WITHDRAWN: Macrophage-targeting Fasudil treatment protects liver from the ischemia/reperfusion injury by promoting M2 macrophage polarization

Macrophages play essential roles in the generation and resolution of inflammation. Ischemia-reperfusion injury (IRI) triggers a systemic inflammatory response and leads to cellular injury and organ failure. During surgical procedures of the liver, such as hepatic resection and liver transplantation, IRI leads to the dysfunction of the liver. Rho-associated protein kinase (ROCK) inhibitors were reported protecting the liver from IRI. However, the systematic administration of ROCK inhibitors causes severe hypotension. Here, using Fasudil carried liposomes, we specifically inhibited the ROCK-II expression in Kupffer cells and blood monocytes. Through this macrophage/monocyte specific treatment of Fasudil, we successfully protected the liver from IRI by shifting Kupffer cells/monocytes from M1/classical to M2/patrolling phenotype in the liver and peripheral blood. Our finding provides novel insights into the macrophage/monocyte-specific drug delivery and the treatment of liver IRI.


Introduction
Ischemia-reperfusion injury (IRI) is one of the most critical complications commonly associated with liver surgery [1], such as hepatic resection and liver transplantation. It comprises a sterile inflammatory response that damage organs. Macrophages/monocytes are vital in innate immunity and inflammation [2], such as phagocytosis [3,4] and resolution [5]. Macrophages/monocytes also play important roles in regulatory immunity, such as antigen presenting [4]. Recent studies showed that there are at least two types of macrophages or monocytes, M1 and M2 macrophages [6] or classical/inflammatory and patrolling/regulatory monocytes [7], respectively. M1 macrophages are the "killer" type macrophages, which promote inflammation and rapidly eliminate the pathogens but also cause tissue damage. M2 macrophages are the "heal" type macrophages, which routinely repair and maintain tissue integrity.
Fasudil is a potent Rho-associated protein kinase (ROCK) inhibitor [8,9,10,11,12,13] and vasodilator and has been used for the treatment of cerebral vasospasm [9] and many other diseases, such as cancer [14,15], allergy [16] and inflammation [17,18]. The systematic administration of Fasudil was reported protecting mice from experimental autoimmune encephalomyelitis (EAE) and shifting macrophages from M1 to M2 [19]. Enhanced M2 macrophage polarization was reported to protect the liver from IRI [20]. However, the systematic administration of ROCK inhibitors, such as Fasudil, causes severe side effects, such as 7 PBS, and were resuspended in 4 mL PBS. The lipid suspension was then forced through a polycarbonate filter with the successively smaller pore size (1000 nm, 500 nm, and 300 nm) mounted on a holder in a heating block (Avanti®mini-extruder). The size distribution of the liposomes was assessed by a Malvern Zetasizer Nano ZS (Malvern Instruments Ltd., UK).

Fluorescence microscopy
Liposomes carrying Fasudil and AF488 was imaged by an IX71 inverted fluorescence microscope

Liposome drug administration
Three administrations of the liposome suspension described above were applied to mice through retro-orbital intravenous (i.v.) injection (10 μL/g, 2 hours apart). One day after administrations, the efficiency and specificity of macrophage/monocyte targeting were assessed by flow cytometry. The CD45 + F4/80 + or CD45 + CD115 + cells were sorted to test the expression of ROCK-II. In the liver IRI experiments, surgeries were performed one day after the first three administrations, and one boost administration of the liposomes (10 μL/g) was performed every day. Liposomes without Fasudil were used as a negative control to avoid any unspecific effects for the liver IRI experiments.

