A role for glia in the progression of Rett’s syndromeDaniel T. Lioy1,2,3, Saurabh K. Garg1,2,3, Caitlin E. Monaghan1,2,3, Jacob Raber2,4,5, Kevin D. Foust6, Brian K. Kaspar6,Petra G. Hirrlinger7, Frank Kirchhoff8,9, John M. Bissonnette2,10,11, Nurit Ballas12 & Gail Mandel1,2,3
Rett’s syndrome (RTT) is an X-chromosome-linked autism spectrumdisorder caused by loss of function of the transcription factor methyl-CpG-binding protein 2 (MeCP2)1. Although MeCP2 is expressed inmost tissues2, loss of MeCP2 expression results primarily in neuro-logical symptoms1,3,4. Earlier studies suggested the idea that RTT isdue exclusively to loss of MeCP2 function in neurons2,4–10. Althoughdefective neurons clearly underlie the aberrant behaviours, we andothers showed recently that the loss of MECP2 from glia negativelyinfluences neurons in a non-cell-autonomous fashion11–13. Here weshow that in globally MeCP2-deficient mice, re-expression of Mecp2preferentially in astrocytes significantly improved locomotion andanxiety levels, restored respiratory abnormalities to a normal pat-tern, and greatly prolonged lifespan compared to globally null mice.Furthermore, restoration of MeCP2 in the mutant astrocytes exerteda non-cell-autonomous positive effect on mutant neurons in vivo,restoring normal dendritic morphology and increasing levels of theexcitatory glutamate transporter VGLUT1. Our study shows thatglia, like neurons, are integral components of the neuropathologyof RTT, and supports the targeting of glia as a strategy for improvingthe associated symptoms.
Global re-expression of Mecp2 postnatally in MeCP2-deficientmice allows normal longevity, rescues motor behaviours and improvesoverall health14. Because the expression of Mecp2 from the neuronaltau locus in early development prevents the appearance of severalRTT-like symptoms9, neurons are probably crucial components in a
rescue. However, previous in vitro studies indicate that astrocyticMeCP2 supports normal neuronal morphology11,12. Therefore, weasked whether astrocytes might also have a role in rescuing RTTneuropathology in vivo.
To this end, we crossed mice harbouring a tamoxifen (TAM)-inducible cre recombinase transgene driven by the human astrocyticglial fibrillary acidic protein (hGFAP) promotor15 (also see refs 16–18)with mice containing a Cre-excisable transcriptional Stop sequence inthe endogenous Mecp2 gene (Mecp2Stop)14. The progeny that inheritedboth alleles are referred to as Mecp2Stop-hGFAPcreT2 mice (Sup-plementary Fig. 1a). We determined the efficiency of astrocyticexcision in ROSA-reporter15 and Mecp2Stop/y-hGFAPcreT2 mice (Sup-plementary Fig. 1b–d, f, g). The percentage of MeCP21GFAP1 astro-cytes was extremely high in caudal brain regions, similar to that ofMecp21/y mice (Fig. 1a and Supplementary Fig. 1e). Re-expression ofMecp2 was not detected in oil-treated Mecp2Stop/y-hGFAPcreT2 mice(Supplementary Fig. 2a). Notably, only a very low percentage (,5%) ofexcision in neurons was detected by immunolabelling, polymerasechain reaction (PCR) analysis of the recombined Stop sequence, andsingle cell immunofluorescence intensity measurements (Fig. 1 andSupplementary Figs 1f, g and 3). This low percentage did not increasewith age (Supplementary Figs 1g and 4), and Mecp2 re-expression wasrestricted to brain (Supplementary Fig. 5c). Overexpression ofMecp2 in rescued astrocytes was not observed (SupplementaryFig. 2b).
1Vollum Institute, Oregon Health and Science University, Portland, Oregon 97239, USA. 2Oregon Health and Science University, Portland, Oregon 97239, USA. 3Howard Hughes Medical Institute. ChevyChase, Maryland 20815, USA. 4Departments of Behavioral Neuroscience and Neurology, Oregon Health and Science University, Portland, Oregon 97239, USA. 5Division of Neuroscience, Oregon NationalPrimate Research Center, Oregon Health and Science University, Beaverton, Oregon 97006, USA. 6Department of Pediatrics, The Ohio State University, Center for Gene Therapy, Nationwide Children’sHospital, Columbus, Ohio 43205, USA. 7Paul-Flechsig-Institute for Brain Research, Leipzig 04109, Germany. 8Department of Neurogenetics, Max Planck Institute of Experimental Medicine, Gottingen,Germany. 9Institute of Physiology, University of Saarland, Homburg 37075, Germany. 10Department of Cell and Developmental Biology, Oregon Health and Science University, Portland, Oregon 97202,USA. 11Department of Obstetrics and Gynecology, Oregon Health and Science University, Portland, Oregon 97202, USA. 12Department of Biochemistry and Cell Biology, State University of New York, StonyBrook, New York 11794, USA.
