Introduction

Upregulation of tumor necrosis factor α (TNF-α), a potent pro-inflammatory cytokine, is recognized to have intimate correlation with the onset of a crowd of neurological disorders, including neuropathic pain [1–3]. Accordingly, anti-TNF-α therapy is currently used to treat arthritis-induced pain in clinic. Furthermore, in the clinical trial, anti-inflammation treatment successfully relieved neuropathic pain caused by spinal cord injury [4]. Nevertheless, patients of painful arthritis receiving anti-TNF-α agents developed severe neurological adverse events, such as multiple sclerosis, optic neuritis, distal acquired demyelinating symmetric neuropathy, Guillain-Barré and Miller Fisher syndromes, which were correlated with demyelination in the central or peripheral nervous system [5–10]. The serious adverse events mentioned above suggest the crucial role of TNF-α in the nervous system in spite of the recognized neurotoxic effect caused by excessive amount of this cytokine, which prompts us to further investigate the role of this cytokine in neurological disorders involving TNF-α upregulation, and to explore the underlying mechanism.

Insulin-like growth factor-1 (IGF-1) is a cytokine which promotes neuronal survival and differentiation, facilitates axonal outgrowth and regeneration, and maintains synaptic connections and normal neuronal functions [11–15]. And during various inflammatory and neurodegenerative diseases, not only is TNF-α upregulated as a pleiotropic mediator of a diverse array of physiologic and neurologic events, but also the expression of IGF-1 is increased as a survival response [16]. Accordingly, further exploration is required to clarify the crosstalk between IGF-1 and TNF-α and the neurobiological effects triggered by interaction of IGF-1 and TNF-α in neuropathic injuries.

Activating transcription factor 3 (ATF3) belongs to the ATF/cyclic AMP responsive element binding family of transcription factors, and ATF3 gene is often described as an adaptive response gene whose activity is usually regulated by stressful stimuli. The mRNA and protein levels of ATF3 are extremely low in neurons at a normal state, but the levels are rapidly upregulated in response to nerve injury [17]. ATF3 is regarded as a unique and reliable neuronal marker of nerve injury in the nervous system [18], and is usually considered to be a neuronal damage marker in essentially all dorsal root ganglion (DRG) neurons with diverse noxious chemical stimuli [19]. Growth-associated protein 43 (GAP-43) is a neurite outgrowth and regeneration marker, upregulation of which is correlated with neuronal regenerative competence after nerve injury and is terminated when the regeneration process is complete [20, 21]. Therefore, in the present study, ATF3, as a neuronal damage marker, and GAP-43, as a neuronal growth and regeneration marker, were employed as indicators for testing the growth and regeneration status of cultured DRG neurons.

Herein, using an in vitro model of primary cultured DRG neurons, we aimed to investigate the interaction between TNF-α and IGF-1 in the peripheral nervous system, with neurite outgrowth, the expression of ATF3 and GAP-43, and the activation status of IGF-1 survival signaling pathway, the phosphatidylinositol 3-kinase (PI3K)/Akt/p70 S6 kinase (S6K) signaling pathway, being detected. And we found that higher concentration of TNF-α (10 ng/mL) exhibited toxic effect on cultured DRG neurons, which was reversed by IGF-1 supplement; while lower concentration of TNF-α (1 ng/mL) promoted the growth of cultured DRG neurons, which could be potentiated by IGF-1 administration. Moreover, the PI3K/Akt/S6K signaling pathway might be the node for TNF-α and IGF-1 crosstalk. This study proposes novel evidence for the association of IGF-1 with TNF-α on neuronal outgrowth status of primary sensory neurons, which is helpful for further interpretation of the neurobiological effects of these cytokines and the underlying mechanism in neurological disorders.

Materials and Methods

DRG Cell Culture Preparation

Newborn rats, taken from the breeding colony of Wistar rats and maintained in the Experimental Animal Center at Shandong University of China, were utilized for DRG cell culture preparation. All animals were taken good care of, and we abided by the National Institute of Health Guide for the Care and Use of Laboratory Animals (eighth edition, 2010; International Standard Book Number-13: 978-0-309-15400-0$4 http://www.nap.edu). All procedures described herein were reviewed by and had prior approval by the Ethical Committee for Animal Experimentation of the Shandong University. The whole surgery was performed under anesthesia, and all efforts were made to minimize suffering of these animals. Under aseptic conditions, the bilateral dorsal root ganglia (DRGs) were removed from each animal, placed in culture medium, and then digested with 0.25% trypsin (Sigma, St. Louis, MO) in D-Hanks solution at 37 °C for 10 min. The suspension of DRG cells was centrifuged at 1 × 103 rpm for 5 min. The supernatant was removed, and the pellet was resuspended in Dulbecco’s Modified Eagle Medium with F-12 supplement (DMEM/F-12) medium (Gibco, Grand Island, NY) by triturating with a sterile modified Pasteur’s glass pipette. Cells were then filtered using a 130 μm filter followed by counting. Dissociated DRG cells were cultured in 24-well clusters at 37 °C with 5% CO2 for 24 h, and maintained in culture medium containing cytosine arabinoside (5 μg/mL) for another 24 h to inhibit growth of non-neuronal cells. Then DRG cells were cultured in different experimental conditions for an additional 24 h before observation. DRG cells for measurement of total neurite length of each neuron were plated at a density of 5 × 104 cells/well, while neurons for double fluorescent labeling were plated at a density of 1 × 105/well. Each well contained a coverslip precoated with poly-l-lysine (0.1 mg/mL, Sigma, St. Louis, MO). DRG cells for real-time PCR and Western blot assay were plated at a density of 5 × 105 cells/mL. DRG cells were cultured in DMEM/F-12 medium with 5% fetal bovine serum (Invitrogen, Life Technologies, Carlsbad, CA) and 20 μl/mL 1× B-27 (Invitrogen, Life Technologies, Carlsbad, CA) at 37 °C with 5% CO2.