Flow cytometry
During assessing the efficiency and specificity of macrophage/monocyte targeting, mice were euthanized by CO 2 inhalation. Peripheral blood and liver were harvested for the assay. The liver single-cell suspension was prepared using the Liver Dissociation Kit according to the manufacturer's instruction.
Erythrocytes in the blood sample or liver cell suspension were lysed using distilled water (RT, 10 seconds). The blood sample or liver cell suspension was incubated with BV421 conjugated anti-mouse CD45 antibody (2 μg/ml) and AF647 conjugated anti-mouse CD115 antibody (2 μg/ml) or AF647 conjugated anti-mouse F4/80 antibody (1 μg/ml), respectively, for 30 minutes on ice. After two washes, cells were re-suspended in FACS buffer (0.5% BSA and 0.05% sodium azide in PBS) and were analyzed on a FACSCalibur (BD Biosciences). Compensation controls were performed using UltraComp eBeads incubated with relative antibodies at RT for 10 minutes.
For sorting of CD115 + or F4/80 + leukocytes from blood samples or liver cell suspensions, respectively, the cells were stained using the same method as described above and the CD45 + CD115 + or CD45 + F4/80 + cells were sorted on a FACSAria (BD Biosciences). The CD115and CD4/80cells were sorted, respectively, as controls for specific targeting.
To assess the number of neutrophils in blood and liver during IRI, we collected peripheral blood samples at different time points and harvested livers 48 hours after the surgery. Samples were prepared as described above and were stained with BV421 conjugated anti-mouse CD45 antibody (2 μg/ml) and PE-conjugated anti-mouse Ly-6G antibody (2 μg/ml) for 30 minutes on ice. After two washes, cells were re-suspended in FACS buffer (0.5% BSA and 0.05% sodium azide in PBS) and were analyzed on a FACSCalibur (BD Biosciences). To assess the amount of classical and patrolling monocytes in blood or M1 and M2 macrophages in liver, we harvested peripheral blood samples and livers 48 hours after the surgery. Samples were prepared as described above. Blood samples were stained with BV421 conjugated anti-mouse CD45 antibody (2 μg/ml), PE-conjugated anti-mouse Ly-6C antibody (2 μg/ml), and AF647 conjugated anti-mouse CD115 antibody (2 μg/ml) for 30 minutes on ice. Intracellular staining was performed for liver single-cell suspensions: cells were fixed by 1% PFA for 10 minutes on ice, permeabilized by 0.5% Triton X-100 for one hour on ice, and treated with blocking buffer containing 2% BSA, 10% goat serum and 0.1% Tween-200 for one hour on ice. Then cells were stained with BV421 conjugated anti-mouse CD45 antibody (2 μg/ml), PE-conjugated anti-iNOS or anti-Arg-1 antibody (2 μg/ml), and AF647 conjugated anti-mouse F4/80 antibody (2 μg/ml) for 30 minutes on ice. After two washes, cells were re-suspended in FACS buffer (0.5% BSA and 0.05% sodium azide in PBS) and were analyzed on a FACSCalibur (BD Biosciences).

Western blot
Western blot was performed as described before [19,31]. Briefly, cell lysates of sorted F4/80 + or CD115 + cells were fractionated by the sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and detected by immunoblotting using rabbit anti-ROCK-II polyclonal antibody (1:1000) and

qRT-PCR
qRT-PCR was performed as described before [31]. Briefly, RNA in sorted F4/80 + or CD115 + cells was extracted using TRIzol. cDNA was synthesized using FastQuant RT Kit (with gDNase) according to the manufacturer's instructions. The expression level of ROCK-II, TNFα, IL-1β, and IL-10 mRNA was measured by Roche Lightcycler 480 (Roche, Nutley, NJ, USA) using SuperRealPreMix Plus (SYBR Green) and normalized to the expression level of GAPDH.

Measurement of serum ALT, AST, TBIL, TNFα, IL-1β, and IL-10
Serum ALT and AST activity were measured by mouse ALT and AST quantitative kits according to the manufacturer's instruction. TBIL, TNFα, IL-1β, and IL-10 were measured by ELISA kits according to the manufacturer's instruction. A Spectra MAX 190 plate reader was used for these measurements.

Cell-free DNA (cfDNA) Measurements
cfDNA was measured directly in sera using SYBR Gold Nucleic Acid Gel Staining, according to the fluorometric method published before [32,33]. Briefly, samples were added to a black 96-well plate (Greiner Bio-One). SYBR Gold was applied (1:10,000), and fluorescence was measured with a microplate fluorometer (BioTek) at an emission wavelength of 530/25 nm and an excitation wavelength of 485/20 nm. Background fluorescence was subtracted by using sera treated with RNase A (100 U) and DNase I (500 U) for five hours at 37°C as a negative control. Commercial Salmon sperm DNA was used as DNA standards.

