Fig 1.
Emergence of GFP-expressing PCs in SCA1 mouse cerebella after transplantation of TRE-GFP-hfMSCs expressing GFP through the Tet-off system.
(a) The TRE-GFP gene was introduced into hfMSCs through a lentivirus (1). The transduced TRE-GFP-hfMSCs were injected into the cerebella of 6–8-month-old SCA1 mice (2). Two weeks after the TRE-GFP-hfMSC injection, the SCA1 mice received an injection of AAV9-L7-HA-mtTA vectors expressing HA-tagged mtTA under the control of a PC-specific L7 promoter (3). (b) Schema depicting GFP expression in PCs expressing mtTA. TRE in the nucleus of TRE-GFP-hfMSCs initiated the transcription of the downstream GFP gene in the presence of mtTA, which was produced only in AAV9-L7-HA-mtTA-infected PCs. (c, d) Immunohistochemistry of cerebellar slices from SCA1 mice treated as shown in the schema (a). The mice were sacrificed 10 weeks after the hfMSC grafting. The slices were double immunolabeled for HA and GFP. HA-tagged mtTA (magenta) was efficiently expressed only in PCs, and some PCs co-expressed GFP (green). (e, f) Enlarged images of PCs immunostained for GFP. Scale bars, 200 μm (c) and 50 μm (d-f).
Fig 2.
Validation of the FLEx-Tet system in mouse cerebella in vivo.
(a) The AAV9-SynImCMV-HA-mtTA-P2A-Cre vector bicistronically expressed HA-tagged mtTA and Cre recombinase under the control of a neuron-specific SynImCMV promoter (1). P2A, which is a ‘self-cleaving’ peptide sequence, was inserted between mtTA and Cre, whereas lentiviral vectors (Lenti-TRE-FLEx-GFP) carried TRE and an inverted GFP sequence flanked by loxP and lox2272 on both sides (2). AAV9 and/or lentiviral vectors were injected into 4-week-old mice. (b) Diagram depicting the FLEx-Tet system. Cre-mediated recombination and inversion of the GFP gene permitted expression of GFP protein in the presence of mtTA. (c-e) Immunohistochemistry of the mice that received an injection of the AAV9-SynImCMV-HA-mtTA-P2A-Cre vector (c), Lenti-TRE-FLEx-GFP vector (d) or the viral mixture (e). Two weeks after the viral injection, the mice were sacrificed, and the cerebellar slices were triple immunostained for GFP, HA and calbindin. Notably, GFP was expressed only in the cerebella of mice that received an injection of the viral mixture. (f) Magnified images of GFP-expressing lobules showing expression of GFP in various types of cortical neurons, including PCs, interneurons and granule cells. Scale bar, 50 μm.
Fig 3.
GFP expression in PCs and putative interneurons after grafting hfMSCs carrying the inverted GFP gene to the SCA1 mouse cerebellum.
(a) Schema depicting the FLEx-Tet system, which permitted us to explore the fusion of hfMSCs with cerebellar neurons. An AAV9-SynImCMV-HA-mtTA-P2A-Cre vector expressing HA-tagged mtTA and Cre through a neuron-specific SynImCMV promoter was injected into the cerebella of 6–8-month-old SCA1 mice (1). A sequence with TRE and inverted GFP genes flanked by loxP and lox2272 was inserted into the genome of hfMSCs by using Lenti-TRE-FLEx-GFP vectors (2). Thus, GFP protein was never produced without Cre recombinase, and the expression increased drastically after co-provision with mtTA. The TRE-FLEx-GFP-hfMSCs (50,000 cells) were then injected into the SCA1 mice 2 weeks after injection of AAV9-SynImCMV-HA-mtTA-P2A-Cre vectors (3). (b-g) Emergence of GFP-expressing neurons in the cerebella of mice that were sacrificed 2 weeks after the TRE-FLEx-GFP-hfMSC injection. The cerebellar slices were immunolabeled for GFP and calbindin, a PC marker. GFP expression was detected in PCs (b, e-g) and cells in the molecular layer (c, d). Arrows in (e-g) indicate a GFP-expressing PCs co-immunostained for calbindin. Because we used a neuron-specific synapsin I promoter with a minimal CMV sequence, Cre and mtTA should have been expressed specifically in neurons, and thus the GFP-expressing cells in the molecular layer (c, d) were presumed to be basket cells (c) and stellate cells (d). Scale bar, 50 μm.
Table 1.
Summary of hfMSC transplantation into the cerebellum of SCA1 mice.
Table 2.
Summary of mice that expressed GFP-labeled cells in the cerebellum.
Fig 4.
A summary graph showing the percent number of mice that yielded GFP-positive cells after the hfMSC grafting to the cerebellum.
Four-week-old and 6–8-month-old wild-type (WT) and SCA1 mice that were treated with the Tet-off system (Tables 1 and 2) were analyzed. All mice were incubated more than 2 weeks and, therefore, those that did not show GFP (+) cells at the time of microscopic observation were regarded as fusion-negative mice. Number on each bar shows number of mice with GFP (+) cells / that of mice examined. Fisher’s exact test showed significantly higher emergence of GFP-positive cells in 6–8-month-old SCA1 mice than 4-week-old and 6–8-month-old wild-type mice and 4-week-old SCA1 mice. **p<0.01, ***p<0.001.
Fig 5.
Degeneration of PC and molecular layer interneurons.
(a, b) Fluorescence images of the cerebellar cortex immunostained for parvalbumin from a 6-month-old wild-type mouse (WT) (a) and an age-matched SCA1 mouse (B05) (b). Black and white arrows show examples of PCs and interneurons, respectively. Scale bar, 50 μm. (c, d) Graphs showing the size of cell bodies of interneurons (WT; n = 110 cells from 3 mice, B05; n = 140 cells from 4 mice) (c) and of PCs (WT; n = 81 PCs from 3 mice, B05; n = 80 PCs from 4 mice) (d). Asterisks indicate a statistically significant difference between the wild-type mice and SCA1 mice; ***p<0.001 by unpaired t-test.