Genistein improves neuropathology and corrects behaviour in a mouse model of neurodegenerative metabolic disease

Abstract

Background: Neurodegenerative metabolic disorders such as mucopolysaccharidosis IIIB (MPSIIIB or Sanfilippo disease) accumulate undegraded substrates in the brain and are often unresponsive to enzyme replacement treatments due to the impermeability of the blood brain barrier to enzyme. MPSIIIB is characterised by behavioural difficulties, cognitive and later motor decline, with death in the second decade of life. Most of these neurodegenerative lysosomal storage diseases lack effective treatments. We recently described significant reductions of accumulated heparan sulphate substrate in liver of a mouse model of MPSIIIB using the tyrosine kinase inhibitor genistein.

Methodology/principal findings: We report here that high doses of genistein aglycone, given continuously over a 9 month period to MPSIIIB mice, significantly reduce lysosomal storage, heparan sulphate substrate and neuroinflammation in the cerebral cortex and hippocampus, resulting in correction of the behavioural defects observed. Improvements in synaptic vesicle protein expression and secondary storage in the cerebral cortex were also observed.

Conclusions/significance: Genistein may prove useful as a substrate reduction agent to delay clinical onset of MPSIIIB and, due to its multimodal action, may provide a treatment adjunct for several other neurodegenerative metabolic diseases.

Figure 1. Primary storage substrates are reduced in brains of MPSIIIB mice after genistein treatment.

(A) 11 month old MPSIIIB and WT, male and female mice with and without long-term genistein treatment were sacrificed and 30 µm coronal sections (numbered 1–4) were cut from each mouse at positions 0.26, −0.46, −1.18 and −1.94 relative to bregma. For cerebral cortex (hatched), two low power fields of view for each section (boxed) were quantified or positive cells counted (8 fields total), whilst for hippocampus (spots), two low power fields from sections 3–4 (boxed) were quantified (4 fields total). All sections were stained together and blinded to ensure consistency. (B) Representative Lysosomal Associated Membrane Protein (LAMP2) staining of cerebral cortex of 11 month old MPSIIIB and WT, male and female mice with and without long-term genistein treatment. This indicates the size of the lysosomal compartment and hence stored material in cells in layers II/III-V/VI of the cerebral cortex. The images correspond to section 2 shown in Figure 1A. Bar  = 100 µm. (C) Quantification of mean LAMP2 staining in cerebral cortex is expressed as a percentage of staining in untreated MPSIIIB mice. (D) Quantification of mean LAMP2 staining of hippocampus. (E) Mean weight of pathological heparan sulphate in the brain per mg protein, measured using the SensiPro assay. (F) High power view of cerebral cortex layer V - box from (B). Bar  = 100 µm. (G) Quantification of mean LAMP2 staining of 2 fields of view from 3 liver sections (6 fields total) is expressed as a percentage of staining in untreated MPSIIIB mice. (H) Mean weight of total glycosaminoglycans in the liver per µg DNA, measured using the Blyscan assay. For all graphs genders were pooled, thus n = 12 mice per group, error bars represent SEM, p values are for Tukey's multiple comparisons test.

Figure 2. Neuroinflammatory markers are reduced in brains of MPSIIIB mice after genistein treatment.

(A) Representative Glial Fibrillary Associated Protein (GFAP) staining of astrocytes (brown) in the cerebral cortex of 11 month old MPSIIIB and WT mice with and without long-term genistein treatment. Sections have been counterstained with haematoxylin to highlight nuclei (blue). The images correspond to section 2 shown in Figure 1A. Boxed areas are enlarged to show individual astrocyte cell bodies. Bar  = 100 µm for low power images and 50 µm for enlargements. (B) The mean number of GFAP positive cells in the cerebral cortex per low power field of view were counted as described in Figure 1. (C) Representative Isolectin B4 staining of microglial cells (brown) in the cerebral cortex. Sections have been counterstained with haematoxylin to highlight nuclei (blue). The images correspond to section 2 shown in Figure 1A. Boxed areas are enlarged to show individual microglial cells. Bar  = 100 µm for low power images and 50 µm for enlargements. (D) The mean number of Isolectin B4 positive cells in the cerebral cortex per low power field of view were counted. For all graphs genders were pooled, thus n = 12 mice per group, error bars represent SEM, p values are for Tukey's multiple comparisons test.

Figure 3. Some improvements were seen in secondary metabolite storage and markers of synaptic transmission in brains of MPSIIIB mice after genistein treatment.

(A) Representative GM2 ganglioside staining in the cerebral cortex of 11 month old MPSIIIB mice with and without long-term genistein treatment at low (layers II/III-V/VI) and high power (layer V). The images correspond to section 2 shown in Figure 1A. Bar  = 100 µm. WT mice have no GM2 staining in the cortex and are not shown. (B) Quantification of mean GM2 staining of cerebral cortex in arbitrary units. (C) Chromatographic profiles of total gangliosides from the fourth brain hemicoronal fifth (rostral to caudal). Each lane represents 3mg wet weight of pooled sections from 4 mice. (D) Representative Vesicle Associated Membrane Protein (VAMP2) staining in the cerebral cortex. VAMP2 is a synaptic vesicle marker required for effective synaptic transmission. The images correspond to section 2 shown in Figure 1A. Bar  = 100 µm. (E) Quantification of mean VAMP2 staining of cerebral cortex in arbitrary units. For all graphs genders were pooled, thus n = 12 mice per group, error bars represent SEM, p values are for Tukey's multiple comparisons test.

Figure 4. Genistein treatment corrects behavioural abnormalities seen in MPSIIIB mice.

(A) 8 month old MPSIIIB and WT, male and female mice with and without long-term genistein treatment were monitored in the open field test of locomotor and exploratory activity for 60 minutes. The arena was divided into 12 squares and the frequency of entry (and total duration- not shown) to the central 4 squares measured to determine responses to danger and thigmotaxis. (B) The average speed in the central area and (C) side squares was also measured, as was (D) total distance travelled, to give an indication of abnormal locomotor activity. (E) Rapid exploratory behaviour was measured by frequency and (F) duration of speed more than 90 mm/s. (G) The number of times and (H) the duration of time spent immobile was also measured as speed under 0.05 mm/s. (I) Latency to cross or fall from a hanging bar is a measure of locomotor activity with some element of cognitive function due to a training period and was tested at 10 months of age. Significant gender*genotype effects were seen in centre and side speed, distance travelled, frequency of speed over 90 mm/s, immobility frequency and duration reflecting greater pathological effect in female MPSIIIB mice. For all graphs genders were pooled, thus n = 12–14 mice per group, error bars represent SEM, p values are for Tukey's multiple comparisons test.