Arterial Thrombosis Models
Carotid Artery
The carotid artery injury is often used as a model to study arterial thrombosis in vivo. In these models, the injury can be induced mechanically, through a laser or chemicals
such as ferric chloride. Depending on the type of injury, several parameters can be
analyzed such as occlusion time, platelet deposition, and thrombus size. Although
the carotid artery is not easily accessible, its large size enables also blood flow
measurements (e.g., using ultrasonic probes).
IVM of carotid artery ligation and ferric chloride (FeCl3) injury are two popular experimental approaches used in assessing arterial thrombosis.
Carotid artery ligation involves mechanical damage caused by tying a suture around
the artery for a given time. The resulting loss of the endothelial cell layer at the
site of injury[1] initiates platelet tethering within the first minutes after endothelial denudation.
Platelet adhesion is mostly firm and irreversible and adherent platelets recruit additional
platelets from the circulation, thus contributing to the growing thrombus. This dynamic
process depends on glycoprotein VI (GPVI), the major platelet collagen receptor and
GPIbα which mediates platelet collagen binding indirectly, through von Willebrand
factor (VWF).[1] Tissue factor microparticles are considered to contribute to thrombus formation,
too.[2] At the sites of vascular injury, plasma VWF promotes platelet adhesion to the extracellular
matrix, and platelet aggregation, in particular as wall shear rates increase.[3]
[4] Common readouts are platelet deposition (platelet count) and thrombus growth (area
covered) over time ([Fig. 2]).
Fig. 2 Intravital imaging of the carotid artery thrombosis model. Mice were anesthetized
and a polyethylene catheter was implanted in the jugular vein. Rhodamine B ex vivo–stained platelets were infused. The left common carotid artery was dissected and
ligated with a polypropylene suture for 5 minutes. Platelets were visualized by in vivo epifluorescence high-speed video microscopy 5 or 30 min after vascular injury.
Ferric chloride injury involves the topical application of ferric chloride solution
to the adventitial surface of the carotid artery. This causes endothelial cell damage
and exposure of the subendothelial matrix resulting in platelet activation, aggregation,
and thrombus formation. The application of FeCl3 to the vessel wall causes major oxidative stress and generation of free radicals
leading to lipid oxidation, destruction of endothelial cells, and subsequent thrombus
formation. More specifically, FeCl3 crosses the endothelium in small vesicles by an endocytic–exocytic pathway, thereby
causing complete endothelial denudation.[5] Microparticles containing ferric iron and tissue factor, surrounded by a tissue
factor-rich endothelial cell membrane bilayer, are also present in the thrombus environment[6] and are in direct contact with fibrin fibers already 2 minutes after injury. Apart
from the direct effect of FeCl3 on the endothelial cell layer, it also caused erythrocyte hemolysis and hemoglobin
oxidation ex vivo which promoted endothelial denudation.[7] However, this could not be reproduced in vivo.[6] Both thrombin generation and platelets were shown to play an important role in the
FeCl3 model.[6] And while there was some debate over the role of the collagen receptor GPVI,[6]
[8] using genetic knockout mice as well as blocking antibodies, it was shown that GPVI
and GPIb contribute to thrombus formation in FeCl3-induced injury models in vivo.[1]
[9] Typically, time to vessel occlusion and thrombus size are quantified and analyzed.
Several factors need to be standardized in this model such as the concentration of
FeCl3 in the solution (usually between 5 and 20%), the application (soaked filter paper
or drop), as well as the incubation time (2–5 minutes). High FeCl3 concentrations can result in vessel opacification hampering the thrombus visualization.
Both techniques have their advantages and limitations. Carotid artery ligation is
a more controlled method in terms of thrombus initiation and is easy to standardize
between animals. Ferric chloride injury on the other hand is more difficult to standardize.
Despite carefully controlling the size and position of the FeCl3-soaked filter paper, thrombus formation can vary[5] with some vessels showing partial occlusion while others exhibiting exuberant thrombus
formation that extends beyond the limits of the filter paper.
