of 41
Gut-seeded
α
-synuclein fibrils promote gut dysfunction and
brain pathology specifically in aged mice
Collin Challis
1
,
Acacia Hori
1,#
,
Timothy R. Sampson
1,2,#
,
Bryan B. Yoo
1
,
Rosemary C.
Challis
1
,
Adam M. Hamilton
2
,
Sarkis K. Mazmanian
1
,
Laura A. Volpicelli-Daley
3
,
Viviana
Gradinaru
1,*
1
Division of Biology & Biological Engineering, California Institute of Technology, Pasadena,
California, USA, 91125
2
Department of Physiology, Emory University School of Medicine, Atlanta, Georgia, USA, 30322
3
Center for Neurodegeneration and Experimental Therapeutics, University of Alabama at
Birmingham, Birmingham, Alabama, USA, 35294
Abstract
Parkinson’s disease (PD) is a synucleinopathy that is characterized by motor dysfunction, death of
midbrain dopaminergic neurons, and accumulation of alpha synuclein (
α
-Syn) aggregates.
Evidence suggests that
α
-Syn aggregation can originate in peripheral tissues and progress to the
brain via autonomic fibers. We tested this by inoculating the duodenal wall of mice with
α
-Syn
preformed fibrils. Following inoculation, we observed gastrointestinal deficits and physiological
changes to the enteric nervous system. We also found that
α
-Syn pathology is reduced by
increased expression of the lysosomal enzyme glucocerebrosidase, using the AAV-PHP.S capsid to
target this protein for peripheral gene transfer. Lastly, inoculation of
α
-Syn fibrils in aged mice,
but not younger mice, resulted in progression of
α
-Syn histopathology to the midbrain and
subsequent motor defects. Our results characterize peripheral synucleinopathy in prodromal PD
and explore cellular mechanisms for the gut-to-brain progression of
α
-Syn pathology.
Users may view, print, copy, and download text and data-mine the content in such documents, for the purposes of academic research,
subject always to the full Conditions of use:http://www.nature.com/authors/editorial_policies/license.html#terms
*
Correspondence and materials request: viviana@caltech.edu,
Contact info:
Prof. Viviana Gradinaru, Ph.D., Professor of
Neuroscience and Biological Engineering, Heritage Principal Investigator, Director, Molecular and Cellular Neuroscience Center of
the Chen Institute, Division of Biology and Biological Engineering, California Institute of Technology, Pasadena, California, USA,
91125, Phone: (626) 395-6813.
#
Authors contributed equally to this work
Author Contributions.
C.C. and V.G. conceptualized the study and developed the research plan. C.C., V.G., and L.A.V. designed the study. L.A.V. generated
the
α
-Syn PFFs and
α
-Syn monomers. C.C performed animal surgeries, tissue clearing, histology, calcium imaging, and retro-orbital
viral injections. C.C., A.H., and T.R.S. performed behavior experiments. C.C., B.B.Y., and R.C.C. performed virus production,
purification, and verification. C.C. and T.R.S. performed protein analysis. C.C., B.B.Y., and R.C.C. performed confocal imaging.
A.M.H. and T.R.S. performed RNA extraction and qPCR analysis. C.C. performed data analysis. S.K.M. provided key reagents and
methods. C.C. and V.G. wrote the manuscript. All authors contributed to discussion. V.G. supervised all work.
Declaration of Interests.
All authors declare no competing interests.
HHS Public Access
Author manuscript
Nat Neurosci
. Author manuscript; available in PMC 2020 August 17.
Published in final edited form as:
Nat Neurosci
. 2020 March ; 23(3): 327–336. doi:10.1038/s41593-020-0589-7.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
INTRODUCTION
Synucleinopathies are neurodegenerative diseases characterized by the aggregation of
insoluble amyloid
α
-Synuclein (
α
-Syn) fibrils
1
.
α
-Syn accumulation in specific cell
populations precipitates distinct clinical phenotypes that serve as the basis for diagnosis of
synucleinopathies. For example, the appearance of
α
-Syn pathology in midbrain
dopaminergic neurons coincides with motor dysfunction in Parkinson’s disease (PD)
2
.
However, mounting evidence suggests that synucleinopathy diagnosis occurs late in the
disease progression and that pathology may originate much earlier in the gastrointestinal
(GI) tract before progressing to the brain
3
. Biopsies of GI tissue from both PD patients and
healthy individuals have found
α
-Syn accumulation in the stomach, duodenum, and
colon
4
,
5
. This accumulation has been observed in cells of the enteric nervous system (ENS),
an organized network of interconnected ganglia comprised of neurons and enteric glial cells
(EGCs) that spans the entire GI tract
6
. One hypothesis is that pathologic
α
-Syn ascends
ENS-innervating vagal fibers to the nodose ganglion and brainstem nuclei
7
. Once in the
brain,
α
-Syn pathology would then propagate through interconnected neurons to reach the
midbrain and cause motor dysfunction. Recent studies show that
α
-Syn fibrils are capable of
progressing through vagal fibers that innervate the gut and can cause behavioral
dysfunction
8
10
. However, how pathologic
α
-Syn impacts the enteric nervous system (ENS),
and the mechanisms that underlie their spread, have not been comprehensively studied.
