Characterization of small fiber pathology in a mouse model of Fabry disease

Abstract

Fabry disease (FD) is a life-threatening X-linked lysosomal storage disorder caused by α-galactosidase A (α-GAL) deficiency. Small fiber pathology and pain are major FD symptoms of unknown pathophysiology. α-GAL deficient mice (GLA KO) age-dependently accumulate globotriaosylceramide (Gb3) in dorsal root ganglion (DRG) neurons paralleled by endoplasmic stress and apoptosis as contributors to skin denervation. Old GLA KO mice show increased TRPV1 protein in DRG neurons and heat hypersensitivity upon i.pl. capsaicin. In turn, GLA KO mice are protected from heat and mechanical hypersensitivity in neuropathic and inflammatory pain models based on reduced neuronal I_h and Na_v1.7 currents. We show that in vitro α-GAL silencing increases intracellular Gb3 accumulation paralleled by loss of Na_v1.7 currents, which is reversed by incubation with agalsidase-α and lucerastat. We provide first evidence of a direct Gb3 effect on neuronal integrity and ion channel function as potential mechanism underlying pain and small fiber pathology in FD.

Keywords: Fabry disease; alpha-galactosidase A; globotriaosylceramide; human biology; medicine; mouse; neuroscience; pain.

Figure 1.. Toluidin blue staining and immunoreaction against globotriaosylceramide and immunoglobulin binding protein of mouse dorsal root ganglia.

Photomicrographs display hematoxylin-eosin staining DRG neurons from young and old GLA KO and WT mice (A–D) and exemplified measured cell area (yellow circles). (E) Quantification of neuronal cell area revealed increased cell size in young GLA KO compared to young WT mice (p<0.01) and in old GLA KO compared to young GLA KO and old WT mice (p<0.001 each). Photomicrographs show toluidin blue staining (F–I) of 0.5 µm semithin sections of dorsal root ganglia (DRG) from young (3 months) and old (≥12 months) wildtype (WT) and α-galactosidase A deficient (GLA KO) mice. Additionally, photomicrographs display immunoreactivity of antibodies against CD77 as a marker for globotriaosylceramide (Gb3) (J–M) and against binding immunoglobulin protein (BiP) (N–O) on 10 µm cryosections of DRG of old GLA KO and WT mice. No deposits were found in DRG neurons of young WT mice (F, arrow), neurons of a young GLA KO mice showed few intraneuronal deposits (G, arrowheads). Similar to young WT mice, there were no deposits in DRG neurons of old WT mice (H, arrow). Old GLA KO mice, however, displayed many deposits in DRG neurons (I, arrowheads). Gb3 load was not different between young GLA KO, young WT, and old WT mice (J–M), while old GLA KO mice displayed increased Gb3 accumulation in DRG neurons (M, arrows) and extraneural structures (M, arrowheads). BiP was homogeneously expressed in DRG neurons of old WT mice sparing the nucleus (N, arrows). Neurons of old GLA KO mice showed increased accumulation of BiP around the nucleus, indicating accumulation in the endoplasmic reticulum (O, arrows). GLA KO: young (3 months; hematoxylin-eosin: male; toluidine: female; CD77: male), old (≥12 months; hematoxylin-eosin: female; toluidine: female; CD77: male). WT: young (3 months; hematoxylin-eosin: male; toluidine: female; CD77: male), old (≥12 months; hematoxylin-eosin: female; toluidine: male; CD77: male). Scale bar hematoxylin-eosin: 50 µm. Scale bar toloudin blue: 10 µm. Scale bar CD77: 50 µm. The non-parametric Mann-Whitney U test was applied for group comparison. **p<0.01; ***p<0.001.

Figure 2.. Reduced intraepidermal nerve fiber density in α-galactosidase A deficient mice and Gb3 distribution in sciatic nerve and skin.

