Macrophage pro-inflammatory cytokine secretion is enhanced following interaction with autologous platelets

Human monocytes were isolated and cultured using techniques similar to those previously described [30, 31]. Briefly, blood from healthy human donors was collected into citrate and peripheral blood mononuclear cells (PBMCs) were isolated by using Lymphoprep (Accurate Chemical) according to the manufacturer's instructions. Monocytes were further isolated by plating the PBMCs on gelatin-coated tissue culture flasks for 45 min at 37°C followed by 10 washes with phosphate buffered saline (PBS) to remove non-adherent lymphocytes. Monocytes were then detached from the flasks by incubation in 10 mM EDTA for 2 min at 37°C. Monocytes (250,000 in 500 μl volume) were then plated in 24-well plates overnight in RPMI 1640 supplemented with 10% fetal bovine serum (FBS) and 10 ng/ml recombinant human GM-CSF (R&D Systems). Monocytes were plated on glass coverslips for scanning electron microscopy (SEM) analysis and plastic tissue-culture plates for TEM analysis and phagocytosis experiments. Media was changed on day 2 and day 5, and after 7 days of culture the MDMs were used for phagocytosis and cytokine assays.

Platelets were isolated from whole blood collected into acid-citrate-dextrose (ACD) from healthy human donors and centrifuged for 15 min at 500 × g to generate platelet-rich plasma (PRP). PRP was pelleted by centrifugation for 10 min at 800 × g and the platelet pellet subsequently washed 2 times in citrated saline (pH 6.8). A portion of the platelet samples were degranulated by incubating 1 ml of platelets (250,000/μl in citrated saline) with 10 μl of 10 μM calcium ionophore A21387 (Sigma) for 15 min with rocking at room temperature, followed by three centrifugational washes with citrated saline. For phagocytosis experiments, platelets were fluorescently labeled with Cell Tracker Green CMFDA (Invitrogen) as previously described [32]. After the final wash, and prior to use in phagocytosis assays, platelets were resuspended in warm serum-free RPMI for 15 min at 37°C.

For analysis of surface P-selectin and phosphatidylserine, platelets (250,000/μl) were first incubated in either citrated saline or serum-free RMPI media for 1 hr at 37°C. A portion of the platelets were activated by including thrombin (King Pharmaceuticals, 1 U/ml final concentration) in the incubation reaction. To detect surface expression of activation markers in each platelet treatment group, a 10 μl aliquot of platelets was stained with either FITC-anti-CD62P (Biolegend) or FITC-Annexin-V (Biolegend) for 30 min at room temperature, after which the cells were fixed and analyzed immediately by flow cytometry. Flow cytometry analysis was performed using a CyAn flow cytometer (Beckman Coulter) and the Summit analysis software.

Dexamethasone-loaded platelets were prepared by incubating 1 ml of platelets (250,000/ul) in citrated saline with 5 ul of dexamethasone (Sigma, 10 mM in DMSO) for 15 min on a rocker at room temperature. Platelets were then washed three times with citrated saline to remove unbound dexamethasone. For subsequent experiments, 5 × 10⁶ dex-platelets were added to each well of macrophages in a 24-well plate.

To prepare apoptotic cells, PBMCs were isolated as above, and following monocyte-adherence to gelatin-coated flasks the non-adherent lymphocytes were collected. Cells were rendered apoptotic (Annexin-V positive) by UV-irradiation for 10 min followed by overnight incubation in RPMI + 10% FBS at 37°C + 5% CO2.

Thirty minutes prior to the start of each experiment, 7-day old MDMs were washed 3 times with PBS and incubated with 500 μl fresh RPMI media. In some experiments the media was supplemented with 10% autologous human serum. A 25 μl aliquot of fluorescently labeled platelets (250,000/μl) was added to each well of macrophages. Platelets and macrophages were co-incubated for 45 min.