Inflammation is linked to the peripheral etiology of PD, as pro-inflammatory cytokines are
elevated in colonic biopsies and stool samples from PD patients
11
. In addition,
in vitro
work
has shown that aggregated
α
-Syn directly activates inflammatory pathways
12
. Macrophages
are recruited as a component of the inflammatory process and are essential for removing
pathogens, including misfolded protein aggregates
13
. The lysosome is a major macrophagic
regulator of protein homeostasis and pathogen clearance, and has been shown to interact
with aggregated
α
-Syn
14
. Glucocerebrosidase (GCase) is a lysosomal enzyme and disease-
associated mutations within its gene,
GBA1
, result in impaired lysosomal function. The
connection between GCase and
α
-Syn was discovered in patients with Gaucher’s disease, a
metabolic disorder caused by loss of GCase function, who presented with elevated levels of
α
-Syn in the brain
15
. Further studies have gone on to demonstrate that impaired GCase
function also results in the abnormal accumulation of
α
-Syn
16
. Interestingly, pathologic
α
-
Syn disrupts trafficking of GCase to the lysosome, which is hypothesized to contribute to a
positive feedback loop for
α
-Syn aggregation
17
. Thus, GCase activity could be critical
during prodromal PD to prevent
α
-Syn accumulation from reaching a pathogenic, self-
propagating threshold.
In this work, we aimed to characterize the mechanisms that underlie synucleinopathy in the
GI tract. We showed that inoculating duodenal intestinal lining with
α
-Syn preformed fibrils
(PFFs) induced the formation of phosphorylated
α
-Syn (p-
α
-Syn)-containing inclusions,
promoted a local inflammatory response, and disrupted ENS connectivity. We also observed
progression of
α
-Syn histopathology from the ENS to the brainstem and reduced levels of
striatal dopamine in aged, but not young mice after PFF inoculation. Finally,
α
-Syn
pathology and ENS dysfunction is recovered by increased expression of GCase, reinforcing
the role of GCase as a critical regulator of pathologic
α
-Syn
18
. Our results reveal
Challis et al.
Page 2
Nat Neurosci
. Author manuscript; available in PMC 2020 August 17.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
mechanisms that contribute to peripheral, non-motor symptoms reflective of prodromal
synucleinopathy and demonstrate how gut-seeded
α
-Syn fibrils promote the progression of
α
-Syn pathology to the brain in an age-dependent manner.
RESULTS
Duodenal
α
-Syn PFF inoculation disrupts GI function and initiates inflammatory responses
One of the most common non-motor symptoms observed in PD is GI dysfunction, which
often precedes the development of motor symptoms
19
. However, it is not fully understood
whether
α
-Syn pathology causes this disruption. To directly test whether pathologic
α
-Syn
interrupts ENS control of GI function, we modeled prodromal peripheral synucleinopathy by
seeding the duodenum of adult (8–10 week old), wild type (WT) mice with mouse
α
-Syn
PFFs
20
via intramuscular injection (Fig. 1a). The use of
α
-Syn PFFs is established as a
model of idiopathic synucleinopathy in the CNS that closely represents human
pathology
21
,
22
. We targeted the duodenum because of its dense innervation by vagal fibers,
which is speculated to be a critical component of the gut-to-brain hypothesis of
synucleinopathy etiology
7
. Duodenal inoculation with
α
-Syn PFFs resulted in time-
dependent GI dysfunction through 120 days post inoculation (dpi), comparable to the
phenotypes reported in transgenic models of synucleinopathy, including the ASO (Thy1-
α
-
Syn overexpressing) line
23
,
24
(Fig. 1b–d, Extended Data Fig. 1a–d). Interestingly,
inoculation with
α
-Syn monomers did not produce an effect, which we observed repeatedly
throughout our study.
To investigate the immediate physiological response to PFF seeding, we evaluated duodenal
tissue lysates against a cytokine panel at 7 dpi and observed increased production of several
pro-inflammatory cytokines (Fig. 1e, Extended Data Fig. 1g–h), including interleukin-6
(IL-6), which was found to be upregulated in colonic tissue from PD paients
11
. Additional
analyses showed a time-dependent increase in duodenal IL-6 production after
α
-Syn PFF
inoculation (Fig. 1f, Extended Data Fig. 1i). Because IL-6 has been demonstrated to promote
enteric neuronal survival
25
, we next wanted to determine how PFF seeding impacts the ENS.
Histological quantification revealed a transient decrease in neuronal volume within duodenal
myenteric ganglia that recovered by 21 dpi (Fig. 1g–h, Extended Data Fig. 2a–c). To
determine whether changes in volume were due to neuronal death followed by repopulation,
we quantified counts of enteric neurons, however we did not observe a statistically
significant decrease at 7 dpi (Extended Data Fig. 2d). Evaluation of myenteric EGCs
revealed a prolonged elevation of EGC volume and an increase in EGC count over time
following
α
-Syn PFF inoculation, indicating a reactive gliosis in response to fibril seeding
(Fig. 1i, Extended Data Fig. 2d). To further understand the changes in EGC volume and cell
count, we labeled with 5-ethynyl-2’-deoxyuridine (EdU), which revealed an immediate
increase in cellular proliferation at 7 dpi following
α
-Syn PFF inoculation compared to
monomer (Extended Data Fig 2e–f). While the majority of these cells were extragangliar
(90.4 ± 1.9%), those that were located in myenteric ganglion were predominantly GFAP
+
(91.2 ± 3.7%).
Activated EGCs, as well as enteric neurons, can mediate cytokine signaling in the gut to
recruit immune cells
6
. Our cytokine screen also found a significant increase in fractalkine
Challis et al.
Page 3
Nat Neurosci
. Author manuscript; available in PMC 2020 August 17.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
(Fig. 1e), which recruits macrophages that support enteric neurons
26
, and macrophage
colony-stimulating factor (MCSF), which promotes macrophage differentiation and
recruitment in the gut
27
. Subsequent analysis revealed a time-dependent increase in
expression of the macrophage marker Iba1 in the duodenum following
α
-Syn PFF
inoculation (Fig. 1j–k). Taken together, these results suggest that gut seeded
α
-Syn fibrils
initiate an inflammatory response and activate immune cell-recruiting cytokines to maintain
a healthy enteric neuronal network.
