Biomolecular Ultrasound Imaging of Phagolysosomal Function
Bill Ling
1
,
Justin Lee
2
,
David Maresca
1
,
Audrey Lee-Gosselin
1
,
Dina Malounda
1
,
Margaret
B. Swift
1
,
Mikhail G. Shapiro
1,*
1.
Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena,
California, 91125, United States
2.
Division of Biology and Bioengineering, California Institute of Technology, Pasadena, California,
91125, United States
Abstract
Phagocytic clearance and lysosomal processing of pathogens and debris are essential functions of
the innate immune system. However, the assessment of these functions
in vivo
is challenging
because most nanoscale contrast agents compatible with non-invasive imaging techniques are
made from non-biodegradable synthetic materials that do not undergo regular lysosomal
degradation. To overcome this challenge, we describe the use of an all-protein contrast agent to
directly visualize and quantify phagocytic and lysosomal activities
in vivo
by ultrasound imaging.
This contrast agent is based on gas vesicles (GVs), a class of air-filled protein nanostructures
naturally expressed by buoyant microbes. Using a combination of ultrasound imaging,
pharmacology, immunohistology and live-cell optical microscopy, we show that after intravenous
injection, GVs are cleared from circulation by liver-resident macrophages. Once internalized, the
GVs undergo lysosomal degradation, resulting in the elimination of their ultrasound contrast. By
non-invasively monitoring the temporal dynamics of GV-generated ultrasound signal in circulation
and in the liver and fitting them with a pharmacokinetic model, we can quantify the rates of
phagocytosis and lysosomal degradation in living animals. We demonstrate the utility of this
method by showing how these rates are perturbed in two models of liver dysfunction: phagocyte
deficiency and non-alcoholic fatty liver disease. The combination of proteolytically-degradable
nanoscale contrast agents and quantitative ultrasound imaging thus enables non-invasive functional
imaging of cellular degradative processes.
Graphical Abstract
*
Corresponding author. mikhail@caltech.edu.
Author Contributions
B.L. and M.G.S. conceived the research. B.L. conducted the
in vivo
imaging experiments with assistance from D. Maresca and A.L.-
G. B.L. and J.L. conducted
in vitro
macrophage experiments. B.L. established and validated the pharmacology and disease models
with assistance from M.B.S. D. Malounda prepared gas vesicles for experiments. B.L. and M.G.S. wrote the paper with input from all
other authors. M.G.S. supervised the research.
Competing interests
The authors declare no competing financial interests.
Supporting Information
This material is available free of charge
via
the internet at
http://pubs.acs.org
.
HHS Public Access
Author manuscript
ACS Nano
. Author manuscript; available in PMC 2021 March 22.
Published in final edited form as:
ACS Nano
. 2020 September 22; 14(9): 12210–12221. doi:10.1021/acsnano.0c05912.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
Keywords
ultrasound; contrast agents; phagocytosis; lysosomes; liver disease; reticuloendothelial system
The reticuloendothelial system (RES), also known as the mononuclear phagocyte system, is
a network of phagocytic immune cells that is essential for organismal development and
homeostasis; malfunctions in this system may lead to increased susceptibility to infections
1
,
2
and are associated with the pathogenesis of a variety of conditions, including
neurodegeneration,
3
,
4
chronic liver disease
5
and many others.
6
Cells of the RES, such as
monocytes, macrophages and dendritic cells, continuously sample their surroundings,
mediating the recognition and clearance of abnormal and senescent cells, debris and foreign
particulates.
7
,
8
Additionally, they interface with the adaptive immune system by presenting
lysosomally-processed antigens to lymphocytes and secreting cytokines to stimulate the
proper inflammatory response.
8
–
10
Phagocytosis and lysosomal degradation are thus vital
processes of RES-mediated immunoregulation.
Non-invasive functional imaging of phagocytosis and lysosomal activities will enable early
detection and monitoring of non-alcoholic fatty liver disease (NAFLD) and other conditions
resulting from RES dysfunction. NAFLD currently affects over 25% of the global
population and its progression is associated with chronic hepatic inflammation.
11
Due to the
large patient population and broad range of outcomes which include hepatitis, cirrhosis,
fibrosis and hepatocellular carcinoma, rapid and non-invasive diagnostic methods are needed
to stratify patients into defined risk groups.
5
,
11
Ultrasound is well suited for this task due to
its wide availability, portability, low operational costs and high tissue penetrance.
12
Based on
in vitro
observations that pro-inflammatory macrophages suppress phagocytosis
13
and
lysosomal degradation,
14
,
15
one would expect livers in patients with NAFLD to exhibit
reduced accumulation and extended persistence of intravenously-administered nanoscale
contrast agents. Indeed, clinical studies have confirmed the former.
16
,
17
However, the latter
cannot be evaluated with currently available technologies because agents compatible with
non-invasive imaging modalities are typically made from synthetic materials which do not
undergo regular lysosomal degradation.
7
,
8
Here, we describe the use of an all-protein nanoscale contrast agent to visualize and quantify
both phagocytic clearance and lysosomal degradation
in vivo
using ultrasound imaging. This
contrast agent is based on gas vesicles (GVs), a class of air-filled protein nanostructures
natively formed inside certain photosynthetic microorganisms as a means to regulate
buoyancy.
18
GVs comprise a rigid, 2 nm-thick protein shell allowing the free exchange of
gas but preventing the internal condensation of liquid water, thereby forming a
thermodynamically stable capsule of air with a hydrodynamic diameter of approximately
Ling et al.
