Gas Vesicle
−
Blood
Interactions
Enhance
Ultrasound
Imaging
Contrast
Bill Ling,
△
Jeong Hoon Ko,
△
Benjamin
Stordy, Yuwei Zhang, Tighe F. Didden, Dina Malounda,
Margaret
B. Swift, Warren C. W. Chan, and Mikhail G. Shapiro
*
Cite This:
Nano
Lett.
2023,
23,
10748−10757
Read
Online
ACCESS
Metrics
& More
Article
Recommendations
*
sı
Supporting
Information
ABSTRACT:
Gas
vesicles
(GVs)
are
genetically
encoded,
air-
filled
protein
nanostructures
of
broad
interest
for
biomedical
research
and
clinical
applications,
acting
as
imaging
and
therapeutic
agents
for
ultrasound,
magnetic
resonance,
and
optical
techniques.
However,
the
biomedical
applications
of
GVs
as
systemically
injectable
nanomaterials
have
been
hindered
by
a lack
of
understanding
of
GVs’
interactions
with
blood
components,
which
can
significantly
impact
in vivo
behavior.
Here,
we
investigate
the
dynamics
of
GVs
in
the
bloodstream
using
a
combination
of
ultrasound
and
optical
imaging,
surface
function-
alization,
flow
cytometry,
and
mass
spectrometry.
We
find
that
erythrocytes
and
serum
proteins
bind
to
GVs
and
shape
their
acoustic
response,
circulation
time,
and
immunogenicity.
We
show
that
by
modifying
the
GV
surface
we
can
alter
these
interactions
and
thereby
modify
GVs’
in vivo
performance.
These
results
provide
critical
insights
for
the
development
of
GVs
as
agents
for
nanomedicine.
KEYWORDS:
gas vesicles,
ultrasound
imaging,
blood,
protein
corona,
surface
engineering
N
anomaterials
are
becoming
increasingly
important
for
biomedical
applications
such
as
drug
delivery,
medical
imaging,
and
diagnostics.
1
In
these
contexts,
nanoparticle
behavior
is
significantly
impacted
by
the
cells
and
proteins
encountered
within
the
bloodstream.
Serum
proteins
rapidly
adsorb
to
nanoparticle
surfaces,
forming
a protein
corona
that
alters
their
physicochemical
properties
and
recognition
by
the
body.
2
−
5
The
corona’s
composition
can
predict
factors
such
as
pharmacokinetics,
biodistribution,
toxicity,
and
cellular
up-
take.
6
−
8
Modification
strategies
often
involve
covering
the
particle
surface
with
hydrophilic
polymers
such
as
poly-
(ethylene
glycol)
(PEG)
and
other
ligands.
9
Additionally,
some
nanomaterials
bind
to
erythrocytes
(RBCs),
affecting
imaging
contrast,
10
biodistribution,
11
and
circulation
time.
12
Gas
vesicles
(GVs)
are
an
emerging
nanomaterial
with
great
potential
as
agents
for
imaging
and
therapy.
13
These
air-filled
protein
nanostructures
are
naturally
produced
by
certain
aquatic
microbes
for
buoyancy
regulation.
14
GVs
comprise
a
2 nm
thick
protein
shell
that
excludes
liquid
water
but
permits
the
dynamic
exchange
of
gas,
forming
a thermodynamically
stable
pocket
of
air
with
nanoscale
dimensions.
14
Acoustic
waves
are
strongly
scattered
at
this
air
−
water
interface,
enabling
GVs
to
produce
robust
ultrasound
contrast
when
injected
into
the
body
15
−
17
or
expressed
in
engineered
cells.
18,19
Furthermore,
they
are
resilient
to
repeated
insonation,
15
easily
tailored
to
target
molecular
markers
20
−
22
or
respond
to
biological
functions,
23,24
and
have
growing
applications
in
therapeutic
ultrasound,
25,26
optical
imaging,
27,28
and
magnetic
resonance
imaging.
29,30
To
effectively
incorpo-
rate
these
capabilities
into
an
injectable
agent,
a
deeper
understanding
of
in vivo
GV
behavior
is needed.
In
this
study,
we
investigate
GV
interactions
with
RBCs
and
serum
proteins,
develop
surface
functionalization
techniques
to
modulate
these
interactions,
and
evaluate
the
downstream
effects
on
the
acoustic
response,
circulation
time,
and
immunogenicity.
