Richard P. Phipps, University of Rochester School of Medicine and Dentistry, Department of Environmental Medicine, 601 Elmwood Ave, Box 850, Rochester, NY 14642, USA. Tel.: +1 585 275 8326; fax: +1 585 506 0239. E-mail: email@example.com
Summary. Background: Platelet production is an intricate process that is poorly understood. Recently, we demonstrated that the natural peroxisome proliferator-activated receptor gamma (PPARγ) ligand, 15-deoxy-Δ12,14 prostaglandin J2 (15d-PGJ2), augments platelet numbers by increasing platelet release from megakaryocytes through the induction of reactive oxygen species (ROS). 15d-PGJ2 can exert effects independent of PPARγ, such as increasing oxidative stress. Heme oxygenase-1 (HO-1) is a potent antioxidant and may influence platelet production. Objectives: To further investigate the influence of 15d-PGJ2 on megakaryocytes and to understand whether HO-1 plays a role in platelet production. Methods: Meg-01 cells (a primary megakaryoblastic cell line) and primary human megakaryocytes derived from cord blood were used to examine the effects of 15d-PGJ2 on HO-1 expression in megakaryocytes and their daughter platelets. The role of HO-1 activity in thrombopoiesis was studied using established in vitro models of platelet production. Results and conclusions: 15d-PGJ2 potently induced HO-1 protein expression in Meg-01 cells and primary human megakaryocytes. The platelets produced from these megakaryocytes also expressed elevated levels of HO-1. 15d-PGJ2-induced HO-1 was independent of PPARγ, but could be replicated using other electrophilic prostaglandins, suggesting that the electrophilic properties of 15d-PGJ2 were important for HO-1 induction. Interestingly, inhibiting HO-1 activity enhanced ROS generation and augmented 15d-PGJ2-induced platelet production, which could be attenuated by antioxidants. These new data reveal that HO-1 negatively regulates thrombopoiesis by inhibiting ROS.
The regulation of thrombopoiesis is poorly understood, but is critically important in maintaining the normal number of properly functioning platelets for hemostasis. We previously reported that 15-deoxy-Δ12,14 prostaglandin J2 (15d-PGJ2) enhances platelet production from megakaryocytes by inducing reactive oxygen species (ROS) . 15d-PGJ2 is a potent natural peroxisome proliferator-activated receptor gamma (PPARγ) ligand that is formed spontaneously from prostaglandin D2 (PGD2) [2,3]. PPARγ is a member of the nuclear hormone superfamily that heterodimerizes with the Retinoid X Receptor (RXR) to induce the expression of genes important for lipid and glucose metabolism, as well as inflammation . We recently demonstrated that PPARγ is expressed in megakaryocytes and platelets and that treatment of platelets with PPARγ ligands such as 15d-PGJ2 attenuated platelet activation .
Certain PPARγ ligands can induce heme oxygenase-1 (HO-1) [6,7]. Heme oxygenases (HO) are believed to protect against the toxicity of free heme by catalyzing the oxidation of heme to ferrous iron, carbon monoxide (CO), and biliverdin. Biliverdin is then converted to bilirubin by biliverdin reductase. Whereas HO-2 production is constitutive, HO-1 is induced by diverse stimuli, especially those that induce ROS, including cigarette smoke, oxidized lipids, and metalloporphyrins . Although the physiologic importance of HO-1 in platelet production has never been described, the clinical manifestation of thrombocytosis in a patient with HO-1 deficiency suggests that HO-1 expression dampens thrombopoiesis [9,10]. We therefore hypothesized that HO-1 would regulate thrombopoiesis. Here, we demonstrate that the inhibition of HO-1 activity enhanced 15d-PGJ2-induced platelet production by enhancing ROS generation. These new data suggest that HO-1 plays an important inhibitory role in thrombopoiesis.
Materials and methods
15d-PGJ2 and rosiglitazone were obtained from Biomol (Plymouth Meeting, PA, USA); Pioglitazone was obtained from ChemPacific (Baltimore, MD, USA); 9, 10 dihydro-15d-PGJ2 (CAY10410), PGJ2, PGD2, 15d-PGD2, and GW9662 were purchased from Cayman Chemical (Ann Arbor, MI, USA); anti-HO-1 (Assay Designs, Ann Arbor, MI, USA); hemin, N- acetylcysteine (NAC), and glutathione reduced ethyl ester (GSH-EE) were all purchased from Sigma (St Louis, MO, USA); and total actin (CP-01) was obtained from Oncogene (Cambridge, MA, USA). 5-(and-6)-carboxy-2’,7’-dichlorodihydrofluorescein diacetate (carboxy-H2DCFDA) was purchased from Molecular Probes (Eugene, OR, USA).
