BDNF Protects Against Neuronal Damage Induced by TNF and β-Amyloid Peptides by Targeting JNK Activation

Abstract

Neuroinflammation, driven by β-amyloid peptide accumulation, plays a critical role in the pathogenesis of Alzheimer’s disease, resulting in neurodegeneration and cognitive decline. Inflammatory cytokines, particularly tumor necrosis factor (TNF), adversely affect neuronal function and survival by counteracting the neuroprotective effects of neurotrophins. Importantly, brain-derived neurotrophic factor (BDNF) has been shown to alleviate the neurotoxic effects of pro-inflammatory cytokines. While the mechanisms through which pro-inflammatory cytokines disrupt BDNF/TrkB signaling are well understood, the specific ways in which BDNF protects neurons from inflammatory damage remain unclear. We present evidence that BDNF reduces cytotoxicity and neuritic damage in cholinergic neurons (SN56) induced by TNF and β-amyloid peptide, through the downregulation of c-Jun N-terminal kinase (JNK) activation. BDNF inhibits TNF-induced JNK activation by stimulating p38 mitogen-activated protein kinase. These findings indicate that BDNF restores neuronal functionality by modulating the signaling pathways of inflammatory cytokines, such as TNF, and highlight potential therapeutic strategies to mitigate neuroinflammation-associated neurodegeneration in Alzheimer’s disease.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11064-026-04740-8.

Introduction

Despite the clear genetic influence on the development of Alzheimer’s disease (AD), genetic components account for only 5.5% of total cases. Age and lifestyle factors also contribute significantly to this condition. Notably, oxidative stress and levels of inflammatory markers increase in the brain with age, correlating with synaptic dysfunction [1]. Recent studies indicate that the accumulation of β-amyloid plaques triggers a chronic inflammatory process with neurodegenerative effects, particularly in brain regions essential for cognitive capacity and memory, such as the hippocampus and cerebral cortex. Cumulative evidence indicates that this inflammatory process is initiated and maintained by microglia [2, 3].

In addition to Toll-like receptors, Nod-like receptors (NLRs) also play a role in recognizing β-amyloid peptides [4, 5]. Exposure of microglia to β-amyloid oligomers activates the NLRP3, ASC, and caspase-1 inflammasome complex, promoting the maturation and release of interleukin 1β (IL-1β) [6]. The absence of NLRP3 or caspase-1 prevents IL-1β maturation in response to β-amyloid peptides [7]. In vivo studies have demonstrated that the lack of caspase-1 decreases β-amyloid plaque formation and enhances learning and memory in various AD animal models, partly due to the restoration of microglial phagocytic capacity and neuronal autophagy [6, 8].

IL-1β is crucial for initiating the neuroinflammatory process in response to β-amyloid peptides, as it sustains tumor necrosis factor (TNF) production [7]. TNF also contributes to neurodegeneration and memory loss [9]. TNF signaling exerts dual and context-dependent effects in the nervous system. While TNFR1 activation is strongly associated with neurodegeneration, its genetic deletion in AD models prevents LTP deficits and memory loss [10–12]. Although TNFR2 has traditionally been considered neuroprotective, limiting Aβ pathology and supporting memory through its effects on β-secretase, Tregs, and microglia [13–18], emerging evidence suggests that TNFR2 can also contribute to neuronal loss and cognitive decline, as observed in fragile X models [19–21]. We recently showed that TNF impairs hippocampal synaptic transmission through TNFR2 signaling [22]. Thus, chronic microglial activation and the pro-inflammatory environment strongly correlate with the neuropathological effects associated with AD.

Although microglia release various factors that can alter neuronal function and promote death, they also secrete neurotrophins that support neuronal survival [23]. Brain-Derived Neurotrophic Factor (BDNF) is particularly important in areas with high synaptic plasticity, promoting axonal growth, synaptogenesis, and neuronal survival through its TrkB receptor by activating signaling pathways involving phospholipase Cγ (PLCγ), phosphatidylinositol 3-kinase (PI3K), and extracellular signal-regulated kinases (ERK) [24]. Post-mortem studies of AD patients have shown reduced BDNF immunoreactivity in neurons with neurofibrillary tangles due to Tau hyperphosphorylation [25].

Various studies have demonstrated that BDNF protects against neuronal death induced by β-amyloid peptides and reverses these toxic effects [26]. Additionally, BDNF administration in AD animal models prevents cell death and neuronal atrophy, mitigating cognitive decline [27]. Conditional overexpression of BDNF in the brains of 5xFAD mice under the GFAP promoter reduces β-amyloid plaque size, increases dendritic spines in hippocampal neurons, and improves memory deficits [28]. These findings suggest that BDNF may alter the pro-inflammatory/anti-inflammatory cytokine balance in the brain through unknown anti-inflammatory mechanisms.

In this study, we show that BDNF reduces the harmful effects of TNF and β-amyloid peptides on SN56 cholinergic neuron survival and neuritic growth by decreasing JNK kinase activity. BDNF’s suppression of TNF-induced JNK activation occurs through p38 MAPK. These results collectively imply that BDNF, by regulating the signaling pathways of inflammatory cytokines like TNF, helps restore neuronal function.

Materials and Methods

Cell Culture

SN56 cells were maintained at 37 °C in a proliferation medium (Advanced DMEM supplemented with 2% fetal calf serum, 1% glutamine, and 1% penicillin/streptomycin) in 100 × 20 mm petri dishes stored in 5% CO₂. To differentiate into a cholinergic phenotype, SN56 cells were plated in 35 × 10 mm petri dishes at a cellular density of 70,000 cells per mL, with 1 mL of proliferation medium for 24 h. Afterward, the medium was removed, and the cells were washed with PBS 1x. Then, SN56 cells were supplemented with 2 mL of differentiation medium (advanced DMEM with 1% glutamine, 1% penicillin/streptomycin, and dibutyryl cAMP at a concentration of 1 mM) for 72 h. The differentiation medium was changed to an “incomplete” medium (only advanced DMEM with 1% glutamine and penicillin/streptomycin) 12 h prior to any treatment.