Photomicrographs show immunoreactivity of antibodies against protein gene product 9.5 (PGP 9.5) as a pan-axonal marker in 40 µm skin sections from footpads of young (3 months) and old (≥12 months) wildtype (WT) and α-galactosidase A deficient (GLA KO) mice (A–D). Arrows indicate single intraepidermal nerve fibers. Boxplots (E) show quantification of intraepidermal nerve fiber density (IENFD). Young WT mice had a higher IENFD compared to young GLA KO and old WT mice (p<0.001, each). Old GLA KO mice showed the most prominent IENFD reduction compared with young GLA KO and old WT mice (p<0.001 each). Additionally, photomicrographs display immunoreactivity of antibodies against CD77 and β-(ΙΙΙ)-tubulin in 10 µm sciatic nerve sections (F–K) and immunoreactivity of antibodies against CD77 and PGP 9.5 in 40 µm skin section (L–Q) of old GLA KO and WT mice. There were no Gb3 depositions detectable. GLA KO: young (3 months, n = 11 male, n = 10 female), old (≥12 months, n = 8 male, n = 11 female). WT: young (3 months, n = 10 male, n = 10 female), old (≥12 months, n = 10 male, n = 9 female). Box plots represent the median value and the upper and lower 25% and 75% quartile. Scale bar: 50 µm. The non-parametric Mann-Whitney U test was applied for group comparison. ***p<0.001.

Figure 3.. More apoptosis and less neurite outgrowth in dorsal root ganglion neurons of old α-galactosidase A deficient mice compared to wildtype mice.

Photomicrographs show the results of a NucView 488 Caspase 3 Enzyme Substrate Assay of cultivated dorsal root ganglion (DRG) neurons from old (≥12 months) wildtype (WT) and α-galactosidase A deficient (GLA KO) mice in the naïve state and after incubation with 500 nM staurosporine (STS) as a positive control (A–D). Empty arrows indicate caspase 3 negative neurons and filled arrows point to caspase 3 positive neurons. Bar graphs show the quantification of caspase 3 positive neurons (E). Cultured DRG neurons of old WT mice in the naïve state displayed a lower percentage of caspase 3 positive neurons than those of old GLA KO mice (p<0.001) and neurons of old WT mice incubated with 500 nM STS (p<0.05). DRG neurons of old GLA KO mice incubated with 500 nM STS showed a higher percentage of caspase 3 positive neurons compared to neurons in the naïve state (p<0.05) and WT positive control neurons (p<0.01). Further, neurite outgrowth was quantified (F). DRG neurons of old WT mice in the naïve state displayed a higher percentage of neurons with neurite outgrowth after 48 hr cultivation compared to neurons from old GLA KO mice (p<0.001). NucView 488 Caspase 3 Enzyme Substrate Assay was performed three times on cultures derived from three different mice of each genotype. GLA KO: old (≥12 months, n = 2 male, one female). WT: old (≥12 months, n = 2 male, one female). Number of neurons analyzed are integrated into the corresponding bar. Scale bar: 50 µm. The non-parametric Mann-Whitney U test for group comparisons was applied. *p<0.05;**p<0.01;***p<0.001.

Figure 4.. Expression, function, and phenotypic reflection of transient receptor potential vanilloid one channels in α-galactosidase A deficient mice.

(A) Boxplots show the results of transient receptor potential vanilloid 1 (TRPV1) channel gene expression in dorsal root ganglia (DRG) of young (3 months) and old (≥12 months) wildtype (WT) and α-galactosidase A deficient (GLA KO) mice. No intergroup difference was found. (B–E) Photomicrographs illustrate immunoreactivity of antibodies against TRPV1 in DRG of young and old WT and GLA KO mice; F) shows the result of quantification. Young and old GLA KO mice showed greater TRPV1 immunoreactivity compared to WT littermates (p<0.001 each). (G) TRPV1 positive neurons were predominantly smaller than 25 µm in diameter. (H, I) Photomicrographs exemplify cultured DRG neurons of an old WT (H) and GLA KO mouse (I). While cultured neurons appeared normal in WT mice (H), intracellular deposits were found in neurons of GLA KO mice (I). In J) a TRPV1 current is exemplified which was recorded from cultured DRG neurons of a young GLA KO mouse after application of 500 nM capsaicin (red bar). (K) TRPV1 current density did not differ between young GLA KO mice and WT littermates. (L) Line charts show thermal withdrawal latencies of old GLA KO and WT mice before and after intraplantar capsaicin injection. Old GLA KO mice displayed increased thermal sensitivty compared to baseline 24 hr after capsaicin injection (p<0.01). GLA KO: young (3 months; gene expression: n = 2 male, n = 4 female; protein expression: n = 12 male, n = 9 female), old (≥12 months; gene expression: n = 4 male, n = 2 female; protein expression: n = 10 male, n = 10 female). WT: young (3 months; gene expression: n = 2 male, n = 4 female; protein expression: n = 9 male, n = 8 female), old (≥12 months; gene expression: n = 3 male, n = 3 female; protein expression: n = 9 male, n = 8 female). TRPV1 currents: Three cells per genotype (GLA KO: n = 1 male, n = 2 female; WT: n = 1 male, n = 2 female) were quantified for calculation of current density. Capsaicin: GLA KO: old (≥12 months; Baseline: n = 33; capsaicin: n = 2 male, n = 6 female). WT old (≥12 months; Baseline: n = 32; capsaicin: n = 3 male, n = 5 female). Behavioral experiments were performed by an observer blinded to the genotype. Scale bar: 50 µm. The non-parametric Mann-Whitney U test for group comparisons was applied. Behavioral data were analyzed using a two-way ANOVA followed by Tukey’s post-hoc test.**p<0.01;***p<0.001.

