-
Cheah et al., Sci. Adv. 2021; 7 : eabb1703 8 January 2021
S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L
E
1 of 13
D I S E A S E S A N D D I S O R D E R S
Gene therapy for tuberous sclerosis complex type 2 in a mouse
model by delivery of AAV9 encoding a condensed form of
tuberinPike-See Cheah1,2*, Shilpa Prabhakar1*, David Yellen1,
Roberta L. Beauchamp3, Xuan Zhang1, Shingo Kasamatsu4,5, Roderick
T. Bronson6, Elizabeth A. Thiele7,8, David J. Kwiatkowski9, Anat
Stemmer-Rachamimov10, Bence György11,12,13, King-Hwa Ling14,15,
Masao Kaneki4,5, Bakhos A. Tannous1, Vijaya Ramesh3, Casey A.
Maguire1, Xandra O. Breakefield1†
Tuberous sclerosis complex (TSC) results from loss of a tumor
suppressor gene - TSC1 or TSC2, encoding hamartin and tuberin,
respectively. These proteins formed a complex to inhibit
mTORC1-mediated cell growth and prolifer-ation. Loss of either
protein leads to overgrowth lesions in many vital organs. Gene
therapy was evaluated in a mouse model of TSC2 using an
adeno-associated virus (AAV) vector carrying the complementary for
a “condensed” form of human tuberin (cTuberin). Functionality of
cTuberin was verified in culture. A mouse model of TSC2 was
generated by AAV-Cre recombinase disruption of Tsc2-floxed alleles
at birth, leading to a shortened lifespan (mean 58 days) and brain
pathology consistent with TSC. When these mice were injected
intravenously on day 21 with AAV9-cTuberin, the mean survival was
extended to 462 days with reduction in brain pathology. This
demon-strates the potential of treating life-threatening TSC2
lesions with a single intravenous injection of AAV9-cTuberin.
INTRODUCTIONTuberous sclerosis complex (TSC) is a hereditary
disease affecting multiple organs with an incidence of about 1 of
5500 (1, 2), resulting from mutations in either TSC1 encoding
hamartin or TSC2 encoding tuberin. Hamartin and tuberin normally
act as a complex to inhibit mTORC1 (mammalian/mechanistic target of
rapamycin complex 1) through guanosine triphosphatase (GTPase)
activating effects on Ras homolog enriched in brain (Rheb) (3).
When a mutation in the corresponding normal TSC1 or TSC2 allele
occurs somatically in susceptible cells, they enlarge and
proliferate causing abnormal de-velopment and tissue lesions. These
secondary mutations can occur prenatally or after birth in
different cell types, and the timing and frequency of these hits
affect the severity of the disease in a stochastic manner.
Neurodevelopmental manifestations are responsible for the greatest
morbidity, including severe, refractory epilepsy and hydrocephalus,
as well as autism (40%), cognitive impairment
(50%), and mental health issues (70%) (4–6). In addition, renal
angiomyolipomas forming later in life can cause life-threatening
hemorrhage and/or renal failure, and pulmonary
lymphangioleio-myomatosis can severely compromise respiratory
function. Current treatments include surgical resection and/or
treatment with rapa-mycin analogs (rapalogs). Although often well
tolerated, rapalogs cause immune suppression (7) and potentially
compromise early brain development (8), and lifelong therapy is
often required. Therefore, there is a clear need to identify other
therapeutic ap-proaches for TSC.
Adeno-associated virus (AAV) vectors have been used widely in
clinical trials for many hereditary diseases with little-to-no
toxicity, long-term action in nondividing cells, and improvement in
symp-toms (9–11). Benefit can be seen after a single injection and
some serotypes, e.g., AAV9, AAVrh8, and AAVrh10, can efficiently
enter the brain, as well as peripheral organs after intravenous
(IV) injec-tion (12, 13). The insert capacity of AAV vectors
is about 4.7 kb (including promoter, transgene, polyadenylation
(poly A) se-quence, and other regulatory elements), and the
complementary DNA (cDNA) for tuberin (5.4 kb) cannot be
accommodated. We generated a cDNA encoding a shorter form of
tuberin, termed cTuberin. We tested its lack of toxicity and
ability to bind to hamartin and Rheb1, as well as to suppress
phosphoS6 kinase activity in cul-tured cells. In a stochastic mouse
model of TSC2 [based on a TSC1 model; (14)], AAV vector encoding
Cre recombinase was intro-duced by intracerebroventricular (ICV)
injection into homozygous Tsc2-floxed mice (15) at postnatal day 0
(P0) typically leading to death at about P58 with enlarged
ventricles. Near-normal life span and reduction of brain pathology
were achieved in most of these animals by a single IV injection of
an AAV9 vector encoding cTuberin under a strong, constitutive
promoter. These studies demonstrate the ability of cTuberin to
suppress overgrowth of tuberin-null cells, including neural cells
and, presumably, other cells in the body, and, hence, support the
preclinical efficacy of AAV-cTuberin for TSC2 lesions.
1Molecular Neurogenetics Unit, Department of Neurology and
Center for Molecu-lar Imaging Research, Department of Radiology,
Massachusetts General Hospital, and Program in Neuroscience,
Harvard Medical School, Boston, MA, USA. 2Depart-ment of Human
Anatomy, Faculty of Medicine and Health Sciences, Universiti Putra
Malaysia, Serdang, Malaysia. 3Center for Genomic Medicine,
Massachusetts General Hospital, Boston, MA, USA. 4Department of
Anesthesia, Critical Care and Pain Medicine, Massachusetts General
Hospital, Harvard Medical School, Charlestown, MA, USA. 5Shriners
Hospitals for Children, Boston, MA, USA. 6Rodent Histopathology
Core Facility, Harvard Medical School, Boston, MA, USA. 7Herscot
Center for Tuberous Sclerosis Complex, Massachusetts General
Hospital, Harvard Medical School, Boston, MA, USA. 8The Pediatric
Epilepsy Program, Department of Neurology, Massachusetts General
Hospital, Boston, MA, USA. 9Brigham and Women’s Hospital, Harvard
Med-ical School, Boston, MA, USA. 10Department of Pathology,
Massachusetts General Hospital, Boston, MA, USA. 11Department of
Neurobiology and Howard Hughes Medical Institute, Harvard Medical
School, Boston, MA, USA. 12Institute of Mo-lecular and Clinical
Ophthalmology, Basel, Switzerland. 13Department of Oph-thalmology,
University of Basel, Basel, Switzerland. 14Department of Genetics,
Harvard Medical School, Boston, MA, USA. 15Department of Biomedical
Science, Faculty of Medicine and Health Sciences, Universiti Putra
Malaysia, Serdang, Malaysia.*These authors contributed equally to
this work as co-first authors.†Corresponding author. Email:
[email protected]
Copyright © 2021 The Authors, some rights reserved; exclusive
licensee American Association for the Advancement of Science. No
claim to original U.S. Government Works. Distributed under a
Creative Commons Attribution NonCommercial License 4.0 (CC
BY-NC).
on July 2, 2021http://advances.sciencem
ag.org/D
ownloaded from
http://advances.sciencemag.org/
-
Cheah et al., Sci. Adv. 2021; 7 : eabb1703 8 January 2021
S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L
E
2 of 13
RESULTScTuberin constructWhereas hamartin is encoded in a cDNA
of 3.5 kb, which fits into an AAV vector (16), the cDNA for tuberin
(5.4 kb) is too large. To generate a potentially functional form of
tuberin encoded in a shorter cDNA, we retained the N-terminal
domain that binds to hamartin and the C-terminal domain containing
GAP (GTPase- activating protein) activity that inhibits Rheb, with
N-terminal region and phosphorylation of the C-terminal region of
tuberin also thought to regulate formation of the complex with
hamartin Fig. 1A (3, 17–20). The potential for cTuberin
to retain some functional activity was supported by findings of
Momose et al. (21) that genomic overexpression of the
C-terminal region of rat tuberin (amino acids 1425 to 1755) can
suppress renal tumors in the Tsc2 Eker rat model. We felt it was
also important to retain the hamartin-binding domain at the N
terminus, as hamartin and tuberin function together as a complex
with Tre2-Bub2-Cdc16 (TBC) 1 domain family, member 7 (TBC1D7) to
accelerate guanosine triphosphate (GTP) to guanosine diphosphate
conversion of Rheb-GTP (3, 22). In addition, this re-quirement
for complex formation for activity might act to limit poten-tial
negative effects of high levels of transgenic cTuberin expression.
cTuberin was thus designed to retain key elements of function,
in-cluding 450 amino acids from the N-terminal region and 292 amino
acids from the C-terminal region, joined by a flexible
serine-glycine linker of 16 amino acids (fig. S1). This cDNA, with
a Kozak sequence, and a C-terminal c-Myc tag was inserted into an
AAV2 backbone under a chicken -actin (CBA) promoter (23), with a
WPRE (wood-chuck hepatitis virus posttranscriptional regulatory
element) and poly A signals (Fig. 1B).
