Decreased ferroportin in hepatocytes promotes macrophages polarize towards an M2-like phenotype and liver fibrosis

Iron release from macrophages is closely regulated by the interaction of hepcidin, a peptide hormone produced by hepatocytes, with the macrophage iron exporter ferroportin (FPN1). However, the functions of FPN1 in hepatocyte secretion and macrophage polarization remain unknown. CD68 immunohistochemical staining and double immunofluorescence staining for F4/80 and Ki67 in transgenic mouse livers showed that the number of macrophages in FPN1−/+ and FPN1−/− mouse livers was significantly increased compared to that in WT (FPN+/+) mice. FPN1 downregulation in hepatic cells increased the levels of the M2 markers CD206, TGF- β, VEGF, MMP-9, Laminin, Collagen, IL-4 and IL-10. Furthermore, the expression of CD16/32 and iNOS, as M1 markers, exhibited the opposite trend. Meanwhile, α-SMA immunohistochemistry and Sirius red staining showed that the trend of liver fibrosis in FPN1−/− mice was more significant than that in control mice. Similarly, in vitro FPN1 knockdown in L02-Sh/L02-SCR liver cell lines yielded similar results. Taken together, we demonstrated that downregulated FPN1 expression in hepatocytes can promote the proliferation and polarization of macrophages, leading to hepatic fibrosis. Above all, the FPN1 axis might provide a potential target for hepatic fibrosis.


FPN1 downregulation in hepatic cells induced hepatic fibrosis.
Macrophage polarization is closely related to fibrosis 15 . Immunohistochemical analysis showed that compared to that in the FPN +/+ group, the expression of Collagen 1, Collagen 4 and matrix metalloproteinase (MMP-9) in FPN1 −/− and FPN1 −/+ mouse livers were enhanced, indicating that the extracellular matrix of liver tissues increased. Furthermore, additional markers of fibrosis were tested. As expected, the levels of Laminin and α-SMA were increased. Extracellular fibrotic deposition, detected by Sirius red staining and Masson staining, further confirmed that the inhibition of FPN1 impaired hepatic fibrosis in vivo (Fig. 3A,B). Glutamic-pyruvic transaminase (ALT) and glutamic oxalacetic transaminase (AST) were regards as important index for liver fibrosis or injury. A lots of ALT or AST are released into the blood and induced increased expression of ALT or AST in serum when there are hepatocyte necrosis, liver fibrosis or injury. As is shown in Fig. 3C,D further confirm that FPN1 downregulation in hepatic cells induced hepatic fibrosis. FPN1 regulated macrophage polarization to the M2 phenotype by inducing IL-10 and TGF-β expression in vivo. As mentioned above, we found that the downregulation of FPN1 expression in mouse liver cells promoted macrophage M2 polarization and fibrosis. To investigate which cytokines were involved in this process, Q-PCR and ELISA were used. The Q-PCR results demonstrated that compared to that in FPN +/+ or FPN1 −/+ mice, the expression of IL-1β, IL-4, IL-10, IL-12β, TGF-α, TGF-β, HGF, VEGF, and IFN-γ was increased in FPN1 −/− hepatocytes (Fig. 4A,B). Though IL-6, IL-12α, TNF-α and CSF-1 showed no significant differences in FPN1 −/− and FPN1 −/+ hepatocytes (Supplementary Fig. S1-4), the expression of IL-10 or TGF-β in FPN1 −/− hepatocytes showed 5.5-fold or 4.2-fold higher than that in FPN +/+ hepatocytes. Their ratio was higher than those of IFN-γ and IL-12β in FPN1 −/− hepatocytes again, which were 3.5-fold and 1.6-fold higher than that in FPN +/+ hepatocytes, suggesting that FPN1 significantly regulated macrophage polarization to the M2 phenotype, compared with M1 phenotype. Moreover, the ELISA results indicated that the levels of IL-10, TGF-β, and IL-4 were significantly increased in FPN1 −/− and FPN1 −/+ mice (Fig. 4C). These cytokines are strong stimulators of M2 polarization and tissue remodeling and repair 16 . These data suggested that FPN1 induced macrophage polarization and fibrosis through IL-10 and TGF-β.