Figure 5.. Expression, function, and phenotypic reflection of hyperpolarization-activated cyclic nucleotide-gated ion channels in α-galactosidase A deficient mice.

(A) Boxplots show the results of potassium/sodium hyperpolarization-activated cyclic nucleotide-gated ion channel 2 (HCN2) gene expression in dorsal root ganglia (DRG) of young (3 months) and old (≥12 months) wildtype (WT) and α-galactosidase A deficient (GLA KO) mice. No intergroup difference was found. (B–E) Photomicrographs illustrate immunoreactivity of antibodies against HCN2 in DRG of young and old WT and GLA KO mice; (F) shows the result of quantification. Old GLA KO and WT mice showed greated HCN2 immunoreactivity compared to young littermates (p<0.05 each) without difference between genotypes. In (G) I_h currents are exemplified recorded from a young GLA KO mouse. (H) I_h current densities did not differ between young GLA KO mice, WT littermates, and old WT mice, while old GLA KO mice displayed markedly reduced I_h current intensities compared to young GLA KO and old WT mice (p<0.001 each). Line charts show thermal withdrawal latencies (I) and mechanial withdrawal thresholds (J) of old GLA KO and WT mice before and after chronic constriction injury (CCI) of the right sciatic nerve. While WT mice showed thermal (p<0.001) and mechanical (p<0.001) hypersensitivity already at day three after CCI lasting up to day 28, old GLA KO mice were spared (p<0.01). GLA KO: young (3 months; gene expression: n = 2 male, n = 4 female; protein expression: n = 2 male, n = 4 female), old (≥12 months; gene expression: n = 4 male, n = 2 female; protein expression: n = 3 male, n = 3 female). WT: young (3 months; gene expression: n = 2 male, n = 4 female; protein expression: n = 4 male, n = 2 female), old (≥12 months; gene expression: n = 3 male, n = 3 female; protein expression: n = 1 male, n = 5 female). I_h currents: At least nine cells per genotype and age-group from at least three different mice each were analyzed. GLA KO young (3 months; n = 4 male, n = 5 female), old (≥12 months; n = 3 male, n = 7 female). WT young (3 months; n = 3 male, n = 6 female), old (≥12 months; n = 4 male, n = 6 female). CCI: GLA KO: old (≥12 months; Baseline: n = 33; CCI: n = 3 male, n = 3 female). WT: old (≥12 months; Baseline: n = 32; CCI: n = 3 male, n = 3 female). Scale bar: 50 µm. The non-parametric Mann-Whitney U test for group comparisons was applied. Behavioral data were analyzed using a two-way ANOVA followed by Tukey’s post-hoc test. *p<0.05;**p<0.01;***p<0.001.

Figure 6.. Expression, function, and phenotypic reflection of voltage-gated soium channel 1.7 in α-galactosidase A deficient mice.