Page 2
ACS Nano
. Author manuscript; available in PMC 2021 March 22.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
250 nm.
19
They are easily isolated from cultures of their native cyanobacterial hosts
20
and
can be expressed heterologously in bacteria
21
,
22
and mammalian cells.
23
Because sound
waves are strongly reflected by air-water interfaces, GVs have been developed as contrast
agents for ultrasound imaging.
19
,
24
–
27
Due to their innate stability, GVs are able to
withstand repeated insonation without loss of contrast.
19
However, when the GV shell is
compromised by mechanical or chemical disruption, the gaseous contents it encloses rapidly
and irreversibly dissolve into the surrounding media, leading to the elimination of ultrasound
contrast.
19
,
21
,
27
Based on their nanoscale dimensions and all-protein composition, which distinguishes them
from other classes of ultrasound contrast agents,
28
–
31
we hypothesized that we could use
GVs as a contrast agent to non-invasively visualize the phagocytic and lysosomal functions
of hepatic macrophages
in vivo
. Previous studies have shown that intravenously injected
GVs are rapidly taken up by the liver.
32
,
33
If this uptake is mediated by macrophages and the
internalized GVs undergo lysosomal proteolysis, this would manifest in the initial transfer of
ultrasound contrast from the bloodstream to the liver, followed by its elimination with
kinetics representative of natural RES clearance and degradation. Measurement of these
processes would thus provide a quantitative picture of the complete phagocytic and
lysosomal degradation pathways. This rate-based approach would improve upon previous
Kupffer cell imaging techniques
16
,
34
–
36
which are limited to the assessment of phagocytosis.
In this study, we test this hypothesis by visualizing the temporal dynamics of GV ultrasound
contrast in the blood and liver, establishing the cellular and molecular pathways mediating
GV uptake and degradation, and developing a pharmacokinetic model to parametrize RES
activity from hemodynamic and liver ultrasound signals. Finally, we demonstrate the
diagnostic utility of functional imaging of macrophage phagolysosomal activity in two
models of liver disease: clodronate-mediated macrophage deficiency and diet-induced
NAFLD.
Results and Discussion
Gas vesicle blood clearance, liver uptake and degradation can be monitored by
ultrasound.
We started by quantifying the kinetics of GV uptake and degradation in healthy C57BL/6
mice (Fig. 1a). We first visualized intravascular GVs with ultrafast power Doppler imaging,
leveraging the ability of intravenously (IV) injected GVs to enhance blood flow contrast.
25
We chose the brain as our target organ due to its practical advantages in mouse experiments:
hemodynamic signals can be conveniently measured through intact skin and skull
25
,
37
and
head-fixation reduces motion artifacts. We acquired images of a single coronal plane at a
center frequency of 15 MHz and frame rate of 0.25 Hz (Fig. 1b). Following a 300-s baseline,
we IV injected 100 μL of purified GVs isolated from
Anabaena flos-aquae
(OD
500
30,
corresponding to 2.1 × 10
11
particles
20
) and tracked the ensuing distribution and clearance
(Fig. 1c). As expected, the introduction of GVs caused a marked increase in hemodynamic
signal, peaking at approximately 100 s after injection, and returning to baseline with an
apparent circulation half-life of 232 s (Fig. 1c, Fig. S1).
Ling et al.
Page 3
ACS Nano
. Author manuscript; available in PMC 2021 March 22.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
Next, we quantified GV uptake and degradation in the liver by imaging this organ during
and after IV injection (Fig. 1a). To maximize GV specificity, we injected GVs modified to
produce non-linear ultrasound contrast
27
and imaged with a non-linear amplitude
modulation (AM) pulse sequence
24
(Fig. 1, d–e). Following injection of 100 μL GVs at
OD
500
30, we observed the accumulation of non-linear contrast in the liver—reaching a
maximum after approximately 10 min—followed by a gradual loss of signal until only 10%
remained at the end of one hour (Fig. 1d). Notably, the maximum occurs just as contrast in
the blood returns to baseline (Fig. S2). The apparent half-life of GVs in the liver—20 min—
is substantially longer than their circulation time, and on a timescale consistent with
lysosomal processing.
38
–
40
To independently confirm liver uptake, we acquired fluorescence
images of mouse organs excised 1h after IV injection of GVs labeled with a far-red
fluorescent dye (Fig. 1f). In line with previous investigations of GV biodistribution,
32
,
33
the
liver was the dominant organ for GV uptake, emitting 81.4% of collected photons. The lungs
(7.8%) and spleen (5.5%) had minor roles in GV clearance, while the heart and kidneys had
no discernible role.
GVs are primarily cleared by liver macrophages.
To identify the cells involved in GV clearance, we performed immunofluorescence imaging
of liver sections obtained from mice perfused 1h after IV injection of fluorescently-labeled
GVs (Fig. 2a). Based on the apparent active degradation of GVs, as suggested by the gradual
decline of liver ultrasound contrast, we hypothesized that GVs would be taken up by Kupffer
cells—resident macrophages lining the hepatic sinusoids which are implicated in the
clearance of many nanoparticles.
7
We tested this hypothesis by defining antibody-stained
F4/80
+
Kupffer cell regions through image segmentation by Ilastik
41
and quantifying the
localization of GVs with respect to these borders (Fig. S3). On average, 60% of GV-
containing pixels resided within Kupffer cells (Fig. 2b).
