Huntingtin acts non cell-autonomously on hippocampal neurogenesis and controls anxiety-related behaviors in adult mouse

Abstract

Huntington's disease (HD) is a fatal neurodegenerative disease, characterized by motor defects and psychiatric symptoms, including mood disorders such as anxiety and depression. HD is caused by an abnormal polyglutamine (polyQ) expansion in the huntingtin (HTT) protein. The development and analysis of various mouse models that express pathogenic polyQ-HTT revealed a link between mutant HTT and the development of anxio-depressive behaviors and various hippocampal neurogenesis defects. However, it is unclear whether such phenotype is linked to alteration of HTT wild-type function in adults. Here, we report the analysis of a new mouse model in which HTT is inducibly deleted from adult mature cortical and hippocampal neurons using the CreER(T2)/Lox system. These mice present defects in both the survival and the dendritic arborization of hippocampal newborn neurons. Our data suggest that these non-cell autonomous effects are linked to defects in both BDNF transport and release upon HTT silencing in hippocampal neurons, and in BDNF/TrkB signaling. The controlled deletion of HTT also had anxiogenic-like effects. Our results implicate endogenous wild-type HTT in adult hippocampal neurogenesis and in the control of mood disorders.

Figure 1. Tamoxifen injection in CaMKCreER ^T2 mice activates Cre expression in mature but not newborn hippocampal neurons.

(A) Design of experiments performed on CaMKCreER ^T2 ;ROSA26R, CaMKCreER ^T2 ; Hdh ^flox/flox and WT; Hdh ^flox/flox mice. (B) Whole mount X-gal staining of a parasagittal brain section of a CaMKCreER ^T2 ;ROSA26R mouse injected with tamoxifen. Scale bar: 500 µm. (C) Immunofluorescence in the hippocampal region of tamoxifen-injected CaMKCreER ^T2 ;ROSA26R mouse with antibodies recognizing β-galactosidase, DCX and NeuN. Scale bar for upper panel: 30 µm ; scale bar for lower panel: 10 µm. (D) Representative Western-blot of hippocampal proteins extracted from mutant and control mice 6 months after tamoxifen injection and incubated with anti-Htt (4C8) and anti-actin antibodies. Data are the mean +/- SEM of the ratios Htt/actin, normalized so that the mean value of controls is equal to 1 (n= 4-5 per group). * p<0.05 (E) Hippocampi of mutant mice 6 months after tamoxifen injection were co-immunostained with antibodies recognizing GFAP or DCX and HTT (4C8). White arrows indicate HTT expression in DCX positive neurons. Scale bar for upper panel : 10 µm; scale bar for lower panel : 8 µm.

Figure 2. Deletion of huntingtin in mature neurons alters survival of newborn hippocampal neurons.

All mice were injected with tamoxifen 6 months before sacrifice. (A) BrdU was injected 2 h before sacrifice into mutant and control mice. Immunofluorescence was performed with an anti-BrdU antibody. Data are the mean +/- SEM of the positive BrdU positive cell counts in DG from 3–4 brains per group. (B) BrdU was injected 21 days before sacrifice into mutant and control mice. Immunofluorescence was performed as in A. Data are the mean +/- SEM of the positive BrdU positive cell counts in DG from 3 brains per group. (C) BrdU was injected 42 days before sacrifice into mutant and control mice. Immunofluorescence was performed as in A. Data are the mean +/- SEM of the positive BrdU positive cell counts in DG from 3–4 brains per group. * p<0.05 ** p<0.01. (D) BrdU (green) and NeuN (red) immunofluorescence in DG of control and mutant mice injected with BrdU 42 days before sacrifice into mutant and control mice. Scale bar : 50 µm.

Figure 3. Deletion of huntingtin in mature neurons alters dendritic arborization of newborn hippocampal neurons.

