Assessing the Biosafety and Antibacterial Efficacy of Graphene and Graphene Oxide for Orthopedic Implant Applications

Abstract

Graphene (G) and graphene oxide (GO) nanoparticles are emerging as promising modifiers for orthopedic implant surfaces, yet their safety and antimicrobial profiles remain incompletely defined. We evaluated in‑vitro cytotoxicity on bone marrow mesenchymal stem cells (BMSCs) and in‑vivo biocompatibility by intramuscular implantation in mice. Results identified 10 µg/mL as the threshold below which G and GO exhibit minimal cytotoxicity. Concentrations above this value induced dose‑dependent cell death, with GO proving more toxic than G. Antimicrobial assays against Staphylococcus aureus revealed that G only inhibits bacterial growth at ≥100 µg/mL, whereas GO shows significant antibacterial activity at ≥50 µg/mL. The 50–100 µg/mL range appears optimal, balancing low cytotoxicity with robust antibacterial efficacy, supporting clinical translation within this dosage window.

Background

Orthopedic implants must combine excellent biosafety with effective infection control, particularly against S. aureus, the most frequent pathogen in implant‑related infections. Conventional therapies often fail, leading to prolonged healing or amputation. Nanotechnology offers new avenues; graphene and its derivatives, with their two‑dimensional carbon lattice, possess unique mechanical, electrical, and chemical properties that can be harnessed for bone regeneration and antibacterial action.

Graphene’s sharp edges can physically disrupt bacterial membranes, while charge transfer mechanisms further inhibit bacterial proliferation. GO, functionalized with oxygen groups, enhances dispersion and can serve as a drug carrier, imaging agent, or antimicrobial surface. However, the same nanostructure that confers therapeutic benefits can also elicit cellular toxicity, especially at higher doses or due to protein corona formation.

Prior studies have reported conflicting cytotoxic thresholds for G and GO. While some data suggest G is relatively benign, GO often displays higher toxicity toward mammalian cells, potentially due to its surface chemistry. Clarifying these dose–response relationships is essential for safe clinical deployment.

Results

In‑Vitro Cytotoxicity of G and GO

Transmission electron microscopy revealed irregular G and GO flakes averaging 30 ± 5.6 nm, with occasional agglomerates (Figure 1a–b). After 7 days of co‑culture, BMSCs maintained a spindle morphology; osteogenic and adipogenic differentiation confirmed multipotency (Figure 1c–e).

G and GO cytotoxicity: TEM images of G (a) and GO (b) show nanonetworks; BMSC morphology (c); calcium deposition (d); lipid accumulation (e); cell viability after exposure to varying concentrations (f). Statistical significance: * p < 0.01 vs. control; # p < 0.01 vs. 10 µg/mL G; ○ p < 0.05 vs. 50 µg/mL G; ☆, □, △ p < 0.01 vs. same concentration G. Correlation coefficients: r² (G) = 0.843; r² (GO) = 0.939.

Cell viability declined in a dose‑dependent manner once concentrations exceeded 10 µg/mL. GO exhibited greater cytotoxicity than G at matched doses, with the disparity widening at higher concentrations (Figure 1f). Scanning electron microscopy corroborated these findings: at 10 µg/mL, BMSCs adhered and spread normally; at 50 µg/mL G, cells shrank and displayed increased surface secretion; at 50 µg/mL GO, cells were markedly deformed and largely non‑viable (Figure 2).

SEM images of BMSCs with G and GO. (a) G 10 µg/mL – healthy cells; (b) G 50 µg/mL – reduced size, elongated microvilli; (c) GO 50 µg/mL – shrunken, deformed cells.

Transmission electron microscopy confirmed internalization of both G and GO. At concentrations above 50 µg/mL, GO induced pronounced intracellular disruption, including disordered organelles and microvilli loss (Figure 3).

TEM images of BMSCs with G and GO. (a) G – moderate structural disruption; (b) GO – extensive cytotoxic changes.

Collectively, 10 µg/mL was identified as the critical safety threshold; GO displayed higher cellular toxicity beyond this level.