To confirm the role of Kupffer cells in GV clearance, we ablated phagocytic cells by IV
administration of 30 mg/kg liposome-encapsulated clodronate
42
(Fig. 2c). 48h later, we
measured GV circulation times with hemodynamic ultrasound (Fig. 2d). Compared to mice
treated with saline-filled control liposomes, clodronate-treated mice had a nearly 7-fold
enhancement in GV circulation time, with half-life increasing from 274 s to 1670 s (Fig.
S2). Our results are in line with previous observations that treatment with 50 mg/kg
clodronate increased the circulation half-life of 100 nm gold nanoparticles 13-fold.
1
Taken
together, our data shows that GVs are mainly filtered from the blood by Kupffer cells.
GVs are degraded in the lysosome following phagocytosis.
Having established their uptake by liver macrophages, we next studied what happens to GVs
following phagocytosis. Macrophages typically internalize nanoparticles into membrane-
bound organelles—phagosomes—that are then trafficked along the phagolysosomal
pathway. During this maturation process, the phagosomes acquire v-ATPase proton pumps
to acidify their contents prior to fusion with the lysosome;
43
this low pH environment is
required for lysosomal enzyme activity. To visualize the movement of GVs along this
pathway
in vitro
, we incubated murine macrophages (RAW264.7) with a dilute suspension
of GVs dually-labeled with Alexa Fluor (AF647) and pHrodo Red—a pH-sensitive dye that
Ling et al.
Page 4
ACS Nano
. Author manuscript; available in PMC 2021 March 22.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
fluoresces weakly at pH 7 and brightly at pH 3—and imaged them with live-cell confocal
optical microscopy (Fig. 3a). Focusing on individual cells, we initially observed punctate
spots of AF647 signal, likely corresponding to GVs concentrated within phagosomes, which
matured over the next several minutes to produce strong pHrodo signal, indicating
acidification of their environment (Fig. 3b). Zooming out to observe population-level
dynamics revealed that the proportion of GVs in acidified compartments, as parametrized by
the ratio of pHrodo to AF647 signal, grew continuously during a 1-hour incubation (Fig. 3,
c–d). This rise was abolished when v-ATPase was inhibited by pretreatment with 100 nM
bafilomycin A1 (BafA1),
44
thereby confirming that GVs undergo phagolysosomal
processing in macrophages.
Lysosomal proteolysis is expected to break down the GV shell, resulting in GV collapse, gas
dissolution and the disappearance of ultrasound contrast. To confirm this effect
in vitro
, we
exposed RAW264.7 cells to GVs for 30 min. At predetermined time intervals, we detached
the cells from their solid substrate and loaded them into an agarose phantom for imaging
with a non-linear cross-propagating amplitude modulation pulse sequence (xAM)
26
(Fig.
3e). In control cells pretreated with 0.01% v/v dimethyl sulfoxide (DMSO), the signal
declined with a half-life of approximately 3 h (Fig. 3, f–g). Conversely, in cells pretreated
with BafA1 to block the activity of the pH-dependent lysosomal enzymes, we observed
signal that persisted for at least 5 h without decay. These results confirm that GVs are
digested within macrophage lysosomes in a process that can be monitored with non-linear
ultrasound imaging. The reason that this process happens somewhat more slowly
in vitro
compared to the liver may be the accelerated rate of phagosome maturation in primary
macrophages.
45
GV pharmacokinetics can be used to monitor disease progression.
The results presented thus far confirm that upon IV injection, GVs are filtered from the
blood by liver macrophages and subsequently catabolized in the lysosome (Fig. 4a). This
process can be described with a two-compartment pharmacokinetic model comprising the
blood and liver, whose rate constants parametrize the concurrent processes of phagocytosis
and lysosomal degradation (Fig. 4b), with contrast enhancement linearly proportional to
intact GV concentration in each compartment (Fig. S4). By fitting this model to the
dynamics of GV ultrasound contrast in the vasculature and liver
in vivo
, we can thus non-
invasively quantify macrophage phagolysosomal function (Fig. 4c, input data shown in Fig.
S5). The assumption that ultrasound signal time courses are representative of true
pharmacokinetics is based on two key observations: GVs are stable under our imaging
parameters, so changes in signal are due to active biological processes; and GVs are
primarily taken up by liver macrophages, with increases in liver AM contrast matched by
decreases in brain Doppler contrast. For simplicity, we further assume each process to be
first-order and neglect the initial distribution dynamics during GV infusion by considering
timepoints occurring after the peak in Doppler signal. Using this approach, we calculated
rates of 0.167 min
−1
and 0.041 min
−1
for uptake and degradation, respectively, in healthy
mice (Fig. 4d, Table S1).
Ling et al.
Page 5
ACS Nano
. Author manuscript; available in PMC 2021 March 22.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
Having established a method to quantify liver macrophage function, we next evaluated its
ability to detect pathological disruption of the RES. First, we administered two doses of
liposomal clodronate—0.40 mg/kg and 30 mg/kg—to partially or fully deplete Kupffer cells
in the liver. Histological evaluation confirmed that 31% of the Kupffer cell population
remained at the lower dose, decreasing to 16% at the higher dose (Fig. 5a; Fig. S6).
Interestingly,
ex vivo
imaging of organ fluorescence revealed that most GVs are still cleared
by the liver (Fig. 5b). However, closer inspection of liver sections with immunofluorescence
showed GVs tending to localize to the sinusoidal margins, suggesting uptake by liver
sinusoidal endothelial cells (LSECs) (Fig. S7). This is consistent with a recent study
showing that LSECs upregulate phagocytic activity upon depletion of nearby Kupffer cells.