cTuberin is expressed in cultured cells with no apparent
toxicityHuman embryonic kidney (HEK) 293T cells were transfected
with plasmids for empty AAV (AAV1-null that contains all the
elements
except the cTuberin cDNA), AAV-CBA–green fluorescent protein
(GFP), or AAV-CBA-cTuberin-Myc to assess the expression level of
cTuberin. In addition to endogenously expressed tuberin (200 kDa),
cTuberin expression at the appropriate molecular weight (MW) of 85
kDa was detected on Western blots using anti-tuberin and anti-Myc
antibodies (Fig. 2A; representative blot, n = 3)
Immunocytochem-istry of 293T cells transfected with different
plasmids demonstrated stronger tuberin immunoreactivity in those
transfected with AAV-CBA-cTuberin-Myc compared to other groups that
expressed only endogenous tuberin (Fig. 2B; representative
micrographs, n = 3). We determined transfection
efficiency of these cells in two ways—microscopically and by flow
cytometry. As it is challenging to dif-ferentiate expression of
endogenous tuberin from cTuberin, image analysis was carried out
microscopically for each well (approximately 2000 cells per well;
n = 3) for 4′,6-diamidino-2-phenylindole (DAPI)– positive
and c-Myc–positive cells, and we determined that 43 ± 2%
of cells were transfected with the AAV-CBA-cTuberin-Myc
(Fig. 2B). Cytotoxicity assays were also performed following
transfection of HEK293T cells with AAV-null, AAV-CBA-GFP, or
AAV-CBA-cTuberin-Myc plasmids to evaluate potential toxicity of
cTuberin. The lactate dehydrogenase (LDH) assay (Dojindo Molecular
Technologies Inc., Rockville, MD, USA) revealed no cytotoxicity in
cTuberin-transfected cells, as compared to controls (Fig. 2C;
n = 3). As a second way to evaluate the extent of
transfection of these 293T cells with AAV-CBA-cTuberin- Myc plasmid
DNA (n = 3), we sorted the c-Myc–positive cells using
flow cytometry, after staining the cells with unlabeled c-Myc
primary antibody followed by Alexa Fluor 647–conjugated secondary
antibody. Compared to the background in nontransfected cells
(4 ± 1%), we detected a marked increase of c-Myc–positive
cells (50 ± 1% or 46% minus the background, similar to
the 43% determined by cell counting) after transfection with the
AAV-CBA-cTuberin-Myc plasmid (P
-
Cheah et al., Sci. Adv. 2021; 7 : eabb1703 8 January 2021
S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L
E
3 of 13
with Flag-hamartin and HA-S6K plasmids. Hamartin and full-length
tuberin coexpression inhibited phosphorylation of S6K T389, as
expected, and similarly, coexpression of hamartin and cTuberin also
decreased pS6K T389 levels (Fig. 4A), supporting the ability
of cTuberin to bind to hamartin and efficaciously inhibit TORC1
activity. Level of pS6K T389 inhibition was quantified rela-tive to
HA-S6K using Fiji/ImageJ. Flag-hamartin cotransfected with AAV-GFP
served as a control (normalized to 1.0), and cotransfec-tion with
full-length tuberin and cTub-Myc revealed a significant inhibition
of S6K T389 phosphorylation by 69 and 56%, respectively
(*P
-
Cheah et al., Sci. Adv. 2021; 7 : eabb1703 8 January 2021
S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L
E
4 of 13
Different cohorts of mice were subjected to body weight
mea-surement and motor function assessment starting at P21/22 for
naïve, noninjected animals, AAV1-Cre ICV injected (1 × 1010 vg/kg)
at P1 only or followed with AAV9-cTuberin injected (8 × 1012 vg/kg)
IV at P21. Body weights of these mice from age 21 to 50 days did
not differ according to treatment (Fig. 5B). Movement was
assessed using an automated rotarod apparatus with accelerating
rotary velocity (4 going to 64 rpm over 2 min) to assess motor
skills of the mice as time of latency to fall. A significant
increase in latency
was observed for the AAV1-Cre + AAV9-cTuberin as
compared to the AAV1-Cre–injected mice and naive mice
(Fig. 5C). During animal handling, two mice of six Tsc2-floxed
AAV-Cre–injected mice (day 41) and two mice of seven Tsc2-floxed
AAV-Cre–injected + AAV-cTuberin– injected mice (one each on days 47
and 50) mani-fested straub (vertical tail), humped back, and/or
motor seizures, which did not, however, compromise their consequent
rotarod per-formances (fig. S2).
Two other approaches were less effective at extending survival
of AAV1-Cre ICV–injected Tsc2-floxed mice. In one, using a similar
time scheme (fig. S3), Tsc2-floxed pups were injected with 1 × 1014
vg/kg AAV1-Cre ICV at P3 and then 3 × 1012 vg/kg of AAV1-cTuberin
(in contrast to AAV9 serotype) IV at P21, with the higher amount of
AAV1-Cre (without cTuberin) leading to death with a mean of 36 days
and survival only being extended by AAV1-cTuberin to a mean of 54
days. This probably reflects the fact that AAV1 is less efficient
at crossing the blood-brain barrier (BBB) than AAV9. In another
experiment, the Tsc2-floxed pups were injected ICV with AAV1-Cre (1
× 1012 vg/kg) at P0, followed by ICV injection (in con-trast to
systemic injection) of 4.5 × 1013 vg/kg of AAV9-cTuberin at P3.
This approach led to median survival of 50 days in Tsc2-floxed mice
without cTuberin injection, while those injected with AAV9-
cTuberin had extended median survival only up to 95 days (fig. S4).
This experiment raises the possibility that other lesions in the
body (in addition to the brain) resulting from ICV injection of
AAV1-Cre were associated with death and were not sufficiently
alleviated by ICV injection of the cTuberin vector and/or that the
high dose AAV-cTuberin injected ICV into P3 pups had some toxicity
(26).
In naïve (normal) Tsc2-floxed mice, the ventricle is lined by a
single layer of ependymal cells (Fig. 6A). Neuropathological
examina-tion at P42 revealed that ICV injection of AAV1-Cre in
Tsc2-floxed mice at P0 led to multiple layers of ependymal and
subependymal cells lining the lateral ventricle (indicating
increased proliferation of these cells) (Fig. 6B, asterisk),
which sometimes appeared as nodules along the ventricular lining
(Fig. 6C). When these AAV1-Cre–injected mice were treated with
AAV9-cTuberin (IV injected at P21), there was apparent regression
of ependymal/subependymal overgrowths
Fig. 3. cTuberin-Myc binds hamartin and Rheb comparably to
full-length tuberin. Representative blot (n = 3 experiments) after
cotransfection of the Myc-tagged cTuberin (AAV-CBA-cTub-Myc) or
full-length tuberin (Myc-FL-tuberin) along with FLAG-tagged
hamartin and HA-tagged GST-Rheb1. Coimmunoprecipitation (co-IP)
using anti-Myc antibody demonstrated that cTuberin-Myc interacts
with both Flag-hamartin and HA-Rheb1 similar to Myc-FL-tuberin.
Conversely, negative con-trol Myc-GSK-3 showed no interaction with
FLAG-hamartin or HA-GST-Rheb1.
Fig. 4. cTuberin-c-Myc inhibits mTORC1 signaling comparably to
full-length tuberin. (A) Full-length Flag-tagged tuberin
(Flag-tuberin), Myc-tagged cTuberin (AAV-cTub-Myc), or AAV-GFP
plasmids were cotransfected into HEK293T cells along with
full-length Flag-tagged hamartin (Flag-hamartin) and HA-tagged
p70S6K (HA-p70S6K), which is phosphorylated at T389 by mTORC1
(latter used as a reporter for mTORC1 activation). Representative
blot (n = 3 experiments) demonstrated similar inhibition levels of
phosphorylated p70S6K (pS6K T389) with either full-length tuberin
or cTub-Myc cotransfected with full-length hamartin. (B)
Quantitation of decrease in S6K T389 phosphorylation was performed
relative to HA-S6K using Fiji/ImageJ. Flag-hamartin cotransfected
with AAV-GFP served as a control (normalized to 1.0), and
cotransfec-tion with full-length tuberin or cTub-Myc revealed
inhibition of 69 or 56%, respectively, representing the results
from three experiments. *P < 0.05.
on July 2, 2021http://advances.sciencem
ag.org/D
ownloaded from
http://advances.sciencemag.org/
-
Cheah et al., Sci. Adv. 2021; 7 : eabb1703 8 January 2021
S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L
E
5 of 13
(Fig. 6D). We also stained these mouse brain sections (P42)
for Ki67 as an indication of cell proliferation. As expected, there
was little-to-no proliferation of ependymal/subependymal cells
lining the ventricles in the naïve brain (Fig. 7A). In
contrast, after AAV1-Cre injection at P0, there was marked
proliferation of these cells, includ-ing apparent migration of
dividing cells into the brain parenchyma (Fig. 7B), also seen
after subsequent IV injection with AAV9-null vector (Fig. 7C).
In contrast, IV injection of the AAV9-cTuberin vector decreased
proliferation and inward migration of Ki67+ ependymal/subependymal
cells (Fig. 7D).
The brain sections (P42) were also immunostained for
phospho-rylated ribosomal protein S6 (pS6). We observed low pS6
expression in the whole brain sections of the noninjected (naïve)
mouse brain (Fig. 8A, top). In contrast, in AAV1-Cre
ICV–injected Tsc2-floxed mice, pS6 expression was intense in many
brain cells [Fig. 8, A (middle) and Bi], with the
pS6-positive cells being significantly larger in size
(Fig. 8Bii) and with a higher pS6 immunofluorescence signal
(Fig. 8Biii). When the AAV1-Cre–injected mice were subjected
to IV injection of the AAV9-cTuberin vector at P21, the pS6
immuno-reactive cells were significantly decreased in average size
by 23%
[P
-
Cheah et al., Sci. Adv. 2021; 7 : eabb1703 8 January 2021
S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L
E
6 of 13
another three Tsc2-floxed animals subjected to no injections
were used as controls. For quantitative polymerase chain reaction
(qPCR) analysis of AAV genomes (probes and primer specific),
50 ng of DNA was used as a template, and primers and probes
were designed to amplify the cTuberin in the infected animal (fig.
