3-Methyladenine Alleviates Experimental Subretinal Fibrosis by Inhibiting Macrophages and M2 Polarization Through the PI3K/Akt Pathway

Qiyu Bo,1,* Mengxi Shen,1,* Meichun Xiao,1 Jian Liang,1,2 Yuanqi Zhai,1,2 Hong Zhu,1–3 Mei Jiang,1,2 Fenghua Wang,1–5 Xueting Luo,1,2 and Xiaodong Sun1–5

Purpose: To explore the effects of 3-methyladenine (3-MA), a selective inhibitor of phosphatidylinositol-3- kinase (PI3K), on experimental subretinal fibrosis (SRF) in mice.
Methods: The SRF mouse model was established by 532 nm laser photocoagulation at each fundus of mice on day 0. 3-MA was administered every 2 days from day 0 to 35. Immunofluorescence of choroidal flat mounts was performed to evaluate the size of SRF area, local macrophages, and polarization, respectively. Besides, Western blot analysis was carried out to assess the expression levels of macrophage polarization-related genes, Arg-1, Ym-1, and transforming growth factor-b2 (TGF-b2). Co-culture and migration experiments were used to demonstrate the inhibitory effect of 3-MA on fibroblasts. The gene knockout and Western blot analysis were used to explore the signal pathways related to macrophage polarization.
Results: Compared with the control group, the 3-MA-treated group showed significantly less size of SRF area. 3-MA treatment reduced both circulating and local macrophages, and counteracted M2 polarization. Moreover, 3-MA inhibited fibroblast recruitment. Mechanistically, we proved that 3-MA inhibits macrophage M2 po- larization by suppressing PI3K/Akt signal pathway rather than the PI3K-autophagy-related signal pathway.
Conclusions: 3-MA exerts antifibrotic effects on experimental SRF by targeting circulating and local macro- phages and M2 polarization, through PI3K/Akt signal pathway. These results support the potential use of 3-MA as a new therapeutic modality for SRF associated with neovascular age-related macular degeneration.

Keywords: subretinal fibrosis, macrophage polarization, 3-MA, PI3K/Akt signal pathway, age-related macular degeneration


ge-ReLATed MAcULAR DegeNeRATIoN (AMD) is a leading cause of vision impairment in the elderly.1,2 Pathologically, choroidal neovascularization (CNV) is in- duced as a tissue repairing process by local vascular endo- thelial growth factor (VEGF), and accordingly, anti-VEGF reagents have been developed as a first-line therapy to treat neovascular AMD (nAMD) patients in the clinic.3,4 How- ever, a long-term follow-up has uncovered significant drawbacks associated with anti-VEGF therapy. Specifically,

subretinal fibrosis (SRF) may develop as a result of repeated anti-VEGF delivery, which potentially contributes to anti- VEGF resistance and visual acuity drop to the base level within 5 years.5–7 Besides, SRF has also been described in untreated or photodynamic therapy-treated AMD eyes.8,9 The occurrence of SRF predicts a poor visual prognosis; therefore, therapeutic strategies for the inhibition of SRF are desired for long-term AMD management.
Previous reports suggested a prominent role for macro- phages and macrophage polarization in CNV.10–12 Macro- phages were also found abundant in subretinal fibrotic

1Department of Ophthalmology, Shanghai General Hospital (Shanghai First People’s Hospital), Shanghai Jiao Tong University School of Medicine, Shanghai, China.
2Shanghai Key Laboratory of Ocular Fundus Diseases, Shanghai, China.
3Shanghai Engineering Center for Visual Science and Photomedicine, Shanghai, China.
4National Clinical Research Center for Eye Diseases, Shanghai, China.
5Shanghai Engineering Center for Precise Diagnosis and Treatment of Eye Diseases, Shanghai, China.
*These authors contributed equally to this work.



tissues of nAMD patients and disease models.13,14 Macro- phages could obtain an array of activation phenotypes de- pending on different signals and be roughly categorized into M1 and M2 types. M1 is inclined to proinflammatory and antiangiogenic properties, while M2 is prone to anti- inflammation and wound healing.15 It is reported that CNV membranes (CNVM) from patients with M2 macrophage infiltration showed more fibrosis than the CNVM from pa- tients with M1 macrophage infiltration.16 Thus, we consid- ered that modulating macrophage polarization and inhibiting M2 polarization might be a novel potential strategy to at- tenuate the fibrosis of nAMD.
3-Methyladenine (3-MA), a selective inhibitor of phosphatidylinositol-3-kinase (PI3K), is also well known as an autophagy inhibitor due to its inhibition role in class III PI3K.17 Recently, 3-MA was reported to diminish fibrosis by abolishing transforming growth factor-b (TGF-b), infil- tration of macrophages, and Wnt signaling pathways in re- nal fibrosis.18 Besides, 3-MA was also shown to inhibit atherosclerotic lesion by modifying macrophages and the immune microenvironment,19 suggesting the specific tar- geting effect of 3-MA on macrophages. In this study, we intend to investigate the efficacy of 3-MA in suppressing fibrosis in laser-induced SRF and its underlying mecha- nisms.

Male C57BL/6J mice, between 6 and 8 weeks of age, weighed 20 – 1 g (Laboratory Animal Center, Shanghai General Hospital, China), were raised at the SPF facility of Shanghai General Hospital under controlled environmental conditions (temperature 22°C – 2°C and relative humidity 60%–70%) and were maintained on a regular 12-h light/12- h dark cycle. All animal experiments conformed to the Statement of the Association for Research in Vision and Ophthalmology for the Use of Animals in Ophthalmic and Vision Research.

