Lipopolysaccharide (LPS), Sirius red F3B, and saturated aqueous solution of picric acid were from Sigma Chemical, Co. Ltd. (St. Louis, MO, USA). Fetal bovine serum (FBS), trypsin, Dulbecco’s modified Eagle medium (DMEM), penicillin, and streptomycin were from Gibco (Carlsbad, CA, USA). Invivofectamine^® 2.0 reagent, in vivo predesigned PlGF siRNA and in vivo non-targeting control (NTC) siRNA were from Life Technologies (Carlsbad, CA, USA). siRNA sequences are provided in the supporting information (Figure S1 in Supplementary Material). Recombinant mouse PlGF-2 protein was from R&D Systems Inc. (Minneapolis, MN, USA).

Male BALB/c mice (8–10 weeks) were purchased from Shanghai Laboratory Animal Research Center (Shanghai, China). The experimental protocol was performed in accordance with the guiding principles for the care and use of laboratory animals approved by the Fudan University Animal Care Committee and all animals received humane care. The animals were kept in an environmentally controlled room (23 ± 2°C, 55 ± 10% humidity) with a 12-h light/dark cycle and allowed free access to food and water. Mice were subjected to bile duct ligation (BDL) to induce liver fibrosis, while controls were sham-operated (SHAM) (27, 28). Mice were randomly distributed in four groups as shown in experimental design (Figure 1A). To deliver each siRNA, in vivo ready siRNAs were mixed with Invivofectamine 2.0 regents and injected in a volume of 100 µl at a dose of 5 mg/kg for three cycles starting 2 days after BDL surgery. Six to ten mice of each group were sacrificed on days 14, 21, and 28 after BDL, respectively; and the livers were removed and cut into small pieces and either snap-frozen in liquid nitrogen for storage at −80°C or fixed in freshly prepared 4% paraformaldehyde for 24 h at 4°C. Mouse sera were isolated to assay for liver functions.

RAW 264.7 murine cells (Sigma, St. Louis, MO, USA) were grown in 150 cm² flasks in DMEM supplemented with 10% FBS, 2 mM l-glutamine, 50 U/ml penicillin, and 50 µg/ml streptomycin. All incubations were performed in cells under the three or four passages. In experiments to assess the effects of PlGF on cells function, cells were transferred to 6-well plates at a density of 2.5 × 10⁵ cells/well in serum-free medium under a humidified 5% CO₂ atmosphere at 37°C for 24 h. Then cells were washed and incubated with PBS vehicle, LPS (100 ng/ml), and recombinant mouse PlGF (rPlGF) at 50 ng/ml for 24 h at 37°C, respectively. To block the PlGF/VEGFR1 signaling, neutralizing antibody against mouse VEGFR1 (R&D Systems Inc., Minneapolis, MN, USA; 10 µg/ml) was applied where indicates. The cells were harvested for immunofluorescence analysis, RNA harvesting, and protein isolation. Staining and quantitative RT-PCR analysis were performed on three independent experiments. All measurements were performed in triplicate wells.

To test migration, cells were investigated using a modified Boyden chamber assay (23, 28). Briefly, RAW 264.7 cells (5 × 10⁴ cells/well) were added to the upper chamber in DMEM without serum and exposed to rPlGF (25, 50, and 100 ng/ml) or LPS or PBS vehicle in the lower chamber. After 24 h of incubation at 37°C, cells on the upper membrane surface were removed and migratory cells on the membrane underside were fixed using 4% paraformaldehyde and stained using Crystal Violet Staining Solution (Beyotime Institute of Biotechnology, Nantong, China). Filter inserts were inverted and the number of migratory cells on the membrane underside was counted manually. The cells’ migration ability was expressed as the average cell number in eight randomly chosen fields at 200× (Olympus BX45, Olympus Corporation, Tokyo, Japan).

Serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) concentrations were determined spectrophotometrically using an automatic biochemical analyzer (Beckman, Fullerton, CA, USA). Hydroxyproline was measured in liver tissue hydrolyzates using the Hydroxyproline Assay Kit (BioVision, CA, USA) according to the manufacturer’s instructions; and the results are expressed as a microgram of hydroxyproline per gram of liver tissue.