Duodenal
α
-Syn PFF inoculation increases
α
-Syn histopathology
Duodenal inoculation with
α
-Syn PFFs induced an increase in
α
-Syn phosphorylation at the
serine 129 (S129P) residue (Fig. 2a–d), which is a known marker of pathologic
α
-Syn
accumulation
20
,
28
that has neurotoxic properties whether present in soluble or insoluble
aggregates
29
,
30
. We observed peak S129P signal at 60 dpi before declining at 120 dpi,
suggesting that protein homeostasis mechanisms are engaged to counter the progression of
α
-Syn pathology. A similar trend was observed in immunoblots using a conformation-
specific antibody that detects
α
-Syn filaments involved in fibril formation (Extended Data
Fig 3). The lysosome is implicated in
α
-Syn homeostasis and enhanced lysosome response
by increased GCase activity promotes the degradation of
α
-Syn aggregates
18
. Interestingly,
pathologic
α
-Syn also inhibits GCase function
17
. Thus, we next determined the effect of
PFF seeding on duodenal GCase. We observed a significant decrease in GCase production at
7 dpi, however this reversed at 21 dpi, and by 120 dpi GCase production was recovered to
levels comparable to WT conditions (Fig. 2c,e).
GBA1
gene transfer partially rescues the pathologic
α
-Syn-induced GI phenotype
Our findings support the previously hypothesized positive feedback loop between pathologic
α
-Syn accumulation and decreased GCase function
17
in the gut. Because we did not observe
permanent pathologic GI effects of
α
-Syn PFF inoculation in adult WT mice, we evaluated
GCase in the ASO mouse model due to its constitutive overproduction of
α
-Syn and
chronically elevated p-
α
-Syn
24
. We also observed significantly decreased levels of duodenal
GCase production in ASO mice (Fig. 2e), which display severe GI deficits
31
(Extended Data
Fig. 1a–d) and have elevated duodenal S129P (Fig. 2b,d). These findings highlight
GBA1
,
the gene encoding GCase, as a therapeutic target for peripheral synucleinopathy. Mutations
in
GBA1
are among the most common risk factors for PD
15
and gene transfer of
GBA1
in
the CNS of mice has been shown to reduce
α
-Syn pathology
32
.
To test whether
GBA1
gene transfer can ameliorate peripheral synucleinopathology, we
packaged a vector system that uses the tetracycline-off transactivator (tTA) to achieve rapid
and robust neuronal expression of GCase (
ihSyn-tTA:TRE-GBA1
) in a novel viral capsid
(AAV-PHP.S) that has high transduction efficiency for the peripheral nervous system and
does not cross the blood brain barrier
33
,
34
(Fig. 2f). Systemic delivery of the vector system
into ASO mice transduced 49.6 ± 10.7% of enteric neurons and achieved long-term recovery
of GCase production in the duodenum (Fig. 2g–h). Gene transfer of
GBA1
also resulted in a
reduction in duodenal p-
α
-Syn at 60 days post viral injection (dpvi) and partially recovered
the GI phenotype observed in ASO mice (Fig. 2i–k, Extended Data Fig. 4). The incomplete
Challis et al.
Page 4
Nat Neurosci
. Author manuscript; available in PMC 2020 August 17.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
functional recovery emphasizes the complexity of synucleinopathy pathology. Together,
these results suggest that gut GCase helps reduce
α
-Syn pathology and recover gut motility.
ENS network connectivity is disrupted by
α
-Syn PFF inoculation and partially rescued by
neuronal
GBA1
gene transfer
It is likely that pathologic
α
-Syn contributes to GI dysfunction by directly affecting ENS
connectivity as
α
-Syn fibrils disrupt neurotransmission prior to significant accumulation
35
.
To explore this, we used an optogenetic strategy to interrogate ENS functional connectivity.
Unlike conventional electrophysiological ENS recording techniques, an optogenetic
approach better maintains the integrity of the duodenal ENS
in situ
. We used AAV-PHP.S to
deliver a tunable triple vector system where an inducer (
ihSyn-tTA
) drives expression of
Channelrhodopsin (
TRE-ChR2-EYFP
) and the red-shifted calcium indicator jRGECO1a
(
TRE-jRGECO1a
) (Fig. 3a). This strategy allowed us to activate a subpopulation of enteric
neurons (i.e. ChR2-EYFP
+
cells) while recording the response activity of individual cells
within the ENS network (i.e. jRGECO1a
+
cells). Calcium imaging has previously been used
to image ENS activity
36
, however, controlled genetic expression of indicators and effectors
in the ENS independent of transgenics has not been previously performed. We injected
different viral doses of each vector to achieve abundant expression of jRGECO1a (41.6 ±
7.5% of PGP9.5
+
neurons) and sparse expression of ChR2-EYFP (8.3 ± 2.9% of PGP9.5
+
neurons) in the ENS for
ex vivo
recording (Fig. 3b–d and Supplementary Video 1). This
ratio allowed us to investigate broad network activity in response to the activation of a small
subset of enteric neurons as an indirect measure of ENS network health. Following
α
-Syn
PFF seeding, the average response of jRGECO1a
+
-only duodenal neurons to sparse enteric
network photostimulation was decreased, suggesting dysfunctional connectivity (Fig. 3e–i
and Extended Data Fig. 5a). Our optogenetic strategy also allowed us to assess the firing
properties of directly photoactivated enteric neurons (i.e. ChR2
+
/jRGECO1a
+
cells), of
which we observed a progressive decrease in the light-activated calcium response, indicating
a reduced ability of these neurons to fire. Duodenal enteric neurons in ASO mice displayed a
similar deficit in network connectivity (Fig. 3j). Following AAV-PHP.S-mediated
GBA1
gene transfer in ASO mice we were able to partially recover enteric network connectivity,
confirming the deleterious effect of
α
-Syn pathology on enteric neuronal physiology and
potential for peripheral
GBA1
as a therapeutic target (Fig. 3j and Extended Data Fig. 5b).