Treatments

Differentiated SN56 cells were treated with TNF (Peprotech; 315–01 A) from previously prepared aliquots in Advanced DMEM at 20 pg/µL and 200 pg/µL concentrations. BDNF (Peprotech; 450-02) treatments were performed using previously prepared aliquots in advanced DMEM medium at 10 ng/µL concentrations. For human β-amyloid 1–42 peptide treatments (GenScript; RP10017) peptides were resuspended in sterile isotonic saline to a concentration of 221 μM and incubated at 37 °C for 24 h to allow oligomerization. The p38 inhibitor (SB203580, Merck; 559389) was used at a concentration of 5 µM diluted in DMSO, 15 min before cells were exposed to TNF or β-amyloid peptides. Rotenone (Sigma-Aldrich 83-79-4) was used as a positive cell death control at 60 µM (when indicated). SN56 cells were differentiated for 72 h. Subsequently, for viability and neurite growth measurements, they were left untreated (-) or treated with BDNF at 30 or 50 ng/mL as indicated in each figure, TNF at 0.1, 1, and 5 ng/mL, and/or human β-Amyloid 1–42 peptide at 10 µM for 24 h. For western blot assays, after the differentiation period, SN56 cells were left untreated (-) or treated with BDNF at 30 or 50 ng/mL (as indicated in each figure) or TNF at 5 ng/mL for the indicated time period in the absence or the presence of the SB203580 p38 inhibitor. The inhibitor was added 15 min before exposure to BDNF or TNF. The assays were carried out with three independent series with two cultures per treatment.

Cellular Viability

SN56 cells were harvested from 35 × 10 mm plates using Trypsin-EDTA 1x (Sigma; T3924) and transferred to 1.5 mL vials. Cell viability was quantified with the trypan blue 0.4% (Gibco; 15250061) exclusion test, cell counting was performed using a Neubauer chamber.

Protein Extraction and Quantification

Cells were collected in 1.5 mL Eppendorf tubes and centrifuged at 21,168 rcf for 1 min. Supernatant was removed, and pellets were resuspended in PBS and centrifuged again. Pellets were stored at -70 °C until used. Pellets were resuspended in 50 µL of lysis buffer (Tris 20 mM pH 7.4, NaCl 137mM, PPiNa 2mM, EDTA 2mM, Triton x-100 at 1%, Glycerol at 10%, DTT 0.5 mM, β-glycerophosphate 25 mM pH 7.4, Na3VO4 200 mM, PMSF 1mM, and the complete protease inhibitor mix (5 mg/mL leupeptin, 5 mg/mL aprotinin and 5 mg/mL anti papain), incubated for 15 min at 4 °C and centrifuge at 18,440 rcf for 15 min at 4 °C. Supernatants were collected and stored at -70 °C until needed. Protein quantification was performed using the Bradford protocol using the iMark BioRad microplate reader.

Western Blot

20 to 30 µg protein extracts were separated by SDS-PAGE on 10% denaturing gels. Proteins were transferred to a nitrocellulose membrane with a current of 100 volts for 90 min with 1 L transfer buffer (Glycine 200 mM, Tris 25mM and Methanol 20%). Membranes were blocked with skimmed milk dissolved with TBS-Tween 0.1% (NaCl 1.5 M, Tris 200 mM, and Tween 20 at 0.1%) for 1 h at room temperature. Membranes were then incubated with primary antibodies for phospho-JNK (cell signaling; Cat. 81E11 at 1:1000 dilution), JNK (cell signaling; Cat. 9252 at 1:1000 dilution), phospho-p38 (cell signaling; Cat. 28B10 at 1:1000 dilution), GAPDH (cell signaling; Cat. 2118 S at 1:1000 dilution) overnight at 4 °C. Primary antibodies were removed, and membranes were washed 3 times for 10 min with TBS-Tween 0.1%, and secondary antibodies were added and incubated at room temperature for 1 h, Goat Anti-Rabbit IgG (Millipore; Cat. 12–348) was used for phospho-JNK (cell signaling; Cat. 4668 S at 1:2000 dilution), JNK (cell signaling; Cat. 9252 S at 1:2000 dilution), GAPDH (cell signaling; Cat. 2118 S at 1:6000 dilution) or Goat Anti-Mouse IgG (Millipore; Cat. 12–349) for phospho-p38 (cell signaling; 28B10 at 1:1000 dilution), after which membranes were washed 3 times for 10 min with TBS-Tween 0.1%. Band signals were detected by chemiluminescence using the C-Digit reader from LiCOR. To determine total JNK levels, blots were performed on stripped membranes used to assess p-JNK levels. Briefly, Nitrocellulose membranes were stripped by incubation in stripping buffer (200 mM glycine, 500 mM NaCl, pH 2.5) for two 20-min washes, followed by two washes with high-salt TBS-T (0.1%) for 10 min each and two additional washes with TBS-T (0.1%) for 10 min each. After stripping, membranes were incubated with ECL reagent to confirm the absence of residual signal before reprobing.

Length and Number of Neurites

Cell cultures were photographed under light microscopy using a Nikon Eclipse TS100 microscope; cell culture dishes were taken out individually for no more than 1 min per dish to avoid effects from ambient conditions. 5 optical fields per condition were selected with roughly equal cell density from each treatment previously mentioned. Images were analyzed for the number of neurites using the Cell Counter plugin for ImageJ, where, depending on the number of neuritic processes, cells were categorized into cells with 3 to 5 neuritic processes and cells with more than 5 neuritic processes. Neuritic length was determined by measuring the entire length, starting from the soma until the end of the projection. The neuritic processes from 10 different cells were measured. We used the NeuronJ [29, 30] plugin for ImageJ where we evaluated 10 cells from each field, and lengths were categorized into groups, minimum length, and maximum length. To minimize bias, images were captured by an independent person who was not involved in performing the experiments and was blinded to treatment groups.

