The COR-AuNPs were obtained by pulsed laser driven CO₂ reduction reaction with a gold target in water enriched with CO₂ by NaOH addition, using an adapted method described elsewhere (Del Rosso et al., 2018; Tahir et al., 2024). In brief, a solution of NaOH at a concentration of 4 mmol.L⁻¹ was prepared in deionized water. A gold target with a purity greater than 99%, supplied by Kurt J. Lesker, was positioned at the base of a Teflon beaker containing 8 mL of the prepared NaOH solution. PLA was performed for a duration of 6 hours in an open air-liquid interface, employing an Nd:YAG source (Q-Smart 850, Quantel United States) operating at a repetition rate of 10 Hz with a temporal width of 5.8 ns, wavelength of 532 nm, and at a fluence on the target of 3.5 J/cm². Following the synthesis process, the amphiphilic copolymer Pluronic F-127 (PF127) was added to the colloidal dispersion reaching a final concentration of 2 mg. mL⁻¹, to enhance the stability of the colloidal system in an environment with high ionic force or contaminants, such as ammonium compounds present in the system water used for acclimatization, and organic matter derived from fish food and biological debris released by the zebrafish themselves, as demonstrated in (Del Rosso et al., 2018).

The size distribution and morphology of the COR-AuNPs was evaluated by transmission electron microscopy (TEM) JEOL Model 2100F. An inductively coupled plasma mass spectrometer (Nexlon 300X Perkin Elmer, United States) was used to determine the metal concentration in the colloidal dispersion. The presence of the CO-Au bond was identified through the use of a micro-Raman spectrometer (HORIBA XploRA), with the sample previously dried on a glass slide to obtain the surface-enhanced Raman spectra (SERS), similarly as reported in (Muniz-Miranda et al., 2011). This technique is employed as a method to obtain the signal related to chemical bonds at remarkably low concentrations by exploiting the surface plasmon resonance of the COR-AuNPs, which leads to an enhancement of the electromagnetic field in a confined area near the COR-AuNPs surface. In optimal conditions, this approach can achieve an enhancement in sensitivity to single molecule levels (Kneipp et al., 1997), thus serving as a pivotal method for investigating low-concentration products.

The stability of COR-AuNPs in the zebrafish water environment was evaluated by monitoring the UV-Vis extinction spectra obtained by a double beam Lambda 950 Perkin Elmer spectrophotometer (United States). After the specimens were acclimatized, aquarium water was collected and used as the nanomaterial environment. The colloidal dispersion at concentration of 42 × 10³ μg.L⁻¹ was added to the fresh water (50:50 v/v), and the UV-Vis spectra was monitored for 6 weeks at room temperature. The colloidal stability of the COR-AuNPs in cell culture medium was evaluated for 24 h by Dynamic Light Scattering measurement using the intensity distribution as a function of the hydrodynamic diameter. The analysis was carried out at 25 °C using an SZ-100 analyzer from Horiba Instruments (Kyoto, Japan), with a scattering angle of 90°, using a 10 mW laser with a wavelength of 532 nm. For the analysis, the COR-AuNPs were dispersed in a phosphate buffer solution (PBS) with pH 7.4, and RPMI + 10% FBS culture medium at ratio of 3:1 (nanoparticle:culture medium/PBS). COR-AuNPs containing 2 mg.mL⁻¹ of the amphiphilic block copolymer Pluronic-F127 (PF127) were used for the test.

