Amniotic Membrane Induces Peroxisome Proliferator-Activated Receptor-γ Positive Alternatively Activated Macrophages

Abstract

Purpose.: Amniotic membrane transplantation (AMT) reportedly improves herpetic stromal keratitis (HSK). Here we studied the role of the amniotic membrane (AM) on macrophages.

Methods.: BALB/c mice with necrotizing HSK received an AMT or tarsorrhaphy (TAR) as control. Apoptosis of F4/80+ cells was determined using the annexinV/7-AAD system. Macrophage invasion was determined using a cornea invasion assay. Cytokine secretion was quantified by ELISA. Arginase activity was measured by bioassay. Expression of nuclear factor (NF)-κB or peroxisome proliferator-activated receptor (PPAR)-γ related proteins was detected by Western blot analysis, and the expression of costimulatory surface molecules or PPAR-γ by flow cytometry. Lipid accumulation was observed by Oil red O and Sudan B staining.

Results.: After AMT apoptotic features of corneal macrophages, but also macrophage invasion increased. IL-6, IL-10, IL-12, TNF-α, and NF-κB content in HSK corneas had decreased with AMT. AMT increased expression of PPAR-γ, arginase 1 and 2, and arginase activity in AM-treated HSK corneas. In vitro, NF-κB, cytokine production, costimulatory molecules (CD80, CD86, CD40), phagocytic capacity, proliferation, viability, and accessory function to herpes simplex virus (HSV)-1 specific draining lymph node (DLN) cells were reduced in bone marrow derived macrophages (BM) cocultured with AM, while CD206, CD204, CD163, and CD68, lipid accumulation in the cytoplasm, PPAR-γ expression, and arginase activity was increased. An increase in viability and proliferation was observed in the presence of AM combined with apoptotic cells, compared with AM alone.

Conclusions.: Based on these results it can be concluded that the action mechanism of AM is associated with modulation of classically activated macrophages into alternatively activated macrophages or macrophage cell death, probably by engaging lipid metabolism and activating the PPAR-γ pathway, consequently curtailing effector T cell functions. Apoptotic cells induced in the environment with AM support the presence and survival of such macrophages.

Herpes simplex infections are very common, affecting up to 90% of humans. The cornea is frequently involved and herpetic stromal keratitis (HSK) can develop, which is among the leading causes of unilateral blindness worldwide. ¹ HSK is an immune-mediated disease, characterized by CD4⁺-mediated cellular responses and associated with the high risk of corneal opacity, edema, neovascularization, and ulceration. ^2,3  

Macrophages derive from hemopoietic progenitors and are distributed to all organs of the body, including the cornea. ⁴ Previous studies showed that macrophages are important effector cells and that they participate in host defense functions, which are mediated through Toll-like receptors and interferon-γ. ⁵  

Macrophage infiltration into the cornea early after herpes simplex virus (HSV)-1 infection promotes more severe HSK, presumably via their function as antigen-presenting cells (APC) and their accessory cell function to T lymphocytes. ^{6 –8} Furthermore, macrophage cytokines such as IL-6, IL-10, IL-12, and TNF-α are found in murine corneas with HSK and represent important factors for the outcome of the corneal infection. ^{9 –11}  

Macrophages may also be activated alternatively (e.g., by IL-4, IL-10, or other stimuli) and exhibit anti-inflammatory functions and promote tissue repair. ⁵ Besides, other investigations have shown that peroxisome-proliferator activated receptor (PPAR)-γ is required for the development of alternatively activated macrophages. ¹²  

Amniotic membrane (AM) is the innermost layer of the placenta, consisting of a basement membrane and an avascular stroma. Human AM is transplanted worldwide to reconstruct the ocular surface, e.g., after chemical burns, ¹³ to prevent scarring ¹⁴ and to improve wound healing. ¹⁵ The AM can induce various anti-inflammatory actions ^16,17 : in vitro studies disclosed that AM is able to regulate chemokine expression in human keratocytes (Bültmann SYL, et al. IOVS 1999;40:ARVO Abstract 3044) and to suppress mixed lymphocytic reactions. ¹⁸ Moreover, potent anti-inflammatory proteins have been found in the AM (e.g., IL-1Ra, and IL-10). ^19,20  

We have previously shown that human amniotic membrane transplantation (AMT) promotes rapid re-epithelialization and strongly reduces corneal inflammation in experimental HSK. ¹⁷ By electron microscopy we found an increase in apoptotic cells and phagocytes after only 12 hours. ²¹ Others have reported that RAW 264.7 macrophages cocultured with AM and IFN-γ undergo a rapid cell apoptosis ²² and showed a decreased amount of TNF-α and IL-6, but an increased IL-10 and decreased expression of major histocompatibility complex (MHC) II, CD80, and CD86 when cocultured with AM extract. ²³  

The AM action mechanism on macrophages that underlies the improvement in HSK has not yet been elucidated; therefore, corneas with HSK treated with AMT or bone marrow derived macrophages (BM) cocultured with AM were investigated to assess macrophage apoptosis and changes in the activation phenotype. The costimulatory function of macrophages to T cells on AM treatment was also assessed. Because apoptotic cells may also support invasion and survival of macrophages we investigated whether this could also be found in macrophages treated with AM. 

Materials and Methods

The study was carried out in accordance with the Institutional Animals Care and Use and Ethics Committee, and with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. ²⁴  

Mice

Female BALB/c mice (6–8 weeks) were purchased from Charles River Wiga (Sulzfeld, Germany) and used for all experiments. 

