High phenylalanine concentrations induce demyelination and microglial activation in mouse cerebellar organotypic slices

Abstract

Phenylketonuria (PKU) is an inborn error of metabolism. Mutations in the enzyme phenylalanine hydroxylase (PAH)-encoding gene lead to a decreased metabolism of the amino acid phenylalanine (Phe). The deficiency in PAH increases Phe levels in blood and brain. Accumulation of Phe can lead to delayed development, psychiatric problems and cognitive impairment. White matter (WM) damage is a neuropathological hallmark of PKU and can be seen even in early detected and treated PKU patients. The mechanisms linking high Phe concentrations to WM abnormalities remain unclear. We tested the effects of high Phe concentrations on myelin in three in vitro models of increasing complexity: two simple cell culture models and one model that preserves local brain tissue architecture, a cerebellar organotypic slice culture prepared from postnatal day (P) 8 CD-1 mice. Various Phe concentrations (0.1-10 mM) and durations of exposure were tested. We found no toxic effect of high Phe in the cell culture models. On the contrary, the treatment promoted the maturation of oligodendrocytes, particularly at the highest, non-physiological Phe concentrations. Exposure of cerebellar organotypic slices to 2.4 mM Phe for 21 days in vitro (DIV), but not 7 or 10 DIV, resulted in a significant decrease in myelin basic protein (MBP), calbindin-stained neurites, and neurites co-stained with MBP. Following exposure to a toxic concentration of Phe, a switch to the control medium for 7 days did not lead to remyelination, while very active remyelination was seen in slices following demyelination with lysolecithin. An enhanced number of microglia, displaying an activated type morphology, was seen after exposure of the slices to 2.4 mM Phe for 10 or 21 DIV. The results suggest that prolonged exposure to high Phe concentrations can induce microglial activation preceding significant disruption of myelin.

Keywords: cerebellum; microglia; organotypic slices; phenylalanine; phenylketonuria; white matter.

FIGURE 1

Timeline of the four cerebellar organotypic slice culture experiments. Following preparation of the cerebellar organotypic slices (OSCs) (first blue line on the left), phenylalanine (Phe) was applied to the slices at 1 day in vitro (DIV) either at 0.1, 0.6, 1.2, or 2.4 mM (in experiment 1, A) or 0.1 or 2.4 mM (in all following experiments, B–D) and then the slices were fixed at different time points depending on the experiment. In all experiments, half of the medium was replaced 24 h after the OSC preparation, and then every 3 days (not shown on the timelines). In experiment 3 (C), Phe was replaced by the control medium at 21 DIV in half the slices treated with Phe.

FIGURE 2

Only after 21 DIV were the highest doses of phenylalanine deleterious to MBP staining. (A) Cerebellar slices were incubated in control or Phenylalanine (Phe)-supplemented medium for 21 days in vitro (DIV) at 0.1, 0.6, 1.2, and 2.4 mM. Results are presented as the area covered by myelin basic protein (MBP) staining as % of the total area of the field of view. Data are plotted as Mean ± SEM of N = 10 slices/treatment group. (B–D) Cerebellar slices were incubated for 21 DIV in a control medium (0.1 mM Phe) or a medium containing 2.4 mM Phe. The graphs show the results of the immunohistochemistry quantification: (B) Mean area covered by MBP staining as % of the total area of the field of view; representative images are shown in panels (Ea,d); (C) Mean area covered by calbindin staining as % of the total area of the field of view; representative images are shown in panels (Eb,e); (D) % of neurites stained for MBP; representative images are shown in panels (Ec,f). Data are plotted as Mean ± SEM of N = 10 slices/group. Scale bar = 20 μm. Statistical symbols indicate significance at ^***P < 0.001.

FIGURE 3

Stopping high phenylalanine for 7 days after 21 DIV of exposure did not result in detectable remyelination. Cerebellar slices were incubated for 21 or 28 days in vitro (DIV) in a control medium or a medium containing phenylalanine (Phe) at 2.4 mM. At 21 DIV, the Phe-enriched medium was replaced with the control medium (0.1 mM Phe) in one of the Phe-supplemented groups (called “Phe 21 stopped”), whereas the other Phe-exposed group was kept in the Phe-enriched medium for another 7 DIV (“Phe 28”). The graphs show the results of the immunohistochemistry quantification: (A) Mean area covered by myelin basic protein (MBP) staining as % of the total area of the field of view; representative images are shown in panels (Da,d,g); (B) Mean area covered by calbindin staining as % of the total area of the field of view; representative images are shown in panels (Db,e,h); (C) % of neurites stained for MBP; representative images are shown in panels (Dc,f,i). Data are plotted as Mean ± SEM of N = 10 slices/group. Scale bar = 20 μm. Statistical symbols indicate significance at *P < 0.05 and ^**P < 0.01.

FIGURE 4

Reactive microglia were seen as early as 10 DIV following continuous treatment with high phenylalanine. Cerebellar slices were incubated for 10 or 21 days in vitro (DIV) in a control medium or a medium containing phenylalanine (Phe) at 2.4 mM. The graphs show the results of the immunohistochemistry quantification. (A) Number of Iba1-positive cells as % of DAPI-positive nuclei from 20 fields of view per slice and 10 slices for each treatment group (N = 10); representative images are shown in panels (Bc,f,i,l); (C) Size analysis at 10 and 21 DIV of 60 microglial cells sampled across 10 slices in each treatment group; corresponding images are shown in panels (Bc,f,i,l). Data are plotted as Mean ± SEM of N = 10 slices/group. The insets (c’, f’, i’, l’) show details of the colocalization of the two stainings and of the morphological differences between treatment groups and time points. Note the different morphologies (amoeboid vs. elongated) and sizes of the cells as a function of their activation state. Scale bar = 20 μm. Statistical symbols indicate significance at ***P < 0.001 of Phe-treated slices vs. control slices and at ^#P < 0.05 and ^###P < 0.001 of 21 DIV vs. 10 DIV.