Mouse model establishment and drug administration
To generate SRF, we established the laser-induced CNV model using a previously described protocol20 and observed it for 35 days. Generally, we carried out laser photocoagulation (532 nm, 120 mW, 100 ms, and 50 mm) at each fundus using a coverslip as contact lens on day 0, and harvested eyes on day 7 and 35 for observation and investigation, respectively. 3- MA powder (M9281; Sigma-Aldrich) was dissolved in 65°C deionized water at a concentration of 30 mg$mL-1 and di- luted 10 times with 0.9% saline for use. Then mice in 3-MA- treated group received an intraperitoneal injection with 15 mg$kg-1 every 2 days from day 0 to 35 according to the half-life and dose described in the literature,21,22 and an equal volume of 0.9% saline was administrated with an intraperi- toneal injection to mice in the control group.

Choroidal flat-mount
and immunofluorescence staining
Choroidal flat-mount and immunofluorescence (IF) staining were performed on day 7 and 35 after laser. Mice

were perfused with cold 4% paraformaldehyde. The anterior segments were removed, and the remaining eye cup was cut 4–6 radial incisions to be flattened. Then the retinal pig- mentepithelium (RPE) cell-choroid complexes were blocked in 5% goat serum albumin with 0.3% Triton X-100 for 1 h at room temperature, and incubated with primary antibodies for FITC-labeled isolectin-B4 (IB4) (FL-1201; Vector Laboratories); Fibronectin (AB2033; Millipore); F4/80 (MCA497R; Abd Serotech); YM-1 (1404; Stem Cell Tech- nology); Arginine-1 (Arg-1, 9819S; Cell Signaling Tech- nology); and a-smooth muscle actin (a-SMA, ab5694; Abcam) at 4°C overnight. The next day, RPE-choroid complexes were washed and incubated in appropriate sec- ondary antibodies at room temperature for 1 h. Next day, RPE-choroid complexes were washed and incubated in appropriate secondary antibodies for 1 h at room tempera- ture. At the end of the process, diamidine phenylindole (DAPI, 1:2000; Vector Laboratories) was used to coun- terstain the nucleus. Images were taken with a fluorescence microscope (Olympus, Tokyo, Japan).

Quantification of macrophages, CNV, and SRF area
To quantify the number of macrophages around fibrosis tissue, DAPI (blue) was used to highlight the nucleus, and F4/80 (red, macrophage marker) was used to display mac- rophages. Stellate-like cells suggesting activated macro- phages and immunoreactive for both F4/80 and DAPI around the subretinal lesions were counted and analyzed.
The CNV and SRF area in choroidal flat mounts on day 7 and 35 after laser treatment were measured by the CellSens Standard software (cell Sens Standard 1.9) that was built in with the fluorescence microscope (Olympus). CNV area was determined by measuring the hyperfluorescent area stained by IB4 (green), and the SRF area was determined by mea- suring the hyperfluorescent area stained by Fibronectin (red). In addition, in some cases, the center of the lesion that was not stained or false-positive stain need to be subtracted. To get comparable results for measuring CNV and fibrosis area, a uniform picture capture method was used (one laser lesion in the center, 200 · magnification, and scale bar = 100 mm). Each fibrosis area was measured at least 3 times by a co-worker who was blind for grouping, and then the average value was used for statistical analysis.

Flow cytometry
The circulating blood of mice was ground in a dish containing red cell lysis solution to obtain single-cell sus- pension. Then the single-cell suspension was stained with the FITC-anti-CD11b antibody (Biolegend). Flow cyto- metry was carried out on a Beckman fluorescence-activated cell sorter.

Western blots
Mice retina-choroid tissues and RAW cells were collected and chopped into homogenate. Protein concentrations were transferred onto a polyvinylidene difluoride membrane ( Merck Millipore, Billerica). After blocking with 5% milk for 1 h, the membranes were incubated with primary anti- bodies for Fibronectin (AB2033; Millipore), Collagen I (ab21286; Abcam), TGF-b2 (ab36495; Abcam), Arg-1 (9819S; Cell Signaling Technology), Arg-1 (sc-271430;


Santa Cruz), YM-1 (1404; Stem Cell), Autophagy-related gene-5 (Atg-5, 12994; Cell Signaling Technology), Beclin-1 (3495; Cell Signaling Technology), p-Akt (S473) (4058; Cell Signaling Technology), pan-Akt (4691; Cell Signaling Technology), GAPDH (10494-1-ap; Proteintech), and Actin (4970; Cell Signaling Technology) at 4°C overnight. The next day, membranes were washed and incubated in ap- propriate secondary antibodies for 1 h at room temperature and visualized using a molecular imaging system (Amer- sham Imager 600; GE Healthcare). GAPDH was used as the loading control.

Cell isolation, culture, and migration assay
The primary fibroblasts were acquired through a tissue block cultivation method. In brief, cornea, lens, and retina were separated from the choroid and the RPE-choroid-sclera complexes were immersed in a 0.25% trypsin solution for 20 min. The RPE was stripped off, and the resting tissue is cut into small pieces and spread evenly on the petri dish and suspended with DMEM containing 1% penicillin/strepto- mycin and 15% fetal bovine serum (FBS). The tissues were maintained in a 5% CO2 incubator at 37°C. Primary fibro- blasts will grow from the tissue pieces. The media were changed every 3 days, and unattached tissues were washed away with phosphate buffer.
The mouse macrophage cell line RAW 264.7 cells and primary fibroblasts were cultured in DMEM containing 10% FBS and 1% penicillin/streptomycin. The RAW cells were treated with IL-4 (100 ng$mL-1; R&D Systems), PI3K in- hibitor LY294002 (25 mM; Absin), or 3-MA (5 mM; Sigma) according to the previous literature.23 Supernatant medium from IL-4-treated or 3-MA-treated RAW cells was used to culture fibroblasts for 24 h, and 6–8 lines were randomly measured to calculate the average distance between cells. Im- age J was used for measurement. Cell migration distance = 0h distance -24 h distance
RNA interference
RAW cells were transfected with double-stranded siRNA or negative control siRNA (non-siRNA) using TranslT-X2 Dynamic Delivery System ( Mirus). Target sequences were as follows: Atg5-siRNA-418 (sense: 5¢-GAGUUGGUAA CUGACAAATT-3¢; antisense: 5¢-UUUGUCAGUUACCA ACGUCTT-3¢); Atg5-siRNA-520 (sense: 5¢-GCAUUAUCC AAUU-GGUUUATT-3¢; antisense: 5¢-UAAACCAAUUGG
AUAAUGCTT-3¢); and negative control (sense: 5¢-UUCU CCGAACGUGUCACGUTT-3¢; antisense: 5¢-ACGUGACA
CGUUCGGAGAATT-3¢). Double-stranded siRNAs were synthesized by Shanghai Gene Pharma (Shanghai, China).

