Dramatic co-activation of WWOX/WOX1 with CREB and NF-kappaB in delayed loss of small dorsal root ganglion neurons upon sciatic nerve transection in rats

Abstract

Background: Tumor suppressor WOX1 (also named WWOX or FOR) is known to participate in neuronal apoptosis in vivo. Here, we investigated the functional role of WOX1 and transcription factors in the delayed loss of axotomized neurons in dorsal root ganglia (DRG) in rats.

Methodology/principal findings: Sciatic nerve transection in rats rapidly induced JNK1 activation and upregulation of mRNA and protein expression of WOX1 in the injured DRG neurons in 30 min. Accumulation of p-WOX1, p-JNK1, p-CREB, p-c-Jun, NF-kappaB and ATF3 in the nuclei of injured neurons took place within hours or the first week of injury. At the second month, dramatic nuclear accumulation of WOX1 with CREB (>65% neurons) and NF-kappaB (40-65%) occurred essentially in small DRG neurons, followed by apoptosis at later months. WOX1 physically interacted with CREB most strongly in the nuclei as determined by FRET analysis. Immunoelectron microscopy revealed the complex formation of p-WOX1 with p-CREB and p-c-Jun in vivo. WOX1 blocked the prosurvival CREB-, CRE-, and AP-1-mediated promoter activation in vitro. In contrast, WOX1 enhanced promoter activation governed by c-Jun, Elk-1 and NF-kappaB. WOX1 directly activated NF-kappaB-regulated promoter via its WW domains. Smad4 and p53 were not involved in the delayed loss of small DRG neurons.

Conclusions/significance: Rapid activation of JNK1 and WOX1 during the acute phase of injury is critical in determining neuronal survival or death, as both proteins functionally antagonize. In the chronic phase, concurrent activation of WOX1, CREB, and NF-kappaB occurs in small neurons just prior to apoptosis. Likely in vivo interactions are: 1) WOX1 inhibits the neuroprotective CREB, which leads to eventual neuronal death, and 2) WOX1 enhances NF-kappaB promoter activation (which turns to be proapoptotic). Evidently, WOX1 is the potential target for drug intervention in mitigating symptoms associated with neuronal injury.

Figure 1. Wwox gene expression post peripheral nerve injury in rats.

(A) As determined by in situ hybridization, there was a rapid upregulation of rat Wwox mRNA in 30 min in the ipsilateral DRG neurons upon axotomy, followed by reduction in 6 hr and then gradual increase from day 1 to month 2. In the contralateral uninjured DRG, the maximal increase in Wwox mRNA is at month 2. Little or no signal was observed for the non-axotomized sham control (top right). No signal for sense-probe reaction was observed (bottom right). White scale bar = 20 µm (see the bars at day 60th). (B) 30–50 cells were randomly selected from each section (3–5 sections used), and the intensity of each individual cell was quantified by ImagePro. Statistical analysis by one-way ANOVA is shown. Contra, contralateral. Ipsi, ipsilateral. (C) Post axotomy for 30 min, there was a significant increase in the expression of Wwox mRNA (picogram per 1 microgram total RNA), as determined by qPCR. The data were normalized to the mRNA levels of GPDH (glycerol-3-phosphate-dehydrogenase) (n = 3, p<0.01; Student's t test).

Figure 2. Axotomy induces accumulation of Tyr33-phosphorylated WOX1 (p-WOX1) in the nuclei of DRG neurons.

(A) Gradual increase in the accumulation of p-WOX1 in the nuclei of DRG neurons, both ipsilateral and contralateral, is shown at each indicated time. See the signal in brown color. Scale bar = 20 µm. (B) Statistical analysis revealed that the injured DRG neurons at the ipsilateral sides possessed significantly greater numbers of nuclear p-WOX1 than those in the contralateral sides (∼450 neurons counted in 3 separate sections). Statistical analysis by one-way ANOVA is shown. Contra, contralateral. Ipsi, ipsilateral.

Figure 3. Accumulation of p-CERB in the nuclei of axotomized DRG neurons.

(A) By immunohistochemistry, sciatic nerve transection induced phosphorylation of CREB (p-CREB) in neurons ipsilateral to injury with time, and that nuclear p-CREB is mostly present in the small neurons. (B) The bar graphs show that p-CREB is expressed at significantly higher levels in the small neurons than in the medium-large neurons (p<0.0001, n = 4, Student's t test). Approximately 500 cells were counted from 4 tissue sections. Contra, contralateral. Ipsi, ipsilateral.