Exposure of Different Agents to DRG Cultures

The DRG neuronal cultures at 48 h of culture age were exposed to different agents. For the concentration grading experiment, cultured DRG neurons were treated with 1 ng/mL, 10 ng/mL, or 100 ng/mL of TNF-α, respectively. For higher concentration of TNF-α (10 ng/mL) effects detection, the neuronal cultures were treated with TNF-α (10 ng/mL), TNF-α (10 ng/mL) + IGF-1 (20 nmol/L), or LY294002 (10 μmol/L) 30 min before treatment with TNF-α (10 ng/mL) + IGF-1 (20 nmol/L), respectively. For lower concentration of TNF-α (1 ng/mL) effects detection, the neuronal cultures were treated with TNF-α (1 ng/mL), TNF-α (1 ng/mL) + IGF-1 (20 nmol/L), or LY294002 (10 μmol/L) 30 min before treatment with TNF-α (1 ng/mL) + IGF-1 (20 nmol/L), respectively. The DRG neurons were continuously cultured in medium as a control.

Total Neurite Length Measurement

Total neurite length of each neuron in different groups was measured by fluorescent labeling of βIII-tubulin. The cells on coverslips were quickly rinsed once in 0.1 mol/L phosphate-buffered saline (PBS) (pH 7.4) to remove medium. The cells were fixed in 4% paraformaldehyde, pH 7.4, for 20 min at 4 °C. After being washed in 0.1 mol/L PBS for 3 times, the cells were blocked with 2% normal goat serum in 0.3% Triton PBS to block non-specific sites and permeabilize the cells. The samples were incubated with mouse monoclonal anti-βIII-tubulin (1:1000, Abcam, Cambridge, MA) overnight at 4 °C. After being washed in 0.1 mol/L PBS for three times, the cells were incubated with goat anti-mouse conjugated to Cy2 (1:100, Abcam, Cambridge, MA) for 45 min in dark. After being washed in 0.1 mol/L PBS, the coverslips were placed on glass slides immediately with anti-fade mounting medium (Santa Cruz Biotechnology, Santa Cruz, CA) and stored at 4 °C prior to observation with a fluorescent microscope (Olympus).

Real Time-PCR Analysis for mRNA Levels of ATF3 and GAP-43

After treatment with different agents for 24 h, the mRNA levels of ATF3 and GAP-43 in cultured DRG neurons were analyzed by real-time PCR. The expression of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA was also determined as an internal control. Total RNA from DRG neuronal cultures was isolated by TRIzol (TakaRa Biotechnology, Dalian, China). cDNA was synthesized using cDNA synthesis kit (Thermo Scientific Molecular Biology, Lithuania, EU) according to the manufacturer’s instructions.

The synthetic oligonucleotide primer sequences for ATF3, GAP-43, and GAPDH were as follows: ATF3 5′-CCT GCA GAA GGA GTC AGA GAA-3′ (coding sense) and 5′-CGT TCT GAG CCC GGA CGA TA-3′ (coding antisense). GAP-43 5′-AAG AAG GAG GGA GAT GGC TCT-3′ (coding sense) and 5′-GAG GAC GGC GAG TTA TCA GTG-3′ (coding antisense). GAPDH 5′-GGC ACA GTC AAG GCT GAG AAT G-3′ (coding sense) and 5′-ATG GTG GTG AAG ACG CCA GTA-3′ (coding antisense).

Real-time PCR was performed using SYBR Green dye (Thermo Scientific Molecular Biology, Lithuania, EU) according to the manufacturer’s instructions. PCR was performed at 50 °C for 2 min, 94 °C for 15 min, followed by 40 cycles at 94 °C for 15 s, 58 °C for 30 s, and 72 °C for 30 s.

A comparative cycle of threshold (Ct) method was used and the relative transcript amount of the target gene was normalized to that of GAPDH using the 2− ΔΔCt method. The final results of real-time PCR were expressed as the ratio of mRNA of the control.