For SEM analysis, the co-cultures were washed once with PBS and fixed in 2% paraformaldehyde and 0.5% glutaraldehyde. The samples were then processed as previously described [33] and examined using a Cambridge S200 scanning electron microscope at 20 kV. For TEM analysis, the co-cultures were washed 5 times with PBS and fixed in 2% paraformaldehyde, processed as previously described [34], and examined using a Leo EM 910 transmission electron microscope. For flow cytometric analysis, warm trypsin/EDTA was then added to the macrophages to remove adherent platelets (confirmed by microscopy) and cells were incubated 15 min at 37°C. Macrophages were then collected and fixed in 1% cold paraformaldehyde and analyzed using a CyAn flow cytometer (Beckman Coulter) and the Summit analysis software. Data is expressed as the percentage of FL1-positive macrophages in a given collection of 10,000 macrophages. Data shown represent the average of at least 3 independent experiments and for each experiment 10,000 macrophages were analyzed.

Latrunculin (Sigma, 1 μg/ml final concentration), used as a pan-phagocytosis inhibitor, was added to a portion of macrophage-containing wells 30 min prior to addition of platelets. Fucoidan (Sigma) was added to macrophages at a final concentration of 250 μg/ml 30 min prior to addition of platelets.

Each well of MDMs was washed 3 times with PBS and incubated with fresh RPMI + 10% autologous human serum. Activated, degranulated, or dexamethasone-loaded platelets (5 × 10⁶) were added to each well in addition to LPS (100 ng/ml). Some samples also received dexamethasone alone at a final concentration of 1 μM. After 24 hrs, supernatants were collected, spun 10 min at 14,000 g, and frozen at -80°C. Cytokines were measured by ELISA using capture and detection antibodies (eBioscience) per the manufacturer's instructions. Cytokines were measured in duplicate and averaged. The amount of protein secreted was normalized to the amount secreted by macrophages treated with LPS alone. Each experiment was performed at least 3 times using 3 different MDM donors. In each experiment, the platelets added were from the same donor as the MDMs. Treatment groups were compared using an unpaired t-test.

To examine the interaction between human MDMs (hMDMs) and autologous platelets, we utilized an in vitro co-culture system consisting of 7-day old hMDMs to which we added freshly isolated autologous platelets. The use of autologous platelets excludes the possibility that platelet-macrophage interactions are the result of an immune response triggered by the recognition of platelets as 'foreign.' The hMDMs and platelets were first co-cultured in serum-free RPMI media and examined by SEM and TEM at various time points to visualize the interaction between these two different cell types. As shown in Figure 1A, we observed platelets interacting with hMDMs during the first hour of co-culture. Platelets near the macrophages became entrapped by a network of macrophage filopodia, and although the macrophages were firmly attached to the coverslip and did not migrate, they appeared to direct groups of filopodia in the direction of nearby platelets that had settled on the dish. Visualization of these cultures suggests that an interaction between human macrophages and autologous activated platelets (AAPs) occurs in vitro, and that it occurs in the absence of serum proteins.

Platelet phagocytosis by the macrophages in our co-culture system was subsequently confirmed by TEM and flow cytometry. Macrophages that were co-incubated with AAPs for one hour developed vacuoles, many of which contained contents that were the same size and shape as platelets (Figure 1B). These phagocytic vacuoles did not appear in control macrophages cultured in the absence of platelets. We then quantified phagocytosis by flow cytometric analysis of macrophage fluorescence after co-incubation with fluorescently labeled platelets and removal of adherent platelets with trypsin. When freshly isolated platelets were incubated in serum-free RPMI media and added in excess to 7-day old macrophages, approximately 50% of the macrophages internalized at least one platelet within 45 min (Figure 2). As expected, pretreatment of the hMDMs with the actin inhibitor latrunculin almost completely blocked phagocytosis, confirming the role of actin polymerization that occurs in all cases of phagocytosis (Figure 2A). The amount of phagocytosis increased if the platelets were pre-stimulated with thrombin (Figure 2B), and was significantly inhibited in the presence of fucoidan, a known competitive inhibitor to Scavenger Receptors [35, 36] (Figure 2C). The presence of 10% autologous human serum had no significant effect on phagocytosis, which excludes the possibility that the platelet-macrophage interaction requires a soluble serum-bound "bridging" molecule. Together these results suggest that phagocytosis of platelets correlates with platelet activation, and that macrophage phagocytosis of autologous platelets may be mediated, at least in part, by scavenger receptors.