Cells and culture conditions
Meg-01 cells were purchased from the American Type Culture Collection (Rockville, MD, USA). Cells were cultured in RPMI-1640 tissue culture medium (Invitrogen, Carlsbad, CA, USA) supplemented with 5% fetal bovine serum (FBS; Invitrogen), 10 mmol L−1 HEPES (N-2-hydroxyethylpiperazine-N’-2-ethanesulfonic acid; Sigma), 2 mmol L−1l-glutamine (Invitrogen), 4.5 g L−1 glucose (Invitrogen), and 50 μg mL−1 gentamicin (Invitrogen).
Human blood platelet isolation
Whole blood was obtained under University of Rochester IRB approval from male and female donors by venipuncture into a citrate phosphate dextrose adenine solution containing collection bag (Baxter Fenwal, Round Lake, IL, USA) or Vacutainer tubes containing buffered citrate sodium (BD Biosciences, Franklin Lakes, NJ, USA). Platelets were then isolated as described [5,11]. On average, 5.5 × 1010 platelets per unit of blood were obtained, and the platelet purity was greater than 99%.
Cells were fixed with Caltag Laboratories fixation medium A (Invitrogen) for 15 min, followed by an incubation with the anti-HO-1 antibody (1:1000) (diluted in Caltag Laboratories permeabilization medium B; Invitrogen) for 1 h at room temperature (RT). Cells were then washed and incubated with a FITC-conjugated goat antirabbit IgG (Jackson Immunoresearch Laboratories Inc., West Grove, PA, USA; 1:400 dilution in Reagent B) for 1 h at RT. Cells were analyzed by flow cytometry or were air-dried on slides and coverslipped in Vectashield Hard Set Mounting Medium (Vector Laboratories Inc., Burlingame, CA, USA). Cells were visualized using an Olympus BX51 light microscope (Olympus, Melville, NY, USA), photographed with a SPOT camera, and analyzed with SPOT RT software (New Hyde Park, NY, USA).
Construction of a lentiviral vector expressing PPARγ siRNA
A lentiviral construct expressing PPARγ siRNA was developed as previously described [1,12].
Western blot analysis
Whole cell lysates were prepared using ELB buffer (15 mmol L−1 HEPES (pH 7), 250 mmol L−1 NaCl, 0.1% Nonidet P-40, 5 mmol L−1 EDTA, 10 mmol L−1 NaF, 0.1 mmol L−1 NA3VO4, 50 μmol L−1 ZnCl2, supplemented with 0.1 mmol L−1 PMSF, 1 mmol L−1 DTT, and mixture of protease and phosphatase inhibitors) and protein was quantified using bicinchoninic acid protein assay (BCA assay kit; Pierce, Rockford, IL, USA). Twenty micrograms of protein was electrophoresed on 10% sodium dodecylsulfate polyacrylamide gel electrophoresis gels, electroblotted onto Immun-blot PVDF membrane (Bio-Rad, Hercules, CA, USA) and blocked with 10% non-fat dry milk in 0.1% Tween 20 (in PBS) for 1 h at RT. Antibodies against HO-1 (1:5000; Assay Designs), PPARγ (1:1000; Biomol) and actin (1:10 000) were used to assess changes in protein levels. Membranes were washed and then incubated with an appropriate horseradish peroxidase-conjugated secondary antibody. The membranes were visualized using enhanced chemiluminescence (NEN Life Science Products, Boston, MA, USA) and developed on Classic X-ray film (Laboratory Product Sales, Rochester, NY, USA).