Western Blot Assay for Protein Levels of ATF3, GAP-43, pAkt, and pS6K

After treatment with different agents for 24 h, the protein levels of ATF3 and GAP-43 in cultured DRG neurons were detected by Western blot assay. To detect the activation of PI3K/Akt/S6K signaling pathway, after incubation of cultured DRG neurons in different conditions for 30 min, the protein levels of pAkt, Akt, S6K, and pS6K in cultured DRG neurons were detected by Western blot assay. Freshly cultured DRG neurons were homogenized in 10 mmol/L Tris homogenization buffer (pH 7.4) with protease inhibitors (Amresco, Solon, OH). The samples were centrifuged at 10,000 g for 20 min and the supernatants were collected for Western blot assay. After determination of the protein concentrations of the supernatants (BCA method, standard: BSA), 50 μg of protein of each sample was loaded onto the 12% SDS gel, separated by electrophoresis, and transferred to a nitrocellulose (NC) membrane. The membranes were blocked in blocking buffer (5% nonfat milk) for 2 h at room temperature, and then were incubated with mouse anti-ATF3 monoclonal IgG (1:1000, Abcam, Cambridge, MA), rabbit anti-GAP-43 polyclonal IgG (1:1000, Abcam, Cambridge, MA), mouse anti-β-actin monoclonal IgG (1:1000, Santa Cruz Biotechnology, Santa Cruz, CA), rabbit anti-pAkt monoclonal IgG (1:1000, Cell Signaling Technology, Danvers, MA), rabbit anti-Akt monoclonal IgG (1:500, Cell Signaling Technology, Danvers, MA), rabbit anti-pS6K monoclonal IgG (1:1000, Cell Signaling Technology, Danvers, MA), or rabbit anti-S6K monoclonal IgG (1:1000, Cell Signaling Technology, Danvers, MA) overnight at 4 °C. After being washed for three times, the membranes were incubated with goat anti-rabbit IgG-HRP (1:6000, Beijing Sequoia Jinqiao Biological Technology Co., Ltd., Beijing, China) or goat anti-mouse IgG-HRP (1:3000, Beijing dingguochangsheng Biotechnology Co., Ltd., Beijing, China). The immunoreactive bands were visualized with an ECL Western blotting detection kit (Millipore Corporation, Billerica, MA) and analyzed quantitatively with ImagJ 1.39u image analysis software. The protein levels of ATF3 and GAP-43 were expressed as the ratio of β-actin levels. The levels of pAkt were expressed as the ratio of the total Akt levels. And the levels of pS6K were expressed as the ratio of the total S6K levels. The final results of the Western blot assay were expressed as a ratio of the expression of the protein of interest in the experiment groups to that in the control group.

Double Fluorescent Labeling of Microtubule-Associated Protein 2 (MAP2) with ATF3 or GAP-43, or GAP-43 with ATF3

After treatment with different agents for 24 h, freshly cultured DRG neurons were processed for double immunofluorescent labeling of MAP2 with ATF3 or GAP-43, or ATF3 with GAP-43. The cells on coverslips were quickly rinsed once in 0.1 mol/L PBS to remove medium. The cells were fixed in 4% paraformaldehyde, pH 7.4, for 20 min at 4 °C. After being washed in 0.1 mol/L PBS for three times, the cells were blocked by 2% normal goat serum in 0.3% Triton PBS to block non-specific sites and permeabilize cells. The samples were incubated by mouse anti-ATF3 monoclonal IgG (1:500, Abcam, Cambridge, MA) or rabbit polyclonal anti-GAP-43 (1:500, Abcam, Hong Kong, China) with chicken polyclonal anti-MAP2 (1:400, Abcam, Cambridge, MA), or mouse anti-ATF3 monoclonal IgG (1:500, Abcam, Cambridge, MA) with rabbit polyclonal anti-GAP-43 (1:500, Abcam, Hong Kong, China) overnight at 4 °C, respectively. After being washed in 0.1 mol/L PBS for three times, the samples were incubated by goat anti-mouse conjugated to Cy3 (1:500, Abcam, Cambridge, MA) or goat anti-rabbit conjugated to Cy3 (1:500, Abcam, Cambridge, MA) with goat anti-chicken conjugated to Cy2 (1:100, Abcam, Cambridge, MA), or goat anti-mouse conjugated to Cy2 (1:500, Abcam, Cambridge, MA) with goat anti-rabbit conjugated to Cy3 (1:500, Abcam, Cambridge, MA), respectively, for 45 min in dark. After being washed in 0.1 mol/L PBS, the coverslips were placed on glass slides immediately with anti-fade mounting medium (Santa Cruz Biotechnology, Santa Cruz, CA) and stored at 4 °C until observed with a fluorescent microscope.