Figure 1 | MeCP2 is restored specifically in GFAP1 astrocytes of TAM-treated Mecp2Stop/y-hGFAPcreT2 mice. a, Efficiencies of Mecp2 re-expression. The numbers above the bars indicate total number of cells counted.b, Genomic PCR analysis of non-recombined (Stop; 4.3 kb) and recombinedamplicons (1.29 kb) of FACS-sorted NeuN1 and NeuN– cells from the wholebrain of a TAM-treated Mecp2Stop/y-hGFAPcreT2 mouse. Genomic DNA wasprepared from 500,000 cells per group. The wild-type (1.25 kb) Mecp2
amplicon is indicated. The b-actin promoter amplicon shows that similaramounts of DNA were present in the reactions. c, Fluorescence-intensityhistogram derived from individual hippocampal pyramidal neurons in tissuesections. Cy2 immunofluorescence intensities of nuclear MeCP2 protein areindicated above the line; DAPI fluorescence intensities of the same neurons areindicated below the line. ALU, arbitrary linear units. n 5 3 mice per genotypeand 100 cells per mouse.
2 8 J U L Y 2 0 1 1 | V O L 4 7 5 | N A T U R E | 4 9 7
To rule out the possibility that the small percentage of neurons, incombination with the low constitutive level of MeCP2 in the Stop mice(Supplementary Fig. 5b, c), might mediate any behavioural changes thatwe would measure, we systemically injected young male Mecp2Stop/y
mice with a suboptimal titre of recombinant MeCP2-AAV9 virus19 orvirus lacking MeCP2. This resulted in physiological levels of MeCP2expression in 2% to 35% of neurons, depending on the brain region(Supplementary Fig. 6a–c). Regardless of genotype, none of the treatedmice showed improvement of RTT-like phenotypes compared to thecontrol-AAV9-injected Mecp22/y mice (Supplementary Fig. 6d–g).Taken together, the results validate the use of the hGFAPcreT2 systemfor dissecting astrocytic contributions to RTT.
The average lifespan of oil-treated Mecp2Stop/y-hGFAPcreT2 andMecp2Stop/y mice was only 3 months14, which is prolonged comparedto Mecp22/y mice3,4 (Supplementary Fig. 5a), and probably due to thesmall amount of MeCP2 protein expressed from the Stop locus(Supplementary Fig. 5b, c). In contrast, nine of eleven TAM-treatedMecp2Stop/y-hGFAPcreT2 mice were alive at 7.5 months, when seven ofthe nine mice were killed for further analysis. The longest-lived mousewas killed at 15 months. The TAM-treated Mecp2Stop/y-hGFAPcreT2mice were also, on average, 20% larger than oil-treated Mecp2Stop/y-hGFAPcreT2 mice (Supplementary Fig. 7a). Using a previouslydescribed observational scoring system14, overall health of the TAM-treated male (Supplementary Fig. 7b) and female (Supplementary Fig. 8)mice stabilized, rather than worsened like the oil-treated controls, andTAM treatment of a highly symptomatic Mecp2Stop/y-hGFAPcreT2mouse reversed symptoms to nearly hGFAPcreT2 values (Supplemen-tary Fig. 9 and Supplementary Movies 1 and 2).
MeCP2-deficient mice are hypoactive1,3,4 and show altered mea-sures of anxiety-related behaviours6. In the home cage (Fig. 2a, b)and open field (Fig. 2c), oil-treated Mecp2Stop/y-hGFAPcreT2 micetravelled only ,20% the distance, and did so at ,20% the velocity,of hGFAPcreT2 control mice. TAM-treated Mecp2Stop/y-hGFAPcreT2mice, however, improved to ,50% the level of hGFAPcreT2 mice inboth measures. Similar improvements were observed in an open fieldtest to measure anxiety. The oil-treated Mecp2Stop/y-hGFAPcreT2 micespent only ,20% as much time in the centre of the cage as hGFAPcreT2mice, whereas TAM-treated Mecp2Stop/y-hGFAPcreT2 mice againimproved to ,50% the level of hGFAPcreT2 mice (Fig. 2d). The ratioof distance travelled in the centre square to total distance was the samefor all genotypes (data not shown). In the elevated zero and plus mazes,Mecp22/y mice consistently show decreased anxiety-related beha-viour20,21. The TAM-treated Mecp2Stop/y-hGFAPcreT2 mice were moreanxious than the oil-treated Mecp2Stop/y-hGFAPcreT2 mice in the ele-vated zero maze, improving up to ,50% the level of the controlhGFAPcreT2 mice (Fig. 2e).
RTT patients and mouse models have abnormal respiration1,22
(Fig. 3a). By 12 weeks, Mecp2Stop/y-hGFAPcreT2 mice had irregularityscores and apnoea rates significantly more severe than hGFAPcreT2controls (Fig. 3a, b and Supplementary Fig. 7c, traces 1 and 2). In con-trast, 2 months after TAM treatment, the respiratory pattern in 10 of 12Mecp2Stop/y-hGFAPcreT2 mice was within the normal range (Fig. 3b).Two mice followed over the subsequent 5-month period maintainedregular breathing patterns (data not shown). In two of three TAM-treated mice, we observed complete reversal to a normal respiratorypattern (Fig. 3a and Supplementary Fig. 7c, traces 3 and 4). The apnoeicfrequency in the third mouse was reduced but did not completely revertto control levels (Supplementary Fig. 7c, trace 5). The improvement inrespiration was consistent with efficient re-expression of Mecp2 inGFAP1 astrocytes within the pre-Botzinger complex of the brainstem,an area implicated in respiratory defects in RTT23 (SupplementaryFig. 7e). Treatment of a Mecp2Stop/y mouse with TAM did not alleviatethe irregular breathing or apnoeic frequency (Supplementary Fig. 7c,trace 2). Oil-treated female Mecp21/Stop-hGFAPcreT2 mice developeda significant number of apnoeas beginning at 4–6 months (Fig. 3c). Theapnoeic breathing was corrected by TAM treatment (Fig. 3c), even in themost severely affected female mouse (Supplementary Fig. 7d).