Duodenal
α
-Syn PFF inoculation in aged mice promotes progression of
α
-Syn
histopathology to the brain
Finally, we determined if PFF-induced seeding of pathologic
α
-Syn in the duodenum
resulted in progression of
α
-Syn pathology to the CNS. We first evaluated p-
α
-Syn in PACT
cleared
37
nodose ganglia, and though the number of S129P
+
neurons had increased, this was
not statistically significant compared to pre-inoculation (Extended Data Fig. 6). Likewise,
we did not observe marked increases in p-
α
-Syn in the brainstem or SNc following duodenal
α
-Syn PFF inoculation (Extended Data Fig. 7a,c,e–f). This was in contrast to tissues from
ASO mice, where p-
α
-Syn was more pronounced in each region (Extended Data Fig.
7b,d,e–f). We also characterized sensorimotor behaviors after
α
-Syn PFF inoculation, and
did observe behavioral impairments in the adhesive removal, pole descent, and beam
Challis et al.
Page 5
Nat Neurosci
. Author manuscript; available in PMC 2020 August 17.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
traversal paradigms at 60 dpi and 90 dpi in the PFF-inoculated cohort, however these effects
reverted by 120 dpi (Extended Data Fig. 8a–f).
Given that
α
-Syn histopathology is not observed in the brain after duodenal PFF inoculation
and that pathological progression peaked at 60 dpi, protein homeostasis mechanisms may be
preventing gut-to-brain progression of
α
-Syn pathology. Aging is considered the principal
risk factor for neurodegenerative disease as cellular mechanisms are unable to efficiently
mitigate protein pathology
38
. In the duodenum of aged mice (16-month old) we observed
decreased GCase production compared to younger adult WT mice (8–10 weeks old), and a
statistically not significant decrease in levels of
GBA1
transcript (Fig. 2c, 4a, and Extended
Data Fig. 1f). These findings implicate a reduced capacity of GCase to degrade pathologic
α
-Syn in the duodenum of aged mice. Indeed, a greater proportion of the aged cohort
presented with p-
α
-Syn
+
duodenal enteric neurons compared to the younger cohort, though
the quantified levels detected were intermediate to those observed in younger adults and
ASO mice (Fig. 2c, 4b–c). The lack of significant difference in GI function suggests that a
diminished capacity for
α
-Syn homeostasis alone does not result in GI pathology (Extended
Data Fig. 1a–d).
Based on the decrease in GCase production, we hypothesized that seeding
α
-Syn PFFs in
the gut of aged mice would exacerbate
α
-Syn-induced pathology compared to younger mice.
Duodenal inoculation with
α
-Syn PFFs in aged mice resulted in a decline in GI function
(Fig. 4d–f) as well as sensorimotor deficits up to 120 dpi (Fig. 4g–i). We employed animal
weight-independent paradigms, as aged mice are significantly heavier than adults (Extended
Data Fig. 1e). Additional behavioral testing determined these deficits were not caused by a
heightened pain response (Extended Data Fig. 8g). Next, we explored the gut-to-brain
progression of
α
-Syn pathology and observed a significant increase in brainstem p-
α
-Syn at
120 dpi in
α
-Syn PFF-inoculated aged mice compared to pre-inoculation (Fig. 4j,l).
However, the increase in the midbrain was not statistically significant and the number of
tyrosine hydroxylase-positive cells was unchanged (Fig. 4k,m, Extended Data Fig. 9).
Because neurotransmission deficits precede accumulation of pathologic
α
-Syn
35
,
39
, we
evaluated total striatal dopamine levels to determine SNc activity
40
. We found a significant
decrease in striatal dopamine in aged mice inoculated with
α
-Syn PFFs, but not in
α
-Syn
PFF-inoculated younger WT mice nor in monomer-inoculated aged mice (Fig. 4n). At 120
dpi, striatal dopamine levels in
α
-Syn PFF-seeded aged mice were comparable to that
observed in 12-month-old ASO mice, but not younger ASO mice, which further implicates
age-related factors in the progression of
α
-Syn pathology.
DISCUSSION
Approximately 90% of PD diagnoses are idiopathic
41
, yet the factors that underlie
pathogenesis remain unclear. Interactions between genetic and environmental factors, such
as toxins, likely trigger pathogenesis by initiating
α
-Syn oligomerization, aggregation, and
propagation
42
. This is supported by observations of
α
-Syn-containing inclusions in
peripheral tissue from PD-patients and otherwise healthy individuals
4
,
5
. Understanding the
impact of pathologic
α
-Syn on the peripheral nervous system is critical for developing novel
diagnostic and therapeutic tools that target early synucleinopathy. Such tools could allow the
Challis et al.
Page 6
Nat Neurosci
. Author manuscript; available in PMC 2020 August 17.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
progression of pathology to be halted early in disorders such as PD before the onset of motor
symptoms. However, the progression and mechanisms of
α
-Syn pathology in the periphery
has not been thoroughly characterized. In this work, we determined that
α
-Syn fibrils
inoculated in the duodenum activate an inflammatory response and disrupt ENS physiology
and GCase function. We also demonstrated age-dependent progression of
α
-Syn
histopathology from the gut to the brain. Finally, increasing levels of GCase reduced
α
-Syn-
induced pathology and functional deficits, suggesting viral-based gene transfer as a potential
therapeutic strategy for idiopathic PD. Our results propose mechanisms that may underlie
the etiology of sporadic PD and highlight
GBA1
as a therapeutic target for prodromal,
peripheral synucleinopathy.