Virus Infection and Clinical Evaluation

HSV-1 (KOS) was prepared as described previously. ²⁴ Mice were anesthetized with ketamine (2 mg) and mepivacaine-hydrochloride (400 ng). The epithelium of the right eye was then scratched in a criss-cross pattern eight times. Five microliters of KOS solution containing 1 × 10⁵ PFU were placed on the cornea. The severity of stromal disease was graded by using an operation microscope (Zeiss, Oberkochen, Germany) on a scale of 0 to 4+, a score of 1+ being consistent with < 25%, 2+ < 50%, 3+ < 75%, and 4+ > 75% corneal opacity with neovascularization, edema, and thinning. ²⁴  

Preparation of Human Amniotic Membrane

Human placenta was obtained after elective cesarean delivery. The amniotic membrane was isolated by blunt dissection, washed in phosphate-buffered saline (PBS) containing antibiotics (50 μg/mL penicillin, 50 μg/mL streptomycin, 100 μg/mL neomycin, and 2.5 μg/mL amphotericin B) and flattened on nitrocellulose paper (Hybond+; Amersham, Buckinghamshire, UK). The AM pieces were stored at −80°C in DMEM/glycerol 1:1 (vol/vol) until use. ^17,25 Immediately before use, AM was thawed, washed three times with sterile medium (RPMI 1640), and cut into approximately 1.5 cm² pieces. 

Cytokine Quantification by ELISA

Corneas (n = 8, each group) were harvested 2 days after AMT or tarsorrhaphy (TAR) and then stored at −80°C. They were thawed, homogenized, and then centrifuged. In the supernatants the content of IL-6, IL-10, IL-12, or TNF-α and also of KC/CXCL-1 (BD Pharmingen, Hamburg, Germany) and MIP-2/CXCL-2 (R&D Systems, Wiesbaden, Germany) by ELISA were determined. ⁷  

Bone Marrow-Derived Macrophages

Bone marrow-derived macrophages (BM) were obtained from femurs of BALB/c mice and were cultured in RPMI 1640 medium containing 10% fetal calf serum and 15% L cell-conditioned media. ^{26 –28} After 5 days in culture, nonadherent cells were removed and adherent cells were harvested for assays. Adherent bone marrow-derived cells were 90% F4/80+ in our experiments (Fig. 1A). To each well of a 24-well plate, 1 × 10⁶ BM in 2 mL RPMI 1640 medium containing 10% fetal calf serum were added. Then 100 ng/mL LPS (Sigma-Aldrich,Taufkirchen, Germany), 100 U/mL IFN-gamma (Biomol, Hamburg, Germany) or medium (control) was added. AM with the epithelial side facing up was placed in the cell culture plate, with BM cultured on the bottom. The cell culture supernatants were harvested after 48 hours and used for ELISA. 

Figure 1.

 

Improvement of corneas with HSK lesions by AMT is associated with macrophage apoptosis. (A) Expression of F4/80 on bone marrow derived cells. Cells were analyzed for an increase of fluorescence with a cytometer (FACSCalibur, Becton Dickinson, San Jose, CA). BM samples on chamber slides 48 hours after treatment with AM. Slides were then stained (Hoechst staining), and were viewed under a fluorescence microscope. (B) BM with medium. (C) BM with LPS and IFN-γ. (D) BM with AM. (E) BM with LPS, IFN-γ, and AM. Many BM showed fragmented nuclei. (F) Eight corneas with TAR or AMT were pooled after 12 and 48 hours, treated with collagenase, and single cell suspensions were stained with CD45 (FITC), annexin V (PE), 7-AAD, and F4/80 (Alexa Fluor 647). CD45+ F4/80+ cells were detected in cornea samples treated with TAR or AMT. Increased numbers of F4/80+ cells after AMT also stained positively for annexin V and the nuclear dye 7-AAD, indicating cell death (early apoptosis, annexin V+/7-AAD−; late apoptosis, annexin V+/7-AAD+). Percentage of apoptotic F4/80+ cells after 48 hours: 1. Experiment: TAR, 32.9%; AMT 54%; 2. Experiment: TAR, 69.9%; AMT 75.2%; 3. Experiment: TAR 54.04%; AMT 61.9%. Viable cells with intact membrane did not stain by annexin V and excluded 7-AAD. (Representative experiment of three with eight mice per group).

Flow Cytometry Analysis of BM

BM were cocultured with AM, and after 48 hours the macrophages were removed from culture dishes using PBS-EDTA on ice. The sections were blocked with 0.05 mg/mL medium (Fc-block; BD Biosciences, Hamburg, Germany) to avoid unspecific staining. Macrophages were characterized with an antibody directed against F4/80 (clone Cl:A3-1; AbD Serotec, Düsseldorf, Germany), CD80 (clone 16-10A1; Biozol, Eching, Germany), CD86 (clone PO3; Biozol), CD40 (clone 3/23; Biozol), CD206 (clone MR5D3; AbD Serotec), CD204, CD163 (sc-33560; Santa Cruz Biotech), CD68 (AbD Serotec), and CD36 (clone 72-1; eBioscience, Frankfurt, Germany). Signs of apoptosis on these cells were detected by employing a kit (Annexin V-PE Apoptosis Detection Kit; BD Pharmingen), which localizes membrane phospholipid phosphatidylserine (PS). By simultaneously staining the nuclei by 7-AAD, intact cells (Annexin V−, 7-AAD−), early (Annexin V+, 7-AAD−), and late apoptotic cells were distinguished (Annexin V+, 7-AAD+). ²⁹  

To measure phagocytotic function, BM were seeded in polypropylene tubes, mixed at a ratio of 1:20 (1 macrophage per 20 zymosan particles, latex beads, or apoptotic splenocytes) with labeled zymosan A (Z2841, Invitrogen - Molecular Probes, Eugene, Oregon) and latex beads (L-4655, Sigma) or labeled with CSFE and AM cocultured splenocytes (AM-splenocytes) and then incubated at 37°C, 5% CO₂ for 3 hours. Cells were analyzed for an increase in green fluorescence due to phagocytosis with a cytometer (FACSCalibur, Becton Dickinson). 