The brains of girls with RTT and affected mice exhibit smaller neur-onal somal sizes and reduced dendritic complexity in someregions1,4,24,25. At ,3.5 months of age, the somal sizes of neurons inhippocampus, cerebellum and cortex were still smaller in TAM-treatedMecp2Stop/y-hGFAPcreT2 mice compared to hGFAPcreT2 controls. At7 months, however, somal size was restored only in brain regions show-ing astrocytic re-expression of Mecp2 (Fig. 4a). Regarding dendriticcomplexity, the Mecp2Stop/y and oil-treated Mecp2Stop/y-hGFAPcreT2mice had ,25% less total number of apical dendrite branchescompared to controls. By 3.5 months of age, however, neurons inTAM-treated Mecp2Stop/y-hGFAPcreT2 mice had a normal number ofbranches and this was sustained with further age (Fig. 4b, c). MeCP2-deficient neurons also show deficits in proteins necessary for excitatoryneurotransmission, such as VGLUT1 (refs 26, 27). We detected ,20%less peri-nuclear VGLUT11 puncta in Mecp2Stop/y and oil-treatedMecp2Stop/y-hGFAPcreT2 mice compared to controls, but the levelsincreased to normal by 3–4 months of age with TAM treatment ofMecp2Stop/y-hGFAPcreT2 mice (Fig. 4d, e). Taken together, the ana-tomical findings indicate that re-expression of Mecp2 in astrocytes can,through a non-cell-autonomous mechanism, positively influence com-ponents of the neurotransmission machinery in vivo.
Our results show that re-expression of Mecp2 in astrocytes amelio-rates overt RTT-like phenotypes in mice. To address the complementaryquestion of the consequences of the removal of MeCP2 from astrocytes,we crossed mice with a floxed Mecp2 allele4 to the same hGFAPcreT2
a b c d
12
12
7
12
12
7
6
6
0
14
28
42
56
70
84
e
Tota
l d
ista
nce t
ravelle
d (m
)
0
***
** *
8
16
24
32
40
48
Tim
e s
pent
in c
entr
e (%
)
0
3.0
6.0
9.0
12.0
0
Avera
ge v
elo
city (m
m s
–1)
0
10
20
30
40
50
60
12
24
36
48
60
72hGFAPcreT2Mecp2Stop/y-hGFAPcreT2+TAM
Mecp2Stop/y-hGFAPcreT2+oil
***
** * ***
** *
***
** *
Tota
l d
ista
nce t
ravelle
d (m
)
Avera
ge v
elo
city (m
m s
–1)
17
13
17
13
6
***
** *
17
13
Tim
e s
pent
in o
pen (%
)
6
*****
4
6
80
70
60
50
40
30
20
10
0
Figure 2 | Partial behavioural rescue after re-expression of MeCP2 inastrocytes. a, Representative activity in a home-cage-like setting. Durationinterval, 5 min. Scale bars indicate 7 inches. b, Locomotor activity histograms ina home-cage-like setting. Duration interval, 10 min. c, Locomotor activity
histograms in an open field. Duration, 20 min. d, Time spent in centre of anopen field. e, Time spent in open portions of an elevated zero maze. Mice aged3–4 months. *P , 0.05, **P , 0.01, ***P , 0.001. All error bars indicate s.e.m.The number of mice analysed is indicated above each bar. b–e, Genotypes as in a.
RESEARCH LETTER
4 9 8 | N A T U R E | V O L 4 7 5 | 2 8 J U L Y 2 0 1 1
Mecp2+/Stop-hGFAPcreT2 + TAMMecp2Stop/y-hGFAPcreT2 + TAM
Mecp2Stop/y-hGFAPcreT2 + oil
Figure 3 | Restoration of normal breathing patterns after re-expression ofMecp2 in astrocytes. a, Representative plethysmographic recordings from afemale RTT patient (modified from ref. 22) and an Mecp2Stop/y-hGFAPcreT2mouse and control. The two middle traces are from the same Mecp2Stop/y-hGFAPcreT2 mouse before and 62 days after TAM treatment (Supplementary
Fig. 7c, trace 3). b, Respiratory irregularity scores and apnoea rates for malemice. c, Same as in b except for female mice. Mice showing at least 1 apnoea perhour were considered for apnoea rates. All error bars indicate s.e.m. *P , 0.05,**P , 0.01, ***P , 0.001. NS, not significant. The number of mice analysed isindicated above each bar.
