09%09%09newWindow.document.open();%0A%09%09%09newWindow.document.write('<img src=%22./Graphics/51_183-t2.jpg%22>');%0A%09%09%09newWindow.document.close();%0A%09%09)
Next, we first attempted to evaluate neurotoxic effects of acrylamide by carrying out neurobehavioral tests, including open-field test and Y-maze test. Open-field test results showed no statistical differences in travelled distance in total area, travelled distance in central area, and time spent in central area (Fig. 3A-E). Y-maze test results showed that there were no significant changes in total entries or spontaneous alteration index in both wild-type and TNF-α KO mice (Fig. 3F-H).
Results of neurobehavior tests in wild-type and TNF-α knockout mice after exposure to acrylamide for 28 days. (A-E) Results of open-field test. (A) Open-field test was performed with a square box (50*50*50 cm). The red rectangle is defined as the center area (25*25*25 cm). The black lines within the box show the trace that the mouse had travelled. (B) Total distance (cm) that mice travelled during 20 min. (C) Distance (cm) that mice travelled in the center area. (D) Percentage of mice travelled in the center compared to the total travelled distance. (E) The time that mice travelled in the center. (F-H) Results of Y-maze test. (F) The Y-maze test was performed by placing the mice into the A-arm of the Y-maze and video was recorded for 10 min. (G) Number of entries into arms. An entry was defined when all four limbs of the mouse are within an arm. (H) Alternative index of mice. An alternation is defined as consecutive entries into all three arms. The percentage of spontaneous alternation (%) was calculated using the formula: number of spontaneous alternation/(total number of arm entries - 2) x 100%. Data are mean ± SD.
Considering that acrylamide exposure induced symptoms such as gait abnormality, ataxia, and paralysis, the above results of no significant changes in general neurobehavior tests such as open-field test and Y-maze test, might be due to that acrylamide exposure affected motor function or neuromuscular function of mice and resulted in altered activity of mice. Therefore, other alternative tools for evaluation of neurotoxicity in acrylamide-exposed mice were explored. The Functional Observational Battery (FOB) is a neurobehavioral assessment tool developed by US-EPA and consists of a series of tests that assess the presence and severity of behavioral and neurologic dysfunction (Moser, 2000). Previous studies showed that landing foot spread test of FOB can be used as marker of acrylamide neurotoxicity (Davuljigari et al., 2021; Zong et al., 2025). In this study, we carried out the grip strength test of FOB for comparison of behavioral changes induced by exposure to acrylamide in wild type and TNF-α KO mice.
Grip strengths for forelimb and all-limb of mice were measured with an animal grip strength meter (Fig. 4). The result showed that exposure to acrylamide at 25 mg/kg bw induced decrease of grip strength of both forelimb and all-limb, in both wild type and TNF-α KO mice. Exposure to acrylamide at 12.5 mg/kg bw did not show significant decrease of forelimb grip strength, both in wild-type and TNF-α KO mice. At 12.5 mg/kg bw exposure, a decrease of all-limb grip strength was only observed in TNF-α KO mice, but not in wild-type mice. It also showed that there is no significant difference between wild type and TNF-α KO mice at same level of acrylamide exposure (p>0.05 by Student’s t-test) (Fig. 4).
Results of grip strength tests in wild-type and TNF-α knockout mice after exposure to acrylamide for 28 days. (A) forelimb and (B) all limb. Grip strength was measured with a grip strength meter (Melquest Ltd.). For each animal the test was repeated five times and the mean after removing the highest and lowest value was calculated and used as the representative value for the individual mouse for statistical analysis. Data are mean ± SD. *p<0.05, **p<0.01, ***p<0.001, ns: not significant.
