PTG depletion removes Lafora bodies and rescues the fatal epilepsy of Lafora disease

Abstract

Lafora disease is the most common teenage-onset neurodegenerative disease, the main teenage-onset form of progressive myoclonus epilepsy (PME), and one of the severest epilepsies. Pathologically, a starch-like compound, polyglucosan, accumulates in neuronal cell bodies and overtakes neuronal small processes, mainly dendrites. Polyglucosan formation is catalyzed by glycogen synthase, which is activated through dephosphorylation by glycogen-associated protein phosphatase-1 (PP1). Here we remove PTG, one of the proteins that target PP1 to glycogen, from mice with Lafora disease. This results in near-complete disappearance of polyglucosans and in resolution of neurodegeneration and myoclonic epilepsy. This work discloses an entryway to treating this fatal epilepsy and potentially other glycogen storage diseases.

Figure 1. LB in brain of 12 month-old LKO and DKO mice.

(a–b) Frontal cortex and hippocampus respectively from a LKO mouse stained with PAS-D. Note abundant LB within the neuropil and in the perikarya of numerous neurons. (c–d) Comparable regions from a DKO mouse. Arrows, examples of LB. All bars, 50 µm.

Figure 2. LB numbers in brain.

(a) Morphometric analysis of granular LB in whole brain and different brain regions. Granular is the histochemical description of the small LB in the neuropil, which by electron microscopy are shown to be in neuronal processes, mainly dendrites. Statistics: p<0.001 in all regions between LKO and DKO (ANOVA); n = 4 per genotype. (b) Morphometric analysis of perikaryal LB. Statistics: p<0.001 between LKO and DKO, except in cerebellum, where the difference is not significant (ANOVA); n = 4 per genotype.

Figure 3. LB in skeletal muscle.

(a) Muscle from a LKO mouse stained with PAS-D. Note presence of numerous LB in many fibers; bar, 100 µm; arrows, LB-replete myofibers; arrowheads, myofibers not containing LB. (b) Comparable field from a DKO mouse; bar, 50 µm. Higher magnification chosen for the DKO example to illustrate lack of even small LB.

Figure 4. Glycogen levels in skeletal muscle and brain in 12-month-old wt, LKO, and DKO mice (µmol glucose/gm tissue).

Skeletal muscle (a) and brain (b).

Figure 5. Gliosis in LKO mice.

(a) Hippocampus of a LKO mouse stained with GFAP. Note the large numbers of GFAP-positive astrocytes. (b) Comparable region from a DKO mouse. Bars, 100 µm. Arrows, astrocytes; arrowheads, gliosis. (c) Counts of GFAP-positive astrocytes. For significance, whole brain p<0.02; hippocampus p<0.001; cerebellum, not significant; frontal cortex, p<0.002 (ANOVA); n = 4–7 per genotype.

Figure 6. Neurodegeneration in LKO mice.

(a–c) Cerebellar Purkinje cells from a wt mouse. Note the smooth appearance of the plasma membrane and the absence of any voids between the cell and the surrounding neuronal processes; N, nucleus; arrows, synapses. (d) Purkinje cell from a LKO mouse. Numerous LBs are seen surrounding the cell (asterisks). (e) Higher power of an LKO Purkinje cell; arrows indicate wrinkling and retraction of plasma membrane. (f) A large LB (asterisk) in close proximity to a degenerating LKO Purkinje cell. (g) A typical DKO Purkinje cell. Note its full normal appearance and smooth plasma membrane. Part of a second normal Purkinje cell is seen to the right of the panel. (h) Higher power of plasma membrane from a DKO Purkinje cell. Note the relative linear appearance of the membrane and the attached synapses (arrows) typical of a normal cell. (i) One of few LB (asterisk) near Purkinje cells detected by EM in DKO.

Figure 7. Myoclonus per minute in wt, DKO, and LKO mice.

n = nine to 12 mice per genotype.