Partial warm liver IRI procedure
The surgical procedure was performed as described before [20,34]. Briefly, mice were fasted for 18 h before experiments and were provided with tap water ad libitum. Mice were anesthetized with ketamine (100 mg/kg, i.m.) and xylazine (10 mg/kg, i.m.). The body temperature of mice was maintained at 37 °C throughout the anesthesia process. The left branches of the portal vein and hepatic artery were clamped to induce complete ischemia of the median and left lobes of the liver. The right lobes remained perfused to prevent venous congestion of the intestine. After 60 min of ischemia, the clamp was removed to allow reperfusion. Sham-operated mice were prepared similarly; however, a clip was not placed on the vasculature leading to the median and left lobes.

Histopathology
The severity of IRI was assessed by hematoxylin and eosin (H&E) staining of liver paraffin sections.
Liver specimens were fixed in formalin and then embedded in paraffin. The specimens were sectioned at 5μm and stained with H&E. Sections were scored from 0 to 4 as described before [35]. The Suzuki scores were assessed by a histopathologist blinded to the sample group. Statistical analysis was performed with Prism 6 (GraphPad). Data were presented as mean ± SD. The means for the data sets were compared using unpaired student t-tests with equal variances. P values less than 0.05 were considered significant.

Macrophages/monocytes were specifically targeted by Fasudil carried liposomes
To realize the macrophage/monocyte specific targeting, we packaged Fasudil ( Figure 1A) together with fluorescence indicator AF488 into lipid bilayer liposomes, which has been widely used in the drug delivery for macrophage/monocyte [23,26]. Size distribution measurements revealed that diameters of the produced liposomes are mainly around 300-350 nm ( Figure 1B). Fluorescence imaging confirmed the liposome size and the loading of AF488 in liposomes, which indicated the loading of Fasudil in these liposomes ( Figure 1C).
To test the targeting efficiency and specificity of liposomes, we used flow cytometry assay after the liposome administration. As the schematic showed in Figure 1D, when liposomes are phagocytosed by targeted cells, the fluorescence indicator AF488, which is resistant to acidity, will be carried. The targeted cells will display green fluorescence (AF488 + , Q1 and Q2 in Figure 1E and F), which can be detected by flow cytometry. From flow cytometry dot plots of intrahepatic leukocytes (CD45 + , Figure 1E), we found out that one day after the three administrations of the Fasudil/AF488 carried liposomes, most AF488 + cells were Kupffer cells (F4/80 + , 86.61 ± 2.68%, Figure 1G) and most Kupffer cells were targeted (AF488 + , 86.44 ± 6.13 %, Figure 1H). Similarly, from flow cytometry dot plots of peripheral blood leukocytes (CD45 + , Figure 1F), most AF488 + cells were monocytes (CD115 + , 80.85 ± 3.38 %, Figure 1G) and most monocytes were targeted (AF488 + , 90.95 ± 3.40 %, Figure 1H). These results demonstrated the good efficiency and specificity of the macrophage/monocyte specific drug delivery by the Fasudil/AF488 carried liposomes.

Suppressed expression of ROCK-II in macrophage/monocytes
Besides evidence that AF488 in the Fasudil/AF488 carried liposomes were specifically delivered to Kupffer cells and blood monocytes, we performed experiments to confirm the Fasudil delivery, which suppresses the expression of ROCK-II in Kupffer cells (