Our work reinforces the role of the lysosomal enzyme GCase as a critical regulator of
α
-Syn
pathology in the periphery. It is known that oligomeric
α
-Syn disrupts lysosomal
trafficking
17
, however the comorbidity of Gaucher’s disease and PD is only 10%
43
,
suggesting that GCase deficiency alone may not be sufficient to initiate
α
-Syn aggregation.
Here, we show that GCase production is significantly reduced in aged mice, which correlates
with an elevation in p-
α
-Syn (Fig. 4a–c). Interestingly, the average age of PD symptom
onset in Gaucher’s patients is up to 6 years earlier than in idiopathic PD patients alone
44
.
Thus, a proposed mechanism for PD pathogenesis is a feedforward loop between GCase and
α
-Syn that results in a gradual accumulation of amyloid
α
-Syn and diminution of GCase
until a pathological threshold is crossed
17
. The lower levels of duodenal GCase we observed
in the aged mice cohort may implicate a susceptibility to the progression of
α
-Syn pathology
following gut
α
-Syn fibril seeding. Whether this interaction occurs in the GI tract has not
been determined. Colonic dysmotility has been observed in a double mutant mouse line
(hSNCA
A53T
/GBA
L444P/+
), however protein interactions within the tissue were not
described
16
. Here, we reported that duodenal
α
-Syn PFF inoculation decreases GCase
production (Fig. 2e), which may be a contributing factor in the peripheral progression of
synucleinopathy. We did observe recovery of GCase production at later time points,
suggesting that otherwise healthy adult mice possess some resilience to pathological insults.
It may be likely that the
α
-Syn PFF dose with which we chose to inoculate allowed this
effect to be observed, as other studies that utilized much higher amounts for inoculation
reported more prominent pathology
9
,
10
. That we see diminished baseline GCase production
in the duodenum of aged mice suggests that an aged population may be more readily
susceptible to
α
-Syn pathology in the GI tract.
In the duodenum of ASO mice, we observed similar pathology to that of
α
-Syn PFF-
inoculated mice, demonstrating the utility of the ASO mouse line in studying peripheral
synucleinopathy. Additionally, because the effects of PFF inoculation reverted after 60 dpi,
we believed the chronicity of
α
-Syn-induced pathology in ASO mice represented a better
opportunity to study peripheral
GBA1
gene transfer as a therapeutic intervention. Indeed, we
were able to use systemic delivery of AAVs packaged with the AAV-PHP.S capsid
33
to
produce broad and uniform non-invasive peripheral gene transfer in ASO mice that
recovered pathologic
α
-Syn-induced GI dysfunction (Fig. 2f–k). This is in line with
previous studies that found that
GBA1
gene transfer reduced levels of pathologic
α
-Syn in
the brains of A53T transgenic mice
32
,
45
. We were unable to completely restore GI function,
indicating that
α
-Syn pathology targets multiple mechanisms or that in the ASO mouse line,
Challis et al.
Page 7
Nat Neurosci
. Author manuscript; available in PMC 2020 August 17.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
genetic intervention must occur before
α
-Syn accumulates beyond an irreversible pathologic
threshold. An effective therapy may require combinatorial treatments, such as the inclusion
of GCase chaperone proteins, and precise timing with respect to progression of pathology.
Additionally, our strategy increased GCase production solely in peripheral neurons; it may
be necessary to target additional cell types such as EGCs and macrophages for a greater
effect. It is also possible that GCase function in the CNS plays a role in gut homeostasis.
After duodenal
α
-Syn PFF inoculation, we observed a disruption in ENS connectivity and
activation of an inflammatory response, both of which may serve as indicators of prodromal
PD. The role of inflammation in synucleinopathies is complex, though evidence suggests it
is an initiating or propagating factor rather than protective
46
. In the ENS, cytokines can be
released by EGCs
47
, and a positive correlation between pro-inflammatory cytokines and
glial markers was observed in colonic biopsies from PD patients
11
. Our findings are in line
with these studies and support
α
-Syn fibril-mediated increases in EGC tone and
inflammation (Fig. 1e–k). Regarding synaptic connectivity, studies have shown that
dopaminergic signaling from the SNc is diminished in early synucleinopathy, prior to the
formation of Lewy aggregates
35
,
39
. We corroborated these findings as we observed
decreased striatal dopamine in aged mice without significantly increased p-
α
-Syn in the SNc
following
α
-Syn PFF inoculation in the duodenum (Fig. 4m–n). The effect of pathologic
α
-
Syn on ENS connectivity is not as well understood. Here, we show that
α
-Syn PFF
inoculation decreased ENS network signaling (Fig. 3g–i), which likely played a role in the
accompanying dysfunctional GI phenotype (Fig. 1b–d). Because
α
-Syn pathology is capable
of progressing transsynaptically, it may also be likely that propagation to innervating
autonomic fibers and the distal GI tract is responsible for altered nutrient and water
homeostasis, leading to the observed change in fecal composition. The presence of
α
-Syn
aggregates may also cause dysfunction in enteroendocrine cells as these cell types also
produce
α
-Syn and modulate gut motility and absorption
48
. These peripheral measurements
may be valuable diagnostic markers the for early detection of synucleinopathies.