FPN1 was involved in macrophage polarization in vitro.
In order to further confirm the role of FPN1 in macrophage polarization, L02 cells were used in vitro to further study FPN1 knockdown using shorthairpin (Sh) or scrambled (Scr) RNA to interfere with endogenous FPN1 (Fig. 5A,B), this was supported by increase of iron in L02-Sh (Fig. 5C). Interestingly, compared with TPH1 + PMA + L02-SCR group, it is showed for the M2 polarization in TPH1 + PMA + L02-Sh group and TPH1 + PMA + IL-4 + IL-13 group (Fig. 5D), indicated decreased FPN1 promoted liver cancer cell M2 polarization. PMA treatment enhanced the adherence of the THP1 cells. Two morphological characteristics of THP-1 cells following differentiation is an increase in cytoplasmic volume and granularity ( Fig. 5D) 17 . M0 macrophages were cocultured with L02-SCR and L02-Sh cells. We detected M2 macrophages with a CD206 antibody. The in vitro results were consistent with the in vivo results. The expression of CD206 was significantly increased in M0 macrophages cocultured with L02-Sh cells (Fig. 5E), suggesting that macrophages exhibited M2 polarization.
Many reports have indicated that the expression of TNF-α and IL-6 are increased in M1 macrophages 18 and that IL-10 and TGF-β expression is induced by M2 polarization 19,20 . Thus, we detected the expression levels of these genes in different liver cells and macrophages. The Q-PCR results showed that many cytokines, such as IL-4, IL-10, TGF-α, TGF-β and VEGF, were successfully induced in L02-Sh cells compared with L02-SCR cells (Fig. 6A,B; Supplementary Fig. S5-8). Interestingly, when L02-Sh cells were cocultured with M0 macrophages, the increasing expression of IL-4, IL-10 and TGF-β were significantly showed in L02-Sh + M0, compared with L02-SCR + M0, on the contrary, these results were not showed in TNF-α and IFN-γ, indicated that the high expression of cytokines in the L02-Sh cells promoted M0 to M2 polarization (Fig. 6C), and this was further confirmed by ELISA analysis (Fig. 6D; Supplementary Fig. S9-11). This indicated decreased FPN1-induced M2 macrophage polarization.

Discussion
Macrophages are among the most versatile cells in the body 21 . Heterogeneity arises due to macrophage differentiation from monocyte precursors, and the phenotype of macrophages is determined by genetic modification as well as specific tissue-related and immune-related stimuli. Macrophages can be labeled by CD68 or F4/80. M1 macrophages are usually characterized by high expression of CD16/32 and proinflammatory cytokines, including IL-1β, IL-12 and TNF-α. M2 macrophages exhibit high expression of CD206 (mannose receptor) and antiinflammatory cytokines (IL-4, IL-10, TGF-β, and VEGF). M2 macrophages can secrete profibrotic factors that can promote the activation of myofibroblasts, the synthesis of extracellular matrix, and the development of fibrosis 22 .