Fasudil carried liposomes protect liver from IRI
To assess the protecting effect of macrophage/monocyte targeting Fasudil liposomes to liver IRI, we tested ALT, AST, and TBIL concentration in serum, which reflect liver injury, in mice with surgical induction of liver IRI after the administration of liposomes or not ( Figure 3A-C). Comparing to mice with sham surgery, which did not show any increase of ALT, AST, and TBIL, the concentration of ALT, AST, and TBIL all increased after liver IRI in mice with Fasudil liposome administration or not. The administration of mock liposomes did not show any effect. In Fasudil liposome treated mice, the increase of ALT, AST, and TBIL were all significantly inhibited two hours after liver IRI. A peak concentration of less than 5,000 U/L ALT and AST was observed in Fasudil liposome treated mice, which is around three-fold less than the peak ALT and AST concentration (~15,000 and ~13,000 U/L, respectively) in liver IRI mice without liposome treatment. The TBIL concentrations continuously increased after IRI. The treatment of Fasudil liposomes significantly inhibited (about three quarters) the increase of serum TBIL.
Another indicator to assess the tissue damage is cfDNA [36]. After measured the cfDNA concentration in serum ( Figure 3D), we found out that in liver IRI mice, the cfDNA concentration in mice treated with Fasudil liposomes is similar to the mice without Fasudil liposome treatment in the first two hours after liver IRI. However, four hours after liver IRI, the cfDNA concentration in Fasudil liposome treat mice was decreased. A significant difference was observed in between mice treated with Fasudil liposomes or not. Sham mice do not show much cfDNA in serum. These results combined with above ALT results indicated that Fasudil liposome treatment reduced the liver tissue damage in the liver IRI mouse model. Neutrophils are vital for inflammation response [37,38] and play critical roles in IRI. During IRI, blood neutrophilia and neutrophil tissue infiltration are frequently observed [39]. Thus, we assessed the dynamics of neutrophil (Ly6G + ) proportion in the blood ( Figure 3E) and the neutrophil (Ly6G + ) infiltration in liver ( Figure 3F) after IRI. We found out that the administration of Fasudil liposomes can prevent neutrophilia after IRI ( Figure 3E). Specifically, a severe neutrophilia was observed in liver IRI induced mice, which neutrophils continuously increased to ~80% of blood leukocytes until 24 hours after IRI. In comparison, the neutrophil percentage in mice treated with Fasudil liposomes showed an inhibited increase in the first eight hours after liver IRI and a decrease after then. The significant difference between these two groups was observed starting at four hours after liver IRI. Sham mice did not show obvious changes in neutrophil percentage. In liver ( Figure 3F), 48 hours after IRI, neutrophil infiltration in Fasudil liposome treat mice (43.68 ± 2.813 % amount all CD45 + intrahepatic leukocytes) was significantly less than that in liver IRI mice without the treatment of Fasudil liposomes (70.87 ± 6.754 %).
We also assessed the severity of IRI by performing H&E histology ( Figure 3G and H). We found that the administration of Fasudil liposomes reduced the Suzuki scores from more than 3 to about 2 in IRI mice ( Figure 3H). Overall, the inflammation after liver IRI, which is indicated by neutrophils, is inhibited by the administration of Fasudil carried liposomes. The release of inflammation cause less hepatic tissue damage and protect the liver from IRI.

Fasudil carried liposomes polarized macrophages/monocytes from M1 to M2 phenotype.
Macrophages/monocytes are key players in the modulation of immunological responses and inflammation [40,41]. The M1 or M2 macrophage phenotype is important for promoting inflammation or healing, respectively [6]. Liver IRI is a severe inflammation condition, which damages the hepatic tissue. To characterize the phenotype changes of macrophages/monocytes after IRI with or without Fasudil liposome administration, we measured the proportion of classical/inflammatory (Ly6C high ) and patrolling/regulatory (Ly6C low ) monocytes in blood leukocytes ( Figure  In blood leukocytes, we found that Fasudil liposome administration significantly reduced the percentage of Ly6C high monocytes from ~2.7% to ~1.6% ( Figure 4A) and significantly increased the percentage of Ly6C low monocytes from ~1.3% to ~2.3% ( Figure 4B). Amount the monocyte proportion, Fasudil liposome administration switched the inflammatory phenotype (67.5% Ly6C high ) to regulatory phenotype (58.4% Ly6C low , Figure 4C). Forty-eight hours after the induction of liver IRI, the percentage of Ly6C high monocytes increased about twice (~1.6% to ~3.0%) or more than twice (~2.7% to ~6.25%) in mice with or without Fasudil liposome treatment, respectively ( Figure 4A). The percentage of Ly6C low monocytes increased more than thrice (~2.3% to ~7.2%) in mice treated with Fasudil liposomes 48 hours after liver IRI. In comparison, 48 hours after liver IRI, the percentage of Ly6C low monocytes significantly but slightly increased (~1.3% to ~1.8%) in mice without Fasudil liposome treatment ( Figure 4B). Amount the monocyte proportion, liver IRI promoted the inflammatory phenotype (67.5% to 77.8% Ly6C high ) in mice without Fasudil liposome treatment but promoted the regulatory phenotype (58.4% to 70.1% Ly6C low ) in mice treated with Fasudil liposomes ( Figure 4C). Sham mice did not show any variation of the monocyte populations.