In this work, we validated previous studies by showing that duodenal
α
-Syn PFF inoculation
in adult mice advanced pathology towards the CNS. However, in younger mice, we observed
termination of this progression in the brainstem (Extended Data Fig. 7). Findings from
Uemura et al., mirror our own as they observed an initial increase in brainstem pathology at
45 dpi followed by a steady decline at 4, 8, and 12 months post-injection
9
. More recent work
by Kim et al., published in June 2019, found that intramuscular injection of
α
-Syn PFF in
the stomach and duodenum resulted in significant midbrain pathology and sensorimotor
deficits at 7 months post-injection
10
. Major differentiating factors between these studies and
ours are different injection locations (duodenum alone in our work; stomach alone in
Uemura et al.; upper duodenum and pyloric stomach in Kim et al.), the time points observed
(4 months post-inoculation in our work; 12 months post-inoculation in Uemura et al.; 10
months post-inoculation in Kim et al.), and, importantly, the amount of
α
-Syn PFF injected
into the GI wall. While Kim et al. introduced 25 μg of material, we selected a modest 6 μg
injection quantity to ask how a small “seeding” dose would specifically affect enteric
physiology, and whether it was sufficient to promote gut-to-brain progression of
α
-Syn
pathology. Comparing our results from younger adult and aged mice suggests that as the
protein homeostatic function in the gut declines during aging
49
, so does the ability to
Challis et al.
Page 8
Nat Neurosci
. Author manuscript; available in PMC 2020 August 17.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
eliminate
α
-Syn aggregates, promoting CNS pathology and sensorimotor deficits (Fig. 4).
Interestingly, Uemura et al., introduced 48 μg of
α
-Syn PFF that did result in brainstem
α
-
Syn pathology, but not a further caudo-rostral progression. A possible factor underlying this
discrepancy is their selection of the stomach for inoculation, which has a much larger
surface area versus the duodenum and differential innervation by autonomic fibers both
compared to the duodenum and between different gastric regions
50
. The parameters we
chose for this study allowed us to better understand the dynamic relationship between
α
-Syn
and GCase in the peripheral nervous system.
In summary, our findings suggest that age-related declines in protein homeostasis, including
diminished GCase function, may promote susceptibility to
α
-Syn pathology in the ENS and
support the gut-to-brain hypothesis of synucleinopathy etiology. Understanding
vulnerabilities in peripheral systems to pathologic
α
-Syn will advance the development of
early detection techniques and therapeutics. Towards this, we demonstrated that
α
-Syn
fibrils directly disrupt ENS connectivity, which may signify a physiological signature of
prodromal synucleinopathy. We also highlighted the potential of using AAV-mediated
peripheral gene transfer for intervention by restoring GCase expression in enteric neurons.
Taken together, our work shifts the focus of neurodegenerative disease etiology to the
peripheral nervous system and expands our understanding of the role the ENS plays in
prodromal synucleinopathy.
METHODS
Animals.
Wild type (WT) C57BL/6N mice were obtained from Charles River (Hollister, CA) and bred
in-house. Adult mice were 8–10 weeks old. Aged mice were WT mice allowed to age to 16
months before experimental manipulation or C57BL/6N mice obtained through the
NIH/NIA Aged Rodent Colony. To generate Thy1-
α
-synuclein overexpressing (ASO) mice,
female BDF1/Thy1-ASO animals heterozygous for the Thy1-
α
-Syn transgene on the X
chromosome were crossed with WT male BDF1 mice
24
,
51
. Male BDF1 were bred by
crossing C57BL/6N females with DBA/2 males (Charles River). Unless otherwise stated,
ASO mice used in experiments were 6 months old. Male mice were used in all experiments
because of insertion of the Thy1-
α
-Syn transgene into the X chromosome, which results in
significantly masked phenotype and pathology
24
. Animals were housed on a reverse light-
dark cycle. Care and experimental manipulation of animals were in accordance with the
National Institutes of Health Guide for the Care and Use of Laboratory Animals and
approved by the Caltech Institutional Animal Care and Use Committee. For each time
course experiment, 1–3 cohorts of animals were used, with littermates randomized to the
appropriate groups prior to surgical manipulation. Mice from each condition were chosen at
random and sacrificed at the indicated time points to collect tissue for analysis and
comparison. Data collection and analysis were not performed blind to the conditions of the
experiments.
Challis et al.
Page 9
Nat Neurosci
. Author manuscript; available in PMC 2020 August 17.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
Tests of gastrointestinal and sensorimotor function.
All assessment was performed between hours 7 and 9 of the dark cycle in laminar-flow
biosafety cabinets within the animal facility. Tests were performed over 2 days. On day one,
the fecal output, beam traversal, pole descent, and wire hang tests were performed. On day
two, the weightlifting, adhesive removal, hind limb tests, and whole gut transit assay were
performed. Animals were sacrificed immediately after the last test for tissue collection.
Fecal output
was performed by placing mice in covered 12 cm diameter × 25 cm height
translucent cylinders for 15 min. Following conclusion of the test, fecal pellets were blotted
on paper towels to absorb urine that may have been expelled during the task. Pellets were
weighed, desiccated at room temperature over 3 days, then reweighed to determine total
weight and water weight per pellet.
Whole gut transit assay
was performed as previously described
52
.
Beam traversal
was performed as previously described
53
with minor modifications. Animals
were trained over 2 days and 8 trials. Mice in the first 2 training trials were guided along the
beam by placing the home cage close to the position of the mouse and encouraged to
traverse by experimenter intervention. By the end of training, animals were adept beam
traversal and did not require additional training during time course experiments.
Pole descent
was performed by placing a 0.5 m long, 1 cm diameter ring stand vertically in
the animal’s home cage. The pole was covered in rubber shelf liner to facilitate grip. Mice
were trained over 6 trials. For the first 3 training trials, animals were placed on the pole,
head towards the cage at 10 cm, 20 cm, and 30 cm heights above the cage floor and allowed
to navigate back to the home cage independently. For the remaining 3 training trials, mice
were placed on the rod with heads directed toward the ceiling and allowed to turn and
descend. Mice were considered trained when, upon being placed on the pole, they
immediately turned and descended. If this criterion was not met, additional training trials
were performed with assistance from the experimenter to encourage descent. The next day,
animals performed 3 trials in a row with 1 min between trials.