S5). cTuberin DNA was not detected in the noninjected control group
(Fig. 8D). Cycle threshold (Ct) values for the Tsc2-floxed
animals injected with AAV1-Cre and AAV9-cTuberin vectors were
readily detectable with approximately 30.8 ± 2.6 and
17.2 ± 0.2 cycles for brain and liver tissue,
respectively (Fig. 8Di). The large difference between the AAV
genomes in brain compared to liver is likely due to both the high
tropism of systemically injected AAV for the liver and the
relatively low dose of vector injected (9 × 1011 vg/kg). To detect
cTuberin transgene expression, total RNA was extracted from the
brains and livers of another set of animals, including noninjected
controls; Tsc2- floxed animals injected with AAV1-Cre only, and
Tsc2-floxed ani-mals injected with AAV1-Cre and AAV9-cTuberin
vectors (n = 3 for all groups), with the dosage of
AAV1-Cre ICV injected at P1 (1 × 1010 vg/kg) or combined with
AAV9-cTuberin injected IV at P21 (1.8 × 1012 vg/kg). Quantitative
reverse transcription PCR (RT-qPCR)
analysis indicated that cTuberin mRNA was undetectable in the
non-injected control group and those injected with AAV1-Cre only.
In contrast, in both brains and livers, we detected cTuberin mRNA
in mice injected with AAV9-cTuberin at levels of Ct
36.8 ± 3 and 34.8 ± 0.5 cycles, respectively
(Fig. 8Dii). We did not detect cTuberin cDNA when reverse
transcriptase was omitted from the RT reac-tion, indicating that we
were detecting bona fide cTuberin mRNA and not sample contamination
with AAV-cTuberin genomes.
DISCUSSIONThis is the first description of an alternative mode
of therapy for TSC type 2 (TSC2) involving gene replacement using
an AAV vec-tor encoding a condensed form of tuberin, termed
cTuberin. We developed a stochastic mouse model for central nervous
system (CNS) lesions in TSC2 in which homozygous Tsc2-floxed
mice (15, 27) are injected ICV in the newborn period (P0 to
P3) with an AAV1 vector expressing Cre recombinase, as described
for our stochastic TSC1 model (14). AAV1-Cre injection in the
Tsc2-floxed model resulted in death at about 58 days. Death
appeared to be due primarily to
Fig. 6. Hematoxylin and eosin histochemical staining of mouse
brain with and without AAV9-cTuberin vector injection following
AAV1-Cre injection. Tsc2-floxed mouse pups were either not injected
(naïve) or injected ICV in both ventricles (1 × 1012 vg/kg) with an
AAV1-Cre vector at P0. At 21 days, some mice were injected IV with
AAV9-cTuberin (9 × 1011 vg/kg) or noninjected. At 42 days, all mice
were euthanized. (A) Naïve, noninjected brain (black arrowhead
indicating the choroid plexus). (B and C) Tsc2-floxed mice with
AAV1-Cre at P0 and no further injection showed (B) proliferation of
ependymal/subependymal cells (asterisk) and (C) subependymal
nodules. (D) Little-to-no subependymal overgrowth was detected in
mice receiving both the P0 AAV1-Cre ICV injection and P21 IV
AAV9-cTuberin injection. Representa-tive images are shown.
Magnification bar, 100 m. CC, corpus callosum; LV, lateral
ventricle.
on July 2, 2021http://advances.sciencem
ag.org/D
ownloaded from
http://advances.sciencemag.org/
-
Cheah et al., Sci. Adv. 2021; 7 : eabb1703 8 January 2021
S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L
E
7 of 13
hydrocephalus caused by ependymal/subependymal overgrowths
blocking cerebrospinal fluid flow, with whole-body pathology
reveal-ing no overt lesions except in the CNS. Although signs of
seizures were noted in a few mice during motor performance
assessment, these animals recovered normal activity. Experiments
showed that IV injection of AAV9-cTuberin vector into this
stochastic Tsc2-floxed mouse model on day 21 extended life span in
most mice (9 of 12) to at least 450 days.
Histochemical/immunohistochemical analysis of the brains supported
a resulting reduction in size of ependymal/subependymal lesions,
decreased proliferation of cells in the sub-ependymal zone, and
reduced phosphorylation of S6 kinase driven by mTOR activity. This
study offers a potential single treatment para-digm for improving
the outcome of patients with TSC2.
Limitations to this stochastic Tsc2 mouse model include the fact
that floxed alleles (before Cre exposure) are normal in function
during prenatal development and that Cre recombinase usually knocks
out both alleles in a cell at once, which is different from the
case in TSC2 patients, most of whom are heterozygous for one mutant
and one normal allele in most-to-all cells in their body. TSC2
heterozygosity itself may compromise some cell functions and
contribute to aspects of the disease phenotype
(1, 28, 29). Further, the model used here is
CNS oriented, with most pathology in the brain; whereas in TSC
patients, a number of organs in addition to the brain are affected.
In addition, this Tsc2 mouse model does not show all the brain
abnor-malities observed in human TSC2, many of which form
prenatally, such as cortical tubers, disorganized cortical
lamination, dysplastic neurons, and giant cells (30). Strengths of
this model are that there is loss of tuberin expression in a number
of different cell types in the brain with variation for animal to
animal, as occurs in patients with TSC. This is in contrast to
commonly used models where Tsc2-floxed mice are mated to mice
expressing Cre recombinase under a cell- specific promoter, e.g.,
the synapsin promoter, in which case most and only neurons lose
expression at embryonic day 12.5 (31).
The central portion of tuberin that was removed to fit coding
sequences into the AAV vector contains a number of phosphorylation
sites that are involved in regulating mTOR activity under some
circum-stances, with three of these sites bearing missense
mutations associated with TSC2, suggesting that they may contribute
to the disease pheno-type or create truncated, nonfunctional
proteins (6). By comparison, there is an ortholog of human tuberin
in Schizosaccharomyces pombe that lacks about 500 amino acids in
the equivalent central region of human tuberin, suggesting that
these sites are dispensable to some
LV
CC
DAPI Ki67 DAPI Ki67
DAPI Ki67DAPI Ki67
AAV1-Cre injected @ P0; AAV9-null @ P21
A B
C D
50 µm
AAV1-Cre injected @ P0; AAV9-cTuberin @ P21
AAV1-Cre injected @ P0Naïve brain (no injection)
**
**
50 µm
50 µm
50 µm
Fig. 7. Ki67 immunostaining of Tsc2-floxed mice brains.
Tsc2-floxed mouse pups were either not injected (naïve) or injected
ICV in both ventricles (1 × 1012 vg/kg) with an AAV1-Cre vector at
P0. At 21 days, some mice were injected IV with AAV9-cTuberin (9 ×
1011 vg/kg), AAV9-null (1 × 1013 vg/kg) or noninjected. At 42 days,
all mice were euthanized. (A) Naïve, noninjected brain reveals
little-to-no staining in the ependymal/subependymal layers. (B)
Tsc2-floxed mice injected with AAV1-Cre vector only showed abnormal
mitotic activity and apparent migration of cells (yellow arrows)
away from the ventricular zone, as well as multiple
ependymal/subependymal layers (green arrowheads) as compared to the
naïve group. (C) Tsc2-floxed mice injected with AAV1-Cre vector
followed by AAV9-null vector showed abnormal mitotic activity of
the cells and thickening of the subventricular zone. (D) The
Tsc2-floxed mice injected with AAV1-Cre and then rescued with the
AAV9-cTuberin vector showed a trend toward normalization of the
ependymal/subependymal layer. The corresponding brain sections were
counterstained with DAPI. The yellow asterisk denotes
autofluores-cence in the choroid plexus. Representative images are
shown. Magnification bar, 100 m.
on July 2, 2021http://advances.sciencem
ag.org/D
ownloaded from
http://advances.sciencemag.org/
-
Cheah et al., Sci. Adv. 2021; 7 : eabb1703 8 January 2021
S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L
E
8 of 13
functions (32). Further, some of the key Akt phosphorylation
sites in mammalian tuberin are not essential in Drosophila (33),
and phosphorylation sites for Akt, ribosomal protein S6 kinase, and
AMP-activated protein kinase (AMPK) in the central region of human
tuberin are not present in Schizosaccharomyces or Dictyostelium
(34), suggesting that these sites may not be critical for function.
Given the critical role of phosphorylation sites in tuberin in
growth factor and cytokine signaling in mammalian cells, one would
anticipate that cTuberin in TSC2-null cells would lack some of
these regulatory con-trols. However, in the Eker rat model of TSC2,
which is prone to renal carcinomas, the C-terminal region alone
(amino acids 1425 to 1755) of rat tuberin suppresses tumor
formation in a dose-dependent manner (35). Fortunately, in TSC2
patients, only a very small fraction of cells in the body suffer
loss of tuberin, and most damage is done by the en-largement and
proliferation of these deficient cells. Thus, if overgrowths can be
suppressed by cTuberin, then that would bring therapeutic benefit
for many of the symptoms of the disease, although the cells would
not be fully “normalized.” So far, in cultured cells, cTuberin has
been shown to bind to hamartin, and overexpression of cTuberin was
not found to be toxic. cTuberin inhibited mTORC1 signaling in
these cells to the same extent as tuberin, supporting the use of
cTuberin as an effective replacement for tuberin for some cell
properties.
