Due to their unique physicochemical properties, nanoparticles (NPs) with an overall size of between 1 and 100 nm in at least one dimension are now widely employed in industrial, biomedical and cosmetic applications, for example, as catalysts and fillers as well as nanoparticulate biomaterials for orthopaedic surgery, UV filters in sunscreens and food additives (Jia and Schüth, 2011; Stark et al., 2015; Bouwmeester et al., 2014). Novel biomedical uses include drug, vaccine and gene delivery and diagnostics and bioimaging (Mabrouk et al., 2021). In vivo, NPs show enhanced solubility and bioavailability, including the ability to cross the blood-brain barrier, to enter the pulmonary system and to undergo enhanced absorption through tight junctions (Kohane, 2007), making them ideal drug delivery vehicles where superior drug bioavailability as well as target-specificity is required (Patra et al., 2018; Eker et al., 2024).

Despite these favourable characteristics, NP toxicity is still a concern. Toxicity is commensurate to the chemical nature of the core material, the NPs’ size and shape (Liu et al., 2022) their ability to generate reactive oxygen species (ROS), disrupt cellular compartments and induce immune responses (Ngobili and Daniele, 2016; Pondman et al., 2023). Hence, standardised methods for biosafety evaluation of newly formulated NPs are urgently required (Drasler et al., 2017; Moya-Andérico et al., 2021). In vitro studies using cell lines have shown that toxicity in response to NP uptake is strongly dependent on cell type (Kroll et al., 2011). For example, identical concentrations of SPIONs induced significant cytotoxicity in neuronal, glial and lung cells, but minimal toxicity in all other tested cell types (Li et al., 2013; Wei et al., 2021). Therefore, in vivo experiments are essential when investigating the biological effects of NPs on the regulation of pH, ionic strength and chemical composition of circulating blood (Moore et al., 2000). Furthermore, NPs have been shown to interact with the immune system and can induce adverse effects, such as hypersensitivity reactions and inflammation (Pondman et al., 2023). Efforts have thus been made to shield NPs from interactions with the immune system. We have shown that binding of soluble complement factors to CNTs enhances their uptake in macrophages and minimises inflammatory cytokine production (Kouser et al., 2018; Pondman et al., 2015). However, difficulties to effectively replicate and fully capture the immune response in in vitro systems necessitates in vivo experiments using animal models (Drasler et al., 2017). Nevertheless, animal models have raised ethical concerns, are costly and not suitable for high throughput screening. In line with the 3Rs - Reduction, Refinement, and Replacement - the adoption of non-animal and invertebrate animal alternatives, where scientifically viable, is now becoming increasingly relevant. Invertebrate models, such as the fruit fly Drosophila melanogaster or the nematode Caenorhabditis elegans, are cost efficient and do not fall under restrictive ethical regulations, thus warranting their widespread use. To this end, the larvae of the invertebrate Galleria Mellonella (GM) (the greater wax moth), have been developed as an alternative and effective tool to study in vivo toxicity. Due to their rapid growth rate, large size, and short life cycle (Cutuli et al., 2019), they can be used to study pathogen virulence, host-pathogen interactions and measure efficacy of various antimicrobial, fungicidal and other agents, including NPs, at an early stage during drug development (Jander et al., 2000; Maslova et al., 2024; Maslova et al., 2020; Opris et al., 2025). Importantly, GM larvae can be maintained at 37 °C (Peleg Anton et al., 2009), thus allowing for nanotoxicity studies to take place at the physiological human body temperature, which is especially important when studying the effect of NPs on infections. Furthermore, the larvae possess a complex innate immune system, closely resembling that of mammals (Hernandez et al., 2019; Gallorini et al., 2024). The cells of the larval haemolymph (haemocytes) are classified into six different types (prohemocytes, plasmatocytes, granular cells, coagulocytes, spherulocytes and oenocytoids) (Tsai et al., 2016), while the humoral immune response consists of opsonisation by complement-like proteins, melanisation and synthesis of lysozyme and small antimicrobial peptides (Trevijano-Contador and Zaragoza, 2018). As such, the utility of this model for studying both immunological and microbiological aspects of host-pathogen interactions has been widely recognised (Menard et al., 2021). Thus, GM larvae can be a useful tool to evaluate the direct immunotoxicity and immunosuppressive activity of engineered NPs and their potential for complementing current immunosuppressive drugs.

