Partial Tmem106b reduction does not correct abnormalities due to progranulin haploinsufficiency

Abstract

Background: Loss of function mutations in progranulin (GRN) are a major cause of frontotemporal dementia (FTD). Progranulin is a secreted glycoprotein that localizes to lysosomes and is critical for proper lysosomal function. Heterozygous GRN mutation carriers develop FTD with TDP-43 pathology and exhibit signs of lysosomal dysfunction in the brain, with increased levels of lysosomal proteins and lipofuscin accumulation. Homozygous GRN mutation carriers develop neuronal ceroid lipofuscinosis (NCL), an earlier-onset lysosomal storage disorder caused by severe lysosomal dysfunction. Multiple genome-wide association studies have shown that risk of FTD in GRN mutation carriers is modified by polymorphisms in TMEM106B, which encodes a lysosomal membrane protein. Risk alleles of TMEM106B may increase TMEM106B levels through a variety of mechanisms. Brains from FTD patients with GRN mutations exhibit increased TMEM106B expression, and protective TMEM106B polymorphisms are associated with decreased TMEM106B expression. Together, these data raise the possibility that reduction of TMEM106B levels may protect against the pathogenic effects of progranulin haploinsufficiency.

Methods: We crossed Tmem106b ^+/- mice with Grn ^+/- mice, which model the progranulin haploinsufficiency of GRN mutation carriers and develop age-dependent social deficits and lysosomal abnormalities in the brain. We tested whether partial Tmem106b reduction could normalize the social deficits and lysosomal abnormalities of Grn ^+/- mice.

Results: Partial reduction of Tmem106b levels did not correct the social deficits of Grn ^+/- mice. Tmem106b reduction also failed to normalize most lysosomal abnormalities of Grn ^+/- mice, except for β-glucuronidase activity, which was suppressed by Tmem106b reduction and increased by progranulin insufficiency.

Conclusions: These data do not support the hypothesis that Tmem106b reduction protects against the pathogenic effects of progranulin haploinsufficiency, but do show that Tmem106b reduction normalizes some lysosomal phenotypes in Grn ^+/- mice.

Keywords: Frontotemporal dementia; Lysosome; Progranulin; TMEM106B.

Fig. 1

Grn^+/−:Tmem106b^+/− Exhibit a 40–50% Reduction in Both Progranulin and Tmem106b Protein Levels. In the frontal cortex of 5–6 month-old mice, the knockout Grn and Tmem106b alleles mediated the expected reduction of Tmem106b (a, ANOVA effect of Tmem106b, p < 0.0001) and progranulin (b, ANOVA effect of Grn, p < 0.0001) protein levels. Similar results were obtained in the hippocampus, with the expected reduction of Tmem106b (c, ANOVA effect of Tmem106b, p < 0.0001) and progranulin (d, ANOVA effect of Grn, p < 0.0001). Tmem106b (e, ANOVA effect of Tmem106b, p < 0.0001) and progranulin (f, ANOVA effect of Grn, p < 0.0001) were also reduced in the frontal cortex of 12 month-old mice. In all cases, the knockout Grn and Tmem106b alleles reduced their target protein levels by ~ 35–45% regardless of the genotype of the other allele. * = p < 0.05, ** = p < 0.01, *** = p < 0.001, **** = p < 0.0001 by Tukey’s post-hoc test. n = 5–12 mice per group for 5- to 6-month-old mice and 10–11 per group for 12-month-old mice

Fig. 2

Tmem106b Reduction Does Not Rescue the Social Dominance Deficits of Grn^+/− Mice. When paired against Grn^+/+:Tmem^+/+ mice, Grn^+/−:Tmem^+/+ mice exhibited the expected losing phenotype as shown by both wins per genotype (a, Binomial test, p = 0.0053, n = 133 matches) and winning percentage (b, Mann-Whitney test, p = 0.0016, n = 45–46 mice per genotype). Grn^+/−:Tmem^+/− mice also exhibited a losing phenotype versus Grn^+/+:Tmem^+/+ mice (c, Binomial test, p < 0.0001, d, Mann-Whitney test, p < 0.0001, n = 30 mice per genotype). Tmem106b reduction did not i change social dominance between either Grn^+/− mice(h, Binomial test, p = 0.6879, i, Mann-Whitney test, p = 0.6198, n = 33 mice per genotype) or Grn^+/+ mice (f, Binomial test, p > 0.9999 g, Mann-Whitney test, p = 0.6914, n = 11 mice per genotype)