Statistical Analysis

The Shapiro-Wilk normality test and Q-Q plots were conducted to evaluate the Gaussian distribution [31]. When there was a non-remarkable deviation from the Gaussian distribution in accordance with the Shapiro-Wilk test, we verified the skewness (for values − 2 > ≈ and ≈ < 2 we used a parametric statistic) [32]. The Brown-Forsythe test was used to assess homogeneity of variance.

Comparisons of means across more than two groups were performed using one-way analysis of variance (ANOVA), followed by Tukey’s post hoc test, after verifying assumptions of normality and homogeneity of variances. In contrast, Welch’s ANOVA followed by Dunnett’s T3 post hoc test was used when unequal variance was found.

Kruskal-Wallis, followed by Dunn’s post hoc test, was used for non-Gaussian distribution data.

For viability and neurite quantification, the data represent the results of three independent experiments, each in duplicate. Data are expressed as mean ± standard deviation (SD) or standard error of the mean (SEM) and were considered significant when p < 0.05. GraphPad Prism 9.0 was used for data analysis.

Results

BDNF Maintains the Viability of Cholinergic Neurons in the Presence of TNF

Previously, we demonstrated that exposure of cholinergic neurons (SN56) to TNF has detrimental effects on cell viability alone [20] or in combination with BDNF at various concentrations of TNF and of BDNF. However, we observed that the deleterious effect of TNF (0.1 ng/mL) was reduced when BDNF was present at 30 ng/mL, compared to 10 ng/mL (Supplementary Fig. 1A). To investigate whether BDNF can protect against inflammatory signals, specifically TNF, we increased the concentration of BDNF to 50 ng/mL while maintaining the same TNF concentrations that were previously found to be detrimental to cholinergic neuron viability. Consistent with our earlier findings, BDNF at a concentration of 30 ng/mL effectively preserved the total number of cells (Fig. 1A) and the viability percentage (Supplementary Fig. 1B) of cholinergic neurons when exposed to a low concentration of 0.1 ng/mL TNF. However, at higher cytokine concentrations (1 and 5 ng/mL), both cell viability and the total cell count decreased, despite the presence of BDNF (Fig. 1A and Supplementary 1). Notably, when the concentration of BDNF was raised to 50 ng/mL, both the total number of cells (Fig. 1A) and cell viability (Supplementary Fig. 1) remained at levels comparable to the control, indicating that higher concentrations of BDNF can exert a protective effect, preventing the death of cholinergic neurons induced by high TNF concentrations. Although the BDNF-mediated increased viability over control cells was not sustained in the presence of 1 ng/mL or 5 ng/mL TNF.

Fig. 1.

BDNF prevents TNF deleterious effect on SN56 neuronal viability and neuritic growth. SN56 cells were differentiated for 72 h. Subsequently, they were left untreated (-) or treated with BDNF or/and TNF at the indicated concentrations for 24 h, and cell viability was analyzed as well as the number and lengths of neuritic growth. A Percentage of cell viability. One-way ANOVA followed by Tukey; F_{9, 20}= 17.59, p = < 0.0001; B Percentage of cells with 3 to 5 neuritic processes. Welch’s ANOVA followed by Dunnett T3; W (DFn, DFd) = 15.76 (9.000, 117.9), p = < 0.0001; C Percentage of cells with more than 5 neuritic processes. Kruskal-Wallis followed by Dunn’s; H = 131.9, p = < 0.0001; D Minimum lengths of neuritic processes. Kruskal-Wallis followed by Dunn’s post hoc test; H = 59.58, p = < 0.0001; E) Maximum lengths of neuritic processes. Kruskal-Wallis followed by Dunn’s post hoc test; H = 76.03, p = < 0.0001. Data represent the results of three independent experiments, each in duplicate (± SD). *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001 vs. Control; +p < 0.05, ++p < 0.01, +++p < 0.001, ++++p < 0.0001 vs. 30 ng/mL BDNF; ♦p < 0.05, ♦♦p < 0.01, ♦♦♦p < 0.001 and ♦♦♦♦p < 0.0001 vs. 50 ng/mL BDNF, #p < 0.05 ##p < 0.01; ###p < 0.001 and ####p < 0.0001 vs. 5 ng/mL TNF; ●p < 0.05 and ●●●p < 0.001 vs. 30 ng/mL BDNF + 0.1 ng/mL TNF; ^p < 0.05 and ^^^^p < 0.0001 vs. 30 ng/mL BDNF + 1 ng/mL TNF; ▼p < 0.05, ▼▼▼▼p < 0.0001 vs. 30 ng/mL BDNF + 5 ng/mL TNF; ○○<0.01p vs. BDNF 50 ng/mL + 0.1 ng/mL TNF

BDNF Protects Cholinergic Neurons Against TNF-Deleterious Effects on Neurite Outgrowth

Since it has been demonstrated that TNF inhibits neurite outgrowth [33–35], we hypothesized that BDNF would preserve cell viability and promote neuritic growth in the presence of TNF. After 24 h of treatment, we captured light microscopy images (Supplementary Fig. 2) and quantified the number and length of neurites. We categorized cells based on the number of processes, 3–5 and more than five processes per cell.

Cells treated with TNF (5 ng/mL) alone showed a significant reduction in the number of cells with 3–5 processes by approximately 30% compared to control cells (Fig. 1B). This loss of processes was not prevented by BDNF at 30 ng/mL. However, BDNF at 50 ng/mL protected against TNF-induced loss, as the number of cells with 3–5 processes remained comparable to the control group (Fig. 1B). The most striking effect was observed in cells with more than five processes, where any TNF treatment, even combined with 30 ng/mL of BDNF, reduced the number of such cells by approximately 50% (Fig. 1C). Remarkably, BDNF at 50 ng/mL significantly preserved the percentage of cells with more than five processes to levels observed in the control cells (Fig. 1C). Overall, BDNF at 50 ng/mL, but not 30 ng/mL, effectively protects cholinergic neurons from TNF-induced neuritic degeneration.