Reverse transcription polymerase chain reaction
Total RNA was isolated from the cells by using the RNeasy RNA isolation kit according to the manufacturer’s instructions (Qiagen, Crawley, UK). RNA was reverse transcribed to cDNA and HO-1 mRNA quantified using the following primers [13,14]: HO-1: 5′-CAGGCAGAGAATGCTGAGTCC-3′ (sense) and 5′-GCTTCACATAGCGCTGCA-3′ (antisense). Human 7S sequences were used as a control. 7S: 5′-ACCACCAGGTTGCCTAAGGA-3′ (sense) and 5′-CACGGGAGTTTTGACCTGCT-3′ (antisense). iQ SYBR Green Supermix (Bio-Rad) was used in the real-time polymerase chain reaction (PCR) assay and results were analyzed with Bio-Rad iCycler software. Values were normalized to 7S and fold-change compared between untreated and 15d-PGJ2-treated Meg-01 cells.
Platelet isolation for Western blot
Ten × 106 Meg-01 cells or primary human megakaryocytes were untreated or treated with 15d-PGJ2 (10 μmol L−1). After 24 h, platelets were isolated by centrifugation at 150 × g for 10 min. Supernatants underwent sequential centrifugation (500 × g for 10 min and 1000 × g for 10 min). Platelet purity (> 99%) was assessed by flow cytometry. The platelet pellet was lysed and western blot was performed as described above.
HO-1 activity assay
Meg-01 cells (1 × 107) were pretreated with SnPPIX (5 μmol L−1) for 1 h and then treated with 15d-PGJ2 (10 μmol L−1) for 12 h. HO-1 activity was measured as previously described . Briefly, Meg-01 lysates were added to a reaction mixture containing hemin as a substrate, mouse liver cytosol as a source of biliverdin reductase and NADPH. The reaction was carried out at 37 °C in the dark and terminated by the addition of 1 mL of chloroform to extract bilirubin, a product of HO degradation. The concentration of bilirubin was determined spectrophotometrically, using the difference in absorbance between 460 and 530 nm with an absorption coefficient of 40 mmol L−1 cm−1.
Platelet production analysis
Cells were pretreated with SnPPIX (5 μmol L−1) and/or GSH-EE (5 mmol L−1) or NAC (1 mmol L−1) for 1–2 h, followed by treatment with either vehicle, hemin (1 μmol L−1), or 15d-PGJ2 (10 μmol L−1) for 24 h. Platelets were then isolated and quantified as previously described .
Reactive oxygen species production
Cells were pretreated with SnPPIX (5 μmol L−1) for 1–2 h. Cells were then treated with either vehicle, hemin (1 μmol L−1), or 15d-PGJ2 (10 μmol L−1) for 6 h. ROS production was measured by adding carboxy-H2DCFDA (10 μmol L−1) to cells for 20 min at RT. The cells were washed and immediately analyzed by flow cytometry.
Megakaryocyte differentiation from human cord blood-derived CD34+ cells
Human CD34+ cord blood cells were obtained from AllCells (Emeryville, CA, USA). Cells were plated at 2.5 × 105 cells per well in a 12-well plate and cultured in serum-free medium as previously described  and supplemented with 100 ng mL−1 of recombinant human thrombopoietin (rhTPO; R&D Systems, Minneapolis, MN, USA). After 14 days in culture, megakaryocytes were identified by staining with a CD61-FITC antibody and analyzed on a BD Biosciences FACSCalibur flow cytometer. Data were analyzed using FlowJo software (Treestar, Ashland, OR, USA).
Statistical significance was determined by paired two-way analysis of variance (anova). All data are represented as the mean ± standard deviation and statistical significance was assigned for P <0.05.
15d-PGJ2 upregulates HO-1 in Meg-01 cells
We first tested whether 15d-PGJ2 would influence HO-1 protein expression in Meg-01 cells. Treatment with 15d-PGJ2 strongly upregulated HO-1 protein in a dose-dependent (1–10 μmol L−1) manner (Fig. 1A). Meg-01 cells treated with 15d-PGJ2 (10 μmol L−1) for 30 min, 8 h or 24 h show an induction of HO-1 protein at both 8 h and 24 h (Fig. 1B). Immunofluorescent staining confirmed these findings and demonstrated that treatment with 10 μmol L−1 of 15d-PGJ2 induced HO-1 protein expression predominantly in the cytoplasm (Fig. 1C).
We also examined HO-1 steady-state mRNA levels after 15d-PGJ2 (10 μmol L−1) treatment using real-time PCR. This showed a > 200-fold induction of HO-1 mRNA by 4 h (Fig. 1D), which peaked by 8 h (> 700-fold) and remained significantly above control (untreated) levels by 12 h (> 200-fold).