0
60
80
100
120
3–4 months
98
141 7
1
79
87
91
140
177
90
90
71
64 63
65 64
96
112
***NS
NS
40
***
0
60
80
100
120 *** ***NS
0
60
80
100
120NS
***
40
20
40
20
ea c
Nis
sl
VG
LU
T1
VG
LU
T1/N
issl
d
NS
0
60
80
100
120
3–4 months 6 months3–4 months 7 months 15 months
***NSNS
NS
37
3140
41
75
54
89
6474
0
60
80
100
120
40
0
60
80
100
120
40
20 20
hGFAPcreT2 + TAM
20
40
b
20
So
ma d
iam
ete
r (%
co
ntr
ol)
Num
ber
of
bra
nches (%
co
ntr
ol)
Num
ber
of
VG
LU
T1
+ p
uncta
per
cell
(% c
ontr
ol)
Mecp2Stop/y-hGFAPcreT2 + TAM
Mecp2Stop/y-hGFAPcreT2
+ TAM
Mecp2Stop/y-hGFAPcreT2
+ oil
Mecp2Stop/y-hGFAPcreT2 + TAM Mecp2Stop/y-hGFAPcreT2 + oil
Mecp2Stop/y-hGFAPcreT2 + oil
Mecp2+/y + TAM Mecp2Stop/y + TAM
hGFAPcreT2+ TAM
hGFAPcreT2 + TAM
Cer
ebellar
Purkinjes C
A1
pyram
idals
Laye
r II/I
II
pyram
idals
Cer
ebellar
Purkinjes C
A1
pyram
idals
Laye
r II/I
II
pyram
idals
Figure 4 | Reversal of neuroanatomical abnormalities by re-expression ofMecp2 in astrocytes. a, Somal diameters of indicated neurons. Control,hGFAPcreT2 1 TAM. b, Representative traces of silver-impregnatedhippocampal CA1 neurons from male mice aged 3–4 months. c, Number ofsilver-impregnated CA1 apical branches in male mice. Control,
Mecp21/y 1 TAM. d, Representative images of Nissl-stained neuronsimmunolabelled for VGLUT1 from medulla oblongata. Scale bar: 10mm(b); 2mm (d). e, Number of VGLUT11 puncta associated with neuronal cellbodies from the medulla oblongata. All error bars indicate s.e.m. ***P , 0.001.NS, not significant. The number of analysed cells is indicated above each bar.
LETTER RESEARCH
2 8 J U L Y 2 0 1 1 | V O L 4 7 5 | N A T U R E | 4 9 9
(ref. 15) line used for the rescue. Recombination efficiencies throughoutthe brain were again higher in caudal compared to rostral regions(Supplementary Fig. 10a, b). The knockout progeny displayed somephenotypes shared with null Mecp2 mice, such as smaller body size,clasped hindlimb posture and irregular breathing (Supplementary Fig.10c, d, f), but their lifespans, locomotion (data not shown) and anxiety-related behaviours were all normal (Supplementary Fig. 10e).Furthermore, loss of MeCP2 from astrocytes did not affect the numberof CA1 apical dendritic branches (Supplementary Fig. 10g). This indi-cates that loss of MeCP2 in astrocytes at postnatal day 21 is unable todisrupt the already established hippocampal neuronal circuitry. In con-trast, loss of MeCP2 from astrocytes and gain of MeCP2 in astrocytesresulted in a strong non-cell-autonomous influence on breathingpattern. Thus, distinct neuronal–glia interactions may underlie hip-pocampal and hindbrain breathing circuitries.
Our results indicate that RTT involves impairments in both neuronsand glia. In familial amyotrophic lateral sclerosis28, neurons and gliaare proposed to have different roles in the disease process, with neu-rons primarily initiating the disease and astrocytes primarily affectingdisease progression. Our results are compatible with this model,because removal of MeCP2 just from astrocytes, at postnatal day 21,results in a subtler phenotype than the global null, and re-expression inastrocytes mainly stabilizes symptoms. Along these lines, the appear-ance of a subset of phenotypes after embryonic removal of MeCP2from subsets of neurons4,6–8,10,29 could be explained by causing diseaseinitiation, and prevention of RTT-like phenotypes after Mecp2 re-expression in embryonic neurons9 could be interpreted as preventingdisease initiation. None of these studies addresses whether it could takeboth MeCP2-deficient neurons and glia to cause disease progression,or whether other non-neuronal cell types, including other glia types,might be involved in the disease process.
Although impaired neurons ultimately underlie nervous system failurein RTT, restoring MeCP2in glia can ameliorate four consistent and robustfeatures of mouse models of RTT: premature lethality, aberrant respira-tion, hypoactivity and decreased dendritic complexity. Future studiesidentifying the key molecules that are restored after glia Mecp2 re-expres-sion may provide further clues into the mechanism of recovery, therebyproviding new potential targets for therapeutic intervention.
METHODS SUMMARYMale mice harbouring an hGFAPcreERT2 transgene were crossed to femaleMecp21/Stop or Mecp21/Jaenisch.Flox mice and the F1 progeny were injected with100 mg kg21 tamoxifen or oil when appropriate. Mice used in astrocyte-rescueexperiments were backcrossed for at least seven generations to a C57BL/6 back-ground. Mice used in astrocyte-knockout experiments were of a FVB/N/C57BL/6background. Histology was performed on transcardially perfused, frozen sections.Behaviour was analysed using CleverSystems StereoScan software. Body plethys-mography was performed on unanaesthetized restrained mice. Statistics wereperformed with Graphpad PRISM V5.0C software. Mouse maintenance, breedingand genotyping, tamoxifen treatments, phenotype scoring, tissue preparation andimmunohistochemistry, fluorescence intensity measurements, FACS, plethysmo-graphy, motor activity and anxiety assessments, western blotting, and statisticswere performed as described in Methods.