We also measured the length of neuritic processes, both minimum and maximum. TNF (5 ng/mL) reduced the minimum length of processes from approximately 40 μm to 32 μm compared to untreated cells (Fig. 1D). BDNF (30 ng/mL and 50 ng/mL) promoted neurite elongation to 55 μm; however, this effect was diminished when TNF was present (Fig. 1D). Despite TNF’s negative impact, BDNF prevented further shortening, as the neurites lengths were like those in untreated cells (Fig. 1D). Regarding maximum neurite length, TNF reduced the length from around 1000 μm to 600 μm compared to controls (Fig. 1E). BDNF at 50 ng/mL restored maximum length, whereas BDNF at 30 ng/mL did not prevent the TNF-induced reduction (Fig. 1E). These findings suggest that BDNF, particularly at 50 ng/mL, protects cholinergic neurons against the neurodegenerative effects of pro-inflammatory cytokines such as TNF.

BDNF Impairs TNF Signaling in Cholinergic Neurons by Promoting JNK Inactivation in a p38 MAPK-Dependent Manner

Our results demonstrate that BDNF mitigates the deleterious effects of TNF on the viability, number, and length of neuritic processes in SN56 cholinergic cells. To elucidate the molecular mechanism by which BDNF antagonizes TNF signaling, leading to neurite degeneration, we evaluated JNK activation downstream of TNF receptor activation. It is well-established that JNK mediates TNF-induced neurodegeneration [36].

As anticipated, exposing SN56 cholinergic neurons to TNF (5 ng/mL) resulted in increased JNK activation, indicated by sustained JNK phosphorylation for up to 30 min relative basal JNK activity found in untreated and cells exposed to BDNF only (Fig. 2A and B). Notably, when neurons were treated simultaneously with TNF (5 ng/mL) and BDNF (30 ng/mL), JNK phosphorylation levels were reduced after 30 min compared to neurons treated with TNF alone (Fig. 2A and B), primarily affecting the p46 JNK isoform.

Fig. 2.

BDNF attenuates TNF-mediated JNK activation. SN56 cells were differentiated for 72 hours. Subsequently, they were left untreated (-) or treated with BDNF or/and TNF at the indicated concentrations for 15 or 30 minutes. “^” Indicates 15 minutes pre-treatment before subsequent exposure to BDNF (30 ng/mL) and/or TNF (5 ng/mL). A Total cell extracts were prepared, and the levels of phosphorylated JNK (pJNK), total JNK, and GAPDH were determined by immunoblot using specific antibodies. B Densitometry and graphical representation of the dynamics of 46 kDa JNK activation (left panel). One-way ANOVA followed by Tukey; F_{9, 20}= 20.98, p = < 0.0001. Densitometry and graphical representation of the dynamics of 54 kDa JNK activation (right panel). One-way ANOVA followed by Tukey; F_{9, 20}= 8.45, p = < 0.0001. Data represent the results of three independent experiments (± SD). Data represent the results of three independent experiments (± SD). ++p < 0.01, +++p < 0.001 and ++++p < 0.0001 vs 30 ng/mL BDNF 15’; ♦p < 0.05 vs 30 ng/mL BDNF 30’; #p < 0.05, ##p < 0.01 and ####p < 0.0001 vs 5 ng/mL TNF 30’; ●p < 0.05 vs 5 ng/mL TNF + 30 ng/mL BDNF 30’. The red dotted line represents basal JNK activity in untreated cells

Importantly, we found that BDNF can attenuate TNF-mediated JNK activation if administered after TNF treatment. Furthermore, preincubating SN56 neurons with BDNF for 15 min before TNF exposure completely reduced JNK phosphorylation levels to those found in untreated neurons, impacting both the p46 and p54 JNK isoforms (Fig. 2A and B). The observation that JNK phosphorylation levels at 15 min post-treatment remain unaffected by BDNF suggests that BDNF does not inhibit initial JNK activation, rather, it accelerates JNK inactivation.

Thus, our data indicate that BDNF reduces the activation levels of TNF-induced JNK kinase, suggesting that BDNF’s antagonistic action against TNF signaling occurs through this pathway.

Considering that p38 MAPK, a kinase involved in the response to cellular stress and inflammatory cytokines, can suppress JNK activation [37], we investigated whether BDNF regulates TNF-induced JNK activity in cholinergic neurons via p38 MAPK. First, we assessed phosphorylated p38 at Thr180 and Tyr182 as markers of activation in SN56 neurons treated with 50 ng/mL of BDNF using specific antibodies. Our results demonstrated oscillatory activation kinetics of p38 in response to BDNF, with initial activation at 5 min, a decrease at 10 min, and reactivation at 15 and 30 min (Fig. 3A). This indicates that BDNF induces p38 activation in cholinergic neurons.

Fig. 3.