Statistical analysis
All data are expressed as mean standard deviation. Sta- tistical analysis was performed using SPSS 20.0. In all studies, a co-worker was blinded to the experimental pro- tocols. Values are expressed as mean – SEM. Statistical analysis was performed using SPSS 20.0. (Version 20; SPSS, Inc., Chicago, IL). Comparisons among groups were carried out with 1-way or 2-way ANOVA, respectively, followed by Tukey’s post hoc test when F achieved P < 0.05 (>2 groups) or Student’s 2-tailed unpaired t-test (2 groups). P < 0.05 was considered statistically significant.

3-MA alleviates CNV and SRF on day 35
To determine an appropriate time point to study the anti- fibrotic role of 3-MA in the laser injury mouse model, we performed Western blots for Fibronectin and Collagen-I (2 makers of fibrosis). We found that both proteins signifi- cantly increased on day 21 and 35, which could be sup- pressed by 3-MA treatment (Fig. 1A). Then to monitor the health of the animals over the 35-day treatment period, we draw a weight curve in Fig. 1B, and we found that during the observation period from day 0 to 35, the weight of the mice in each group increased steadily, with no statistically sig- nificant difference between the 2 groups. To evaluate the antifibrotic role of 3-MA, we performed IF staining on choroidal flat mounts. On day 7, there was no significant difference in the area of CNV (green, IB4) and SRF (red, Fibronectin) between the 2 groups (0.032 – 0.003 mm2 and
0.030 – 0.002 mm2, respectively, P > 0.05). However, on day 35, the SRF area of 3-MA-treated group lessened mainly (0.047 – 0.019 mm2 and 0.017 – 0.003 mm2, respectively, P < 0.01), and there was also less CNV area on choroidal flat mounts of the 3-MA-treated eyes (0.011 – 0.001 mm2 and
0.002 – 0.0003 mm2, respectively, P < 0.01) (Fig. 1C, D).
The above data suggest that the longtime usage of 3-MA is needed to play the role of antifibrosis and day 35 can be used as an appropriate time point for testing the antifibrotic role of 3-MA.

3-MA decreases the population of both local and circulating macrophages
Macrophages are critical regulators for fibrosis develop- ment24; therefore, we investigated whether 3-MA affects macrophage population around the fibrosis area. We labeled RPE/choroid flat mounts with Abs against F4/80 to repre- sent macrophages, and IF staining showed that 3-MA treatment reduced the number of macrophages around the disciform scar on day 35 after laser (41.7 – 9.8 and 8.8 – 5.5, respectively, P < 0.01) (Fig. 1E). Since experimental studies have shown that 70%–90% of macrophages at laser damage are blood derived,25 we analyzed the percentage of circu- lating monocyte from mice treated with or without 3-MA by flow cytometry. We found that 3-MA treatment reduced the percentage of circulating monocyte in the whole blood (5.53% – 0.94% and 3.48% – 0.49%, respectively, P < 0.05) (Fig. 1F).

3-MA treatment inhibits macrophage M2 polarization
To determine which type of macrophages surrounds the fibrosis area, we performed IF staining for YM-1and Arg-1 (both M2 markers) on choroidal flat mounts on day 35 after laser. Colocalization of F4/80 and YM-1 or Arg-1 showed that most macrophages around the SRF lesion on day 35 after laser were M2 macrophages (Fig. 2A).
For further confirmation, we tested the expression levels of unique genes (YM-1 and Arg-1) of M2 macrophages by Western blot analysis. Compared with the controls, the ex- pression of Arg-1 and YM-1, both upregulated in eyes from the solely laser-treated group and downregulated signifi- cantly in eyes from 3-MA-treated group (Fig. 2B). To

Downloaded by UPPSALA UNIVERSITETSBIBLIOTEK from www.liebertpub.com at 06/21/20. For personal use only.

FIG. 1. 3-MA alleviates SRF and macrophage population on day 35 after laser. (A) Western blots showed a time-dependent alteration of Fibronectin and Collagen I expressions and 3-MA inhibited their upregulation on day 21 and 35. n = 4 eyes per group. **P < 0.01. (B) Weight curve of mice during the treatment period. n = 6 mice per group.
(C) Immunohistochemistry representative images of CNV (stained by IB4) and SRF (stained by Fibronectin) on the choroid flat mounts on day 7 and 35 after laser (Scale bars 50 mm). (D) The mean size of CNV and SRF area on day 7 and 35 after laser and in 3-MA-treated group. n = 8 eyes per group. **P < 0.01. (E) IF staining for F4/80 (red, macrophage marker) in laser-injured eyes with or without 3-MA treatment. Nuclei are counterstained with DAPI (blue). Stellate-like cells immunoreactive for both F4/80 and DAPI (white dots) in the subretinal lesion were counted and analyzed. n = 10 eyes per group. **P < 0.01 (Scale bars 100 mm). (F) Circulating blood was collected from mice with or without 3-MA treatment. Percentages of CD11b (monocyte marker)-positive cell subsets were analyzed by flow cytometer. Similar results were obtained in 3 independent experiments. n = 3 mice per group. *P < 0.05. DAPI, diamidine phenylindole; IF, immunofluorescence; 3-MA, 3-methyladenine; SRF, subretinal fibrosis. Color images are available online.