Quantitative Analysis of the Percentage of ATF3- or GAP-43-Expressing Neurons, or GAP-43-Expressing Neurons Coexpressing ATF3

ATF3-immunoreactive (IR) neurons, GAP-43-IR neurons, or GAP-43-IR neurons coexpressing ATF3 were observed under a fluorescent microscope with 20× objective lens. ATF3-IR neurons, GAP-43-IR neurons, or GAP-43-IR neurons coexpressing ATF3 in five visual fields in the central part of each coverslip were counted as the positive neurons in each sample. One visual field was in the center of the coverslip. The other four visual fields were just adjacent to, but did not overlap with, the central visual field at the upper, lower, left, and right side, respectively. To quantify the percentage of ATF3-IR or GAP-43-IR neurons, MAP2-IR neurons in the same visual field were counted as the total neurons in each sample. To obtain the percentage of GAP-43-IR neurons coexpressing ATF3, GAP-43-IR neurons in the same visual field were counted as the total neurons in each sample.

Statistical Analysis

All data were reported as mean ± SD. All the data were processed for verifying normality test for Variable. The data with abnormal distribution were analyzed with non-parametric test. If the data were of normal distribution, statistical analysis was evaluated with one way analysis of variance followed by Student–Newman–Keuls test (homogeneity of variance) or Dunnett’s T3 test (heterogeneity of variance) for significance to compare the differences among various groups. Statistical analysis was executed using SPSS (version 17.0) software with P value <0.05 employed to delineate significance for analysis of all results.

Results

TNF-α Triggers Diverse Effects on Neurite Outgrowth at Different Concentrations

Total neurite length of each neuron directly reflects the neurite elongation in DRG neuronal cultures. DRG neurons at 48 h post-culture were exposed to different concentrations of TNF-α for an additional 24 h, and then the total neurite length of each neuron was measured by fluorescent labeling of βIII-tubulin. Results showed that with higher concentrations of TNF-α (10 and 100 ng/mL) challenge, the total neurite length of each neuron decreased significantly in a dose dependent manner as compared with the control group. Interestingly, in contrast, lower concentration of TNF-α (1 ng/mL) exposure increased the total neurite length of each neuron (Fig. 1a, b). These results indicate that TNF-α at different concentrations could elicit diverse biological effects, with higher concentrations of TNF-α inhibiting neurite outgrowth and relatively lower concentration of TNF-α promoting neurite sprouting. According to the data above, we conducted the following experiments with 10 or 1 ng/mL of TNF-α to investigate the interaction between IGF-1 and TNF-α.

Fig. 1
figure 1

TNF-α elicited diverse effects on total neurite length of each DRG neuron at different concentrations. Panel a Neurite outgrowth with βIII-tubulin immunofluorescent staining; Panel b Quantification of total neurite length of each neuron. Scale bar 50 μm. Bar graphs with error bars represent mean ± SD (n = 5). **P < 0.01, ***P < 0.001

Full size image

The Effects of Higher (10 ng/mL) or Lower (1 ng/mL) Concentration of TNF-α with or without IGF-1 Administration on Activation of PI3K/Akt/S6K Signaling Pathway

In the nervous system, TNF-α not only directly foments signals of death, but also indirectly promotes neurodegeneration or triggers the death of neurons by impairing essential components of the IGF-1 survival response [22]. Here, we investigated the activation status of PI3K/Akt/S6K signaling pathway in primary cultured DRG neurons exposed to different agents for 30 min with western blotting assay. Results showed that higher concentration of TNF-α (10 ng/mL) treatment transiently downregulated the levels of pAkt and pS6K, which was reversed by IGF-1 administration. The restoration of pAkt and pS6K levels was blocked by PI3K inhibitor LY294002 (Fig. 2a–c). Lower concentration of TNF-α (1 ng/mL) treatment had no significant effect on the level of pAkt or pS6K, while IGF-1 treatment in the presence of lower concentration of TNF-α (1 ng/mL) promoted phosphorylation of Akt and S6K. The effects of IGF-1 were blocked by PI3K inhibitor LY294002 (Fig. 2d–f). These results suggest that the activation status of PI3K/Akt/S6K signaling pathway might be the node for IGF-1 and TNF-α interaction, which helps to further clarify the underlying mechanism of the functions of the two cytokines.

Fig. 2
figure 2

Activation of PI3K/Akt/S6K signaling pathway analysis. Panel a Immunoreactive bands in different experimental conditions with higher concentration of TNF-α (10 ng/mL); Panel b Quantification of pAkt protein levels in different experimental conditions with higher concentration of TNF-α (10 ng/mL); Panel c Quantification of pS6K protein levels in different experimental conditions with higher concentration of TNF-α (10 ng/mL); Panel d Immunoreactive bands in different experimental conditions with lower concentration of TNF-α (1 ng/mL); Panel e Quantification of pAkt protein levels in different experimental conditions with lower concentration of TNF-α (1 ng/mL); Panel f Quantification of pS6K protein levels in different experimental conditions with lower concentration of TNF-α (1 ng/mL). Bar graphs with error bars represent mean ± SD (n = 5). *P < 0.05, **P < 0.01, ***P < 0.001