One drawback of both models presented is that neither mechanical nor ferric iron-induced
injuries are (patho)physiologic injuries. In addition, in both models, the injury
is caused to healthy blood vessels, which does not mimic the human disease, as the
main reason for arterial thrombosis is atherosclerotic plaque rupture. To overcome
this issue, Kuijpers and colleagues developed a model in Apoe-/-
mice fed with high fat diet to investigate thrombus formation after atherosclerotic
plaque rupture using an ultrasound probe.[10]
[11] In this model, atherosclerotic arteries showed loose thrombi consisting of platelets,
erythrocytes, collagen, and fibrin. Rupture was induced by ultrasound application
and the following platelet adhesion and thrombus formation was analyzed using intravital
fluorescence microscopy.
Studying Thromboinflammation Using (Transient) Middle Cerebral Artery Occlusion Models
Ischemic stroke in humans is commonly caused by blockade of the middle cerebral artery
(MCA). Therefore, MCA occlusion (MCAO) in rodents is frequently used to model ischemic
stroke and to study its pathophysiology and potential treatments. In the late 1980s,
Koizumi et al. and Longa et al. described the first MCAO models in rats that were performed without craniotomy.[12]
[13] Animals were anesthetized and carotid arteries were dissected and the common carotid
artery (CCA) was located next to the trachea. Ligatures were tied around the proximal
CCA and the external carotid artery (ECA). A tourniquet was loosely tied around the
distal CCA. A vessel clip was used to cut the blood flow at the distal CCA. Next,
an incision was made between the ligature at the proximal CCA and the vessel clip
at the distal CCA. A (silicon-tipped) filament was inserted into the CCA, the vessel
clip removed, and the filament advanced into the internal carotid artery and further
intracranially to occlude the MCA at its origin.[14] The end of the filament was secured by tightening the tourniquet and left in place
for 30 to 120 minutes to induce transient ischemia and allow subsequent reperfusion.
Alternatively, the filament stayed in place for 24 hours to cause permanent occlusion.[15] The filament-based transient (t)MCAO model is now the most common technique to model
ischemic stroke in rodents and causes neuronal cell death, glial cell activation,
and blood brain barrier damage.[16] However, significant practice is needed to reproduce comparable infarct volumes.
Typical outcome measures after (t)MCAO include infarct volume, histology/immune fluorescence
staining, and scoring of the neurological and motor function. To assess infarct size,
brain sections are stained using triphenyltetrazolium chloride (TTC) and the ischemic
(whitish) areas quantified using image analysis software. The Bederson test is commonly
used to assess neurologic functions,[17] whereas the string or grip test allows to score motor function.[18]
Restoring blood flow in acute stroke (e.g., through thrombolysis or mechanical thrombectomy)
remains the main treatment goal in the clinic. However, especially when reperfusion
is achieved too late, further damage and infarct growth is frequently observed—a phenomenon
termed reperfusion injury (RI). During RI, reactive oxygen species (ROS) are generated which directly damage cells
and spark inflammation (e.g., through the release of damage-associated molecular patterns
[DAMPs])[19] that leads to the upregulation of cell adhesion molecules and leukocyte recruitment.
In addition, platelets adhere to and become activated at the site of vascular damage,
thereby increasing the risk of secondary thrombotic events.[20] Thus, both thrombotic and inflammatory pathways are active following brain ischemia.
The interplay and amplification between these two pathways has been termed thromboinflammation.[21] Using the tMCAO model in combination with genetically modified animals has allowed
to unravel key players in thromboinflammation after stroke such as platelet receptors
GPIb and GPVI[22] as well as soluble factors released from platelet alpha and dense granules.[23]
[24]
Dong and colleagues assessed the impact of intravenous administration of Resolvin
D2 (RvD2)-loaded nanovesicles[25] on inflammation after ischemic stroke using the tMCAO model. Nanovesicles were labeled
with the lipid dye DiR (1,1′-dioctadecyl-3, 3, 3′, 3′-tetramethylindotricarbocyanine
iodide) which emits light in the near-infrared spectrum. Ex vivo analysis of brain
sections using an IVIS imaging system at different time points post reperfusion showed
that the nanovesicles specifically accumulated in the ischemic brain hemisphere.
