Analysis of the fibroblast growth factor system reveals alterations in a mouse model of spinal muscular atrophy

Abstract

The monogenetic disease Spinal Muscular Atrophy (SMA) is characterized by a progressive loss of motoneurons leading to muscle weakness and atrophy due to severe reduction of the Survival of Motoneuron (SMN) protein. Several models of SMA show deficits in neurite outgrowth and maintenance of neuromuscular junction (NMJ) structure. Survival of motoneurons, axonal outgrowth and formation of NMJ is controlled by neurotrophic factors such as the Fibroblast Growth Factor (FGF) system. Besides their classical role as extracellular ligands, some FGFs exert also intracellular functions controlling neuronal differentiation. We have previously shown that intracellular FGF-2 binds to SMN and regulates the number of a subtype of nuclear bodies which are reduced in SMA patients. In the light of these findings, we systematically analyzed the FGF-system comprising five canonical receptors and 22 ligands in a severe mouse model of SMA. In this study, we demonstrate widespread alterations of the FGF-system in both muscle and spinal cord. Importantly, FGF-receptor 1 is upregulated in spinal cord at a pre-symptomatic stage as well as in a mouse motoneuron-like cell-line NSC34 based model of SMA. Consistent with that, phosphorylations of FGFR-downstream targets Akt and ERK are increased. Moreover, ERK hyper-phosphorylation is functionally linked to FGFR-1 as revealed by receptor inhibition experiments. Our study shows that the FGF system is dysregulated at an early stage in SMA and may contribute to the SMA pathogenesis.

Figure 1. Expression and regulation of FGFs in muscle and C2C12-cells.

(A) Transcript abundances relative to GAPDH in control muscle and scrambled siRNA transfected C2C12-cells. FGF-mRNA levels were measured by qRT-PCR in pooled samples of P5 control mice tissues and pooled samples of scrambled siRNA transfected cells relative to GAPDH as internal control. Transcript abundances were calculated from ΔC_T-values. Unpaired t-tests of technical repetitions of measurements in cell culture compared to tissue abundances were performed (n = 2; n.s. = non significant, *p<0.05, **p<0.01, ***p<0.001). Bars represent means with standard deviations (SD). (B) Fold-change of FGF transcript levels in SMA-mice muscle and C2C12 cells after SMN-knockdown. FGF transcript concentration was measured by qRT-PCR in SMA-mice muscle and SMN-siRNA treated C2C12-cells. Transcript levels of SMA-mice and control animals were measured at P1 and P5. C2C12-cells were either transfected with SMN- or control scrambled-siRNA in 4 independent experiments (n = 4) with three replicates in each group. The knockdown for each experiment was monitored by western-blot analysis (Fig. S1). The fold-changes were calculated against each corresponding control group. Fold-changes of SMA-mice spinal cords were calculated against transcript levels of control mice of the same age. Results of SMN siRNA treated cells were compared to scrambled RNA transfected cells of the same experiment. SMA-mice transcript levels were tested against control animals by a Mann-Whitney test (n≥5, * p<0.05, ** p<0.01). mRNA-levels of SMN siRNA transfected cells were compared to scrambled siRNA transfected cells by a repeated measurements two way ANOVA (n = 4). Bars for fold changes represent means with standard error of mean (SEM).

Figure 2. Expression and regulation of FGFs in spinal cord and NSC34-cells.