Subependymal nodules (SENs) occur in 10 to 15% of children with
TSC, usually appearing after birth and being more severe in TSC2
than TSC1 (36–38). SENs can enlarge into subependymal giant cell
astrocytomas (SEGAs) during the first decade of life causing
obstruction of cerebrospinal fluid flow, potentially leading to
life- threatening hydrocephalus, as well as endocrinopathy and
visual impairment (36, 37, 39, 40). Under optimal
care, infants and children with TSC are monitored for subependymal
lesions by magnetic res-onance imaging (MRI) every 6 to 12 months.
The two current standards of care are neurosurgical removal of
SEGAs through craniotomy, which can be associated with significant
morbidity (37), or treatment with rapalogs, which inhibit mTOR
activity. Rapalogs have proven effective in reducing lesion size,
but they require continuous treatment and have limited access to
the brain after peripheral administration. Potential problems with
this class of drugs include a compromise of immune function (41),
interference with white matter integrity (42), and possible
interference with brain development in early childhood (43). In
several studies, the mTOR
B
Nai
veA
AV1-
Cre
on
ly
AAV
1-C
re @
P0;
A
AV9-
cTub
erin
@
P21
High magnification of insetsDAPI pS6
Inverted imageA
C AAV1V1-1-Cre injected @ P0Noninjectedninjected
TSC2
pS6
GAPDH 0 . 00 . 5
1 . 0
1 . 5
2 . 0
2 . 5
Pr
ote
in e
xp
re
ss
ion
le
ve
l
re
lati
ve
to
GA
PD
H
A A V 1 - C r e @ P 0 − + − +
*
*
T S C 2 p S 6
D ii. Transgene-encoded mRNAi. AAV genomes
0
1 0
2 0
3 0
4 0
No
rm
al
iz
ed
C
t
n . d . n . d . n . d . n . d .
No
ni n
j ec t
e d
P0 ,
AA
V1 -
Cr e
P0 ,
AA
V1 -
Cr e
; P
2 1,
AA
V9 -
c Tu
be r
i n
No
ni n
j ec t
e d
P0 ,
AA
V1 -
Cr e
P0 ,
AA
V1 -
Cr e
; P
2 1,
AA
V9 -
c Tu
be r
i n
B r a i n L i v e r
2 6
2 8
3 0
3 2
3 4
3 6
3 8
4 0
No
rm
ali
ze
d C
t
B r a i n L i v e r
n . d . n . d .
No n
i nj e
c te d
P 0, A
A V1 -
C re ;
P2 1
,
A AV 9
- cT u
b er in
No n
i nj e
c te d
P 0, A
A V1 -
C re ;
P2 1
,
A AV 9
- cT u
b er in
0 k
5 0 k
1 0 0 k
1 5 0 k
2 0 0 k
2 5 0 k
Inte
ns
ity
(p
ixe
ls) *
( i i i ) p S 6 p u n c t a i n t e n s i t y
n . d .
0
1
2
3
No
rma
liz
ed
nu
mb
er
of
pS
6 P
un
cta
(a
.u.)
n . d .
( i ) T o t a l p S 6 p u n c t a a / a r e a
0
1 0
2 0
3 0
4 0
5 0
Siz
e (
um
2)
n . d .
*
( i i ) p S 6 p u n c t a s i z e
AAV1
-Cre
@ P
0;@
P21
AAV1
-Cre
@ P
0;@
P21
AAV1
-Cre
@ P
0;@
P21
Fig. 8. pS6 immunostaining of Tsc2-floxed mice brains.
Tsc2-floxed mouse pups were either not injected (naïve) or injected
ICV (1 × 1012 vg/kg) with an AAV1-Cre vector at P0. At P21, some
mice were injected IV with AAV9-cTuberin (9 × 1011 vg/kg) or
noninjected. All were euthanized at P42. (A) Whole mouse brain
sections from naïve, AAV1-Cre, and AAV1-Cre+ AAV9-cTuberin injected
mice stained for pS6 and DAPI. Representative whole brain sections
(scale bar, 1 mm; eight-bit–thresholded in-verted images) indicated
absence of pS6 puncta in naïve group. In other groups, pS6 puncta
appeared as darkened spots within the cerebral cortex and caudate
putamen; high magnification inset images (scale bar, 100 m;
12-bit–thresholded inverted images). (B) pS6 analysis included
puncta density (i), size (ii), and intensity (iii). *P < 0.05; n
= 3. a.u., arbitrary units. (C) Compared to naïve pups,
immunoblotting demonstrated AAV1-Cre–mediated decrease of tuberin
(54%) and increase in pS6 (76%) in Tsc2-floxed mice injected with
AAV1-Cre, relative to glyceraldehyde-3-phosphate dehydrogenase
(GAPDH) with naïve brain as control (normalized to 1.0; *P <
0.05; n = 3). (Di) Ct values for biodistribution of AAV vector
genomes in the brain and liver measured by qPCR. (Dii) Ct value of
GAPDH and cTuberin cDNAs in brains of naïve animals in-jected with
AAV1-Cre only or with AAV1-Cre and AAV9-cTuberin. n.d., not
determined.
on July 2, 2021http://advances.sciencem
ag.org/D
ownloaded from
http://advances.sciencemag.org/
-
Cheah et al., Sci. Adv. 2021; 7 : eabb1703 8 January 2021
S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L
E
9 of 13
pathway has been found to be critical to neurodevelopment,
includ-ing neuronal growth, axonal guidance, synapse formation, and
my-elination (44–46). Inhibition of mTOR by rapalogs may contribute
to the observed memory dysfunction following prenatal/postnatal
drug treatment in Tsc mouse models (47) and the behavioral
abnor-malities in wild-type mice treated prenatally with rapamycin
(48). Some physicians do not recommend the use of these drugs in
children or pregnant women as “long-term effects on growth and
develop-ment in pediatric patients are not fully known” (43).
Although in at least one study, rapalog treatment was reported to
have no signifi-cant effect on neurocognitive function or behavior
in children with TSC (49).
Our premise is that current therapies for children with TSC may
have associated morbidity resulting in the potential for decreased
mental functions. Another therapeutic approach would be
intravascular administration of an AAV vector that can cross the
BBB encoding a replacement “gene” for the mutant TSC1 or TSC2
alleles. Since SENs are slow growing, there would be time to
monitor their size by MRI over several months and leave open the
opportunity to administer standard-of-care treatment, as needed. It
is hoped that gene re-placement therapy might reduce use of more
problematic standard-of-care procedures in young children and
provide long-lasting benefit with a single administration. Certain
serotypes of AAV, such as AAV9, are able to penetrate the BBB as
well as deliver to peripheral tissues (13). Thus, with IV delivery,
“extra copies” of the replace-ment gene would be provided to
multiple tissues, including brain, kidney, liver, and lungs, which
might reduce the likelihood that so-matic mutations in TSC genes
later in life would lead to disruptive hamartomas.
Advantages of AAV gene therapy are the potential for a single
vector injection yielding long-term transgene expression in
non-dividing cells. It is assumed that once a tuberin analog is
delivered to cells in TSC2 lesions, they would shrink and stop
dividing and, hence, retain transgene expression. Gene therapy may
be a viable option for infants/children with TSC to reduce
potential compro-mise of brain functions caused by congenital
lesions and secondary sequelae of these lesions. AAV9 vectors have
been used in young mice with spinal muscular atrophy (SMA) for gene
replacement of the survival motor neuron (SMN) protein using both
IV (50) and intrathecal (51) gene delivery. An AAV9-SMN drug,
Zolgensma (Novartis), is now U.S. Food and Drug
Administration–approved for IV treatment of babies/children with
SMA. Two critical aspects of successful gene therapy with AAV
vectors are as follows: (i) a known target, in the case of TSC2
loss of function of tuberin; and (ii) no toxicity resulting from
overexpression of the replacement protein, since levels of
expression cannot at present be regulated. There is a predicted
reduced chance of toxicity of cTuberin as it should only be active
in a 1:1 complex with hamartin, and hamartin levels are normal in
TSC2 null cells (52), with cTuberin not bound to hamartin
presumably being degraded. So far, no toxic effects of cTuberin
expression have been observed in cells in culture or in mice.
Clinical trials should be facilitated by the ability to image
reduced lesion size within months by MRI due to shrinking of cell
volume and inhibition of cell proliferation, as was found in the
rapalog trial for renal angiomyolipomas (53). Typically, AAV
vectors are just administered once due to previous exposure to the
AAV virus in life eliciting an immune response to the capsid and
reduc-ing secondary transduction (54). If replacement is
insufficient to reduce symptoms or new TSC2 null lesions arise
later in life after
AAV gene replacement, it would still be possible to treat
patients with rapalogs or possibly exoAAV (55). These studies
support the potential of AAV gene therapy for TSC2, which might be
especially useful in infants and children where drug inhibition of
the mTOR pathway may interfere with early brain development.
MATERIALS AND METHODSAAV vector design and packagingThe AAV
vector plasmid, AAV-CBA-Cre-BGHpA, was derived as described in
Prabhakar et al. (16). These AAV vectors carry AAV2 inverted
terminal repeat elements, and gene expression is controlled by a
hybrid promoter (CBA) composed of the cytomegalovirus (CMV)
immediate/early gene enhancer fused to the -actin pro-moter (23).
To increase the efficiency of cTuberin translation (for future use
in human gene therapy approach), cDNA encoding cTuberin was human
codon-optimized before gene synthesis by GenScript Biotech
(Piscataway, NJ, USA). AAV vector plasmid, AAV-CBA-cTuberin-c-Myc,
was derived from the plasmid pAAV-CBA-W (56). This vector contains
the CBA promoter driving cTuberin, followed by a WPRE and both SV40
and bovine growth hormone (BGH) polyadenylation (poly A) signal
sequences. Our cTuberin construct contains the following: ACC
(Kozak sequence) :: amino acids 1 to 450 of human tuberin::gly/ser
linker :: amino acids 1515 to 1807 of human tuberin :: c-Myc
tag = 2307 bp encod-ing an 85-kD protein (fig. S1). The
pAAV-CBA-W, which contains the CBA promoter, WPRE, and poly A
sequences, but no transgene, served as AAV-null in our studies.
