Murine models of acute neuronopathic Gaucher disease

Abstract

Gaucher disease (GD) is an autosomal recessive lysosomal storage disorder caused by mutations in the glucosidase, beta, acid (GBA) gene that encodes the lysosomal enzyme glucosylceramidase (GCase). GCase deficiency leads to characteristic visceral pathology and, in some patients, lethal neurological manifestations. Here, we report the generation of mouse models with the severe neuronopathic form of GD. To circumvent the lethal skin phenotype observed in several of the previous GCase-deficient animals, we genetically engineered a mouse model with strong reduction in GCase activity in all tissues except the skin. These mice exhibit rapid motor dysfunction associated with severe neurodegeneration and apoptotic cell death within the brain, reminiscent of neuronopathic GD. In addition, we have created a second mouse model, in which GCase deficiency is restricted to neural and glial cell progenitors and progeny. These mice develop similar pathology as the first mouse model, but with a delayed onset and slower disease progression, which indicates that GCase deficiency within microglial cells that are of hematopoietic origin is not the primary determinant of the CNS pathology. These findings also demonstrate that normal microglial cells cannot rescue this neurodegenerative disease. These mouse models have significant implications for the development of therapy for patients with neuronopathic GD.

Fig. 1.

Aberrant splicing of gba mRNA causes GCase deficiency in K14-lnl/lnl mice. (A) Schematic view of the generation of the K14-lnl/lnl mice; mice harbouring the lnl allele in intron 8 of the gba gene were bred with K14-Cre mice (21). K14-mediated expression of Cre enabled removal of the lnl cassette in the epidermis of the K14-lnl/lnl mice. Black bars represent gba exons, gray bars represent metaxin exons, open triangles represent loxP sites, and P1, P2, and P3 denote primer locations. (B) (Upper Left) The lethal skin phenotype in the newborn gba^lnl/lnl mouse was similar to the phenotype of previous GCase-deficient mice (17, 19, 23). (Upper Right) Newborn littermate control [representative of both gba^lnl/+; K14-Cre (termed K14-lnl/wt) and gba^+/+; K14-Cre (termed K14-wt)]. (Lower) K14-lnl/lnl mouse at end stage (constant paralysis, ≈2 weeks old). (C) Representative PCR showing removal of the lnl cassette [visualized as presence of a wild-type (wt) band] in the skin of K14-lnl/lnl mice. (D) Retention of the lnl cassette in intron 8 of the gba gene causes aberrant splicing. (Left) A prominent splice-variant (blue box) was obtained through RT-PCR on brain, spleen, and liver specimens from K14-lnl/lnl mice. (Right) Schematic representation of the variant sequence. Correctly spliced gba mRNA was also present in these mice (red box). Actin was used as loading control. S, stop codon; ex, exon; int, intron; Neo, neomycin resistance gene. (E) GCase activity was severely reduced in brain, spleen, and liver of the K14-lnl/lnl mice (n = 4–7) compared with both K14-lnl/wt (n = 5–8) and K14-wt mice (n = 5). GCase activity in K14-wt mice was set to 1. All mice were 2 weeks old at the time of analysis. Error bars represent SD. (F) K14-lnl/lnl mice (n = 7) had elevated levels of Glccer in brain, spleen, and liver compared with K14-lnl/wt (n = 3) and K14-wt (n = 3) mice. Error bars represent SD.

Fig. 2.

Cell loss in brains of GCase deficient mice. Histological analyses of K14-lnl/wt (control) mice (A–D) and K14-lnl/lnl mice (E–L). Nissl staining of cryosections of brains from K14-lnl/lnl mice displayed normal brain architecture (E) similar to the control mice (A). Extensive loss of neurons was observed in the cortex (F) and thalamus (H). C and G are higher magnifications of the cortical layer V (indicated by squares labeled “C” and “G”) from B and F, respectively. Note the loss of large pyramidal neurons in cortical layer V (G). (H) The thalamic region of mutant mice also showed a severe loss of neurons. (I) A large number of pyknotic cells (arrows) were scattered throughout the brain. (J) Neurons surrounded by microglia processes of one or several phagocytic cells with ameboid shapes (arrows) were frequently detected in regions with abundant apoptotic cell death. Note the presence of huge vacuoles (arrows) inside large neurons of the motor trigeminal nucleus (K) and pons region (L) of mutant mice. All K14-lnl/lnl mice and controls were ≈2 weeks old at analysis. In B and F, I–VI indicate the cortical layers. (Scale bars: A, 600 μm; B, 200 μm; C, D, K, and L, 30 μm; I, 20 μm; J, 10 μm. Scale of E matches A, F matches B, G matches C, and H matches D.)

Fig. 3.

Neuropathology and increased apoptotic cell death in GCase-deficient mice. Histological stainings for TUNEL (A–F) and activated caspase 3 (G–L). Compared with control mice (A–C and G–I), apoptotic cell death was abundantly increased in thalamus (D and J), dendate gyrus (DG) (E and K), and cerebellum (F and L) in K14-lnl/lnl mice. (M–O) FluoroJade B labeling was observed in K14-lnl/lnl mice (N and O) but not in control animals (M). Degenerating neurons were scattered throughout the brain in GCase-deficient mice as shown for the cortex (N) and hippocampus (O). (P and Q) In contrast to controls (P), calbindin-labeled Purkinje neurons were reduced in number and exhibited swollen dendrites in the mutant mice (Q). (R) The presence of activated caspase 3-positive cells with typical neuronal morphologies indicates apoptotic neuron death in K14-lnl/lnl mice. (S-X) Oil red O staining was used to detect lipid accumulation in control (S and V) and mutant (T, U, W, and X) mice. Regions in the medulla (S–U) and the cerebellum (V–X) displayed a marked accumulation of lipids in mutant animals. U and X are higher magnifications of T and W, respectively, and reveal the presence of lipids in both axons and cell bodies (arrows indicate neuronal cell bodies). Note the selective accumulation of lipids in both the cell bodies (arrows) and dendrites of Purkinje cells in the cerebellum of GCase deficient mice (W and X). Only background staining with nonspecific oil red O precipitates was observed in control brain sections (S and V). All K14-lnl/lnl mice and controls were ≈2 weeks old at analysis. [Scale bars: A (applies to A–L, V, and W), 100 μm; M (applies to M–Q, S, and T), 60 μm; R, U, and X, 30 μm.]

Fig. 4.

Normal microglial function in the brain of Nestin-flox/flox mice. Immunostaining with microglia markers: Iba-1 (A–L) and mac-2 (M–X) in indicated brain regions of K14-lnl/lnl (A–F and M–R) and Nestin-flox/flox (G–L and S–X) mice. Iba-1 stains both resting and activated microglia, whereas mac-2 is more selective for activated/proliferating microglia (–43). No mac-2 staining was observed in control animals. Higher magnification of the thalamic region shows the presence of lipid-engorged microglia (arrows) in K14-lnl/lnl mice (E, F, Q, and R) but not in Nestin-flox/flox mice (K, L, W, and X). Note the higher activation of microglia in Nestin-flox/flox animals (G–L and S–X) compared with K14-lnl/lnl mice (A–F and M–R), suggesting a proliferation defect of microglia in K14-lnl/lnl mice. All K14-lnl/lnl mice and Nestin-flox/flox mice were analyzed at end-stage paralysis, corresponding to an age of ≈2 and ≈3 weeks, respectively. (Scale bars: A, 600 μm; B–E, 80 μm; F, 20 μm. Scale within columns is uniform.)