(A) Representative images of DCX immunostained slices in the upper blade of DG of control and mutant mice 6 months after tamoxifen injection. Scale bar : 50 µm. (B) Sholl analysis was performed on 12 randomly chosen DCX positive neurons in DG for each control and mutant mice (n=4 mice per group). (C) The total length of the dendrites of 12 randomly chosen DCX positive neurons in DG for each control and mutant mice was measured using NeuroLucida software (n=4 mice per group). * p<0.05 ** p<0.01 *** p<0.005.

Figure 4. Deletion of huntingtin in mature neurons alters hippocampal Akt and Erk phosphorylation without affecting BDNF production.

(A) Representative Western-blots of hippocampal proteins extracted from mutant and control mice 6 months after being injected with tamoxifen and incubated with phospho-ERK, ERK, phospho-Akt, Akt, BDNF or actin recognizing antibodies. (B–D) Data are the mean +/- SEM of the ratios phospho-ERK/ERK, phospho-Akt/Akt, mature BDNF/actin obtained by Western-blot densitometric analysis, normalized so that the mean value of controls is equal to 1 (n= 4-5 per group). * p<0.05 ** p<0.01 *** p<0.005.

Figure 5. Downregulation of huntingtin expression alters BDNF vesicular trafficking and downregulates BDNF secretion in hippocampal neurons.

(A,B) Rat primary hippocampal neurons were electroporated with BDNF-mCherry and scrambled RNA (scRNA) or siRNA targeting Htt (siHtt) or shRNA targeting luciferase (shLuc) as a control or shRNA targeting Htt (shHtt). Three days later, proteins were extracted and Western-blots were incubated with anti-Htt (4C8) and anti-tubulin antibodies. Data are the mean +/- SEM of 3 independent experiments per group, normalized to the value obtained for the scHtt group or the shLuc group. (C, D) Three days after electroporation with the constructs described above, the movements of BDNF containing vesicles were observed by videomicroscopy and analyzed with kymographs. Data are the mean +/- SEM of the anterograde velocity (C) or percentage of time a vesicle is pausing (velocity less than 0.05 µm/s) (D) in 18-27 neurons per group analyzed in 3 independent experiments. * p<0.05 ** p<0.01 *** p<0.005 (E) Representative kymographs of BDNF-mCherry vesicles in primary hippocampal neurons expressing shLuc or shHtt. (F) Rat primary hippocampal neurons were electroporated with BDNF and scrambled RNA (scRNA) or siRNA targeting Htt (siHtt) or shRNA targeting luciferase (shLuc) as a control or shRNA targeting Htt (shHtt). Three days later, a first depolarization allowed the cellular release of BDNF in the medium (K1) and depleted the internal BDNF vesicular store. After 30 min of recovery, a second depolarization was applied and a new sample of the medium was isolated (K2). Concentrations of BDNF in K1 and K2 samples were determined by ELISA. Data are the mean +/- SEM of 3 independent experiments per group, normalized to the value obtained for the scHtt group. * p<0.05 ** p<0.01 *** p<0.005.

Figure 6. Deletion of huntingtin in mature cortical and hippocampal neurons triggers modifications of anxiety-related behaviors.

(A) Depressive-like behavior was assessed by the forced swim test (FST) where mice are placed into plastic buckets filled with water and the immobility durations are measured. (B–C) Mixed anxio-depressive behavior was assessed by the NSF test where starving mice are placed in a corner of a plastic box containing a pellet of food in the center. The latency to feed is timed. The data are presented with means +/- SEM and as a percentage of mice that did not feed per group during the 10 min of the test. (D–E) Anxiety-related behavior was assessed by placing mice in the center of a Plexiglas open field box, and time and number of entries in the center were measured during 30 minutes. (F) In the previous open field test, the distance travelled in the center and the total distance travelled were measured and the percentage of distance travelled in the center over the total distance travelled was calculated. (G, H) Anxiety-like behaviour was assessed by placing mice in the center of an elevated plus maze (EPM) and time and the number of entries in open arms were measured during 5 minutes. For all the tests, n=12-14 per group. * p<0.05 ** p<0.01 *** p<0.005 for comparisons between genotypes.