We
characterize
GVs’
protein
corona
and
identify
molecular
pathways
governing
their
in vivo
fate.
This
analysis
offers
valuable
insights
for
the
ongoing
development
and
optimization
of
injectable
nanoparticles
and
GV-based
agents.
We
began
by
studying
the
behavior
of
GVs
after
intravenous
(IV)
administration.
We
visualized
circulating
GVs
with
ultrafast
power
Doppler
ultrasound
imaging,
leveraging
their
ability
to
enhance
blood
flow
contrast.
16
Targeting
a single
coronal
plane
in
the
mouse
brain,
we
acquired
images
at
a
Received:
July
25,
2023
Revised:
November
10,
2023
Accepted:
November
13,
2023
Published:
November
20,
2023
Letter
pubs.acs.org/NanoLett
©
2023
The
Authors.
Published
by
American
Chemical
Society
10748
https://doi.org/10.1021/acs.nanolett.3c02780
Nano
Lett.
2023,
23,
10748
−
10757
This article is licensed under CC-BY 4.0
15.625
MHz
center
frequency
and
a
0.25
Hz
frame
rate
(Figure
1A).
After
a 5 min
baseline,
we
injected
100
μ
L
of
5.7
nM
GVs
purified
from
Anabaena
flos-aquae
31
(Ana)
and
monitored
the
ensuing
changes
in
hemodynamic
signal.
In
healthy
BALB/c
mice,
contrast
reached
an
initial
peak
within
1
min,
followed
by
a larger
peak
3.5
min
later,
and
then
returned
to
baseline
over
approximately
30
min
as
GVs
were
cleared
by
the
liver
24
(Figure
1B).
Intensities
at
the
first
peak
were
consistent
across
trials
but
varied
significantly
at
the
second
peak
(Figure
1C).
We
next
investigated
the
cause
of
this
dual-peak
phenomenon.
We
hypothesized
that
the
first
peak
was
due
to
dispersion
of
free-floating
GVs
throughout
the
bloodstream,
as
its
onset
time
matched
the
vascular
distribution
kinetics
of
other
nanoparticle
and
small
molecule
contrast
agents.
32,33
Furthermore,
the
intensity
of
the
first
peak
correlated
linearly
with
injected
dose,
suggesting
that
it is
directly
governed
by
GV
concentrations
in
the
blood
(Figure
S1).
In
comparison,
the
intensity
of
the
second
peak
appeared
to
plateau
at
higher
doses,
suggesting
a binding
interaction.
We
hypothesized
that
the
second
peak
arose
from
an
increase
in
acoustic
backscatter
due
to
GV
clustering.
15
We
observed
similar
contrast
enhancement
dynamics
in
immuno-
competent
BALB/c
(Figure
1B)
and
immunocompromised
NSG
mice
(Figure
S2),
which
lack
both
B cells
and
T cells,
and
therefore
suspected
an
antibody-independent
mechanism
such
Figure
1.
GV
adsorption
to
RBCs
contributes
to
a second
peak
of
the
hemodynamic
ultrasound
contrast.
(A)
Diagram
of
the
in vivo
imaging
setup.
Intravascular
dynamics
of
IV
injected
GVs
were
visualized
by
transcranial
ultrafast
power
Doppler
imaging
of
the
brain.
(B)
Time
courses
of
Doppler
signal
enhancement
in
immunocompetent
BALB/c
mice
following
injection
of
100
μ
L
of
5.7
nM
GVs.
N
= 5.
Dashed
gray
line,
time
of
injection
(300
s);
dashed
blue
line,
peak
1 (350
s);
dashed
red
line,
peak
2 (480
s).
(C)
Signal
enhancement
at
the
indicated
peaks
in
time
courses
from
panel
B.
Points
from
the
same
trial
are
connected.
N
= 5.
Paired
t
test,
(
**
p
< 0.01).
(D)
Diagram
of
the
RBC
binding
assay.
Ultrasound
imaging:
RBCs
were
incubated
with
GVs
modified
to
produce
nonlinear
signal,
washed
by
centrifugation,
and
loaded
into
an
agarose
phantom
for
nonlinear
AM
imaging.
Flow
cytometry:
RBCs
were
incubated
with
fluorescently
labeled
GVs,
washed
by
centrifugation,
stained
with
anti-TER-
119,
and
analyzed
by
flow
cytometry.