Degeneration of noradrenergic neurons has been reported to be involved in various neurodegenerative disorders (Weinshenker, 2018) as well as neurotoxicity induced by exposure to occupational and environmental chemicals (Abd El Naby et al., 2023). In this study, to compare the histopathological effects of acrylamide in wild-type and TNF-α KO mice, noradrenergic-immunoreactive axons were stained by anti-noradrenaline transport (NAT) antibody and the length of axons were quantified by ImageJ, in cortex regions including S1HL (primary somatosensory cortex, hindlimb region), S1BF (primary somatosensory cortex, barrel field region), S1FL (primary somatosensory cortex, forelimb region) (Ekuban et al., 2021). The results showed that exposure to acrylamide at 25 mg/kg bw significantly decreased the density of noradrenergic axons of S1FL and S1HL regions, both in TNF-α KO and wild-type mice, while 12.5 mg/kg bw showed no statistical significance in either TNF-α KO or wild-type mice (Fig. 5). Simple regression analysis showed significant negative linear trend (p<0.05) in noradrenergic axons densities of S1FL, S1HL, and S1BF regions for both genotypes (Table 3). Student’s t-test showed significant differences at 12.5 mg/kg bw between wild-type and TNF-α KO mice for both S1FL and S1HL regions (p<0.05), suggesting alleviation effects of TNF-α ablation at 12.5 mg/kg bw, but not at 25 mg/kg bw.
Changes of density of noradrenergic axons in the cortex of wild-type and TNF-α knockout mice after exposure to acrylamide for 28 days. (A) S1FL region. (B) S1HL region. (C) S1BF region. Data are mean ± SD. *p<0.05.
;%0A%09%09%09newWindow.document.open();%0A%09%09%09newWindow.document.write('<img src=%22./Graphics/51_183-t3.jpg%22>');%0A%09%09%09newWindow.document.close();%0A%09%09)
We next compared the expression of major inflammatory cytokines and inflammatory markers in wild-type and TNF-α KO mice. Our previous study already showed that after exposure to acrylamide, TNF-α expression showed a dose-dependent increase in wild-type mice (Zong et al., 2025). In this study, the results further showed that transforming growth factor-beta (TGF-β) also had a dose-dependent increase and a significant increase was observed at 25 mg/kg bw exposure in wild-type mice; and TNF-α deletion successfully suppressed upregulation of TGF-β (Fig. 6A). Expression of NF-κB showed dose-dependent increase with significant changes at 12.5 or 25 mg/kg bw in wild type mice, which was also suppressed by TNF-α deletion (Fig. 6B). On the other hand, IL-6 showed a dose-dependent increase and a significant increase at 25 mg/kg bw exposure, but TNF-α deletion could not suppress the upregulation of IL-6 at 25 mg/kg bw (Fig. 6C). Expression of iNOS showed dose-dependent decrease and significant decrease at 25 mg/kg bw, and TNF-α deletion could not suppress the downregulation of iNOS at 25 mg/kg bw (Fig. 6D). IL-1β and cyclooxygenase-2 (COX2) showed no significant changes (Fig. 6E, 6F).
Relative expression levels of cytokines in wild-type and TNF-α knockout mice after exposure to acrylamide for 28 days. (A) TGF-β, transforming growth factor beta. (B) NF-κB, nuclear factor kappa-light-chain-enhancer of activated B cells. (C) IL-6, interleukin 6. (D) iNOS, inducible nitric oxide synthase. (E) IL-1β, Interleukin-1 beta. (F) COX2, cyclooxygenase-2. Data are mean ± SD. *p<0.05.
Oxidative stress has been reported to be also one of the regulating pathways of acrylamide induced neurotoxicity. Next, we compared the expression of oxidative stress related genes in wild-type and TNF-α KO mice.
At 25 mg/kg bw, expressions of Nrf2-suppressing gene (Kelch-like ECH-associated protein 1) Keap1, and antioxidant genes heme oxygenase-1 (HO1), glutamate-cysteine ligase catalytic subunit (Gclc), glutamate-cysteine ligase modifier subunit (Gclm), superoxide dismutase 1 (Sod1), catalase (Cat), metallothionein 1 (MT-1) were significantly increased in wild type; and TNF-α deletion suppressed acrylamide-induced upregulation of the above oxidative stress-related genes (Fig. 7B, 7C, 7E, 7F, 7G, 7H, 7J). Gstm showed no acrylamide-induced change in wild type mice, but TNF-α deletion significantly decreased Gstm expression (Fig. 7I).
Relative expression levels of genes of oxidative stress pathway in wild-type and TNF-α knockout mice after exposure to acrylamide for 28 days. (A) Nrf2, nuclear factor erythroid 2-related factor 2. (B) Keap1, Kelch-like ECH-associated protein 1. (C) HO1, heme oxygenase-1. (D) Nqo1, NAD(P)H quinone dehydrogenase 1. (E) Gclc, glutamate-cysteine ligase catalytic subunit. (F) Gclm, glutamate-cysteine ligase modifier Subunit. (G) Sod1, superoxide dismutase 1. (H) Cat, catalase. (I) Gstm, glutathione s-transferase Mu 1. (J) MT-1, metallothionein 1. (K) Gsr, glutathione reductase. Data are mean ± SD. *p<0.05.