(A) Transcript abundances relative to GAPDH in control mice spinal cords and scrambled siRNA transfected NSC34-cells. FGF mRNA levels were measured by qRT-PCR in pooled samples of P5 control mice tissues and scrambled siRNA transfected cells relative to GAPDH as internal control. Transcript abundances were calculated from ΔC_T-values. Unpaired t-tests of technical repetitions of measurements in cell-culture compared to tissue abundances were performed (n = 2; n.s. = non significant, *p<0.05, **p<0.01, ***p<0.001). Bars represent means with standard deviations (SD). (B) Fold-change of FGF transcript levels in SMA-mice spinal cords and NSC34 cells after SMN-knockdown. FGF transcript concentration was measured by qRT-PCR in SMA-mice spinal cords and SMN-siRNA treated NSC34-cells. Transcript levels of SMA-mice and control animals were measured at P1, P5 and P8. NSC34-cells were either transfected with SMN- or control scrambled-siRNA in four independent experiments (n = 4) with three replicates in each group. The knockdown for each experiment was monitored by western-blot analysis (Fig. 3). The fold-changes were calculated in comparison to each corresponding control group. Fold-changes of SMA-mice spinal cords were calculated compared to transcript levels of control mice of the same age. Results of SMN siRNA treated cells were calculated compared to scrambled RNA transfected cells of the same experiment. SMA-mice transcript levels were tested against control animals by a Mann-Whitney test (n≥5, *p<0.05, **p<0.01). mRNA-levels of SMN siRNA transfected cells were tested against scrambled siRNA transfected cells by repeated measurements two way ANOVA (n = 4, **p = 0.0046). Bars for fold-changes represent means with standard error of mean (SEM).

Figure 3. Western blot analysis of NSC34-cells after SMN-knockdown.

(A) Phosphorylation of FGF-downstream targets Akt and ERK was analyzed in siRNA treated (si) and scrambled siRNA (scr) transfected cells. Phospho-antibodies against Akt (pAkt, S473), and ERK1/2 (pERK, T202,T204) were used to quantify changes in phosphorylation levels compared to non-phosphorylated Akt, ERK. Four independent experiments with three replications were performed. (B) Densitometrical measurements of SMN-bands normalized by α-tubulin showed an efficient knockdown of 20±5% in comparison to control-siRNA-transfected cells. (C) pAkt normalized to non-phosphorylated Akt was upregulated by a factor of 2.8±0.7. (D) pERK normalized to non-phospho-ERK was upregulated by a factor of 2.1±0.3. (E) The functional link between FGFR-1 signaling and ERK-hyperphosphorylation was analyzed by application of the specific FGFR-1 inhibitor PD173074 (50 µM). Additionally, FGF-2 was added to the medium in a final concentration of 50 ng/ml. (F) Densitometrical measurements of pERK normalized to non-phospho ERK revealed an upregulation of 10.6±3.1 fold in scrambled siRNA transfected cells under FGF-2 incubation compared to scrambled siRNA transfected control conditions. When transfected with SMN siRNA, the pERK level rises up to 23.7±0.6 fold change. These differences disappeared after PD173074 incubation. Bars and values represent means with standard error of mean (SEM). Significance was tested via repeated measurements two-way ANOVA (B, C, D, n = 4, ***p<0.0001, **p<0.01, *p<0.05) and paired ratio t-test (F, n = 6 for control conditions and n = 3 for remaining conditions, **p<0.01, *p<0.05).

Figure 4. Crosstalk of neurotrophic factor signaling and ROCK pathway leading to neurite outgrowth – involvement of SMN protein.

FGF-signaling promotes neurite outgrowth by small GTPase-dependent-, ERK- and Akt-pathways. Activation of the small GTPases Rac and Cdc42 and inhibition of the negative effector RhoA promotes outgrowth by posttranslational mechanisms. The ERK and Akt pathways, however, change transcriptional profiles towards an outgrowth promoting state. Both pathways functionally interact with RhoA-downstream-target Rho-kinase (ROCK) as a central node , . ROCK thereby is important for integration and processing signals of the two major pathways involved in neurite outgrowth. Interestingly, recent findings of our group suggest a role of SMN-Profilin2a interaction in ROCK-pathway dependent outgrowth. SMN-reduction thereby leads to changes in phosphorylation-dependent regulation of ROCK-downstream targets indicated by red arrows . Moreover, SMA-patient derived SMN-S230L-mutation does not interact with Prof2a and acts dominant negative on neurite outgrowth when overexpressed in vitro . In this study, FGFR-1 was found to be upregulated under SMN-knockdown and consistently Akt and ERK showed hyper-phosphorylations. Thereby a sequestration of ROCK via Prof2a binding might enhance ERK-hyper-phosphorylation in a FGFR dependent manner.