(E)
Acoustic
detection
of
adsorbed
GVs.
Left:
representative
ultrasound
images
of
RBCs
mixed
with
GVs.
AM
signal
is overlaid
on
a B-mode
image
to
show
sample
outlines.
Scale
bars:
1 mm.
Right:
mean
AM
signal
intensity
within
each
well.
N
= 18
−
60.
Error
bars:
±
SEM.
Welch’s
t
test
(
****
p
< 0.0001).
(F)
Flow
cytometric
detection
of
GVs
adsorbed
to
RBCs.
Left:
representative
scatter
plots
of
washed
RBCs,
gated
for
single
cells.
The
gating
strategy
is shown
in
Figure
S5.
Right:
mean
fluorescence
of
TER-119+
cell
population.
RBC,
N
=
11;
RBC+Ana,
N
= 18.
Error
bars:
±
SEM.
Welch’s
t
test
(
****
p
< 0.0001).
Nano
Letters
pubs.acs.org/NanoLett
Letter
https://doi.org/10.1021/acs.nanolett.3c02780
Nano
Lett.
2023,
23,
10748
−
10757
10749
Figure
2.
Surface
passivation
reduces
RBC
binding
and
extends
the
circulation
time.
(A)
Reaction
scheme
for
GV
functionalization.
Alkynes
were
conjugated
to
lysines
on
the
GV
surface,
and
polymers
were
attached
through
a CuAAC
reaction.
(B)
DLS
measurements
of
the
hydrodynamic
diameter.
N
= 8
−
12.
Error
bars:
±
SEM.
Welch’s
t
test
(
*
p
< 0.05;
****
p
< 0.0001;
n.s,
p
≥
0.05).
(C)
Zeta-potential
measurements
of
engineered
GVs.
N
= 5
−
11.
Error
bars:
±
SEM.
Welch’s
t
test
(
*
p
< 0.05;
****
p
< 0.0001).
(D)
Normalized
optical
density
at
600
nm
as
a function
of
hydrostatic
pressure.
N
= 4. Thick
lines:
mean;
shaded
areas:
±
SEM
(E)
Left:
representative
B-mode
images
of
Ana
and
Ana-PEG
embedded
in
an
agarose
phantom.
Right:
mean
B-mode
signal
intensities
were
within
each
well.
N
= 48.
Error
bars:
±
SEM.
Welch’s
t
test
(n.s.,
p
≥
0.05).
(F)
Flow
cytometric
detection
of
fluorescently
labeled
Ana-PEG
adsorbed
to
RBCs.
Left:
representative
dot
plot
of
washed
RBCs,
gated
for
single
cells.
RBCs
are
stained
with
anti-TER-119.
Right:
mean
fluorescence
of
TER-119+
cell
population.
N
= 18.
Ana
and
RBC-only
controls
from
Figure
1F
are
shown
as
a reference.
Error
bars:
±
SEM.
Welch’s
t
test
(
*
p
< 0.05;
****
p
< 0.0001).
(G)
Acoustic
detection
of
Ana-PEG
modified
to
produce
nonlinear
contrast.
Left:
representative
ultrasound
images
of
RBCs
were
mixed
with
Ana-PEG.
The
AM
signal
is overlaid
on
a B-mode
image.
Right:
AM
signal
intensities
were
normalized
to
their
respective
GV-only
samples.
N
=
6.
Normalized
data
from
Figure
1E
are
shown
for
comparison.
Error
bars:
±
SEM.
Welch’s
t
test
(
**
p
< 0.01;
****
p
< 0.0001).
(H)
Time
courses
of
ultrafast
power
Doppler
signal
enhancement
following
injection
of
Ana-PEG
into
BALB/c
mice.
N
=
5.
Dashed
line
is
the
time
of
injection
(300
s).
(I)
Half-life
of
GV-induced
signal
enhancement
calculated
by
fitting
time
courses
in
Figure
1B
(Ana)
and
Figure
2H
(Ana-PEG)
to
an
exponential
decay
function.
Error
bars:
±
SEM.
Welch’s
t
test
(
*
p
< 0.05).
Nano
Letters
pubs.acs.org/NanoLett
Letter
https://doi.org/10.1021/acs.nanolett.3c02780
Nano
Lett.
2023,
23,
10748
−
10757
10750