To study the mechanism on a cellular level, the authors set up intravital microscopy
of the mouse brain using a cranial window. Mice were anesthetized and a dorsal midline
incision was made to expose the skull. A high-speed drill was used to remove a circular
area of the skull bone above the right brain hemisphere. A glass cover slip was placed
on the top of the dura mater and super glue was applied to secure it in place.[25]
[26] Lastly, a bar was mounted on the head to later attach the mouse to the microscope
imaging stage. After surgery, mice were allowed to recover before the tMCAO procedure.
Seven days after implantation of the cranial window, mice were imaged using an intravital
microscope (e.g., laser scanning confocal microscope or multiphoton microscope if
deeper tissue penetration is desired). DiR-labeled nanovesicles were injected 1 hour
after reperfusion, followed by BSA-Cy5 and Ly-6G-Alexa Fluor 488 antibody to label
the vessel lumen and neutrophils, respectively. Using this setup, the authors visualized
the binding of nanovesicles to inflamed brain vasculature in real time and also observed
the reduced recruitment of neutrophils after treatment with RvD2-loaded nanovesicles.[25]
In another study, Desilles and colleagues performed partial dura-sparing craniotomy
in rats followed by 60 minutes of tMCAO. Intravital fluorescence microscopy was used
to visualize platelets and leukocytes as well as fibrinogen during MCAO and after
recanalization in pial vessels.[27] During MCAO, they observed fibrin(ogen) deposition in postcapillaries colocalizing
with leukocyte and platelet accumulation. These microthrombi then caused full vessel
occlusion in one or more vessels per field of view. Platelets and leukocytes remained
attached even after reperfusion. On the other hand, Göb et al. used light sheet fluorescence microscopy after tMCAO and found that tissue damage
peaked at 8 hours after reperfusion, while thrombi were observed only at later time
points, thereby suggesting that thrombosis was not the driving force of secondary
infarct growth after stroke.[28]
Ishikawa et al. investigated the role of CD40 and CD40L in cerebromicrovascular dysfunction and tissue
damage following transient MCAO using a cranial window and intravital microscopy.
They performed a craniotomy 1 mm posterior from the bregma and 4 mm lateral from the
midline leaving the dura mater intact. A cover glass was placed over the exposed brain
tissue. Both CD40- and CD40L-deficient mice showed reduced infarct volumes after tMCAO
compared to control mice after 1 hour of MCAO followed by 4 hours of reperfusion.[29] Platelets were isolated from donor mice and labeled with carboxyfluorescein diacetate
succinimidyl ester (CFSE). 108-labeled platelets were infused into recipient mice over 5 minutes via the femoral
vein. Rhodamine 6G was used to label endogenous leukocytes. Using intravital microscopy
they found that both platelet and leukocyte adhesion was reduced in CD40/CD40L-deficient
animals.
These studies show that the tMCAO model combined with intravital microscopy, transgenic
mice, and appropriate labeling allows to obtain insight into the pathomechanism of
ischemic stroke and RI on the cellular level.
Mesentery
Due to their easy access, mesenteric vessels are very well suited for intravital microscopy
and different ways to induce thrombus formation exist. One common method uses application
of FeCl3 to the mesenteric arterioles (for a detailed protocol see Kuijpers and Heemskerk[11] and Bonnard and Hagemeyer[30]). Briefly, after anesthetizing the mice, staining or stained cells are injected
either via the tail or the jugular vein and the intestine is carefully exposed. Injury
is caused by placing a FeCl3-soaked filter paper on the arteriole of interest. In this model, thrombus growth
was shown to depend on platelet P2Y12 ADP receptor,[31] GPVI,[32] and VWF.[33] The method is rather simple, yet many variables need to be optimized. The reproducibility
is affected by the size of the vessels, which often vary even among mice of the same
age. Vessel-associated fat may compromise reproducibility, as it affects the extent
of the injury[30] and therefore the use of younger animals is preferred. A second approach which is
easier to standardize is the laser injury model; however, the thrombotic effect is
less pronounced compared to chemical injury.[33]
Apart from chemical or laser damage in mesenteric arterioles, there is the model of
in vivo gut ischemia/reperfusion (I/R) injury which is often investigated by inflicting damage
to the superior mesenteric artery—the most important vessel for blood supply to the
colon in rodents.[34] This model is used to mimic I/R-induced intestinal injury often occurring in bowel
ischemia, abdominal aortic aneurism repair, and abdominal compartment syndrome.[35] Intestinal I/R starts with ischemia which results in hypoxia and malnutrition, leading
to energy metabolism disorders.[36] Lack of ATP impacts a range of ATP-consuming processes such as cation pumps and
F-actin polymerization. Consequently, endothelial and epithelial barriers break down.