7
Based on these results, we expected that GVs would circulate longer in the blood in
clodronate-treated animals due to diminished phagocytic potential, and that their residence
time in the liver would increase due to less efficient lysosomal degradation by non-
macrophage cells. Indeed, fitting our model to the normalized hemodynamic Doppler (Fig.
5c) and liver AM (Fig. 5d) signal time courses yielded uptake and degradation rates
substantially lower than those of healthy mice (Fig. 5e). Specifically, phagocytosis rates
were reduced by 66% and 82% at the low and high doses of clodronate, while proteolysis
rates were reduced by 27% and 57%, respectively. Notably, phagocytosis rates were
proportional to the macrophage population.
For our second model of RES dysfunction, we imaged mice with NAFLD. This disease is
characterized by liver infiltration of pro-inflammatory M1-polarized macrophages
5
,
46
which
have lower phagocytic
13
,
17
,
34
,
47
and lysosomal activities
15
than the normally anti-
inflammatory Kupffer cells.
48
We induced NAFLD by feeding mice with a methionine- and
choline-deficient (MCD) diet
5
,
49
and performed ultrasound imaging after 4 weeks of this
treatment (Fig. 5f). Histological evaluation confirmed the appearance of widespread
steatosis, a hallmark of NAFLD (Fig. 5f). In line with our hypothesis, diseased mice had
significantly suppressed phagocytic and lysosomal functions: uptake rate was reduced by
35% while degradation rate was reduced by 58% (Fig. 5, g–i). We verified that these
differences are not due to saturation of the smaller livers of MCD mice
50
by GVs (Fig. S8).
When we simulated therapeutic intervention by reverting to a control diet for 3 additional
weeks, the steatosis subsided (Fig. 5f) and phagolysosomal activity returned to its original
level (Fig. 5, g–i). Compared to age-matched litter-mate controls, these “recovered” mice
showed a slight decrease in degradation rate but no discrepancies in uptake rate (Fig. S9,
Table S1). Taken together, our results demonstrate the capability of GV-enhanced ultrasound
to non-invasively visualize macrophage malfunction as a biomarker of disease.
Conclusions
GVs are advantageously positioned to image
in vivo
phagolysosomal function due to their
inherent stability at ambient conditions, susceptibility to natural proteolytic degradation and
dependence on shell integrity for ultrasound contrast. When combined with a simple
pharmacokinetic model, GV imaging makes it possible to parametrize macrophage activity
in terms of phagocytosis and lysosomal degradation rates, clearly delineating healthy and
disease states, as demonstrated in two models of RES deficiency.
Ling et al.
Page 6
ACS Nano
. Author manuscript; available in PMC 2021 March 22.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
The diagnostic power of macrophage functional imaging arises from the dependence of
phagolysosomal kinetics on cellular phenotype which, in turn, reflects the local tissue and
inflammatory microenvironment. Moving forward, this capability could be refined by
application of GVs engineered to display surface ligands,
27
as phenotype-specific responses
to certain particle-bound domains may augment differences in degradative behavior.
51
Methods to alter GV biodistribution would enable targeting and functional assessment of
macrophages in tissues other than the liver. Additionally, the ability to genetically express
GVs
23
could enable study of intracellular proteolytic processes, such as autophagy and
proteasomal degradation.
To maximize the translational utility of this technology, three aspects could be improved.
First, imaging parameters should be optimized for clinical use. In this study, we separately
acquired ultrafast Doppler and non-linear AM images to maximize signal specificity.
However, simultaneous multiplexed imaging of blood and liver signals would greatly
streamline diagnostic use. This could be accomplished by intercalating amplitude
modulation images with Doppler images of the liver, enabling GV quantitation in both
compartments with a single, stationary transducer. Second, while GV administration at doses
similar to those used in our experiments does not result in acute, adverse health effects in
mice,
19
clinical translation would require formal studies of dose-limiting and long-term
toxicity. In addition, to support long-term monitoring of individual subjects, it would be
useful to better understand the immunogenicity of GVs and the impact of repeated
injections, as the development of antibodies may skew clearance kinetics.
52
Finally, in some
applications it may be useful to image GVs with other imaging modalities, such as magnetic
resonance imaging
28
,
29
and optical coherence tomography;
53
adaptation of phagolysosomal
imaging to these modalities would facilitate applications where the efficacy of ultrasound
may be limited.
In summary, the combination of nanoscale, lysosomally-degradable contrast agents and
quantitative ultrasound imaging enables non-invasive assessment of macrophage function as
a disease-relevant biomarker. This technology will broaden the diagnostic capabilities of
biomolecular ultrasound and motivate further methods for non-invasive characterization of
cellular function.
Methods
GV preparation and quantification
Native gas vesicles (GVs) were isolated from
Anabaena flos-aquae
as previously described.
20
Concentrations were measured by optical density (OD) at 500 nm using a
spectrophotometer (NanoDrop ND-1000, Thermo Scientific). Stripped GVs were prepared
by treatment of native GVs with 6M urea solution followed by two rounds of centrifugally-
assisted flotation and removal of the subnatant.
20
Fluorescently-labeled gas vesicles were
prepared by mixing GVs at OD 10 in 1x phosphate-buffered saline (PBS) with 6 μM Alexa
Fluor 647 NHS Ester (Invitrogen, prepared as 10 mM solution in dimethyl sulfoxide).
Dually-labeled GVs were prepared by mixing GVs at OD10 with 6 μM pHrodo Red
succinimidyl ester (Invitrogen, prepared as 10 mM solution in dimethyl sulfoxide) and 18
μM Alexa Fluor 647 NHS Ester (Invitrogen, prepared as 10 mM solution in dimethyl
Ling et al.