In‑Vivo Cytotoxicity

Muscle tissue implanted with G preserved normal architecture, with parallel myofibrils and intact nuclei (Figure 4a–b). In contrast, GO‑treated tissue exhibited fractured transverse lines and signs of atrophy and necrosis, indicating higher in‑vivo toxicity (Figure 4c).

HE‑stained muscle sections: (a) control – intact fibers; (b) G – preserved structure; (c) GO – disrupted fibers and necrosis.

Antibacterial Properties

In‑Vitro

Bioluminescent imaging of Staphylococcus aureus (Xen‑29) demonstrated dose‑dependent inhibition by both G and GO. G required ≥100 µg/mL for significant reduction; GO achieved inhibition at ≥50 µg/mL. Photon intensity (PI) curves confirmed stronger antibacterial action of GO (Figure 5).

Photon intensity of Xen‑29 in vitro: (a) time course; (b–d) statistical correlations. Significance markers: * p < 0.01 vs. control; # p < 0.01 vs. control; ☆, □, △, ○ p < 0.01 vs. same concentration G.

In‑Vivo

In mouse muscle, GO at 100 µg/mL produced a markedly lower PI at both 0 and 24 h, whereas G did not differ significantly from control (Figure 6). Thus, GO retained robust antibacterial efficacy in vivo, while G’s effect was negligible at this dose.

Photon intensity in vivo: (a) time course; (b–c) PI at 0 and 24 h. Significance: # p < 0.01 vs. control.

Discussion

The study confirms that both G and GO exhibit dose‑dependent cytotoxicity toward BMSCs and skeletal muscle, with GO consistently more toxic. Antibacterial activity likewise follows a dose response, with GO outperforming G in both in‑vitro and in‑vivo settings. The 50–100 µg/mL window appears to balance minimal cytotoxicity with significant antibacterial action, making it a promising target range for implant surface modification.

Mechanistic insights suggest that G’s sharp edges and charge transfer properties disrupt bacterial membranes, while GO’s surface oxygen groups facilitate stronger interactions with microbial cells and possibly induce reactive oxygen species production. The higher mammalian toxicity of GO may stem from protein corona formation and oxidative stress, underscoring the need for surface modifications that mitigate adverse effects without compromising antimicrobial function.

Future work should explore GO functionalization to reduce cytotoxicity, assess long‑term biodegradation, and evaluate effects on other tissues. Understanding the interplay between physicochemical attributes and biological responses will be crucial for advancing G and GO‑based nanomedicines.

Methods

Animals

Male Sprague‑Dawley rats (4 weeks) provided BMSCs; male Balb/C mice were used for implantation studies. All procedures complied with institutional animal care guidelines.

Graphene and Graphene Oxide Preparation

Commercially sourced G and GO (1–2 layers) were dispersed in ethanol, PBS, or saline, then ultrasonicated for 2 h to achieve a 1 mg/mL stock solution.

Cell Culture and Cytotoxicity Assays

BMSCs were expanded to passages 3–5, then exposed to 0, 10, 50, 100, 500, or 1000 µg/mL of G or GO for 24 h. AlamarBlue® assays quantified metabolic activity; SEM and TEM provided morphological insights.

In‑Vivo Implantation

G or GO (10 µg/mL) was injected into the medial femoral muscle of mice; tissues were harvested after 7 days for HE staining.

Antibacterial Testing

Bioluminescent S. aureus (Xen‑29) was cultured at 10⁷ cfu/mL and co‑treated with G or GO at the same concentration ranges. Photon intensity was recorded at 0, 8, and 24 h (in‑vitro) and at 0 and 24 h (in‑vivo) using the IVIS Lumina II system.

Data Analysis

Results are expressed as mean ± SD. Student’s t‑test and one‑way ANOVA evaluated statistical significance; p < 0.05 was considered significant.

Conclusions

Graphene and graphene oxide both demonstrate dose‑dependent cytotoxicity and antibacterial effects. GO’s superior antibacterial potency at 50–100 µg/mL, coupled with manageable toxicity, positions it as a strong candidate for orthopedic implant coatings, provided that surface modifications can further reduce cellular adverse responses.