Figure 4. Dramatic co-activation of WOX1, CREB and NF-κB in small neurons post axotomy for 2 months.

Accumulation of WOX1 and transcription factors in the nuclei of injured and control DRG neurons is shown in a time-course experiment. Of particular note is that dramatic co-activation of WOX1 (>65% of cells), CREB (>65%) and NF-κB (40–65%) occurred in small neurons at month 2 post-injury. Criteria for calculating nuclear localization is shown (top panel). Approximately 150 cells were counted from 4 tissue sections at 200× magnification. n.d. = not done. Contra, contralateral (non-injured side). Ipsi, ipsilateral (injured side).

Figure 5. WOX1 modulates promoter activation driven by multiple transcription factors.

(A) A schematic diagram for the Gal4-based promoter activation assay using luciferase reporter is shown. (B–D) EGFP-WOX1 was transiently overexpressed in HEK-293 fibroblasts, in the presence of a specific promoter construct and a luciferase reporter , . In control cells, EGFP and Gal4-CMV vectors were used, which showed no promoter activation. Transiently overexpressed WOX1 significantly enhanced the promoter activity regulated by c-Jun and Elk-1 (p<0.0005, n = 3, Student's t test), but blocked the activation of promoter elements responsive to transcription factors CREB, CRE, and AP-1 (p<0.005, n = 3). (E) Neuroblastoma SK-N-SH cells were transfected with an indicated construct for ECFP-tagged WOX1 or ECFP alone, in the presence or absence of an NF-κB promoter (using GFP as reporter). The wild type WOX1 significantly increased promoter activation by 4.69 fold (mean±standard deviation, n = 10), compared to ECFP alone (p<0.00001). By domain mapping, the N-terminal first and second WW domains of WOX1 dramatically enhanced the activation of promoter by 7.90 fold. However, SDR domain had no effect. A schematic structure of WOX1 is shown. (F) In control experiments, no promoter activation was observed using a negative, a positive, and an NF-κB promoter only in transfecting SK-N-SH cells.

Figure 6. The WW domain area of WOX1 binds CREB most strongly in the nucleus.

Rat DRG neurons were cultured overnight and transfected with WOX1-DsRed and CREB-EGFP. (A) In controls, EGFP did not bind DsRed, as revealed by very low binding energy (FRETc; see color scale). Scale bar = 20 µm. (B,C) Both the full-length WOX1 and the WW domain area bound CREB most strongly in the nuclei rather than in the cytoplasm. N = nucleus, C = cytoplasm. (D) Exposure of neurons to DFO (500 µM) for 2 hours to induce hypoxic conditions resulted in increased binding of WOX1 with CREB in both cytoplasm and nuclei. (E) The bar graph shows the increased binding when both proteins are in the nuclei, and DFO significantly increased the binding (average ± standard deviations, n = 7; Student's t tests for all WOX1/CREB interactions versus GFP/DsRed).

Figure 7. Axotomy-induced complex formation of p-WOX1 with p-CREB in vivo determined by immunoelectron microscopy (Immuno-EM).

Complex formation of p-WOX1 and p-CREB in the injured DRG neurons was greater in the ipsilateral side (b,d) than in the contralateral side (a,c). p-CREB, 20-nm anti-rabbit immunogold IgG particles (large); p-WOX1, 10-nm immunogold anti-goat IgG particles (small). Note the presence of p-WOX1/p-CREB complexes in the nucleus, nuclear envelope and cytosol in the ipsilateral side. Selected areas from a,b were magnified and shown in c,d.

Figure 8. CREB enhances the apoptotic function of WOX1.

(A) SK-N-SH neuroblastoma cells were transfected with a non-apoptotic dose of WOX1, in the presence of various amounts of CREB by electroporation. These cells were cultured for 48 hr. Cell cycle analysis by flow cytometry showed that CREB enhanced the apoptotic function of WOX1 (see the increased cell population at the SubG1 but reduced population at G0/G1 phases). CREB alone had no effect on apoptosis. (B) Under similar conditions, when cells were transfected with CREB and a dominant negative WOX1 (dn-WOX1), little or no apoptosis occurred. dn-WOX1 blocks WOX1 phosphorylation at Tyr33 . As a positive control of apoptosis, staurosporine (stauro) was used.