2025 Volume 50 Issue 11 Pages 617-625
Background: Acute inflammation is induced by lipopolysaccharide (LPS), accompanied by activation of platelets. Carbon monoxide (CO), an endogenous bioactive gas, has been shown to bind to mitochondria and exert anti-inflammatory effects. In this study, we investigated the effect of CO on the mitochondrial membrane potential of platelets activated by LPS. Methods: To elucidate the mechanism of the LPS-induced platelet response, human platelets were stimulated with LPS (10 μg/mL). Human platelet concentrates were divided into four groups: Untreated (Control), LPS-treated (LPS), LPS and CO-dissolved solution-treated (LPS + CO), and LPS and exogenous carbon monoxide releasing molecule-2-treated (LPS + CORM-2) groups. After 30 minutes, lactate levels and mitochondrial membrane potential (ΔΨm) in the platelets were measured. Morphological changes of the platelets were also observed using transmission electron microscopy. Results: In the LPS group, the proportion of platelets with depolarized ΔΨm increased, accompanied by elevated lactate levels compared with the control group. On the other hand, in the LPS+CO and LPS+CORM-2 groups, the proportion of depolarized platelets did not significantly increase, and lactate levels were not significantly elevated. Morphologically, elongating pseudopods and cell condensation were observed in the LPS group, however, these changes were not induced in the LPS+CO and LPS+CORM-2 groups. Conclusion: These results suggest that CO prevents a decrease in the platelet ΔΨm and thereby inhibits platelet activation by LPS treatment.
Lipopolysaccharide (LPS) is a major component of the cell wall of Gram-negative bacteria and is a potent immune stimulant. LPS triggers an inflammatory response, especially in association with infection, and activates platelets directly or via other immune cells (Galgano et al., 2022). LPS induces platelet aggregation, thereby promoting thrombus (blood clot) formation. LPS also activates the coagulation cascade, which causes excessive platelet consumption (Damien et al., 2015) and a consequent decrease in platelet count (thrombocytopenia) (Garraud et al., 2023). Furthermore, the inflammatory response induced by LPS shortens the lifespan of platelets and accelerates their degradation (Page et al., 2022).
Although carbon monoxide (CO) is generally recognized as a poisonous gas, it has been demonstrated that platelets contain an enzyme that produces CO, an endogenous biologically active gas (Motterlini and Otterbein, 2010). It has become clear that CO controls platelet aggregation and platelet function (Motterlini et al., 2002). Moreover, CO reduces platelet reactivity and is an anticoagulant; however, it exerts some cardioprotective and procoagulant properties (Russo et al., 2023). In a study in which CO was administered, it was found that exogenous CO can influence platelet function and regulate platelet aggregation reactions in vitro (Motterlini et al., 2003). Lipid-soluble metal carbonyl complex tricarbonyldichlororuthenium (II) dimer ([Ru(CO)3Cl2]2), also termed exogenous carbon monoxide releasing molecule-2 (CORM-2), was the first compound to make this technology feasible (Seixas et al., 2015). CORM-2 has an ability to facilitate the pharmaceutical use of CO by delivering it to tissues and organs of interest (Liu et al., 2013). CO is toxic in high concentrations, but CORM-2 releases it in a controlled way, allowing us to obtain its beneficial effects with avoiding its toxicity (Sun and Chen, 2009). Studies have shown that CORM-2 can suppress LPS-induced inflammatory responses in endothelial cells in the umbilical vein, peripheral blood mononuclear cells, and macrophages (Fei et al., 2012). Similar studies have confirmed that CO from CORM-2 promotes recovery in mice models of lethal endotoxemia and sepsis induced by LPS and cecal ligation and puncture (Leytin et al., 2009). However, the detailed mechanism by which CO prevents LPS-induced platelet activation of platelets remains unclear.
Besides their primary role in ATP generation, mitochondria are involved in other important processes that contribute to platelet function and signaling, such as reactive oxygen species generation, Ca2+ homeostasis, and apoptosis (Melchinger et al., 2019). Platelets have relatively few mitochondria (4–8 mitochondria per platelet) and lack a nucleus; thus, they cannot replenish the mitochondrial proteins encoded by nuclear DNA (Ajanel et al., 2023). This indicates that small changes in platelet mitochondrial dynamics can remarkably impact platelet function and hemostasis (Melchinger et al., 2019). Mitochondrial membrane potential (ΔΨm) reflects mitochondrial function and is an indicator of mitochondrial energy status. Mitochondrial dysfunction has been implicated in the pathogenesis of multiple-organ dysfunction syndrome (Yamakawa et al., 2013). In a relatively small cohort study consisting of patients with septic shock reported that ΔΨm depolarization in monocytes is enhanced in the early phase of septic shock (Leytin et al., 2009). Mitochondrial damage or dysfunction can considerably reduce platelet viability and increase the risk of thrombotic vascular events (Melchinger et al., 2019). Furthermore, CO can exhibit cytoprotective and anti-inflammatory functions; however, the underlying mechanisms remain unclear. The purpose of this study was to clarify whether exogenously administered CO prevents a decline in platelet ΔΨm during in vitro LPS stimulation.
PCs used in this study were obtained from the Japanese Red Cross Society based on the “Partial Revision of the Approach to Restrictions on Blood Collection, etc. as Defined in Article 12 of the Act on Securing a Stable Supply of Safe Blood Products” (Notification of the Director, Blood Products Division, Pharmaceutical and Consumer Health Bureau, Ministry of Health, Labor and Welfare, No. 0826-3, August 26, 2020). The PCs (Irradiated Concentrated Platelets-LR, 1 unit approximately 20 mL:Ir-PC-LR-20) (2.0 × 1011/mL) were collected according to the Japanese Red Cross Donor Selection Guidelines (Takami et al., 2019) and contains <5% plasma. The prepared PCs were stored at 20°C–24°C on a flatbed agitator (60 cycles/min) for at least 3 hr before use and used within approximately 3 days of collection.
Preparation of bicarbonated Ringer’s solution supplemented with acid citrate-dextrose and adenineBicarbonated Ringer’s solution supplemented with acid citrate-dextrose and adenine (BRS-A) was prepared by adding 25 mL acid citrate-dextrose and adenine (ACD-A) (Kawasumi Laboratories, Inc., Tokyo, Japan) to 500 mL BICANATE injection solution (Otsuka Pharmaceutical Factory, Inc., Tokushima, Japan) at a ratio of 1:20.
BICANATE injection solution is a 500-mL solution containing 2.92 g NaCl, 0.15 g KCl, 0.11 g CaCl2·2 H2O, 0.10 g MgCl2·6 H2O, 1.175 g NaHCO3, and 0.10 g Na3-citrate·2 H2O in a 500 mL solution. The ACD-A solution comprised a 500-mL solution containing 11.0 g C6H5Na3O7 ·2 H2O, 4.0 g C6H8O7·H2O, and 11.0 g glucose in a 500 mL solution (Oikawa et al., 2013).
Preparation of the CO-dissolved solution in BRS-ATo remove composition of air, it was eliminated from the BRS-A preparation using ACD-A-BICANATE solution. The BRS-A solution was mixed with an equal volume of CO contained in a syringe. To prepare the CO-dissolved solution, the BRS-A buffer was mixed with an equal volume of CO contained in a syringe at a 1:1 ratio and incubated for 1 hr at room temperature (20–24°C) under constant agitation (60 cycles/min) to ensure maximal solubility. The solution was used within 30 min. Although the CO concentration was not directly measured, saturation was achieved by keeping temperature and mixing conditions the same each time. Fresh CO-dissolved solution was prepared for each experiment, and consistent exposure conditions (temperature and agitation) were maintained to ensure reproducibility across all samples. CO exists predominantly as an intact diatomic molecule in neutral aqueous solution (pH 7.2–7.4)(Awoonor-Williams and Rowley, 2016). The pH of the CO-dissolved solution was measured and confirmed to be within the physiological range (pH 7.2–7.4) before use.
Preparation of CORM-2 solutionCORM-2 and dimethyl sulfoxide (DMSO) were purchased from Sigma-Aldrich (St. Louis, MO, USA). CORM-2 was solubilized in DMSO to obtain a 5-mM stock solution and used at 50 μM as a final concentration in the experiments (Sawle et al., 2005).
Study designLPS was used to activate platelets in vitro. LPS O111:B4 (Sigma-Aldrich, USA) derived from Escherichia coli, was used in this study (Galgano et al., 2022). The final LPS concentration used in the experiments was 10 µg/mL, as described previously (Damien et al., 2015). The prepared PCs were divided into four treatment groups: untreated (control), LPS-treated (LPS), LPS and CO-dissolved solution-treated (LPS + CO), and LPS and CORM-2-treated (LPS + CORM-2) groups. As a preliminary experiment, platelets were treated with CO-dissolved solution or CORM-2 without LPS to evaluate the effects of CO or CORM-2 alone. These treatments did not cause any noticeable changes in ΔΨm, lactate, or glucose levels compared to the untreated control group. In each group, 600 μL PCs were dispersed evenly into 400 μL of BRS-A with or without LPS, dissolved CO, or CORM-2. After 30 min later, mitochondrial membrane potential (ΔΨm) and glucose and lactate levels were measured in the platelets of each group were measured (n = 5). Furthermore, morphological changes in those platelets were observed using transmission electron microscopy (TEM).
Assessment of obtained PCs by flow cytometryFirst, we assessed whether our gating strategies were optimized for determining the numbers of platelets. Untreated PC samples were stored at room temperature for 3 hours before use. The platelet count percentages in the PCs were then determined. The percentage of platelets in these samples remained unchanged from those in the fresh samples. Next, we assessed the accuracy of the flow cytometry platelet count approach. The number of platelets in each sample (n = 3) were quantitated using the chosen flow rate approach and an automated cell counter (50,000 cells). Expression of the platelet surface marker CD41α (GPαIIb) was determined by flow cytometry, as described previously (Liu et al., 2013). Anti CD41α antibodies were purchased from BD Bioscience- Pharmingen (San Jose, CA, USA). The data were collected and analyzed using Cell Quest software version 4.0 (BD Bioscience- Pharmingen). Before the experiment, the flow cytometer was optimized using CaliBrite beads and FACS Comp software (BD Bioscience-Pharmingen). The forward scatter (FSC), side scatter (SSC), and fluorescence parameters were set at a logarithmic scale (Fig.1). A threshold for the FSC was set to differentiate debris and noise signals from true events. The platelet gate was first established on an FSC-vs.-SSC dot plot. The number of platelets was calculated using a flow rate based approach, as previously described (Martyanov et al., 2020).
Percentages of CD41α-positive human platelets in each experimental group. The CD41α positive rate was analyzed relative to all platelet concentrates. (a) Platelets were stained with FITC-conjugated CD41α antibody and isotype control (IgG), showing a CD41α-positive population of 96.6% (FITC-A, Iso type). (b) Surface marker staining confirmed a platelet purity of >90%. Dots indicate individual values.
For glycolysis analysis, lactate and glucose levels in the PCs of the 4 groups were measured by colorimetric and fluorometric detection using a Glucose Assay Kit-WST (Dojindo, Japan) and Lactate Assay Kit-WST (Dojindo, Japan) according to the manufacturer’s instructions (Takashima et al., 2020). A standard curve of the intensity of the color (O.D.) was plotted to the concentration of standards. The lactate and glucose concentrations in each sample were interpolated from this standard curve.
Measurement of ΔΨm in PCsMitotracker Green FM (Invitrogen™, USA) was used to react with platelets according to the manufacturer’s instructions and then incubated in the dark for 30 min at 37°C. After incubation, platelets were fixed by adding 600 uL of HEPES saline 4% formaldehyde for 10 min, protected from light. The activity of ΔΨm was evaluated according to the manufacturer’s instructions.
Samples were analyzed using a flow cytometry system (BD FACS LyricTM, USA). Fluorescent beads (BD® CS&T beads) were added daily to ensure the stability of the system. After setting the appropriate threshold for the FSC, 50,000 events were acquired in a life gate (Yamakawa et al., 2013).
Observation of platelets by Transmission Electron Microscope (TEM)The platelets in BRS-A were treated with and without LPS, dissolved CO, or CORM-2 as described above to prepare the treatment protocol for each treatment group. The platelets were incubated for 30 min at 24°C and then fixed in suspension with 2.5% glutaraldehyde for 30 min, and pelleted by centrifugation (800 × g, 10 min). The precipitates were washed, postfixed in 1% osmium tetroxide, dehydrated, and embedded in epoxy resin. Ultrathin sections were sliced using a Leica UC7 ultramicrotome and contrasted using 2% uranyl acetate and lead citrate (Almhanawi et al., 2017). Grids were viewed using a JEM-1400 PLUS transmission electron microscope (Japan Electron Optics Laboratory Co., Ltd.).
Ethics statementThe study was performed in accordance with the provisions of the Declaration of Helsinki. Ethical approval was granted by the Ethics Committee of Tokyo Medical University (ethics approval code: T2021-0196).
Statistical analysisData were analyzed and plotted in graphs using GraphPad Prism version 5.0 (GraphPad Software Inc.; San Diego, CA, USA). A nonparametric Kruskal–Wallis test combined with a Dunn’s multiple comparisons test (Dunn’s) was conducted to test for differences between groups. Analysis of continuous data was performed using Friedman test combined with Dunnett’s multiple comparison test. p < 0.05 was considered statistically significant.
To investigate the metabolic effects of CO and CORM-2 on human platelets, the cells treated with LPS and subsequently exposed to respective solutions were prepared. flow cytometry analyses revealed that no differences were found among the 4 experimental groups in terms of positive percentage and mean fluorescence intensity of CD41α (Fig. 1). Therefore, the ΔΨm, lactate and glucose concentrations in each group, were then assessed.
GlycolysisThe LPS group showed a significant increase of lactate production (p = 0.0031) (Fig. 2a) but no change in glucose consumption (Fig. 2b) when compared with the control group. LPS-treated platelets showed a significant increase in lactate levels, indicative of enhanced glycolysis. However, the lactate production in the LPS + CO and LPS + CORM-2 were not significantly elevated than the control group. Furthermore, their lactate production seemed lower than the LPS groups, although the differences were not significant. Although lactate levels increased with LPS, glucose levels did not change significantly, possibly due to a compensatory increase in glucose uptake mediated by the activation or upregulation of glucose transporters (GLUT) during platelet activation (Karpatkin, 1967; Warshaw et al., 1966), or by the utilization of intracellular glycogen stores. The glucose level was significantly lower in the LPS + CO group among the 4 groups (p = 0.0037).
Lactate and glucose levels in human platelets from each experimental group. Concentrations of Lactate (a) and glucose (b) on Control, LPS, LPS + CO, and LPS + CORM-2 groups. Values are expressed as the mean ± S.E.M of data from 5 samples per group. **p < 0.01. Dots indicate individual values.
In the LPS group, the percentage of platelets with depolarized Δψm, determined by reduced Mitotracker fluorescence intensity, was significantly lower than that in the control group (p = 0.0133) (Fig. 3). In contrast, the proportion of depolarized platelets did not significantly change in the LPS + CO (p = 0.9145) and LPS + CORM-2 groups (p = 0.2733) compared with the control group.
Percentage of platelets with depolarized mitochondrial membrane potential (ΔΨm) in each experimental group. The percentage of platelets with depolarized ΔΨm was evaluated by reduced Mitotracker Green FM fluorescence intensity using flow cytometry. The LPS group showed a significantly lower proportion of depolarized platelets compared with the control group, whereas no significant decrease was observed in the LPS + CO and LPS + CORM-2 groups. Values are expressed as the mean ± S.E.M of data from 5 samples per group. *p < 0.05. Dots indicate individual values.
TEM visualization at 12,000× magnification showed morphological changes in platelets (Fig. 4a). In the LPS group, platelet activation was indicated by morphological changes, such as elongating pseudopods and irreversible cell condensation (Fig. 4b). In contrast, in the LPS + CO and LPS + CORM-2 groups, no cell pseudopodia or protrusions were observed although the number of open canalicular system was increased (Figs. 4c and 4d). Glycogen granules were observed in the control group. Morphological features were assessed qualitatively, as structural heterogeneity and the limited scope of TEM sampling precluded reliable quantitative analysis.
Electron microscopy of human platelets from each experimental group. Representative micrographs of human platelets from Control (a), LPS (b), LPS + CO (c), and LPS + CORM-2 (d) groups (Magnification ×12,000). The scale bar is 500 nm. Abbreviation and symbols show Glycogen (Gly), α-granule (*), Mitochondria (▷), Open canalicular system (▶), and Pseudopod (→).
This study demonstrated that CO prevented a decrease of platelet ΔΨm and morphologically preserved the cells after LPS treatment. This study is the first to show the CO-induced protection of ΔΨm in human platelets.
In general, activated platelets increase their glucose uptake, glycolysis, and glucose oxidation, and consume stored glycogen (Karpatkin, 1967; Warshaw et al., 1966). Although LPS-treated platelets showed increased lactate production, glucose levels remained unchanged, suggesting upregulated glycolysis without increased glucose uptake, possibly due to saturation of glucose transporters or mobilization of intracellular glycogen stores (Karpatkin, 1967). In contrast, the LPS + CO group exhibited a significant decrease in glucose levels without a corresponding increase in lactase, suggesting glucose utilization in mitochondrial energy production (Melchinger et al., 2019; Takashima et al., 2020). This response may reflect a protective effect of CO. In contrast, no such decrease was observed in the LPS+CORM-2 group, likely due to the slower and solvent-dependent release of CO from CORM-2, which may have resulted in insufficient CO availability (Motterlini and Otterbein, 2010). This metabolic shift may reflect a protective response induced by CO to preserve mitochondrial function and reduce oxidative stress. In the LPS + CORM-2 group, ΔΨm depolarization was attenuated whereas glucose levels did not significantly decrease. Therefore, there is a possibility that action mechanism of CORM-2 in protective effect for platelet activation partially differed from that of dissolved CO; however, the reason remains unclear. In the present study, CO concentrations were not directly quantified in either preparation. We will further investigate difference of the action mechanisms between CORM-2 and dissolved CO in further studies. Interestingly, CO preserved mitochondrial membrane potential while suppressing lactate accumulation, suggesting that platelet energy metabolism was redirected from glycolysis to alternative pathways. This effect may be explained by CO-mediated regulation of mitochondrial processes, such as mild uncoupling, promotion of oxidative phosphorylation, and increased activity of mitochondrial enzymes (Motterlini and Otterbein, 2010). CO has also been reported to reduce upregulated glycolysis indirectly via inhibition of key enzymes or restoration of mitochondrial respiration (Melchinger et al., 2019; Takashima et al., 2020). Therefore, the observed shift does not indicate a pathological abnormality in aerobic metabolism, but rather a regulated metabolic adaptation to inflammatory stress. Such a protective regulated metabolic shift may serve to preserve mitochondrial integrity and energy homeostasis in activated platelets. Lactate is the primary end-product of glycolysis and is released in large quantities from sites of inflammation sites. Changes in ΔΨm indicate the activation of cell organelles (Yamakawa et al., 2013). Furthermore, exogenous CO effectively inhibited an LPS-induced increase in platelet adhesion and aggregation, in addition to ATP secretion (Russo et al., 2023). Reportedly, CO released from CORMs inhibits LPS-induced pathological changes in platelets (Adach and Olas, 2020; Liu et al., 2013). In this study, the level of platelet ΔΨm seemed inversely correlated with the level of blood glycolysis, suggesting that this parameter represents an abnormality in platelet aerobic metabolism in platelets. LPS activates platelets primarily through TLR4 signaling, which triggers downstream inflammatory pathways and promotes metabolic changes including increased glycolysis (Russo et al., 2023). In this study, LPS treatment led to decreased ΔΨm and increased lactate levels, while CO and CORM-2 mitigated these effects. This suggests that CO preserves mitochondrial function and prevents excessive platelet activation. Our results are consistent with prior studies indicating that CO preserves mitochondrial integrity and influences platelet behavior in inflammatory settings (Russo et al., 2023). The results of ΔΨm showed a significant reduction in the LPS group compared to the control group. On the other hand, the LPS + CO and LPS + CORM-2 groups tended to have higher ΔΨm than the LPS group. This demonstrates that the treatments with CO-dissolved solution or CORM-2 treatment suppressed the LPS-induced activation of platelet organelles. Although the LPS group showed a lower proportion of platelets with depolarized Δψm compared with the control group, this does not contradict the concept of mitochondrial impairment. Early platelet activation has been reported to involve transient mitochondrial hyperpolarization or retention of Mitotracker fluorescence signals (Melchinger et al., 2019). These phenomena can mask depolarization and yield an apparently reduced ratio. When interpreted together with the significant lactate accumulation and the ultrastructural abnormalities observed under TEM, these findings consistently support the conclusion that LPS impairs mitochondrial function. Importantly, the absence of such changes in the LPS + CO and LPS + CORM-2 groups suggests that CO preserved mitochondrial stability and limited platelet activation under LPS stimulation, consistent with previous reports (Rondina et al., 2013; Yamakawa et al., 2013). Furthermore, we noted the presence of extending cell membrane projections, which are highly prominent indicators of platelet activation and an irreversible process (Almhanawi et al., 2017). In Fig 4, both LPS + CO and LPS + CORM-2 groups showed an increased number of open canalicular systems (OCS), which are typically linked to platelet activation. However, our TEM analysis did not reveal a clear increase in granule release in these groups. This suggests that CO helps maintain platelet structure and mitochondrial integrity, thereby limiting full activation (Selvadurai and Hamilton, 2018). These ultrastructural features, together with the metabolic data, support the notion that CO modulates platelet responses in a protective manner. Similar dissociations between structural and metabolic changes have been reported under oxidative or anti-inflammatory conditions (Rondina et al., 2013). TEM further revealed that the development of cell membrane projections inhibited platelet activation. DMSO is one of the most commonly used solvents for hydrophobic substances in biological experiments but is a toxic agent, although the influence of DMSO on respiration, and thus on arterial blood oxygenation, is still unclear and contentious (Takeda et al., 2016). CORM-2 is dissolved in DMSO. Therefore, it should be noted in determining the dose and concentration of DMSO when administered in vivo (Lin et al., 2004). On the other hand, CO-dissolved solution changes CO concentration in proportion to pressure according to Henry's law with no use of DMSO. Therefore, the sustained release of CO using CO-dissolved solution may be more safe and useful than using CORM-2.
The mitochondrial dysfunction of platelets results in their apoptosis in several diseases, including sepsis (Dewitte et al., 2017). Apoptotic platelets clot 50–100 times faster than normal platelets owing to the presence of phosphatidylserine on the platelet surface, which serves as a catalytic site for the assembly of clotting enzymes and thrombin generation (Lannan et al., 2014). Further studies are needed to establish the effects of exogenous CO administration using a CO-dissolved solution that can control the release of CO in apoptotic platelets.
In conclusion, the beneficial effects of CO on human platelets after LPS treatment may be mediated by preservation of the cell ΔΨm. Further development of exogenous CO administration using CO-dissolved solution that can control and release CO may enable the inhibition of platelet response during septic shock.
The authors thank Yuki Ogawa and Yuka Mituya for their contributions to this study. The authors would like to thank Enago (www.enago.jp) for the English language review.
Conflict of interestThe authors declare that there is no conflict of interest.