In all experiments, the left liver lobe was excised and fixed with 10% neutral-buffered formalin, embedded in paraffin, and cut into 5-µm thick sections for histological analysis or IHC. Liver sections were stained with H&E and Sirius red according to standard procedures. Portal inflammation was graded with a 0–3 scale as described previously (22, 27). Fibrosis was quantified using ImageJ software 1.49 (NIH, Bethesda, MD, USA) on 10 non-contiguous Sirius red-stained sections and by the Scheuer modified histological activity index scoring system (5, 22, 27). A liver pathologist without knowledge of the treatment group examined histology. Protocol for IHC is described in detail in the supplementary experimental procedures (Figure S1 in Supplementary Material).

The dissected liver tissues retrieved were fixed in 4% paraformaldehyde solution for 30 min, washed with PBS (pH 7.4), embedded in optimum cutting temperature tissue compound (OCT compound, Sakura, Japan), and frozen at −80°C for 1 day. Then the sections (10 µm in thickness) were cut with a cryotome Cryostat (Leica, CM 1900, Germany) and placed on slides for immunofluorescence staining. Blocking was performed in PBS with 3% BSA. The slides were incubated with antibody von Willebrandfactor (vWF) (Dako North American, Inc., Carpentaria, CA, USA), F4/80, α-SMA, or CD31 (Abcam, Cambridge, CA, USA) at the dilution of 1:100 overnight at 4°C and, subsequently, incubated with antibody PlGF or VEGFR1 (Abcam, Cambridge, CA, USA) at the dilution of 1:200 for 1 h at room temperature (RT) in case of double-staining. Alexa Fluor 594 donkey anti-mouse and Alexa Fluor 488 Donkey anti-Rabbit secondary antibodies (Yeasen Biotech, Shanghai, China) were incubated at 1:200 in PBS for 1 h at RT. After washing with Tris-buffered saline for three times, the cell nuclei were counterstained with Dapi-Fluoromount-G™ (SouthernBiotech, Birmingham, AL, USA). Finally, the stained tissues were analyzed by fluorescence microscopy (BX51, Olympus, Japan).

RAW 264.7 cells plated on 24-well plates and cultured on cover glass slips were fixed and permeabilized for 10 min in 4% paraformaldehyde, 0.2% TritonX-100 in PBS. Non-specific binding was blocked with 3% BSA for 1 h at RT, and then the cells were incubated with primary antibodies for F4/80 (dilution 1:200) and VEGFR1 (dilution 1:100) overnight at 4°C. After washing twice in PBS, the cells were incubated with fluorescein-labeled secondary antibody for 1 h at RT in the dark. The nuclei were stained with DAPI in the dark for 40 min at RT. The slides were washed twice with PBS, covered with DABCO (Sigma-Aldrich, St. Louis, MO, USA), and imaged by fluorescence microscopy (IX51, Olympus, Japan).

The number of α-SMA-, Desmin-positive cells, and the intensity of collagen III immunostaining in tissue sections were quantified using five random non-overlapping fields (100×) of each slide and determined for six animals in each group, and the area of staining was calculated as a percentage of the total area using the software NIH ImageJ 1.49 as described previously (5, 22).

For quantification of the numbers of hepatic macrophages in sections, six non-overlapping randomly selected fields of view per slide at 400× magnifications (F4/80⁺ cells) or 200× magnifications (CD68⁺ and Ly6C⁺ cells) were examined and expressed as cells per fields of view; and five mice of each group were examined (22, 29, 30).

Microvascular density in the liver tissue was assessed by determining the count of CD31-labeled ECs in five areas from each liver section at 200× magnification and is expressed as the number of CD31-positive vessels per field (22, 30). The vWF-positive cells were quantified using NIH ImageJ software; and five non-overlapping randomly selected fields of view per slide at 200× magnifications and eight mice of each group were examined (29, 30).