Flow Cytometry Analysis of Inflammatory Cells in the Cornea

HSK corneas with AMT or TAR were incubated in RPMI medium at 4°C for 2 hours. The endothelia and Descemet's membrane of the corneas were removed, and eight corneas per group were pooled and digested for 2 hours at 37°C in 80 to 100 U/cornea collagenase type 1 (Sigma-Aldrich, Germany). The single corneal cells were then pretreated by using a 40 μm cell strainer. The flow-through was then washed and incubated with anti-CD16/CD32 (FCγIII/II receptor-block, BD Pharmingen), anti-CD45 (FITC, clone 30-F11; BD Pharmingen, Heidelberg, Germany), Annexin-5 (PE), 7-AAD (BD Pharmingen), and anti-F4/80 (Alexa-Fluor 647, clone CI:A3-1, AbD Serotec). Unstained specimen and isotype controls (BD Pharmingen) were used to exclude unspecific binding of antibody. The suspension was analyzed by flow cytometry (FACSCalibur, Becton Dickinson). ³⁰  

Proliferation Assay with 3H-Thymidine

Macrophage proliferation was assessed in the BM collected from uninfected animals, using 10⁴ cells per well with medium, lipopolysaccharides (LPS) 100 ng/mL, or IFN-γ (100 U/mL). 

Draining lymph node cells (DLN) obtained from HSV-1 infected mice (day 14 postinfection) were used for a 3H-thymidine uptake assay to measure lymphocyte proliferation with or without AM treated BM (ratio 1:10). ⁷  

MTT Viability Assay

To determine the effect of AM on the viability of BM, the cells (1 × 10⁶ BM/mL) were treated with medium (control) or AM (1.5 cm in diameter) for 48 hours. Some wells were additionally treated with LPS (100ng/mL) or IFN-γ (100 U/mL). To determine cell viability, MTT solution prepared by dissolving MTT (5 mg/mL) in RPMI 1640 was added and incubated for 3 hours. Afterward, cells and dye crystals were dissolved by adding N,N-dimethylformamide. Absorbance was measured at 570 nm (reference at 690 nm) in an ELISA reader (MRX-ELISA; Dynatech Laboratories, Sussex, UK). The results were expressed as optical density (OD). ²¹  

Ex Vivo Leukocyte Migration Assay

The ex vivo leukocyte migration assay was performed in an adapted fashion, as described previously. ³¹ Corneas were excised in a sterile environment, aided by a dissection microscope, from mice with HSK at 12 or 48 hours with TAR or AMT. After the endothelium and Descemet's membrane were removed, the remaining corneas were placed in a U-bottomed 96-well plate with the epithelium side down. CSFE-labeled BM (1 × 10⁵ BM per 50 μL) from naive BALB/c mice were seeded on the inner part of each corneal cup. The corneas (each group n = 6) were incubated with the CSFE-labeled BM at 37°C for 3 hours and then washed extensively in RPMI. The whole mounted corneas were then examined by using a fluorescence microscope (magnification, ×200) and the fluorescent cells in the corneas were counted. 

Preparation of Apoptotic Splenocytes

Apoptosis was induced in splenic cells from noninfected mice by cocultivating them with AM for 24 hours. Viable cells were determined by MTT test. ³² The cells were used separately for coculturing experiments and for the proliferation assay at a ratio of 1:20 (1 macrophage per 20 splenocytes). 

Oil Red O Staining

Oil Red O and Sudan Black B staining were performed to stain the intracellular triglyceride content. BM treated with medium or LPS (100 ng/mL) and IFN-γ (100 U/mL) were then photographed using optics microscopy (BX40; Olympus, Tokyo, Japan). ³³  

Arginase Assay

The arginase activity of macrophages and in HSK-corneal samples was determined using an arginase assay kit (DARG-200; BioAssay Systems, Köln, Germany) according to the manufacturers' description. 

Western Blot Analysis

Protein expression in corneal specimens and isolated BM were determined by using Western blot analysis. The same quantity of protein (15 μg) from each corneal specimen as detected by the Bradford method was electrophorized at 4°C under nonreducing conditions in SDS-PAGE 10% polyacrylamide gels (Biorad, München, Germany) for 2 hours at 120 V. Macrophages were lysed in lysate buffer (20 mM Tris-HCl, pH 7.5), 140 mM NaCl, 50 mg/mL deoxycholate, 0.1% SDS, 1% Triton X-100, 10% glycerol, 1 mM Na3VO4, 1 mM DTT, 1 mM PMSF, 1 μM pepstatin, and 10 μM leupeptin). 

Proteins were eletrophoretically transferred from the samples to the nitrocellulose membrane (Biorad) for 2 hours at 110 mA. Nonspecific binding sites were blocked with 5% (wt/vol) BSA in TBS-T buffer (50 mM Tris-HCl; pH 7.0; 0.15 M NaCl; and 0.05% Tween) at RT for 1 hour; membranes were then incubated with primary antibody overnight at 4°C with agitation. A nuclear factor (NF)-κB pathway kit (cat no: #9936; Cell Signaling Technology, Frankfurt, Germany) was used, containing antibodies directed against IkB kinase (IKK)a, IKKb, phospho-IKKα/β, NF-κB p65, phospho-NF-κB p65, IkBa, phospho-IkB-α, and β-actin as control. 

Statistical Analysis

Student's t-test was carried out to test the differences between the experimental groups with respect to cell numbers in the histologic and immunohistochemical staining, the leukocyte migration assay, the MTT viability assay, and the ELISA. One-way ANOVA with Tukey post hoc tests were applied to determine the difference between the proliferative responses. P < 0.05 was considered statistically significant. 

Results

AMT Improves Herpetic Stromal Keratitis

In agreement with a previous report, corneas with TAR contained many inflammatory cells and an abundance of viable polymorphonuclear neutrophil (PMN)-like cells. ¹⁷ In the corneas treated with AMT, fewer inflammatory cells were found and many PMN-like cells showed an abnormal, round morphology and blebbing. Groups of macrophage-like cells with pseudopodia directed to the PMN-like cells could also be found. ²¹  

AM Induces Apoptosis of Macrophages In Vitro and in HSK Corneas

In vitro studies revealed that RAW 264.7 macrophages activated with IFN-γ and cocultured with AM stromal matrix undergo apoptosis. ²² BM cultivated in L-cell conditioned medium for 5 days showed strong expression of the surface markers F4/80 as detected by flow cytometry, indicating that the cell suspension contained constantly > 90% F4/80+ cells. BM cocultured with AM and LPS/IFN-γ for 48 hours and stained with Hoechst staining reagent showed typical apoptotic features (e.g., fragmented nuclei), while BM cultured without AM showed no such morphology (Fig. 1B–E). Additionally, BM cell death increased after cocultivation with AM or AM with LPS (Table 1) as determined by flow cytometry. We also tested the cell membrane integrity of BM after cocultivation with AM by trypan blue exclusion. The results showed that cell membrane degradation was not induced by AM on BM in vitro even when the cells were activated with LPS or IFN-γ (data not shown). 