Restoration of blood flow, however, results in greater tissue injury[37] due to ROS production and ischemia RI.
Acute mesenteric infarction is characterized by a disrupted gut barrier and dysregulation
of the host immune response.[38]
[39] Bacteria or bacterial products leak through the compromised barrier from the gut
lumen to the circulation and promote leukocyte adhesion on the endothelial cells of
the mesenteric venules.[40]
[41] The process is more complex than FeCl3-induced injury as it is based on arterial occlusion. After identification of the
superior mesenteric artery, it is carefully occluded with a small vascular clamp.
It is important to keep the exposed intestine moist and to avoid trauma by touching
it, otherwise this would lead to unintended platelet and leukocyte adhesion that is
not caused by the blood stasis. The time of occlusion as well as the time of reperfusion
can be varied and afterwards the mesenteric venules are observed by microscopy and
platelet and leukocyte rolling and adhesion can be quantified ([Fig. 3]).
Fig. 3 Intravital imaging of mesenteric ischemia reperfusion injury. Mice were anesthetized
and a polyethylene catheter was implanted in the jugular vein. Rhodamine B ex vivo–stained platelets were infused. Following a midline laparotomy, the superior mesenteric
artery was identified and occluded with a small vascular clamp. After 60 minutes of
ischemia, reperfusion was allowed. Before and immediately after ischemia-reperfusion,
the entire small intestine was carefully taken out of the abdomen. Platelets and leukocytes
(stained with acridine orange infusion via the catheter shortly before imaging) were
visualized in situ by in vivo epifluorescence high-speed video microscopy in the mesenterial venules before and
after the vascular injury.
Cremaster Arteriole Laser Injury Model
By combining laser-induced injury and state-of-the-art imaging techniques, it is possible
to investigate the kinetics and molecular mechanisms of in vivo thrombus formation. Injury is achieved by using a focused laser pulse to the vessel
wall of an arteriole inside the cremaster muscle. With this technique, it is possible
to control the place and time of thrombus formation in vivo. The accumulation of the main components of thrombi—platelets and fibrin—can be visualized
easily by using fluorophore-conjugated antibodies combined with intravital microscopy.
Recent advances in this technique also allow the analysis of the thrombi architecture,
signaling pathways, calcium mobilization, blood flow velocity around the thrombus
and even the formation of multiple thrombi in close distance to each other.[42]
[43]
To perform this model, mice are anesthetized and a catheter is placed into the jugular
vein for application of additional anesthesia and fluorophore-conjugated antibodies
or labelled cells.[44] Externalization of the testis and anus as well as removal of excessive tissue and
fat allows free access to the cremaster muscle. Afterward, the cremaster muscle needs
to be expanded to a cover-slip and fixed with a pin to the imaging tray.[45] Visualization of platelets, fibrin, and other factors can be achieved either by
the administration of fluorophore-conjugated antibodies or by intravenous injection
of exogenous labelled platelets.[46]
The injury is usually induced using a pulsed nitrogen laser at 440 nm.[42] The power of the laser should be adjusted based on the desired size of the injury,
the diameter of the arteriole, the thickness of the muscle, and the size of the intended
thrombus. Thrombus-forming injuries to the vessel wall can already be induced with
one to two pulses,[46]
[47] but depending on the laser power, also up to 10 pulses, could be necessary.[48] The laser is focused on the luminal surface of the arterioles.[49] Some research groups perform their laser-induced injury by using repeated, very
short laser pulses (3–5 ns) with a distance of 500 nm between each in a two-by-two
pattern. With a total of 24 pulses/injuries, this method is gentle and no subendothelial
collagen is exposed during the vascular injury. The injury is achieved by laser-mediated
activation of the endothelium.[42]
[50] The groups of Atkinson et al. and Dubois et al. showed that there was no difference between a single injury per mouse versus multiple
upstream injuries.[46]
[47]