Inverted wire hang
was performed as previously described
54
with minor modifications.
Animals were placed on a steel mesh screen and timed until they released their grip or
remained on the screen for 5 min. Mice performed 3 trials, allowing at least 15 min between
each trial.
Weightlifting paradigm
was performed as previously described
54
.
Adhesive removal
was performed as previously described
53
. Mice performed 2 trials with 5
min between each trial.
Hot plate test.
Evaluation of nociception was adapted from previous work
55
,
56
. Mice were habituated to the
unheated hot plate apparatus for 5 min. Hot plate was then heated to 55°C and verified by
probe thermometer. Mice were placed one at a time into the chamber, timed, and recorded
for playback. Each trial initiated when all paws touched the surface. Trial was terminated
Challis et al.
Page 10
Nat Neurosci
. Author manuscript; available in PMC 2020 August 17.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
and time recorded upon first occurrence of the following endpoint behaviors: hindpaw
licking, vocalizing, or jumping. A cutoff time of 30 sec was established as the endpoint in
the event the mouse did not exhibit these behaviors.
α
-Syn PFF and monomer preparation.
Purification of recombinant, untagged mouse
α
-Syn and PFF generation was performed as
previously described
20
.
α
-Syn monomers were subjected to a Pierce
Limulus
Amebocyte
Lysate (LAL) high capacity endotoxin removal resin. Levels of endotoxin were measured to
be 0.017 Units/μg of protein as determined using a Pierce LAL endotoxin quantitation kit.
Duodenal intramuscular injections.
Mice were anesthetized with 1–4% isoflurane and placed in a supine position on a self-
regulating heating pad. The hair over the abdomen was removed and a 2 cm incision was
made along the midline. The duodenal intestinal lining was directly injected at 2 sites, 1 cm
apart, with a 10 μl Hamilton syringe equipped with a 36 GA beveled needle (World
Precision Instruments, Sarasota, FL). Each site was injected with 3 μl of saline containing 1
μg/μl of protein (
α
-Syn PFF,
α
-Syn monomer, or BSA). Injection into the gastric lining was
confirmed with 0.2% Fast Green FCF (Sigma-Aldrich, F7252). Once injections were made,
the duodenum was carefully replaced and the skin was sutured. Mice were sacrificed at 7,
21, 60, and 120 days post inoculation.
Duodenal histology.
Prior to sacrifice, mice were injected intraperitoneally with 200 units of heparin in saline.
After 15 min, mice were anesthetized with pentobarbital and transcardially perfused with
cold phosphate buffered saline (PBS) followed by 4% paraformaldehyde in PBS. The
gastrointestinal tract was removed and post-fixed for 12–16 h in 4% PFA at 4°C, then
washed and stored in PBS with 0.05% sodium azide. A 2 cm segment of the duodenum was
excised just after the pylorus and cleared of connective tissue. Antigen retrieval was
performed by boiling the tissue in sodium citrate buffer (10 mM sodium citrate, 0.05%
Tween-20, pH 6.0) at 95–100°C for 20 min followed by a 10 min wash under running tap
water at room temperature. Two lengthwise cuts were made on each duodenum, one along
the mesentery and the other along the opposite side, and the two pieces were blocked using
the Mouse on Mouse kit (Vector Labs, Burlingame, CA) diluted in PBS with 0.1% Triton
X-100 (TX100). Whole duodenal sections were incubated in primary antibodies
(Supplementary Table 1) diluted in PBS with 3% normal donkey serum (NDS) and 0.1%
TX100 at room temperature overnight on an orbital shaker. Individual duodenal pieces were
washed in 4 × 10 ml PBS buffer exchanges across 6 h before incubation in secondary
antibodies diluted in PBS with 3% NDS at room temperature overnight on an orbital shaker.
Duodenal pieces were again washed in 4 × 10 ml PBS buffer exchanges for 6 h and then
incubated in a refractive index matching solution (RIMS) overnight at room temperature on
an orbital shaker. Tissue was mounted serosa facing up on a slide affixed with a 0.5 mm
spacer (SunJin Labs, Hsinchu City, Taiwan) and chamber filled with additional RIMS, and
coverslipped for confocal imaging (LSM 880, Zeiss, Oberkochen, Germany).
Challis et al.
Page 11
Nat Neurosci
. Author manuscript; available in PMC 2020 August 17.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
EdU staining.
Adult mice received a single intraperitoneal injection of 5-Ethynyl-2
-deoxyuridine (EdU,
Sigma Aldrich, 900584) at a concentration of 100 mg/kg, 24 hours after inoculation with
α
-
Syn PFF or monomer. At 7 dpi, mice were sacrificed and tissue was fixed as described
above. Incorporation of EdU was visualized using Click-iT EdU Cell Proliferation Kit for
Imaging (ThermoFisher, C10337).
Nodose ganglia tissue clearing and histology.
Nodose ganglia were removed from transcardially-perfused mice and podst-fixed for 12–16
h in 4% PFA at 4°C. The passive CLARITY technique (PACT) was performed on nodose
ganglia as previously described
37
. Overnight incubations in 3% NDS and 0.1% TX100 in
PBS at room temperature were performed for blocking and primary antibody
(Supplementary Table 1) steps. Ganglia were washed in 3 × 10 mL PBS over 2 h before
incubation in secondary in PBS with 3% NDS. Ganglia were again washed in 3 × 10 ml PBS
over 2 h before incubated in RIMS overnight at room temperature on an orbital shaker.
Ganglia were mounted on a slide affixed with a 0.5 mm spacer and chamber filled with
additional RIMS, and coverslipped for confocal imaging.