Page 7
ACS Nano
. Author manuscript; available in PMC 2021 March 22.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
sulfoxide). After rotating in the dark at 25°C for 1 h, the reactions were quenched with Tris-
HCl. Prior to use, all GVs were buffer exchanged into 1x PBS by two rounds of overnight
dialysis through a regenerated cellulose membrane (12–14 kD MWCO, Repligen).
Cell culture
RAW264.7 (TIB-71) and HEK293T (CRL-3216) cells were ordered from the American
Type Culture Collection (ATCC). Cells were cultured on tissue culture treated 10-cm dishes
in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% Fetal Bovine
Serum and 1% penicillin/streptomycin.
Lentiviral transduction.—
Plasmid constructs were designed with SnapGene (GSL
Biotech) and assembled with Gibson Assembly reagents from New England Biolabs.
Briefly, mWasabi
54
was inserted downstream of a 20-AA palmitoylation tag from GAP43
and expressed under the EF-1
α
promoter (gift from Dan I. Piraner
55
). This plasmid was then
transfected along with third-generation lentiviral vector and helper plasmids (kind gifts from
the laboratory of David Baltimore) into HEK293T cells using polyethyleneimine (PEI).
Following a 12 h incubation, PEI-containing media was replaced with fresh media
supplemented with 10 mM sodium butyrate (Sigma Aldrich). Viral particles were
concentrated 48 h later
via
ultracentrifugation. RAW264.7 cells were transduced by
spinfection. Briefly, concentrated virus was added to non-tissue culture treated 24-well
plates coated with RetroNectin (Takara Bio). Following centrifugation (2,000xg, 2h), 4e5
RAW264.7 cells in 1 mL media were added to each well. The plates were spun again at
900xg for 50 min before transferring to the incubator. The brightest 10% of cells were
selected with a BD FACSAria III (BD Biosciences) at the City of Hope Analytical
Cytometry Core Facility.
Preparation of fibronectin-treated cover slips.—
Ethanol sterilized square (22 mm ×
22 mm) #1.5H glass cover slips (Thorlabs) were individually placed into the wells of a 6-
well plate and immersed in 2 mL PBS containing 10 μg fibronectin from bovine plasma
(Sigma Aldrich) for 2h at room temperature. The fibronectin solution was then aspirated and
the plates stored at 4°C until use. Sterile glass-bottom 35mm dishes (MatTek) were similarly
coated with 2.5 μg fibronectin in 500 μL PBS.
Animal preparation and disease models
All
in vivo
experiments were performed on male C57BL/6J mice (The Jackson Laboratory)
under protocols approved by the Institutional Animal Care and Use Committee at the
California Institute of Technology.
Macrophage depletion.—
Liposome-encapsulated clodronate (Clodrosome, Encapsula
NanoSciences) was administered through the lateral tail vein 48 h prior to imaging. Mice
receiving a dose of 30 mg/kg were injected with undiluted liposome suspension, while mice
receiving the lower dose of 0.40 mg/kg were injected with liposomes diluted 1:100 with
sterile saline. Control mice were injected with the equivalent volume of undiluted PBS
liposomes (Encapsome, Encapsula NanoSciences).
Ling et al.
Page 8
ACS Nano
. Author manuscript; available in PMC 2021 March 22.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
Diet-induced nonalcoholic fatty liver disease.—
8-week old mice were free fed with
either a methionine and choline deficient diet (5ADJ, TestDiet) or control diet (5CC7,
TestDiet) for up to 4 weeks. Afterwards, all mice were fed the control diet for an additional
3 weeks. Because this dietary protocol often results in dramatic weight loss, the mice were
monitored weekly for signs of adverse health. GV pharmacokinetics were measured at 2
weeks, 4 weeks and at the conclusion of the study. Immediately after ultrasound imaging,
the mice were fixed
via
sequential transcardial perfusion of PBS and 10% neutral buffered
formalin (Sigma Aldrich), and the livers were removed for histological assessment by the
UCLA Translational Pathology Core Laboratory. Briefly, 4-μm sections were cut from
paraffin-embedded organs, stained with hematoxylin & eosin, and imaged at 20x with a
Leica Aperio slide scanner.
Ultrasound imaging
Transcranial ultrafast Doppler imaging.—
Mice (8–10 weeks old) were maintained
under 1.5% isoflurane anesthesia on a temperature-controlled imaging platform with a rectal
probe (Stoelting Co.). Following head depilation (Nair) and insertion of a catheter with a 30-
g needle into the lateral tail vein (fixed in place with GLUture), the mice were head-fixed in
a stereotaxic frame inside a light- and sound-proofed box on an optical table. A 16 MHz
transducer (Vermon) connected to a programmable ultrasound scanner (Verasonics Vantage)
was coupled to the head through a column of ultrasound gel (centrifuged at 2000xg, 10 min
to remove bubbles). The transducer was positioned to capture a full coronal section at an
arbitrary plane along the rostrocaudal axis. Once the internal temperature of the mouse
stabilized at 37°C, power Doppler images were acquired every 4 s for up to 60 min using a
previously described functional ultrasound script with slight modifications.