BDNF impairs TNF-mediated JNK activation via p38 MAPK. A SN56 cells were differentiated for 72 h. Subsequently, treatment with BDNF was conducted at the indicated time points (for the 0-minute time point, no treatment was administered). Total cell extracts were prepared, and the levels of phosphorylated p38 (p-p38) and GAPDH, used as a loading control, were determined by immunoblot using specific antibodies (upper panel). Densitometry and graphical representation of the dynamics of p38 activation (lower panel). Data represent the results of three independent experiments (± SD). *p < 0.05 and **p < 0.01 vs. Control using Kruskal-Wallis followed by Dunn’s post hoc test; H = 17.19, p = 0.009. B After the differentiation period, SN56 cells were left untreated (-) or treated with BDNF (50 ng/mL) or TNF (5 ng/mL) at the indicated concentrations for 30 min in the absence or the presence of the SB203580 p38 inhibitor. “^” Indicates 15 min pre-treatment before subsequent exposure to BDNF or TNF. SB203580 (5 µM) was added 15 min before the treatments with BDNF and TNF or the pre-treatments. Total cell extracts were prepared, and the levels of phosphorylated JNK (pJNK), total JNK, and GAPDH were determined by immunoblot using specific antibodies (upper panel). Densitometry and graphical representation of the activation dynamics of 46 kDa (left lower panel). One-way ANOVA followed by Tukey; F 10,22 = 32.78, p < 0.0001. Densitometry and graphical representation of the activation dynamics of 54 kDa JNK (right lower panel). One-way ANOVA followed by Tukey; F 10,22 = 56.08, p < 0.0001. Data represent the results of three independent experiments (± SD). ♦p < 0.05 and ♦♦♦♦p < 0.0001 vs. BDNF; #p < 0.05, ##p < 0.01, ###p < 0.001 and ####p < 0.0001 vs. TNF; ●●●●p < 0.0001 vs. BDNF + TNF; *p < 0.05 and ****p < 0.0001 vs. Control + SB203580; ^^^^p < 0.0001 vs. TNF 15´+ BDNF 30´ + SB203580; ▼▼▼▼p < 0.0001 vs. BDNF 15´+TNF 30´ + SB203580. Untreated cells were normalized to 1 as indicated with a red dotted line

Consistent with p38 MAPK negatively regulating JNK activity [37], we found that pharmacological inhibition of p38 restored TNF-induced JNK phosphorylation levels in neurons treated with BDNF, regardless of whether BDNF was administered before, simultaneously, or after TNF (Fig. 3B). Both p46 and p54 JNK isoforms showed enhanced phosphorylation levels in the presence of the SB203580 p38 MAPK inhibitor (Fig. 3B). These results suggest that BDNF-mediated activation of p38 MAPK is crucial for JNK inactivation and for impairing TNF neurodegenerative signals.

Next, we investigated whether p38 MAPK mediates the protective effect of BDNF against TNF-induced neuronal death. SN56 cholinergic neurons were treated with BDNF (50 ng/mL) and TNF (5 ng/mL) in the presence or absence of the specific p38 inhibitor (SB203580) to assess cell viability. Consistent with previous findings, BDNF treatment attenuated TNF-induced neuronal death, as indicated by similar cell viability (Fig. 4A) and total cell counts (Supplementary Fig. 3) compared to untreated control neurons. However, administration of the SB203580 p38 MAPK inhibitor completely abolished the protective effects of BDNF on cell viability, significantly decreasing total cell numbers and viability percentages compared to untreated cells. Notably, the neuronal death observed in the presence of the p38 inhibitor in cells treated with both BDNF and TNF mirrored that of cells treated only with TNF (Fig. 4A and Supplementary Fig. 3). These results underscore the crucial role of p38 MAPK in mediating the protective effects of BDNF against TNF-induced neuronal death.

Fig. 4.

BDNF protection against TNF-mediated deleterious effects on SN56 neuronal survival and neurite growth involves p38 MAPK activity. SN56 cells were differentiated for 72 h. Subsequently, they were left untreated (-) or treated with BDNF and/or TNF at the indicated concentrations for 24 h in the absence or presence of the SB203580 p38 inhibitor. SB203580 (5 µM) was administered 15 min before BDNF or TNF treatments, respectively. Cell viability and the number and lengths of neuritic growth were analyzed. A Percentage of cell viability. One-way ANOVA followed by Tukey; F_{7, 16}= 51.85, p = < 0.0001; B Percentage of cells with 3 to 5 neuritic processes. Kruskal-Wallis followed by Dunn’s post hoc test; H = 27.21, p = 0.0003; C Percentage of cells with more than 5 neuritic processes. Kruskal-Wallis followed by Dunn’s post hoc test; H = 80.86, p = < 0.0001; D Minimum lengths of neuritic processes. Kruskal-Wallis followed by Dunn’s post hoc test; H = 32.62, p = < 0.0001; E Maximum lengths of neuritic processes. Kruskal-Wallis followed by Dunn’s to compare the mean rank of each column; H = 68.04, p = < 0.0001. Subsequent pairwise comparison for BDNF 50 ng/mL + TNF 5 ng/mL vs. BDNF 50 ng/mL + TNF 5 ng/mL + SB203580 using Kruskal-Wallis showed p = 0.055. Data represent three independent experiments, each in duplicate (± SD or ± SEM). *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001 vs. Control; ^♦p < 0.05, ^♦♦p < 0.01, ^♦♦♦p < 0.001 and ^♦♦♦♦p < 0.0001 vs. 50 ng/mL BDNF; ^#p < 0.05, ^##p < 0.01 and ^####p < 0.0001 vs. 5 ng/mL TNF; ^●p < 0.05 and ^●●●●p < 0.0001 vs. 50 ng/mL BDNF + 5 ng/mL TNF; ^○○○○p < 0.0001 vs. Control + SB203580; ^^p < 0.01 and ^^^^p < 0.0001 vs. 50 ng/mL BDNF + SB203580

We also evaluated the role of p38 MAPK in the BDNF signaling pathway that mitigates TNF’s negative effects on neurite homeostasis in cholinergic neurons. Representative images for each experimental condition were obtained (Supplementary Fig. 4), and the number and length of neurites were quantified. Consistent with our previous findings, cells treated only with TNF showed a significant reduction in the number of cells with 3–5 process relative to cells treated with BDNF (Fig. 4B). Accordingly, the combination of the p38 inhibitor with BDNF and TNF abolished BDNF’s protective effect, resulting in a clear reduction in the number of cells with 3–5 processes compared to those treated with both BDNF and TNF without the inhibitor (Fig. 4B). Interestingly, the p38 inhibitor also led to a decrease in the number of processes per cell, particularly for those with more than five processes, compared to cells treated only with BDNF (4 C). These results indicate that p38 signaling is involved in the protective effects promoted by BDNF on neuritic growth.

