Core features of frontotemporal dementia recapitulated in progranulin knockout mice

Abstract

Frontotemporal dementia (FTD) is typified by behavioral and cognitive changes manifested as altered social comportment and impaired memory performance. To investigate the neurodegenerative consequences of progranulin gene (GRN) mutations, which cause an inherited form of FTD, we used previously generated progranulin knockout mice (Grn-/-). Specifically, we characterized two cohorts of early and later middle-aged wild type and knockout mice using a battery of tests to assess neurological integrity and behavioral phenotypes analogous to FTD. The Grn-/- mice exhibited reduced social engagement and learning and memory deficits. Immunohistochemical approaches were used to demonstrate the presence of lesions characteristic of frontotemporal lobar degeneration (FTLD) with GRN mutation including ubiquitination, microgliosis, and reactive astrocytosis, the pathological substrate of FTD. Importantly, Grn-/- mice also have decreased overall survival compared to Grn+/+ mice. These data suggest that the Grn-/- mouse reproduces some core features of FTD with respect to behavior, pathology, and survival. This murine model may serve as a valuable in vivo model of FTLD with GRN mutation through which molecular mechanisms underlying the disease can be further dissected.

Fig. 1. Kaplan-Meier survival plots for Grn mice

Survival curve of Grn-/- mice showed a median survival of 514 days.

Fig. 2. Grn-/- mice displayed normal levels of locomotor activity and exploratory behavior and only mild sensorimotor deficits at older ages

(A) Grn-/- and Grn+/+ mice showed similar levels of general locomotor activity (total ambulations) in both younger (9-12 months) and older (15-18 months) cohorts. Both groups showed significant habituation of ambulatory activity over the test session in each cohort (*p < 0.0003). (B) The Grn-/- and Grn+/+ groups in both cohorts also showed similar levels of exploratory behavior in terms of vertical rearing. However, both groups showed significant habituation only in the younger cohort (*p = 0.025 and 0.003, respectively) while only the Grn+/+ mice showed evidence of habituation in the older cohort (†p = 0.038). (C) No differences were observed between the Grn-/- and Grn+/+ mice on 5 out of the 6 sensorimotor measures in either cohort similar to that represented in the data from the ledge test. Numbers in parentheses above bars represent sample sizes for each group. (D) However, although the groups performed similarly on the inverted screen test at an earlier age, the Grn-/- mice in the older cohort were significantly (*p = 0.003) impaired on the inverted screen test suggesting that they may have deficits in strength at this age.

Fig. 3. Grn-/- male mice were not hyperaggressive but showed evidence of altered social interactions

(A) Grn-/- male mice were significantly less reactive to handling on average across three daily test sessions compared to Grn+/+ controls (**p = 0.009), with differences being significant during session 3 (*p = 0.013) and during session 2 (†p = 0.020). (B) However, no reliable differences were observed between the Grn-/- and Grn+/+ male groups on an aggression index (duration of fighting + biting) derived from data collected during the resident-intruder test. (C) In contrast, the Grn-/­male mice displayed a significantly decreased duration of pawing during nonaggressive interactions (**p=0.005), with differences being significant during session 3 (*p = 0.004), and very large during sessions 1 and 2 (†p < 0.025). (D) Male Grn-/- mice also spent a significantly decreased duration of time following the intruder mouse compared to the Grn+/+ group across test sessions (**p = 0.018) with differences being significant during session 3 (*p = 0.004). (E) The Grn-/- mice spent significantly more time alone compared to the Grn+/+ group (**p = 0.004) with significant differences being observed during sessions 1 (*p = 0.006) and 2 (*p = 0.012) and substantial differences also being found for session 3 (†p = 0.026).

Fig. 4. Grn-/- and Grn+/+ mice showed similar olfactory preferences for different odorants

(A) The Grn-/- and Grn+/+ mice made more pokes into holes that contained fresh bedding, which was a familiar odorant (*p = 0.0001 and †p = 0.044, respectively), relative to empty holes. (B) However, neither group showed a significant preference for a novel odorant (coconut extract). (C) In contrast, the Grn-/- and Grn+/+ groups showed a striking preference for a familiar food odorant while under food-restricted conditions as indicated by both groups showing significantly increased levels of poking into the odorant holes compared to empty holes (*p < 0.00005).