The SaOS-2 cell line (BCRJ-0217) was obtained from the Rio de Janeiro Cell Bank (https://bcrj.org.br/) packed in frozen ampoules and kept in liquid nitrogen. After thawing, cells were grown in McCoy’s 5A medium (Sigma-Aldrich) supplemented with 10% FBS (BioNutrientes) in an incubator at 37 °C in a humidified environment (CellXpert^® C170i, Eppendorf, Hamburg, Germany) with 5% CO2 atmosphere. Sterility was ensured by tests for bacteria, fungi, and mycoplasmas. For bacteria and fungi, the cell culture supernatant was placed in thioglycolate (TIO) (Acumedia, Baltimore, United States) and tryptic soy broth (TSB) (Acumedia, Baltimore, United States) and incubated aerobically for 14 days at 22.5 °C ± 2.5 °C and 32.5 °C ± 2.5 °C, respectively. Mycoplasma contamination in the cell supernatants was investigated by bioluminescence using the MycoAlertTM PLUS Mycoplasma Detection Kit (MycoAlert^®, Lonza, Verviers, Belgium) and all samples were confirmed to be free of contamination. The cytotoxicity was evaluated according to ISO 19007:2018 – in vitro MTS assay for measuring the cytotoxic effect of nanoparticles. SaOS-2 cells were seeded in 96-well plates at a density of 3 × 10⁴ cells per well corresponding to a density of about 93.750 cells per cm² (each well ∼ 0,32 cm²) and maintained for 24 h in McCoy’s 5A medium (Sigma-Aldrich) supplemented with 10% FBS (BioNutrientes) in an incubator at 37 °C with a 5% CO₂ atmosphere. Subsequently, the cells were treated with COR-AuNPs at concentrations of 5 and 10 × 10³ μg.L⁻¹. After 24 h of exposure, the culture medium was discarded, and 200 µL of 0.01 M PBS (Sigma-Aldrich) was added (3×) to wash the cells and remove excess nanoparticles. To assess cell viability, the commercial CellTiter 96^® AQueous Non-Radioactive Cell Proliferation Assay (MTS) (Promega) was used. The MTS reagent was prepared in phenol red-free culture medium, and 120 µL was added to each well across the plate. The incubation time was 1 h at 37 °C, after which absorbance was measured at 490 nm using a Biotek Synergy H4 microplate reader (Aligent, Santa Clara, California, United States). After the initial absorbance reading, the supernatant was collected, subjected to centrifugation at 25.155 RCF for 90 min, and absorbance was measured again. This step was designed to minimize potential interferences. There was no significant difference between the readings obtained before and after centrifugation. The negative control consisted of cells cultured in complete medium alone (vehicle). The positive control was performed with 20% DMSO to induce cell death.

Endothelial Colony Forming Cells (ECFCs) were isolated from Human Umbilical Cells (>50 mL) of healthy newborns, as described elsewhere (Margheri et al., 2011; Armanetti et al., 2021), with informed consent from the mothers (R711-D) and observing the Italian legislation (art. 2, paragraph 1, letter f, decree of 18 November 2009). The cells were analyzed by flow-cytometry for expression of the antigens CD45, CD34, CD31, CD105, ULEX, vWF, KDR, and uPAR. ECFCs were cultured in 10% fetal bovine serum (FBS, Euroclone) supplemented EGM2 cell culture media, incubated at 37 °C with 5% CO₂ saturation.

ECFCs were seeded in 6 cm dishes at a density of 2 × 10⁴ cells/cm² and cultured in a humidified incubator at 37 °C with 5% CO₂. After cell attachment, they were incubated with EGM-2 medium (3 mL/well) containing COR-AuNPs at concentrations of 5 × 10³ and 10 × 10³ μg·L⁻¹ for 24 h. Cytotoxicity was determined by trypan blue staining: 20 µL of cell suspensions were resuspended with an equal volume of 0.4% (w/v) trypan blue solution prepared in 0.81% NaCl and 0.06% (w/v) dibasic potassium phosphate. Viable and non-viable cells (trypan blue positive) were counted separately using a dual-chamber hemocytometer and a light microscope. Optical microscopy was used to assess qualitative intracellular uptake of COR-AuNPs. Due to their strong light-scattering properties and high electron density, AuNPs appear as dark, punctate regions within the cytoplasm under brightfield illumination. Cellular images were acquired through the EVOS xl core microscope (AMG, Advanced Microscopy Group).