Furthermore, GM larvae are a good model for infection with human pathogenic bacteria, such as Pseudomonas aeruginosa and Acinetobacter baumannii, both of which are leading causes of nosocomial infections in cystic fibrosis patients, burn victims, and other immunocompromised individuals (Jander et al., 2000; Maslova et al., 2020).

In this study, we determined the in vivo biological safety of a range of nanoparticles, both with and without treatment with the immunosuppressive drug cyclophosphamide. We utilised flow cytometry as an accurate and reliable method for the quantification of total circulating haemocytes and employed cryo-sectioning and histochemistry to determine the systemic effects of injected NPs in GM larvae. Larval survival was studied with Kaplan Meier survival curves and reactive oxygen generation in vivo was assessed by measuring the concentration of 4-hydroxynonenal (4-HNE), a secondary product of lipoperoxidation (Poli and Schaur, 2000). Challenging the larvae with sub-lethal doses of highly virulent strains of P. aeruginosa and A. baumannii enabled us to study early immune responses to invasion of these two clinically important pathogens after immune suppression with cyclophosphamide and treatment with GNPs.

Our findings highlight the value of GM larvae in the study of the innate immune response to injected NPs and its usefulness for high-throughput investigation of NP immunotoxicity. Additionally, we discuss potential clinical applications of the tested nanoparticles to directly modulate the innate immune response, with the potential to complement current immunosuppressive therapies.

With the rapid expansion of nanotechnology, efficient and accurate toxicity assays are essential for the use of NPs in the clinic. Although rodent models provide valuable insights, they are costly and limited by ethical concerns. In this study, GM larvae were used as a non-rodent, in vivo alternative to assess nanotoxicity under immune suppression. We demonstrated here the usefulness of the GM model to study NP toxicity and responses to nosocomial infection under immunosuppression. CTX, one of the most used agents in cancer chemotherapy and for the treatment of immunological disorders, robustly reduces THC in GM larvae within 24 h with only mild overall toxicity (Figure 2B).

Due to the similarities between the innate immune cellular components in GM larvae and mammals, larval THC was used as an indicator of NP immunotoxicity/immunosuppression. Significant levels of immunotoxicity were observed for both SPION variants (Figure 4A); however, it was more significant for An-SPIONs, both in control and CTX-treated larvae and translated into significant larval death 72 h post-inoculation (Figure 3A). This difference in toxicity is most likely due to the lack of biocompatible coating in An-SPIONs, as various surface modifications of magnetite NPs are known to increase their biocompatibility (Singh et al., 2010). Furthermore, the immunotoxic effect could be due to increased levels of cellular oxidative stress, alterations of the cytoskeleton and genotoxicity in response to NP uptake (Singh et al., 2010). Fluorescent microscopy imaging shows that GM larval haemocytes avidly internalise NPs (Figure 6A) with internalised SPIONs localised throughout the cytoplasm and in close proximity to the cell nucleus. Interestingly, data reviewed by Shah and Dobrovolskaia indicates that iron oxide nanoparticle accumulation in rat alveolar macrophages and RAW264.7 macrophage like cells induced widespread reprogramming of genes involved in oxidative stress and inflammation (Shah and Dobrovolskaia, 2018). However, Stroh et al. showed that oxidative stress is a transient event, linked to free iron concentration in the cells, which does not lead to long-term cytotoxicity (Stroh et al., 2004).

