Disturbance of cellular homeostasis as a molecular risk evaluation of human endothelial cells exposed to nanoparticles

Even though application of nanoparticles in medicine seems to provide unique solutions for drug delivery and diagnosis diseases, understanding interactions between nanoscale materials and biological systems is imperative. Therefore, this study determined the effect of different types of nanoparticles (NPs) on human endothelial cells and examined the types of toxicity responses they can induce. Four different types of NPs were tested (PLA/MMT/TRASTUZUMAB, PLA/EDTMP, PLGA/MDP, and Pluronic F127 MICELLES), representing three putative areas of application: anticancer therapy, scintigraphy, and cosmetology. The experiments were performed on immortalized human umbilical vein endothelial cells (HUVEC-STs). Light contrast phase microscopy as well as cell viability assays showed that only Pluronic F127 MICELLES decreased the number of HUVEC-STs in contrast to PLA/MMT/TRASTUZUMAB, PLA/EDTMP, and PLGA/MDP NPs, which altered cell morphology, but not their confluency. The tested NPs induced not only DNA strand-breaks and alkali-labile sites, but also internucleosomal DNA fragmentation, visualized as a DNA ladder pattern typical of apoptosis. Moreover, generation of free radicals and subsequent mitochondrial membrane potential collapse showed the significance of free radical production during interactions between NPs and endothelial cells. High concentrations of NPs had different degrees of toxicity in human endothelial cells and affected cell proliferation, redox homeostasis, and triggered mitochondrial dysfunction.

www.nature.com/scientificreports/ of macromolecules to the underlying interstitium and cells. Many substances (drugs, antibodies, and hormones) interact with the vessel wall and modulate both the metabolism and function of the endothelial barrier 5 . For this reason, the endothelium is extremely important in nanotoxicological studies, mainly due to the possible interactions between ECs and nanosubstances. In addition, the use of NPs can modulate endothelial leakiness and have been increasingly employed in recent studies to overcome vascular barriers, improve drug delivery, and enhance cellular uptake of nanodrugs 6 . The influence of nanosubstances on the endothelial function is a key factor for better understanding of the potential in vivo NP effects 7 . It is a huge challenge to find the best cellular model for the evaluation of nanomaterial effect on the human blood vessel lumen. There are many variables that should be considered when human umbilical vein endothelial cells (HUVECs) are used in an in vitro model to assess NP toxicity. Consequently, NP physicochemical properties (size, shape, solubility, and surface charge) define the diversity of EC viability and proliferation as molecular responses to NP exposure 8 . For example, the polyvalent surface of NPs may induce cross-linking of cellular receptors, initiate signaling processes, induce structural alterations at the cell surface, and interfere with normal endothelial function 9 . Moreover, the physicochemical properties of nanosubstances can cause undesirable toxic effects associated with long-term application of NPs.
HUVECs were developed via fusion of primary cell cultures with permanent cell lines. Gene and protein profile analysis has indicated that immortalized cells often retain most of the important endothelial markers 10,11 . However, there are other cellular models that are crucial for understanding the EC reaction to nanomaterials application. For instance, human aortic endothelial cells (HAECs) are a good choice for in vitro cell culture if the research is targeted on restenosis after vascular intervention triggered by NPs 12 . Some data have described a more complicated system with microfluidic channels lined with hCMEC/D3 cells, constructed to examine the effect of NP shape on particle adhesion 13 . On the other hand, computational fluid dynamics (CFD) is a well-known in silico approach to find the connection between NP properties and their ability to modulate immune system interactions and blood clearance profile 14 .
The present study aimed to verify the hypothesis that NPs can reveal cytotoxic properties and modulate endothelial cell homeostasis even though they do not always reduce cell viability. Therefore, this basic and cognitive study investigated different types of molecular responses induced by endothelial NP interactions. Four types of NPs were used (Table 1, Fig. 1): the monoclonal antibody, TRASTUZUMAB (PLA/ MMT/TRA) modified with polylactic acetate (PLA) and montmorillonite (MMT), which specifically targets HER2 receptor extracellular domain 15 ; ethylene diamine N,N,N' ,N' tetramethylene phosphonic acid NPs coated with PLA (PLA/EDTMP) 16 , which are used in breast and bone cancer treatment; silicon NPs PLGA/MDP modified with methylene diphosphate (MDP) and coupled to poly lactic-co-glycolic acid (PLGA) 17 , which can be used in scintigraphy; and Pluronic F127 MICELLES (Pluronic F127 Ms), that can be administered dermally 18 , intranasally, or intravenously 19 . HUVEC-STs were used as a cellular model. Thus, multi-dimensional analyses were performed to evaluate interactions between NPs and ECs. Phase contrast microscopy demonstrated that NPs altered the morphology of HUVEC-STs. NPs also induced DNA damage that did not always correspond to inhibition of cellular proliferation and EC viability. The cytotoxic effect of the investigated nanosubstances was studied for up to 72 h with particular emphasis on the induction of programmed cell death, including reactive oxygen species (ROS) production. Mitochondrial stress monitoring was used to determine the importance of redox homeostasis disturbances in ECs triggered by NP interactions.