Shortly after laser-induced injury, thrombus formation can be observed using a bright-field
microscope or fluorescence microscope based on the aim of the experiment. Usually,
the whole procedure is performed in an intravital microscopic setting. Basic measurements
include the increase in fluorescence signal by labelled platelets and fibrin at the
site of the injury. Also, factors such as thrombin and tissue factor can be labelled
and their impact on thrombus formation observed.[49] By measuring the thrombus formation over time, a thorough understanding of the kinetics
under different conditions (based on the model) can be achieved. For instance, labeling
platelets, fibrin and activated αIIbβ3 integrin can give information on the kinetics of αIIbβ3 integrin activation as well as the size and stability of the thrombi.[51] To reduce concurrent activation, αIIbβ3 integrin inhibitors such as eptifibatide can be used.[47] To investigate the external environment of the thrombus like the velocity of red
blood cells, externally labelled red blood cells can be injected and their flow around
the thrombus measured with intravital microscopy and analyzed using image analysis
software such as Slidebook, Volocity, or Imaris.[50]
For the analysis of thrombus formation kinetics and the influence of emboli on downstream
thrombi, a standard bright-field microscope should be sufficient. Also, the diameter
of the vessel up- and downstream to the injury can be assessed.[50] For a more in-depth analysis of the thrombi components, the kinetics of platelet
and fibrin binding (taking the median fluorescence intensity and by co-localization
of the platelet and fibrin signal) or the influence of tissue factor, thrombin, or
αIIbβ3 integrin, a multichannel fluorescence microscope is essential.[52]
[53] The architecture of thrombi can later be inspected using confocal microscopy on
thin sections of the microvasculature or thrombi.[43]
[46] With intravital microscopy, it is also possible to record videos of the thrombus
formation.
One limitation of this technique is the high variability between single injuries.
The degree of injury is highly dependent on the vessel itself, the thickness of the
muscle, and the laser power.[42]
Venous Thrombosis
A very well characterized model to study venous thrombosis in vivo is the stenosis model of the inferior vena cava (IVC). This model mimics deep vein
thrombosis (DVT), occurring most frequently in lower limbs which could result in pulmonary
embolism if the thrombus detaches from its initial location and occludes the vasculature
of the lungs. These two diseases together are some of the most common vascular disease
worldwide.[54] DVT is usually caused by cancer, trauma, or blood stasis,[55] and once the thrombus gets detached from its initial location and reaches the lungs,
it leads to pulmonary embolism. In the in vivo model of DVT, thrombosis can be induced either by stenosis, ferric chloride, or laser
injury.[56]
Although the application of ferric chloride on the IVC can chemically induce thrombosis,[57] thrombus formation is caused by nonphysiologic processes, which are not representative
of the human disease, where thrombosis is usually initiated by blood stasis.[58] Nevertheless, ferric chloride induces a rapid leukocyte- and platelet-rich thrombus
formation already 1 minute after application.[57] Such fast response is different from human disease progression, but enables the
quick experimental observation of thrombus growth.[57] DVT in the IVC can also be induced by stenosis or complete occlusion models (for
protocols see Payne and Brill[59] and Wrobleski et al.
[60]).
During the vena cava stenosis model, a space holder is secured on the vessel with
a permanent narrowing ligature, restricting the flow. Removal of the wire leads to
flow restriction but not occlusion, mimicking human disease progression. The thrombus
growth can be observed periodically under the microscope, after infusion of the appropriate
fluorescently labeled cells, dyes, or antibodies. However, the in vivo observation time is limited to a certain timeframe during thrombus formation since
extensive imaging sessions might increase the risk of nonphysiologic reactions. At
the end of the experiment, the occlusive thrombus can be characterization ex vivo.
The ex vivo thrombus analysis proved that inflammatory processes and DVT are closely linked.