To assess carbon monoxide (CO) release from the nanoparticles, we quantified carboxyhemoglobin (COHb) levels using a specific ELISA assay. COHb, formed by the binding of CO to hemoglobin, serves as a reliable indicator of intracellular CO presence. For this purpose, we employed human chronic myeloid leukemia K562 cells (DSMZ), which were induced toward erythroid differentiation using 1 µmol.L⁻¹ Imatinib for 5 days (Jacquel et al., 2003). These cells act as erythroid progenitors capable of synthesizing hemoglobin, providing a biologically relevant model for evaluating CO delivery. Following the treatment, the cells were exposed to 15 × 10³ μg.L⁻¹ of COR-AuNPs, or to 100 μM solution of CORM-2 (Sigma-Aldrich). Cell lysates were analyzed using the COHb ELISA kit (96-well format, MyBioSource, MBS7254040) according to the manufacturer’s instructions, enabling sensitive and specific detection of CO-bound hemoglobin within a cellular context that mimics erythropoiesis. The instructions include a final centrifugation process (at 1000 RCF for 15 min at 2 °C–8 °C), which eliminate the potential interference of the extinction of the AuNPs in the measurement.

In vitro capillary morphogenesis was performed in tissue culture wells with Matrigel (BD Biosciences) coating. A total of 18.000 cells were seeded per well, corresponding to a density of 60.000 cells/cm² (well growth area: 0.3 cm²) in 2% FBS supplemented EGM-2 culture medium, and incubated at 37 °C at 5% CO2 atmosphere. The pictures were acquired at regular intervals at EVOS optical microscope (Thermo Fisher Scientific, Monza, Italy). The Angiogenesis Analyzer tool of ImageJ software19 provided the statistical analysis for each experimental condition tested quantifying nodes, master junctions and meshes Nodes are identified as pixels that have at least three neighbors, corresponding to a bifurcation. Master junctions are element junctions linking at least three segments. They delimited the master segments”. Meshes are the polygon structures reinforced with more than one layer of cells in their walls and have also been referred to by other authors as a “Honeycomb formation”.

This study was approved by the Animal Use Ethics Committee of the Oswaldo Cruz Institute (protocol 035/2022 and license L-001/2023-A1, valid until 2026). Adult zebrafish (D. rerio) used in the experiments had an average weight of 0.4 g and length of 3.5 cm. Fish were maintained in glass aquaria with dechlorinated water (pH 7.0–7.6; 23 °C–27 °C), under controlled ammonia levels and a 14:10 h light:dark photoperiod. They were fed with Alcon^® Neon feed (Alcon, Camboriú, Brazil) via automatic digital feeders. Acclimation lasted 7 days, in accordance with ABNT NBR 15088 guidelines (ABNT NBR, 2022).

Following acclimation, males and females (4.17 ± 0.35 cm; 2.26 ± 0.45 g) were randomly transferred to 36 L exposure aquaria (49.5 × 30 × 24.5 cm), at a density of 1 g of fish per liter. Eighteen fish were placed in each tank, with three replicates per group, totaling 324 animals. Six fish per group were allocated for genotoxicity analysis via the comet assay.

For the acute exposure test, a COR-AuNPs stock dispersion was serially diluted to obtain final concentrations of 5, 10, 20, 35, and 75 µg·L⁻¹. All concentrations were sublethal and based on previous toxicological data. Souza et al. (2021) reported complete lethality in zebrafish at concentrations >300 µg·L⁻¹ of gold nanorods (AuNRs), while 75 µg·L⁻¹ was considered sublethal. Thus, 75 µg·L⁻¹ was adopted as the highest test concentration. Additionally, Gong et al. (2016) demonstrated physiological effects without lethality at 60 µg·L⁻¹ of cobalt in CO-releasing molecules (CORMs), supporting the sublethal design of this experiment. Despite differences in metal core composition, the common mechanism of CO release ensures biological comparability and relevance.