Liver samples were homogenized in RIPA lysis buffer by adding protease inhibitor Cocktail (Roche) and phosphatase inhibitors Cocktail (Sigma), and then centrifuged at 10,000 g at 4°C for 20 min. Protein extraction from macrophage cells was as previously described (22). The protein concentration was measured using the Bicinchoninic Acid Protein Colorimetric Assay kits (BMI, Shanghai, China) with BSA as standard. Equivalent aliquots of protein samples (40 µg) were separated by electrophoresis on 7.5–12% SDS-PAGE gels and transferred onto polyvinylidenedifluoride membranes. The membrane was then incubated in blocking buffer (5% non-fat milk powder in TBST) for 3 h followed by incubation with primary antibody in TBST overnight at 4°C with the specific primary antibodies against PlGF, α-SMA, VEGFR1, CD31, TNF-α, IL-1β, TLR4, TLR9, HIF-1α, MCP-1, VCAM-1, ICAM-1 (all from Abcam, Cambridge, CA, USA), and CXCL10 (R&D Systems Inc., Minneapolis, MN, USA) at 1:1,000 dilution. The membrane was washed with TBST and then incubated with goat anti-rabbit, anti-mouse, or anti-rat secondary antibodies (Biotech Well, Shanghai, China; 1:1,500 dilution) for 2 h at RT. GAPDH or β-actin (Cell Signaling Technology, Boston, MA, USA; 1:5,000 dilution) was used as internal control, respectively. After washing off the unbound antibody with TBST, the expression of the antibody-linked protein was determined by an ECL™ Western Blotting Detection Reagents (Amersham Pharmacia Biotech Inc., Piscataway, NJ, USA). The densitometric analysis was performed with ImageJ.

Total RNA was extracted from frozen liver tissues (caudate lobe) and cultured cells using Trizol reagent (Life Technologies, Grand Island, NY, USA) following manufacturer’s protocol. RNA was extracted reverse-transcribed with random hexamers and avian myeloblastosis virus reverse transcriptase using a commercial kit (Perfect Real Time, SYBR^® PrimeScriP™TaKaRa, Japan). Quantitative RT-PCR was performed for assessment of mRNA expression on an ABI Prism 7500 Sequence Detection system (Applied Biosystems, Tokyo, Japan) according to the manufacturer’s protocol. Probes and primers for target genes were purchased from Sangon Biotech Co., Ltd. (Shanghai, China, Table S1 in Supplementary Material). SYBR green gene expression assays were used to quantify target genes. The relative changes normalized to GAPDH mRNA using the formula 2^−ΔΔCt, where ΔΔCt represents ΔCt values normalized with the mean ΔCt of control samples.

Data are expressed as mean ± SD. Statistical analyses were performed by using Graphpad Prism7 software (La Jolla, CA, USA). Comparisons between two independent groups were performed using a two-sample t-test. Comparisons between multiple groups were performed by one-way analysis of variance with post hoc Tukey’s multiple comparison tests or by two-tailed unpaired Student’s t-tests. A P value less than 0.05 was considered statistical significance.

To examine the role of PlGF in chronic liver injury and fibrosis, we have used a well-established animal model of liver fibrosis induced by BDL (Figure 1A). As shown in Figure 1B, IHC staining revealed that PlGF expression was undetectable in SHAM mice and dramatically increased in non-parenchymal cells of the fibrotic liver as fibrosis progression and minimally in hepatocytes, particularly remarkable at the portal tracts and fibrous septa at 28 days of BDL. Notably, in livers of BDL mice, PlGF immunofluorescence co-localized with α-SMA and in cells located hepatic sinusoids, suggesting that PlGF expression is upregulated in profibrotic myofibroblasts; and we also noted that PlGF slightly expressed in macrophages (F4/80⁺) and in hepatic sinusoidal EC in sinusoids (Figure 1C).

To identify the role of PlGF in liver inflammation and fibrosis, we silenced PlGF in vivo using a chemically synthesized short, double-stranded RNA, which having well-defined structure with a phosphorylated 5′ end and hydroxylated 3′ ends with two overhanging siRNA to target hepatic PlGF expression. After 2 days of BDL, mice were injected with PlGF-specific or NTC siRNA by Invivofectamine reagent (Figure 1A). Efficiency of knockdown of PlGF in vivo by using siRNA was assessed by IHC and Western blotting; and the IHC staining signal of PlGF was obviously weak in fibrotic livers from PlGF siRNA-treated mice when compared with the livers from control and NTC siRNA-treated mice (Figure 1B). Consistent with our histological finding, Western blot results confirmed that targeted siRNA treatment resulted in a significant decrease in intrahepatic PlGF expression at different stages of disease progression (days 14, 21, and 28 after BDL) (Figure 1D). Moreover, to further ascertain the effect of siRNA-mediated suppression of PlGF expression in vivo, we also analyzed liver PlGF mRNA levels at 14, 21, and 28 days after BDL. Our results demonstrated that the levels of PlGF mRNA expression were gradually increased following BDL; which were significantly downregulated at their corresponding time points by PlGF siRNA administration (Figure 1E).