Downloaded by UPPSALA UNIVERSITETSBIBLIOTEK from www.liebertpub.com at 06/21/20. For personal use only.

FIG. 2. 3-MA treatment inhibits macrophage M2 polarization. (A) IF staining for YM-1 (green; a M2 macrophage marker), Arg-1 (green; a M2 macrophage marker), F4/80 (red; macrophage marker), and DAPI (blue; a marker for cell nucleus) on the choroid flat mounts on day 35 after laser. Stellate-like cells immunoreactive for both YM-1 (or Arg-1) and F4/80 were considered M2 macrophages (white asterisks). (Scale bars 100 mm). (B) Western blots show 3-MA treatment decreased the expression of Arg-1and YM-1 in RPE choroidal-scleral complexes on day 35. n = 3 samples per group. *P < 0.05, **P < 0.01; Similar results were obtained in 3 independent experiments. (C) Histological section through a lesion on day 35 after laser. Macrophages (stained by F4/80) located in inner retina, choroidal vasculature (white arrows), as well as the subretinal space and YM-1 or Arg-1 positive macrophages (white asterisks) decreased in 3-MA-treated group. n = 16 lesions per group. (Scale bars: 100 mm). (D) Average number of cells per tissue section immunoreactive for both YM-1 and F4/80 or solely immunoreactive for F4/80. n = 16 lesions per group. **P < 0.01. (E) Average number of cells per tissue section immunoreactive for both Arg-1 and F4/80 or solely immunoreactive for F4/80. n = 16 lesions per group. **P < 0.01. RPE, retinal pigment epithelium. Color images are available online.


investigate whether 3-MA affects macrophage subtypes, we performed IF on retinal histologic sections on day 35 after laser. 3-MA treatment reduced the number of macrophages per retinal section (P < 0.01). More importantly, there are many macrophages (F4/80 positive cells) immunoreactive for YM-1 (67% per section) and Arg-1 (78% per section), whereas in the 3-MA-treated sections, YM-1 and Arg-1 signaling can rarely be found, although there was still a small amount of F4/80 positive cells (Fig. 2C–E).

3-MA treatment inhibits SRF by suppressing fibroblasts recruitment
Macrophages are near the collagen-producing myofibro- blasts, and macrophages produce profibrotic mediators that directly activate fibroblasts, such as TGF-b2.26 To demon- strate the role of M2 macrophages in SRF, we examined the relationship of location and population between a-SMA+ fibroblasts and F4/80+ macrophages. We found that mac- rophages were closer to the center of the SRF lesion, and fibroblasts were adjacent to macrophages and located at the periphery of the SRF lesion. In addition, 3-MA treatment decreased macrophage and fibroblast population, accompa- nied by the decreased size of the SRF area (Fig. 3A).
To verify the profibrotic effect of M2 macrophages in vitro, IL-4 was used to induce M2 polarization in RAW 264.723 and TGF-b2 expressions increased after exposure to IL-4 for 24 h, but 3-MA treatment downregulated TGF-b2 expression levels (Fig. 3B). To illustrate whether 3-MA affects fibroblast actions through M2 macrophages, we examined the effects of conditioned medium from macro- phages exposed to IL-4 and 3-MA in vitro on fibroblast behavior. As shown in Fig. 3C, after choroid fibroblasts were cultured with the conditioned medium from macro- phages for 24 h, we observed increased migration rate in IL-4-treated group, but decreased migration rate in 3-MA- treated group (Fig. 3C).

3-MA inhibits macrophage polarization through PI3K/Akt signaling pathway
As 3-MA was a well-known autophagy inhibitor, thus, we examined the effect of autophagy inhibition on macrophage polarization. To ensure specificity of the knockdown, 2 Atg- 5 siRNA with different sequences was used. We tested the decreased expression of Atg5 and Beclin-1, which verified the knockdown efficiency and inhibition of autophagy level. However, Arg-1 and YM-1 expression did not decrease, but increased after silencing Atg-5 gene expression (Fig. 4A), indicating that the suppressing role of 3-MA on M2 polar- ization of macrophages is not through the autophagy sig- naling pathway.
It is well established that activation of the PI3K/Akt sig- naling cascade plays a pivotal role in macrophage activation and M1/M2 polarization.27 Considering that 3-MA can in- hibit not only the III PI3K-autophagy pathway but also the I PI3K/Akt pathway, more importantly, it has been proved that 3-MA blocks class I PI3K persistently, whereas it suppressed class III PI3K transiently28; thus, we detected the expression of phosphorylated-Akt (p-Akt) in RPE-choroid complexes of mice. We found that p-Akt expression was significantly up- regulated on day 35 after laser photocoagulation, but down- regulated obviously in the 3-MA-treated group (Fig. 4B),

indicating that the antifibrosis effect of 3-MA is indeed re- lated to the Akt activation. Next, to prove that 3-MA affects macrophage polarization through the PI3K/Akt signaling pathway, an inhibitor of PI3K (LY294002) was used in in vitro experiment. Western blots analysis showed IL-4- treated RAW cells have upregulated expression of Arg-1 and YM-1; however, LY294002 treatment inhibited the IL- 4-induced upregulations of Arg-1 and YM-1 (Fig. 4C), which confirmed the inhibitory effect of PI3K pathway on M2 polarization of macrophages. Finally, we demonstrated that 3-MA treatment could inhibit both M2 macrophage polarization (Arg-1 and YM-1) and Akt activation in RAW cells (Fig. 4D), which further demonstrates that the inhi- bition of 3-MA on macrophage M2 polarization is closely related to its inhibition of PI3K/AKT pathway.