Full size image

IGF-1 Reverses Neurite Retraction Induced by Higher Concentration of TNF-α (10 ng/mL), while Promotes Neurite Elongation Triggered by Lower Concentration of TNF-α (1 ng/mL)

DRG neurons at 48 h post-culture were exposed to different agents for an additional 24 h and then the total neurite length of each neuron was measured by fluorescent labeling of βIII-tubulin. Higher concentration of TNF-α (10 ng/mL) exposure decreased the total neurite length of each neuron. IGF-1 incubation prevented neurite retraction induced by higher concentration of TNF-α (10 ng/mL), which was blocked by pretreatment with PI3K inhibitor LY294002 (Fig. 3a, b). In contrast, lower concentration of TNF-α (1 ng/mL) administration increased the total neurite length of each neuron, which was potentiated by IGF-1 supplement. The growth-promoting effect of IGF-1 was inhibited by pretreatment with PI3K inhibitor LY294002 (Fig. 3c, d). These data imply the beneficial effects of IGF-1 in the presence of TNF-α on cultured DRG neurons that IGF-1 rescues neurons from growth inhibition induced by excessive amount of TNF-α, and that when there is relatively small amount of TNF-α, though still at a pathological level, IGF-1 assists the growth-promoting effect of this pro-inflammatory cytokine.

Fig. 3
figure 3figure 3

Neurite outgrowth analysis with βIII-tubulin immunofluorescent staining. Panel a Neurite outgrowth with higher concentration of TNF-α (10 ng/mL); Panel b Quantification of total neurite length of each neuron with higher concentration of TNF-α (10 ng/mL); Panel c Neurite outgrowth with lower concentration of TNF-α (1 ng/mL); Panel d Quantification of total neurite length of each neuron with lower concentration of TNF-α (1 ng/mL). Scale bar 50 μm. Bar graphs with error bars represent mean ± SD (n = 5). ***P < 0.001

Full size image

IGF-1 Interacts with Higher (10 ng/mL) or Lower (1 ng/mL) Concentration of TNF-α on ATF3 and GAP-43 mRNA Expression in an Antagonistic or Synergetic Pattern

ATF3 is considered to be a neuronal injury marker of DRG neurons, while GAP-43 is regarded as a molecule responsible for neurite outgrowth and regeneration. We examined alterations of ATF3 and GAP-43 mRNA expressions, which reflected the influence of TNF-α and IGF-1 on ATF3 and GAP-43 expressions at transcriptional level. Results showed that higher concentration of TNF-α (10 ng/mL) exposure significantly increased ATF3 mRNA level and decreased GAP-43 mRNA level in DRG neurons. IGF-1 blocked the effects of higher concentration of TNF-α (10 ng/mL). The effects of IGF-1 could be inhibited by LY294002 (Fig. 4a, b). Interestingly, lower concentration of TNF-α (1 ng/mL) challenge promoted the mRNA expression of GAP-43 and suppressed the mRNA expression of ATF3, which was potentiated by IGF-1 administration. PI3K inhibitor LY294002 supplement blocked the effects of IGF-1 (Fig. 4c, d). These results suggest that IGF-1 could interact with TNF-α by modulating ATF3 and GAP-43 expression at transcriptional level through regulating PI3K/Akt/S6K signaling pathway.

Fig. 4
figure 4

Real-time PCR analysis of ATF3 mRNA and GAP-43 mRNA expressions in DRG neurons at different experimental conditions. Panel a ATF3 mRNA levels with higher concentration of TNF-α (10 ng/mL); Panel b GAP-43 mRNA levels with higher concentration of TNF-α (10 ng/mL); Panel c ATF3 mRNA levels with lower concentration of TNF-α (1 ng/mL); Panel d GAP-43 mRNA levels with lower concentration of TNF-α (1 ng/mL). Bar graphs with error bars represent mean ± SD (n = 5). *P < 0.05, **P < 0.01, ***P < 0.001

Full size image

IGF-1 Interacts with Higher (10 ng/mL) or Lower (1 ng/mL) Concentration of TNF-α on ATF3 and GAP-43 Protein Expression in an Antagonistic or Synergetic Pattern

Herein, ATF3 and GAP-43 protein expressions in primary cultured DRG neurons, which were exposed to different agents, were examined with Western blot assay to detect the influence of TNF-α and IGF-1 on ATF3 and GAP-43 expressions at translational level. ATF3 protein expression was induced and GAP-43 protein expression was reduced by higher concentration of TNF-α (10 ng/mL) challenge. IGF-1 treatment reversed the actions of higher concentration of TNF-α (10 ng/mL) insult, and the effects of IGF-1 in the presence of higher concentration of TNF-α (10 ng/mL) were suppressed by PI3K inhibitor LY294002 (Fig. 5a–c). In contrast, lower concentration of TNF-α (1 ng/mL) challenge reduced ATF3 expression and induced GAP-43 expression at protein level. IGF-1 treatment potentiated the effects of lower concentration of TNF-α (1 ng/mL), which was suppressed by PI3K inhibitor LY294002 (Fig. 5d–f). The alterations of ATF3 expression were counter to, while the alterations of GAP-43 expression were parallel to that of total neurite length of each neuron, indicating that TNF-α and IGF-1 might influence neurite elongation through regulating ATF3 and GAP-43 expression.