Morphological analysis by H&E staining of liver sections from BDL mice revealed distortion of the normal architecture, with a marked aggregation of lymphocytes, severe hepatocytes necrosis, and proliferation of bile ductules. Mice presented with remarkable fibrosis (stage 3 or 4) showing the characteristic pattern of extensive portal–portal and portal–central fibrosis linkage following 4 weeks of BDL; whereas the SHAM mice shown normal architecture (Figures 2A,B). Impressively, however, PlGF silencing in BDL mice exhibited thinner septa, mild liver fibrosis, and more preserved hepatic parenchyma (Figures 2A,B). Moreover, in BDL mice, PlGF silencing decreased the severity of hepatic inflammation compared with those of NTC siRNA-treated group (1.60 ± 0.52 vs. 1.00 ± 0.47, respectively, P = 0.021; Figure 2C). This was indeed also supported by the findings that serum ALT and AST levels were decreased in BDL mice receiving PlGF siRNA (Figure 2D).

As revealed by the histological analysis of liver sections, there was a lower mean fibrosis score in BDL mice receiving PlGF siRNA treatment compared with those of the mice receiving NTC siRNA treatment (2.2 ± 0.8 vs. 3.2 ± 0.8, P = 0.011; Figure 2E). This was further confirmed by Sirius red-stained area analysis and hepatic hydroxyproline content, showing that PlGF silencing to BDL mice resulted in a 53.2% reduction in Sirius red-stained area (5.64 ± 1.21% vs. 12.05 ± 2.44%, P < 0.0001; Figure 2F) and a 49.5% reduction in hepatic hydroxyproline content (Figure 2G) compared to those animals receiving NTC siRNA. In addition, IHC evaluation showed that the deposition of collagen III was increased in the portal tracts, septa, and perisinusoidal spaces of the lobules in BDL mice, whereas PlGF siRNA treatment attenuated collagen III accumulation in livers (Figure 2H). These findings were supported by quantification of collagen III positive areas showing a decrease in the areas by 46.5% in fibrotic liver sections from PlGF siRNA-treated mice compared those from NTC siRNA-treated animals (Figure 2I). We also examined the gene expression of collagen 1α1 and collagen 3α1, suggesting the levels of both genes were significantly downregulated by siRNA-mediated PlGF knockdown in BDL mice (Figure 2J).

Taken together, these results suggested that PlGF silencing led to a significant reduction in BDL-induced liver inflammation and fibrogenesis in mice.

Activated HSCs are considered central ECM-producing cells within the injured liver and also involved in the pathologic angiogenesis and vascular remodeling (22, 23). We found that PlGF silencing induced a significantly reduction in BDL-induced expression of the marker of activated HSCs, α-SMA, and Desmin as shown by IHC analysis (Figure 3A). Moreover, computer-assisted semiquantitative analysis demonstrated that the number of α-SMA- and Desmin-positive cells was significantly lower in livers from PlGF siRNA-treated BDL mice than those from NTC siRNA-treated BDL mice (Figure 3B). These findings were substantiated by quantitative RT-PCR experiments, suggesting the levels of α-SMA and Desmin mRNA transcript in fibrotic livers were correspondingly reduced following PlGF knockdown (Figure 3C). In addition, we also examined α-SMA protein expression by western blotting, indicating PlGF silencing with siRNA inhibited the α-SMA protein expression in vivo (Figure 3D). Collectively, these in vivo data indicated that PlGF silencing efficiently inhibited myofibroblastic activation of HSC during BDL-induced liver injury and fibrosis.