Results
Physiochemical characterization of NPs. The examined NPs had a hydrodynamic diameter ranging from 140 to 278 nm with a near neutral or low positive charge, a zeta potential of + 0.4 to + 7.9 mV (Table 1), and polydispersity index (PDI) varying from 0.19 to 0.30. The PDI values confirmed the medium dispersity of all NPs. NP entrapment efficacy varied from 70% (PLA/MDP) to 100% (Pluronic F127 Ms), supporting the efficient encapsulation of all drugs used during the NP preparation ( Table 1). The NP size was stable upon incubation in PBS when stored at room temperature for more than 24 h (data not shown).
Various sensitivities of human ECs to NP treatment. When particle cytotoxicity was compared by adding NPs at increasing concentrations to HUVEC-STs, a striking difference in NP cytotoxicity was observed Table 1. Physicochemical properties of investigated particles: size distribution, zeta-potential, PDI, and entrapment efficacy of PLA/MMT/TRA, PLA/EDTMP, PLGA/MDP, and Pluronic F127 Ms. a The size from each nanosystem has been confirmed by TEM or AFM analysis and the results were similar with the DLS analysis.

Nanoparticles
Application NP size a (nm) ζ potential (meV) PDI Entrapment efficacy (% of drug)  Fig. 2A,B). This is in agreement with our previous analyses of NP cytotoxicity in peripheral blood mononuclear cells 20 . The morphological changes of human ECs induced by 24-h exposure to 100 μg/mL of PLA/ EDTMP, PLGA/MDP, and PLA/MMT/TRA or 0.025 μg/mL of Pluronic F127 Ms were visualized in parallel. The microscopy images showed that Pluronic F127 Ms had the highest toxic effect (cell detachment), whereas PLA/ MMT/TRA, PLA/EDTMP, and PLGA/MDP at this time point mostly led to cell rounding and did not reduce cell numbers (Fig. 2C). Cell proliferation rate is a measurement of cell sensitivity to various types of external factors. Therefore, it was determined whether the examined NPs had a toxic effect on HUVEC-ST proliferation by counting the cells after trypan blue staining for up to 72 h. A reduction in cell number was observed after HUVEC-STs were treated with Pluronic F127 Ms (Fig. 2D). A severe reduction in cell proliferation when HUVEC-STs were incubated with PLA/MMT/TRA, PLA/EDTMP, and PLGA/MDP for up to 24 h was not observed. It should be noted that a significant reduction was observed, although with a slightly different extent and kinetics, when HUVEC-STs were incubated with a higher concentration of PLA/MMT/TRA, PLA/EDTMP, and PLGA/MDP (250 μg/mL) for up to 72 h (Fig. 2E).
Genotoxic effect of NPs. A reduction in proliferation rate (Fig. 2D,E) in cell cultures is often a hallmark of altered cellular homeostasis when DNA damage is induced 21 . Therefore, the study evaluated whether the investigated NPs triggered DNA single-and double-strand breaks as well as alkali-labile sites on the basis of degradable DNA percentage. When HUVEC-STs were treated with Pluronic F127 Ms for up to 24 h (Fig. 3B), an increase in DNA percentage in the comet tail was noted. Moreover, significant differences between the extent of DNA damage induced by PLA/EDTMP, PLGA/MDP, and PLA/MMT/TRA were detected 24, 48, and 72 h after exposure of ECs to 100 μg/mL of NPs. On the other hand, a nearly two-fold decrease in DNA percentage in the comet tails (Fig. 3C) was noted in cells treated with Pluronic F127 Ms if incubation continued for up to 72 h, suggesting that HUVEC-STs needed a longer time for DNA repair machinery to be activated as a consequence of cellular stress initiation by NPs. Because human ECs were able to cope with the accumulation of DNA damage induced by PLA/EDTMP, PLGA/MDP, and PLA/MMT/TRA, it was necessary to determine how early the genes involved in a response to one or more DNA lesions could be activated. Human cells contain highly conserved pathways to signal and repair DNA damage during constant exposure to exogenous and endogenous DNA-damaging agents (Fig. 3A). ATM and ATR are near the top of these signalling networks, which are often named the "sentries" at the gate of genome stability. Therefore, the present study evaluated whether the expression of these proteins can be induced in human ECs by treatment with the tested NPs. ATM and ATR kinase transcription was increased when the cells were treated with Pluronic F127 Ms for up to 24 h (Fig. 3D). Neither PLGA/MDP nor PLA/MMT/TRA triggered extensive transcription of ATR or ATM mRNA. However, when HUVEC-STs were treated with PLA/  Internucleosomal DNA fragmentation of ECs is part of DNA lesions caused by NPs. As described above, HUVEC-STs have been shown to be vulnerable to genotoxic effects of the investigated NPs. As the comet assay has been broadened to include the detection of DNA fragmentation in cells undergoing apoptosis 23 , it was first confirmed whether apoptosis-dependent DNA fragmentation was induced by the investigated NPs. Indeed, the population of TUNEL-positive cells (those that exhibited single-and double-stranded DNA fragments with possible label-free 3′ OH ends) increased in a time-dependent manner with all tested NPs (Fig. 4). The maximum increase in the percentage of apoptotic cells with DNA strand-breaks was detected 72 h after PLGA/MDP and Pluronic F127 Ms treatment. These data correlated well with oligosomal DNA fragmentation monitored during the DNA ladder assay. The characteristic DNA ladder-like pattern of the internucleosomal DNA cleavage (Fig. 4B, Supplementary Fig. S1) was detected when HUVEC-STs were incubated for up to 72 h with all tested NPs (100 μg/mL of PLA/EDTMP, PLGA/MDP, and PLA/MMT/TRA or 0.025 μg/mL of Pluronic F127 Ms). Cellular stress response proteins are activated when DNA damage is initiated by various stressful conditions. One of these response proteins downstream of apoptosis-induced DNA damage is the phosphorylation of histone H2AX. PLA/EDTMP, PLGA/MDP, and Pluronic F127 Ms triggered histone H2AX phosphorylation after just 24 h of NP incubation (Fig. 4C). On the other hand, a statistically significant effect of all NPs was observed when www.nature.com/scientificreports/ HUVEC-STs were exposed to NPs for up to 48 h. Interestingly, and in support of this, NP treatment led to an increase in histone H2AX transcription (Fig. 4D), as the first, cellular response to DNA lesions.