Table 1.

 

Induction of Apoptosis in BM by AM In Vitro

	Medium (%)	LPS (%)	IFN-γ (%)
Medium
Late apoptotic	60.6	56.2	66.9
Early apoptotic	18.6	11.9	18.7
Necrotic	2.6	1.4	2.1
AM*
Late apoptotic	74.8	64.8	66.7
Early apoptotic	15.2	25.6	16.5
Necrotic	1.3	0.3	2.7

*  The results of one experiment representative of two are shown.

We therefore hypothesized that F4/80+ macrophages in AMT-treated corneas would be composed of macrophages with apoptotic properties, whereas corneas with TAR would be composed mainly of viable macrophages. After 48 hours, we found that the corneas with HSK and AMT contained more F4/80+ cells displaying apoptotic features (positively stained for annexin V/7-AAD) than corneas treated with TAR (Fig. 1F). Therefore, HSK improvement after AMT is associated with an increased apoptotic death of macrophages. 

Enhanced Macrophage (BM) Migration into AMT-Treated Murine Corneas with HSK

We recently reported that MIP-2/CXCL-2 is upregulated in HSK corneas after AMT. ²¹ As MIP-2/CXCL-2 is known to induce PMN and macrophage infiltration in tissues, we addressed the question of whether AMT modulates macrophage infiltration into HSK corneas. To this end, groups of BALB/c mice with 4+ HSK received TAR or AMT. The corneas were excised and used for an ex vivo leukocyte migration assay (Fig. 2A). ³¹ The number of CSFE+ macrophages that invaded the HSK corneas was evaluated by immunofluorescence. After 48 hours significantly more CSFE-positive BM were found in corneas treated with AMT than in those treated with TAR (Fig. 2B). 

Figure 2.

 

BM infiltration into the HSV-1 infected cornea with HSK after AMT. Mice with 4+ HSK were treated with TAR or AMT. Mice were euthanized 48 hours later. (A) The corneas were collected, the endothelium and Descemet's membrane were removed, and the corneas were used, epithelium side down, for an ex vivo invasion assay of CSFE-labeled BM macrophages. ³¹ (B) Number of CSFE-positive cells in corneas with HSK and AMT or TAR. The results show an increase in CSFE+ BM in the HSK corneas 48 hours after AMT. Invasion of BM into corneas with TAR or AMT were not significantly increased 12 hours after AMT (data not shown). Representative experiments of two with six mice per group. (C) Analysis of the supernatants collected from macrophages after cocultivation with apoptotic splenocytes, AM, or apoptotic splenocytes and AM. The results show that macrophages produce increased levels of MIP-2/CXCL-2 and KC/CXCL-1 when cells were cocultivated with apoptotic splenocytes. AM slightly decreased the content of MIP-2/CXCL-2 or KC/CXC-1 when cocultured with BM and apoptotic cells. The data are from a representative experiment of two, with data expressed as mean ± SEM of triplicate wells (*P < 0.05). Med, Med + AM; Med + AC − AM +AC: statistically not significant.

It was previously demonstrated that the chemokine MIP-2/CXCL-2 supports invasion of PMN/macrophages into HSV-1-infected corneas ³⁴ and that macrophages are important producers of MIP-2/CXCL-2 when apoptotic cells are present. ³⁵ In our experiments BM produced high amounts of MIP-2/CXCL-2 and KC/CXCL-1 when they were cocultured with apoptotic splenocytes. BM with AM alone induced only a slight increase in KC/CXCL-1. Chemokine levels were only slightly lower in BM cultured with apoptotic cells together with AM than in BM incubated only with apoptotic cells (Fig. 2C). 

Taken together, these observations demonstrate that BM invasion is increased in HSK corneas treated with AMT. The chemokine data imply that paracrine factors released by apoptotic cells rather than factors liberated from AM support chemokine production in macrophages. 

Changes in Signaling Pathways After AM Treatment of Corneas with HSK and in BM Cocultured with AM

Innate immune responses in the cornea participate in the development of immunopathological conditions. ³⁶ ELISA analysis revealed that AMT reduced the levels of IL-6, IL-10, IL-12, and TNF-α in HSK corneas (Fig. 3A, * P < 0.05). Because levels of macrophage-related cytokines were lower in corneas with AMT, we tested if AM would also affect macrophage cytokine production in cell culture conditions. Accordingly, the production of TNF-α, IL-10, and IL-12 was also lower in AM-treated BM than in BM without AM. A decrease in IL-6, IL-10, IL-12, and TNF-α production was observed when BM plus AM were stimulated with LPS, or LPS and IFN-γ (Fig. 3B). 

Figure 3.

 

Changes in signaling pathways in corneas with HSK after AMT and in BM cocultured with AM. (A) Murine corneas with HSK with TAR or AMT (48 hours). AMT-treated mice showed a decreased amount of TNF-α, IL-6, IL-10, and IL-12 in the corneas. The data are from a representative experiment of two, with eight mice per group. (B) Analysis of the supernatants collected from BM after cocultivation with medium or AM. The results show that BM produced less IL-6, IL-10, IL-12, and TNF-α after stimulation with LPS or LPS and IFN-γ when BM were cocultivated with AM. Data show a representative experiment of two as mean ± SEM of triplicate wells. Pooled cornea samples (n = 8) were analyzed with antibodies against NF-κB. (C) AMT strongly reduces the NF-κB pathway in HSK cornea with AMT. Representative Western blot analysis results with eight mice per group in TAR and AMT (48 hours) treated HSK corneas. (D) BM coculturing experiments with AM indicate that AM reduced expression and phosphorylation of IKKα/β subunits, reduced P-IkBα, and decreased NF-κB and P-NF-κB. (E) F4/80+ BM were used for flow cytometry analysis using antibodies targeting P-NF-κB p65 (Ser536) (93H1), confirming decreased levels of P-NF-κB.