In this study, we found that the systemic application of 3-MA could alleviate the size of experimental SRF. Be- sides, 3-MA decreased both local and circulating macro- phages, and counteracted M2 polarization and fibroblast migration. Mechanistically, we proved that the inhibitory effects of 3-MA on experimental SRF might involve in- hibited activation of the PI3K/Akt signaling pathway. To- gether, our data demonstrate the potential treatment effect of 3-MA in SRF prevention and treatment.
Macrophage is a vital participator in fibrosis develop- ment. He and Marneros10 ever showed that ablation of macrophages after laser injury inhibits the formation of this fibrovascular scaffold, indicating that macrophages are in- volved in the progression of SRF as early as the onset of injury and CNV development. Furthermore, M2 macro- phage was proved to be involved in the development of SRF in a delayed but sustained pattern in the laser injury mod- el.29,30 In this study, we found abundant macrophages around the SRF area and macrophages were closer to the center of the SRF lesion; however, fibroblasts were adjacent to macrophages and located at the periphery of the SRF lesion, and we further proved that M2 macrophages could promote fibroblast migration. Therefore, we inferred that macrophages around the SRF area (local or surrounding macrophages) may play a role in recruiting collagen- producing fibroblasts through macrophage-derived TGF-b. The antifibrotic effect of 3-MA might be partly due to in- hibition of the surrounding macrophages and macrophage- recruited fibroblasts.
On the other hand, studies have shown that 70%–90% of macrophages at laser damage are blood derived.25 In addi- tion, Bao et al. ever reported that 3-MA could suppress the infiltration of macrophages and lymphocytes in injured kidneys,18 suggesting 3-MA may inhibit SRF by intervening circulating macrophages. Thus, we measured the population of macrophages in circulation. Interestingly, we found that 3-MA treatment decreased the proportion of mononuclear macrophages in circulation. Consistently, several studies have confirmed the treatment effect of systemic application of 3-MA in various diseases through inhibiting macrophage infiltration, including renal fibrosis,18 pulmonary fibrosis,31 and arteriosclerosis.19 These studies suggest that macrophages are targeted by 3-MA. Systemic inhibition of macrophage infiltration of 3-MA is an absolute advantage of 3-MA since Brockmann et al.25 once proved that local partial depletion

Downloaded by UPPSALA UNIVERSITETSBIBLIOTEK from www.liebertpub.com at 06/21/20. For personal use only.

FIG. 3. 3-MA treatment inhibits SRF by suppressing M2 macrophage-recruited fibroblasts. (A) Triple IF staining for a-smooth muscle cell (a-SMA, fibroblast marker), F4/80 (macrophage marker), and DAPI (cell nucleus marker) on the choroid flat mounts on day 35 after laser. 3-MA treatment reduced surrounding macrophages and fibroblasts. n = 8 eyes per group. **P < 0.01 (Scale bars 100 mm) (B) IL-4-treated macrophages have upregulated TGF-b2 expression levels, and 3-MA treatment suppresses TGF-b2 expression levels. n = 3 samples per group. **P < 0.01. Similar results were obtained in 3 independent experiments. (C) IL-4-treated macrophage medium promoted fibroblast migration, and 3-MA treatment decreased fibroblast migration. n = 3 samples per group. (Scale bars 200 mm) **P < 0.01. Similar results were obtained in 3 independent experiments. a-SMA, a-smooth muscle actin; TGF-b2, transforming growth factor-b2. Color images are available online.

Downloaded by UPPSALA UNIVERSITETSBIBLIOTEK from www.liebertpub.com at 06/21/20. For personal use only.

FIG. 4. 3-MA inhibits macrophage polarization through PI3K/Akt signaling pathway. (A) Western blots show knockdown of Atg-5 by Atg5-siRNA-418 or Atg-5 siRNA-520 increased Arg-1 and YM-1 expression. n = 3 samples per group. **P < 0.01. (B) Western blots show 3-MA treatment suppressed p-Akt expression level in eyes of mice on day 35 after laser. n = 3 samples per group. **P < 0.01. (C) Western blots show PI3K inhibitor LY294002 treatment can inhibit Arg-1 and YM-1 expression induced by IL-4. n = 3 samples per group. **P < 0.01. (D) Western blots show 3-MA treatment suppressed p-Akt, Arg-1, and YM-1 expression levels upregulated by IL-4 stimulation. n = 3 samples per group.
*P < 0.05, **P < 0.01. Similar results were obtained in 3 independent experiments. p-Akt, phosphorylated-Akt; PI3K, phosphatidylinositol-3-kinase. Color images are available online.