Apoptosis induction in HUVEC-STs after NP treatment.
Intrigued by the possibility of DNA fragmentation caused by different cytotoxic effects of NPs, the present study aimed to identify the fraction of apoptotic and necrotic dead cells (Fig. 5A) using a mixture of two fluorochromes Hoechst 33258/PI. The four cell populations: live, early apoptotic, late apoptotic, and necrotic were distinguished during fluorescence microscopy observations (Fig. 5B). A time-dependent difference between the investigated NPs was not detected. It appeared that HUVEC-STs exhibited the same resistance pattern to NPs. Slight differences between the NPs were noted when incubation was prolonged for up to 72 h. These data are in agreement with cytotoxicity studies performed using the neutral red and Alamar Blue assays. Apoptosis induction was further detected via fluorogenic CellEvent caspase-3 substrate staining. The highest caspase-3 activity we detected after 24 and 48 h of treatment with all investigated NPs (Fig. 5C). In contrast to PLA/MMT/TRA, PLA/EDTMP, and PLGA/MDP, the maximal cleavage of caspase-3 substrate was observed when cells were incubated with Pluronic F127 Ms for up to 48 h. These data are in agreement with caspase-3 mRNA transcription measured after 24 h (Fig. 5D). It should be noted that the expression, as well as transcription, was the highest in PLGA/MDP-treated samples. As an increase in PARP functionality may be a consequence of alterations in caspase-3 activity, PARP mRNA level was measured after 24 h of incubation with NPs. All investigated NPs altered mRNA PARP transcription, which was directly related to the initiation of DNA repair pathways (Fig. 5E).

Mitochondrial stress is triggered in ECs via NP properties. ROS generation related to the oxidation
chain mitochondrial metabolism is one of the markers confirming apoptosis induction (Fig. 6A). Increased free radical production was observed in NP-treated cells as a consequence of oxidative stress generation. ROS production was considerably greater in cells incubated with PLA/EDTMP, PLGA/MDP, and Pluronic F127 Ms, reaching a maximal level after 48 h of incubation ( Fig. 6B). At this time point, a two-fold and 1.7-fold increase was noted in cells incubated with PLA/EDTMP and PLGA/MDP or Pluronic F127 Ms, respectively. On the other hand, the highest level of ROS production in HUVEC-STs treated with PLA/MMT/TRA was noted after It has been previously described that the role of mitochondria in programmed cell death induction is related to the changes in mitochondrial membrane potential (MMP, Δψ m ) 24 . Early MMP depletion was measured using JC-1-stained cells in a fluorescence microplate reader assay. FCCP was used as a positive control (Fig. 6D), which is an ionophore that disrupts ATP synthesis. The collapse of Δψ m in human ECs was observed after 24 and 48 h of treatment with PLA/MMT/TRA, PLA/EDTMP, and PLGA/MDP (Fig. 6C). Interestingly, an increase in JC-1 fluorescence was observed when incubation with these three NPs was prolonged for up to 72 h. At this time point, MMP increased by about 50% when HUVEC-STs were cultured with PLA/EDTMP and PLGA/MDP, which may suggest hyperpolarization of the inner mitochondrial membrane.