Full Methods and any associated references are available in the online version ofthe paper at www.nature.com/nature.
Received 5 April; accepted 18 May 2011.
Published online 29 June 2011.
1. Chahrour, M. & Zoghbi, H. Y. The story of Rett syndrome: from clinic toneurobiology. Neuron 56, 422–437 (2007).
2. Shahbazian, M. D., Antalffy, B., Armstrong, D. L. & Zoghbi, H. Y. Insight into Rettsyndrome:MeCP2 levels display tissue- and cell-specific differencesand correlatewith neuronal maturation. Hum. Mol. Genet. 11, 115 (2002).
3. Guy, J., Hendrich, B., Holmes, M., Martin, J. E. & Bird, A. A mouse Mecp2-nullmutationcausesneurological symptoms thatmimicRett syndrome. NatureGenet.27, 322–326 (2001).
4. Chen, R. Z., Akbarian, S., Tudor, M.& Jaenisch, R. Deficiency of methyl-CpG bindingprotein-2 in CNS neurons results in a Rett-like phenotype in mice. Nature Genet.27, 327–331 (2001).
5. Kishi, N. & Macklis, J. D. MECP2 is progressively expressed in post-migratoryneuronsand is involved inneuronal maturation rather thancell fatedecisions.Mol.Cell. Neurosci. 27, 306–321 (2004).
6. McGill, B. E. et al. Enhanced anxiety and stress-induced corticosterone release areassociated with increased Crh expression in a mouse model of Rett syndrome.Proc. Natl Acad. Sci. USA 103, 18267–18272 (2006).
7. Fyffe, S. L.et al. DeletionofMecp2 inSim1-expressing neurons reveals a critical rolefor MeCP2 in feeding behavior, aggression, and the response to stress. Neuron 59,947–958 (2008).
8. Samaco, R. C. et al. Loss of MeCP2 in aminergic neurons causes cell-autonomousdefects in neurotransmitter synthesis and specific behavioral abnormalities. Proc.Natl Acad. Sci. USA 106, 21966–21971 (2009).
9. Luikenhuis, S., Giacometti, E., Beard, C. F. & Jaenisch, R. Expression of MeCP2 inpostmitotic neurons rescues Rett syndrome in mice. Proc. Natl Acad. Sci. USA 101,6033–6038 (2004).
10. Chao, H. T. et al. Dysfunction in GABA signalling mediates autism-like stereotypiesand Rett syndrome phenotypes. Nature 468, 263–269 (2010).
11. Ballas, N., Lioy, D. T., Grunseich, C. & Mandel, G. Non-cell autonomous influence ofMeCP2-deficient glia on neuronal dendritic morphology. Nature Neurosci. 12,311–317 (2009).
12. Maezawa, I., Swanberg, S., Harvey, D., LaSalle, J. M. & Jin, L. W. Rett syndromeastrocytes are abnormal and spread MeCP2 deficiency through Gap junctions.J. Neurosci. 29, 5051–5061 (2009).
13. Maezawa, I. & Jin, L. W. Rett syndrome microglia damage dendrites and synapsesby the elevated release of glutamate. J. Neurosci. 30, 5346–5356 (2010).
14. Guy, J., Gan, J., Selfridge, J., Cobb, S. & Bird, A. Reversal of neurological defects in amouse model of Rett syndrome. Science 315, 1143–1147 (2007).
15. Hirrlinger, P.G., Scheller, A., Braun, C., Hirrlinger, J. & Kirchhoff, F. Temporal controlof gene recombination in astrocytes by transgenic expression of the tamoxifen-inducible DNA recombinase variant CreERT2. Glia 54, 11–20 (2006).
16. Brenner, M., Kisseberth, W. C., Su, Y., Besnard, F. & Messing, A. GFAP promoterdirects astrocyte-specific expression in transgenic mice. J. Neurosci. 14,1030–1037 (1994).
17. Chow, L., Zhang, J. & Baker, S. J. Temporal control of gene recombination inastrocytes by transgenic expression of the tamoxifen-inducible DNA recombinasevariant CreERT2. Transgenic Res. 17, 919–928 (2008).
18. Casper, K. B., Jones, K. & McCarthy, K. D. Characterization of astrocyte-specificconditional knockouts. Genesis 45, 292–299 (2007).
19. Foust, K. D. et al. Intravascular AAV9 preferentially targets neonatal neurons andadult astrocytes. Nature Biotechnol. 27, 59–65 (2008).
20. Stearns, N. A. et al. Behavioral and anatomical abnormalities in Mecp2 mutantmice: A model for Rett syndrome. Neuroscience 146, 907–921 (2007).
21. Pelka, G. J. et al. Brain Mecp2 deficiency is associated with learning and cognitivedeficits andalteredgeneactivity in thehippocampal regionofmice.129,887–898(2006).