AAV1 and AAV9 serotype vectors were produced by transient
cotransfection of HEK293T cells by calcium phosphate precipita-tion
method of vector plasmids (e.g., AAV-CBA-cTuberin-Myc), adenoviral
helper plasmid pAdF6, and a plasmid encoding AAV9 (pAR9) or AAV1
(pXR1) rep and capsid genes, as previously de-scribed (57). All AAV
vectors carried the identity of all PCR-amplified sequences as
confirmed by sequencing. Briefly, AAV vectors were purified by
iodixanol density gradient centrifugation. The virus- containing
fractions were concentrated using Amicon Ultra 100-kDa molecular
weight cut-offs (MWCO) centrifugal devices (EMD Millipore,
Billerica, MA, USA), and the titer vector genomes (vg) per
milliliter was determined by quantitative real-time PCR
amplification with primers and TaqMan probe specific for the BGH
poly A signal.
Cell cultureHEK293T cells [American Type Culture Collection
(ATCC)] and COS-7 cells (ATCC, Manassas, VA, USA) were cultured in
Dulbecco’s modified Eagle’s medium (DMEM; Thermo Fisher Scientific,
Hampton, NH, USA) supplemented with 10% fetal bovine serum (FBS;
Gemini Bio Products, West Sacramento, CA, USA) and 1%
penicillin/streptomycin (Thermo Fisher Scientific). The cell
cultures were periodically screened to ensure they are free from
mycoplasma contamination using the PCR Mycoplasma Detection Kit
(ABM, G238, Richmond, BC, Canada).
Detection of cell cytotoxicity using LDH release assayHEK293T
cells were seeded in 96-well plates (10,000 cells per well) and,
after 24 hours, transfected with various plasmid DNAs (AAV-null,
AAV-GFP, and AAV-cTuberin) at 250 ng/10,000 cells using
Lipo-fectamine 2000, according to the manufacturer’s instructions
(Life Technologies, Carlsbad, CA, USA) in Opti-MEM (Life
Technologies).
on July 2, 2021http://advances.sciencem
ag.org/D
ownloaded from
http://advances.sciencemag.org/
-
Cheah et al., Sci. Adv. 2021; 7 : eabb1703 8 January 2021
S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L
E
10 of 13
Six hours later, transfection media was removed and replaced
with DMEM (10% FBS and 1% Penicillin-Streptomycin solution), and
cells were allowed to grow for 72 hours. One group of cells was
treated with potent proteasome inhibitor Bortezomib (VELCADE;
Mil-lennium Pharmaceuticals Inc., Cambridge, MA, USA) (58) at 250
nM for 72 hours, as a positive control for toxicity. Cellular
toxicity caused by plasmid DNA transfection was assessed by
quantification of extracellular LDH activity using LDH assay
kit-WST (Dojindo Molecular Technologies Inc.), following the
manufacturer’s instruc-tions. Briefly, the supernatant for each
transfected or treated sample was collected and incubated with
substrate for 30 min at 37°C. Fol-lowing incubation, stop
solution was added, and absorbance was measured at 490 nm.
Western blotsBriefly, cultured cells were harvested in lysis
buffer [50 mM Hepes (pH 8.0), 150 mM NaCl, 2 mM EDTA, 2.5% sodium
dodecyl sulfate, 2% CHAPS, 2.5 mM sucrose, 10% glycerol, 10 mM
sodium fluoride, 2 mM sodium vanadate, 1 mM phenylmethylsulfonyl
fluoride (Sigma- Aldrich, St. Louis, MO, USA), 10 mM sodium
pyrophosphate, and protease inhibitor cocktail (P8340,
Sigma-Aldrich)]. After sonica-tion and incubation at 8°C for 10
min, the samples were centrifuged at 14,000g for 30 min at
8°C. Equal amounts of protein, determined by a detergent-compatible
protein assay kit (Bio-Rad Laboratories, Hercules, CA, USA), were
boiled for 5 min in Laemmli sample buffer (Bio-Rad), separated
by SDS–polyacrylamide gel electrophoresis (PAGE), and transferred
onto nitrocellulose membranes (Bio-Rad). Equal protein loading was
confirmed by Ponceau S staining. The membranes were blocked in 2%
blocking reagent (GE Healthcare, Pittsburgh, PA, USA) for 1 hour at
room temperature (RT) and in-cubated with primary antibodies
overnight at 4°C. Anti-tuberin/TSC2 (#3612), anti–phospho-S6
(#2211), anti-S6 (#2212), anti-Myc (clone 9B11, #2276) (Cell
Signaling Technology, Danvers, MA, USA), anti–-actin (#A5441),
anti-FLAG (clone M2, #F1804) (Sigma- Aldrich), anti-HA (clone F-7,
sc-7392, Santa Cruz Biotechnology, Dallas, TX, USA), and
anti–glyceraldehyde-3-phosphate dehydro-genase (GAPDH) (#CB1001,
EMD Millipore) were used as primary antibodies. Anti-rabbit or
anti-mouse immunoglobulin G antibody conjugated with horseradish
peroxidase was used as a secondary antibody (Thermo Fisher
Scientific). Enhanced chemiluminescence reagent, Lumigen ECL Ultra
(TMA-6) (Lumigen, Southfield, MI, USA), was used to detect the
antigen-antibody complexes.
Immunoprecipitation and HA-p70S6K reporter assaysFor
immunoprecipitations, COS-7 cells were transfected with plasmid
vectors—AAV empty, AAV-CBA-cTuberin-Myc, pcDNA-hamartin- FLAG (V.
Ramesh laboratory), pReceiver-M09/tuberin-Myc (catalog no.
EX-Z5884-M09, GeneCopoeia, Rockville, MD, USA),
pCMV-Tag3A-Myc-GSK-3 (GSK-3 sequence was cloned into pCMV- Tag3A
vector; catalog no. 211173-51, Agilent Technologies, Santa Clara,
CA, USA), and pRK5-HA-GST-Rheb1 [catalog no. 19310, Addgene,
Watertown, MA, USA; provided by Sancak et al. (59)] using
Lipofectamine 2000 (Life Technologies). Cells were lysed with
ice-cold phosphate-buffered saline (PBS) (pH 7.4) containing 1%
Triton X-100, 2 mM EDTA, 10 mM sodium pyrophosphate, 1 mM
phenylmethylsulfonyl fluoride, 2 mM sodium vanadate, 10 mM sodium
fluoride, and proteinase inhibitors cocktail (Sigma-Aldrich).
Lysates were centrifuged at 15,000 rpm for 10 min at 4°C,
and pro-tein concentration was measured using the Bradford protein
assay
(Bio-Rad). One milligram of lysates was incubated with 2 g of
anti-Myc-tag antibody (catalog no. 16286-1-AP, Proteintech,
Rosemont, IL, USA) in the presence of Protein A/G Agarose (Santa
Cruz Bio-technology) at 4°C overnight. After washing twice with
ice-cold modified PBS buffer (pH 7.4) (287 mM NaCl, 2.7 mM KCl, 10
mM Na2HPO4, 1.8 mM KH2PO4, 0.05% Triton X-100, and 1 mM EDTA),
resins were incubated in 30 l of 0.2 M glycine-HCl buffer (pH
2.5) (Polysciences Inc. Warrington, PA, USA) at RT for 15 min, and
then the supernatants were collected and neutralized by adding an
equal amount of 1 M tris-HCl (pH 8.0) (Sigma-Aldrich). To
increase stringency during the washing, NaCl concentration was
increased from 137 to 287 mM in the modified PBS buffer to reduce
ionic protein interaction. Eluted immunoprecipitates or whole-cell
lysates were separated by SDS-PAGE and analyzed by immunoblotting
with antibodies specific for Myc-tag (dilution 1:5000) (catalog no.
2276, Cell Signaling Technology), FLAG-tag (1:25,000) (catalog no.
F1804, Sigma-Aldrich), and HA-tag (1:3000) (catalog no. sc-7392,
Santa Cruz Biotechnology). Anti-mouse antibody conjugated with
horseradish peroxidase (Thermo Fisher Scientific) was used as a
secondary anti-body (dilution 1:25,000). Enhanced chemiluminescence
reagent, Lumigen ECL Ultra (TMA-6) (Lumigen, Southfield, MI, USA),
was used to detect the antigen-antibody complexes.
To assess the functional activity of AAV-cTuberin-Myc, we
cotransfected HEK293T cells, as previously described with minor
modifications (60). Plasmids included HA-tagged p70S6 kinase
(HA-p70S6K) (60), which is phosphorylated (pS6K T389) by mTORC1 and
was used as a reporter for mTORC1 activation, and Flag-tagged
hamartin (Flag-hamartin) (60), along with AAV-cTuberin-Myc.
Full-length Flag-tagged tuberin (Flag-tuberin) (60) was used as a
positive control, and AAV-GFP was used as a negative control.
Transfections were carried out for 48 hours using Lipofectamine
2000. Cell lysates were prepared using radioimmunoprecipitation
assay lysis buffer, and immunoblotting was performed, as described
(60). Briefly, proteins were separated on a Novex 4 to 12%
tris-glycine gradient gel (Life Technologies) followed by transfer
to 0.45 M nitrocellulose membrane (Bio-Rad). Antibodies included M2
anti- Flag mouse monoclonal (Sigma-Aldrich), anti-hamartin and
anti- pS6K (T389) (Cell Signaling Technology), anti-Myc mouse
monoclonal (9E10, University of Iowa Hybridoma Bank), and anti-HA
mouse monoclonal (HA.11, BioLegend/Covance, San Diego, CA,
USA).