Fig. 5
figure 5

Western blot assay of ATF3 and GAP-43 protein expressions in DRG neurons at different experimental conditions. Panel a ATF3 and GAP-43 protein levels with higher concentration of TNF-α (10 ng/mL); Panel b Quantification of ATF3 protein levels with higher concentration of TNF-α (10 ng/mL); Panel c Quantification of GAP-43 protein levels with higher concentration of TNF-α (10 ng/mL); Panel d ATF3 and GAP-43 protein levels with lower concentration of TNF-α (1 ng/mL); Panel e Quantification of ATF3 protein levels with lower concentration of TNF-α (1 ng/mL); Panel f Quantification of GAP-43 protein levels with lower concentration of TNF-α (1 ng/mL). Bar graphs with error bars represent mean ± SD (n = 5). **P < 0.01, ***P < 0.001

Full size image

IGF-1 Counteracts the Effect of Higher Concentration of TNF-α (10 ng/mL), while Potentiates the Effect of Lower Concentration of TNF-α (1 ng/mL) on ATF3 and GAP-43 Expression In Situ

In the previous sections, the results unveiled the effects of IGF-1 on ATF3 and GAP-43 mRNA and protein levels in the presence of higher (10 ng/mL) or lower (1 ng/mL) concentration of TNF-α in cultured DRG neurons via regulation of PI3K/Akt/S6K signaling pathway, which prompted us to further investigate the alterations of the expressions of the two molecules in situ by double fluorescent labeling of MAP2 with ATF3 or GAP-43. Immunofluorescent labeling analysis revealed that higher concentration of TNF-α (10 ng/mL) exposure significantly increased the percentage of ATF3-IR neurons and decreased the percentage of GAP-43-IR neurons. IGF-1 inhibited the effects of higher concentration of TNF-α (10 ng/mL) on the alterations of the proportions of ATF3-IR and GAP-43-IR neurons, which was blocked by PI3K inhibitor LY294002 (Fig. 6a–d). The percentage of ATF3-IR neurons was decreased by lower concentration of TNF-α (1 ng/mL) and the percentage of GAP-43-IR neurons was increase. IGF-1 assisted lower concentration of TNF-α (1 ng/mL) in modulating ATF3 and GAP-43 expressions in DRG neurons, and the effect was partially eliminated by pretreatment with PI3K inhibitor LY294002 (Fig. 6e–h). The expressions of ATF3 and GAP-43 in situ were consistent with the expressions of the two molecules at mRNA and protein levels, further confirming that the alterations of ATF3 and GAP-43 expressions play a role in the crosstalk between IGF-1 and different concentrations of TNF-α.

Fig. 6
figure 6figure 6figure 6figure 6

Double fluorescent labeling of MAP2 with ATF3 or GAP-43 in DRG neurons. Panel a ATF3 expression in situ with higher concentration of TNF-α (10 ng/mL); Panel b Quantification of the percentage of ATF3-IR neurons with higher concentration of TNF-α (10 ng/mL); Panel c GAP-43 expression in situ with higher concentration of TNF-α (10 ng/mL); Panel d Quantification of the percentage of GAP-43-IR neurons with higher concentration of TNF-α (10 ng/mL); Panel e ATF3 expression in situ with lower concentration of TNF-α (1 ng/mL); Panel f Quantification of the percentage of ATF3-IR neurons with lower concentration of TNF-α (1 ng/mL); Panel g GAP-43 expression in situ with lower concentration of TNF-α (1 ng/mL); Panel h Quantification of the percentage of GAP-43-IR neurons with lower concentration of TNF-α (1 ng/mL). Scale bar 50 μm. Bar graphs with error bars represent mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001

Full size image

IGF-1 Antagonizes Higher Concentration of TNF-α (10 ng/mL) while Potentiates Lower Concentration of TNF-α (1 ng/mL) on Modulating ATF3 and GAP-43 Colocalization