On the contrary, at 12.5 mg/kg bw, TNF-α deletion enhanced upregulation of NAD(P)H quinone dehydrogenase 1 (Nqo1), Gclm, Sod1, Cat, and glutathione-disulfide reductase (Gsr). At 12.5 mg/kg bw (Fig. 7D, 7F, 7G, 7H, 7K), the expression of Nrf2, Keap1, HO1, glutathione S-transferase mu 1 (Gstm), and MT-1 did not change in wild type or TNF-α mice (Fig. 7A, 7B, 7C, 7I, 7J).
In this study, neurotoxicity of acrylamide was compared in wild-type and TNF-α KO mice, demonstrating that TNF-α deletion alleviated noradrenergic axon degeneration in primary somatosensory cortex S1FL and S1HL regions at 12.5 mg/kg bw, but could not alleviate at 25 mg/kg bw, in agreement with results of our recent study in secondary somatosensory cortex (S2 region) (Zong et al., 2025). Further molecular study found genetic TNF-α deletion suppressed acrylamide-induced upregulation of TGF-β and NF-κB, but did not suppress upregulation of IL-6 at 25 mg/kg bw, suggesting IL-6 might play important alternative roles in the above unmitigated neurotoxicity at high concentration of acrylamide. At 12.5 mg/kg bw, no change of IL-6 expression in both wild-type and TNF-α KO mice suggests that at low acrylamide concentration, the involvement of IL-6 may be limited; meanwhile, TNF-α deletion alleviated acrylamide neurotoxicity at 12.5 mg/kg bw suggests that TNF-α may play important role at relative low acrylamide concentration. On the other hand, at 25 mg/kg bw, upregulation of IL-6 was observed even in TNF-α KO mice, suggesting TNF-α absence didn’t eliminate IL-6 upregulation induced by 25 mg/kg bw acrylamide. These results generated a hypothesis on the possible alternative role of IL-6 independent of TNF-α. Moreover, this study showed dipolar effects of TNF-α deletion on oxidative stress genes, i.e., 1) at 25 mg/kg bw, TNF-α deletion suppressed upregulation of Nrf2-suppressing gene Keap1 and genes of oxidative stress pathway (HO-1, Gclc, Gclm, Sod, Cat, Gstm, MT-1); 2) On the contrary, at 12.5 mg/kg bw, TNF-α deletion accelerated upregulation of Nqo1, Gclm, Sod1, Cat, Gsr.
Genetic ablation of TNF-α alleviated acrylamide induced neurotoxicity, particularly at low concentration exposure. This indicates an important role of TNF-α in neurotoxicity of acrylamide at low concentration. A large population is likely to face the risk of low concentration of acrylamide exposure, such as via daily dietary exposure. Related to this realistic exposure condition, our results provide helpful evidence in understanding mechanism of acrylamide neurotoxicity, and provided valuable strategies in preventing and treating. Targeting TNF-α’s has also been explored in neurodegenerative disease by various studies, given the important role of TNF-α in neuroinflammation and neurodegeneration. Epidemiological studies report that patients treated with the TNF-α blocker Etanercept had a ~60–70% lower risk of developing Alzheimer's Disease (AD) (Torres-Acosta et al., 2020). Clinical trials in AD also reported modest cognitive improvements with Etanercept treatment, and animal models of dementia have shown that preventive and intervention anti-inflammatory therapies (infliximab, etanercept, rapamycin, thalidomide, curcumin, and celastrol) to lower TNF-α have shown improvements in brain pathology and cognitive function in AD rodent models (Decourt et al., 2017).