We also measured neuritic process lengths, which displayed patterns similar to the previous experiments. Treatment with TNF alone significantly decreased the minimum process length compared to control and BDNF-treated cells; in this case, BDNF did not mitigate the TNF-induced reduction, as lengths were comparable to those of TNF-only treated cells (Fig. 4D). Interestingly when the p38 inhibitor was applied, the minimum lengths of control and BDNF-treated cells slightly decreased compared to untreated cells (Fig. 4D). However, combined treatments of BDNF and TNF in the presence of the SB203580 inhibitor prevented the decrease in minimum neurite lengths compared to controls without the inhibitor, resembling lengths of untreated cells and those treated with BDNF alone (Fig. 4D).

Regarding maximum process lengths, we observed a clear reduction in response to TNF treatment (Fig. 4E), which was prevented by BDNF, as those cells exposed to both BDNF and TNF exhibited lengths comparable to control cells or BDNF-treated cells (Fig. 4E). Notably, the presence of the p38 MAPK inhibitor did not impact the protective effect of BDNF on maximum length of neurites (Fig. 4E), suggesting that p38 does not significantly influence the BDNF signaling pathway that counteracts TNF’s negative effects on neuritic process length. Overall, these findings underscore the role of p38 MAPK in mediating the antagonistic actions of BDNF against TNF-mediated neurodegeneration. Inhibition of p38 MAPK impairs BDNF’s beneficial effects on neuronal viability and, to a lesser extent, neurite growth, achieved through p38 MAPK-mediated inactivation of JNK.

Effects of β-amyloid Peptides on SN56 Neurons

Given that β-amyloid peptides contribute to initiating and maintaining inflammation in brain regions associated with memory and learning [38, 39], and prior studies have demonstrated that β-amyloid peptides reduce the survival of primary cortical neurons [26], we aimed to determine whether the detrimental effects of β-amyloid peptides on SN56 cells could be mitigated by BDNF, similar to the protective effects observed against TNF-mediated neuronal damage.

Our results show that exposing SN56 neurons to β-amyloid peptides at a concentration of 10 µM significantly decreased cell viability after 24 h (Fig. 5A and Supplementary Fig. 5A). However, the administration of BDNF at a concentration of 50 ng/mL alongside the peptides maintained cell viability at levels comparable to untreated control cells. Notably, this protective effect was lost when the cells were pre-treated with the p38 inhibitor SB203580 (Fig. 5A and Supplementary Fig. 5A). These findings suggest that while β-amyloid peptides adversely affect SN56 neuronal viability, BDNF can partially rescue these detrimental effects.

Fig. 5.

BDNF protection against β-amyloid peptides-mediated deleterious effects on SN56 neurons survival and neurite growth involves p38 MAPK activity. SN56 cells were differentiated for 72 h. Subsequently, they were left untreated (-) or treated with BDNF and/or β-amyloid peptides (βA) at the indicated concentrations for 24 h in the absence or the presence of the SB203580 p38 inhibitor. Cell viability and the number and lengths of neuritic growth were analyzed. A Percentage of cell viability. One-way ANOVA followed by Tukey; F_{4, 10}= 33.21, p = < 0.0001; B Percentage of cells with 3 to 5 neuritic processes. One-way ANOVA followed by Tukey; F_{4, 145}= 5.19, p = 0.0006; C Percentage of cells with more than 5 neuritic processes. Welch’s ANOVA followed by Dunnett T3; W (DFn, DFd) = 30.54 (4.000, 70.70), p = < 0.0001; D Minimum lengths of neuritic processes. Kruskal-Wallis followed by Dunn’s post hoc test; H = 14.65, p = 0.0055; E Maximum lengths of neuritic processes. Kruskal-Wallis followed by Dunn’s post hoc test; H = 24.71, p = < 0.0001. SB203580 (5 µM) was administered 15 min before BDNF and β-Amyloid treatment. Data represent three independent experiments, each in duplicate (± SD or ± SEM). **p < 0.01, ***p < 0.001 and ****p < 0.0001 vs. Control; ^♦♦p < 0.01 and ^♦♦♦♦p < 0.0001 vs. BDNF; ^#p < 0.05 and ^##p < 0.01 vs. βA

Furthermore, we examined the impact of β-amyloid peptide treatments on neurite growth. Representative images for each experimental condition were obtained based on viability assays. Cells treated with β-amyloid peptides displayed fewer processes per cell, and these processes were also shorter (Supplementary Fig. 5B). In contrast, cells exposed to β-amyloid peptides in the presence of BDNF exhibited a greater number of processes, which were longer (Supplementary Fig. 5B). Interestingly, unlike TNF effects, exposure to β-amyloid peptides did not affect the number of cells with 3–5 processes when compared to untreated cells or those exposed to BDNF (Fig. 5B). Additionally, the p38 inhibitor did not alter the number of cells with 3–5 processes in combination with BDNF and β-amyloid peptides compared to those treated with the combination of BDNF and β-amyloid peptides in the absence of the inhibitor (Fig. 5B). Nonetheless, the significant reduction in the number of cells with more than five processes resulting from β-amyloid exposure was partially reversed by BDNF (Fig. 5C). However, this effect was not mediated by p38 MAPK, as cells pre-treated with the inhibitor and subjected to the combined treatment of BDNF and β-amyloid peptides exhibited processes compared to cells treated without the inhibitor (Fig. 5C), indicating that, unlike TNF signaling, the mechanism by which β-amyloid peptides impairs neurite growth is insensitive to BDNF mediated p38 MAPK activation.