Our Western blot analyses show reduced expression of the NF-κB in HSK corneas treated for 48 hours with AMT. This was also noted for phospho-NF-κB, IKK-α, IKK-β, P-IKK-β, IkBα, and P-IkBα (Fig. 3C). 

In BM with AM we noted a decreased expression of IKKα, IKKβ, P-IKK α/β, P-IkBα, NF-κB, and P-NF-κB. The differences were even more pronounced when BM were treated also with LPS or IFN-γ. Less P-NF-κB was also found in BM by flow cytometric analysis after cocultivation with AM (Figs. 3D, 3E). 

Taken together, the results indicate that AM influences the corneal NF-κB pathways. Furthermore, AM decreases activation of the NF-κB pathway in BM, probably via an interaction of AM-related factors with the IKK subunits IKKα/β, resulting in suppressed phosphorylation of P-IkBα and decreased phosphorylation of NF-κB. Because activation of macrophages enhanced the effect of AM on macrophages, upstream activation elements (e.g., via toll-like receptor (TLR)4) might also contribute to this effect (Fig. 3D). 

Amniotic Membrane Reduces the Macrophage Accessory Function to DLN Cells

Previous studies have identified a close link between the IL-1, IL-6, or IL-12 secretion of macrophages and their accessory function to T lymphocytes. After exposing BM to AM, HSV-1-specific proliferation of cocultured DLN cells, IFN-γ content, and the autocrine T cell survival factor IL-2 decreased; this was also observed after pretreatment of BM with the HSV antigen (Figs. 4A, 4B). 

Figure 4.

 

Influence of AM cocultivation on BM costimulatory function to DLN cells. BM were cocultured with medium or AM and used as antigen-presenting cells to DLN cells obtained from HSV-1 infected mice (ratio BM to DLN cells: 1:20). (A) The proliferative response of DLN cells was decreased when the macrophages were cocultured with amniotic membrane beforehand and this correlated with a decrease in IL-2 and IFN-γ in the supernatant. (B) A similar result was obtained when BM were also cultured with HSV-1 antigen. Data are expressed as mean ± SEM (*P < 0.05) of six wells. (C) Surface molecule expression on macrophages cocultured with medium or AM. F4/80+ BM with AM cocultivation expressed less CD80, CD86, CD40, and CD69 on the cell surface compared with BM without AM. Data show one representative experiment of three with similar results, counting 2 × 10⁴ cells in each sample.

Additionally, the expression of costimulatory molecules CD80, CD86, and CD40 was decreased on the cell surface of AM-treated BM (Fig. 4C), even when BM were stimulated with LPS or IFN-γ. AM also decreased expression of the activation marker CD69. These results indicate that the activation stage of BM is directly modulated by AM. The presence of other corneal cells, including keratocytes, is not required for this effect of AM on macrophages. Collectively, these findings suggest that the accessory function of macrophages to isolated DLN cells is reduced after contact with AM. 

Marker for Macrophage Alternative Activation on BM Cocultured with Amniotic Membrane

We then investigated whether BM express altered mannose, scavenger, or macrosialin receptors on the cell surface, which are important for diverse phagocytic functions but which are also known to be indicators of alternative activation (CD206, CD204, CD163) of macrophages. The flow cytometry data indicate that CD206 and CD204 expression was increased after AM cocultivation when cells were stimulated with LPS or IFN-γ (Fig. 5). Expression of the markers CD163 and CD68 was increased on AM macrophages in general. Collectively, in BM AM induces an increased expression of cell surface markers linked with characteristics of alternatively activated macrophages functions. 

Figure 5.

 

Influence of AM cocultivation of BM on expression of CD206, CD204, CD163, and CD68. Surface molecule expression on BM cocultured with AM or with medium only. F4/80+ BM express more CD206 after LPS or IFN-γ treatment and AM. CD204, CD163, and CD68 were increased on BM treated with AM and medium, LPS, and IFN-γ. Data show a representative experiment of three for each antigen with similar results, counting 2 × 10⁴ cells in each sample (*P < 0.05).

Amniotic Membrane Decreased the Phagocytic Function of Macrophages

Next, we addressed the question of whether AM also affects the phagocytic function of BM. Apoptotic cell phagocytosis of BM treated with AM was markedly lower than in macrophages without prior AM treatment, even when the BM were activated with LPS or IFN-γ (Fig. 6A). A similar result was also observed when FITC-zymosan or FITC latex beads were used for the assay (Fig. 6B). These results indicated that phagocytosis of apoptotic cells or large particles in BM is inhibited by AM. 

Figure 6.

 

Influence of AM cocultivation of BM phagocytic function. F4/80+ BM were cocultured with medium or AM with or without LPS or IFN-γ. (A) CSFE-labeled apoptotic splenocytes were used for phagocytic analysis of BM 48 hours later (ratio 1 BM to 20 splenocytes). (B) In some of the experiments FITC-zymosan or FITC latex beads were used instead of CSFE splenocytes. The results indicate that macrophage function of taking up particles is reduced after cocultivation with AM. Data show a representative experiment of two (*P < 0.05).

AM Cocultivation Induces Expression Linked with Lipid Metabolism

Surface expression of CD206, CD204, CD163, and CD68 was modulated by AM and these surface molecules were known to mediate uptake of lipids or lipid acids; therefore, we performed Oil Red O and Sudan Black B staining to determine whether BM accelerates lipid vesicle content when cocultured with AM. After cultivating BM with medium, LPS, or IFN-γ, neutral lipids, as measured by Oil Red O, were not detected intracellularly, while after cocultivating BM with AM, high amounts of lipids could be found in the cytoplasm (Figs. 7A, 7B). The lipid load increased further on stimulation with either LPS or IFN-γ (data not shown) or both. This was also found by Sudan Black B staining. These results indicate that AM cocultivation with BM promotes strong lipid accumulation in the cells. 

