Selective inhibition of extracellular signal-regulated kinases 1/2 blocks nerve growth factor to brain-derived neurotrophic factor signaling and suppresses the development of and reverses already established pain behavior in rats

Abstract

Brain-derived neurotrophic factor (BDNF) plays a key role in the development of pathological pain. Although it is known that nerve growth factor (NGF) induces BDNF mRNA through extracellular signal-regulated kinases (ERK), whether ERK1/2 or ERK5, two closely related members of the ERK family, mediate this signal is still unclear because classical MEK inhibitors block both pathways. We studied the involvement of ERK-signaling in NGF induction of BDNF in PC12 cells, cultured dorsal root ganglia neurons, and in rats subjected to neuropathic pain models using ERK1/2- and ERK5-specific tools. Selective activation of ERK1/2 upregulated BDNF mRNA in PC12 cells, whereas selective ERK5 activation did not. AZD6244, a potent selective inhibitor of ERK1/2 activation, blocked NGF induction of BDNF mRNA in vitro suggesting that NGF induction of BDNF is mediated by ERK1/2. siRNA experiments indicated that both ERK1 or ERK2 can signal suggesting that both pathways must be blocked to prevent NGF-induced increase in BDNF mRNA. I.p. injection of AZD6244 prevented the development of pain in rats subjected to the chronic constriction injury and reversed already established pain in the spared nerve injury model. Immunohistochemical studies showed decreased phospho-ERK1/2-immunoreactivity in dorsal root ganglia and BDNF immunoreactivity in ipsilateral spinal dorsal horn in the drug-treated rats. Our results suggest the possible use of AZD6244, already in human clinical trials as an anticancer agent, for the treatment of pathological pain.

Figure 1. NGF stimulation increases BDNF mRNA and activates both ERK5 and ERK1/2

A. Phosphorylation of ERKs after NGF stimulation. Cultured PC12 cells were stimulated with 100 ng/mL of NGF for indicated durations. The cell lysate was harvested at each time points and equal amount of proteins were loaded onto gels. Phosphorylated and total ERKs were detected by immunoblotting. B. Changes in expression of BDNF mRNA after stimulation with NGF. Cultured PC12 cells were stimulated with 100 ng/mL of NGF for indicated durations. Expression levels of BDNF mRNA were quantified by real-time RT-PCR. Each value was expressed relative to the baseline. n = 4 in each group. *p < 0.05 vs. baseline.

Figure 2. Selective activation of ERK1/2 increases BDNF mRNA

A. Phosphorylation of ERKs by constitutively-active (CA) MEK1 and MEK5. HEK293T cells were transiently transfected with CA-MEK1 or MEK5 along with wild-type ERK5 (ERK5wt). Epidermal growth factor (EGF) stimulation was done at 10 ng/mL for 5 min. Phosphorylated and total ERK were detected by immunoblotting. B. Effects of CA-MEK1 and MEK5 on expression of BDNF mRNA. PC12 cells were infected with CA-MEK1 or MEK5 lentivirus for 4 days. NGF stimulation was done at 100 ng/mL for 6 h. Expression of BDNF mRNA were quantified by real-time RT-PCR. Each value was expressed relative to the baseline. n = 6 in each group. *p < 0.05 vs. control. †p < 0.05 vs. CA-MEK5.

Figure 3. AZD6244 is a selective inhibitor of ERK1/2 and blocks NGF-induced increase in BDNF mRNA

A. Inhibitory effect of AZD6244 on phosphorylation of ERKs. PC12 cells were treated with AZD6244 at indicated concentrations for 1 h before NGF stimulation. Cells were stimulated with NGF at 100 ng/mL for 5 min. Phosphorylated form and total ERK were detected by immunoblotting. B. Densitometry was performed on pERKs. Each value was normalized to total ERK and expressed relative to baseline. n = 3 in each group. *p < 0.0001 vs. control. C. Effect of AZD6244 on expression of BDNF mRNA in PC12 cells. PC12 cells were pre-treated with 3 μM of AZD6244 or vehicle for 1 h before NGF stimulation. NGF stimulation was done at 100 ng/mL for 6 h. Expression of BDNF mRNA were quantified by real-time RT-PCR. Each value was expressed relative to baseline. n = 4 in each group. *p < 0.01 vs. control. D. Effect of AZD6244 on expression of BDNF mRNA in cultured primary DRG neurons. The neurons were pre-treated with 3 μM of AZD6244 or vehicle for 1 h before NGF stimulation. NGF stimulation was done at 100 ng/mL for 6 h. Expression of BDNF mRNA were quantified by real-time RT-PCR. Each value was expressed relative to baseline. n = 3 in each group. *p < 0.05 vs. control.

Figure 4. Either ERK1 or ERK2 can mediate the NGF-induced increase in BDNF mRNA

A. Selective knockdown of ERK1 and ERK2 confirmed by western blotting. PC12 cells were harvested 48 h after siRNA transfection and protein contents for ERK1, ERK2, ERK5, and GAPDH probed in a Western blot. Note the selective knock down of ERK1/2 by their respective siRNA without effect on ERK5. B. Effect of siRNA targeting ERK1 or ERK2 on expression of BDNF mRNA. PC12 cells were transfected with siRNA and 48 h later stimulated with 100 ng/mL of NGF for 6 h. Expression of BDNF mRNA were quantified by real-time RT-PCR. Each value was expressed relative to baseline. n = 3 in each group. *p < 0.05 vs. control.