Detection of cTuberin using Western blotHEK293T cells were
seeded in a six-well plate (500,000 cells per well) for 24 hours.
The cells were then transfected with plasmid DNAs (AAV-null,
AAV-GFP, and AAV-cTuberin) at 2.5 g/500,000 cells using
Lipofectamine 2000 in Opti-MEM. Six hours later, transfec-tion
media was removed and replaced with DMEM (10% FBS and 1% PS), and
cells were grown for 72 hours. Cells were washed twice in PBS, and
proteins were extracted with protein extraction solution (PRO-PREP,
iNtRON Biotechnology, Korea) for 20 min at −20°C. The cell
lysates were centrifuged at 14,000g at 4°C. Protein concen-trations
of cell lysates were determined using a Bio-Rad protein assay kit.
Equal amounts of protein (20 g) were separated using 4 to 12%
precast NuPAGE bis-tris SDS-PAGE gels (Invitrogen) and trans-ferred
onto nitrocellulose membranes (Thermo Fisher Scientific Inc.,
Rockford, IL, USA). Membranes were blocked for 1 hour in
tris-buffered saline (TBS) with 0.1% Tween 20 and 5% nonfat dry
milk, followed by an overnight incubation with primary antibody to
tuberin (#3990, 1:1000 dilution, Cell Signaling Technology diluted
in
on July 2, 2021http://advances.sciencem
ag.org/D
ownloaded from
http://advances.sciencemag.org/
-
Cheah et al., Sci. Adv. 2021; 7 : eabb1703 8 January 2021
S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L
E
11 of 13
the same buffer at 4°C). On the next day, the membranes were
washed with TBS with 0.1% Tween 20 (three times, 5 min each)
followed by incubation with the appropriate horseradish
peroxidase–conjugated secondary antibodies (Jackson ImmunoResearch
Laboratories, West Grove, PA, USA) for 1 hour at RT. An enhanced
chemiluminescence kit (Pierce ECL Western Blotting Substrate,
Thermo Fisher Scien-tific, Waltham, MA, USA) was used to detect
protein expression. The optical density of each band was determined
on Western blots scanned with a G:Box (Syngene, Cambridge, UK).
DNA and RNA extraction, cDNA synthesis, and qPCR for DNABrains
and livers were flash-frozen to determine AAV genome
bio-distribution and expression of transgene mRNA. Genomic and AAV
vector DNA was isolated using the Qiagen DNeasy Blood and Tissue
Kit (catalog no. 69504) according to the manufacturer’s
instruction. Total RNA was extracted using the Qiagen RNeasy Lipid
Tissue Mini Kit (catalog no.74804) and Qiagen RNeasy Mini Kit
(catalog no. 74104), with additional on-column deoxyribonuclease
(DNase) digestion with the Qiagen RNase-free DNase set (catalog no.
79254) to ensure digestion of AAV-cTuberin genomes. Then, extracted
RNA was converted to cDNA using the SuperScript VILO cDNA Synthesis
Master Mix (Thermo Fisher Scientific, catalog no. 11754-050),
ac-cording to the manufacturer’s protocol. A no-RT set of samples
for the AAV-cTuberin group was included to confirm detection of
cDNA derived from cTuberin mRNA and not contaminating AAV-cTuberin
genomes. Using 50-ng genomic DNA as template, TaqMan qPCR was
performed using custom TaqMan probe and primers to 3′ end of
cTuberin and c-Myc tag of the transgene expression cassette
(for-ward primer, 5′-AGCCAACACCAGGATACGAA-3′; reverse primer,
5′-GCTAATCAGCTTCTGCTCCAC-3′; probe, 5′-FAM-
AGCG-GCTGATCTCCTCCGTGG-MGB-3′) (fig. S5). For each sample, a
separate qPCR was performed using TaqMan probe and primer sets
(Thermo Fisher Scientific, assay ID Mm01180221_g1, gene symbol
Gm12070) that detects GAPDH genomic DNA, to ensure equal ge-nomic
DNA input for each sample. For each organ/tissue, the AAV vector
genome copies for each sample were adjusted by taking into account
any differences in GAPDH Ct values using the following formula:
(AAV vector genome copies)/(2Ct). The Ct value was calculated as
GAPDH Ct value (sample of interest) − average GAPDH Ct value
(sample with highest Ct value). Data were expressed as AAV vector
genomes per 50 ng of genomic DNA.
Animals and injectionsExperimental research protocols were
approved by the Institutional Animal Care and Use Committee for the
Massachusetts General Hospital (MGH) following the guidelines of
the National Institutes of Health for the Care and Use of
Laboratory Animals. Experiments were performed on Tsc2c/c-floxed
mice [Tsc2-floxed; (61)]. These mice have a normal, healthy life
span. In response to Cre recombinase, the Tsc2c/c alleles are
converted to null alleles. For vector injections, in the neonatal
period (P0 to P3), pups were cryo-anesthetized and injected with 1
to 2 l of viral vector AAV1-CBA-Cre into each ce-rebral lateral
ventricle with a glass micropipette (70 to 100 mm in diameter
at the tip) using a Narishige IM300 microinjector at a rate of 2.4
psi/s (Narshige International, East Meadow, NY, USA). Mice were
then placed on a warming pad and returned to their mothers after
regaining normal color and full activity typical of newborn mice.
At 3 weeks of age (P21), mice were anesthetized with isoflurane
(Baxter Healthcare, Deerfield, IL, USA) inhalation [3.5%
isoflurane
in an induction chamber and then maintained anesthetized with 2
to 3% isoflurane and oxygen (1 to 2 liters/min) for the dura-tion
of the injection]. AAV vectors were injected retro-orbitally into
the vasculature in a volume of 60 l (AAV1 or AAV9) of AAV-cTuberin-
Myc using a 0.3-ml insulin syringe over less than 2 min (62)
or noninjected.
Body weight measurement and assessment of motor activityEighteen
measurements of the body weight of the animals were re-corded from
P23 to P50. To assess motor coordination, animals were placed on an
automated rotarod apparatus (Harvard Appara-tus, Holliston, MA,
USA) using accelerated velocities (4 to 64 rpm over
120 s). Each animal was assessed three times with 5-min rest
intervals in each session for nine sessions 3 to 4 days apart. For
each assessment, the time ended when the mouse fell off the
treadmill or when the time interval elapsed. All functional
assessment tests were performed blinded with respect to the mouse
genotype.
Detection of the presence of c-Myc using
immunocytochemistryHEK293T cells were seeded on coverslip coated
with poly-d-lysine (25,000 cells per coverslip) for 24 hours. The
cells were then trans-fected with plasmid DNAs (AAV-null, AAV-GFP,
and AAV-cTuberin) at 250 ng/25,000 cells using Lipofectamine
2000 in Opti-MEM. Six hours later, transfection media was
removed and replaced with DMEM (10% FBS and 1% PS), and cells were
grown for 72 hours. The cells were fixed with 4% paraformaldehyde
(PFA) (Boston BioProducts, Ashland, MA, USA) for 10 min at RT
followed by perme-abilization using 0.01% Triton X-100
(Sigma-Aldrich) in PBS (PBST) for 10 min at RT. The cells were
then blocked with 3% bovine serum albumin (BSA) in PBST for 1 hour
at RT, followed by overnight incubation with primary antibodies at
4°C [primary antibodies: c-Myc (1:400 dilution; 9E10, Life
Technologies)] and GFP (1:400 dilution; A11122, Life Technologies).
The cells were then washed three times for 5 min in PBST and
incubated with secondary anti-body (goat anti-mouse 488, Jackson
ImmunoResearch Laboratories) (1:400 dilution), for 1 hour at RT.
The cells were washed three times for 5 min using PBST,
mounted with Vectashield containing DAPI (Vector Laboratories,
Burlingame, CA, USA). Note that, unfor-tunately, we were not able
to detect cMyc in brain sections using several sources of c-Myc
antibodies.
Histology and immunohistochemistryThe mouse brains were
harvested and subjected for standard his-tological processing as
described (14). Five-micrometer sections were stained with
hematoxylin and eosin. For frozen sections, adult mice were
euthanized using ketamine/xylazine (100:10) (Akorn Inc., Lake
Forest, IL, USA) followed by transcardiac perfusion with 1× PBS and
4% PFA in PBS overnight at 4°C, cryo-protected with 25% su-crose in
PBS, and embedded in optimal cutting temperature medium (catalog
no. 4583, Tissue Teck). Brain sections were prepared in 10-mm
coronal sections and were blocked in 10% BSA in 1×
PBS + 0.3% Triton X-100 for 1 hour at RT and subsequently
incubated with rabbit anti-Ki67 (1:1000; #ab15580, Abcam) or rabbit
anti-phospho-S6 ribosomal protein (Ser235/236) (1:400; #2211, Cell
Signaling Tech-nology) overnight at 4°C. Following three washes in
0.1× PBS, the sections were incubated with secondary antibody Alexa
555 (1:400; Jackson ImmunoResearch Laboratories) for 1 hour at RT.