22. Weese-Mayer, D. E. et al. Autonomic nervous system dysregulation: breathing andheart rate perturbation during wakefulness in young girls with Rett syndrome.Pediatr. Res. 60, 443–449 (2006).
23. Viemari, J. et al. Mecp2 deficiency disrupts norepinephrine and respiratorysystems in mice. J. Neurosci. 25, 11521–11530 (2005).
24. Bauman, M. L., Kemper, T. L. & Arin, D. M. Pervasive neuroanatomic abnormalitiesof the brain in three cases of Rett’s syndrome. Neurology 45, 1581–1586 (1995).
25. Armstrong, D. D. Neuropathology of Rett syndrome. J. Child Neurol. 20, 747–753(2005).
26. Chao, H. T., Zoghbi, H. Y. & Rosenmund, C. MeCP2 controls excitatory synapticstrength by regulating glutamatergic synapse number. Neuron 56, 58–65 (2007).
27. Marchetto, M. L. A model for neural development and treatment of Rett syndromeusing human induced pluripotent stem cells. Cell 143, 527–539 (2010).
28. Ilieva, H., Polymenidou, M. & Cleveland, D. W. Non-cell autonomous toxicity inneurodegenerative disorders: ALS and beyond. J. Cell Biol. 187, 761–772 (2009).
29. Gemelli, T. et al. Postnatal loss of methyl-CpG binding protein 2 in the forebrain issufficient to mediate behavioral aspects of Rett syndrome in mice. Biol. Psychol.59, 468–476 (2006).
Supplementary Information is linked to the online version of the paper atwww.nature.com/nature.
Acknowledgements We thank P. Brehm, R. H. Goodman, C. Bond, M. McGinley andC.Mandel-Brehm for discussions;P.Micha, J. Eng, S. Knopp and T. Shaffer for technicalsupport; and M. Murtha for generating the CBA/CMV-MECP2 construct. ViraPur, LLCgenerated the AAV9 virus. The work was supported by grants from the NationalInstitutes of Health (G.M. and N.B.), International Rett Syndrome Foundation (N.B. andJ.M.B.), Rett Syndrome Research Trust (G.M. and B.K.K.), Oregon Brain Institute (D.T.L.),and OHSU Cell and Developmental Biology Training Program (D.T.L.). G.M. is anInvestigator of the Howard Hughes Medical Institute.
Author Contributions D.T.L., S.K.G., J.R., J.M.B., N.B. and G M. designed the astrocyteknockout and rescue experiments. B.K.K. and K.D.F. helped design the AAV9experiments.D.T.L., S.K.G.,C.E.M. andJ.M.B. performedtheexperiments.P.G.H. andF.K.provided the hGFAPcreT2 transgenic mice. D.T.L., S.K.G., N.B. and G.M. wrote themanuscript with input from the other co-authors.
Author Information Reprints and permissions information is available atwww.nature.com/reprints. The authors declare no competing financial interests.Readers are welcome to comment on the online version of this article atwww.nature.com/nature. Correspondence and requests for materials should beaddressed to G.M. ([email protected]).
RESEARCH LETTER
5 0 0 | N A T U R E | V O L 4 7 5 | 2 8 J U L Y 2 0 1 1
METHODSAll animal studies were approved by the Oregon Health and Science UniversityInstitutional Animal Care and Use Committee.Maintenance, breeding and genotyping of mice. Mice were group housed withlittermates in standard housing on a 12:12 h light:dark cycle. For rescue experi-ments, hGFAPcreT2 mice were backcrossed for eight generations to the C57BL/6strain. hGFAPcreT2 mice used for knockout experiments were on a FVB/N/C57BL/6 background. Mecp2Stop (ref. 14) and Mecp2Bird.knockout (Mecp2B.Null)(ref. 3) mice were obtained from Jackson Laboratories and were also on aC57BL/6 background. Mecp2Jaenisch.Flox (Mecp2J.Flox) (ref. 4) mice were obtainedfrom the Mutant Mouse Medical Resource Center at University of California,Davis and were also on a C57BL/6 background. Male hemizygous hGFAPcreT2mice were crossed to female Mecp21/Stop mice to yield male and female Mecp2Stop-hGFAPcreT2, Mecp2Stop, Mecp21/y and hGFAPcreT2 genotypes. The floxed Stopsequence was identified from tail biopsies using the following primers: common59-AACAGTGCCAGCTGCTCTTC-39, WT 59-CTGTATCCTTGGGTCAAGCTG-39, and mutant 59-GCCAGAGGCCACTTGTGTAG-39. The hGFAPcreT2sequence was identified with the following primers: 59-CAGGTTGGAGAGGAGACGCATCA-39, 59-CGTTGCATCGACCGGTAATGCAGGC-39.Note that this primer set is specific for the hGFAPcre locus and does not recognizeother cre loci. The Jaenich floxed Mecp2 genotype for astrocyte knockout wasidentified using primers 59-CACCACAGAAGTACTATGATC-39 and 59-CTAGGTAAGAGCTCTTGTTGA-39.Tamoxifen treatments. TAM (Sigma) was made fresh weekly by dissolving in90% sunflower seed oil/10% ethanol solution by bath sonication for 20–30 min at4 uC with intermittent vortexing. Final concentration of TAM was 20 mg ml21.Three-to-four-week-old mice were injected intraperitoneally with 100 mg kg21
with TAM or oil once daily for 8 days. Occasionally mice began to show signsof discomfort during TAM treatments (that is, decreased mobility, increasedtremors, dehydration, rough coat, or gasping). In that case, the remaining TAMtreatments were given once every 3–4 days until all mice received 8 injections. Thelongest interval over which TAM was given was 15 days. Additional TAM treat-ments to symptomatic mice did not rescue survival.Phenotype scoring. Mice were removed from their home cage and placed onto ametal laminar flow hood for observation.