Leukocyte adhesion occurs as early as 1 hour after IVC stenosis and increases dynamically
on the thrombosed vein wall with tumor necrosis factor-α playing a pivotal role in
neutrophil extravasation. During thrombus maturation (day 6), neutrophils are accompanied
by monocytes and lymphocytes.[61] An in vivo model with 80% flow restriction leads to the formation of small thrombi after 6 to
12 hours with occlusion happening between 24 and 48 hours in 60% of the mice.[62]
The observation of early thrombus formation during murine DVT with intravital microscopy
(e.g., 6 hours after stenosis) showed that similarly with the ex vivo analysis, the process is dominated by leukocyte adhesion.[62] The exposure of adhesive molecules such as P-selectin and VWF factor proves that
the activation of the endothelium is a response to depressed blood flow.[62] The inflammatory response leads to activation of the coagulation cascade initiated
by leukocyte-derived tissue factor, which subsequently results in fibrin formation.
Neutrophils promote the growing thrombus further by releasing neutrophil extracellular
traps (NETs) 3 hours after flow restriction. In contrast to arterial thrombosis, where
platelets dominated the thrombus growth within minutes of stenosis, in this model,
platelets are adhering either on the activated endothelium or on leukocytes, but are
outnumbered by the inflammatory cells, which also proceed this process. Yet, the interaction
between platelets and neutrophils is critical for thrombus development. Inhibition
of platelet G protein–coupled receptors and immunoreceptor tyrosine-based activation
motif signaling results in reduced platelet adhesion to the inflamed vascular wall,
impaired neutrophil activation, and thus reduced stenosis-induced thrombus formation.[63]
The vena cava stenosis model has several limitations. To better mimic the human disease,
the thrombus formation should occur only due to the flow restriction and not based
on endothelial denudation. Therefore, all mice that show bleeding events during the
stenosis should be excluded from the study. Also, the thrombus occurrence often varies
or is completely absent.[62] The vascular anatomy in mice is highly variable. There are different numbers of
lateral and dorsal branches on the vena cava, even among mice of the same strain,
providing alternative circulatory paths for the venous flow. It has been shown that
the number of lateral branches is often the reason for the absence of thrombus formation
in this model.[64]
[65] Interestingly, the thrombus occurrence does not depend on the flow in these branches,
rather on the distance of the confluent side branch from the stenosis suture. Ligation
of these branches does not impact the thrombus size or the variability of thrombus
formation[66] and reduces the variability of the model but induces endothelial injury.[67] In order to visualize the early stage of thrombus formation in the completely occluded
IVC in vivo, the side branches need to be intact so that the flow persists.
Liver Thrombosis Models
Infection-Mediated Liver Thrombosis
Thrombosis is a common complication of systemic infection and can be associated with
multiorgan damage. Inflammation following infection activates platelets, which may
accompany endothelial damage and thrombus development. This process is frequently
referred to as immunothrombosis.[68]
Here, we highlight different infection-induced thrombosis models focusing our attention
on thrombi formation in the liver and the techniques that have been used, so far,
to study this phenomenon. Depending on the model, liver thrombosis can be induced
following the injection of a single antigen such as LPS and alpha toxin, or entire
bacteria in mice.[69]
Studies by Hitchcock et al. as well as Beristain-Covarrubias et al. focused on inflammation-driven thrombosis induced by Salmonella Typhimurium (STm) infection of mice.[70]
[71] By means of tissue staining and flow cytometric analysis of liver tissue, they observed
a novel pathway of thrombosis. After infection, inflammation directly induces thrombosis
by upregulating podoplanin on monocytes/macrophages resulting in CLEC-2-mediated platelet
activation.[70] Immunohistological staining also showed that thrombus formation in the liver was
occurring 7 days after infection and persisted even after the infection was resolved.