We also assessed the effect of β-amyloid peptides on the lengths of SN56 dendritic processes. Treatment with β-amyloid peptides significantly reduced the minimum length of the processes compared to BDNF-treated cells. In contrast, combined treatments (β-amyloid peptides and BDNF) exhibited lengths comparable to those of cells treated with BDNF alone (Fig. 5D). Interestingly, the p38 MAPK inhibitor had no effect on the protective effect of BDNF against β-amyloid exposure, as the minimum lengths observed in these cells were comparable to those observed in cells treated with the β-amyloid and BDNF combination in the absence of the inhibitor (Fig. 5D). In contrast, we observed that for the maximum neuritic lengths, the p38 MAPK inhibitor impaired the BDNF ability to prevent TNF-induced neurite shortening (Fig. 5E). Our data suggest that p38 MAPK plays a significant role in the signaling cascade triggered by BDNF resulting in neuroprotection against the negative impact of β-amyloid peptides on neuronal survival and neurite homeostasis.

Given that β-amyloid peptides induce the expression of inflammatory genes through JNK phosphorylation [28], and that TNF-induced JNK activation is suppressed by p38 MAPK in response to BDNF (Figs. 2 and 3), we investigated whether BDNF employs a similar mechanism to prevent β-amyloid–mediated JNK activation. To this end, SN56 cholinergic cells were treated with β-amyloid peptides alone or in combination with BDNF (15-minute preincubation), with or without a p38 inhibitor. As expected, JNK phosphorylation levels increased following treatment with β-amyloid peptides. Notably, preincubation with BDNF before β-amyloid exposure significantly reduced JNK activation in the 54 kDa isoform and partially in the 46 kDa isoform (Fig. 6A and B). In contrast, inhibition of p38 reversed this effect, resulting in increased 54 kDa JNK isoform activation compared to cells treated with BDNF in the absence of the inhibitor (Fig. 6A and B). Additionally, the levels of the phosphorylated 54 kDa JNK isoform were restored to those of cells treated only with β-amyloid peptides; this effect was not observed on the 46 kDa JNK isoform (Fig. 6A and B). These findings confirm that the neurodegenerative effects of β-amyloid peptides are mediated through JNK and further demonstrate that p38 MAP kinase–dependent inactivation of JNK is a critical mechanism by which BDNF exerts its neuroprotective effects.

Fig. 6.

BDNF impairs β-amyloid-mediated JNK activation via p38 MAPK. SN56 cells were differentiated for 72 h. Subsequently, they were left untreated (-) or treated with β-amyloid peptides for 5 min or/and BDNF (administered 15 min before with β-amyloid peptides) in the absence or presence of p38 inhibitor (SB203580, 5 µM). SB203580 was added 15 min before the treatments with BDNF and β-amyloid peptides. A Total cell extracts were prepared, and the levels of phosphorylated JNK (pJNK), total JNK, and GAPDH were determined by immunoblot using specific antibodies. B Densitometry and graphical representation of the activation dynamics of the 46 kDa JNK (Upper panel). One-way ANOVA followed by Tukey; F 4,10 = 20.10, p < 0.0001. Densitometry and graphical representation of the activation dynamics of the 54 kDa JNK (Lower panel). One-way ANOVA followed by Tukey; F 4,10 = 33.83, p < 0.0001. Data represent the results of three independent experiments (± SD). **p < 0.01 and ***p < 0.001 vs. βA; ♦♦p < 0.01, ♦♦♦p < 0.001 and ♦♦♦♦p < 0.0001 vs. βA + BDNF; ⁺p < 0.05 and ⁺⁺p < 0.01 vs. Control + SB203580; ###p < 0.001 vs. βA + SB203580. Untreated cells were normalized to 1 as indicated with a red dotted line

Discussion

It’s customary to name new molecules based on their function and the tissue where they were first discovered. However, it has become evident that such names often fail to capture the full spectrum of a molecule’s functions. For instance, cytokines, once believed to primarily coordinate inflammation and immune responses, have also been found to act as neuromodulators [40]. Conversely, well-established neurotransmitters can modulate immune cell functions [41], and even hormones traditionally associated with roles outside the immune system may exhibit pro-inflammatory properties [42]. Neurotrophins, including brain-derived neurotrophic factor (BDNF), are no exception, as recent evidence suggests that, in addition to their classical neurotrophic effects, they also play a role in regulating inflammation [43].

BDNF, well-known for its neuroprotective function, has demonstrated anti-inflammatory properties in several experimental settings. For example, intracerebroventricular administration of BDNF in a model of meningitis induced by Streptococcus pneumoniae has been shown to reduce the levels of pro-inflammatory cytokines such as TNF, IL-1β, and IL-6, while increasing anti-inflammatory cytokines like IL-10 in the cortex and hippocampus, ultimately leading to reduced neuronal death [44]. Additionally, in vitro studies indicate that BDNF treatment can prevent the accumulation of cyclooxygenase-2 (Cox-2), an enzyme critical for sustaining inflammatory responses in the central nervous system (CNS) [45], and suppress LPS-induced cytokine production in cortical primary microglia cultures [46].

In the context of Alzheimer’s disease (AD), recent research has shown that enhancing BDNF expression in astrocytes, controlled by the GFAP promoter responsive to inflammatory signals, can effectively mitigate neuroinflammation and cognitive impairment in the 5XFAD familial AD mouse model [28]. These findings collectively suggest that BDNF exerts an anti-inflammatory effect by shifting the balance between pro-inflammatory and anti-inflammatory cytokines in the brain.

While the molecular mechanisms by which pro-inflammatory cytokines disrupt BDNF/TrkB neuroprotective signaling are well-documented [47, 48], the precise ways in which BDNF mitigates inflammation remain largely unexplored. Our study provides evidence for the first time that BDNF prevents TNF-induced neurodegenerative signals in cholinergic neurons by activating the p38 MAPK pathway, which in turn attenuates TNF-induced JNK activation.