The involvement of FPN1 in immune regulation is complex. In recent years, studies have shown that FPN1 is an anti-oncogene in breast cancer and myeloma, but not in the liver 23  The two main contributors to hepatic fibrosis are excessive fibroblast proliferation and matrix accumulation 22 . The onset of these processes is usually preceded by an acute inflammatory response. Alpha-smooth muscle actin (α-SMA) is a marker of the activation of hepatic stellate cells 16 . Additionally, the activation of hepatic stellate cells can promote transformation into muscle fibrosis cells and ultimately promote the formation of liver fibrosis. TGF-β is a known mediator of fibrotic remodeling and matrix accumulation, and the secretion of TGF-β by fibroblasts can directly promote collagen synthesis and maturation 25 . This is consistent with the increase in collagen in FPN1 knockout mice. Vascular endothelial growth factor (VEGF) can promote the proliferation of vascular endothelial cells and is also the strongest angiogenic factor. The increased expression of vascular endothelial growth factor in the livers of transgenic mice may be due to hypoxia in local pathological tissues, which stimulates the activation of hepatic stellate cells and then produces a variety of cytokines, resulting in increased expression of vascular endothelial growth factor. Additionally, VEGF can induce endothelial cells to express plasminogen activator and matrix collagenase and then promote peripheral vascular growth 16 . The significant correlation between the levels of VEGF and TGF-β in mouse livers is mainly due to the direct activation of vascular endothelial growth factor when the expression of TGF-β increases in vivo 26 . Hepatocyte growth factor (HGF) is a multifunctional antifibrotic factor involved in kidney development, acute injury and regeneration 27,28 . Increased expression of HGF and MMP-9 can inhibit ECM production 29 . FPN1 downregulation in hepatic cells increases the expression of Collagen, Laminin, α-SMA, TGF-β and VEGF, and induces a state of immune tolerance in the livers of mice 30 . Handa P 's groups have shown that dietary iron overload of C57Bl/6 mice led to hepatic macrophage M1 activation and stimulated hepatic fibrogenesis 31 . However, there is a dynamic balance between M1 and M2 macrophages in the process of fibrosis. The main differences between dietary iron overload model and our hepatocyte specific FPN1 knockout mice model was iron content in macrophages depending on iron homeostasis in liver microenvironment, resulting in entirely different polarization.
IL-4, IL-10 and TGF-β can inhibit the expression of TNF-α, IL-6 and IL-12, leading to the development of a liver microenvironment that favors the M2 polarization of macrophages 32,33 . IL-1β can attenuate collagen deposition mediated by TGF-β, which indicates that it has a long-term antifibrotic effect on some tissues 34 . The increased expression of IL-1β may attenuate tissue remodeling mediated by TGF-β in the absence of FPN1. Though decreased FPN1 also induce the increase of IFN-γ and IL-12β, their magnitude of increase was lower than that in IL-10 and TGF-β. This indicated that M2 phenotype macrophages were mainly showed in FPN1 −/− mice. In addition, no significant different TNF-α and IFN-γ protein expression between L02-Sh and L02-SCR cells is www.nature.com/scientificreports/ not accordance with significant difference between FPN1 −/− mice and FPN1 +/+ mice, suggested that the protein of TNF-α and IFN-γ can easily degraded in vitro. Together, the elevated levels of IL-4, IL-10 and TGF-β observed in our FPN1 −/− animal models may be a response to an increase in M2-mediated remodeling processes that occur in the absence of FPN1. THP-1 can usually be induced by PMA to obtain macrophages 35 . Furthermore, IL-4 + IL-13 would transfer M0 to M2 (Supplementary Fig. S12). Interestingly, decreased FPN1 in L02 cells promoted M2 polarization of THP-1 cells as measured by increased CD206, indicating IL-4 and IL-13 cytokines play a key role again.
Previous studies on FPN1 were mainly related to iron 24 . In this study, we found that FPN1 is also closely related to macrophage proliferation, polarization and fibrosis. Reduced expression of FPN1 in hepatocytes promotes iron accumulation in hepatocytes. This result is not contradiction with previous studies, which have shown that iron-overloaded macrophages exhibit an M1 phenotype 36 , because elevated iron was showed in hepatocytes, or not macrophage in FPN1 −/− mice. Whether it acts macrophage would be further studied. Whether the decrease in FPN1 expression further leads to hepatocellular carcinoma after the formation of the M2 environment in the liver remains to be further studied. It will be very interesting to use conditional knockout mouse models to detect the physiological role of FPN1 in hepatocellular carcinoma in the future. www.nature.com/scientificreports/ In conclusion, the downregulation of FPN1 in hepatocytes is beneficial to the proliferation and polarization of macrophages to the M2-like phenotype in the liver, which may lead to fibrosis (Fig. 7). Potential drugs that can upregulate the expression of FPN1 may be new directions in the treatment of hepatic fibrosis.