Figure 5. Effects of AZD6244 on the phosphorylation of ERK1/2 in peripheral blood mononuclear cells (PBMC) assessed by flow cytometry

A. Osmotically-shocked partially lysed methanol fixed peripheral blood was gated by forward and side scatter to define the population corresponding to PBMCs. B–E. Representative histograms of anti-pERK1/2 antibody labeled PBMCs. Baseline phosphorylation in unstimulated PBMCs (B), after ex vivo stimulation with 200 nM PMA for 10 min (C), as in C but pre-treated with 1 or 5 μM AZD6244 before PMA stimulation (D, E). F. Summary bar graph of the percentage of pERK1/2 positive PBMCs demonstrating the inhibition of PMA-stimulated pEKR1/2 by AZD6244. n = 3 in each group. G. In vivo effect of AZD6244 on phosphorylation of ERK1/2 in PBMC. Blood was harvested before the intraperitoneal injection of AZD6244, 15 minutes or 24 hrs after a single injection and subjected to the ex vivo PMA stimulation assay. n = 2 for each group.

Figure 6. The effects of AZD6244 on painful behaviors in rats subjected to CCI and SNI

A. The thermal stimulation withdrawal latencies were measured before subjecting the rats to CCI, and at Days 3. AZD6244 (1 mg/kg, i.p.) was administered daily for 3 days after surgery until the behavioral assessment. The values were expressed as differences from the baseline. n = 10–11 in each group. **p < 0.01 vs. baseline. B. The mechanical withdrawal thresholds were measured before and at Day 3 after the CCI. n = 5 in each group. **p < 0.01 vs. baseline. C. The mechanical withdrawal thresholds were measured before and at Day 3, 4, 6, 7, 8 after the SNI. n = 6 in each group. *p < 0.05, ** p < 0.01, *** p < 0.001 vs. baseline. Arrows indicate AZD6244 administration.

Figure 7. AZD6244 inhibits phosphorylation of ERK1/2 in the ipsilateral DRGs in rats subjected to CCI

A–D. Photomicrographs showing pERK1/2-immunoreactivity in ipsilateral (A, C) and contralateral (B, D) L4 DRG from rats with CCI. The rats were treated with vehicle (A, B) and AZD6244 (C, D). Scale bar in D = 100 μm. E. Representative all-points pixel intensity histograms in control, contralateral and ipsilateral L4 DRG neurons. F. Weighted intensity count (Σ pixel × intensity) of pERK1/2- immunoreactivity above the threshold. The DRGs were harvested at 3 days after CCI. n = 3 in each group.

Figure 8. AZD6244 reduces the number of BDNF-immunoreactive DRG neurons

A. Photomicrographs of BDNF-IR neurons in L4 DRGs from DMSO or AZD6244 treated rats. Scale bar = 100 μm. B. A fluorescent image of a DRG section (Day 3 after CCI ipsilateral to the injury) (top), pixel intensity distribution plot (middle) where the vertical line indicates the threshold chosen, and the resulting processed image used for cell counting (bottom). C. Summary bar plot of BDNF-IR neuron counts (mean percentage ± S.E.M.) for the stated conditions. * P < 0.001. At least 100 neurons from random fields captured from 3 different sections where analyzed for the cell counting. All slides were immuno-stained in parallel and the images captured in an identical manner with identical exposure times. The threshold chosen based on the most BDNF-IR section (i.e. ipsilateral side in DMSO treated rats) was stringent and may have resulted in a underestimation of the total BDNF-IR neurons counted.

Figure 9. AZD6244 decreases BDNF-immunoreactivity (IR) in ipsilateral spinal dorsal horn

A, B. Photomicrographs showing BDNF-IR in L4–5 spinal dorsal horn in rats with CCI. The rats were treated with vehicle (A) and AZD6244 (B). Scale bar in B = 200 μm. C. All-points-pixel histogram of laminae I, II and a slightly larger area including background. The threshold was determined as the intersection of the two histograms. Pixels above the threshold were counted as specific BDNF-IR. D. Histograms of the weighted intensity of BDNF-IR in laminae I and II (Σ pixel × intensity). The calculated weighted intensity measures were normalized to the contralateral spinal dorsal horn. The spinal cords were harvested at 3 days after CCI. n = 3 in each group. *p < 0.05 vs. contralateral.

Figure 10. Double-immunostaining of BDNF and other selected cellular markers in spinal dorsal horn

Distribution of BDNF (green) and marker proteins (red) in the ipsilateral spinal dorsal horn in rats subjected to CCI. High-powered images were taken by confocal microscopy (B–F). The red fluorescence labels CGRP (A, B), Synaptophysin (C), NeuN (D), OX-42 (E) and GFAP (F) immunoreactivity. Scale bar in A = 200 μm and B = 20 μm.