Foxo3a induces motoneuron death through the Fas pathway in cooperation with JNK

Abstract

Background: Programmed cell death of motoneurons in the developing spinal cord is thought to be regulated through the availability of target-derived neurotrophic factors. When deprived of trophic support, embryonic spinal motoneurons in vitro over-express FasL, a ligand activating a Fas-mediated death pathway. How trophic factors regulate the expression of FasL is presently unclear, but two regulators of FasL, FOXO3a (FKHRL1) and JNK have been described to play a role in other cell types. Thus, their potential function in motoneurons was investigated in this study.

Results: We show here that as a result of removal of neurotrophic factors and the consequent reduction in signalling through the PI3K/Akt pathway, Foxo3a translocates from the cytoplasm to the nucleus where it triggers cell death. Death is reduced in Fas and FasL mutant motoneurons and in the presence of JNK inhibitors indicating that a significant part of it requires activation of the Fas/FasL pathway through JNK.

Conclusions: Therefore, in motoneurons as in other cell types, FOXO transcriptional regulators provide an important link between other signalling pathways and the cell death machinery.

Figure 1

Presence of endogenous Foxo3a in motoneurons. Motoneurons from E12.5 mice spinal cords were cultured in the presence of a cocktail of neurotrophic factors (NTFs) and immunostained using an antibody against total Foxo3a, demonstrating a clearly cytoplasmic localization of endogenous Foxo3a in these conditions. (A) Immunolabelling of Foxo3a; (B) Immunolabelling combined with DAPI staining to visualize the nuclei. Scale bar: 5 microns

Figure 2

Subcellular localization of Foxo3a reflects its state of phosphorylation by Akt. (A, B): Motoneurons were electroporated with constructs encoding HA-tagged version of wildtype Foxo3a (wt Foxo3a; A) or the constitutively active triple-mutant form (TM Foxo3a; B), together with another vector encoding GFP. They were cultured for 24 hr either in the presence of NTFs, or in their absence. In the latter case, they were treated with the PI3K inhibitor LY 294002 (100 μM) for the last 90 min before fixation. HA-wt-Foxo3a and HA-TM-Foxo3a were detected staining for the HA tag. Characteristic labelled cells are illustrated for each condition, together with nuclear labelling using DAPI. Scale bar = 5microns. (C) Quantification of subcellular localization in A and B. Approximately 20 motoneurons were analysed in each condition in duplicate wells; results of two independent experiments are shown as mean ± range.

Figure 3

Non-phosphorylatable triple-mutant Foxo3a triggers motoneuron death. (A) Purified motoneurons were coelectroporated with constructs for constitutively active (ca) Akt or empty vector, together with another vector encoding GFP. The survival of transduced motoneurons was evaluated after 2 d in culture with NTFs. Akt ca further enhances the effects of NTFs on motoneuron survival. Results are mean ± SD of 2 independent experiments. (B) Using the same protocol, motoneurons were electroporated with vectors encoding either HA-wt-Foxo3a or HA-TM-Foxo3a, together with the GFP vector. They were cultured in the presence or the absence of NTFs and survival was evaluated 2 d later. Overexpression of TM-Foxo-3a leads to reduced survival in each condition. Results are mean ± SD of 12 wells from 6 independent experiments. The 100 % corresponds to the number of motoneurons electroporated with HA-wt-Foxo3a. Differences between overexpression of HA-TM-Foxo3a and HA-wt-Foxo3a were found significant using Student's t-test (p < 0.01).

Figure 4

Constitutively active Foxo3a induces motoneuron death in a Fas-dependent manner. HA-wt-Foxo3a and HA-TM-Foxo3a were overexpressed by electroporation in control motoneurons. In parallel, HA-TM-Foxo3a was expressed in mutant motoneurons with reduced capacity for Fas signalling, lpr mice being deficient for Fas and gld mice for FasL. Cells were cultured in the presence of NTFs and grown for 2 d before counting numbers of surviving transduced motoneurons. The absence of Fas signalling reduces the ability of HA-TM Foxo3a to trigger cell death. Results are means ± SD of 4–6 wells in 2–3 independent experiments. The 100 % corresponds to the number of wt or mutant motoneurons electroporated with HA-wt-Foxo3a. Differences with the TM-Foxo3a effect in lpr and gld compared to control neurons were found significant using Student's t-test (p < 0.001 in lpr and in gld).

Figure 5

The killing effect of TM Foxo3a involves the JNK pathway.Motoneurons were electroporated with HA-wt-Foxo3a and cultured for 2 d with NTFs in the presence or not of a JNK inhibitor, L-JNKl1 (1 μM). Blocking the JNK pathway in motoneuron reduces the killing effect of TM-Foxo3a by about 20 %. Results are mean ± SD of 6 wells in 3 independent experiments. The 100 % corresponds to the number of motoneurons electroporated with HA-wt-Foxo3a and untreated. Differences with the TM-Foxo3a effect in L-JNKl1-treated compared with control neurons were found significant using Student's t-test (p < 0.001).