For mobility: 0 5 as wild type; 1 5 reduced movement when compared to wildtype, with extended freezing periods or extended delay to movement when firstplaced on the surface; 2 5 lack of spontaneous movement when placed on thesurface.
For gait: 0 5 as wild type; 1 5 hindlimbs spread wider than wild type whenambulating and/or a lowered pelvis when ambulating; 2 5 lack of full strides byhindlimbs resulting in a dragging of hindquarters.
For limb posture: 0 5 hindlimbs splay outward when suspended by the tail;1 5 one hindlimb is pulled into the body or forelimbs are stiff and splayed outwardwithout motion; 2 5 one hindlimb is pulled into the body and forelimbs are stiffand splayed outward without motion and might form a widened bowl shape orboth hindlimbs are pulled into the body with or without abnormal forelimbposture.
For tremor: 0 5 no tremor; 1 5 intermittent mild tremor; 2 5 continuous tremoror intermittent violent tremor.
For general condition: 0 5 shiny coat, clear and opened eyes, normal bodystance; 1 5 dull or squinty eyes, dull or ungroomed coat, somewhat hunchedstance; 2 5 piloerection, hunched stance.Tissue preparation, immunohistochemistry, cell counts and neuronal somameasurements. Mice were anaesthetized by intraperitoneal injection of Avertin(2-2-2 Tribromoethanol) and killed by transcardial perfusion of 4% parafomalde-hyde in phosphate-buffered saline. Brains were post-fixed overnight and thenequilibrated in 30% sucrose overnight at 4 uC. Sagittal sections (40mm) were cutat –20 uC using a cryostat (Leica) and stored at –20 uC. Sections were immunola-belled overnight at 4 uC using the following primary antibodies: rabbit-MeCP2(1:400, Covance), mouse-GFAP (1:400, Abcam), chicken-GFAP (1:400, Abcam),mouse-NeuN (1:200, Millipore), goat-somatostatin (1:200, Santa Cruz), rabbit-EGFP (1:100, Millipore), sheep-VGLUT1 (1:200, Abcam). Nissl staining (at either594 nm or 640 nm) was performed as instructed by the manufacturer(NeuroTrace, Invitrogen).
Appropriate Alexa Fluor secondary antibodies (1:500, Molecular Probes) orCy5 were used for 1 h at room temperature. DAPI was present in the ProLongGold Antifade (Invitrogen) mounting reagent. All images were collected on a Zeissconfocal laser scanning LSM 510 microscope and an Olympus confocal laserscanning FW1000 microscope.
MeCP2 expressing cells were identified as follows: nuclei of astrocytes (GFAP1
at 594 nm or 640 nm) and neurons (NeuN1 at 594 nm; somatostatin1 at 594 nm
or 640 nm; GFAP2 at 640 nm/Nissl1 at 594 nm) were first identified by DAPIstaining. Cells with clearly identified nuclei were then assessed for MeCP2expression by analysing 505 nm signal (excitation: 488 nm) in the nucleus.MeCP2 antibody specificity was previously confirmed11 and re-confirmed byimmunostaining and western blot of samples taken from male Mecp2Bird.knockout
mice3. Cell counts are expressed as the percentage of total astrocytes or neuronalpopulations in specific brain regions that are MeCP21.
Somal diameters of Nissl-stained neurons were determined by averaging thelengths of the long and short axes across the cell body. Long and short axes wereperpendicular to each other. Only cells with a clearly visible DAPI-stained nucleuswere considered. Every fourth serial section was used. Only after all cell diameterswere collected was the genotype of each section revealed to the experimenter.
Golgi staining was performed using the FD Rapid GolgiStaining Kit accordingto the manufacturer’s instructions (FD NeuroTechnologies, catalogue numberPK401). Tissue was vibratome sectioned at 200mm. Hippocampal CA1 pyramidalneuron apical branches were analysed using an inverted bright-field microscope at320 magnification by two separate experimenters blind to the genotypes.