Another study focusing on STm infection showed that thrombosis can have different
kinetics. By immunohistological and immunofluorescent staining, they characterized
both liver and spleen and observed that thrombi formation can occur sequentially in
multiple tissues, opening the doors for a possible targeted therapeutic treatment.[71]
During staphylococcal sepsis, patients develop microvascular thrombosis with a tendency
toward systemic bleeding, possibly caused by abnormalities in platelet function.[72]
[73]
[74] Intravenous injection of alpha toxin has been used to establish a mouse model of
staphylococcal sepsis.[75]
[76]
[77] Surewaard and colleagues took advantage of spinning disk intravital microscopy to
investigate the role of Staphylococcus aureus infection on the microvasculature.[77] This powerful technique allows to study the interactions of microbes with different
host cells, such as immune cells and platelets in real time. Mice were anesthetized
and a tail vein catheter was inserted to inject fluorescent antibodies and to maintain
anesthesia. An incision is made along the midline of the abdomen to expose underlying
organs. The mouse is placed on its right side onto the microscope stage and one liver
lobe is carefully guided onto the microscope cover glass.[78] To model acute sepsis, alpha-toxin, culture supernatants, or live bacteria, were
injected intravenously and platelet aggregation and bacteria were observed and analyzed
using image analysis software to quantify the amount of platelet aggregation and thrombi
formation ([Fig. 4]). The approach used in this study demonstrated that intravenous alpha-toxin injection
results in rapid platelet aggregation with subsequent micro-thrombi formation in the
microcirculation. These aggregates accumulated in the liver sinusoids and kidney glomeruli
causing multiorgan dysfunction. Furthermore, they observed that treatment with an
alpha-toxin neutralizing antibody prevented platelet aggregation and subsequent damage,
without affecting platelet contribution to eradicating S. aureus infection.[77]
Fig. 4 Intravital imaging of sepsis-associated liver thrombosis. Mice were anesthetized
and a jugular vein catheter was inserted to inject fluorescent antibodies and maintain
anesthesia. Image acquisition was performed using an inverted spinning disk confocal
microscope. Kupffer cells and platelets were visualized using fluorescently labeled
antibodies against F4/80 and CD49b, respectively. Mice were intravenously injected
with 5 × 107 colony-forming units (CFUs) of GFP-expressing Staphylococcus aureus. Shown are images before and 30 min post S. aureus injection.
Hepatic Ischemia Reperfusion Injury
Liver resection and liver transplantation sometimes represent the only available treatment
for patients with liver disease.[79] Hepatic I/R injury is a common complication of these procedures and is one of the
primary causes of early organ dysfunction and failure after the procedure. I/R injury
is a pathological process involving ischemia-mediated cellular damage followed by
a paradoxical exacerbation upon reperfusion of the liver.[80]
[81] This phenomenon engages many different cell types and several signaling pathways.
During the first phase, the ischemic insult exposes hepatic cells to oxygen deprivation,
pH changes, and ATP depletion. This condition forces the cells to rely on anaerobic
metabolism for energy production.[82] Consequently, ROS production is enhanced, leading to organelle damage, hepatic cell
injury, and death.[83] The following reperfusion causes a profound inflammation which aggravated hepatocellular
damage.[84]
[85]
[86] Even though I/R injury has been widely investigated, its mechanisms remain largely
unclear. In this context, animal models are valuable tools for understanding the physiopathology
of hepatic I/R injury, clarifying the molecular mechanisms involved and discovering
novel therapeutic targets and drugs. DAMPs released from dead cells after I/R injury
activate Kupffer cells leading to the release of chemokines and cytokines and result
in the recruitment of immune cells such as circulating monocytes, neutrophils, and
T cells to promote hepatocellular injury.[87] Zhang et al. studying a mouse model of I/R injury showed by flow cytometry, that Kupffer cells
were a major source of CCL2, which attracts CCR2+ neutrophils into the ischemia–reperfusion-stressed liver.[88] Kupffer cells also produce CXCL1, CXCL2, and CXCL8 upon liver injury.[89]
[90] A study by de Oliveira et al. used confocal intravital microscopy to image neutrophil migration in the liver of
Lysm-eGFP mice subjected to I/R injury.[91] These reporter mice express the green fluorescent protein (eGFP) mainly in neutrophils.[92] Mice are anesthetized, a midline laparotomy is performed, and mice are subjected
to I/R by occluding the pedicle of the left and median lobes of the liver, which contains
the bile duct, hepatic artery, and portal vein, using an atraumatic clamp. After 60 minutes
of ischemia, the clamp is removed and reperfusion is allowed. Neutrophil accumulation
in the liver is observed using intravital microscopy of GFP-bright cells, excluding
low–GFP-expressing mononuclear cells.[93] After different reperfusion times, the mice are anesthetized and the liver is carefully
exposed. Sytox orange is injected intravenously to stain DNA and show necrotic cells.