Fasudil carried liposomes lead to a shift in cytokine expression profiles.
To assess the cytokine expression after the treatment of Fasudil carried liposomes in liver IRI, we measured the amount of pro-inflammatory cytokines TNFα and IL-1β and anti-inflammatory cytokine IL-10 in the serum of IRI mice treated with liposomes with or without Fasudil. Sham mice treated with mock liposomes were used as controls. We found that the IRI of liver induced a predominant increase of TNFα ( Figure 5A) and IL-1β ( Figure 5B) compared to sham controls. The macrophage/monocyte targeting Fasudil treatment significantly inhibited the increase of serum TNFα and IL-1β in mice with the liver IRI. Mice with the liver IRI showed an increase of serum IL-10 ( Figure 5C) compared to sham controls. This increase is significantly boosted by the treatment of Fasudil-carrying liposomes in mice with the liver IRI. The treatment of control liposomes did not affect the cytokine amount in serum of mice with the liver IRI.
We also tested mRNA expression of these cytokines in sorted Kupffer cells (CD45 + F4/80 + ) using qRT-PCR. Similar to serum cytokines, we found out that the mRNA expressions of TNFα, IL-1β, and IL-10 in Kupffer cells were increased in mice with the liver IRI compared to sham controls ( Figure 5D-F).
The treatment of Fasudil-carrying liposomes resulted in a significantly lower mRNA expression of TNFα ( Figure 5D) and IL-1β ( Figure 5E) but a significantly higher mRNA expression of IL-10 ( Figure 5F) in Kupffer cells compared to those in liver IRI mice without treatments. The treatment of control liposomes did not affect the cytokine expression in Kupffer cells of mice with the liver IRI.

Discussion
In this study, we specifically targeted macrophages/monocytes by packaging Fasudil in liposomes.
Efficiency and specificity of the targeting were confirmed by the gain-of-fluorescence of macrophages/monocytes from the co-packaged AF488 indicator. The ROCK-II inhibition in macrophages/monocytes was demonstrated by reduced protein and mRNA expression. The polarization from M1 to M2 phenotype in blood monocytes and Kupffer cells was established by flow cytometry assay.
These shift from inflammatory to regulatory phenotype reduced the inflammatory responsesneutrophilia in blood and neutrophil infiltration in hepatic tissue, resulting in a protective effect of liver from IRI.
Many attempts have been made to protect the liver from IRI. Most of them are based on regulating the M1/M2 macrophage responses. IL-10, an important chemokine which can enhance M2 macrophage polarization, was demonstrated to protect the liver from IRI [42]. iNOS inhibitor, which can inhibit M1 macrophage response, was reported as having a curative effect in liver IRI [43]. Resolvin D1, which is an endogenous pro-resolving lipid mediator, was reported to polarizing macrophage to M2 phenotype and inhibit liver IRI [20]. ROCK inhibitors were showed a protective effect in liver IRI in the previous study [44,45]. But they thought that effect is hepatic stellate cell dependent. ROCK is a kinase belonging to the AGC (PKA/ PKG/PKC) family of serine-threonine kinases. ROCK is a key regulator of actin organization and thus a regulator of cell migration [46]. The function of ROCK in macrophage regulation is contradictory [19,47]. Our studies demonstrated that ROCK inhibitor could polarize macrophage to M2 phenotype and protect the liver from IRI. Because our study is the first to use a method specific targeting macrophages/monocytes, we may conclude that the previous studies, which showed decrease M2 macrophage after ROCK inhibitor treatment, might be an indirect effect caused by the systematic administration.

Availability of data and materials
All contributing authors agree and declare that the data and material is available by contacting the corresponding author L.L.

Competing interests
All contributing authors declare that they have no conflict of interest.

Authors' Contributions
Experiments were designed by Y.X. and L.L. Experiments were performed by Y.X., D.Z., and P.D. Data analysis was performed by Y.X. The manuscript was written by Y.X. and L.L. The project was supervised by L.L. All authors (Y.X., D.Z., P.D. and L.L.) discussed the results and revised the manuscript.