In addition to TNF-α, our study showed interesting results of roles of other cytokines in the inflammatory pathway. We found suppressive effects of TNF-α deletion on acrylamide-induced upregulation of TGF-β and NF-κB–related signaling, yet IL-6 remained elevated, especially at 25 mg/kg bw. IL-6 is a pleiotropic cytokine that plays complex roles in CNS development, synaptic plasticity and neurodegeneration, and numerous studies have highlighted its ability to sustain or modulate neuroinflammation and axonal degeneration independently of TNF-α (Donnelly and Popovich, 2008; Erta et al., 2012; Kummer et al., 2021; Petković and Castellano, 2016). In this context, the persistence of IL-6 despite suppression of TGF-β and NF-κB in TNF-α KO mice suggests that IL-6 may act as an alternative pro-inflammatory axis that can maintain neurotoxicity at higher acrylamide doses. Although TNF-α is a potent inducer of IL-6 and helps organize a complex cytokine network, yet under inflammatory conditions TNF-α and IL-6 also display partial redundancy and non-overlapping functions, as illustrated by models and clinical data in which IL-6–targeted therapies remain effective in patients or settings that are insufficiently controlled by TNF inhibitors, and combined TNF/IL-6 blockade provides broader suppression of pathogenic T cell and granulocyte responses than either mono-therapy (Belle et al., 2017; Kwon et al., 2013; Liu et al., 2024; Pons-Espinal et al., 2024; Vitcheva et al., 2011). It has been reported that the production of IL-6 is supported by TNF-α-independent signaling redundancy under certain inflammatory contexts (De Cesaris et al., 1998; Hirano, 2021; Tosato and Jones, 1990). For example, ROS can activate mitogen-activated protein kinase pathways (MAPK) signaling —including extracellular signal-regulated kinase (ERK), c-Jun N-terminal kinase (JNK) and p38 MAPK—thereby converging on IL-6 regulatory elements. In cardiac fibroblasts, ROS-sensitive ERK/p38 activation drives IL-6 transcription through a cAMP response element (CRE) in the IL-6 promoter, while NF-κB signaling is not activated (Sano et al., 2001). Moreover, ROS is also a key trigger for the NLRP3 inflammasome, which promotes caspase 1–dependent processing of pro-IL-1β into mature IL-1β and IL-1β then acts through IL-1 receptor signaling to stimulate IL-6 production in multiple stromal/vascular cell types, establishing a TNF-α-independent IL-6 amplification loop (Hirano, 2021; Tosato and Jones, 1990). It has also been reported that IL-6 expression can be gated at the chromatin and epigenetic level of the IL-6 locus. In human neutrophils, the IL-6 locus is constitutively maintained in an inactive chromatin configuration, and stimuli that trigger chromatin remodeling—such as Toll-like receptor 8 ligands (TLR8)—enable robust IL-6 transcription and secretion (Zimmermann et al., 2015). Once this chromatin “lock” is opened, diverse downstream transcriptional programs (including MAPK-driven factors) can more efficiently engage IL-6 regulatory regions, providing a mechanistic basis for apparent TNF-α dispensability in IL-6 induction under selected contexts (Hirano, 2021; Zimmermann et al., 2015). However, the exact molecular mechanism remains largely unknown and thus further study is needed to reveal the role of IL-6, as well as the interactions among TNF-α, IL-6 and Nrf2 pathways. This also emphasized that future interventions of co-targeting IL-6 signaling in addition to TNF-α may provide valuable data to understand the role of proinflammatory signaling in acrylamide neurotoxicity.
In the present study, deletion of TNF-α produced an unexpected, dose-dependent “dipolar” modulation of acrylamide-induced oxidative-stress genes. On one hand, at the higher acrylamide dose (25 mg/kg bw), TNF-α deletion suppressed induction of Keap1, HO-1, Gclc/Gclm, Sod, Cat, Gstm and MT-1, indicating that full activation of the broad antioxidant and detoxification program requires basal TNF-α signaling. On the other hand, at the lower acrylamide dose (12.5 mg/kg bw), TNF-α deletion resulted in enhanced induction of Nrf2-linked genes including Nqo1, Gclm, Sod1, Cat and Gsr. As Nrf2 is known to be a master regulator of cellular defense against electrophiles and oxidative stress, directly controlling the expression of antioxidant and proinflammatory genes [38–41], it is possible to generate a hypothesis that Nrf2 plays a role in the dipolar role of TNF-α, although the present study does not provide any evidence for involvement of Nrf2 at mRNA level in changes in antioxidant gene expression. It has been reported that acrylamide is a soft electrophile that preferentially targets cysteine residues of proteins such as reduced-type glutathione (GSH), which is an antioxidant, inducing depletion of GSH and significantly increases reactive oxygen species (ROS) levels (Tagkalidou et al., 2025). On the other hand, oxidative stress or electrophilic reagents such as acrylamide has been reported to directly modify specific cysteine residues on Keap1, which is an inhibitory protein of Nrf2, resulting in the release of activated Nrf2 and translocation into nucleus to initiate the transcription of a series of downstream antioxidant enzymes and Phase II detoxifying enzymes, such as heme oxygenase-1 HO-1, GCLC and GCLM (Zhang et al., 2011). This mechanism may explain that in this study even without significant induction of Nrf2 at mRNA level, acrylamide can trigger oxidative stress through its electrophilic properties by attacking cysteine protein residual on keap1, and subsequently induced release of Nrf2 and the expression of downstream antioxidant enzymes, such as HO-1, Nqo1, Gclc, Gclm, Sod1, and Cat.