The sec-tions were then washed three times with 1× PBS and mounted
with DAPI mounting medium (Vectashield, #H-1200).
on July 2, 2021http://advances.sciencem
ag.org/D
ownloaded from
http://advances.sciencemag.org/
-
Cheah et al., Sci. Adv. 2021; 7 : eabb1703 8 January 2021
S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L
E
12 of 13
pS6 puncta analysisWhole mouse brain sections immunostained for
pS6 (biological triplicates for each group, three coronal sections
per mouse) were imaged using a Nikon Ti2 inverted microscope
equipped with W1 Yokogawa Spinning disk scanhead with 50-m
pinholes, a Toptica 4 laser launch, and an Andor Zyla 4.2 Plus
sCMOS monochrome camera. The slides were mounted on a Nikon linear
encoded motorized stage, and the mouse whole brain sections were
scanned using Plan Apo 20×/0.8 differential interference contrast
(DIC) I objective lens objective lens at 405 nm for DAPI
(100-ms exposure) and 561 nm for pS6 staining (100-ms
exposure). Signals were col-lected using a Semrock
di01-t405/488/568/647 dichroic mirror and Chroma 455/50 or
605/52 nm emission filters. Images were captured using NIS AR
5.02 acquisition software and 12-bit gain four-camera setting. A
series of images were captured and stitched together us-ing
blending algorithm with 15% overlap among images.
Stitched images were analyzed in Fiji, an open source image
pro-cessing package based on ImageJ (63). All images were
thresholded within the 80 to 800 tonal range for both DAPI and pS6
staining. An outline was manually drawn to delineate choroid
plexuses, ventricles, large empty spots, and meninges from the
whole mouse brain sec-tion image. These regions are known to
contain significant amounts of autofluorescence and therefore were
excluded from downstream analysis. Within the confined region of
interests (ROIs), we mea-sured the area for the whole brain
section. To identify pS6 puncta size and intensity within them, the
thresholded pS6 channel image was converted into eight-bit image
and further thresholded within the 70 to 255 tonal range.
Subsequently, particle analysis was per-formed to identify any
puncta within 5 to 200 m2 and 0.1 to 1.0 circularity parameters.
The area for each punctum was measured. These puncta ROIs were then
used to identify raw integrated density on original unthresholded
12-bit brain section images. Normalized pS6 puncta number of a
brain section was calculated by dividing the total number of pS6
puncta by the brain section area.
Statistical analysisAll analyses of survival curves (Mantel-Cox
test and log-rank test) were performed using GraphPad Prism
software (GraphPad Software Inc., La Jolla, CA, USA). Flow
cytometry analysis on c-Myc–positive cells was analyzed using
unpaired t test. Western blot anal-ysis on pS6 and tuberin
expression levels in the mouse brain and PS6 puncta parameters were
analyzed using unpaired t test. LDH cytotoxicity assay and Western
blot analysis on relative levels of S6K T389 phosphorylation were
analyzed using one-way analysis of variance (ANOVA) test. P values
of
-
Cheah et al., Sci. Adv. 2021; 7 : eabb1703 8 January 2021
S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L
E
13 of 13
patients display hypoexcitability and synaptic deficits
associated with reduced FMRP levels and reversed by rapamycin. Mol.
Psychiatry 23, 2167–2183 (2018).
29. K. D. Winden, M. Sundberg, C. Yang, S. M. A. Wafa, S. Dwyer,
P.-F. Chen, E. D. Buttermore, M. Sahin, Biallelic mutations in TSC2
lead to abnormalities associated with cortical tubers in human
iPSC-derived neurons. J. Neurosci. 39, 9294–9305 (2019).
30. V. Ruppe, P. Dilsiz, C. S. Reiss, C. Carlson, O. Devinsky,
D. Zagzag, H. L. Weiner, D. M. Talos, Developmental brain
abnormalities in tuberous sclerosis complex: A comparative tissue
analysis of cortical tubers and perituberal cortex. Epilepsia 55,
539–550 (2014).
31. L. Meikle, D. M. Talos, H. Onda, K. Pollizzi, A. Rotenberg,
M. Sahin, F. E. Jensen, D. J. Kwiatkowski, A mouse model of
tuberous sclerosis: Neuronal loss of Tsc1 causes dysplastic and
ectopic neurons, reduced myelination, seizure activity, and limited
survival. J. Neurosci. 27, 5546–5558 (2007).
32. C. C. Dibble, B. D. Manning, The TSC1-TSC2 complex: A key
signal-integrating node upstream of TOR. Enzymes 28, 21–48
(2010).
33. J. Dong, D. Pan, Tsc2 is not a critical target of Akt during
normal Drosophila development. Genes Dev. 18, 2479–2484 (2004).
34. J. Serfontein, R. E. R. Nisbet, C. J. Howe, P. J. de Vries,
Conservation of structural and functional elements of TSC1 and
TSC2: A bioinformatic comparison across animal models. Behav.
Genet. 41, 349–356 (2011).
35. S. Momose, T. Kobayashi, H. Mitani, M. Hirabayashi, K. Ito,
M. Ueda, Y.-i. Nabeshima, O. Hino, Identification of the coding
sequences responsible for Tsc2-mediated tumor suppression using a
transgenic rat system. Hum. Mol. Genet. 11, 2997–3006 (2002).
36. F. J. DiMario Jr., Brain abnormalities in tuberous sclerosis
complex. J. Child Neurol. 19, 650–657 (2004).
37. K. Kotulska, J. Borkowska, M. Roszkowski, M. Mandera, P.
Daszkiewicz, K. Drabik, E. Jurkiewicz, M. Larysz-Brysz, K. Nowak,
W. Grajkowska, D. Domańska-Pakieła, S. Jóźwiak, Surgical treatment
of subependymal giant cell astrocytoma in tuberous sclerosis
complex patients. Pediatr. Neurol. 50, 307–312 (2014).
38. I. E. Overwater, R. Swenker, E. L. van der Ende, K. B. M.
Hanemaayer, M. Hoogeveen-Westerveld, A. M. van Eeghen, M. H.
Lequin, A. M. W. van den Ouweland, H. A. Moll, M. Nellist, M.-C. Y.
de Wit, Genotype and brain pathology phenotype in children with
tuberous sclerosis complex. Eur. J. Hum. Genet. 24, 1688–1695
(2016).
39. K. C. Ess, C. A. Kamp, B. P. Tu, D. H. Gutmann,
Developmental origin of subependymal giant cell astrocytoma in
tuberous sclerosis complex. Neurology 64, 1446–1449 (2005).
40. J. Jozwiak, S. Jozwiak, P. Skopinski, Immunohistochemical
and microscopic studies on giant cells in tuberous sclerosis.
Histol. Histopathol. 20, 1321–1326 (2005).
41. J. D. Powell, K. N. Pollizzi, E. B. Heikamp, M. R. Horton,
Regulation of immune responses by mTOR. Annu. Rev. Immunol. 30,
39–68 (2012).
42. J. M. Peters, A. Prohl, K. Kapur, A. Nath, B. Scherrer, S.
Clancy, S. P. Prabhu, M. Sahin, D. N. Franz, S. K. Warfield, D. A.
Krueger, Longitudinal effects of everolimus on white matter
diffusion in tuberous sclerosis complex. Pediatr. Neurol. 90, 24–30
(2019).
43. A. Jeong, M. Wong, mTOR inhibitors in children: Current
indications and future directions in neurology. Curr. Neurol.
Neurosci. Rep. 16, 102 (2016).
44. P. Tsai, M. Sahin, Mechanisms of neurocognitive dysfunction
and therapeutic considerations in tuberous sclerosis complex. Curr.
Opin. Neurol. 24, 106–113 (2011).
45. D. M. Feliciano, T. V. Lin, N. W. Hartman, C. M. Bartley, C.
Kubera, L. Hsieh, C. Lafourcade, R. A. O’Keefe, A. Bordey, A
circuitry and biochemical basis for tuberous sclerosis symptoms:
From epilepsy to neurocognitive deficits. Int. J. Dev. Neurosci.
31, 667–678 (2013).
46. A. M. Sokolov, C. M. Seluzicki, M. C. Morton, D. M.
Feliciano, Dendrite growth and the effect of ectopic Rheb
expression on cortical neurons. Neurosci. Lett. 671, 140–147
(2018).
47. S. W. Way, N. S. Rozas, H. C. Wu, J. M. Kenna III, R. M.
Reith, S. S. Hashmi, P. K. Dash, M. J. Gambello, The differential
effects of prenatal and/or postnatal rapamycin on
neurodevelopmental defects and cognition in a neuroglial mouse
model of tuberous sclerosis complex. Hum. Mol. Genet. 21, 3226–3236
(2012).
48. P. T. Tsai, E. Greene-Colozzi, J. Goto, S. Anderl, D. J.
Kwiatkowski, M. Sahin, Prenatal rapamycin results in early and late
behavioral abnormalities in wildtype C57BL/6 mice. Behav. Genet.
43, 51–59 (2013).
49. D. A. Krueger, A. Sadhwani, A. W. Byars, P. J. de Vries, D.
N. Franz, V. H. Whittemore, R. Filip-Dhima, D. Murray, K. Kapur, M.
Sahin, Everolimus for treatment of tuberous sclerosis
complex-associated neuropsychiatric disorders. Ann. Clin. Transl.
Neurol. 4, 877–887 (2017).
50. A. K. Bevan, S. Duque, K. D. Foust, P. R. Morales, L. Braun,
L. Schmelzer, C. M. Chan, M. M. Crate, L. G. Chicoine, B. D. Coley,
P. N. Porensky, S. J. Kolb, J. R. Mendell, A. H. M. Burghes, B. K.
Kaspar, Systemic gene delivery in large species for targeting
spinal cord, brain, and peripheral tissues for pediatric disorders.
Mol. Ther. 19, 1971–1980 (2011).
51. K. Meyer, L. Ferraiuolo, L. Schmelzer, L. Braun, V. M.
Govern, S. Likhite, O. Michels, A. Govoni, J. Fitzgerald, P.