HO-1 upregulation by 15d-PGJ2 is independent of PPARγ
Because 15d-PGJ2 is a metabolite of PGD2, we next evaluated the ability of PGD2 and two other metabolites, PGJ2 and 15d-PGD2 , to enhance HO-1 expression. Figure 2A demonstrates that PGD2, PGJ2, and 15d-PGJ2 all induced HO-1 protein expression in Meg-01 cells, whereas there was no induction with 15d-PGD2. HO-2 expression remained relatively unchanged (Fig. 2A).
We next assessed the role of PPARγ in HO-1 induction by 15d-PGJ2 because 15d-PGJ2 activates PPARγ in Meg-01 cells . Meg-01 cells were treated with natural and synthetic PPARγ ligands including 15d-PGJ2 (10 μmol L−1), rosiglitazone (10 μmol L−1), and pioglitazone (10 μmol L−1), as well as 9,10 dihydo-15d-PGJ2, a structural analog of 15d-PGJ2 that lacks the electrophilic carbon [6,7,18]. Hemin (1 μmol L−1), a well-described inducer of HO-1 , was used as a positive control. Figure 2B demonstrates that only 15d-PGJ2 and hemin induced HO-1 in Meg-01 cells.
To further evaluate the role of PPARγ in 15d-PGJ2-induced HO-1, Meg-01 cells were infected with a lentiviral PPARγ siRNA to knock down PPARγ expression and demonstrated a > 90% reduction in PPARγ protein (Fig. 2C). However, reducing PPARγ expression failed to attenuate 15d-PGJ2-induced HO-1 expression (Fig. 2D, lanes 4 and 6). In addition, treatment of Meg-01 cells with an irreversible PPARγ antagonist (GW9662) 1 h prior to 15d-PGJ2 addition failed to attenuate 15d-PGJ2-induced HO-1 (Fig. 2D, lane 2). These complementary approaches both indicate that HO-1 induction by 15d-PGJ2 is independent of PPARγ.
Daughter platelets derived from 15d-PGJ2-treated Meg-01 cells have elevated HO-1 protein levels
To investigate whether platelets express HO-1, platelets were isolated from five healthy donors (D1–D5) and western blotted for HO-1. Figure 3A demonstrates that human platelets from the five donors express variable levels of HO-1 protein. We next investigated whether platelets derived from untreated or 15d-PGJ2-treated Meg-01 cells also express HO-1. Interestingly, platelets produced from 15d-PGJ2-treated Meg-01 cells highly expressed HO-1 (Fig. 3B), whereas platelets produced from untreated Meg-01 cell cultures lacked HO-1. These data demonstrate that HO-1 is expressed in freshly isolated and culture-derived platelets.
Inhibition of HO-1 enhances platelet production in Meg-01 cells
The function of HO-1 in human platelets is unknown. To determine if the 15d-PGJ2-mediated increase in HO-1 was associated with an increase in HO-1 activity, Meg-01 cells were incubated with 15d-PGJ2 in the presence or absence of tin protoporphyrin-IX (SnPPIX), a well-described competitive inhibitor of human HO activity [20,21]. Meg-01 cells exhibited a basal level of HO activity, which was augmented by 15d-PGJ2 treatment (Fig. 4A). SnPPIX alone decreased HO activity and addition of SnPPIX (5 μmol L−1) prevented the 15d-PGJ2-mediated increase.
To further investigate the role of HO-1 in platelet production, we treated Meg-01 cells with SnPPIX to inhibit HO-1 activity. Meg-01 cells treated with SnPPIX prior to 15d-PGJ2, exhibited significantly enhanced platelet production compared to 15d-PGJ2 treatment alone (Fig. 4B), indicating that HO-1 regulates platelet production.