Crosstalk between TNF-α and Nrf2 is not simple. Genetic ablation of Nrf2 exacerbates neurotoxic effects of acrylamide with downregulation of antioxidant genes and upregulation of proinflammatory genes (Ekuban et al., 2021); while Nrf2 activation by sulforaphane attenuated acrylamide-induced neuropathy in mice with upregulation of antioxidant genes and downregulation of proinflammatory genes (Davuljigari et al., 2021). Previous studies in cardiomyocytes show that physiological TNF-α signaling via TNF receptors supports basal Nrf2 activation and maintenance of antioxidant genes, whereas genetic ablation of TNF receptors reduces expression of Nrf2 targets such as Nqo1, HO-1 and G6pd (Shanmugam et al., 2016). However, other reports show that pharmacological disruption of TNF-α signaling can itself provoke an antioxidant program, exemplified by the anti-TNF antibody certolizumab pegol, which induces Nrf2 nuclear translocation, HO-1 upregulation and subsequent suppression of LPS-driven ROS and IL-1β production in human monocytes (Boyer et al., 2016). Further studies are needed to clarify possible involvement of Nrf2 in the dipolar role of TNF-α on antioxidant gene expression.
Our results are consistent with previous studies showing that high-dose acrylamide may produce specific characteristics of neurotoxicity. Lopachin and colleagues indicated that high-dose acrylamide produces a central–peripheral distal axonopathy and recovery can be slow or incomplete after exposure cessation (LoPachin et al., 2003). Previous studies also reported that lower concentrations of acrylamide selectively form adducts with Cys152 site of protein, but at higher concentrations acrylamide reacted with more sites such as Cys156 and Cys247 (Tilson and Cabe, 1979). Our study is also consistent with previous reports indicate that high-dose acrylamide exposure causes deficits that are not easily alleviated by pharmacologic interventions: in vivo NAC (N-acetylcysteine) administration failed to protect against neurotoxicity in rats after exposure to 50 mg/kg bw acrylamide (Kaji et al., 1989; Wispriyono et al., 1999); Nrf2 activation by sulforaphane only partially improved pathology while functional deficits persisted at high dose (Davuljigari et al., 2021); and vitamin E did not protect during active acrylamide feeding but mainly accelerated recovery after exposure cessation (Rahangadale et al., 2012). Under certain occupational or environmental exposure condition, the exposure can be extremely high. For example, a report of cases of human intoxication of acrylamide in drinking water suggested that the exposure dose of acrylamide can be as high as 400ppm (Igisu et al., 1975). Occupational exposure of workers in small factories manufacturing acrylamide was also reported and the workers were exposed to high concentrations of 27-30% w/v aqueous solution of acrylamide via thermal contact, resulting in symptoms of cerebellar dysfunction followed by polyneuropathy (He et al., 1989). Moreover, neurotoxicity of environmental electrophiles is considered to be accumulative, which will enlarge the risk of exposure under long-term effects even at low concentration (Lopachin and Gavin, 2008). The results in this study showed that preventive or treatment strategy may lose function at high dose exposure, highlighting the health risk of exposure to high concentration under certain occupational or environmental exposure conditions.
In this study, routine neurobehavior tests such as open-field and Y-maze test did not show significant changes after exposure to acrylamide. These results revealed that routine neurobehavior tests that depend on movement of animal might have limited efficiency in evaluating neurotoxicity of acrylamide at certain concentrations. This is potentially because acrylamide affects the motor function and neuromuscular function of animals, and our results for the first time provided solid experimental evidence. Indeed, measurement of grip strength, which is one of the FOB tests recommended by US-EPA, showed significantly decreased grip strength for forelimbs and all limbs after exposure to 25 mg/kg bw acrylamide for 28 days. In addition, our previous study reported that landing foot spread test, which is also one test of FOB, also can be used as a tool to evaluate neurotoxicity of acrylamide (Zong et al., 2025). Our study shows that grip strength measurement is also a tool for monitoring neurobehavior effects in neurotoxicity of environmental chemicals such as acrylamide.