Figure 7.

 

AM induces neutral lipid accumulation in BM and induces gene expression linked with lipid metabolism. (A) Air-dried section of BM with or without AM were stained with Oil Red-O. Neutral lipid accumulation (red color) was detected in BM cocultured with AM. Activation of BM with LPS and IFN-γ significantly increased Oil Red-O staining in AM-treated BM (arrows). (B) Accumulation of neutral lipids in BM with AM could also be found by staining with Sudan Black B (arrows). (C) Air-dried sections of BM cocultured with AM were stained with mAb targeting PPAR-γ. Red staining indicates PPAR-γ-positive macrophages (arrows). (D) Arginase activity was measured in cell lysate obtained from BM after AM treatment. The results show that arginase activitiy is increased after AM cocultivation. (E) Increased expression of PPAR-γ, arginase 1, and arginase 2 in corneas with HSK when corneas were treated with AMT. A representative experiment of two is shown for (A–E). (F) Increased arginase activity in corneas with HSK and AMT compared with corneas with HSK with TAR. The data are from a representative experiment of two with twelve mice per group (P < 0.005).

Expression of the PPAR-γ in BM Cocultured with AM and in HSK Corneas Treated with AMT

PPAR-γ is a nuclear receptor which regulates the expression of genes linked with lipid metabolism. It is well documented that uptake and internalization of oxidized lipoproteins activate PPAR-γ expression and activity in macrophages, which promotes macrophage differentiation and uptake of oxidized LDL. ^37,38 Immunohistochemical staining in our study using antibodies targeting PPAR-γ indicate that expression is enhanced in BM cultured with AM together with medium, LPS, IFN-γ, or both (Fig. 7C). Western blot analyses show that the expression of PPAR-γ1/3 and PPAR-γ2 was higher in HSK corneas treated with AM (Fig. 7E). An increased expression of PPAR-γ could also be found in BM cocultured with AM as determined by flow cytometric observation (Fig. 8C). 

Figure 8.

 

Influence of apoptotic cells on BM proliferation and survival. BM treated with AM showed decreased survival and proliferative response. (A) Proliferative response of BM with medium or AM. In some settings, the BM were also cocultured with apoptotic splenocytes. The results show that AM downregulation of proliferative response could be restored when macrophages were cocultured with necrotic (data not shown) or apoptotic splenocytes (P < 0.05). (B) The MTT conversion assay after cocultivation of BM with medium, apoptotic cells, AM, or both. The MTT test demonstrated that apoptotic cells increased macrophage survival with or without further activation with LPS and IFN-γ. (C, D) Flow cytometric analysis of PPAR-γ and CD36 expression in BM with or without AM or apoptotic cells. The results suggest that PPAR-γ and CD36 expression was decreased after cocultivation of BM with AM and apoptotic cells. The data are from a representative experiment of two (for A–D).

These results show that AM induces mechanism of action in BM or in corneas with HSK linked with increased PPAR-γ expression. 

Amniotic Membrane Increase of Arginase Expression and Activity in Corneas with HSK and of BM

It was previously reported that PPAR-γ/δ are also able to induce arginase 1, ³⁹ which is an essential suppressive mediator of alternatively activated macrophages. ⁴⁰ To examine the activity of arginase, BM were cultured with AM together with medium only, LPS, or IFN-γ. The results showed that the arginase activity in BM was slightly higher after coculture with AM (Fig. 7D). Furthermore, Western blot analysis findings indicate that arginase 1 and 2 expression was higher in AMT corneas (Fig. 7E). Additionally, the activity of arginase was significantly higher in HSK corneas treated with AM than in the control group of corneas treated with TAR (Fig. 7F). Taken together, the data show that AM increases activity of arginase in macrophages and in corneas with AMT and induces enhanced expression of arginase 1 and 2 in corneas with AMT. 

Presence of Apoptotic Cells in the Amniotic Membrane Environment Controls Macrophage Proliferation and Survival

We then investigated whether AM directly influences the proliferative response and viability of BM. Indeed, AM profoundly reduced macrophage proliferation and viability, as measured by 3H-thymidine uptake and MTT test, and this was further decreased when the cells were activated with LPS or IFN-γ. In contrast, if apoptotic splenocytes were added, the viability and proliferative response of AM-treated macrophages increased (Figs. 8A, 8B). 

According to our flow cytometric observations the expression of the scavenger receptor CD36 (scavenger receptor class B) and PPAR-γ decreased when BM were cocultured together with AM and apoptotic cells (Figs. 8C, 8D). This result indicates that uptake of AM-related lipids or apoptotic cells may be lower when both factors are present. 

Taken together, these results suggest that AM reduces proliferation and viability of macrophages. However, in the presence of dead cell bodies, macrophages are able to upregulate proliferation and viability, likely by a competing mechanism of AM-related lipids or apoptotic cells to PPAR-γ and CD36, and also other factors that are important for phagocytosis, ultimately avoiding apoptotic cell death of macrophages (Figs. 8C, 8D). 

Discussion

AMT can be used to rapidly improve the severity of murine and human necrotizing HSK, as shown previously. ^17,41 PMN and lymphocyte apoptosis can be observed in HSK corneas after human AMT. ^21,32 Our present study now shows that macrophage apoptosis also increased on AMT treatment in murine HSK corneas. This could be reproduced in vitro. A rapid cell apoptosis was previously described in RAW 264.7 macrophages on AM treatment and stimulation. ^22,23 Macrophage cell death by AMT may be important to downregulate inflammation in the cornea. 