The exposure period lasted 96 h, with solution renewal at 48 h to maintain contaminant levels, establishing a semi-static system (Ge et al., 2015). Control groups were maintained under identical conditions, but without COR-AuNPs.

At the end of the 96-h exposure, fish were euthanized in accordance with ethical guidelines established by the Brazilian National Council for the Control of Animal Experimentation (CONCEA, 2015), and aligned with international standards (Wendy Underwood, 2020; National Cancer Institute, 2023). The procedure was performed by immersion in an ice-water solution (5:1 ratio), maintained between 0 °C and 4 °C. Death was confirmed by the absence of opercular (gill) movements, and the animals were kept in the solution for at least 10 min to ensure complete loss of vital signs. This method was chosen to avoid the use of chemical anesthetics, which could interfere with subsequent biochemical analyses, as previously described by Borski and Hodson (2003).

After euthanasia, liver and brain samples were collected, and 30 mg of each tissue were selected. A total of 1,500 µL of 50 mmol·L⁻¹ potassium phosphate buffer (pH 7.0) was added. The buffer was prepared using ultrapure water, monobasic potassium phosphate (KH₂PO₄; Sigma-Aldrich, St. Louis, MO, United States) and dibasic sodium phosphate (Na₂HPO₄; Merck, Darmstadt, Germany). The tissue was homogenized for 60 s using a stainless steel spatula. Samples were then centrifuged at 9.400 RCF for 10 min at 4 °C using a 5430R centrifuge (Eppendorf, Hamburg, Germany). The resulting supernatant was transferred to new microtubes (Eppendorf, Hamburg, Germany) and stored at −80 °C until further analysis. The protocol was adapted from Yan et al. (2015) and Yan et al. (2016).

Genotoxicity was assessed using the alkaline comet assay, adapted from OECD Test Guideline 489 for the zebrafish (D. rerio) model (OECD Science, 2016). Six fish per experimental group were used, and blood was collected by caudal vein puncture, forming pools by contaminant concentration. Each pool (10 µL) was mixed with 3 µL of heparin (Parinex^®; Hipolabor, Belo Horizonte, Brazil) and embedded in low melting point agarose (LMP agarose; Sigma-Aldrich, St. Louis, MO, United States) at 37 °C. The mixture was applied onto slides previously coated with normal melting point agarose (NMP agarose; Sigma-Aldrich) at 60 °C. After solidification (4 °C ± 2 °C), slides were lysed for 24 h at 4 °C ± 2 °C in a buffer containing sodium chloride (2.5 mol·L⁻¹; Synth, Diadema, Brazil), EDTA (0.1 mol·L⁻¹; Vetec, Duque de Caxias, Brazil), Tris (0.01 mol·L⁻¹; Sigma-Aldrich), Triton X-100 (Sigma-Aldrich), and dimethyl sulfoxide (DMSO; Dinâmica, Indaiatuba, Brazil). Electrophoresis was conducted in alkaline buffer (10 mol·L⁻¹ sodium hydroxide and 0.2 mmol·L^-1 EDTA) at 0 °C–4 °C, with 20 min of DNA unwinding followed by 20 min of electrophoresis at 25 V and 300 mA. Slides were then neutralized with PBS (pH 7.4), fixed in absolute ethanol (Dinâmica), stained with silver nitrate (Sigma-Aldrich), and dried at room temperature. Comet images were captured using an optical microscope (Kasvi–Motic, ×400 magnification) and analyzed using Motic Image Plus 2.0^® software. At least 200 nucleoids per slide (five slides per group) were evaluated. DNA damage was assessed based on tail intensity, tail moment, and classified using a five-grade scale (0–4), according to Noroozi et al. (1998). Negative controls consisted of animals maintained in aquarium water under the same experimental conditions, without exposure to COR-AuNPs (see Supplementary Figure S1 in Supplementary Material). Although a positive control was not included in this experiment, our group routinely validates the assay using methyl methanesulfonate (MMS; Sigma-Aldrich) in various models, ensuring methodological reliability. This assay was performed following a protocol previously validated by our group (Sales Junior et al., 2020; 2021; 2024), in accordance with the Minimum Information for Reporting on the Comet Assay (MIRCA) guidelines (Møller et al., 2020), ensuring reproducibility and transparency of the results (Supplementary Table S1 on Supplementary Material).