The latest research showed that upregulation of ATF3 indicated the responses of injured axons regeneration [23]. Moreover, it has been reported that the nerve injury marker ATF3 could coexpress with the nerve regeneration marker GAP-43 [24]. In order to detect the effects of TNF-α and IGF-1 on ATF3 and GAP-43 colocalization and further explore the roles of the two molecules in the interaction between TNF-α and IGF-1, we conducted double fluorescent labeling of GAP-43 and ATF3. Results showed that higher concentration of TNF-α (10 ng/mL) insult increased colocalization of GAP-43 and ATF3, which was reversed by IGF-1 treatment. The effect of IGF-1 was blocked by pretreatment with PI3K inhibitor LY294002 (Fig. 7a, b). In contrast with the effect of higher concentration of TNF-α (10 ng/mL) and IGF-1, lower concentration of TNF-α (1 ng/mL) downregulated the colocalization level of GAP43 and ATF3. IGF-1 treatment further decreased colocalization of the two molecules, which was inhibited by PI3K inhibitor LY294002 (Fig. 7c, d). These results indicate that the sharp increase of ATF3 expression in injured neurons reflects the intrinsic regenerative capacity. Besides, ATF3 elevation in injured neurons might help to restore the suppressed expression of GAP-43 and permit the regeneration of these neurons. On the other hand, ATF3 plays only a minor role in the elevation of GAP-43 expression in growing neurons.

Fig. 7
figure 7figure 7

Double fluorescent labeling of GAP-43 with ATF3 in DRG neurons. Panel a Colocalization of GAP-43 and ATF3 in DRG neurons with higher concentration of TNF-α (10 ng/mL); Panel b Quantification of the percentage of GAP-43-IR neurons coexpressing ATF3 with higher concentration of TNF-α (10 ng/mL); Panel c Colocalization of GAP-43 and ATF3 in DRG neurons with lower concentration of TNF-α (10 ng/mL); Panel d Quantification of the percentage of GAP-43-IR neurons coexpressing ATF3 with lower concentration of TNF-α (1 ng/mL). Scale bar 50 μm. Bar graphs with error bars represent mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001

Full size image

Discussion

TNF-α is a potent pro-inflammatory cytokine, which exerts pleiotropic effects on various cell types regulating cell apoptosis, proliferation, differentiation, activation, and etc [25]. In the nervous system, it is most commonly proposed that TNF-α plays a deleterious role, since upregulation of TNF-α is closely related to the pathogenesis and progression of numerous neurological disorders, such as neuropathic pain, Alzheimer’s disease [26, 27], Parkinson’s Disease [26, 28], and etc. Among the mentioned disorders, neuropathic pain could be elicited by elevation of TNF-α in DRGs, which leads to functional abnormality or lesion of DRG neurons [1, 2, 29]. Nevertheless, accumulating evidences indicate the neuroprotective, function-maintaining and growth-promoting potential of TNF-α. It has been reported that TNF-α preconditioning protected neurons against amyloid beta1-42 insults [30], and that TNF-α was essential to maintaining synaptic plasticity and could further influence learning and memory formation [31]. Moreover, TNF-α application promoted neurite elongation in cultured adult sensory neurons [32]. The existing reports on the role of TNF-α in the nervous system are seemingly contradictory. However, in the present study, we found that TNF-α at different concentrations could elicit diverse biological effects in DRG neurons. Results showed that higher concentrations of TNF-α (10 and 100 ng/mL) induced axonal retraction, while relatively lower concentration of TNF-α (1 ng/mL) triggered axonal elongation. Since we applied TNF-α in this experiment at pathological concentrations, this research indicates the possible beneficial influence of TNF-α in neurological disorders at relatively low levels that slightly upregulated TNF-α might trigger a growth-promoting effect and promote the self-recovery of neurons from intrinsic or ectogenic insults.

Recent studies demonstrated that IGF-1 exhibited powerful neuroprotective effects in the nervous system [33–35]. In this study, we focused on the interaction of IGF-1with either higher concentration of TNF-α (10 ng/mL) or lower concentration of TNF-α (1 ng/mL), and intended to figure out the mechanism responsible for the interaction. And we found that higher concentration of TNF-α (10 ng/mL) challenge transiently suppressed the activation of Akt and its downstream molecule S6K, which promotes the synthesis of numerous growth-associated proteins when activated. IGF-1 treatment in the presence of TNF-α restored the levels of pAkt and pS6K, which was blocked by PI3K inhibitor LY294002. Interestingly, lower concentration of TNF-α (1 ng/mL) had no significant effect on the activation status of Akt or S6K, while IGF-1 administration triggered the phosphorylation of these two kinases. These data imply that the node of the crosstalk between TNF-α and IGF-1 might be the activation status of prosurvival PI3K/Akt/S6K signaling pathway. However, the mechanism underlying the growth-promoting effect of lower concentration of TNF-α (1 ng/mL) on DRG neurons is currently unknown, and we speculate that there might be two possible reasons responsible for the beneficial effect of lower concentration of TNF-α (1 ng/mL). Firstly, the activation of extracellular signal-regulated kinase 1/2 (ERK1/2) triggered by TNF-α stimulation might play a role in the beneficial facet of lower concentration of TNF-α (1 ng/mL). Secondly, since it has been recently reported that TNF receptors (TNFRs) exhibited antithetic effects in the central nervous system that most of the pro-inflammatory actions of TNF-α were mediated by TNFR1, while TNFR2 signaling possessed neuroprotective characteristics [36–38], the diverse effects of higher (10 ng/mL) and lower (1 ng/mL) concentrations of TNF-α might attribute to the different activation status of these two receptors when stimulated by different concentrations of TNF-α in DRG neurons.