However, apoptotic cells must be removed, for example, by phagocytosis during the healing process, to avoid secondary necrosis. A variety of chemoattractants for phagocytic cells secreted by apoptotic cells have been described. IL-8 mediates leukocyte infiltration into inflamed tissues, with endothelial monocyte-activating polypeptide II (EMAP II), a fragment of aminoacyl-tRNA synthetase, mimicking its action. ⁴² Apoptotic cells may also secrete increasing amounts of thrombospondin-1 or release the lipid mediator sphingosine-1-phosphate for the recognition of apoptotic cells or to mediate macrophage migration. ^43,44 Apoptotic cells are also involved in resolving inflammation by secreting lactoferrin, which further inhibits PMN migration in vitro and in vivo. ⁴⁵  

In our ex vivo migration experiments an enhanced BM migration into AMT-treated murine HSK corneas was observed. This was also associated with an increase in MIP-2/CXCL-2 in the cornea, while the expression of other chemokines was downregulated. ²¹ The assay may therefore support our observation that macrophages migrate into the HSK corneas with AMT, likewise to clear the corneas from apoptotic cell bodies. ²¹  

We demonstrated further that apoptotic cells are able to induce MIP-2/CXCL-2 and KC/CXCL-1 in BM, and with AM we found a slightly decreasing chemokine expression, suggesting that apoptotic splenocytes treated with AM in vitro induce chemokine expression in macrophages, AM counteracting this effect only slightly. 

We further noted a significant downregulation of the levels of the proinflammatory cytokines TNF-α, IL-6, and IL-12 and of the anti-inflammatory cytokine IL-10 in the murine corneas with HSK after AMT and also in supernatants collected from BM after coculture with AM. The decreased cytokine content in the corneas and in BM may represent an important action mechanism of AM, as these cytokines are critical for the pathogenesis of HSK and for the outcome of infection. ⁹ After AMT we also found that NF-κB expression was highly downregulated in the corneal samples and in BM in vitro. As the corneal experiments were performed with the entire tissue, the observations reflect the whole corneal environment, and not specifically the macrophage fraction. Expression of IKK-α, IKK-β, and p65 (RelA) subunit of NF-κB was also downregulated in IFN-γ-activated macrophages cultured on AM as shown previously. ²² There is evidence that NF-κB is produced by corneal epithelial cells after HSV-1 infection of the cornea, and this was accompanied by transcriptional expression of IL-6, IL-8, tumor necrosis factor (TNF)-α, and interferon (IFN)-β. ⁴⁶ NF-κB has been shown to be an important inhibitor of pathogen-induced apoptosis in macrophages in vitro. ⁴⁷ NF-κB plays an important role in activating inflammatory and innate immune responses. ^48,49  

Interestingly, AM impaired BM in functioning as accessory cells to DLN cells. In our study, we noted that the expression of CD80, CD86, CD40, and CD69 in BM cocultured with AM was decreased, suggesting that AM treatment reduced the activation and costimulatory function of macrophages. Furthermore, proliferation of DLN cells from HSV-1-infected animals was decreased and lower levels of the autocrine survival factor IL-2 and of IFN-γ were measured in the supernatants when cultured with AM-treated BM compared with BM. Therefore AM may directly influence activation and survival of T lymphocytes as shown previously. ³² The present data show that AM may also influence activation of T lymphocytes via the function of macrophages. 

CD80, CD86, and CD40 costimulate T cells during antigen presentation by APC. Previous studies have demonstrated that CD40 receptor activation in T lymphocytes in vitro induces significant expression of bcl-xL, a potent antiapoptotic member of the Bcl-familiy. ⁵⁰ Furthermore, CD40 receptor stimulation induces multiple other signaling pathways, including NF-κB. ⁵¹ Our results are also in agreement with preceding observations that distinct macrophage populations are able to produce components (e.g., from the extracellular matrix) that exert indirect regulatory effects on the immune response by producing cytokines and suppressing clonal expansion of neighboring lymphocytes. ⁵²  

We found that the expression of CD206, CD204, CD163, and CD68 on BM was generally increased after cocultivation with AM; this effect was even stronger after further stimulation with LPS or IFN-γ. CD206 is upregulated on IL-4 stimulation, which introduced the concept of alternative activation of macrophages. ⁵³ The scavenger receptor CD204 and the haptoglobin-hemoglobin scavenger receptor CD163 were suggested, together with CD206, to be other markers for alternatively activated macrophages. ^54,55 The mouse antigen CD68 mediates phagocytosis of oxidized low density lipoproteins. CD36 was identified as an oxidized LDL receptor, but is also thought to be implicated in cell adhesion, phagocytosis of apoptotic cells, and metabolism of long-chain fatty acids. In line with increased expression of these surface receptors, we found lipid vesicle accumulation inside BM cocultured with AM, as observed by Oil Red O and Sudan Black B staining. However, phagocytosis of dead cells, latex beads, or zymosan was decreased after cocultivation with AM, which may be related to these lipid vesicles, probably via a competing uptake mechanism. Indeed, previous reports have shown that lipid accumulation can also cause macrophage cell death. In mouse peritoneal macrophages oxLDL initiates an apoptotic program ⁵⁶ while human macrophages die by necrosis. ⁵⁷  

Peroxisome proliferator-activated receptor (PPAR) has been found to regulate diverse aspects of lipid metabolism, including fatty acid oxidation, fat cell development, lipoprotein metabolism, and glucose homeostasis. Our results now show that PPAR-γ expression is higher after treating murine corneas with AMT, or when BM were treated with AM. 