of macrophages did not reveal a significant reduction in CNV areas; however, systemic inhibition of macrophages works, which suggested the critical role of systemic mac- rophages in CNV and SRF. More interestingly, Subhi et al.32 recently found that the population of circulating monocytes increased with the number of anti-VEGF injec- tions in patients with neovascular AMD. The increased population of macrophages in the circulation may partly explain why patients with various anti-VEGF therapy are more likely to develop SRF.6,33
Another vital mechanism associated with the antifibrotic effect of 3-MA might be its inhibitory role in M2 macro- phage polarization. A pathological study showed that CNVM from patients with M2 macrophage infiltration showed more fibrosis than CNVM from patients with M1 macrophage infiltration.16 Likewise, this study showed that subretinal macrophages recruited around experimental SRF are predominantly M2 macrophages (Fig. 3). 3-MA could inhibit M2 polarization-related markers and M2 macrophage population, indicating the critical role of M2 macrophages in SRF and providing a potential target for future anti-SRF therapy. Our previous published studies showed that local targeting of M2 polarization could alleviate CNV develop- ment34; therefore, we believe that local intervention of macrophage M2 polarization may inhibit SRF of mice too. However, for AMD patients, local intervention alone might not be enough. It has been reported that compared with age- matched controls, peripheral blood mononuclear cells (PBMCs) extracted from whole blood of AMD patients have significantly higher VEGF level35; in addition, activated M1 (induced by IFNg and LPS) and M2 (induced IL-4 and IL- 13) macrophages from these PBMCs are both associated with a proangiogenic gene and protein expression profile,36 indicating the proangiogenic characteristics of circulating monocytes of AMD patients. Moreover, studies found that certain subsets of circulating monocytes are closely related to the pathological process of AMD, for example, the CD11b+CD200+ circulating monocyte37 or the CD14+CD16+ circulating monocyte.35 Taken together, the circulating monocytes and local M2 macrophages are both vital factors to be concerned with in future antifibrosis treatment, and 3-MA, with the advantage of targeting both circulating and local macrophages, would be a good supplement for anti-VEGF treatment.
3-MA was widely considered an autophagy inhibitor due to its inhibition in class III PI3K. To explore the mechanism of 3-MA in macrophage polarization, we inhibited autop- hagy by knocking out autophagy-related gene Atg-5 and found that autophagy inhibition upregulated the expression of YM-1 and Arg-1. It suggests that autophagy inhibition is a promoter rather than an inhibitor of M2 polarization, in- directly indicating that 3-MA inhibiting macrophage M2 polarization is not through inhibition of class III PI3K- autophagy pathway. Autophagy was also reported to sup- press isoprenaline-induced M2 macrophage polarization in the tumor microenvironment of breast cancer.38 On the other hand, 3-MA is also a class I PI3K inhibitor and it has been reported that 3-MA has a dual role on PI3K: 3-MA blocks class I PI3K persistently, whereas suppresses class III PI3K transiently,28 which highlights the advantage of 3-MA in inhibiting class I PI3K/Akt pathway.
3-MA, an analog of adenosine triphosphate (ATP) in structure, was reported to compete with ATP for a binding

site on PI3K in the cytoplasm, leading to class I PI3K losing the ability to phosphorylate PIP2, thus blocking the PI3- K/Akt pathway.39 The PI3K/Akt signaling pathway modu- lates multiple cellular processes and participates in the pathological process of AMD. Inhibition of PI3K/Akt sig- naling pathway could significantly suppress vascular leak- age and CNV in experimental CNV.40,41 It is also well established that PI3K/Akt signaling cascade plays a pivotal role in promoting macrophage activation and M1/M2 po- larization.27 The class I PI3K activation has been reported as an essential step toward M2 activation of macrophages in response to IL-442 and Akt activation is required for M2 activation because Akt inhibition abrogates the upregulation of M2 genes.43 According to our Western blot analysis, protein markers (Arg-1 and YM-1) of M2 macrophages were indeed inhibited by PI3K inhibitor treatment. In ad- dition, 3-MA suppressed the expression of both p-Akt and M2 markers of macrophages. Therefore, we speculated that 3-MA might inhibit macrophage M2 polarization by in- hibiting class I PI3K/Akt pathway.
However, whether 3-MA is beneficial to SRF in AMD patients needs further investigations since other risk factors for AMD, such as age- and RPE-related factors, were not taken into account in our present studies. Some studies insist that age affects the expression pattern of macrophage cy- tokines, and younger macrophages have antiangiogenic properties, but older macrophages do not. Besides, old mice demonstrated significantly more neovascularization com- pared to young mice.44,45 Therefore, we speculate that older mice may have more severe SRF than younger mice. Al- though we had demonstrated the antifibrotic role of 3-MA through inhibition of macrophage using young mice (6–8 weeks), retesting it using the old mouse (16–18 weeks) may make our results more robust, which is one of the limitations of our research. Second, we detected high expression of TGF-b in IL-4-stimulated RAW cells, whose culture solu- tion promoted fibroblast migration, but we could not attri- bute all credit to macrophages, although they showed significant changes before and after 3-MA treatment. Pre- vious literature has shown that RPE cells can be transformed into fibroblasts through the epithelial-mesenchymal transi- tion to promote SRF.46 In addition to macrophages, RPE and fibroblasts can secrete TGF-b and other profibrotic factors.47 However, our study did not explore the effect of 3-MA on other local cells such as RPE; thus, the antifibrosis mecha- nism of 3-MA needs further research.
In conclusion, in this study, we proved that 3-MA pro- foundly inhibited circulating and local macrophage amounts, M2 polarization, and fibroblast recruitment to al- leviate SRF. Mechanistically, we showed that 3-MA in- hibited SRF and the macrophage M2 polarization through the PI3K/Akt pathway. Our findings suggest that 3-MA may have promising therapeutic applications for the treatment in SRF.

Author Disclosure Statement
No competing financial interests exist.

Funding Information
This study was supported by the National Natural Science Foundation of China (81730026); National Science and Technology Major Project of China (2017ZX09304010);

10 BO ET AL.

Project of Shanghai Hospital Development Center (SHDC12016105); Science and Technology Commission of Shanghai Municipality (17411953000, 16140900800, and 0303N17001); National New Drugs Program (2018 ZX09301029001); National Key R&D Program (2016YFC 0904800 and 2017YFA0105301), and Eastern Young Pro- gram (QD2016003).