All statistical analyses were performed using GraphPad Prism software (version 9.5.1; GraphPad Software, San Diego, CA, United States). Data normality was assessed using the Shapiro–Wilk test (α = 0.05). For datasets with normal distribution, one-way ANOVA followed by Dunnett’s multiple comparisons test was applied. In cases where normality was not confirmed, non-parametric analyses were conducted using the Kruskal–Wallis test followed by Dunn’s post hoc test. All biomarker datasets from the zebrafish experiments failed to meet the normality assumption and were therefore analyzed using non-parametric methods. Results are expressed as median ± standard error of the mean (SEM), and statistical significance was set at p < 0.05. For the SaOS-2 assays, data normality was verified using the Shapiro–Wilk test, followed by one-way ANOVA with multiple comparisons. Results are presented as mean percentage values relative to non-exposed control wells (considered as 100%), which were cultured in medium alone.

The PLA process resulted in the formation of well-defined spherical nanoparticles with a narrow distribution, measuring less than 10 nm (Figure 1A). As the synthesis process was performed in batch without a water flowing system, the nanoparticles were also subjected to laser fragmentation in liquid subsequent to their production, resulting in the formation of ultrasmall nanoparticles with an average radius of 1.9 nm, as previously reported (Tahir et al., 2024).

As stated in the introduction, the pulsed laser driven CO₂ reduction reaction results to the synthesis of gold nanoparticles rich in CO (Del Rosso et al., 2016) and organometallic nanocomposites constituted by gold nanoclusters embedded in formic, acetic and lactic acids, with concentrations of about 0.3, 1.4, and 0.2 ppm, respectively (Tahir et al., 2024). The UV-Vis spectra of the pristine colloidal dispersion of nanoparticles exhibit the characteristic localized surface plasmon resonance (LSPR) peak at 514 nm, which shifts to 524 nm after the addition of the amphiphilic copolymer block Pluronic-F127 (PF127) and the subsequent incubation in the zebrafish aquarium. As visible in Figure 1B, the shape and full width half maximum (FWHM) of the LSPR curve is not affected by the zebrafish water environment, and we only notice a slight decrease in the intensity of the maximum, probably associated to a small portion of nanoparticles (less than 10%) precipitated after the incubation in the aquarium. The SERS measurement demonstrated the existence of gold-carbonyl (Au-CO) chemical bonds, thereby confirming the presence of CO associated with the COR-AuNPs (Figure 1C).

The copolymer PF127 was utilized as a method to enhance the stability of the COR-AuNPs, as demonstrated in previous study (Del Rosso et al., 2018). In addition to the afforementioned effect, PF127 has applications as a drug delivery system (Moore et al., 2000), facilitating the penetration of nanocomposites into cancer cells (Elagin et al., 2014; Simon et al., 2015). In this study, we assess the stability of the NPs in zebrafish aquarium water by monitoring the variation in intensity of the extinction peak of the COR-AuNPs (Figure 2A) over a 6-weeks period. It can be observed that the NPs were markedly stable, with the peak extinction reducing by only 6% over the course of the experiment (1 week), and by only 10% within 6-weeks (Figure 2B), while the FWHM had a variation of 0.8% and 7.6% indicating minimal agglomeration after their dispersion in the zebrafish water (Sangwan and Seth, 2022). The physicochemical stability of the COR-AuNPs is a fundamental requirement for their applicability in biological systems (Sharma et al., 2021), as instability can lead to aggregation, which in turn alters the nanoparticle’s biodistribution, bioavailability, and potential toxicity (Abbasi et al., 2023).