To investigate and measure cellular ROS production in the larvae, we conducted an in vivo assay to quantify cellular ROS in response to NP exposure and/or cellular uptake (Zhang et al., 2011). A 4-HNE-lipid peroxidation ELISA was used for the first time in conjunction with GM larval biological samples. This assay showed that at the used concentrations, SPIONs only lead to limited ROS generation 24 h post-treatment (Figure 5). Our results are thus in line with the in vitro data obtained by Stroh et al. (2004). However, our results clearly demonstrate a negative effect on larval haemocytes that is exacerbated by CTX, which could be due to increased SPION uptake in CTX-treated larvae. This suggests that cellular responses, including endoplasmic reticulum stress, mitochondrial damage, and autophagy are affecting haemocyte survival (Park et al., 2014). An increase in ROS production similar to treatment with SPIONs was observed in larvae injected with 100 nm GNPs (Figure 5). The 100 nm GNPs also induced a comparable effect to the An-SPIONs on THC and larval survival, reflecting the lack of biocompatible coating and the overall negative charge of these NPs (Figures 3C, 4C). In general, GNPs showed a size-dependent effect on THC, with both 60 nm and 100 nm GNPs inducing a strong THC depletion. This immunotoxic effect has also been observed in vitro with human peripheral blood lymphocytes and murine splenic lymphocytes, which were significantly inhibited by GNPs at a concentration of 200 mg/kg (Devanabanda et al., 2016).

While significant decreases in larval THC were observed across all tested GNPs, only the 20 and 60 nm GNPs induced significant larval death, whereas the 100 nm variant showed only mild systemic toxicity (Figure 3C). This is in line with the study by Vecchio et al. that assessed toxicity of citrate capped GNPs of different sizes (5, 15, 40 and 80 nm) in D. melanogaster (Vecchio et al., 2012). While the flies were ingesting the GNPs with their food, the results clearly demonstrate a size dependent toxicity with a strong reduction of Drosophila lifespan, while a concentration dependent toxicity was observed for the presence of DNA fragmentation and changes in expression of genes linked to stress response, DNA damage and apoptosis (Vecchio et al., 2012). Furthermore, Isoda et al. assessed toxicity in mice treated with 10, 50 and 100 nm GNPs and their results showed that 10 and 50 nm GNPs cause kidney damage, while the 100 nm GNPs did not cause any systemic effect (Isoda et al., 2020). In our experiments, all GNP variants showed enhanced systemic toxicity after immunosuppression with CTX (Figure 3C). However, the effect varied and was highest for the 100 nm GNP. This could be due to the fact that these GNPs are more efficiently removed in immune competent larvae whereas in immunosuppressed larvae, the lack of immune cell uptake and systemic removal could induce greater levels of systemic toxicity.

The highest level of 4-HNE was observed in the CMC-CNT treated samples, with 4-HNE concentrations comparable to the positive control. Conversely, in the unmodified CNTs 4-HNE increases were negligeable. This data is in line with a study by Hsieh and Jafvert demonstrating that coated single-wall CNTs greatly enhance the rate of superoxide formation, which is not simply based on charge, but most likely due to the increased ability of the coated CNTs to scavenge ROS and in turn to generate greater quantities of more stable ROS (Hsieh and Jafvert, 2015). Furthermore, CMC-CNTs were more avidly taken up by the larval haemocytes than the other variants (Figure 5B), increasing the likelihood of cellular ROS production in an effort to clear them from the cells (Qi et al., 2024). Additionally, Meunier et al. have shown that phagocytosis of unmodified double-walled CNTs by human monocytes is necessary to induce IL-1β secretion, which stimulates a strong inflammatory response (Meunier et al., 2012). This inflammatory response is higher in CNTs with reduced solubility (Dumortier et al., 2006) and most likely explains the lack of THC reduction that we observed in larvae treated with the CNT variants (Figure 4B). Interestingly, CTX pre-treatment did not suppress THC (Figure 4B), indicating that there was a strong activation of the larval innate immune system by all tested CNT variants leading to a systemic inflammatory response. Despite a lack of effect on THC in CNT-treated larvae, CTX pre-treatment reduced larval death, an effect that was more pronounced in larvae treated with the unmodified CNTs (from 54% to 15%), whereas for both CMC-CNT and ox-CNTs, this protective effect was more limited. These results highlight the importance of CNT dispersion not only on larval death (27% CMC-CNTs and 20% ox-CNTs compared to 54.7% for the unmodified CNTs), but also on the innate immune response (Figure 4B). Furthermore, an increased clearance of CNTs by haemocytes could contribute to the increased larval survival in CTX-treated larvae. In mammals, CTX was shown to enhance the population of monocytes undergoing active endocytosis, thus leading to enhanced clearance by liver and spleen of IgG-sensitized erythrocytes (Ziccheddu et al., 2013; Palermo et al., 1991; Giordano and Isturiz, 1983).