25
Briefly, the
pulse sequence consisted of 11 tilted plane waves (varying from −10 to 10 degrees), each
containing 8-half-cycle emissions at a voltage of 15V (900 kPa peak positive pressure
measured in free water tank). An ensemble of 250 coherently compounded frames, collected
at a framerate of 500 Hz, was then processed through a singular value decomposition filter
to isolate blood signals from tissue motion and generate a single power Doppler image. 300
s after the start of imaging, 100 μL OD30 native GVs were infused over 10 s by syringe
pump. Pixel-wise signal enhancement was calculated as the ratio of intensity at each time
point relative to its mean intensity in the first 75 frames. Time courses were then extracted
by averaging signal enhancement within a manually defined region of interest encompassing
the whole brain, processed with a 10-sample moving mean filter and normalized to the
global maximum.
Liver amplitude modulation imaging.—
Mice (10–14 weeks) were maintained under
2% isoflurane anesthesia on a mouse heating pad controlled by a rectal probe (TCAT-2LV,
Physitemp Instruments). After depilation of the abdomen (Nair) and insertion of a 30-g tail
vein catheter, the mice were secured in a supine position with surgical tape. Ultrasound
imaging was performed with an 18 MHz, 128-element linear array transducer (L22–14v,
Verasonics) mounted on a custom-made manual translation stage and positioned such that
the liver was at approximately 8 mm in depth. Once the internal temperature of the mouse
stabilized at 37°C, B-mode and amplitude modulation images were simultaneously acquired
every 4 s for up to 90 min. All images were reconstructed from 128 focused beam ray lines.
Ling et al.
Page 9
ACS Nano
. Author manuscript; available in PMC 2021 March 22.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
Each ray line was transmitted at 18 MHz from a 32-element active aperture with a focal
depth of 8 mm and peak positive pressure of 600 kPa (measured in free water tank). B-mode
images were reconstructed from a single pulse, while amplitude modulation was
implemented by first transmitting a single pulse from the full active aperture, followed by
two pulses where the even and odd elements in the active aperture are sequentially silenced.
24
Stripped GVs (OD 30, 100 μL) were manually injected as a bolus after 300 s. Image
processing and display were performed by internal Verasonics programs. Time courses were
calculated as the average signal intensity within a manually defined rectangular region of
interest encompassing the liver. To enable comparison, the time courses were smoothed by
robust locally weighted-regression using linear least squares, baseline corrected with respect
to the first 75 time points and normalized to the global maximum.
In vitro macrophage imaging.—
Wild-type RAW264.7 cells were seeded onto
fibronectin-coated cover slips (2e6 cells/2mL DMEM). After 24 h, the culture media was
exchanged with fresh DMEM containing bafilomycin A1 (100 nM) or vehicle (0.01% v/v
DMSO). Media of the same composition was used for all subsequent steps. Following a 1 h
pretreatment, a GV suspension composed of 320 μL fresh media and 80 μL stripped GVs
(OD10 in PBS) was dropped at the center of a UV-sterilized Parafilm-lined 6-well plate and
a cover slip was floated on top, cell-side down. This GV suspension was freshly prepared
immediately prior to uptake. After incubation at 37°C for 30 min, the cover slips were
transferred to pre-warmed fresh media and incubated for the desired amount of time. The
media was then aspirated and the cover slips were gently washed once with 2 mL room
temperature PBS. Cells were detached with 500 μL 0.25% trypsin-EDTA (Genesee
Scientific), neutralized with 1 mL media, and pelleted by centrifugation (300×g, 5 min,
4°C). From this point on, special care was taken to minimize exposure of the cells to
temperatures above 4°C. The pellet was washed once with 1.4 mL ice-cold PBS and
resuspended in 50 μL cold serum-free DMEM with 25 mM HEPES before loading into an
ultrasound phantom (1% agarose in PBS). Cell densities were manually counted by
hemocytometer.
The phantoms were imaged with a 128-element linear array transducer (L10–4v, Verasonics)
mounted on a custom manual translation stage using a previously described cross-
propagating amplitude modulation pulse sequence
26
that was modified to acquire single
frames. Briefly, each frame consisted of 64 ray lines transmitted at 4V (400 kPa peak
positive pressure in water) and 6 MHz from a 65-element aperture. Within the active
aperture, amplitude modulation was implemented by sequentially sending a plane wave
angled at 19.5° from the first 32 elements, a plane wave angled at −19.5° from the last 32
elements, followed by simultaneous emission of both plane waves. The first 3 frames were
saved along with a post-collapse image (after 10 insonations at 30 V). Signal intensities
were extracted from manually selected circular regions of interest with diameters of 1.8 mm,
baseline corrected by subtraction of signal from the post-collapse image, and adjusted for
cell density. The time courses from each run were then normalized to the mean intensity
from the samples harvested immediately after uptake (t=0).
Ling et al.
Page 10
ACS Nano
. Author manuscript; available in PMC 2021 March 22.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
GV contrast measurement.—
Phantoms were constructed as previously described.
19
Briefly, phantoms were made by embedding stripped GVs in 1% agarose in PBS and imaged
with the same parameters used for liver imaging. Signal intensities were extracted from
manually defined regions of interest.
Fluorescence imaging
Whole organ fluorescence.—
Mice were prepared as described above for transcranial
neuroimaging, with the only modification being that the GVs were fluorescently-labeled
with Alexa Fluor 647. Ninety minutes after GV injection, the mice were transcardially
perfused with 30 mL of cold heparinized PBS (10 U/mL, Sigma Aldrich). The heart, lungs,
kidneys, spleen, and liver were then carefully excised and stored in ice-cold Fluorobrite
DMEM (Gibco) prior to analysis. Images were acquired on a Bio-Rad ChemiDoc MP
imaging system using red epi-illumination and a 695/55 nm filter with an exposure time of
0.5 s. Integrated intensities were then calculated using the built-in “Analyze Particles”
function in FIJI.