Morales, K. D. Foust, J. R. Mendell, A. H. M. Burghes, B. K.
Kaspar, Improving single injection CSF delivery of AAV9-mediated
gene therapy for SMA: A dose-response study in mice and nonhuman
primates. Mol. Ther. 23, 477–487 (2015).
52. H. Zhang, G. Cicchetti, H. Onda, H. B. Koon, K. Asrican, N.
Bajraszewski, F. Vazquez, C. L. Carpenter, D. J. Kwiatkowski, Loss
of Tsc1/Tsc2 activates mTOR and disrupts PI3K-Akt signaling through
downregulation of PDGFR. J. Clin. Invest. 112, 1223–1233
(2003).
53. R. A. Sheth, A. S. Feldman, E. Paul, E. A. Thiele, T. G.
Walker, Angiographic and volumetric effects of mammalian target of
rapamycin inhibitors on angiomyolipomas in tuberous sclerosis.
World J. Radiol. 8, 308–315 (2016).
54. G. Gernoux, J. M. Wilson, C. Mueller, Regulatory and
exhausted T cell responses to AAV capsid. Hum. Gene Ther. 28,
338–349 (2017).
55. A. Meliani, F. Boisgerault, Z. Fitzpatrick, S. Marmier, C.
Leborgne, F. Collaud, M. S. Sola, S. Charles, G. Ronzitti, A.
Vignaud, L. van Wittenberghe, B. Marolleau, F. Jouen, S. Tan, O.
Boyer, O. Christophe, A. R. Brisson, C. A. Maguire, F. Mingozzi,
Enhanced liver gene transfer and evasion of preexisting humoral
immunity with exosome-enveloped AAV vectors. Blood Adv. 1,
2019–2031 (2017).
56. M. Sena-Esteves, J. C. Tebbets, S. Steffens, T.
Crombleholme, A. W. Flake, Optimized large-scale production of high
titer lentivirus vector pseudotypes. J. Virol. Methods 122, 131–139
(2004).
57. M. L. D. Broekman, L. A. Comer, B. T. Hyman, M.
Sena-Esteves, Adeno-associated virus vectors serotyped with AAV8
capsids are more efficient than AAV-1 or −2 serotypes for
widespread gene delivery to the neonatal mouse brain. Neuroscience
138, 501–510 (2006).
58. R. Koschny, C. Boehm, M. R. Sprick, T. L. Haas, H. Holland,
L.-X. Xu, W. Krupp, W. C. Mueller, M. Bauer, T. Koschny, M. Keller,
P. Sinn, J. Meixensberger, H. Walczak, T. M. Ganten, Bortezomib
sensitizes primary meningioma cells to TRAIL-induced apoptosis by
enhancing formation of the death-inducing signaling complex. J.
Neuropathol. Exp. Neurol. 73, 1034–1046 (2014).
59. Y. Sancak, T. R. Peterson, Y. D. Shaul, R. A. Lindquist, C.
C. Thoreen, L. Bar-Peled, D. M. Sabatini, The Rag GTPases bind
raptor and mediate amino acid signaling to mTORC1. Science 320,
1496–1501 (2008).
60. S. Prabhakar, X. Zhang, J. Goto, S. Han, C. Lai, R. Bronson,
M. Sena-Esteves, V. Ramesh, A. Stemmer-Rachamimov, D. J.
Kwiatkowski, X. O. Breakefield, Survival benefit and phenotypic
improvement by hamartin gene therapy in a tuberous sclerosis mouse
brain model. Neurobiol. Dis. 82, 22–31 (2015).
61. H. Onda, A. Lueck, P. W. Marks, H. B. Warren, D. J.
Kwiatkowski, Tsc2+/− mice develop tumors in multiple sites that
express gelsolin and are influenced by genetic background. J. Clin.
Invest. 104, 687–695 (1999).
62. T. Yardeni, M. Eckhaus, H. D. Morris, M. Huizing, S.
Hoogstraten-Miller, Retro-orbital injections in mice. Lab. Anim.
(NY) 40, 155–160 (2011).
63. J. Schindelin, I. Arganda-Carreras, E. Frise, V. Kaynig, M.
Longair, T. Pietzsch, S. Preibisch, C. Rueden, S. Saalfeld, B.
Schmid, J.-Y. Tinevez, D. J. White, V. Hartenstein, K. Eliceiri, P.
Tomancak, A. Cardona, Fiji: An open-source platform for
biological-image analysis. Nat. Methods 9, 676–682 (2012).
Acknowledgments: We thank S. McDavitt for editorial assistance,
M. F. Lee (Medical Photographer in Pathology Media Laboratory, MGH)
for imaging training, M. Zinter (Vector Core, MGH, Charlestown, MA,
USA) for AAV vector packaging, and M. Whalen for the use of the
rotarod. Funding: This work was supported by DOD Army Grant
W81XWH-13-1-0076 (to X.O.B.), NIH R01GM115552 (to M.K.), NIH NIDCD
R01DC017117-01A1 (to C.A.M.), NIH NINDS 1R61NS108232 (to X.O.B.,
C.A.M., and V.R.), and NIH NS109540 (to V.R.). We would like to
acknowledge the MGH Vector Core for the production of viral vectors
(supported by NIH/NINDS P30NS045776; B.A.T.) and P. M. Llopis,
Microscopy Resources on the North Quad (MicRoN), Harvard Medical
School, NRB-Longwood, MA, USA. Author contributions: X.O.B., S.P.,
D.Y., C.A.M., and M.K. conceived and designed the experiments.
S.P., P.-S.C., R.L.B., X.Z., and S.K. performed the experiments.
S.P., P.-S.C., K.-H.L., and S.K. analyzed the data. S.P., P.-S.C.,
D.Y., B.A.T., E.A.T., X.Z., R.L.B., R.T.B., D.J.K., A.S.-R., B.G.,
K.-H.L., V.R., M.K., C.A.M., and X.O.B. wrote and edited the paper.
Competing interests: X.O.B., S.P., D.Y., and C.A.M. have filed a
provisional patent application for the cTuberin construct. C.A.M.
has a financial interest in Chameleon Biosciences Inc., a company
developing an enveloped AAV vector platform technology for repeated
dosing of systemic gene therapy. X.O.B., V.R., and C.A.M.’s
interests are reviewed and managed by MGH and Partners HealthCare
in accordance with their competing interest policies. All other
authors declare that they have no competing interests. Data and
materials availability: All data needed to evaluate the conclusions
in the paper are present in the paper and/or the Supplementary
Materials. Plasmid requests can be provided by MGH pending
scientific review and a completed material transfer agreement.
Requests for the plasmid should be submitted to C.A.M. at
[email protected].
Submitted 4 February 2020Accepted 18 November 2020Published 8
January 202110.1126/sciadv.abb1703
Citation: P.-S. Cheah, S. Prabhakar, D. Yellen, R. L. Beauchamp,
X. Zhang, S. Kasamatsu, R. T. Bronson, E. A. Thiele, D. J.
Kwiatkowski, A. Stemmer-Rachamimov, B. György, K.-H. Ling, M.
Kaneki, B. A. Tannous, V. Ramesh, C. A. Maguire, X. O. Breakefield,
Gene therapy for tuberous sclerosis complex type 2 in a mouse model
by delivery of AAV9 encoding a condensed form of tuberin. Sci. Adv.
7, eabb1703 (2021).
on July 2, 2021http://advances.sciencem
ag.org/D
ownloaded from
http://advances.sciencemag.org/
-
encoding a condensed form of tuberinGene therapy for tuberous
sclerosis complex type 2 in a mouse model by delivery of AAV9
Kaneki, Bakhos A. Tannous, Vijaya Ramesh, Casey A. Maguire and
Xandra O. BreakefieldBronson, Elizabeth A. Thiele, David J.
Kwiatkowski, Anat Stemmer-Rachamimov, Bence György, King-Hwa Ling,
Masao Pike-See Cheah, Shilpa Prabhakar, David Yellen, Roberta L.
Beauchamp, Xuan Zhang, Shingo Kasamatsu, Roderick T.
DOI: 10.1126/sciadv.abb1703 (2), eabb1703.7Sci Adv
ARTICLE TOOLS
http://advances.sciencemag.org/content/7/2/eabb1703
MATERIALSSUPPLEMENTARY
http://advances.sciencemag.org/content/suppl/2021/01/04/7.2.eabb1703.DC1
REFERENCES
http://advances.sciencemag.org/content/7/2/eabb1703#BIBLThis
article cites 62 articles, 8 of which you can access for free
PERMISSIONS
http://www.sciencemag.org/help/reprints-and-permissions
Terms of ServiceUse of this article is subject to the
is a registered trademark of AAAS.Science AdvancesYork Avenue
NW, Washington, DC 20005. The title (ISSN 2375-2548) is published
by the American Association for the Advancement of Science, 1200
NewScience Advances
License 4.0 (CC BY-NC).Science. No claim to original U.S.
Government Works. Distributed under a Creative Commons Attribution
NonCommercial Copyright © 2021 The Authors, some rights reserved;
exclusive licensee American Association for the Advancement of
on July 2, 2021http://advances.sciencem
ag.org/D
ownloaded from
http://advances.sciencemag.org/content/7/2/eabb1703http://advances.sciencemag.org/content/suppl/2021/01/04/7.2.eabb1703.DC1http://advances.sciencemag.org/content/7/2/eabb1703#BIBLhttp://www.sciencemag.org/help/reprints-and-permissionshttp://www.sciencemag.org/about/terms-servicehttp://advances.sciencemag.org/