The Δψ m changes are often one of the factors that can confirm mitochondrial homeostasis disturbances. Indeed, during mitochondrial stress, many molecules are imported from the intermembrane space or matrix to the cytosol. Therefore, it was determined whether the mRNA level of Bax, Bcl-2, cytochrome c (Cyt c), and AIF changed as a consequence of transcription alteration of these genes. Neither Bcl-2 nor Bax transcription was changed after incubation with NPs for up to 24 h, apart from Pluronic F127 Ms, which triggered a 4.5-fold increase in Bcl-2 transcription (Fig. 6E). However, a significant increase was noted for Cyt c and AIF mRNA www.nature.com/scientificreports/ under the same treatment conditions, when all NPs induced an increase in gene transcription of at least two-fold, indicating that apoptosis was induced before cell rupture.

Discussion
As NPs can easily penetrate into the body, they can be transported to various organs or tissues and affect them in many different ways. There are many chemical modifications used during the synthesis of well-known NPs that increase their circulation time, limit aggregation, or decrease their association with untargeted serum and tissue proteins 25 . However, several reports have indicated that the strategies used to coat NPs may negatively influence the performance of NPs as drug carriers 26,27 . Human blood vessels are one of the first barriers that NPs interact with. Thus, it is necessary to understand the adverse effects of NPs on human ECs that cover the lumen of arteries, veins, capillaries, and venules.
In the present study, four different types of NPs were used, including PLA/EDTMP, PLGA/MDP, PLA/MMT/ TRA, and Pluronic F127 Ms, to study cellular responses and human EC sensitivity to various nanomaterials. It is worth mentioning that both nanomaterials and their components can have cytotoxic properties during the evaluation of NP toxicity. For example, the major side effects of 153 Sm-EDTMP NPs are a reduction of platelets and leucocytes in peripheral blood caused by β-particles from the radiopharmaceutical attached to the bone matrix 16 .
In the present study, Pluronic F127 Ms were the most toxic after 72 h of incubation, as determined using both resazurin oxidation and neutral red assays. However, the effect was greater during the assessment of neutral red Results represent the means ± SD of three independent experiments. *p < 0.05 for significant differences between NP-treated and untreated control cells. (D) Caspase-3 and PARP transcription levels (relative to HPRT1) in HUVEC-STs exposed to examined nanosubstances: PLA/EDTMP, PLGA/MDP, and PLA/MMT/TRA (100 μg/mL) or Pluronic F127 Ms (0.025 μg/mL) for 24 h. Asterisks refer to significant differences (*p < 0.05, **p < 0.01, and **p < 0.001; n = 3) in transcription levels in NP-treated cells compared to untreated cells. www.nature.com/scientificreports/ accumulation inside the lysosomes. This may suggest that Pluronic F127 Ms cytotoxicity is related to a decrease in lysosome functionality or lysosome damage. Moreover, the low viability of HUVEC-STs after NM treatment also showed that it is essential to evaluate the cytotoxicity index of various nanomaterials. Interestingly, it was noted that neither the Alamar Blue test nor the neutral red assay showed a reduction in HUVEC-ST viability after treatment with PLA/EDTMP, PLGA/MDP, and PLA/MMT/TRA. As demonstrated by Helal-Neto et al. 28 , HUVEC exposition to non-loaded NP does not interfere with their viability. Controversially, in experiments carried out with bacterial strains, Guedes et al. 29 have indicated high MDP toxicity. Our data have demonstrated how crucial it is to carry out different assays when studying the cytotoxic effect of NPs on normal cells 30 . Different HUVEC-ST sensitivity to the four different NP treatments may be associated with their various physicochemical properties. Thus, it was suggested that NPs with a positive charge bind to the negatively charged cell surface 31 . Consequently, these positively charged NPs can be imported into the cell more efficiently than negatively charged particles 32 . This hypothesis was confirmed by Huhn et al. 33 , who showed higher cytotoxicity of positively charged gold NPs in HUVECs compared to negatively charged gold NPs. The critical role of NP surface charge has also been proposed with a reference to NP absorption by serum proteins 34 . This phenomenon, known as the "corona effect", significantly increases the interactions between NPs and HUVECs and reduces toxicity 35 . DLS measurement results indicated that the examined NPs had a slightly positive charge and were marginally greater than 200 nm in size, with the exception of Pluronic F127 Ms, which had a diameter of 140 nm.
The size of NPs is an important factor in achieving an efficient desired effect, such as accumulation in tumor tissue or toxicity. Different toxicity profiles of PLA/MMT/TRA, PLA/EDTMP, PLGA/MDP, and Pluronic F127 . Fluorescence ratio of JC-1 dimers to JC-1 monomers in control was assumed to be 100%. Results are presented as means ± SD of four experiments. *p < 0.05, **p < 0.01 for significant differences between NP-treated and untreated control cells (taken as 100%); (E) Bax, Bcl2, Cyt c, and AIF gene transcript expression (relative to HPRT1 housekeeping gene) in HUVEC-STs exposed to examined nanosubstances: PLA/EDTMP, PLGA/MDP, and PLA/MMT/TRA (100 μg/mL) or Pluronic F127 Ms (0.025 μg/mL). Asterisks refer to level of significant (**p < 0.01, ***p < 0.001; n = 3) difference in expression in cells treated with investigated NPs compared to untreated cells. www.nature.com/scientificreports/ Ms were observed not only in the viability assays, but also when HUVEC-ST proliferation rate was monitored following NP treatment for up to 72 h. Interestingly, the data obtained from the comet assay, one of the most popular assays used in toxicology studies to detect DNA damage 21 , revealed that all tested NPs possessed genotoxic properties, even though only Pluronic F127 Ms decreased HUVEC-ST viability. This high genotoxicity was confirmed by Mattos et al. 36 , who reported that 99mTc-MDP induced DNA strand breaks in rat blood cells.