Fig. 3

Increased Enzyme Activity is Associated with Increased Enzyme Levels in Grn^+/− and Grn^−/− Mice. Grn^−/− mice exhibited elevated HexA activity in the frontal cortex at age 2–3 months (a, ANOVA effect of Grn, p < 0.0001, Dunnett’s post-hoc test, p = 0.0001), which was associated with reduced HexA protein levels (b, ANOVA effect of Grn, p = 0.0003, Dunnett’s post-hoc test, p = 0.0007). While Grn^+/− mice did not exhibit elevated HexA activity at 2–3 months of age, 20 month-old Grn^+/− mice exhibited elevated HexA (a, t test, p = 0.037) and GCase (c, t test, p = 0.0408) activity in the frontal cortex. Consistent with this increased activity, Grn^+/− mice exhibited elevated HexA (b, t test, p = 0.0262) and GCase (d, t test, p = 0.0222) protein levels. * = p < 0.05, *** = p < 0.001, **** = p < 0.0001. n = 8–19 mice per genotype

Fig. 4

Tmem106b Reduction Does Not Rescue Most Lysosomal Phenotypes of Grn^+/− Mice. Frontal cortex samples from 5- to 6-month-old Grn^+/− mice exhibited no differences from wild-type in LAMP-1 levels (a) or GCase activity (c), but increased HexA activity (b, **** = ANOVA effect of Grn, p < 0.0001) and Gusb activity (d, ANOVA effect of Grn, p = 0.0009). Tmem106b reduction had no significant effect on LAMP-1, HexA, or GCase in 5- to 6-month-old mice, but significantly reduced Gusb activity (ANOVA effect of Tmem106b, p = 0.0031, * = p < 0.05 by Tukey’s post-hoc test). Frontal cortex samples from 12-month-old Grn^+/− mice exhibited increased LAMP-1 levels (e, ** = ANOVA effect of Grn, p = 0.0072), HexA activity (f, **** = ANOVA effect of Grn, p < 0.0001), GCase activity (g, * = ANOVA effect of Grn, p = 0.0179), and Gusb activity (h, ANOVA effect of Grn, p < 0.0001). As in the younger mice, Tmem106b reduction had no significant effect on LAMP-1, HexA, or GCase in 12 month-old mice, but significantly reduced Gusb activity (ANOVA effect of Tmem106b, p = 0.0010, * = p < 0.05, **** = p < 0.0001 by Tukey’s post-hoc test). n = 12 mice per group for 5- to 6-month-old mice and 11–12 mice per group for 12-month-old mice

Fig. 5

Tmem106b Reduction Suppresses Activity of β-Glucuronidase. Tmem106b reduction had no significant effect on LAMP-1 levels (a) or GCase activity (c) in the frontal cortex of 2- to 5-month-old mice. HexA activity was significantly reduced in Tmem106b^+/− mice (b, ANOVA effect of Tmem106b, p = 0.0175, ** = p = 0.0095 by Dunnett’s post-hoc test). Gusb activity was significantly reduced in both Tmem106b^+/− and Tmem106b^−/− mice (d, ANOVA effect of Tmem106b, p = 0.0007, ** = p < 0.01, *** = p < 0.001 by Dunnett’s post-hoc test). n = 6–7 mice per genotype

Fig. 6

Proposed Model of the Effects of Progranulin and Tmem106b in Grn^+/−:Tmem106b^+/− Mice. The current study and others [35], indicate that Tmem106b reduction in progranulin insufficient mice normalizes some lysosomal abnormalities induced by progranulin insufficiency, but fails to rescue the most abnormalities caused by progranulin insufficiency. In this model, progranulin insufficiency causes lysosomal dysfunction, which may then cause the social deficits of Grn^+/− mice. In addition, levels of many lysosomal proteins are increased, probably as a result of increased lysosomal gene expression via transcription factor EB. This increase in lysosomal proteins may also contribute to social behavior deficits, or may be a parallel phenomenon resulting from the underlying lysosomal dysfunction. In contrast, Tmem106b reduction suppresses the expression of many lysosomal genes (Figs. 3, 4 and [35]). Genes affected by both progranulin insufficiency and Tmem106b reduction, such as Gusb (Fig. 3d, h), may have normalized activity in Grn^+/−:Tmem106b^+/− mice, but lysosomal dysfunction and social deficits remain intact in Grn^+/−:Tmem106b^+/− mice