We previously reported that ROS were important for 15d-PGJ2-induced platelet production . Therefore, we hypothesized that HO-1 activity inhibits platelet production by protecting against 15d-PGJ2-induced oxidative stress. To test this, carboxy-H2DCFDA was used to determine the effects of 15d-PGJ2 and SnPPIX on ROS generation. SnPPIX alone had no effect on ROS production (Fig. 4C), whereas treating Meg-01 cell with SnPPIX prior to 15d-PGJ2, significantly enhanced ROS production (P <0.05) compared to 15d-PGJ2 treatment alone (Fig. 4C). Interestingly, addition of antioxidants such as GSH-EE (5 mmol L−1) or NAC (1 mmol L−1), in conjunction with SnPPIX prior to 15d-PGJ2 addition, significantly reduced platelet production (Fig. 4D). These data support our hypothesis that the antioxidant properties of HO-1 are important for blunting platelet production.
15d-PGJ2-induced HO-1 in primary human megakaryocytes is inhibitory to platelet production
We next characterized thrombopoiesis using primary human megakaryocytes. These cells expressed a basal level of HO-1 that was dose-dependently induced by 15d-PGJ2 (Fig. 5A). In addition, platelets produced from these cells also expressed elevated levels of HO-1 protein after treatment with 15d-PGJ2 (Fig. 5B). Pretreating primary human megakaryocytes with SnPPIX prior to 15d-PGJ2 addition significantly enhanced platelet production when compared to 15d-PGJ2 treatment alone (Fig. 5C). 15d-PGJ2 induced ROS in primary human megakaryocytes (Fig. 5D). Treating with SnPPIX prior to 15d-PGJ2 significantly enhanced ROS production compared to 15d-PGJ2 treatment alone (Fig. 5D). Treatment with the antioxidant, GSH-EE (5 mmol L−1) or NAC (1 mmol L−1), in conjunction with SnPPIX prior to 15d-PGJ2 significantly reduced enhancement of platelet production (Fig. 5E). Collectively, these data provide compelling evidence that HO-1 dampens ROS generation and inhibits platelet release.
Maturation of megakaryocytes and platelet production are complex processes orchestrated by cytokines, growth factors and transcription factors. Recently, our laboratory discovered that 15d-PGJ2 is a potent inducer of ROS and that ROS are important for platelet release . The effects of 15d-PGJ2 on both ROS generation and platelet production were independent of PPARγ . Here, we demonstrate that 15d-PGJ2 also increases HO-1 protein expression in megakaryocytes, as well as daughter platelets independent of PPARγ. Moreover, inhibition of HO-1 activity significantly increased 15d-PGJ2-induced ROS generation and platelet production. This platelet production was attenuated with antioxidant treatment. These new findings indicate that redox balance plays a role in regulating platelet production and that HO-1 inhibits thrombopoiesis.
15d-PGJ2 elicits a broad spectrum of biologic events that include the induction of antioxidant enzymes, including HO-1. 15d-PGJ2 increases HO-1 expression in macrophages, glial cells, and myofibroblasts [22–24]. Herein, we showed that 15d-PGJ2 potently increased HO-1 protein expression in Meg-01 cells and in primary human megakaryocytes (Figs. 1A and 5A) as well as HO-1 activity in Meg-01 cells (Fig. 4A). Interestingly, Meg-01 cells lack basal HO-1 expression (Fig. 1A), whereas primary human megakaryocytes modestly express HO-1 (Fig. 5A). We speculate that the difference in basal HO-1 expression between Meg-01 cells and primary human megakaryocytes is a result of the stage of maturation, as several key platelet proteins are induced during megakaryopoiesis .
The ability of the PPARγ ligand 15d-PGJ2 to increase HO-1 was independent of PPARγ. Inhibiting PPARγ utilizing the PPARγ antagonist, GW9662, or inhibiting PPARγ expression had little effect on the ability of 15d-PGJ2 to augment HO-1. We speculate that the PPARγ-independent effect of 15d-PGJ2 relates to its electrophilic properties. PGJ2, 15d-PGJ2, and PGD2 all induced HO-1 (Fig. 2A). PGJ2 and 15d-PGJ2 both possess an α,β-unsaturated ketone within a cyclopentenone ring, rendering both molecules highly electrophilic and chemically reactive in nature . Both PGD2 and 15d-PGD2 lack this α,β-unsaturated ketone. However, PGD2 is metabolized to electrophilic PGJ2 and 15d-PGJ2 inside the cell . In contrast, a 15d-PGJ2 structural analog, 9,10 dihydro-15d-PGJ2 (CAY10410), and other non-electrophilic PPARγ ligands, rosiglitazone, pioglitazone, and 15d-PGD2, all failed to augment HO-1 (Figs. 2A and 2B). Therefore, the electophilic properties of 15d-PGJ2 are important for HO-1 induction.