Animals. FPN1 flox/flox conditional gene knockout mice were constructed by Nancy Andrews of the USA. Two LoxP sites were inserted into exon 6/7 of the SLC40A1 gene, and the mice were maintained on a 129/SvEvTac background. SLC40A1-LoxP transgenic mice (129S-Slc40a1tm2Nca/J) were purchased from Jackson Laboratories (USA), backcrossed on a C57BL/6 background and bred in-house. When they were bred with liver cellspecific promoter (Alb-Cre) mice, all of the offspring obtained exhibited hepatocyte-specific knockout. Mice with downregulation of the FPN1 gene were used as the model mice for this experiment. Unless indicated, agematched animals were used at 8-12 weeks of age. Homozygote (FPN1 −/− ), heterozygote (FPN1 −/+ ) and wild type (FPN1 +/+ ) mice were obtained by mating, and the genotypes were determined by agarose gel electrophoresis. All animals in this study were raised in SPF-level animal rooms and fed a standard diet. The housing of the animals and the experiment procedures were carried out in accordance with the Guide for the Care and Use of Laboratory Animals (United States National Institutes of Health) and were approved by the Ethical Committee for Care and Use of Laboratory Animals of Guangzhou Medical University. This study was carried out in compliance with the ARRIVE guidelines 2.0.
Real-time quantitative PCR. Total RNA was isolated from THP-1 cells, M0 macrophages, and M0 macrophages cocultured with L02-SCR or L02-Sh using TRIzol according to the manufacturer's instructions. One microgram of RNA was reverse transcribed into cDNA using a First Strand cDNA Synthesis Kit. cDNA was quantified using the Applied Biosystems Step-One Real-Time PCR system with a SYBR Green Real-time PCR Master Mix kit. The following primer sequences that were used are shown in Table 1 (It is showed in Supplementary information). GAPDH served as the housekeeping gene. www.nature.com/scientificreports/ incubated with HRP-conjugated secondary antibodies at room temperature for 1.5 h. Following each step, the membranes were washed five times with TBS-T for 5 min. Finally, the blots were developed using the enhanced chemiluminescence system. Immunohistochemical analysis. Immunohistochemical analysis was played according to the procedure reported by Hau-Wen Li 39 . Deparaffinization of the tissue sections was executed using a graded ethanol series, and then 0.3% hydrogen peroxide was used to block endogenous peroxidase activity. Antigen retrieval was achieved by placing the sections in 10 mmol/L citrate buffer (pH 6.0) and heating in an autoclave for 5 min. The sections were then rinsed with phosphate-buffered saline (PBS, pH 7.2) and incubated with 10% nonimmunized goat serum for 30 min at room temperature to inhibit nonspecific binding. The tissue sections were then incubated at 4 °C overnight with antibodies. After 24 h, the slides were incubated with horseradish peroxidase (HRP)-conjugated anti-mouse/rabbit secondary antibodies for 30 min at 25 °C. Then, staining with 3,3-diaminobenzidine (DAB) was carried out for 90 s at room temperature; hematoxylin staining was performed as a control.
Histology. Tissue from mice was fixed with 4% paraformaldehyde, embedded in paraffin, and sectioned  Statistical analysis. Statistical analysis was performed using GraphPad Prism software version 6.0. All data are expressed as the mean ± standard deviation (SD). Comparisons between two groups were performed using Student's t test, and three or more groups were evaluated for significance using one-way ANOVA combined with Bonferroni's post hoc test. A P value < 0.05 or 0.001 considered statistically significant.
Ethical approval. The housing of the animals and the experiment procedures were carried out in accordance with the Guide for the Care and Use of Laboratory Animals (United States National Institutes of Health) and were approved by the Ethical Committee for Care and Use of Laboratory Animals of Guangzhou Medical University. This article does not contain any studies with human participants performed by any of the authors.