The fatal consequences of DNA damage are often referred to cellular uptake of NPs. There are several endocytic mechanisms(e.g. clathrin mediated, clathrin-independent uptake) known with a membrane invagination diameter of less than 150 nm, whereas macropinocytosis is the most likely mechanism for uptake of NPs above this size 37 . Nevertheless, caveolae which are commonly implicated in trans-endothelial transport have diameters of 80 nm or less 38 , may be excluded here, as the possible mechanism of endocytosis, because they are smaller than all of four tested NPs. Interestingly, the most likely genotoxic properties of Pluronic F127 Ms can be partially related to the size of NPs and their reactivity. As particle size decreases, the particle unit of mass and overall surface area increases, because surface atoms have a tendency to possess high energy bonds. This larger surface area enhances catalytic activity, and consequently may lead to the oxidative stress and indirect induction of DNA lesions 39 .
On the other hand, cytotoxicity measurement and proliferation assay results were in contrast to the genotoxic profile of the investigated NPs. In support of these data, Calarco et al. 40 have noted that micelles based on polyethyleneimine (PEI)-PLGA polymer induced DNA damage in HUVECs without a significant effect on cell viability, which showed that DNA damage can be a sensitive marker for NP-triggered toxicity.
It has been previously shown that several types of NPs, such as micelles based on polymers, silica NPs, or carbon nanotubes, can trigger DNA damage in HUVECs [41][42][43] . In agreement with these results, NPs in the present study induced DNA lesions after 24 and 48 h of treatment. When NP incubation was continued for up to 72 h, a decrease in DNA percentage was noted in the comet tail, suggesting that DNA repair had started.
There are many signalling pathways that are central to the maintenance of genome integrity. The role of ATM and ATR kinases has been well described in DNA repair, cell cycle regulation, and apoptosis 22 . There was a relationship between ATM and ATR gene mRNA level and NP toxicity in the present study. An increase in transcription of these two kinases suggested that ATM and ATR activity was rising due to the need to maintain genomic stability during exposure to DNA-damaging NPs. ATM and ATR kinases normally phosphorylate an overlapping set of DNA repair or checkpoint targets (p53, CHK2, NBS1, and BRCA1). By doing so, they ensure that the cells accurately repair DNA damage before DNA replication or cell division occurs 44 .
The results obtained in the present study clearly demonstrated that DNA lesions triggered by NPs can be interpreted as being due to the induction of DNA strand breaks and/or formation of alkali-labile sites, which can be transformed into strand breaks 23 . The breaks or alkali-labile sites may be the result of programmed cell death induced by the tested nanomaterials. Confirmation that the investigated NPs induced single or double-strand breaks was provided by the TUNEL assay and histone H2AX phosphorylation experiments. Many external factors such nanosubstances may lead to increased H2AX phosphorylation due to death-associated DNA fragmentation, especially long periods after drug treatment or in the case of agents that induce rapid apoptosis 45 . The characteristic oligonucleosomal fragmentation observed in the DNA ladder and the fraction of TUNELpositive cells revealed that single-strand breaks were the highest during incubation. Interestingly, there were no evident differences between proapoptotic NP properties when considering Hoechst/PI double staining. The greatest population of necrotic cells that appeared after the NM treatment confirmed that these particles have very toxic properties. However, after analysing caspase-3 activity and its transcription after 24 h of incubation, the highest increase of these two parameters was observed in PLGA/MDP-treated samples. When NP incubation was prolonged to 48 h, the maximal caspase-3 activity was triggered by Pluronic F127 Ms, which was in agreement with PARP gene transcription measured when cells were treated with NPs for 24 h. Keeping in mind the cascade of molecular events where caspase-3 participates in PARP fragmentation, it can be concluded that Pluronic F127 Ms need more time to induce apoptosis in HUVEC-ST culture. Intriguingly, promising results for synergistic therapy between proteasome inhibitor PS-341 and 153 Sm-EDTMP were revealed by Goel et al. Apoptosis induction (pro-caspase-3 and PARP cleavage) was observed in highly resistant myeloma cell lines after 24 h of combined treatment, even though neoplastic cell viability was 68% 46 .