To investigate the vascular changes in PlGF silencing, we conducted studies to determine hepatic neovascularization and the expression of angiogenic factors in livers. The EC marker CD31 was expressed in the endothelium of the veins and in the central veins in livers of sham-operated mice, but not along the sinusoids; and challenge mice by BDL at 28 days led to a markedly increased number of CD31-positive vessels in livers (Figure 4A). However, CD31-positive EC staining in BDL siPlGF liver sections was significantly less than in the BDL siNTC group as evidenced in mean microvessels density (67.9 ± 8.1 vs. 41.1 ± 6.4/per field, P < 0.001) (Figure 4B). These results were further supported by the levels of CD31 mRNA and protein expression in livers; showing that hepatic angiogenesis was inhibited by PlGF silencing in fibrotic mice (Figures 4E,F). Similar histologic pattern was observed in vWF staining of tissues, indicating upregulated in livers from BDL mice; however, BDL mice receiving PlGF-siRNA exhibited decrease in the intensity of vWF staining and vWF-positive vessels area in livers (Figures 4C,D), consistent with the decrease in vWF gene expression (Figure 4F). In addition, we noted a significant decrease in expression of HIF-1α in fibrotic animals treated with PlGF siRNA as shown by IHC (Figure 4G) and Western blot analysis for HIF-1α confirmed the morphological changes observed (Figure S1 in Supplementary Material). Similarly, hepatic HIF-1α mRNA levels in BDL mice were also significantly reduced by PlGF silencing with siRNA in vivo (Figure 4H). Together, these results indicated that PlGF silencing effectively attenuates pathologic vascular changes that occur in response to BDL.

To investigate the mechanism of PlGF silencing in the pathogenesis of hepatic inflammation and fibrosis-associated angiogenesis, we explored the markers of monocytes and macrophages infiltration in fibrotic livers. Compared to the SHAM mice, IHC staining for the macrophage markers indeed revealed BDL-enhanced infiltration of F4/80⁺ or CD68⁺ macrophages into fibrotic livers. Remarkably, however, the increase of hepatic macrophages infiltration was significantly reduced by PlGF siRNA treatment in BDL mice when compared with those by NTC siRNA treatment mice (Figure 5A); and the results were further confirmed by quantification of the F4/80⁺ or CD68⁺ staining cells (Figure 5B). Moreover, these results consistent with the genes expression of F4/80 and CD68, demonstrating PlGF silencing in BDL mice strikingly decreased the upregulated F4/80 and CD68 mRNA levels (Figure 5C). In addition, the total number of Ly6C⁺ cells, a marker for BM-derived circulating peripheral blood monocytes, was also significantly higher in fibrotic livers in BDL mice than normal livers from the SHAM mice. However, PlGF silencing led to reduced Ly6C⁺ macrophage infiltration (Figures 5A,B). Similarly, hepatic Ly6C mRNA expression in BDL mice was also inhibited by PlGF silencing with siRNA in vivo (Figure 5C). Taken together, these results suggested that siRNA-mediated PlGF knockdown significantly reduced hepatic macrophage recruitment to the livers, which being responsible for attenuating liver inflammation and fibrosis.

To investigate whether PlGF mediates liver inflammation through switching macrophages subtypes and regulating their function, we examined the expression of pro-inflammatory cytokines associated with M1 macrophages in the liver of fibrotic mice, such as TNF-α, IL-1β, and MCP-1. Our results demonstrated that hepatic expression of TNF-α, IL-1β, and MCP-1 mRNA was strongly upregulated in BDL fibrosis models. However, BDL mice receiving siPlGF reduced TNF-α, IL-1β, and MCP-1 mRNA by 6.1-fold, 7.0-fold, and 3.5-fold, respectively (Figure 6A). Those findings were supported by our Western blot analysis, demonstrating that the increase of these chemokines in fibrotic liver was indeed attenuated by PlGF silencing (Figure 6B).

In addition, we also examined the expression of TLR4 and TLR9 in livers. IHC data showed that weak constitutive expressions of TLR4 or TLR9 on sinusoidal ECs of SHAM mice livers, with hepatocytes showing no or only slight expression (Figure 6C). After 4 weeks BDL, increased TLR4 or TLR9 expression in livers was markedly observed in the periportal and interlobular septa, as well as increased expression on interstitial space between hepatocytes. However, giving PlGF siRNA to BDL mice resulted in moderate staining for TLR4 and TLR9 (Figure 6C). Consistent with these results, the expression of TLR4 and TLR9 protein was obviously upregulated in livers of BDL mice; however, siPlGF treatment to BDL mice decreased hepatic TLR4 and TLR9 expression when compared with vehicle treatment (Figure 6D). Similar results were seen in TLR4 and TLR9 mRNA expression, indicating PlGF silencing markedly reduced both gene expression (Figure 6E).