We showed that BDNF protects cholinergic SN56 neurons from TNF-induced cell death and preserves neurite stability. SN56 neurons were susceptible to TNF-induced cell death at concentrations as low as 1 ng/mL, with lower concentrations (0.1 ng/ml) proving insufficient to induce neuronal death. Interestingly, BDNF at 30 ng/mL did not protect cholinergic neurons from the negative effects of TNF when used at one or five ng/ml. However, increasing BDNF concentration to 50 ng/mL effectively counteracted TNF-induced cell death. These findings underscore that the survival or death of SN56 neurons depends on a subtle balance between cytokine and neurotrophic factor signaling, influenced by the concentration of each factor.

Healthy axons and neurites are crucial for neuronal function; unrepaired damage can lead to cell death [49]. BDNF plays a key neuroprotective role by promoting axonal and neurite homeostasis. Accordingly, BDNF counteracts the detrimental effects of TNF on dendrites. SN56 neurons treated with TNF showed reduced numbers and lengths of neurites compared to control cells, an effect reversed by BDNF, particularly at the higher concentration of 50 ng/mL.

Consistent with BDNF’s neuroprotective functions against the detrimental effects of β-amyloid peptides on neuronal survival and function [26], our findings demonstrate that BDNF effectively protects SN56 cholinergic neurons from cell death and neurite damage induced by these peptides. The most pronounced neuroprotective effects were observed at a concentration of 50 ng/mL, highlighting the importance of BDNF signal strength in counteracting harmful stimuli such as TNF and β-amyloid peptides while preserving neuronal homeostasis.

Crosstalk between signaling pathways is crucial in determining cellular responses to environmental conditions. Inflammatory signaling can disrupt insulin signaling, resulting in insulin resistance [50], while inhibition of BDNF signaling can impair synaptic transmission, contributing to cognitive deficits [51]. Given that JNK activation is a key factor mediating the neurodegenerative effects of both TNF and β-amyloid peptides and considering that p38 MAPK negatively regulates JNK activity [37], we speculate that BDNF’s interaction with the TrkB receptor may activate p38 MAPK, thus mitigating JNK activation and reducing neurodegeneration associated with TNF and β-amyloid peptides. Supporting this hypothesis, we observed that BDNF signaling in SN56 neurons leads to p38 MAPK activation. Moreover, exposure to BDNF reduced JNK activity in TNF-treated neurons, whereas inhibition of p38 MAPK with SB203580 resulted in sustained JNK activation. Interestingly, p38 MAPK has been shown to cross-inhibit JNK activity through feedback mechanisms [37]. This raises the possibility that the oscillatory p38 activation in response to BDNF observed in our study contributes to neuroprotection by dynamically buffering JNK signaling downstream of TNF and β-amyloid peptides. Although we did not directly evaluate whether BDNF signaling dominates over TNF or whether p38 MAPK activation is sustained during co-treatment, most of the biological outcomes assessed in this study were measured 24 h after exposure. This suggests that BDNF promotes long-term protective effects, likely established by signaling events triggered shortly after receptor engagement. Future studies will be necessary to determine the temporal dynamics and hierarchical interactions between BDNF- and TNF-mediated pathways that contribute to neuroprotection.

Our data, along with evidence that β-amyloid and TNF have a toxic synergistic effect on adult rat cortical neurons [52], highlights the importance of sustained BDNF signaling for maintaining neuronal homeostasis in the presence of neurotoxic insults and emphasizes p38 MAPK as a key neuroprotective pathway. It is important to note that β-amyloid toxicity engages multiple signaling pathways, including inflammasome activation, oxidative stress, ER stress, and mitochondrial dysfunction, and therefore, the focus on JNK/p38 likely underestimates the broader neuroprotective actions of BDNF. Moreover, the differential effect of β-amyloid on neurite numbers compared with TNF suggests distinct mechanisms of toxicity that warrant further investigation in future studies. Furthermore, additional signaling pathways may act synergistically to sustain neuroprotection, as both BDNF and its precursor, proBDNF, can activate alternative cascades that are essential for cellular remodeling, plasticity, and survival [53, 54]. Supporting this idea is the fact that PeproTech, on its website, indicates a molecular weight of 27 kDa for the BDNF protein used in these experiments (catalog 450-02), suggesting that the preparation likely contained a mixture of proBDNF and mature BDNF.

While apoptosis is often the first mechanism associated with neuronal cell death, emerging research indicates that neurons may also undergo pyroptosis or necroptosis [55]. Notably, axonal damage can trigger neuronal death through pyroptosis, mediated by caspase-3-dependent Gasdermin E processing and subsequent membrane pore formation [56]. Notably, while the TNF and TNFR1 signaling pathway is traditionally associated with MLK activation, which leads to necroptosis, JNK—an important downstream target of MLK—may also promote pyroptosis under certain cellular conditions [57]. Supplementary video data from our study suggest that TNF induces axonal damage before neuronal death. While we did not directly assess GSDME activation, these findings support its potential involvement, which warrants further investigation in the future.

A limitation of this study is the use of SN56 cholinergic neurons, which do not fully recapitulate the complexity of primary neuronal networks or the contribution of glial cells; therefore, future studies will extend these findings to primary neurons, co-culture systems, and in vivo models to better reflect Alzheimer’s disease pathophysiology.

In summary, our research demonstrates that BDNF counteracts JNK activation induced by harmful neuronal signals, such as inflammatory cytokines or β-amyloid peptides, in a p38 MAPK-dependent manner. This protective mechanism promotes neurite homeostasis and prevents cholinergic neuron death, underscoring a potential therapeutic avenue to counteract chronic inflammatory activation in patients with Alzheimer’s disease.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

A. R.-O. Investigation, Data analysis, Writing-original draft, Writing-Review & Editing. J. L. A. Data analysis, Review & Editing. L. P.-M. Supervision, Data analysis, Review & Editing. G. P-A Conceptualization, Data analysis, Supervision, Writing-Review & Editing. All authors revised the manuscript and approved its final version.

Data Availability

No datasets were generated or analysed during the current study.

Declarations

Competing Interests

The authors declare no competing interests.