Neuronal soma VGLUT11 puncta were counted in the medulla oblongataunder 363 magnification. Only neuronal somas showing VGLUT11 stainingwere considered. The experimenter was blind to the tissue genotypes.Fluorescence intensity measurements. Cells with clear nuclei were identified byDAPI fluorescence. MeCP2 signal for only these cells were considered. MeCP2signal for this analysis was not amplified. Rather, a Cy2 secondary antibody wasused (collected at 505 nm), directed directly against the primary anti-MeCP2antibody. All images were captured using an AxioCam HRc (Zeiss) at exactlythe same exposure. Raw pixel intensities associated with the DAPI and MeCP2signals were measured separately in Photoshop. The genotypes of the data wererevealed to the experimenter only after all data were collected and analysed.FACS and semi-quantitative PCR. Whole brains were dissected from 6–8-week-old mice and tissue was minced in small pieces in pre-cooled dissociation medium(80 mM Na2SO4, 30 mM K2SO4, 0.25 mM CaCl2, 20 mM glucose, 10 mM MgCl2,0.001% phenol red and 10 mM HEPES pH 7.5). The tissue was dissociated inmedium containing 40 U ml21 papain (Worthington) for 45 min at 37 uC. Thetissue was washed twice in dissociation buffer before transferring to deactivationbuffer (DMEM, 0.5 mg ml21 DNase I and 10% FBS). Sequential trituration wascarried out using 10-, 5- and 1-ml pipette tips. Debris was allowed to settle for2 min. Supernatant was filtered through a 40mm cell strainer before cells wereharvested at 1,000 r.p.m. for 10 min at 4 uC and re-suspended in Dulbecco’s PBS(DPBS). To fix, cells were treated with 1% formaldehyde for 15 min at 25 uC. Cellswere washed twice with DPBS and then permeabilized in buffer (PBS, 0.2% TritonX-100 and 10% FBS) for 30 min at 25 uC. To identify the NeuN1 cells in thepreparation, cells were probed with anti-mouse NeuN antibody for 30 min at25 uC. Preparation was probed with anti-mouse IgG-Alexa-488 secondary anti-body (Invitrogen). Cells were washed twice with PBS containing 0.2% TritonX-100 before re-suspending in DPBS. To sort the cells using FACS, cells wereagain passed through a 40mm filter and subjected to FACS. The sorted cells werecollected at 8,000 r.p.m. for 10 min and genomic DNA was prepared from NeuN1
and NeuN2 fractions using the QIAamp DNA kit (Qiagen). Genomic PCR for theMecp2 locus was carried out using the following oligonucleotides: forwardMECP2-U2 59-GTTCAGAATCAGGGGAGCAGCCC-39 and reverse upexIII-R3 59-CCTTGGGTCAAGCTGGGGCC-39. For genomic PCR of the b-actin pro-moter, the following oligonucleotides were used: forward 59-CCCAACACACCTAGCAAATTAGAACCAC-39 and reverse 59-CCTGGATTGAATGGACAGAGAGTCACT-39. PCR products were analysed on a 1% ethidium-bromide-stainedagarose gel.Plethysmography. Respiratory parameters were determined in a body plethys-mograph. Individual unanaesthetized animals were placed in a 65-ml chamberwith their head exposed through a close-fitting hole in parafilm. A pneumotacho-graph was connected to the chamber and a differential pressure transducer (ModelPT5A, Grass Instrument). The pressure signal was integrated to give tidal volume.Volume changes were calibrated by injecting known amounts of air into thechamber. The analogue signal from the transducer was amplified, converted todigital, displayed on a monitor, and stored to disk by computer for later analysis.Apnoea was defined as an expiratory time of 1.0 s or greater. Irregularity score wasdetermined from: absolute (TTOTn 2 TTOTn 1 1)/(TTOTn 1 1).Motor activity and anxiety assessment. Motor activity and anxiety tests werecarried out at the same time of day (12.00 to 18.00) and in the same dedicatedobservation room. Mice were placed singly into an observation box, which wasakin to a new home cage, for a total of 20 min, or a standard open field box for20 min with side-viewing and top-viewing cameras (Clever Systems), or an ele-vated Zero maze for 5 min with top-viewing cameras (Clever Systems). Mice wereallowed to acclimatize to the observation box for the first 10 min and the next10 min of recording was analysed on a Dell computer. Activity traces were
acquired in real time using StereoScan Software (Clever Systems). The mice couldnot see the experimenter during recordings. Mice were never tested in the threearenas on the same day.Western blot. Mice were killed by decapitation, and brains immediately isolatedand homogenized on ice in nuclear lysis buffer containing 2-mercaptoethanol.Lysates were boiled for 5 min and separated on a denaturing 10% acrylamide gel.We used an antibody to MeCP2 as described above and a mouse-tubulin antibody(Sigma). Horseradish peroxidase-conjugated secondary antibodies were used andchemically activated with the Western Lighting Chemiluminescent System(PerkinElmer Life Sciences).AAV9 production and injections. AAV9 was produced by transient transfectionprocedures using a double-stranded AAV2-ITR based vector system as previously
described19. MeCP2 expression was driven from a chicken-b-actin promoter withCMV enhancer. AAV9 virus was titred by quantitative PCR, and stored aspreviously described19. MeCP2-AAV9 or empty AAV9 (control-AAV9) wasinjected via the tail vein at 1 3 1012 viral particles in a volume of 300ml. Injectedmice included symptomatic Mecp2Stop/y, Mecp2B.Null/y, or Mecp2J.Null/y micebetween 4 and 8 weeks old.Statistics. All behaviour tests were analysed using two-way ANOVAs followed,when appropriate (P , 0.05), by Newman–Keuls post-hoc test. Soma size mea-surements were analysed using unpaired two-tailed t-tests. All other morpho-logical measurements were analysed using two-way ANOVAs followed, whenappropriate (P , 0.05), by Tukey’s post-hoc test. Statistical analyses were per-formed using PRISM software.
