Engrailed-2 and inflammation convergently and independently impinge on cerebellar Purkinje cell differentiation

Abstract

Autism spectrum disorders (ASD) have a complex pathogenesis thought to include both genetic and extrinsic factors. Among the latter, inflammation of the developing brain has recently gained growing attention. However, how genetic predisposition and inflammation might converge to precipitate autistic behavior remains elusive. Cerebellar structure and function are well known to be affected in autism. We therefore used cerebellar slice cultures to probe whether inflammatory stimulation and (over)expression of the autism susceptibility gene Engrailed-2 interact in shaping differentiation of Purkinje cells, key organizers of cerebellar histogenesis and function. We show that lipopolysaccharide treatment reduces Purkinje cell dendritogenesis and that this effect is enhanced by over-expression of Engrailed-2 in these cells. The effects of lipopolysaccharide can be blocked by inhibiting microglia proliferation and also by blocking tumor necrosis factor alpha receptor signaling, suggesting microglia and tumor necrosis factor alpha are major players in this scenario. These findings identify Purkinje cells as a potential integrator of genetic and environmental signals that lead to an autism-associated morphology.

Keywords: Autism; Cerebellum; Engrailed; Inflammation; LPS; Microglia; Purkinje cell differentiation; Slice culture; Tumor necrosis factor alpha.

Fig. 1

Engrailed-2 overexpression affects dendritogenesis of PCs developing in slice cultures. Slices prepared from 6-day-old mice and cultured for 6 days were immunohistochemically stained for CALB1 (green) and EN2 (red). Note the simple and sparsely ramified dendrites of Purkinje cells of L7En-2 mice (B) as compared to cells derived from FVB/N wildtype mice (A). Staining for EN2 (red) reveals overexpression of this transcription factor in Purkinje cells of L7En-2 derived slices (red arrow). Endogenous EN2 expression in numerous non-Purkinje cells is visible in cultures from both genotypes. Bar is 100 µm for overviews and 25 µm for inserts

Fig. 2

Effect of different concentrations of LPS/IFG on PC arborization in wildtype and L7En-2 PCs. Cerebellar slice cultures were prepared from 6-day-old FVB/N wildtype (A–C) or L7En-2^tg transgenic mice (D, E), treated with either low (lc, 10 ng/ml in B, E) or high concentration (hc, 100 ng/ml in C, F) of each LPS and IFG and fixed at DIV 6. In A and D, typical untreated PCs of either genotype are shown. In both genotypes, the overall complexity of the dendritic tree seems to be diminished by inflammation

Fig. 3

Additive effects of LPS/IFG treatment and EN2 overexpression on PC dendritogenesis. Cerebellar slice cultures prepared from 6-day-old FVB/N wildtype (A–C) or L7En-2^tg transgenic mice were treated with low (lc) or high (hc) concentrations of LPS/IFG and analyzed at DIV 6. Both genotype and inflammatory stimulation reduced total dendritic length (A) and number of dendritic terminals (B). Consistently, numbers of PC dendrites reaching a higher branch levels are reduced by EN2 expression and inflammatory stimulation (C). This may be taken from the left shift of the Hill curves describing the distribution of segment levels due to inflammatory stimulation (for wildtype, p_ct,lc, p_ct,hc, p_lc,hc, all < 0.002; for L7En-2, p_ct,lc, p_ct,hc, p_lc,hc, all < 0.02) and to the genotype. Genotype effects were assessed at identical treatment levels (p_ct, p_{LI lc}, p_LIhc, all < 0.0003). To facilitate comparison, PC numbers are given as percentages of the total number of PCs analyzed. Statistics were calculated on original values. Segment lengths of wildtype PC dendrites were similar across levels and not influenced by inflammatory stimulation (D; all p-values larger than 0.1). In L7En-2 PCs, inflammation increased segment lengths at levels 1, where the low dose of LPS resulted in a significant increase (p_ct,lc = 0.0236). The same was true for level 2 (p_ct,lc = 0.0191). An even further increase was seen with high dose of LPS (p_ct,hc, = 0.0021). At level 3 the change approached the threshold of significance (p_ct,hc = 0.0682) in tg PCs. No differences were seen at level 4. Significance levels for genotype differences are indicated by light gray symbols, and those for differences due to LPS/IGF-treatment are indicated by black symbols). Two-way ANOVA *p < 0.05, **p < 0.01, ***p < 0.001; n_{ct, wt} = 41, n_{lc, wt} = 26, n_{hc, wt} = 32, n_{ct, tg,} = 26, n_{lc, tg} = 26, n_{hc, tg} = 11)

Fig. 4

Blocking LPS-induced microglia proliferation restores PC dendritogenesis. Cerebellar slice cultures prepared from 6-day-old C57Bl6/J mice were treated with LPS (L) and PLX3397 (P) and fixed at DIV 6. Morphometric analysis revealed a clear difference in total dendritic length, number of branch levels and number of terminals between LPS and untreated cultures (ct), and also between LPS- and PLX-treated (L + P) vs. LPS treated (L) cultures. Overall, parameters of ct and L + P cultures were rather more similar to each other than to those of LPS-treated cells (“ns” not significant, ** p < 0.01, *** p < 0.001 (n_ct = 41, n_L = 32, n_L+P = 24)

Fig. 5

Inflammation enhances PC dendritogenesis in slice cultures derived from newborn mice. Cerebellar slice cultures were prepared from newborn C57Bl6/J wildtype mice, treated with LPS (L) or LPS and PLX (L + P) and fixed at DIV 10 (bar is 20 µm). Control cultures (ct) were not treated. A visual comparison already demonstrated clear differences in PC arborization between untreated (A) and LPS treated PCs (B). Morphometric analysis revealed a clear difference in total dendritic length (C), and in the number of terminals (D) between untreated and LPS treated PCs as well as between LPS and LPS + PLX treated PCs. (“ns” not significant, ***p < 0.001; n_ct, = 25, n_L = 37, n _L+P = 34)

Fig. 6

TNF alpha receptor blockage inhibits LPS mediated reduction of PC arborization in P6 cultures. Slice cultures prepared from 6-day-old C57Bl6/J wildtype mice were treated with LPS (L) and R-7050 (R) to block TNFaR signaling. As revealed by Western blotting, the supernatant of LPS treated cultures contained increased levels of monomeric TNFa (A). The band at about 60–70 kDal probably reflects weak non-specific binding of TNFa antibody to albumin and maybe used as a loading control. Albumin is quantitatively the predominant protein in the culture medium, and the lower panel shows albumin levels as assessed by Commassie staining (Alb/Cm). Total dendritic length, number of terminals and dendritic branching of LPS treated PCs were significantly different from control and LPS and R-7050 (L + R) treated cells (*p < 0.05). In contrast, controls and L + R treated cells could not be distinguished (all p > 0.28). n_ct,, = 21, n_L = 20, n_L+R = 18

Fig. 7

Inhibition of LPS induced effects on PC dendritogenesis by TNFaR blocker R-7050. Cerebellar slice cultures prepared from newborn C57Bl6/J mice were treated with LPS (L) or LPS and R-7050 (L + R) and fixed at DIV 10. Morphometric analysis revealed a clear difference in total dendritic length (A), and in the number of terminals (B) between untreated (ct) and LPS-treated PCs as well as between LPS- and LPS/R-7050-treated PCs. (*p < 0.05, **p < 0.01, ***p < 0.001; n_ct,, = 12, n_L = 12, n_L+R = 12)