There are many molecules like free radicals that act as second messengers and contribute to apoptosis induction. ROS production is also a hallmark of oxidative stress and an important mechanism of NP cytotoxicity. Numerous studies have established an important role for ROS in cell growth and division, e.g., low doses of ROS are necessary for cell proliferation, whereas high concentrations are inhibitory and apoptotic. In ECs, ROS function as second messengers to activate multiple intracellular proteins and enzymes, including the epidermal growth factor receptor, c-Src, p38 mitogen-activated protein kinase, Ras, and Akt/protein kinase B [47][48][49] . Direct induction of ROS in HUVEC-ST culture might lead to oxidative damage of DNA, proteins, or membranes. However, it must be noted that in addition to exogenous agents, endogenous cell components may also directly oxidize H 2 DCF-DA without the involvement of ROS as described for Cyt c 50,51 . Interestingly, it has been shown that the presence of antioxidant NAC can attenuate the adverse effects of Ag NPs and quantum dots (QDs) in HUVECs 52,53 . This finding was in contrast to the present results that showed the protective effect of NAC, which can be an extracellular source of amino acids essential for the synthesis of cytoplasmic scavengers. Perturbations of redox homeostasis are often related to mitochondrial dysfunction. As autonomous organelles, mitochondria are involved in ATP production. Damage to the mitochondrial oxidation chain in the endothelium results in an increase in oxidative stress parameters and has been implicated in cardiovascular disease 54 .
It has been previously shown that silica NPs and CdTe QDs provoke MMP collapse and release of mitochondrial Cyt c in HUVECs 41,52,55 . In the present study, the increase in Bcl2, AIF, and Cyt c gene transcription confirmed the hypothesis that long-term NP treatment induces mitochondria-dependent apoptosis, which is an intrinsic apoptosis pathway. www.nature.com/scientificreports/ MMP disruption is characterized by the opening of permeability transition pores (PT) in the inner mitochondrial membrane. In the present study, an early decrease in MMPs 24 h after PLA/MMT/TRA and PLA/ EDTMP treatment of HUVEC-ST cultures was observed. It seems that apoptosis induced by these NPs was due to disturbance in the generation of ROS necessary for cell proliferation. Interestingly, MMP hyperpolarisation, observed after 72 h of incubation with PLA/EDTMP, PLGA/MDP, or Pluronic F127 Ms, usually occurs before activation of caspases and phosphatidylserine externalization 56 , which may suggest secondary induction of apoptosis by these NPs 57 . Pluronic F127 Ms were 500-fold more toxic to HUVEC-STs than the remaining tested NPs PLA/MMT/TRA, PLA/EDTMP, and PLGA/MDP. Even though these three tested NPs did not decrease cell viability, they displayed a genotoxic effect and induced DNA damage. The formation of single and double-strand breaks was related to NP-triggered apoptosis induction. ROS production and perturbations of cellular redox homeostasis led to mitochondrial stress and contributed to the intrinsic apoptosis pathway.
Even though interactions of nanoparticles with e.g. cancer cells have been extensively studied, there is still not sufficient knowledge available about how the NPs behave in the body and how fast they may escape from blood and reach the target tissue. Firstly, it is essential to discover it could be interesting to study what is the effect of blood pressure, hemodynamic flow in blood vessels and velocity on interactions between the NPs and the vessel wall 58 . Besides, human blood is composed of plasma (containing salts, lipids, proteins, vitamins, hormones, and water) and importantly, a variety of immune cells (e.g., monocytes, neutrophils, B cells, dendritic cells (DCs), T cells, and (NK) cells 59 . For the future, understanding such important effect of blood pressure, hemodynamic flow in blood vessels as well as the velocity on interactions between the NPs and human blood vessels is our foundational goal.
In conclusion, the present study demonstrated that NPs potentially used in biomedicine displayed different levels of cytotoxicity in HUVEC-STs. The investigated NPs induced DNA damage, mainly oligonucleosomal DNA fragmentation, characteristic of programmed cell death. HUVEC-STs showed morphological markers of apoptosis following PLA/EDTMP, PLGA/MDP, and PLA//MMT/TRA treatment. Incubation with Pluronic F127 Ms caused typical cell necrosis and marked disruption of the cellular membrane. These NPs were able to generate ROS, which mediated a decrease in MMP and triggered the expression of genes involved in the intrinsic apoptosis pathway. These results demonstrated that the mechanism of action of the investigated NPs was highly type-dependent and their effect should be examined in other types of normal cells. Finally, the data underscored the necessity of a broad range of in vitro experiments in order to provide an in-depth prediction of NP cellular toxicity and their potential future directions.
Cell lines. HUVEC-STs (immortalized by transfection with both SV40 large/small T antigens and catalytic human telomerase subunit) was a kind gift from Dr. Claudine Kieda (University of Orleans, France) via a personal mediation by Prof. G. Bartosz (Department of Molecular Biophysics, University of Lodz, Poland). This cell line was established and characterized at the University of Rome Tor Vergata as described by Tentori et al. 60 . Cells were cultured in Opti-MEM medium supplemented with 3.5% fetal bovine serum, penicillin (100 U/mL), and streptomycin (100 mg/mL). Cells were grown at 37 °C in standard conditions (100% humidity, 37 °C, 95% normal air, and an atmosphere with 5% CO 2 ). The cells were periodically tested for mycoplasma contamination and observed daily by microscopy for growth control. Logarithmic cell growth phase was maintained in all experiments. www.nature.com/scientificreports/ PLGA/MDP NP production. Double emulsification technique was used for the production of PLGA/MDP NPs. A total of 1 mL of PVA (1% w/v; Sigma-Aldrich), 15 mg of MDP (ABX Advanced Biochemical Compounds, Germany), and 2 mL of PLGA (Sigma-Aldrich) solution (120 mg dissolved in 2 mL of dichloromethane) were sonicated (UP100H, Hielscher, Teltow, Germany) for 2 min at 55 W. Then, 50 mL of PVA (0.8% w/v; Sigma-Aldrich) were added and sonicated again for 5 min at 55 W resulting in a "water/oil/water" (W/O/W) solution.