1. Schmidt-Erfurth, U., Chong, V., Loewenstein, A., et al. Guidelines for the management of neovascular age-related macular degeneration by the European Society of Retina Specialists (EURETINA). Br. J. Ophthalmol. 98:1144– 1167, 2014.
2. Lim, L.S., Mitchell, P., Seddon, J.M., Holz, F.G., and Wong, T.Y. Age-related macular degeneration. Lancet. 379:1728–1738, 2012.
3. D’Souza, P., Ranjan, R., Babu, U., Kanakath, A.V., and Saravanan, V.R. Inflammatory choroidal neovascular membrane: long-term visual and anatomical outcomes after intravitreal anti-vascular endothelial growth factor therapy. Retina. 38:1307–1315, 2018.
4. Bhisitkul, R.B., Desai, S.J., Boyer, D.S., Sadda, S.R., and Zhang, K. Fellow eye comparisons for 7-year outcomes in ranibizumab-treated AMD subjects from ANCHOR, MARINA, and HORIZON (SEVEN-UP Study). Ophthal- mology. 123:1269–1277, 2016.
5. Hwang, J.C., Del, P.L., Freund, K.B., Chang, S., and Ir- anmanesh, R. Development of subretinal fibrosis after anti- VEGF treatment in neovascular age-related macular de- generation. Ophthalmic Surg. Lasers Imaging. 42:6–11, 2011.
6. Daniel, E., Toth, C.A., Grunwald, J.E., et al. Risk of scar in the comparison of age-related macular degeneration treat- ments trials. Ophthalmology. 121:656–666, 2014.
7. Daniel, E., Pan, W., Ying, G.S., et al. Development and course of scars in the comparison of age-related macular degeneration treatments trials. Ophthalmology. 125:1037– 1046, 2018.
8. Ruiz-Moreno, J.M., and Montero, J.A. Subretinal fibrosis after photodynamic therapy in subfoveal choroidal neo- vascularisation in highly myopic eyes. Br. J. Ophthalmol. 87:856–859, 2003.
9. Ahn, S.J., Park, K.H., and Woo, S.J. Subretinal fibrosis after antivascular endothelial growth factor therapy in eyes with myopic choroidal neovascularization. Retina. 36: 2140–2149, 2016.
10. He, L., and Marneros, A.G. Macrophages are essential for the early wound healing response and the formation of a fibrovascular scar. Am. J. Pathol. 182:2407–2417, 2013.
11. Grunin, M., Hagbi-Levi, S., and Chowers, I. The role of monocytes and macrophages in age-related macular de- generation. Adv. Exp. Med. Biol. 801:199–205, 2014.
12. Chinnery, H.R., McMenamin, P.G., and Dando, S.J. Mac- rophage physiology in the eye. Pflugers Arch. 469:501– 515, 2017.
13. Cherepanoff, S., McMenamin, P., Gillies, M.C., Kettle, E., and Sarks, S.H. Bruch’s membrane and choroidal macro- phages in early and advanced age-related macular degen- eration. Br. J. Ophthalmol. 94:918–925, 2010.
14. Zhang, R., Liu, Z., Zhang, H., Zhang, Y., and Lin, D. The COX-2-selective antagonist (NS-398) inhibits choroidal neovascularization and subretinal fibrosis. PLoS One. 11: e146808, 2016.

15. Italiani, P., and Boraschi, D. From monocytes to M1/M2 macrophages: phenotypical vs. functional differentiation. Front. Immunol. 5:514, 2014.
16. Cao, X., Shen, D., Patel, M.M., et al. Macrophage polari- zation in the maculae of age-related macular degeneration: a pilot study. Pathol. Int. 61:528–535, 2011.
17. Seglen, P.O., and Gordon, P.B. 3-Methyladenine: specific inhibitor of autophagic/lysosomal protein degradation in isolated rat hepatocytes. Proc. Natl. Acad. Sci. U. S. A. 79: 1889–1892, 1982.
18. Bao, J., Shi, Y., Tao, M., et al. Pharmacological inhibition of autophagy by 3-MA attenuates hyperuricemic nephrop- athy. Clin. Sci. (Lond). 132:2299–2322, 2018.
19. Dai, S., Wang, B., Li, W., et al. Systemic application of 3- methyladenine markedly inhibited atherosclerotic lesion in ApoE(-/-) mice by modulating autophagy, foam cell for- mation and immune-negative molecules. Cell Death Dis. 7: e2498, 2016.
20. Zhang, P., Wang, H., Luo, X., et al. MicroRNA-155 in- hibits polarization of macrophages to M2-type and sup- presses choroidal neovascularization. Inflammation. 41: 143–153, 2018.
21. Wu, Y., Zhang, Y., Wang, L., Diao, Z., and Liu, W. The role of autophagy in kidney inflammatory injury via the NF-kappaB route induced by LPS. Int. J. Med. Sci. 12:655– 667, 2015.
22. Lee, Y.R., Wang, P.S., Wang, J.R., and Liu, H.S. En- terovirus 71-induced autophagy increases viral replication and pathogenesis in a suckling mouse model. J. Biomed. Sci. 21:80, 2014.
23. Gordon, S., and Martinez, F.O. Alternative activation of macrophages: mechanism and functions. Immunity. 32: 593–604, 2010.
24. Wynn, T.A., and Barron, L. Macrophages: master regula- tors of inflammation and fibrosis. Semin Liver Dis. 30:245– 257, 2010.
25. Brockmann, C., Kociok, N., Dege, S., et al. Local partial depletion of CD11b(+) cells and their influence on cho- roidal neovascularization using the CD11b-HSVTK mouse model. Acta. Ophthalmol. 96:e789–e796, 2018.
26. Lech, M., and Anders, H.J. Macrophages and fibrosis: how resident and infiltrating mononuclear phagocytes orches- trate all phases of tissue injury and repair. Biochim. Bio- phys. Acta. 1832:989–997, 2013.
27. Vergadi, E., Ieronymaki, E., Lyroni, K., Vaporidi, K., and Tsatsanis, C. Akt signaling pathway in macrophage acti- vation and M1/M2 polarization. J. Immunol. 198:1006– 1014, 2017.
28. Wu, Y.T., Tan, H.L., Shui, G., et al. Dual role of 3- methyladenine in modulation of autophagy via different temporal patterns of inhibition on class I and III phos- phoinositide 3-kinase. J. Biol. Chem. 285:10850–10861, 2010.
29. Peng, X., Xiao, H., Tang, M., et al. Mechanism of fibrosis inhibition in laser induced choroidal neovascularization by doxycycline. Exp. Eye Res. 176:88–97, 2018.
30. Yang, Y., Liu, F., Tang, M., et al. Macrophage polarization in experimental and clinical choroidal neovascularization. Sci. Rep. 6:30933, 2016.
31. Liu, H., Cheng, Y., Yang, J., et al. BBC3 in macrophages promoted pulmonary fibrosis development through induc- ing autophagy during silicosis. Cell Death Dis. 8:e2657, 2017.
32. Subhi, Y., Krogh, N.M., Molbech, C.R., et al. Association of CD11b+ monocytes and anti-vascular endothelial growth