We also used flow cytometry to more accurately examine the level of platelet activation in different culture conditions. Platelets were analyzed for expression of P-selectin, an alpha granule component expressed during early platelet activation, and phosphatidylserine, a membrane lipid exposed on the surface of completely (and irreversibly) activated platelets. Incubation in serum-free media alone for 1 hour resulted in an approximately ten-fold increase in P-selectin expression but did not induce surface expression of phosphatidylserine (Table 1). Treatment of platelets with thrombin, known to cause complete degranulation and irreversible platelet activation [23, 37], resulted in even higher levels of P-selectin and also increased surface expression of phosphatidylserine (Table 1). Interestingly, the data in Figures 1 and 2 showing phagocytosis of platelets incubated only in RPMI media suggests that only partial platelet activation, in the absence of complete degranulation or phosphatidylserine exposure, is sufficient to trigger phagocytosis. Although phagocytosis was enhanced when the platelets did express phosphatidylserine, we conclude that surface exposure of phosphatidylserine is not an absolute requirement for phagocytosis of platelets.

The hypothesis that macrophage phagocytosis of activated platelets results in an inflammatory response that differs from the response following phagocytosis of apoptotic cells was tested by measuring the secretion of cytokines following addition of platelets or apoptotic cells to LPS-stimulated hMDMs. Autologous platelets in two different activation states were used in the co-culture experiments: platelets that were "partially activated" (with surface exposure of measurable quantities of P-selectin and CD40L) by preparing in serum-free media or "irreversibly activated" (with phosphatidylserine exposure in addition to P-selectin) by treatment with the calcium ionophore A23187 [37].

The inflammatory response of the hMDMs was assessed by measuring the levels of TNF-α, IL-6, and IL-23 after incubation with autologous primed platelets, autologous activated platelets, or control apoptotic leukocytes in the presence of LPS for 24 hours. As an additional control, we analyzed "platelets-only" cultures using the same media and incubation times as the platelet-macrophage co-cultures and were unable to detect any TNF-α, IL-6, or IL-23 in platelets alone (with or without LPS, data not shown). We therefore conclude that the cytokines secreted in this system are macrophage-derived, and in each experiment the cytokine levels were normalized to the amount of cytokine secreted by hMDMs incubated with LPS alone.

When compared to LPS stimulation alone, macrophage co-incubation with apoptotic cells inhibited LPS-induced secretion of all three pro-inflammatory cytokines (Figure 3A, grey bars). However, co-incubation with primed or activated platelets enhanced macrophage secretion of TNF-α, IL-6, and IL-23. Induction of pro-inflammatory cytokines in the presence of platelets was 20-60% higher than the levels obtained by LPS treatment alone. Furthermore, the macrophage cytokine secretion was enhanced to a similar degree after co-incubation with both partially activated and degranulated platelets (Figure 3A). These data suggest activated platelets enhance LPS-induced macrophage cytokine secretion even when they present phosphatidylserine to the macrophage.

Based on the knowledge that platelets can bind glucocorticoids via glucocorticoid receptors [38], we tested the hypothesis that glucocorticoid-bound platelets would be less inflammatory than platelets that are activated, but otherwise unmodified. Platelets were incubated with dexamethasone, then unbound glucocorticoid was removed by washing the platelets in citrated saline. Although it is unlikely that all of the dexamethasone becomes bound to the platelets, complete retention of the dexamethasone would yield a final concentration in the DEX-plt stock of 50 μM and a final (effective) concentration in the platelet-macrophage co-cultures of 1 μM. As shown in Figure 3B (striped bars), the levels of cytokines produced after co-culture with dexamethasone-loaded platelets were inhibited to 30-50% of the levels produced by stimulation with LPS alone. The dexamethasone-loaded platelets had a similar effect on cytokine secretion as 1 μM dexamethasone alone. These results indicate that the pro-inflammatory platelet effect on macrophage activation can be reversed by pre-loading the platelets with glucocorticoids.