Imaging analysis software is used to analyze neutrophil count, location, and migration
distance. Time points between 1 to 48 hours after reperfusion were chosen for examination.
They observed that treating mice with reparixin, a CXCR1/2 receptor antagonist, significantly
decreased neutrophil activation and infiltration in the liver. Additionally, reparixin
was able to reduce the reperfusion-associated tissue damage, suggesting that the blockage
of CXCR1/2 could be a possible therapy for patients undergoing liver surgery.[91]
The first evidence of the involvement of platelets in the development of I/R-induced
hepatic injury was obtained by Cywes et al. who showed platelet infiltration in the perfused rat liver using electron microscopy
and in the human liver using immunohistochemistry.[94]
[95]
[96] Later, Khandoga et al., using intravital microscopy, observed that hepatic I/R induced both transient interaction
and permanent adhesion of platelets to the post-ischemic liver sinusoids. The number
of these contacts during early reperfusion depended on ischemia time.[97] P-selectin is a major adhesion molecule involved in platelet and leukocyte adherence
to the endothelium.[98] Mice deficient for P-selectin display reduced platelet and neutrophil accumulation,
a decrease in post-ischemic liver injury, and improved survival following warm ischemia.[96]
[99] Treatment with the selective PAR-4 antagonist TcY-NH2 attenuated I/R-induced platelet
and CD4+ T-cell recruitment, improved sinusoidal perfusion, and reduced apoptotic injury.[100] Also, antifibrinogen antibody administration inhibited platelet adhesion and decreased
short-term liver injury in a mouse model of liver I/R injury.[101] In response to I/R, platelets become activated and release factors that promote
liver injury as well as hepatic regeneration such as platelet-activating factor (PAF),
cytokines, nitric oxide (NO), transforming growth factor-β (TGF-β), and serotonin.
PAF, also produced by Kupffer cells and liver sinusoidal endothelial cells, acts on
neutrophils boosting ROS generation, further contributing to the amplification of
the neutrophil response.[102] NO produced by platelets combined with oxygen-free radical formation upon reoxygenation
of the ischemic liver can lead to the production of peroxynitrite, which can support
programmed cell death in endothelial cells.[103]
[104]
[105] Platelet-derived serotonin has been exhibited to promote hepatocellular proliferation
and tissue repair, also in the context of hepatic ischemia in mice.[106]
[107] Nocito et al. assessed the impact of platelets in hepatic I/R injury using mouse models of impaired
platelet function and platelet depletion via clopidogrel administration and injection
of anti-CD41 antibody, respectively. Partial hepatic ischemia was induced for 60 minutes,
and different time points of reperfusion were evaluated. Interestingly, they observed
that impairment of platelet function and platelet depletion had no direct effect on
I/R injury, while liver regeneration and repair were significantly impaired in platelet-depleted
animals.[107]
Another study conducted by Zhang et al. focused on the possible interplay between neutrophils and platelets in liver I/R
injury.[108] In this study, I/R was performed by occluding the left and median liver lobes with
a microvascular clamp for 90 minutes, followed by reperfusion times of 6, 12, and
24 hours. Interestingly, flow cytometric analysis showed that mice subjected to liver
I/R displayed systemic platelet activation and increased platelet-neutrophil aggregates.
Liver I/R resulted also in injury in distant organs such as kidney and lung with increased
NETs and platelet-rich microthrombi formation. Blocking NETs by DNase treatment decreased
organ damage as well as thrombi formation in mice. Moreover, using a platelet-specific
TLR4 KO mouse model, they observed that following liver I/R, mice had reduced distant
organ injury with decreased circulating platelet activation and platelet-neutrophil
aggregates.[108]