Immunofluorescence.—
Mice were prepared as described above for whole organ
fluorescence. Ninety minutes after GV injection, the mice were transcardially perfused with
30 mL of cold heparinized PBS, followed immediately by 20 mL 10% neutral buffered
formalin. The liver and spleen were removed and immersed in formalin overnight (4°C).
Each organ was then sectioned with a vibrating microtome (75 μm, Compresstome,
Precisionary Instruments). Slices were blocked and permeabilized (2h, rt) with PBS
containing 10% goat serum (Sigma Aldrich), 0.2% Triton X-100 (Fisher Scientific), and
0.1% sodium azide (Sigma Aldrich). Each slice was stained for macrophages with rat anti-
mouse F4/80 (BioLegend, 1:200 dilution, overnight, 4°C) and Alexa Fluor 594 goat anti-rat
IgG secondary antibody (2h, rt, 1:400 dilution). The sections were mounted with ProLong
Diamond with DAPI (Invitrogen) and allowed to harden overnight before imaging with a
Zeiss LSM 800 confocal microscope through a 10x or 20x objective. Imaging parameters
prioritized signal specificity over speed.
Confocal microscopy images of entire liver slices were background subtracted in FIJI (20
px, rolling ball method). Randomly selected 500 px by 500 px regions of interest—
simulating the sampling of arbitrary fields of view -were exported to Ilastik
41
for processing.
The “Density Counting” workflow was used to count macrophages (Fig. S6). Images were
also segmented into macrophage and non-macrophage regions with the “Pixel
Classification” workflow and loaded into MATLAB for colocalization analysis (Fig. S3).
Live-cell imaging.—
1e5 RAW264.7 cells expressing palmitoylated mWasabi were seeded
on fibronectin-treated 35mm glass-bottom dishes. After 24 h, the culture media was
exchanged with serum-free Fluorobrite DMEM containing 25 mM HEPES and either 100
nM bafilomycin A1 (Cayman Chemical) or vehicle (0.01% v/v DMSO). Following a 1 h
incubation, this media was replaced with a 200 μL freshly-prepared suspension of OD 1.2
dually-labeled GVs. The well was then sealed with a UV-sterilized 18mm circular glass
cover slip and inverted for 5 min at 37°C to allow for contact and uptake.
Ling et al.
Page 11
ACS Nano
. Author manuscript; available in PMC 2021 March 22.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
Laser scanning confocal images were acquired every 2 min for 1h on a Zeiss LSM 800
microscope with a large incubation chamber maintained at 37°C. High magnification images
were acquired through a 63x oil immersion objective. Population level images were acquired
through a 20x objective. In both cases, acquisition parameters were set to optimize speed.
Image files were loaded into FIJI, visualized by maximum intensity projection, de-speckled
with a 1-px median filter and quantified by integration of signal intensities across the entire
field of view.
Pharmacokinetic modeling
A two-compartment pharmacokinetic model was implemented in MATLAB as the following
system of ordinary differential equations:
dB
dt
= −
k
1
B
(1)
dL
dt
=
k
1
k
c
B
−
k
2
L
(2)
Where B represents GV contrast in the blood and L represents GV contrast in the liver.
These variables were then directly parametrized with normalized Doppler and AM signal
time courses, respectively, and the constants were derived by non-linear least squares curve
fitting with initial values of 0 and bounds of 0 to 1.
k
1
and
k
2
represent rates of phagocytosis
and lysosomal degradation, respectively.
k
c
is a constant relating the blood Doppler signal to
the liver nonlinear signal. Input data were all distinct combinations of Doppler and AM time
courses from each biological condition. Output values are tabulated in Table S1.
Statistical analysis
Sample sizes were chosen based on preliminary experiments to yield sufficient power for the
proposed comparisons. Statistical methods are described in applicable figure captions.
Data and code availability
All gas vesicles, plasmids, data and code are available from the authors upon reasonable
request.
Supplementary Material
Refer to Web version on PubMed Central for supplementary material.
Acknowledgements
The authors thank A. Collazo and the Caltech Biological Imaging Facility of the Beckman Institute for assistance
with optical microscopy; the City of Hope Analytical Cytometry Core Facility for assistance with cell sorting; the
UCLA Translational Pathology Core Laboratory for assistance with tissue histology; J. Szablowski for helpful
advice on tissue immunofluorescence; L. Frankiw and D. Baltimore for lentiviral plasmids and assistance with
macrophage cell lines; and D. Piraner, A. Lakshmanan and D. Wu for fruitful discussions. This research was
supported by the National Institutes of Health (grant R01-EB018975 to M.G.S.) and the Human Frontier Science
Program (grant RGP0050/2016 to M.G.S.). B.L was supported by the NIH/NRSA Pre-Doctoral Training Grant
(T32GM07616) and the Caltech Center for Environmental and Microbial Interactions. J.L. was supported by the
Paul and Daisy Soros Fellowship. D.M. was supported by the Human Frontier Science Program Cross-Disciplinary
Ling et al.
Page 12
ACS Nano
. Author manuscript; available in PMC 2021 March 22.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
Fellowship (LT000637/2016). Related research in the Shapiro laboratory is supported by the Pew Charitable Trust,
the David and Lucile Packard Foundation and the Heritage Medical Research Institute.
References
1. Tavares AJ; Poon W; Zhang Y-N; Dai Q; Besla R; Ding D; Ouyang B; Li A; Chen J; Zheng G;
Robbins C; Chan WCW, Effect of Removing Kupffer Cells on Nanoparticle Tumor Delivery. Proc.