(A) Boxplots show the results of voltage-gated sodium channel 1.7 (Na_v1.7) gene expression in dorsal root ganglia (DRG) of young (3 months) and old (≥12 months) wildtype (WT) and α-galactosidase A deficient (GLA KO) mice. No intergroup difference was found. (B) Na_v1.7 protein levels, as investigated by enzyme-linked immunosorbent assay, did not differ between genotypes and age-groups. (C) Graphs display exemplified sodium currents of old GLA KO and WT mice at −40 mV. Comparing young GLA KO mice with young and old WT mice, there no difference was found in current densities of sodium currents (D), while sodium currents of old GLA KO mice were markedly reduced compared to young GLA KO (p<0.001) and old WT mice (p<0.001). At baseline (E, black trace) sodium currents showed fast inactivation kinetics. After adding 100 nM tetrodotoxin (TTX) to the bath solution, sodium currents were completely blocked (E, red trace). Sodium currents recovered completely after washout with bath solution (E, grey trace). Line charts show thermal withdrawal latencies (F) and mechanical withdrawal thresholds (G) of young and old GLA KO and WT mice before and after intraplantar injection of complete Freund`s adjuvant (CFA). While young and old WT mice showed thermal and mechanical hypersensitivity already one hour after CFA injection lasting up to day seven, old GLA KO mice were spared from heat hypersensitivity (p<0.001). GLA KO: young (3 months; gene expression: n = 2 male, n = 4 female; protein expression: n = 3 male, n = 1 female), old (≥12 months; gene expression: n = 4 male, n = 2 female; protein expression: n = 2 male, n = 2 female). WT: young (3 months; gene expression: n = 2 male, n = 4 female; protein expression: n = 2 male, n = 2 female), old (≥12 months; gene expression: n = 3 male, n = 3 female; protein expression: n = 2 male, n = 2 female). Sodium currents: At least nine cells per genotype and age-group from at least three different mice each were analyzed. GLA KO young (3 months; n = 4 male, n = 5 female), old (≥12 months; n = 3 male, n = 7 female). WT young (3 months; n = 3 male, n = 6 female), old (≥12 months; n = 4 male, n = 6 female). CFA: GLA KO: young (3 months; n = 4 male, n = 2 female), old (≥12 months; Baseline: n = 33; CFA: n = 6 male, n = 6 female). WT: young (3 months; n = 4 male, n = 2 female), old (≥12 months; Baseline 32; CFA: n = 6 male, n = 6 female). Scale bar: 50 µm. The non-parametric Mann-Whitney U test for group comparisons was applied. Behavioral data were analyzed using a two-way ANOVA followed by Tukey’s post-hoc test. *p<0.05;***p<0.001.

Figure 7.. Knock-down of α-galactosidase A in human embryonic kidney 293 cells expressing voltage-gated sodium channel 1.7.

Photomicrographs show immunoreactivity of antibodies against CD77 as a marker for globotriaosylceramide (Gb3) accumulation in human embryonic kidney 293 (HEK) cells expressing voltage-gated sodium channel 1.7 (Na_v1.7) after one week of transfection with control small hairpin RNA (shRNA) (control HEK cells) (A-C, empty arrows), shRNA against α-galactosidase A (shRNA HEK cells) (D-F, arrows), and after 24 hr of incubation with agalsidase-alpha (G-I, empty arrows) and lucerastat (J-L, empty arrows). (M) Exemplified sodium currents of HEK cells transfected with control shRNA (black) and shRNA (red). (N) shRNA HEK cells displayed a marked reduction of Na_v1.7 currents compared to control shRNA HEK cells (p<0.01). Treatment with agalsidase-α (p<0.05) and lucerastat (p<0.01) restored Na_v1.7 currents. Na_v1.7 currents were not different between shRNA treated HEK cells incubated with agalsidase-α, or lucerastat and control cells. Control: n = 16; shRNA: n = 16; shRNA+ 24 hr agalsidase- α: n = 6; lucerastat: n = 11. Bar graphs represent the mean and standard error of the mean and at least three biological replicates. Scale bar 50 µm. The non-parametric Mann-Whitney U test for group comparison was applied. *p<0.05, **p<0.01.

Author response image 1.. Mechanical sensitivity after injection of complete Freund`s adjuvant.

Line chart displays mechanical withdrawal thresholds from old (≥12 months) wildtype (WT) and α-galactosidase A deficient (GLA KO) mice, calculated as percent change from baseline. Old WT mice were slightly more sensitive to mechanical stimulus, which was not significant except for 48 hours after CFA injection compared to old GLA KO mice. Data were analyzed using two-way ANOVA following Tukey’s post-hoc test after data transformation using Johnson’s procedure.