We provide compelling evidence that HO-1 expression by megakaryocytes inhibits platelet production. The inhibition of HO-1 activity in combination with 15d-PGJ2 treatment significantly enhanced platelet production by megakaryocytes (Figs. 4B and 5C). We have previously shown that culture-derived platelets generated from the Meg-01 cells and CD34+ cord blood express platelet markers and are functional . We hypothesized that the antioxidant properties of HO-1 dampen platelet release from megakaryocytes. We previously reported that ROS generation was important for platelet production . Further, antioxidants such as NAC and GSH-EE attenuated 15d-PGJ2-enhanced platelet production . Here, we show that inhibiting HO-1 activity in conjunction with 15d-PGJ2 treatment significantly increased ROS generation compared to 15d-PGJ2 treatment alone (Figs. 4C and 5D). Interestingly, human HO-1 deficiency is characterized by both elevated levels of blood platelets and oxidative injury [9,10]. This is consistent with our data showing that inhibiting HO-1 activity increases both platelet number and ROS. Furthermore, Farkas et al.  recently demonstrated a negative association between HO-1 mRNA expression in venous blood and platelet number in preterm infants with respiratory distress syndrome. As human platelets have different levels of HO-1 protein expression (Fig. 3A), we speculate that levels of HO-1 in human donor platelets may correlate with platelet number.
Because we demonstrated the ability of human megakaryocytes to induce HO-1 protein expression, we next wanted to determine if the HO-1 protein could be packaged into platelets during thrombopoiesis. Interestingly, the platelets derived from primary human megakaryocytes expressed a basal level of HO-1 similar to freshly isolated human platelets (Fig. 5B). Culture-derived Meg-01 platelets lacked basal HO-1 expression (Fig. 3B). Importantly, the platelets produced from both Meg-01 cells and primary human megakaryocytes following 15d-PGJ2 treatment exhibited elevated levels of HO-1 protein (Figs. 3B and 5B). This demonstrates that alterations in key megakaryocyte proteins can be packaged into platelets and hence could influence platelet function.
It has been suggested that HO-1 induction contributes to the anti-inflammatory activities of 15d-PGJ2. For example, HO-1 induction by 15d-PGJ2 reduces fibrotic activity of myofibroblasts and myocardial infarct size in a rat model of acute myocardial infarction [24,28]. The ability of 15d-PGJ2 to elevate platelet HO-1 by elevating megakaryocyte HO-1 levels may have important implications for inflammation, such as reducing the risk of thrombotic events. Hemin-induced HO-1 in the carotid arterial wall inhibited platelet-dependent thrombus formation . HO-1 knockout mice exhibited accelerated, occlusive arterial thrombosis when compared to wild-type mice . Therefore, we speculate that platelet-specific HO-1 can also dampen thrombus formation. Future experiments will address whether or not inhibiting HO-1 activity will affect platelet function. Studies of HO-1 activity in patients with diseases involving pathological platelet production/destruction (e.g. idiopathic thrombocytopenic purpura, liver disease) or pathologically increased activity (deep vein thrombosis, acute myocardial infarction) will be of interest.
The complex developmental biology of platelets coupled with the prevalence of disorders characterized by platelet activation, thrombocytosis and/or thrombocytopenia warrants research to understand mechanisms of thrombosis and thrombopoiesis. We have furthered the definition of the role of ROS in platelet production by providing strong evidence for an inhibitory role of HO-1 during thrombopoiesis. We have demonstrated that 15d-PGJ2 can increase platelet HO-1 levels by increasing megakaryocyte HO-1 levels. This suggests that 15d-PGJ2, by increasing platelet HO-1 levels, may be a promising antiplatelet therapy. Importantly, the idea that the platelet proteome and, ultimately, platelet function may be altered by influencing megakaryocyte protein levels suggests the megakaryocyte as a new target for platelet-directed pharmacotherapies.
This work was supported by the National Institutes of Health Grants T32ES07026, T32HL007152, HL078603, HL086367, ES01247 and the PhRMA Foundation.
Disclosure of Conflicts of Interests
The authors state that they have no conflict of interest.