A dynamic light scattering (DLS) analysis of COR-AuNPs in phosphate buffered saline (PBS) and McCoy’s 5A cell culture medium supplemented with fetal bovine serum (FBS) measurements were taken at 0 h for both conditions, and after 24 h for COR-AuNPs in McCoy’s 5A + FBS. Peaks centered at 13 nm and 49 nm were observed for COR-AuNPs in PBS and cell culture medium, respectively, indicating a possible interaction with FBS. Following a 24-hour period of incubation in cell medium, we observed a small shift of the peak to 51 nm, indicating that the COR-AuNPs functionalized by PF127 are stable in complex biological media (Supplementary Figure S2 on Supplementary Material).

The characterization of the COR-AuNPs was completed by measuring the release of CO in cells using the Elisa COHb test on human chronic myeloid leukemia cells (K562). The results of CO release are reported in Figure 3, showing the formation of COHb when K562 cells were treated with both CORM-2 and COR-AuNPs. The CO release profiles exhibited significant variation over the duration of one hour. The release rate of CORM-2 was found to be accelerated, and after a duration of 6 h, the concentration of COHb was observed to be elevated in the presence of COR-AuNPs. The ensuing results illustrate the disparities in the time-dependent mechanisms of CO release for the structures in question. Indeed, extant literature on the subject indicates that the release of carbon monoxide (CO) by CORM-2 occurs extracellularly and commences within minutes of its dispersion in the culture medium (Southam et al., 2021). Conversely, the COR-AuNPs necessitated a duration exceeding 60 min for the identification of COHb, thereby indicating an intracellular initiation of the CO release process.

The laser-synthesized COR-AuNPs presented herein offer several advantages over most commercial CORMs that frequently contain toxic metals (e.g., Cr, B, Fe, Mn, Co, Mo, Ru, W, Re, Ir) (Herrmann, 1990; Motterlini et al., 2002; Nobre et al., 2007; Zhang et al., 2009; Crook et al., 2011; Kautz et al., 2016), require protection from light, must be stored at −20 °C (Pizarro et al., 2009; Crook et al., 2011), and present low solubility in water (Motterlini et al., 2002; Bannenberg and Vieira, 2009). In contrast, the synthesis of COR-AuNPs employs a one-step, eco-friendly technique that utilizes solely gold, water and aqueous CO₂. These nanoparticles exhibit stability at room atmosphere and can be stored at room temperature on the shelf, as evidenced by the COHb levels measured after a few months of storage, that remained consistent, indicating the preservation of the CO release property (Chillàet al., 2025).

Considering the prospective applications of CO delivery systems in the treatment of vascular diseases (Dulak et al., 2008; Choi et al., 2022), we assessed the cytotoxicity of the nanoparticles in endothelial colony forming cells (ECFCs), a subtype of Endothelial Progenitor cells (Figures 4A,B), as well as in the osteogenic sarcoma cell line SaOS-2.

To further investigate their impact on angiogenesis, a capillary morphogenesis assay was performed using ECFCs. This assay evaluated both nanoparticle uptake (Figure 4C) and their effect on the formation of capillary-like structures (Figure 4D). A clear increase in the number of master junctions, meshes, and master nodes (Figure 4E) was observed. As illustrated in Figures 4A,B, the cells were exposed to vehicle or COR-AuNPs at concentrations of 5 and 10 × 10³ μg.L⁻¹, and no toxicity was observed after 24 h of incubation. The outcomes can be attributed to the laser synthesis process (Tahir et al., 2024), which does not involve the use of toxic solvent or reagents, and relies exclusively on the physical interaction between light, the target, and the molecules in the liquid environment. These results are consistent with our previous studies on RAW267.4, NCTC and HMVEC cells (Tahir et al., 2024).