Histopathological studies showed that TNF-α KO rescued the noradrenergic axon degeneration in the secondary somatosensory cortex of 12.5 mg/kg bw acrylamide-exposed groups, even the behavior observation such as grip strength did not find significant changes. This is consistent with serials of our previous studies that demonstrated specific noradrenaline axon degeneration as a sensitive and robust biomarker for acrylamide and 1-bromopropane toxicity. Exposure to acrylamide induced noradrenergic axon degeneration in hippocampus and prefrontal cortex (PFC) of rat brain (Zhang et al., 2020), primary/secondary somatosensory cortex (Davuljigari et al., 2021; Fergany et al., 2023; Zong et al., 2025), dorsal medial prefrontal cortex (d-mPFC) and ventral medial prefrontal cortex (v-mPFC) of mice brain (Ekuban et al., 2021). Exposure to 1-bromopropane induced noradrenergic axon degeneration in prefrontal cortex of rats (Mohideen et al., 2011). Axonal retraction and degeneration are common features across neurodegenerative diseases, often emerging early before overt neuronal loss is detected (Ikezu et al., 2020; Salvadores et al., 2020). Our results in cerebral cortex of mice are consistent with these results and further suggest that changes in neurite polarity might be a more sensitive readout than neuron death. Together with our previous studies, this study highlights noradrenergic axon degeneration as a robust marker for monitoring neurotoxicity induced by environmental electrophiles or other chemicals.
One limitation of the study is that TNF-α KO mouse is a genetic ablation model, in which TNF-α is lacking during development stages of mice. Although genetic TNF-α ablation did not affect the survival of animals (Taniguchi et al., 1997), it is unknown whether alternative pathways or mechanisms happened during development periods of animals, or under stimulation by exposure to external factors. Another limitation is that this study focused on histopathological and biochemical changes in cerebral cortex after acrylamide exposure. It has been reported that exposure to acrylamide also induced damage to cerebellum at synapse protein and vesicle levels (Zhang et al., 2017), as well as Purkinje cell and axon damage in the cerebellum of rat (Lehning et al., 2002). Further investigation on the cerebellum will provide valuable information to understand mechanism of acrylamide neurotoxicity.
In conclusions, our results demonstrated that the impact of TNF-α deficiency differed across acrylamide doses, suggesting that the interplay between inflammatory signaling (TNF-α and IL-6) and the oxidative-stress response (Nrf2) may shift in a dose-dependent manner. At a low dose, TNF-α deletion partially attenuated neurotoxicity, whereas at a high dose the electrophilic/oxidative burden may engage compensatory pathways—potentially including IL-6 induction and Nrf2/ARE-centered defense programs—even under TNF-α-deficient conditions, thereby limiting the apparent protection. Importantly, these interpretations are tentative and intended as a conceptual framework; further studies are needed to define the underlying mechanisms, including whether IL-6 upregulation is an upstream driver of neurotoxicity or a secondary consequence of injury, and to clarify the crosstalk among TNF-α, IL-6, and Nrf2 signaling in acrylamide neurotoxicity.
The authors are grateful to Ms. Satoko Arai and Ms. Yuko Uozumi for the excellent secretarial support.
FundingThis research was supported by Japan Society for the Promotion of Science: Grant-in-Aid for Early-Career Scientists #21K17277, and Grant-in-Aid for Scientific Research (B) 19H04279.
Conflict of interestThe authors declare that there is no conflict of interest.
Data availabilityThe data in this study are included in the article/supplementary materials. Contact the corresponding author(s) directly to request the underlying data.
Author contributionsWriting-original draft: C.Z.. Writing- review & editing: C.Z., G.I.. Conceptualization: C.Z., Y.I., S.I, G.I.. Methodology: C.Z., S.H., G.I.. Investigation: C.Z., S.H.. Visualization: C.Z.. Supervision: G.I.. Funding acquisition: C.Z., G.I.. Resources: C.Z., Y.I., S.O., S.I., G.I..
Ethical approval and consent to participateNot applicable.
Patient consent for publicationNot applicable.