To further understand the link between PlGF knockdown and the reduction in inflammatory infiltrate, the expression of pro-inflammatory adhesive molecules, such as CXCL10, ICAM-1, and VCAM-1 in the vasculature of fibrotic mice was also analyzed. We found that the levels of CXCL10, VCAM-1, and ICAM-1 mRNA expression in livers were markedly enhanced in BDL mice received siNTC compared with SHAM mice, but these increase in livers were attenuated by PlGF siRNA treatment to BDL mice (Figure 7A). Meanwhile, those finding were supported by our western blotting, demonstrating that the increase of these chemokines in BDL-induced fibrotic livers was indeed attenuated by PlGF silencing (Figure 7B).

Taken together, these results suggest that PlGF might play a key role in the activation of Kuffer cells/macrophages in liver upon chronic injury and substantially produce a variety of pro-inflammatory cytokines and chemokines.

Placental growth factor exclusively binds to VEGFR1 and not VEGFR2 (17, 18), we, therefore, investigated whether VEGFR1 signaling in macrophages mediated the role of PlGF in liver inflammation and fibrosis in vivo. First, we investigated the cellular source of VEGFR1 in fibrotic livers; and our double staining of liver sections for VEGFR1 and CD31, F4/80, or α-SMA revealed obviously increased expression VEGFR1 in those non-parenchymal cells in mice of BDL, whereas there has weak expression in those cells in livers from SHAM mice (Figure 8A). Next, we further examined levels of VEGFR1 in fibrotic livers at 4 weeks of BDL mice by both quantitative RT-PCR and Western blot analysis, respectively. Our results showed that a marked increase in VEGFR1 mRNA and protein expression was demonstrated with the development of hepatic fibrosis in BDL mice compared with SHAM control (Figures 8B,C). However, PlGF silencing significantly downregulated the expression of VEGFR1 at gene levels and at protein levels in fibrotic livers when compared to NTC siRNA-treated fibrotic mice (Figures 8B,C).

To further confirm whether VEGFR1 expression on macrophages was involved in macrophages recruitment or activation upon liver injury, we tested the migratory response and inflammatory properties of mouse macrophages RAW 264.7 cell line using an in vitro transmigration assay with recombinant mouse PlGF (rPlGF). First, we evaluated the effect of PlGF on VEGFR1 expression from macrophages, we selected LPS as positive control since LPS is known to activate macrophages in vitro. Indeed, rPlGF stimulated VEGFR1 expression in RAW 264.7 cells as shown by double immunofluorescent staining (Figure 8D) and Western blotting (Figure 8E), consistent with increased expression of VGFR1 in livers from BDL mice. Similarly, the levels of VEGFR1 mRNA expression were also increased by rPlGF in a similar manner in LPS-treated (Figure 8F). Second, RAW 264.7 cells were treated for 24 h with rPlGF (50 ng/ml) and there was 6.98-fold increase in RAW 264.7 cell migration toward rPlGF in Boyden assays compared to cells migrating toward vehicle (Figures 9A,B). The 24-h incubation time and 50 ng/ml contents of PlGF were chosen on the basis of the results of a pilot studies (Figure S2 in Supplementary Material). Moreover, to determine whether PlGF could activate macrophages to generate cytokines, MCP-1, TNF-α, and IL-1β, we utilized an in vitro assay to mimic this situation. rPlGF treatment of RAW 264.7 cells for 24 h showed that the expression of MCP-1, TNF-α, and IL-1β mRNA were obviously increased as shown in our quantitative RT-PCR results (Figure 9C). Finally, to further confirm the role of PlGF in the migration and activation of macrophages, we blocked PlGF/VEGFR1 signaling axis with a specific VEGFR1 neutralizing antibody that was added to cultured cells. We found that the upregulation of inflammatory cytokines (MCP-1, TNF-α, and IL-1β) in macrophages at rPlGF challenge was reduced by VEGFR1 neutralizing antibody (Figure 9C), suggesting that PlGF/VEGFR1 signaling axis was strongly involved in activation of macrophages. Notably, the migratory capacity of macrophages was also significantly inhibited (52.2% reduction in mean cells number) while PlGF/VEGFR1 signaling was blocked (Figures 9A,B). Together, these results show clearly that PlGF promotes macrophage recruitment and activation upon liver injury via VEGFR1.