Finally the residual organic solvent was removed from the NPs by vacuum evaporation (Rotavapor R114, Buchi, Postfach, Switzerland) at 30 °C for 1 h and recovered by ultracentrifugation (15,000 rpm) at 25 °C for 20 min (Centrifugal Beckman Coulter J 25, Brea, California, USA). The NP properties were analysed just after synthesis by AMF microscopy 63 .
PLA/EDTMP NP production. For PLA/EDTMP NP production a double emulsion technique was used. Briefly, PLA (2.5% w/v; Sigma-Aldrich) was dissolved in 3 mL of methylene chloride (Sigma-Aldrich). A total of 200 µL of PVA (0.1% w/v; Sigma Aldrich) and EDTMP (4% w/v; ABX Advanced Biochemical Compounds, Germany) were added to this solution, followed by sonication (GEX 600 Sonicator Ultrasonic Processor) for 1 min at 0 °C and 55 W. Then, 4 mL of PVA (0.7% w/v) were added and sonicated again for 2 min (55 W) at 0 °C. The residual organic solvent was evaporated (Rotavapor R114, Buchi, Postfach, Switzerland) under reduced pressure for 20 min at 37 °C. Finally the NPs were recovered by ultra-centrifugation (20,000 rpm) at 25 °C for 10 min (Centrifugal Beckman Coulter J 25, Brea, CA, USA) and AMF images were taken if NPs were ready for experiments referred to as human endothelium cells 64 .
Pluronic F127 Ms production. The Pluronic F127 Ms was prepared using dispersed Pluronics F127 (Sigma-Aldrich) at a concentration of 12% (w/w) in water. Briefly, 100 mL of water were added to 12 mg of Pluronics F127 and gently stirred using a magnetic bar (Magnetic Stirrer, IKA, C-MAG HS-7) for 3 min and then processed for 3 min using an ultrasonic processor (UP100H, Hielscher, Power: 60%, Cycle: 1) in an ice bath at 10 °C. After the synthesis, transmission electron microscopy (TEM) analysis of the obtained Pluronic F127 Ms was carried out 65 .

Characterization of investigated NPs.
To demonstrate the breadth and variety of nanosubstance effects in human ECs, four different NPs were used in cancer therapy, theranostics, and cosmetology (Table 1, Fig. 1). NP zeta potential (ζ), PDI, and size (hydrodynamic diameter) were measured in 0.01 M phosphate buffer at pH 7.4 (PBS) and 25 °C using dynamic light scattering (DLS) and a Zetasizer Nano ZS (Malvern Instruments, UK). The NP suspensions were diluted to 1 mg/mL with PBS before analysis and the measurements were performed in triplicate maintaining the device in automatic mode. The purification steps were performed before the in vitro experiments. Sterilizing filtration was used for NPs sized < 220 nm, whereas for particles with greater sizes, NPs were lyophilized and then resuspended in sterile water. The final step referred to the analyzed NP description of entrapment efficacy (% of encapsulated drug) calculation described previously 66-68 . Cytotoxicity assays. Cytotoxicity was estimated by performing the neutral red assay (using a dye that passes through intact plasma membranes of viable cells and is accumulated in lysosomes), Alamar Blue assay (by measuring resazurin oxidation), and observing resulting cell morphology changes under a microscope.
Neutral red assay. Cells in 96-well plates were treated with PLAMMT/TRA, PLA/EDTMP, PLGA/MDP (concentration range 0-500 μg/mL), or Pluronic F127 Ms (concentration range 0-1 μg/mL) for 72 h. Thereafter, the medium was removed and replaced with a neutral red solution (0.5 mg/mL). The incubation was continued for a further 3 h under cell growth conditions. At the end of incubation, neural red imported into fully functionalized live cell lysosomes was dissolved in the extractant solution (51% H 2 O, 48% ethanol, and 1% acetic acid). The absorbance was measured at 550 nm (analytical wavelength) and 620 nm (reference wavelength) with a microplate reader (Awareness Technology Inc., USA).
Alamar blue assay. Cells (2 × 10 3 ) in 100 μL of culture medium per well were seeded into black 96-well microtiter plates with flat-bottomed wells 24 h before the experiment and then exposed in triplicate to different concentrations of PLA/MMT/TRA, PLA/EDTMP, PLGA/MDP, or Pluronic F127 Ms for 72 h (5% CO 2 , 37 °C, 100% humidity). Cell medium was aspirated and replaced with 100 μL of resazurin solution at the end of incubation. Incubation was continued for a further 3 h and then fluorescence was measured using a microplate reader (Fluoroskan Ascent FL, Sweden) with excitation and emission wavelengths of 530 nm and 570 nm, respectively.