Natl. Acad. Sci. U. S. A 2017, 114, E10871–E10880. [PubMed: 29208719]
2. Berg RD; Levitte S; O’Sullivan MP; O’Leary SM; Cambier C; Cameron J; Takaki KK; Moens CB;
Tobin DM; Keane J; Ramakrishnan L, Lysosomal Disorders Drive Susceptibility to Tuberculosis by
Compromising Macrophage Migration. Cell 2016, 165, 139–152. [PubMed: 27015311]
3. Block ML; Zecca L; Hong J-S, Microglia-Mediated Neurotoxicity: Uncovering the Molecular
Mechanisms. Nat. Rev. Neurosci 2007, 8, 57–69. [PubMed: 17180163]
4. Wang C; Yue H; Hu Z; Shen Y; Ma J; Li J; Wang X-D; Wang L; Sun B; Shi P; Wang L; Gu Y,
Microglia Mediate Forgetting via Complement-Dependent Synaptic Elimination. Science 2020,
367, 688–694. [PubMed: 32029629]
5. Kazankov K; Jørgensen SMD; Thomsen KL; Møller HJ; Vilstrup H; George J; Schuppan D;
Grønbæk H, The Role of Macrophages in Nonalcoholic Fatty Liver Disease and Nonalcoholic
Steatohepatitis. Nat. Rev. Gastroenterol. Hepatol 2019, 16, 145–159. [PubMed: 30482910]
6. Netea MG; Balkwill F; Chonchol M; Cominelli F; Donath MY; Giamarellos-Bourboulis EJ;
Golenbock D; Gresnight MS; Heneka MT; Hoffman HM; Hotchkiss R; Joosten LAB; Kastner DL;
Korte M; Latz E; Libby P; Mandrup-Poulsen T; Mantovani A; Mills KHG; Nowak KL, et al., A
Guiding Map for Inflammation. Nat. Immunol 2017, 18, 826–831. [PubMed: 28722720]
7. Poon W; Zhang Y-N; Ouyang B; Kingston BR; Wu JLY; Wilhelm S; Chan WCW, Elimination
Pathways of Nanoparticles. ACS Nano 2019, 13, 5785–5798. [PubMed: 30990673]
8. Tsoi KM; MacParland SA; Ma X-Z; Spetzler VN; Echeverri J; Ouyang B; Fadel SM; Sykes EA;
Goldaracena N; Kaths JM; Conneely JB; Alman BA; Selzner M; Ostrowski MA; Adeyi OA; Zilman
A; McGilvray ID; Chan WCW, Mechanism of Hard-Nanomaterial Clearance by the Liver. Nat.
Mater 2016, 15, 1212–1221. [PubMed: 27525571]
9. Chow A; Brown BD; Merad M, Studying the Mononuclear Phagocyte System in the Molecular Age.
Nat. Rev. Immunol 2011, 11, 788–798. [PubMed: 22025056]
10. Blanco E; Shen H; Ferrari M, Principles of Nanoparticle Design for Overcoming Biological
Barriers to Drug Delivery. Nat. Biotechnol 2015, 33, 941–951. [PubMed: 26348965]
11. Anstee QM; Reeves HL; Kotsiliti E; Govaere O; Heikenwalder M, From NASH to HCC: Current
Concepts and Future Challenges. Nat. Rev. Gastroenterol. Hepatol 2019, 16, 411–428. [PubMed:
31028350]
12. Maresca D; Lakshmanan A; Abedi M; Bar-Zion A; Farhadi A; Lu GJ; Szablowski JO; Yoo S;
Shapiro MG, Biomolecular Ultrasound and Sonogenetics. Annu. Rev. Chem. Biomol. Eng 2018, 9,
229–252. [PubMed: 29579400]
13. MacParland SA; Tsoi KM; Ouyang B; Ma X-Z; Manuel J; Fawaz A; Ostrowski MA; Alman BA;
Chan WCW; McGilvray ID, Phenotype Determines Nanoparticle Uptake by Human Macrophages
from Liver and Blood. ACS Nano 2017, 11, 2428–2443. [PubMed: 28040885]
14. Russell DG; VanderVen BC; Glennie S; Mwandumba H; Heyderman RS, The Macrophage
Marches on Its Phagosome: Dynamic Assays of Phagosome Function. Nat. Rev. Immunol 2009, 9,
594–600. [PubMed: 19590530]
15. Canton J; Khezri R; Glogauer M; Grinstein S, Contrasting Phagosome pH Regulation and
Maturation in Human M1 and M2 Macrophages. Mol. Biol. Cell 2014, 25, 3330–3341. [PubMed:
25165138]
16. Iijima H; Moriyasu F; Tsuchiya K; Suzuki S; Yoshida M; Shimizu M; Sasaki S; Nishiguchi S;
Maeyama S, Decrease in Accumulation of Ultrasound Contrast Microbubbles in Non-Alcoholic
Steatohepatitis. Hepatol. Res 2007, 37, 722–730. [PubMed: 17559420]
17. Yoshikawa S; Iijima H; Saito M; Tanaka H; Imanishi H; Yoshimoto N; Yoshimoto T; Futatsugi-
Yumikura S; Nakanishi K; Tsujimura T; Nishigama T; Kudo A; Arii S; Nishiguchi S, Crucial Role
of Impaired Kupffer Cell Phagocytosis on the Decreased Sonazoid-Enhanced Echogenicity in a
Ling et al.
Page 13
ACS Nano
. Author manuscript; available in PMC 2021 March 22.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript