Cardiovascular disease remains the leading cause of death worldwide, and thrombosis, the formation of blood clots inside blood vessels, sits at the center of heart attacks, strokes, and many other vascular events. While existing anticoagulant and antiplatelet drugs save lives, they carry risks of bleeding and are not always suitable for long-term preventive use. This has fueled intense interest in dietary compounds and botanical extracts that might gently modulate the vascular environment and reduce the likelihood of clots forming in the first place. The new research, led by Jin Tae Kim of the Korea Research Institute of Bioscience and Biotechnology and Hui Mang Son of Chung-Ang University, with corresponding authors Ho Jin Lee of Seoul National University and Hong Jin Lee of Chung-Ang University, adds mangosteen pericarp to the growing list of food-derived candidates worth serious scientific attention.

The research team focused on a 70 percent ethanol extract of the mangosteen pericarp, designated GMPE70, which they selected because it contains a relatively high concentration of α-mangostin, the most studied xanthone in the fruit’s rind. Xanthones are a class of polyphenolic compounds with well-documented antioxidant and anti-inflammatory properties, and α-mangostin in particular has attracted attention for its immunomodulatory effects, its ability to inhibit inflammatory signaling, and its reported benefits in metabolic and cardiovascular contexts. What makes the new study distinctive is that the investigators did not simply test the whole extract or the purified compound in isolation. Instead, they compared GMPE70 directly with α-mangostin at the exact concentration present in the extract, allowing them to determine whether the observed effects could be attributed to α-mangostin alone or whether other constituents of the pericarp contribute.

At the heart of the study lies the endothelium, the thin layer of cells lining the interior surface of blood vessels. A healthy endothelium maintains smooth blood flow, resists unnecessary leukocyte adhesion, and produces nitric oxide, a molecule that keeps vessels relaxed and prevents platelets from sticking together. When the endothelium becomes dysfunctional, often through chronic inflammation, oxidative stress, or metabolic disease, it transforms into a surface that actively promotes clot formation. Inflammatory signaling pathways, most notably the transcription factor nuclear factor kappa B, or NF-κB, switch on genes encoding adhesion molecules such as intercellular adhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule-1 (VCAM-1). These molecules act as molecular Velcro, snagging circulating immune cells and platelets and initiating the cascade that can culminate in an occlusive thrombus.

Using EA.hy926 cells, a widely used human endothelial cell line, the researchers demonstrated that both GMPE70 and α-mangostin at its concentration in the extract suppressed NF-κB activation. This suppression had downstream consequences: the expression of ICAM-1 and VCAM-1, the two key endothelial adhesion molecules, was significantly reduced. Simultaneously, both treatments enhanced the phosphorylation of endothelial nitric oxide synthase (eNOS), the enzyme responsible for producing nitric oxide in blood vessels. Increased eNOS phosphorylation generally means increased NO bioavailability, which promotes vasodilation, inhibits platelet aggregation, and maintains an antithrombotic surface. In other words, the mangosteen compounds appeared to push endothelial cells away from a pro-inflammatory, pro-thrombotic state and back toward a healthy, anticoagulant phenotype.

Perhaps the most visually compelling evidence came from adhesion assays in which the researchers incubated EA.hy926 endothelial cells together with THP-1 monocytes, a human monocytic cell line frequently used to model leukocyte behavior. When endothelial cells are activated, monocytes adhere to their surface, mimicking the early stages of vascular inflammation and thrombus initiation. GMPE70 and α-mangostin both significantly reduced this monocyte-endothelial interaction, confirming that the molecular changes seen at the level of gene and protein expression translated into functional consequences at the cellular level. The team also found that expression of LFA-1, the major leukocyte counter-receptor that binds ICAM-1, was suppressed in the monocytes. This dual effect, reducing both the ligands on the endothelial surface and the receptors on the immune cells, suggests a coordinated dampening of the adhesive dialogue between the two cell types.

To move beyond the Petri dish, the researchers employed a well-established animal model of thrombosis in which rats are injected with collagen and epinephrine. This combination triggers rapid platelet aggregation and widespread vascular occlusion, and it has been used extensively to evaluate the antithrombotic potential of natural products and synthetic compounds alike. When rats received GMPE70 or α-mangostin, the degree of vascular occlusion was significantly attenuated compared to untreated controls. Molecular analysis of vascular tissue revealed that the gene expression of adhesion molecules was regulated by the treatments, mirroring the in vitro findings and suggesting that the mechanisms observed in cultured cells operate in living organisms as well.

The equivalence between the whole extract and purified α-mangostin is one of the study’s most interesting findings. If the effects of GMPE70 could be fully explained by its α-mangostin content alone, this would simplify quality control and standardization for any future nutraceutical or functional food application. It also raises the possibility that α-mangostin serves as the primary pharmacologically active principle of the pericarp with respect to vascular protection, with other xanthones and phenolic compounds playing secondary or synergistic roles that remain to be fully characterized.

The findings align with a broader body of research on α-mangostin. Previous studies have shown that the compound improves endothelial dysfunction in diabetic mouse models by inhibiting the acid sphingomyelinase/ceramide pathway, attenuates blood pressure and reverses vascular remodeling in hypertensive rats by balancing the renin-angiotensin system axes, and induces vasorelaxation through interactions with large-conductance calcium-activated potassium channels. A 2025 systematic review and meta-analysis further documented lipid-lowering effects of α-mangostin in hyperlipidemic animal models. The new study extends this literature by demonstrating, for the first time in a collagen and epinephrine-induced thrombosis model, that both the whole pericarp extract and its principal xanthone can attenuate vascular occlusion and modulate the adhesive program of the vessel wall.

What remains to be determined is whether the doses used in the animal experiments can be translated into realistic human intake levels and whether long-term supplementation with mangosteen pericarp extract would produce meaningful reductions in cardiovascular events. Human clinical trials would need to address questions of bioavailability, since xanthones are known to undergo extensive metabolism, and potential interactions with conventional antithrombotic medications. The authors caution that their work is a foundation rather than a prescription, but they argue that Garcinia mangostana pericarp extract represents a promising candidate for the prevention of thrombosis and the improvement of blood flow.

The research was supported by the National Research Foundation of Korea and by Chung-Ang University Research Scholarship Grants. As interest in food-based approaches to cardiovascular prevention continues to grow, the humble mangosteen, a fruit whose rind has been discarded as waste for centuries, may find itself at the center of a new chapter in vascular health research, one in which the boundary between food and medicine becomes increasingly productive.

The work addresses a central challenge in carbon capture: finding solid materials that can bind CO₂ efficiently without requiring the large energy inputs associated with regenerating liquid solvents. Adsorption systems work by allowing gas molecules to attach to a solid surface, either through relatively weak physical forces or stronger chemical interactions. Graphene oxide is attractive because its carbon sheets contain hydroxyl, epoxy, carbonyl and carboxyl groups that can interact with CO₂. Yet its performance depends heavily on how intensely the graphite is oxidized, how the sheets restack, and how many accessible pores and chemically active sites remain after synthesis. Excessive oxidation can create defects and collapse useful structure, while insufficient oxidation can leave too few polar sites to attract the quadrupolar CO₂ molecule.

The researchers prepared graphene oxide using a modified Hummers method, an established chemical route in which graphite is oxidized by potassium permanganate in concentrated sulfuric and phosphoric acids. They varied the later preparation temperature to 60, 80 and 100 °C, producing samples called GO 60, GO 80 and GO 100. The use of phosphoric acid and the omission of sodium nitrate follow greener modifications intended to reduce the toxic nitrogen oxides associated with the original Hummers process. The oxidized material was washed repeatedly to remove residual acids and minerals, then dried. In a second step, the team extracted compounds from powdered mangosteen peel using ethanol, evaporated the solvent and mixed the dried extract with graphene oxide in water at 60 °C. The resulting composites were designated GO 60+MP, GO 80+MP and GO 100+MP.

Mangosteen peel is more than a source of biomass: it contains phenolic compounds, flavonoids, tannins, anthocyanins and xanthones, including mangostin derivatives. Many of these molecules carry hydroxyl groups and aromatic rings. Those chemical features could alter graphene oxide in two important ways. First, they can increase the polarity of the surface, strengthening interactions with CO₂. Second, the bulky organic molecules can prevent graphene sheets from packing too tightly, creating or exposing pathways through which gas molecules can diffuse. Microscopy images indicated that the extract adhered to the graphene oxide, producing more folds, wrinkles and visibly heterogeneous regions. The researchers interpret these features as evidence that mangosteen-derived compounds changed the texture and accessibility of the carbon sheets, although the study does not establish the precise molecular bonding arrangement of every extract component.

Several characterization techniques supported the chemical changes. Infrared spectra of the untreated graphene oxide showed signals associated with carbonyl and carboxyl groups, aromatic carbon–carbon bonds, epoxy bridges and alkoxy or alcohol groups. After mangosteen impregnation, the broad hydroxyl signal became more pronounced, consistent with the addition of phenolic compounds and an increase in surface polarity. Raman spectroscopy showed that the disorder-to-graphite ratio, known as the ID/IG ratio, rose from 1.45 to 1.55 as the graphene oxide preparation temperature increased from 60 to 100 °C. After adding the peel extract, the ratio increased further to 1.75, 1.76 and 1.86 for the three temperatures. In graphene-based materials, a higher ID/IG ratio generally indicates more structural disorder or a greater number of defect sites. Such defects are not automatically beneficial, but they can expose reactive edges and create additional locations where gas molecules can bind.

The temperature itself had a pronounced effect on the material’s composition and structure. Energy-dispersive X-ray analysis found that the oxygen content of pristine graphene oxide fell from 31.77 atomic percent in GO 60 to 24.53 percent in GO 80 and 14.03 percent in GO 100, according to the study’s discussion. X-ray diffraction measurements likewise indicated shrinking interlayer distances at higher preparation temperatures, consistent with the loss of oxygen-containing groups between carbon layers. GO 60 had a specific surface area of 52.64 square metres per gram, compared with 12.03 and 14.62 square metres per gram for GO 80 and GO 100. The researchers suggest that 60 °C created a balance: enough oxidation to expand and functionalize the layers, but not so much thermal damage that the structure became less accessible. Adding mangosteen extract raised the measured surface area of all three materials, with GO 60+MP reaching 71.27 square metres per gram.

Nitrogen adsorption measurements revealed why the GO 60 composite performed particularly well. GO 60+MP displayed a type IV adsorption isotherm with a pronounced H2 hysteresis loop, a pattern commonly associated with mesopores and complex pore networks. Mesopores range from 2 to 50 nanometres across and can provide a compromise between high surface area and rapid gas transport. The composite also contained larger voids and slit-like spaces created by partially restacked graphene oxide sheets. Its average pore diameter was 1.37 nanometres, smaller than the 1.97-nanometre average measured for GO 60, suggesting that the extract narrowed some openings while increasing the overall accessible surface. At higher synthesis temperatures, the composites exhibited weaker type III isotherms and lower surface areas—32.98 square metres per gram for GO 80+MP and 27.04 square metres per gram for GO 100+MP.

The most direct test used a fixed-bed reactor packed with 3 grams of adsorbent between layers of quartz wool. Before each experiment, the material was heated under nitrogen, cooled to 30 °C and exposed to a simulated gas stream containing nitrogen and carbon dioxide. The reported testing conditions include a gas flow of 80 millilitres per minute and CO₂ concentrations of either 20 percent in the experimental setup or 15 percent in the capacity comparison and cycling tests. Breakthrough curves track the ratio of outlet to inlet CO₂ concentration: when the ratio is low, the bed is retaining most of the gas; when it approaches one, the material is becoming saturated. Pure graphene oxide reached equilibrium after roughly 5 to 7 minutes, depending on preparation temperature. With mangosteen extract, the corresponding times extended to about 19 to 30 minutes, with GO 60+MP producing the broadest breakthrough curve. Its measured capacity was 1.86 millimoles per gram, compared with 0.51 millimoles per gram for GO 60.

The proposed explanation combines pore structure with surface chemistry rather than attributing the improvement to a single mechanism. CO₂ is a linear molecule with a substantial quadrupole moment, meaning its charge distribution allows it to interact with localized electric fields. Oxygen-containing groups on graphene oxide create surface dipoles that can attract CO₂, while hydroxyl and phenolic groups may support hydrogen-bond-like interactions. Aromatic regions in both graphene oxide and the mangosteen compounds can also interact with CO₂ through its quadrupole and the electron clouds of the rings. At the same time, mesoporous channels can speed diffusion toward internal sites. Kinetic modelling supports a mixed mechanism. Untreated graphene oxide was best described by the Avrami fractional-order model, with correlation coefficients of 0.97 to 0.99, consistent with heterogeneous sites and multiple adsorption pathways. After impregnation, pseudo-first- and pseudo-second-order models both fitted the data well, with coefficients near 0.99, suggesting that the surface treatment changed the apparent distribution and behaviour of adsorption sites rather than converting the process into purely physical or purely chemical capture.

For a carbon-capture material to be useful beyond the laboratory, it must release the stored gas without losing capacity. The researchers regenerated the composites by heating them to 200 °C under pure nitrogen and repeated adsorption and desorption three times. GO 60+MP showed complete regeneration by the third cycle, while the measured capacities remained relatively consistent across the tested cycles. That result is encouraging, but it is not yet evidence of industrial readiness. The experiments used a small fixed bed, a synthetic CO₂–nitrogen mixture and only three cycles; real flue gas contains water vapour, oxygen, sulfur compounds and other contaminants that can compete for active sites or degrade the biomass-derived coating. The chemical oxidation route also uses strong acids and oxidants, so the environmental advantages of recycling mangosteen peel will ultimately depend on reagent recovery, waste treatment, scale-up and the durability of the composite over hundreds or thousands of cycles. Even with those limitations, the study demonstrates a compelling design principle: agricultural waste can supply functional molecules that tune a graphene-based surface, while a moderate synthesis temperature preserves the porous architecture needed to capture CO₂ efficiently.