factor injections in treatment of neovascular age-related macular degeneration and polypoidal choroidal vasculo- pathy. JAMA Ophthalmol. 137:515–522, 2019.
33. Barikian, A., Mahfoud, Z., Abdulaal, M., Safar, A., and Bashshur, Z.F. Induction with intravitreal bevacizumab every two weeks in the management of neovascular age- related macular degeneration. Am. J. Ophthalmol. 159:131– 137, 2015.
34. Xu, N., Bo, Q., Shao, R., et al. Chitinase-3-like-1 promotes M2 macrophage differentiation and induces choroidal neovascularization in neovascular age-related macular de- generation. Invest. Ophthalmol. Vis. Sci. 60:4596–4605, 2019.
35. Chen, M., Lechner, J., Zhao, J., et al. STAT3 Activation in circulating monocytes contributes to neovascular age- related macular degeneration. Curr. Mol. Med. 16:412–423, 2016.
36. Hagbi-Levi, S., Grunin, M., Jaouni, T., et al. Proangiogenic characteristics of activated macrophages from patients with age-related macular degeneration. Neurobiol. Aging. 51: 71–82, 2017.
37. Singh, A., Falk, M.K., Hviid, T.V., and Sorensen, T.L. Increased expression of CD200 on circulating CD11b+ monocytes in patients with neovascular age-related macular degeneration. Ophthalmology. 120:1029–1037, 2013.
38. Shan, M., Qin, J., Jin, F., et al. Autophagy suppresses isoprenaline-induced M2 macrophage polarization via the ROS/ERK and mTOR signaling pathway. Free Radic. Biol. Med. 110:432–443, 2017.
39. Zou, Z., Zhang, J., Zhang, H., et al. 3-Methyladenine can depress drug efflux transporters via blocking the PI3K- AKT-mTOR pathway thus sensitizing MDR cancer to chemotherapy. J. Drug. Target. 22:839–848, 2014.
40. Guo, J., Luo, X., Liang, J., Xiao, M., and Sun, X. Anti- angiogenic effects of doxazosin on experimental choroidal neovascularization in mice. J. Ocul. Pharmacol. Ther. 33: 50–56, 2017.
41. Ma, J., Sun, Y., Lopez, F.J., et al. Blockage of PI3K/mTOR pathways inhibits laser-induced choroidal neovasculariza- tion and improves outcomes relative to VEGF-A suppres- sion alone. Invest. Ophthalmol. Vis. Sci. 57:3138–3144, 2016.
42. Weisser, S.B., McLarren, K.W., Voglmaier, N., et al. Al- ternative activation of macrophages by IL-4 requires SHIP degradation. Eur. J. Immunol. 41:1742–1753, 2011.

43. Ruckerl, D., Jenkins, S.J., Laqtom, N.N., et al. Induction of IL-4Ralpha-dependent microRNAs identifies PI3K/Akt signaling as essential for IL-4-driven murine macrophage proliferation in vivo. Blood. 120:2307–2316, 2012.
44. Zhao, H., Roychoudhury, J., Doggett, T.A., Apte, R.S., and Ferguson, T.A. Age-dependent changes in FasL (CD95L) modulate macrophage function in a model of age-related macular degeneration. Invest Ophthalmol Vis. Sci. 54: 5321–5331, 2013.
45. Dace, D.S., and Apte, R.S. Effect of senescence on mac- rophage polarization and angiogenesis. Rejuvenation Res. 11:177–185, 2008.
46. Ishikawa, K., Sreekumar, P.G., Spee, C., et al. alphaB- Crystallin regulates subretinal fibrosis by modulation of epithelial-mesenchymal transition. Am. J. Pathol. 186:859– 873, 2016.
47. Ishikawa, K., Kannan, R., and Hinton, D.R. Molecular mechanisms of subretinal fibrosis in age-related macular degeneration. Exp. Eye. Res. 142:19–25, 2016.

Received: September 18, 2019
Accepted: March 9, 2020

Address correspondence to: Prof. Xiaodong Sun Department of Ophthalmology Shanghai General Hospital
(Shanghai First People’s Hospital) Shanghai Jiao Tong University School of Medicine
100 Haining Road
Shanghai 200080
China E-mail: [email protected]
Dr. Xueting Luo Department of Ophthalmology Shanghai General Hospital (Shanghai First People’s Hospital)
Shanghai Jiao Tong University School of Medicine
100 Haining Road
Shanghai 200080
China E-mail: [email protected]