Cell morphology. Cell morphology alterations were analysed after 24-h exposure to NPs. Images were acquired using an inverted microscope (Olympus IX70, Japan) equipped with a 20 × objective and a Digital Sight camera (Olympus, Tokyo, Japan).
Growth inhibition assay. The cells used to determine the growth rate were seeded at a density of 0.5 × 10 5 cells per sample and cultured for up to four days. The number of viable cells was evaluated every 24 h using the trypan blue exclusion method.
Comet assay. To measure DNA damage, the comet assay was performed under alkaline conditions according to the slightly modified procedure described by Singh et al. 69  RNA extraction and quantitative real-time polymerase chain reaction (qRT-PCR). ATM, ATR, AIF, Bax, Bcl-2, Casp-3, H2AX, and PARP (the primer sequences are listed in Table 2) mRNA expression levels were measured by qRT-PCR as described previously 71 . Total RNA was extracted using the TRI Reagent (Sigma-Aldrich, USA) according to the manufacturer instructions, and reverse transcription to cDNA was carried out using the TaqMan Reverse Transcription Reagents (Thermo Scientific, USA  Apoptosis and necrosis detection. The number of cells in various stages of cell death were analysed by double staining with Hoechst 33258 and PI as described previously 72

ROS formation assay.
To measure intracellular ROS formation, fluorescent probe dichlorodihydrofluorescein diacetate (H 2 DCF-DA) was used. H 2 DCF-DA was diffused into cells and deacetylated by cellular esterase to non-fluorescent 2′,7′-dichlorodihydrofluorescein, which was rapidly oxidized by ROS to highly fluorescent 2′,7′-dichlorofluorescein. Fluorescence intensity is proportional to ROS levels within cell cytosol. Briefly, cells were seeded in 96-well plates and incubated with NPs for 24, 48, and 72 h. In some experiments, 1-h cell preincubation with 3 mM antioxidant (N-acetyl cysteine, NAC) was performed before the NPs were added and incubation was continued for the required period of time under the same conditions. Subsequently, the cells were incubated with 5 μM H 2 DCF-DA at 37 °C for 30 min 74 and ROS fluorescence (DCF) was measured using a Fluoroskan Ascent FL microplate reader (Labsystems, Sweden). In addition, DNA content was measured in NP-treated samples in order to avoid measurement errors in fluorescence intensity caused by cell detachment in the well 75 . After measuring H 2 DCFDA fluorescence, the probes were removed by gentle aspiration, 100 μL of deionized water were added to the appropriate wells, and the microplate with cells was frozen at − 70 °C. Before measurement, the microplate with cells was thawed at room temperature and RNA was digested with RNAse. Then, 100 μL of 5 μM PI were added, the plate was incubated for 15 min at room temperature in the dark, shaken, and the fluorescence was read at 535/617 nm by a Fluoroskan Ascent FL microplate reader (Labsystems, Sweden).
Mitochondrial membrane potential (ΔΨm). HUVEC-STs were seeded into 96-well microplates. After 24 h, various concentrations of investigated NPs (100 μg/mL of PLA/MMT/TRA, PLA/EDTMP, PLGA/MDP, and 0.025 μg/mL of Pluronic F127 Ms) or a concentration range of 0.01-5 μM for FCCP, an uncoupling mitochondrial agent (positive control), were added to the wells. The cells were incubated with the analyzed NPs or FCCP for 24, 48, and 72 h. The medium was removed at the end of the treatment, and the cells were incubated in the dark with 5 μM JC-1 in HBSS for 30 min at 37 °C. JC-1 is a fluorescent carbocyanine dye, which accumulates in the mitochondrial membrane in two forms (monomers or dimers), depending on mitochondrial membrane potential. JC-1 monomers show maximum fluorescence excitation and emission at 485 and 538 nm wavelengths, respectively 76 . The fluorescence of both JC-1 monomers and dimers was measured on a Fluoroskan Ascent FL microplate reader (Labsystems, Sweden) using filter pairs of 530/590 nm (dimers) and 485/538 nm (monomers). Statistical analysis. The data were presented as the mean ± SD from three independent experiments and estimated for normal distribution with the Shapiro-Wilk test. The sample size was calculated for type I and type II statistical errors of 0.05 and 0.01, respectively. Subsequently, homogeneity variance was verified with Levene's test. Evaluation of statistically significant differences between control and NP-treated samples was performed using univariate analysis of variance and Tukey's test with a post hoc analysis as described previously 71 . A p-value of 0.05 was considered significant. Statistical analysis was performed using STATISTICA.PL software v.12 (Stat-Soft, Poland). In addition the viability curves were prepared using the GraphPad Prism 5.0 software (GraphPad Inc., USA).