Fig. 5. Grn-/- mice exhibited mild spatial learning and memory deficits in the Morris water maze

(A) The Grn-/- and Grn+/+ groups performed similarly in terms of escape path length during the cued trials in the water maze, suggesting that the Grn-/-mice did not have impaired nonassociative functions that would affect performance on the subsequent place condition. (B) However, on average across the blocks of trials, the Grn-/- mice showed significant acquisition deficits in terms of path length during the place trials compared to the Grn+/+ controls (**p = 0.012). Differences were significant during block 3 (*p = 0.005), and very large during blocks 2 (†p = 0.033) and 5 (††p = 0.018). (C) The Grn-/- mice swam significantly more slowly on average across the blocks of trials compared to the Grn+/+ control mice (**p = 0.0001), and differences between groups were significant for blocks 1, 2, 3, and 5 (*p < 0.003), and large for block 4 (†p = 0.015). (D) During the probe trial, the Grn-/- group made significantly fewer crossings over the location where the platform had been located compared to the Grn+/+ mice (*p = 0.045) suggesting a deficit in the retention of the exact position of the platform. (E) However, the Grn-/- mice performed as well as Grn+/+ controls on a less highly-resolved retention of the platform in terms of showing a significant spatial bias for the target quadrant that contained the platform. Both groups spent significantly more time in the target quadrant compared to each of the other pool quadrants (*p < 0.00005).

Fig. 6. Grn-/- mice displayed impaired acquisition performance on a learning set protocol in the water maze

(A) A significant effect of Trial was found with regard to the path length of the mice from the second cohort during the learning set protocol in the water maze, (*p = 0.048). In addition, a significant effect of Trial (1 vs 4) was found in the analysis of the data from the Grn+/+ mice (†p = 0.024) but not from the Grn-/-group suggesting some acquisition of the learning set task in the Grn+/+ mice but not in the Grn-/- group. (B) For the Grn-/- mice the effect of Trial (trial 1 vs trial 4) was not significant indicating that there was no improvement in performance from trial 1 to trial 4, although performance on both trials improved across test days. (C) In contrast, the effect of Trial was significant in the Grn+/+ mice (**p = 0.047) showing that average path length was significantly decreased between trial 1 and trial 4 thus providing additional evidence for acquisition of the learning set task on the part of the Grn+/+ controls.

Fig. 7. Increased ubiquitination observed in 24-month old Grn-/- mice

Grn-/- mice (A, D, G; n=4) have an increased number of ubiquitinated structures, largely in the neuropil, compared to age-matched Grn+/+ mice (B, E, H; n=4) in the (A-C) hippocampus, (D-F) cortex, and (G-I) thalamus. Scale bar 200 μm in (A, B), 100 μm in (D), 50 μm in (E), and 50 μm in (G, H). Percent thresholding shown in C, F, and I (Grn-/- in white, Grn+/+ in black) as mean ± SEM. Statistics denoted as p>0.05, ns; p<0.01, **; and p<0.001, ***.

Fig. 8. Increased microgliosis observed in 24-month old Grn-/- mice

Grn-/- mice (A, D, G; n=4) have an increase in the number and size of IBA-1-immunoreactive microglial cells compared to age-matched Grn+/+ mice (B, E, H; n=4) in the (A-C) hippocampus, (D-F) cortex, and (G-I) thalamus. Scale bar 200 μm in (A, B), 100 μm in (D, E), and 100 μm in (G, H). Percent thresholding shown in C, F, and I (Grn-/- in white, Grn+/+ in black) as mean ± SEM. Statistics denoted as p<0.05, * and p<0.01, **.

Fig. 9. Increased astrocytosis observed in 24-month old Grn-/- mice

Grn-/- mice (A, D, G; n=4) have increased GFAP immunoreactivity (number, cell body size, and number and size of processes) as compared to age-matched Grn+/+ mice (B, E, H; n=4) in the (A-C) hippocampus, (D-F) cortex, and (G-I) thalamus. Scale bar 200 μm in (A, B), 100 μm in (D, E), and 200 μm in (G, H). Percent thresholding shown in C, F, and I (Grn-/- in white, Grn+/+ in black) as mean ± SEM. Statistics denoted as p<0.001, ***.