Total neurite length of each neuron is considered to reflect the growth status of cultured neurons. Here, results illustrated that higher concentration of TNF-α (10 ng/mL) challenge led to significantly reduced total neurite length of each neuron, which was reversed in the presence of IGF-1. Blockage of PI3K/Akt/S6K signaling pathway with LY294002 inhibited the growth-promoting effect of IGF-1. In contrast, neurite sprouting appeared to be facilitated by lower concentration of TNF-α (1 ng/mL) administration, and the growth-promoting effect was potentiated by IGF-1, which was inhibited by LY294002. On account that the level of GAP-43, a molecule which participates in neurite outgrowth during development [39] and axon regeneration after injury [40], could also mirror the status of neuronal growth, we further detected GAP-43 mRNA level, GAP-43 protein level and GAP-43 expression in situ in DRG neurons. These data were in line with the detection of total neurite length of each neuron. GAP-43 expression declined when treated with higher concentration of TNF-α (10 ng/mL), and IGF-1 treatment in the presence of higher concentration of TNF-α (10 ng/mL) restituted GAP-43 expression, which was blocked by LY294002. In contrast, treating cultured neurons with lower concentration of TNF-α (1 ng/mL) elevated GAP-43 mRNA level, GAP-43 protein level, and GAP-43 expression in situ in DRG neurons. The elevations were further potentiated in the presence of IGF-1, which was inhibited by pretreatment with PI3K inhibitor LY294002. These results intimate the neuroprotective effect of IGF-1 treatment on higher concentration TNF-α-induced neurotoxicity and the potentiating effect of IGF-1 administration on lower concentration TNF-α-facilitated neuronal growth, and further unveil the crucial role of PI3K/Akt/S6K signaling pathway in the interaction of TNF-α and IGF-1.

ATF3 is a basic leucine zipper transcription factor induced rapidly by various stressful stimuli [41]. Commonly, ATF3 is regarded as a neuronal injury maker. However, ATF3 has been recently reported to play a crucial role in neuronal survival, axonal elongation, and axonal regeneration [42–44]. In this study, the elevated expression of ATF3 induced by higher concentration of TNF-α (10 ng/mL) was reduced by IGF-1 treatment along with restored neurite outgrowth, while the downregulated expression of ATF3 decreased by lower concentration of TNF-α (1 ng/mL) further declined with IGF-1 supplement along with further facilitated neurite sprouting. These results indicate that ATF3 plays a minimal role in relatively healthy DRG neurons, which is consistent to its low expression in intact DRG neurons, whereas upregulation of ATF3 expression might be an ultimate intrinsic neuroprotective action promoting neuronal recovery from environmental or endogenous insults. If the toxicity overweighs the rescue, the neuron finally dies; if the rescue outweighs the toxicity, the neuron recuperates, and the level of ATF3 decreases since it has fulfilled its duty. Moreover, in this study, pretreatment with LY294002 reversed the effect of IGF-1 on ATF3 expression in the presence of TNF-α, which suggests that activation of PI3K/Akt/S6K signaling pathway might inhibit the expression of ATF3.

We then conducted double fluorescence labeling of GAP-43 and ATF3, and found that neurons coexpressing GAP-43 and ATF3, which were regarded as injured neurons possessing regeneration potential, increased significantly challenged by higher concentration of TNF-α (10 ng/mL). As a consequence, the majority of GAP-43-IR neurons in the presence of higher concentration of TNF-α (10 ng/mL) were ATF3-IR neurons. IGF-1 treatment reduced the elevated percentage of GAP-43-IR neurons coexpressing ATF3. In contrast, lower concentration of TNF-α (1 ng/mL) administration decreased GAP-43 and ATF3 colocalization level, and IGF-1 treatment further assisted the decrease. Pretreatment with PI3K inhibitor LY294002 blocked the effect of IGF-1. These data further confirm that though serving as a neuronal damage marker, ATF3 plays a role in DRG neurons as an ultimate intrinsic neuroprotective molecule.

Taken together, results of this study showed that IGF-1, through restoration or activation of prosurvival PI3K/Akt/S6K signaling pathway and thereby modulating GAP-43 and ATF3 expression, counteracted higher concentration TNF-α-(10 ng/mL)-induced neurotoxicity, or potentiated lower concentration TNF-α-(1 ng/mL)-facilitated neuronal growth, respectively. These data help to unveil the association of IGF-1 with TNF-α and the biological effects produced by interaction of IGF-1 and TNF-α on neuronal outgrowth status of primary sensory neurons, which may provide new information for interpretation of the molecular mechanisms underlying neurological disorders.