The chemokine CXCL12 promotes survival of postmitotic neurons by regulating Rb protein

Abstract

Postmitotic neurons need to keep their cell cycle under control to survive and maintain a differentiated state. This study aims to test the hypothesis that the chemokine CXCL12 regulates neuronal survival and differentiation by promoting Rb function, as suggested by previous studies showing that CXCL12 protects neurons from apoptosis induced by Rb loss. To this end, the effect of CXCL12 on Rb expression and transcriptional activity and the role of Rb in CXCL12-induced neuronal survival were studied. CXCL12 increases Rb protein and RNA levels in rat cortical neurons. The chemokine also stimulates an exogenous Rb promoter expressed in these neurons and counteracts the inhibition of the Rb promoter induced by E2F1 overexpression. Furthermore CXCL12 stimulates Rb activity as a transcription repressor. The effects of CXCL12 are mediated by its specific receptor CXCR4, and do not require the presence of glia. Finally, shRNA studies show that Rb expression is crucial to the neuroprotective activity of CXCL12 as indicated by NMDA-neurotoxicity assays. These findings suggest that proper CXCR4 stimulation in the mature CNS can prevent impairment of the Rb-E2F pathway and support neuronal survival. This is important to maintain CNS integrity in physiological conditions and prevent neuronal injury and loss typical of many neurodegenerative and neuroinflammatory conditions.

Figure 1

CXCL12 increases nuclear Rb levels in cortical neurons via CXCR4 stimulation. Cortical cultures were treated with CXCL12 (20 nM) for the indicated time before collecting neuronal extracts for immunoblots. In (a) equal amounts of protein (30 μg per lane) were loaded for each fraction – that is cytosolic extracts (CE) and nuclear extracts (NE). NeuN and actin were used as nuclear and cytosolic markers, respectively. Densitometric analysis of three independent experiments is reported in the graph as mean±S.E.M. of band density units (*P<0.05 versus control). The immunoblots on the bottom (b) are from additional studies with neurons treated with the specific CXCR4 antagonist, AMD 3100 (100 ng/ml), or the protein synthesis inhibitor, cycloheximide (CHX; 1 μg/ml). The inhibitors were added to the culture before exposure to CXCL12 (20 nM); at the end of the treatment (5 h) total cell lysates were extracted and immunoblotted for Rb and actin (b)

Figure 2

Generation of GFP-Rb fusion protein and their expression in cortical neurons. GFP-Rb fusion vectors were generated by cloning full-length Rb coding sequence as well as its truncations in a GFP expression vector at 3′ of GFP coding sequence (a). Neurons were imaged 24 h after transfection with either the GFP-Rb 1–928 (b; upper panels), or the truncated Rb 1–602 lacking NLS (b; lower panels). A panel showing distribution of the GFP is included for comparison (large image in b). The neuronal marker, β-tubulin III (red) was used to identify the neuronal bodies and the nuclear dye Hoechst 33342 (blue) to visualize the nuclei (b). The micrograph in (c) shows a live neuron expressing full-length GFP-Rb in the nucleus imaged several days after transfection. Scale bar =25 μm

Figure 3

Effect of CXCL12 on Rb in HOS cells. HOS cells were transfected with either full-length GFP-Rb, or truncated GFP-Rb (1–602) expression vectors, treated with CXCL12 (20 nM, 3 h), fixed, and imaged. CXCL12 enhances expression of full-length GFP-Rb in the nucleus (top panel). The graph shows average pixel intensity from the nuclei of control and CXCL12-treated cells (n =11 in each group; mean±S.E.M.). GFP-Rb (1–602) remains in the cytosol of all cells analyzed, that is w/or w/o CXCL12 (lower panel). Cells were stained with phalloidin (red) and Hoechst 33342 (blue). Scale bar =25 μm

Figure 4

CXCL12 stimulates the Rb promoter in neurons. Neurons were treated with CXCL12 (20 mM, 5 h), 24 h after transfection with a luciferase construct containing the human Rb promoter (a). Data from three different experiments (each run in triplicate) are reported as mean±S.E.M. of RLU (*P<0.05 versus control). In (b), neurons were transfected either with the Rb promoter construct alone or with both the Rb promoter construct and a GFP-E2F1 expression vector and then treated with CXCL12 as reported above (mean±S.E.M.; *P<0.05 versus control; n =3). The panels in (c) report results from RT-PCR (top, n =2) and qPCR (bottom, n = 4) studies performed on control neurons and neurons treated with CXCL12 (20 nM, 3 h). The qPCR graphs show relative changes in Rb expression normalized to the housekeeping gene GAPDH in the absence or presence of exogenous E2F1 (*P<0.01 versus control)

Figure 5

CXCL12 induces the expression of a transcriptionally active Rb that regulates its prosurvival effect in neurons. Neurons were transfected with the Rb-TA-Luc plasmid and treated with CXCL12 (20 nM; 24 h post transfection) for the indicated time (a). Data are shown as mean±S.E.M. (*P<0.05 versus control; n = 3). Images in (b) are from neurons transfected with the Rb shRNA vector, psi(Rb), and a GFP expression vector, fixed after 24 h and immunostained for Rb (red); Hoechst 33342 (blue) was used to stain neuronal nuclei. Arrows point to an example of transfected neuron identified by GFP. Immunoblots from neurons transfected with psi(Rb) or a control vector carrying a scrambled sequence are also shown in (b). The western blot shows three separate psi(Rb) transfections in the same gel. The densitometric analysis from five independent experiments is reported in the graph (mean±S.E.M.; *P<0.05 versus scrambled). The graph in (c) shows survival data from cortical neurons transfected with either psi(Rb) or a control vector carrying scrambled sequence (along with a GFP expression vector as transfection marker). After 24 h post transfection, the neurons were treated with NMDA (100 μM) with or without CXCL12 (20 nM); data are expressed as mean±S.E.M. from three independent experiments containing at least three coverslips per experiment (*P<0.01 versus control; only this is shown in graph for clarity purposes). Multiple comparison analyses showed that both scrambled and psi(Rb) NMDA are different than controls (P<0.01); the control scrambled and control psi(Rb) are not different; within the scrambled group, CXCL12 and CLCL12 + NMDA are not different than control, but different than NMDA alone (P<0.01); within the psi(Rb) group the CXCL12 and CXCL12 +NMDA are different than control (P<0.05) and not different than NMDA alone; the scrambled NMDA +CXCL12 is different than the psi(Rb) NMDA + CXCL12 and the scrambled CXCL12 is different than the psi(Rb) CXCL12 (P<0.05). Scale bars =50 μm

Figure 6

Effect of NMDA and CXCL12 on Rb/pRb. Cytosolic and nuclear extracts (a, b) or total cell extracts (c) were prepared from neurons treated with NMDA (100 μM) and/or CXCL12 (20 nM) as previously described. Neurons were exposed to NMDA for 20 min and collected at the indicated intervals (0.5–5 h) post treatment. CXCL12 was added 10 min before NMDA. Equal amount of proteins (30 μg) were loaded in each lane; gels were blotted for total Rb, pRb (Ser 807/811), actin, and NeuN