In addition to ECFCs, we investigated the effects of COR-AuNPs on SaOS-2 osteosarcoma cells, that did not show cytotoxic potential since the cell solutions offered relative viability greater than 70% in all conditions tested after 24 h of exposure (Figure 4B), when compared to the negative control (untreated cells). Considering that effective bone regeneration is closely linked to neovascularization (Dalle Carbonare et al., 2025), this approach aims to explore the potential applicability of these nanoparticles in regenerative therapies across different tissue types.

The application of metallic nanostructures in biological models (cells and animals) has generated a substantial body of literature, often with conflicting findings due to variations in synthesis methods and experimental models. The use of different concentration ranges in in vitro and in vivo assays reflects model-specific sensitivities and endpoints. Furthermore, it is essential to consider the physicochemical properties of the nanoparticles, which are inherently determined by the synthesis method. Table 1 provides a direct comparison between the synthesis methods used in prior studies and the current approach.

PLA was chosen for this study due to its unique synthesis process, which does not require the use of reagents or surfactants. This distinguishes it from other chemical methods, such as the Brust–Schiffrin protocol (Brust et al., 1994), or ionic liquid-based syntheses (Kim et al., 2004), which do require the use of these substances. These methods still require complex purification to remove residual chemicals, in contrast to the relatively uncomplicated process involved in PLA synthesis.

In comparison with biologically mediated synthesis routes, PLA offers superior reproducibility and better control over key synthesis parameters, including particle size distribution, morphology, and chemical composition. This is particularly especially pronounced in the context of biogenic approaches, which frequently yield nanoparticle formulations that are characterized high polydispersity, a broad spectrum of morphological variants, and inconsistent surface chemistry. For instance, Noruzi et al. (2011) and Shankar et al. (2004) reported the formation of heterogeneous mixtures of gold nanostructures when using biological extracts as reducing and capping agents.

Conversely, laser ablation in liquids facilitates the direct synthesis of nanoparticles in a pristine environment, circumventing the introduction of extraneous contaminants and ensuring greater batch consistency, a prerequisite for biomedical and toxicological investigations.

Furthermore, the physical and chemical characteristics of the nanostructure have been demonstrated to exert a substantial influence on its biological response when exposed to cells or organisms. It has been observed that surface chemistry plays a pivotal role in determining the biological effects of AuNPs. Studies on cellular responses has shown that commonly used stabilizers, such as citrate, exhibit negligible toxicity in most cases (Connor et al., 2005; Mironava et al., 2010; Tsai et al., 2013), whereas charged surfactants, particularly those with a positive charge, significantly reduce biocompatibility (Schaeublin et al., 2011). For instance, AuNPs functionalized with CTAB (cetyltrimethylammonium bromide) exhibited higher levels of cellular toxicity in comparison to gold nanorods (AuNRs) in both the Human Dermal Fibroblast and U87 cell lines (Bhamidipati and Fabris, 2017). Furthermore, when CTAB-stabilized AuNRs were exposed to HeLa cells, they exhibited severe toxicity, which was mitigated by replacing the surfactant (Mallick et al., 2013). These findings underscore the dominant role of surface ligands and charge in modulating the biocompatibility of AuNPs.

A comparable trend is evident in zebrafish models, where charged surfactants have been associated with augmented toxicity (Truong et al., 2013). The morphology of AuNPs has been demonstrated to play a significant role in their toxicological profile, particularly at in vivo models. In zebrafish, exposure to non-spheroidal nanoparticles resulted in a more pronounced toxic response compared to exposure to spherical particles (Patibandla et al., 2018; Souza et al., 2021). While analogous patterns are discerned in cell models, in vivo systems manifest a heightened sensitivity to structural disparities. A number of studies have suggested that AuNPs may exert neuroprotective effects, while others have reported embryotoxicity, including morphological abnormalities and impaired development (Patibandla et al., 2018; Verma et al., 2018; Torres et al., 2021). These findings suggest that nanoparticle shape can significantly influence biological interactions, especially during early developmental stages.

The impact of particle size on the toxicity of AuNPs is a complex matter, as there is a divergence of opinions on the subject. Some evidence has suggested that larger nanoparticles induce greater cytotoxic effects (Mironava et al., 2010), while others have suggested that smaller nanoparticles exhibit higher toxicity due to their greater surface area and reactivity (Broda et al., 2016). These discrepancies underscore the intricacy of size-dependent effects and the impact of confounding variables, such as surface chemistry, agglomeration state and synthesis method. Despite the presence of conflicting findings regarding morphology and size, a clear and consistent trend across studies is the dose-dependent increase in cytotoxic effects. It has been observed that elevated concentrations of AuNPs frequently correlate with augmented toxicity, both cellular and systemic (Connor et al., 2005; Mironava et al., 2010; Schaeublin et al., 2011; Tsai et al., 2013; Broda et al., 2016; Bhamidipati and Fabris, 2017). In zebrafish, for instance, exposure to AuNP resulted in both dose-dependent toxicity and tissue-specific bioaccumulation in the brain, muscles, gills, and gastrointestinal tract, evaluated over a 60-day period (Geffroy et al., 2012). This accumulation may be associated with long-term toxic effects, including modulation in the expression of genes related to DNA repair and detoxification pathways (Geffroy et al., 2012; Dedeh et al., 2015). Consequently, further studies with the COR-AuNPs synthesized in the present work are warranted to evaluate longer exposure periods and potential cumulative effects.

Another relevant aspect concerns the selection of sublethal endpoins. Torres et al. (2021); Sadhukhan et al. (2022) reported that AuNPs can modulate endogenous antioxidant defenses and influence the activity of key enzymes in zebrafish. Importantly, (Torres et al., 2021), observed that these parameters returned to baseline after exposure, indicating a transient and reversible sublethal biological response (Torres et al., 2021). In line with these findings, our study also focused on sublethal endpoints, evaluating sensitive oxidative stress biomarkers and early indicators of cellular damage, including lipid peroxidation (MDA), protein carbonylation (PTC), and DNA damage (comet assay), and found no significant alterations following acute exposure to COR-AuNPs.

This study offers a significant advancement in the field by assessing the toxicity of gold nanoparticles produced by pulsed laser ablation, enriched with oxocarbons, in zebrafish. Notably, the synthesis approach employed in this research is devoid of chemical contaminants, a notable feature that enhances the study’s rigor and reliability. In the literature, Gong et al. (2016) reported that cobalt-based CORMs did not induce significant toxicity in zebrafish embryos at concentrations below 1.0 µM (equivalent to about 60.0 µg·L⁻¹ of Co) Gong et al. (2016). Comparatively, in the present study, COR-AuNPs were tested at concentrations ranging from 5 to 75 µg·L⁻¹, without causing mortality or significant alterations in oxidative stress and genotoxicity biomarkers. These results reinforce that CO-releasing nanomaterials, when properly stabilized and free of toxic ligands, may exhibit a safe profile at concentration ranges similar to those previously established for cobalt-based CORMs.

Therefore, the present study demonstrates that COR-AuNPs produced by pulsed laser ablation exhibit high biocompatibility in zebrafish, with no significant acute toxicity or sublethal effects observed at the concentrations that were tested. These results are consistent with previous findings for cobalt-based CORMs and highlight the critical influence of synthesis method, surface chemistry, and particle size on the biological responses to metallic nanoparticles. Furthermore, subsequent studies should encompass conventional gold nanoparticles devoid of CO-release properties to distinguish the effects attributable to CO from those of the nanoparticle core.