An earlier study showed that transient transfection ligand activation of PPAR-γ results in apoptosis induction of unactivated differentiated macrophages by negatively interfering with the antiapoptotic NF-κB signaling pathway. ⁵⁸ The receptor is also known to regulate the behavior of noninflammatory cells (i.e., fibrogenic reaction or cell proliferation during wound healing). ⁵⁹ PPAR are ligand-dependent transcription factors that heterodimerize with the retinoid X receptor (RXR). ⁶⁰ Early studies showed that PPAR-γ promotes macrophage gene expression and uptake of oxLDL. ³⁸ PPAR-γ signal suppresses the inflammatory reaction by immune cells, including macrophages, in vitro and also has therapeutic effects. ⁵⁹ It has previously been shown that PPAR-γ overexpression suppresses the fibrinogenic reaction in cultured mouse ocular fibroblasts and macrophages by inhibiting nuclear translocation of the phosphorylated smads, and consequently prevented excess scarring in an alkali-burned mouse cornea. ⁶¹ PPAR-γ activation also promotes infiltration of alternatively activated macrophages into adipose tissue, and this was associated with downregulation of classically activated macrophage markers, including IL-18, and characteristically with upregulation of alternatively activated macrophages (arginase 1, IL-10). ⁶² PPAR-γ increased arginase expression and inhibited expression of proinflammatory genes, including cytokines and inducible nitric oxide synthase (iNOS). ⁶³ Taken together, improvement in inflammation and wound healing after AMT is associated with PPAR-γ upregulation. 

Concerning the placenta, PPARs, particularly PPAR-γ, are essential for multiple physiological functions of the trophoblastic and amniotic parts, which are important for fetal protection. The pathophysiology of gestational diseases often involves PPAR pathways (e.g., chorioamnionitis, gestational diabetes, etc.). Indeed, the term labor is associated with the production of proinflammatory cytokines (e.g., IL-1β, IL-6, IL-8, IL-10, and TNF-α), which are known to induce uterine contractions. Natural ligands of PPAR-γ have been demonstrated to inhibit the secretion of IL-6, IL-8, and TNF-α in amnion and chorion, ⁶⁴ demonstrating the role of PPARs in regulating the inflammatory response in human gestational tissues and cells. ^65,66  

We observed that the biologic activity of arginase was increased in cultured BM cocultured with AM and in HSK corneas treated with AMT. An increased expression of arginase 1 and 2 was observed in HSK corneas after AMT. Arginases catalyze the hydrolysis of L-arginine to l-ornithine and urea to generate l-proline, which serves as a substrate for collagen synthesis and polyamines to stimulate cell proliferation. Both isoforms are constitutively expressed in murine macrophages. ⁶⁷ Former studies demonstrated that arginase 1 is induced by Th2-derived cytokines in macrophages ⁶⁸ ; it is considered to be one of the hallmarks of alternative macrophage activation. ⁶⁹ Macrophages have been previously classified in the M1 and M2 lineages. The M1 designation was used for the classical activated macrophages (for host defense), and M2 designation for alternatively activated macrophages. ⁶⁹ M2 macrophages encompass cells with dramatic differences in their biochemistry and physiology, namely for immune regulation and wound healing. ⁵  

Classically activated macrophages are induced by TLR and IFN-γ signaling, and produce high IL-12, but low IL-10. The regulatory M2 macrophages can be induced by IL-10, apoptotic cells, immune complexes, and diverse tumors to produce low IL-12, but high IL-10 levels. Alternatively activated wound healing macrophages are induced by IL-4 or other stimuli. They produce large amounts of arginase-1, but low IL-12 and IL-10. Analogous to wound healing macrophages, macrophages of healthy (nonobese humans) in the adipose tissue produce little cytokine content, but express high amounts of arginase. ⁷⁰ They also have adipocyte function and maintain sensitivity to insulin. ⁷⁰ The nuclear receptor PPAR-γ seems to be an important regulator of this macrophage phenotype; there have been several reports correlating the alternative activation state of macrophages with PPAR-γ activation and macrophage cell death. ⁷¹ Therefore, we assume that different macrophage subpopulations may be present during the course of HSK, and may be influenced by AM. 

We also investigated how AM on BM affected proliferative response and survival as determined by uptake of 3H+ thymidine and MTT test. The results indicate that AM downregulates the BM proliferative response and survival and this was enhanced by macrophage activation. This downregulation of BM proliferation and survival could be reversed, in part, by the presence of necrotic (data not shown) or apoptotic cells. 

Previous studies have found that phagocytosis of apoptotic cells leads to secretion of growth and autocrine survival factors by phagocytes. ⁷² Furthermore macrophages cocultured with apoptotic cells may produce diverse chemokines on coculturing with apoptotic cells. ^73,74 Soluble factors, released by the apoptotic cells are responsible for the effects. ⁷⁵ When macrophages take up apoptotic cell bodies they may be induced to M2/regulatory macrophages to produce anti-inflammatory cytokines such as IL-10 and/or TGF-β. ⁷⁶  

Although a rapid clearance of apoptotic cell bodies is important to prevent the release of potentially cytotoxic or antigenic content into the extracellular matrix to prevent inflammation and tissue injury, we observed a decrease of efferocytosis, when macrophages and AM were cocultured with apoptotic cells. This finding was supported by the fact that the marker CD36, a well known surface molecule that facilitates nonopsonic phagocytosis, was decreased, when amniotic membrane and apoptotic cells were both present. Additionally, PPAR-γ was decreased in macrophages cocultured with AM and apoptotic cells. 

It may be speculated that lipid factors released from the amniotic membrane are in competition with apoptotic cells for their interaction with macrophages. They may share a similar surface molecule or pathway. The reduction of CD36 on the macrophage cell surface could be important to prolong their survival. 

Taken together, our present results indicate that alternative activation phenotype and apoptosis of macrophages are probably among the important effects involved in the anti-inflammatory action of AM, probably by engaging lipid metabolism and activating the PPAR-γ pathway. Our study indicates that macrophages have some survival advantages in environment with AM compared with other leukocytes (PMN or lymphocytes) that is amplified by the presence of apoptotic cells that also support the invasion of new macrophages. 

Footnotes

 Supported by DFG Grants Ba 2248/1-1, Ba 2248/1-2, He 1877/12-2; and the Ernst and Berta Grimmke Foundation.

Footnotes

 Disclosure: D. Bauer, None; M. Hennig, None; S. Wasmuth, None; H. Baehler, None; M. Busch, None; K.-P. Steuhl, None; S. Thanos, None; A. Heiligenhaus, None