In addition to innate immune-mediated inflammatory responses, the observed systemic toxicity (Figure 3B) could also be due to the in vivo accumulation of CNTs, contributing to larval death. Histochemistry is commonly carried out on paraffin embedded whole body sections of GM larvae to investigate host-pathogen interactions (Perdoni et al., 2014), progression of pathogen infection in vivo (Djainal et al., 2020) and pathogen invasiveness with cryo-imaging (Sheehan et al., 2018). In this study, we utilised for the first time histological analysis on GM larval cryo-sections stained with H&E, which show CNT aggregates localised in close association with the digestive tract, striate muscle fascicles and the Malpighian tubules (Figure 7).

To measure the effect of immunosuppression caused by NPs in the context of bacterial infection, we used 60 nm GNPs. These NPs, which are close in size to those used in commercial lateral flow assays, showed robust immunosuppression (Kim et al., 2016). GM larvae are highly susceptible to P. aeruginosa PA14 infections whereas higher CFU are tolerated for A. baumannii AB5075 (Tsai et al., 2016). When the GM larval innate immune system was challenged with these nosocomial infection causing bacteria, a significant increase in larval THC was observed (Figure 8A). For P. aeruginosa, this increase was not observed after CTX- or 60 nm GNP-induced immunosuppression whereas A. baumannii induced THC increases even after immune suppression (Figure 8A). This result highlights similarities to human infection and findings in rodent models, where A. baumannii triggers a strong increase in cytokine release (Wang H. et al., 2024). P. aeruginosa infections on the other hand show a strong induction of neutrophil recruitment to the site of infection with extensive formation of neutrophil extracellular traps (NETs) (Kamoshida et al., 2015). Interestingly, we saw an increase in extracellular material resembling the NETs observed in mammalian host surrounding haemocytes isolated from PA14 treated larvae especially after pre-treatment with 60 nm GNPs (Figure 8C). At the tested concentrations, reduction in larval survival after immunosuppression was only observed in P. aeruginosa infected larvae, which, however, did not reach statistical significance (Figure 8B). It is important to note that we used sublethal bacterial doses for both bacteria species, which allowed us to investigate the immediate early effects of infection (Ambrosi et al., 2020), including THC count.

NP-induced immunosuppression through direct immune interaction is often overlooked. Metal oxide NPs, like SPIONs, show immunosuppressive and anti-inflammatory properties - e.g., IONPs reducing humoral immune responses (Shah and Dobrovolskaia, 2018). Noble metal NPs, such as GNPs, interact with both innate and adaptive immune components, but their immunosuppressive mechanisms remain underexplored (Ngobili and Daniele, 2016; Wang Y.-L. et al., 2024). Depending on the context, NP-induced immunosuppression can be beneficial (e.g., enhancing drug efficacy, treating autoimmune diseases, improving transplant tolerance) or harmful (e.g., reducing host defences against infections or cancer) (Pondman et al., 2023). It is thus important to evaluate the immunosuppressive effect of novel NP formulations. Our results confirm the importance of NP coating to reduce toxicity. However, they also demonstrate that not all coatings have the anticipated effect as further physical features, such as dispersibility and tissue uptake, are important factors to take into consideration when developing novel nanotherapeutics. While our study sheds light on the immediate early effects of NP immunotoxicity, more work needs to be done to elucidate the molecular mechanisms leading to the observed immunosuppression elicited by GNPs and SPIONs and the long-term effects of immunosuppression in the larvae. Future work could include a quantification of the NP retention in the larvae including development defects.

In summary, our study shows that GM larvae provide a valuable intermediate model to assay NP in vivo toxicity, enabling more accurate predictions of their behaviour in whole organisms and their immunomodulating potential.