Brain section histology.
Brains were extracted from transcardially perfused mice and post-fixed in 4% PFA in PBS
for 16–20 h at 4°C. Serial sections 70–80 μm thick were cut in sets of 4 on a vibratome
(Leica Biosystems, VT1200, Wetzlar, Germany) in PBS. For each of the brainstem or
substantia nigra, three sections were mounted on Superfrost slides and allowed to dry for
antibody labeling. Staining was performed directly on the slides in a slide moisture chamber.
Slides were blocked with 3% NDS and 0.1% TX100 in PBS for 2 h at room temperature and
incubated in primary antibody (Supplementary Table 1) in the same buffer overnight at 4°C.
Slides were washed 3× with PBS and then incubated with secondary antibodies diluted in
PBS with 3% NDS. Slides were again washed 3× in PBS, mounted with Vectashield (Vector
Labs), and coverslipped for confocal imaging.
Stereological quantification.
The number of TH-positive cells in midbrain sections was determined manually from
maximum projection confocal images. Boundaries of the SNc were determined by TH signal
threshold detection using ImageJ and manually excluding non-SNc regions within the
section. Density was determined by quantifying number of cells per SNc volume (area
multiplied by the thickness of the slice).
Tissue lysates and protein quantification.
Tissues were placed in RIPA buffer (Millipore) diluted in PBS with a protease inhibitor
cocktail (Thermo Fisher, A32963) and homogenized in 2 ml Lysing Matrix D tubes (MP
Biomedicals, Santa Ana, CA) using a Mini-Beadbeater (Biospec, Bartlesville, OH).
Homogenates were centrifuged and the protein concentration of the supernatant was
determined with a BCA assay (Thermo Fisher). The cyotkine array (Abcam, ab133999) was
performed according to manufacturer’s instructions using 250μg of total protein from fresh
Challis et al.
Page 12
Nat Neurosci
. Author manuscript; available in PMC 2020 August 17.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
lysate. ELISA assays for interleukin-6 (IL-6) (Thermo Fisher) and dopamine (Eagle
Biosciences, Nashua, NH) were performed according to manufacturers’ instructions from
fresh lysate. For Western blots, 35 μg of protein from fresh or previously frozen lysate was
separated by 4–20% SDS-PAGE (Bio-Rad, Hercules, CA) and transferred to PVDF
membranes (Bio-Rad). For dot blot quantification of
α
-Syn filament, 1 μg of tissue
homogenate was spotted in 1 μL volume aliquots onto 0.45 μm nitrocellulose membrane.
Tissue from 1–2 WT mice was always included for intra-membrane comparisons. On orbital
shakers, membranes were blocked with 5% blocking solution (Bio-Rad) in Tris-buffered
saline with 0.1% Tween-20 (TBS-T) for 2 h at room temperature then incubated with
primary antibodies (Supplementary Table 1) in TBS-T at 4°C overnight. Membranes were
washed 3× in TBS-T and incubated with fluorescent or horseradish peroxidase (HRP)-
conjugated secondaries in TBS-T for 2 h at room temperature on an orbital shaker.
Membranes were again rinsed 3× in TBS-T. A Bio-Rad ChemiDoc MP was used to detect
fluorescence or HRP-conjugated (via Clarity Max chemiluminescent substrate [Bio-Rad])
secondaries. Densitometry analysis was performed in ImageJ or Image-Lab (Bio-Rad), as
described below.
RNA quantification.
Duodenal samples in RNAlater were homogenized in TRI Reagent (Zymo Research) and a
Tissue LyserII (Retsch) and RNA was isolated with DirectZol RNA extraction kit (Zymo
Research) via manufacturer’s instructions. cDNA was synthesized using a RevertAid first
strand cDNA synthesis kit (Thermo Fischer Scientific) according to manufacturer’s
instructions.
Real-time quantitative PCR was performed using SybrGreen qPCR master mix (Applied
Biosystems) on an ABI 7900 Prism qPCR instrument. cDNA was probed with the following
primers from Harvard PrimerBank:
GAPDH
– Forward 5
TGG CCT TCC GTG TTC CTA
C 3
Reverse 5
GAG TTG CTG TTG AAG TCG CA 3
;
GBA1
– Forward 5
GCC AGG
CTC ATC GGA TTC TTC 3
Reverse 5
CAC GGG GTC AAG AGA GTC AC 3
. Target
genes were normalized to
GAPDH
and fold change calculated by ΔΔCT relative to adult
control samples. Data is graphed as the average of technical triplicates.
Viral constructs.
The
pAAV-ihSyn-tTA
construct was developed in the Gradinaru lab
33
and is available on
Addgene (99120, Cambridge, MA). The
pAAV-TRE-GBA1-IRES-EGFP
,
pAAV-TRE-
jRGECO1a-NES
, and
pAAV-TRE-hChR2-EYFP
plasmids were cloned as described below.
The
pAAV-TRE-EGFP
was obtained from Addgene (89875). All viruses were packaged
with the AAV-PHP.S capsid (pUCmini-iCAP-PHP.S, Addgene, 103006) in-house, as
previously described
34
.
pAAV-TRE-GBA1-IRES-EGFP
was created by isolating the
GBA1
gene transcript variant 1
(NM_008094) from
a
plasmid obtained from Origene (MC200608, Rockville, MD) via
restriction digest with AsiSI and BsiWI. The
IRES-EGFP
site was PCR cloned to add 5’
BsiWI and 3’ EcoRI sites from a plasmid obtained from Addgene (20672). The TRE
promoter was PCR cloned from an Addgene plasmid (99113) to add a 5’ XbaI and 3’ AsiSi
Challis et al.
Page 13
Nat Neurosci
. Author manuscript; available in PMC 2020 August 17.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript