Operationalization and application of “early warning signs” to screen nanomaterials for harmful properties

Abstract

In 2001 the European Environment Agency (EEA) published a report that analyzed 14 cases of technological developments that later on turned out to have negative side-effects and they identified 12 “late lessons” for current and future policy-makers to bear in mind when initiating new technological endeavors. This paper explores how the first lesson – “Acknowledge and respond to ignorance, uncertainty and risk in technology appraisal” could be applied to screen nanomaterials. In cases of ignorance, uncertainty and risk, the EEA recommends paying particular attention to important warning signs such as novelty, persistency, whether materials are readily dispersed in the environment, and whether they bioaccumulate or lead to potentially irreversible action. Through an analysis of these criteria using five well-known nanomaterials (titanium dioxide, carbon nanotubes, liposomes, poly(lactic-co-glycolic acid) and nanoscale zero-valent iron), it was found that only nanoTiO₂ fulfils all the five criteria. Depending on the length of the nanotubes, carbon nanotubes fulfil 3 or 4 criteria whereas liposomes, poly(lactic-co-glycolic acid), nanoscale zero-valent iron fulfil only one criteria. Finally, we discuss how these warning signs can be used by different stakeholders such as nanomaterial researchers and developers, companies and regulators to design benign nanomaterials, communicate what is known about nano-risks and decide on whether to implement precautionary regulatory measures.

---

Environmental impact

There is an urgent need for a set of simple and robust warning signs for evaluating the risks of nanomaterials without complete information to inform risk governance efforts. In 2001 the European Environment Agency (EEA) published a report, Late Lessons from Early Warnings: The Precautionary Principle 1896–2000, which explored 14 case studies, all of which demonstrated how not heeding early warnings had led to a failure to protect human health and the environment (EEA 2001). The EEA recommended paying particular attention to important warning signs such as novelty, persistency, whether the material is readily dispersed, whether it bioaccumulates and whether its use leads to potentially irreversible action. In this article we explore how the EEA warnings signs may be used to screen nanomaterials for human health and environmental risks and we use liposomes, poly(lactic-co-glycolic acid) (PLGA), nanoscale zero-valent iron (nZVI), titanium dioxide (TiO₂), and carbon nanotubes (CNT) as illustrative examples. Based on our analysis, we conclude that the five warning signs can easily be used to screen nanomaterials for harmful properties and we furthermore show that these warning signs can be used by different stakeholders to design benign nanomaterials, communicate what is known about risks related to nanomaterials and decide on whether to implement precautionary regulatory measures.

1. Introduction

The questions of when and how to apply the precautionary principle have been controversial in the past, especially regarding potential risks pervaded with scientific uncertainty and ignorance.^1,2 In Europe, precautionary rationales have been mobilized and contested in regard to the scientific evidence supporting decisions of (de jure or de facto) moratoriums on various chemicals (such as phthalates and brominated flame retardants), beef from Great Britain, animal growth promoters, and genetically modified food. The challenge of finding an appropriate role for precaution, however, has become more relevant than ever with the emergence of new technology oriented research fields such as nanotechnology and synthetic biology, the health and environmental risks of which are often completely unknown. In this realm, we find ourselves at the borderline between enlightened ignorance and complete ignorance about what these risks could be.³

In 2001 the European Environment Agency (EEA) published a report, Late Lessons from Early Warnings: The Precautionary Principle 1896–2000, which explored 14 case studies, all of which demonstrated how not heeding early warnings had led to a failure to protect human health and the environment.² Covering topics as diverse as asbestos, chlorofluorocarbons, non-ionizing radiation and ‘mad cow disease’, the EEA report examined the delay between the emergence of scientific evidence of harm and the action being taken to reduce risks in each case. To avoid repeating past mistakes as new technologies are developed, the expert group identified 12 “late lessons”. This paper explores how the first lesson – “Acknowledge and respond to ignorance, uncertainty and risk in technology appraisal” could be applied in the emerging technological field of nanotechnology. In cases of ignorance, the EEA recommends being proactive, alert and humble about the state of the scientific evidence indicating harm as well as paying particular attention to important warning signs such as novelty, persistency, whether the material is readily dispersed, whether it bioaccumulates and whether its use leads to potentially irreversible action.

The potential risks of especially nanomaterials are as diverse as the field of nanotechnology, and it has yet to be determined as to which inherent properties of nanomaterials determine their hazard potential and fate and behavior in the environment. Current risk assessment procedures for chemicals require substantial information about physico-chemical properties that we know are not applicable or meaningful for nanomaterials. Furthermore, the generation of hazard and exposure information based on guidelines and standard for chemicals has repetitively been identified as being cumbersome and difficult.^4–7

There is an urgent need for a set of simple and robust warning signs for evaluating the risks of nanomaterials without complete information to inform risk governance efforts including, among other things, efforts to assess hazards, exposure and design regulations and monitoring and prioritization of research efforts.

In this article we explore how the EEA warnings signs may be used to screen nanomaterials for human health and environmental risks and we use liposomes, poly(lactic-co-glycolic acid) (PLGA), nanoscale zero-valent iron (nZVI), titanium dioxide (TiO₂), and carbon nanotubes (CNT) as illustrative examples. These nanomaterials were chosen in order to capture the diversity of nanomaterials and challenge the applicability of the five warning signs suggested by the EEA. For each warning sign we will first discuss how it can be understood and operationalized as limited guidance is provided in this regard in the original publication by the EEA and the process by which one can determine whether a given nanomaterial fulfils the warning sign or not. We then describe how the warning sign applies to one illustrative nanomaterial in depth and to the four other nanomaterials in less detail. Finally, we will discuss how these warning signs can be used by different stakeholders such as researchers, developers, companies and regulators to design benign nanomaterials, communicate what is known about nanorisks and decide on whether and how to implement precautionary regulatory measures.

2. Novelty

Based on past experiences with chemicals such as halocarbons, polychlorinated biphenyls (PCBs) and methyl tert-butyl ether (MTBE) the EEA panel stated that the very novelty of these chemicals might be taken as a warning sign.

In the case of halocarbons and PCBs and methyl tert-butyl ether, “novelty” lies in the fact that huge amounts of these chemicals were suddenly used for new applications (spray, refrigerants, PVC plastics, paints, adhesives, lubricants, anti-knocking agent, etc.) and subsequently released into the atmosphere and the environment.²

The EEA does not define the term “novelty”; however this is of key importance in order to determine whether an application of nanotechnology or a given nanomaterial is indeed novel. Without specifying how the term “novelty” should be defined, many definitions of nanotechnology require either “novel applications” and/or that nanomaterials exhibit “novel” properties compared to bulk materials.^8–10 No single exhaustive taxonomy exists for novel materials and as noted by UK Royal Commission on Environmental Pollution⁴ it is unlikely that one is possible or even necessarily desirable. Nevertheless, the UK Royal Commission on Environmental Pollution⁴ distinguished between four types of novel materials:

1. new materials hitherto unused or rarely used on an industrial scale;

2. new forms of existing materials with characteristics that differ significantly from familiar or naturally occurring forms e.g. silver and gold;

3. new applications for existing materials or existing technological products formulated in a new way e.g. cerium oxide used as a fuel additive;

4. new pathways and destinations for familiar materials that may enter the environment in forms different from their manufacture and envisaged use.⁴

In the following we intend to categorize five exemplary materials according to these four overall criteria of novelty.

2.1. Novelty of liposomes

Nano-sized liposomes represent an example of a nanomaterial that falls into the fourth RCEP category of novel materials i.e. nanomaterials “that display new forms of existing materials with characteristics that differ significantly from familiar or naturally occurring forms”.⁴ Liposomes are spherical, self-closed structures formed by one or several concentric lipid bilayers with aqueous phase inside or between the bilayers. Liposomes were first used as a model to study biophysical properties of cell membranes, but have since been used in a variety of applications such as medicines, bioengineering, cosmetics, agro-food and water based paints, glues and resins. In the last 30 years liposomes have also been getting more attention as pharmaceutical carriers in drug formulation. Processed at the nanoscale the ∼100 nm liposomes penetrate biological barriers and flow in the systemic circulation more easily, giving nano-sized liposomes several new advantages as a tool in drug formulation^11,12 and therein lies their novelty.

Liposomes can be used to change the physicochemical properties of active pharmaceutical drugs, either by encapsulation of a hydrophilic drug into the aqueous core or by incorporation of a lipophilic drug into the lipid bilayer. Encapsulation or incorporation protects the drug from degradation and can increase the half-life of the drug in the organism.¹³ Additional liposomes can control drug delivery to a specific area of the body, and in this way increase the therapeutic index of the drug.¹³ A lot of focus has also been in the development of vaccines using liposomes, since liposomes can be used to induce adjuvant effects.¹⁴ Epaxal, a vaccine for hepatitis A, was the first commercially available liposomal drug product to be approved in the 1990's. Today several drug loaded nano-sized liposomes for intravenous administration are on the market, e.g. AmBisome for treatment of fungal infection, where amphotericin B is formulated in 100 nm conventional liposomes. Another available product is Doxil for treatment of cancer, where doxorubicin citrate is formulated in 100 nm liposomes modified with polymer coating. The liposomes are associated with the cells by endocytosis. During endocytosis the liposomes enter the endosomes, where the liposomes are cleaved enzymatically, the phospholipids are hydrolysed and the drug is released into the cytoplasm of the cell.¹⁵ Another widespread application of nano-sized liposomes is in topically applied creams and ointments, where liposomes are utilized to deliver active ingredients into the skin, and increase hydration of the skin. Liposomes have been explored for delivery of calcipotriol into barrier impaired skin for treatment of skin diseases.¹⁶ Although, it has been reported that several cosmetic creams and gels containing liposomes are commercially available, it is unclear how widespread use of liposomes actually is, as claims regarding the presence of intact liposomes in the final product are difficult to validate.¹⁷

2.2. Novelty of carbon nanotubes, TiO₂, nZVI and PLGA

Carbon nanotubes represent an example of a nanomaterial that would fall into the first category of novel materials defined by RCEP⁴ as “(1) new materials hitherto unused or rarely used on an industrial scale”. CNTs were first discovered by Iijima¹⁸ and are often said to be the strongest and stiffest materials known in terms of tensile strength and elastic modulus.¹⁹ Carbon nanotubes (CNT) are high aspect ratio cylindrical nanostructures consisting of one or more layers of carbon atoms and can vary in diameter and length with diameters up to 20 nm and a length ranging from a few micrometers up to millimeters. Surface functionalization can lead to a wide variety of CNTs that vary in their physico-chemical properties as well as biological activity. The strength and flexibility of CNTs suggest that they will have an important role in high-strength material applications such as composite reinforcement or cable components.¹⁹ Other current uses include everyday items like clothes, sports gear and military combat jackets.²⁰

A good example of a “new form of existing materials with characteristics that differ significantly from familiar or naturally occurring forms” (RCEP category 2) is the use of nanoscale zero-valent iron (nZVI) for in situ remediation, which has gained increased attention in recent years as a promising environmentally beneficial nanotechnology. For nZVI the nanoscale size provides significantly increased surface area, leading to higher reactivity per mass unit compared to larger iron particles. This has been found to subsequently enhance contaminant degradation reactions. Furthermore, the nano-sized particles can be prepared as suspensions to be pumped directly into contamination plumes in groundwater aquifers allowing for in situ remediation without the often very costly excavations needed for traditional remediation with zero-valent iron. Today nZVI has been applied for in situ soil and groundwater remediation at contaminated sites in order to degrade a range of contaminants (e.g. PCBs, chlorinated organic solvents, organochlorine pesticides).²¹ Scientists and engineers are increasingly interested in developing various coatings of nZVI in order to better control its reactivity and migration to the environment as discussed further in subsequent sections.

TiO₂ nanoparticles are a good example of a nanomaterial that would fall into the third RCEP category of novel categories, as the use of TiO₂ is nothing new whereas new applications are enabled by formulating TiO₂ in the nanoscale. TiO₂ nanomaterials have the ability to scatter electro-magnetic radiation, which can be manipulated with by changing size, surface treatment and lattice doping. A wide range of applications already exists for TiO₂ nanoparticles e.g. nanoscale TiO₂ is widely used in sunscreens and cosmetics due to the UV-absorption of the material. It is also used in paints and for water treatment as a photo-catalyst producing reactive oxygen species that degrades organic contaminants. Other uses include ointments, toothpaste, catalysts, semi-conductors, etc.^22,23

Finally, biodegradable polymeric particles such as poly(lactide-co-glycolides) (PLGA) could be categorized as “(4) new pathways and destinations for familiar materials that may enter the environment in forms different from their manufacture and envisaged use”⁴ since polymer particles can be engineered into the nanoscale to protect them from premature degradation in vivo, have a specific drug-release rate, or target specific tissues or cells (e.g. specific immunological cell types).^24–27 PLGA has been used in clinical applications at least since 1975,²⁸ the use of the material in itself is not particularly novel. However, a major potential advantage of using PLGA nanoparticles (as compared to microparticles) in drug delivery is that their small size enables them to transgress biological membranes (e.g. epithelial lining and cell membranes), allowing for (passive or active) distribution to desired tissues and cell types.²⁵ Recent immunological research indicates that nanoparticles, depending on their size, may target different immunologically significant cell types that are regarded as important in the design of improved vaccines (i.e. dendritic cells),²⁹ and that PLGA nanoparticles – but not microparticles – allow for the targeting of such cell types.³⁰

3. Persistency of nanomaterials

The second warning sign that the EEA suggests is persistency. Persistency is normally defined as the time for which a substance remains in the environment before it is physically removed or chemically or biologically transformed.³¹ In contrast to “novelty” some fairly firm criteria exist for a given substance or material to be called persistent and these are well embedded in the existing regulatory regimes of organic chemicals. Traditionally the term is related to degradability and in the EU regulation persistency is expressed in terms of media dependent half-lives. This means that a substance is classified as persistent if: T_½ > 60 days in marine water, T_½ > 40 days in fresh or estuarine water, T_½ > 180 days in marine sediment, T_½ > 120 days in fresh or estuarine sediment or T_½ > 120 days in soil. A substance is termed very persistent if T_½ > 60 days in marine, fresh or estuarine water, T_½ > 180 days in marine, fresh or estuarine sediment or T_½ > 180 days in soil.³²

However, since most nanomaterials are inorganic, degradation based criteria for persistency is in most cases not relevant. On the one hand inorganic nanomaterials can be claimed to be persistent per se since the elements cannot be degraded. In this way all inorganic nanomaterials would be classified as persistent; however attention should be given to the fact that the definition of persistency also involves “chemical transformations”. A number of inorganic nanomaterials may be reactive and be transformed to other materials or other forms of the same element depending on the redox conditions and ionic strength of the environmental media into which they are released or where they end up (e.g. uncoated nZVI will be transformed to iron-oxides under aerobic conditions, nano-silver may dissociate to silver-ions in freshwater). These “new” forms may or may not be nanoscale, due to processes like dissolution, oxidation/reduction, agglomeration and aggregation. Thus, only for non-reactive inorganic nanomaterials should applying the term “persistent” be considered whereas reactive inorganic nanomaterials that undergo transformation in the environment or in the human body should not be called persistent. For trace organic contaminants the term “pseudo-persistency” has been introduced to describe the occurrence of ready biodegradable compounds that, because of a continuous (low) discharge, remain in the environment.³³ It may be tempting to describe “pseudo-persistency” as an early warning sign for a number of nanomaterials, since these are expected to be emitted continuously in low concentrations as a result of leaching from consumer products. This term is, however, not suited for this purpose in our view as “pseudo-persistency” depends more on the exposure scenario than on the inherent properties of the material. For nanomaterials the exposure scenarios are at present very uncertain and there is a risk of generating too many false positives if “pseudo-persistency” were to be introduced as a generally applicable early warning sign for nanomaterials.³⁴

3.1. Persistency of PLGA

PLGA degrades in an aqueous environment (in vitro or in vivo) through hydrolysis of the ester bonds that form the backbone of the polymer, releasing the lactic acid and glycolic acid of which it is constituted. In vivo, these acids are eventually metabolized or excreted.³⁵ The degradation rate depends on factors such as co-polymer composition (i.e. the ratio of lactic to glycolic acid), temperature, site of administration, additives (including drugs in delivery systems), size, surface charge and shape of the device or particle, porosity and pH.^24,36–42 Rat experiments have shown that particles made of polymers with a 50 : 50 lactide/glycolide ratios have the fastest degradation rate (e.g. about two months) while polymers of skewed ratios degrade slower. Microparticles made of pure lactide have been found to have degradation rates of more than 1 year in rat implants.^26,43 To the best of our knowledge information on in vivo degradation rates in poikilothermic (i.e. “cold-blooded”) organisms is not available currently. In such organisms, e.g. fish, degradation of the particles may depend more on the organism's immune system than the physical (and temperature) dependent hydrolysis process that will take place in any aqueous environment.

Little research seems to have been conducted to establish degradation rates for PLGA in the environment under realistic environmental conditions, including at temperatures below 5 °C. Dunne et al.⁴⁴ reported no polymer mass loss, but a loss in shape or coalescence of PLGA micro-/nanoparticles was observed after 175 days at 5 °C indicating that PLGA may degrade much less rapidly in low temperature aquatic environments (e.g. as compared to in mammalian bodies or at equivalent temperatures). While the hydrolysis of PLGA may be rather slow in cold water, it nevertheless seems likely that PLGA nanoparticles in an aquatic environment should not be termed as persistent.

3.2. Persistency of TiO₂, nZVI, liposomes and CNT

The degree to which, for instance, nZVI may be persistent in the environment, and especially the extent to which it may be transformed either through biotic or abiotic mechanisms, is not yet clear. For instance, there have only been a handful of studies so far which have investigated this area for uncoated or coated nZVI, although no studies have directly investigated potential persistency per se.²¹ Nonetheless, some authors have documented that some types of bacteria may be able to respire on iron oxide nanoparticles and subsequently alter their surfaces.⁴⁵ Other authors have also suggested that these local dissolution processes of iron nanoparticles through microbial processes may subsequently limit nZVI potential for migration in the environment,⁴⁶ although no suggestions were made in regard to its persistency. When nZVI is used for injection in contaminant plumes or hot spots, the redox conditions will most likely be anoxic (perhaps even anaerobic) and the transformation of nZVI under these conditions has not been described. However, upon transport to aerobic zones (e.g. down gradient of the hot-spot or on the fringe of the contamination plume) nZVI will be oxidised, forming iron-oxides (see also Section 4.1 Dispersibility of nZVI in the paper), and therefore nZVI can be claimed not to be persistent under realistic environmental conditions. It should be noted that the coatings often applied to facilitate the transport and increase the reactivity of nZVI may alter this picture. However, only one study has investigated the potential for different nZVI coatings to desorb.⁴⁷ In this study, the authors found that one polyelectrolyte coating could desorb relatively slowly (less than 30% wt after 4 months) and concluded that the lifetime of these coatings may be approximately a few years. However, biodegradation of some coatings may also be relatively plausible since it has been documented that some microbes may use a variety of substances as food sources in sub-surface environments.⁴⁵

Liposomes and PLGA represent examples of nanotechnology designs that are not persistent and for which increased stability (or persistency) is what developers are trying to achieve in order to increase longer circulation time in the body and/or shelf life of e.g. a pharmaceutical or a cosmetic product. Liposomes mainly consist of biodegradable phospholipids, but different components might be added to adjust the properties of the bilayer such as biodegradable polymer, ethanol, surfactants or cholesterol. One way to increase the stability of liposomes is to coat them with hydrophilic lipopolymers, and long chains of polyethylene glycol (PEG) are commonly used to increase in vivo and in vitro stability.^14,48In vitro PEG chains introduce a sterical barrier between adjacent liposomal lipid bilayers which prevent fusion and aggregation upon storage.⁴⁹In vivo PEG chains reduce the interaction with plasma proteins in the blood, thereby increasing the circulation of liposomes in the blood stream after intravenous (i.v.) administration.^50,51

CNTs are often said to be among the least biodegradable man-made materials known.⁴ Only one paper refers to the possible degradation of SWCNTs via enzymatic catalysis.⁵² After incubation of SWCNTs with a natural horseradish peroxidase (HRP) and low concentrations of H₂O₂ (40 μM) at 4 °C over 12 weeks under static conditions Allen et al.⁵² found indications of degradation of the nanotube structure. These results lead to the suggestion that plant peroxidases may have a role in CNT degradation along with material type and physico-chemical conditions. CNTs can be coated and functionalized and although the CNTs might not be degradable, the coating and functionalization might be so, which again may alter the degradability behavior significantly.⁶ In a study on Daphnia magna Roberts et al.⁵³ showed that water-soluble, lysophosphatidylcholine-coated SWCNTs underwent biological modification upon ingestion as the Daphnia stripped off the lysophosphatidylcholine from the particle surface thereby decreasing the solubility of the SWCNTs. In the presence of Daphnia, initial concentrations of 2.5–20 mg L⁻¹ were decreased by around 50% whereas SWCNTs remained in solution without the D. magna. This finding may be of importance for the evaluation of the fate of modified CNTs after release into the environment.⁶

4. Readily dispersed?

The third warning signal that the EEA suggest concerns the extent to which a substance or material is readily dispersed. Dispersibility is a broad term for which a clear-cut definition is difficult to make; however the intention in the EEA report is to use this as an indicator of whether or not widespread environmental distribution is likely.

Whether a given nanomaterial is readily dispersed depends on a number of factors such as distribution in the air, water, and soil including issues of water soluble versus insoluble materials. Soluble materials generally move through aqueous environments, whereas insoluble particles have different transport mechanisms and may be more easily retained in soil or sediments. In this regard it is important to notice that one of the greatest opportunities of nanotechnology is that it enables engineering of increased solubility through manipulation of surface chemistry and surface charge of nanomaterials.⁵⁴ Under laboratory conditions it has been found that hydrophobic nanoparticles such as C₆₀, CNTs, pure nanometals and nanometal oxides aggregate rapidly in water^55,56 as a result of electric double layer compression.⁵⁷ This indicates that these will precipitate out of the air or water column and end up in the soil or the sediment and hence they would not be readily dispersed. Knowledge to date indicates that in many cases, rather than remaining intact, nanomaterials will tend to aggregate, agglomerate or become associated with other dissolved, colloidal or particulate matter present in the environment.⁵⁸

4.1. Dispersibility of nZVI

The extent to which nZVI may be readily dispersed in the environment is perhaps one of the most well understood parameters within an environmental risk context for nZVI, although it is still relatively unclear to what extent nZVI may migrate in different environmental contexts or become easily dispersed.²¹ The current state-of-the-art knowledge regarding environmental migration of uncoated nZVI is that it is expected to have limited migration (a few centimeters) in most cases^59,60 due to aggregation or agglomeration processes and attaching or depositing onto other surfaces in the environment such as soil or organic particles.⁴⁶ The migration of nZVI in the environment is also strongly dependent on environmental parameters, such as e.g. pH, groundwater geochemistry, oxidation–reduction potential and the nature of the aquifer including groundwater flow rates.^61,62 While bare or uncoated nZVI used in in situ injections are not likely to disperse over great distances based on currently available knowledge, it should also be mentioned that there may be greater chances for environmental ‘hot spots’ of higher concentrations compared to if nZVI had high dispersivity in the environment which would lead to larger areas with low concentrations of nZVI.²¹

Due to these limited migration distances, nZVI has been intentionally engineered to have different applied coatings which would enable increased migration distances in order to have better contact with the environmental contaminants. From the perspective of successfully remediating contaminated soil and groundwater, it is advantageous to ensure better contact between the nZVI nanoparticles and the contaminants in order to allow for sufficient degradation processes to occur. In some cases, where coated or ‘stabilized’ nZVI has been used, it has been estimated that nZVI may migrate up to tens or hundreds of meters in certain environments (i.e. consolidated sandy aquifers),⁵⁹ and possibly remain mobile up to 8 months under some hydrogeological conditions.⁴⁷ It should be stated, however, that these studies have been conducted under laboratory settings as opposed to field-scale conditions, which have natural heterogeneities, which would also likely affect migration and dispersivity in the environment.²¹

In addition to the use of engineered coatings, natural coatings has to be considered as well.⁶³ However there are no studies, which have investigated this for nZVI thus far.

4.2. Dispersibility of carbon nanotubes, TiO₂, liposomes and PLGA

A number of factors have been found to influence the transport of carbon nanotubes, TiO₂, liposomes and PLGA such as aggregation, ionic strength, organic content of the soil, and functionalization (i.e. hydrophilic groups enhance water solubility and thus transport in porous media).

The factors affecting the stability of MWCNTs in natural waters are similar to those identified for C₆₀ (ref. 64 and 65) and in particular, NOM (e.g. humic acids) in water can stabilize MWCNT dispersions for over 1 month, but the NOM adsorption efficiency is dependent on the aromatic hydrocarbon content.⁶⁶ Finally, absorption of other materials present within water bodies (such as bacteria) has also been shown to affect solubility of CNTs. For instance Roberts et al.⁵³ have found that daphnids decreased the solubility of water-soluble, lysophosphatidylcholine-coated SWCNT upon ingestion by stripping off the lysophosphatidylcholine from the particle surface. With regard to soil transport, Lecoanet and Wiesner⁶⁷ have observed that SWCNTs display fairly high mobility compared to other nanomaterials such as functionalized C₆₀ and nanometals as they migrated about 10 m in unfractured sand aquifers. Jaisi et al.⁶⁸ found carboxyl-modified SWCNT deposition to be correlated with ionic strength or the addition of calcium ions when studying SWCNT transport in quartz.

Only one study has so far reported quantitative information on transport of nanomaterials in the environment, tracing leached synthetic TiO₂ nanoparticles from paint on house facades into receiving water bodies.⁶⁹ Kaegi et al.⁶⁹ compared the leaching of TiO₂ nanoparticles from a newly painted model facade and a real facade. Leaching from a newly painted model facade resulted in a titanium concentration of 600 μg L⁻¹ with 90% of the nanoparticles with a size range between 20 and 300 nm. In contrast, the real facade exposed for two years to weather resulted in a leaching of 10 μg L⁻¹ for titanium, while the urban run-off contained 8 μg L⁻¹. The size of particles in the real facade was around 20–300 nm for 90% of the particles, while in urban run-off 50% of particles are <300 nm. As with CNTs, absorption of other materials present within water bodies (such as bacteria) has also been shown to affect transport of metal oxide nanoparticles.⁷⁰ Clearance of metal oxide nanoparticle within aquatic systems is also significantly altered by surface charge and addition of dispersion stabilizing surfactant.⁷⁰ TiO₂ transport was investigated by Lecoanet and Wieser⁶⁷ who found that transport is a function of Darcy velocity, with 55% and 77% of the particles being transported through the porous media under low and high flow velocities, respectively.

Although liposomes might affect the dispersibility of a given drug, the likelihood of being readily dispersed is low as the liposomes can degrade before they are excreted from or washed off the human body or a short while after they enter the environment. Little is known about dispersibility of liposomes in cosmetics, but it is not likely that the liposomes enter the environment as intact nanosized vesicles as surfactants during wash will disturb the hydrophobic interactions.

Little is known about the environmental fate and behavior of PLGA nanoparticles. These issues become particularly pertinent for potential applications of PLGA nanoparticles in agriculture^71,72 or aquaculture⁷³ for which large-scale usage would be conceivable. The dispersibility of PLGA nanoparticles depends on issues like their tendency to aggregate in aquatic environments,⁷⁴ issues of which little or no information is currently available.⁷³

5. Bioaccumulative

Bioaccumulation is the net result of an organism's uptake, distribution and elimination of a given substance or material, when all exposure routes are taken into account, i.e. air, water, soil/sediment or food, is higher than the elimination rate.³¹ Similar to persistency, fairly firm criteria exist for when a given substance or materials are classified as bioaccumulative, which are also well incorporated in the existing regulatory regimes of chemicals. For instance in Europe, a bioconcentration factor (BCF, i.e. the relationship between the concentration of a substance in biological tissue and concentration in water) greater than 2000 L kg⁻¹ a substance leads to a classification as bioaccumulative and a BCF > 5000 L kg⁻¹ as very bioaccumulative.³²

The bioconcentration factor for benthic and terrestrial invertebrates is expressed as biota-to-soil/sediment accumulation factor (BSAF). As a general rule BSAF values of 0.5 and higher are interpreted as an indication of high bioaccumulation. Lower BSAF values should however not be used to conclude that a substance has a low bioaccumulation potential as a low uptake from sediment or soil does not imply a low aquatic BCF value.³²

Whether or not a given nanomaterial will bioaccumulate will depend on properties such as environmental distribution (e.g. solubility in lipid or water) and degradability. For instance, carbon nanotubes are known to be non-biodegradable, insoluble in water and lipophilic (i.e. preference for entering fatty cell membranes) which indicates that carbon nanotubes are likely to bioaccumulate.⁴

5.1. Bioaccumulation of TiO₂

Only a limited number of studies exist on bioaccumulation of TiO₂ in environmentally relevant species. Zhang et al.⁷⁵ observed a significant bioaccumulation in carp (Cyprinus carpio) after exposure to TiO₂ nanoparticles in concentrations of 3 mg L⁻¹ and 10 mg L⁻¹. Concentrations in carp were found to be 2.1 mg g⁻¹ and 5.8 mg g⁻¹ after 25 days and the BCF values at equilibrium are 675.5 and 595.4 respectively. Significant TiO₂ accumulation was observed to occur primarily in the viscera and gills of fish.⁷⁵ The bioconcentration factors in the visceral organs were approximately 2100 at 3 mg L⁻¹ and approximately 1400 at 10 mg L⁻¹. In contrast to Zhang et al.,⁷⁵ Federici et al.⁷⁶ reported no accumulation in the rainbow trout after exposure of up to 1 mg L⁻¹ nanoTiO₂ for 14 days. In the study by Federici et al.⁷⁶ 80% of the water in the fish tank was changed every 12 hours and hence the rainbow trout were exposed to lower concentrations of nanoTiO₂ compared to the carp in the study by Zhang et al.⁷⁵ These two studies furthermore differed with regard to water temperature and carp and trout differ in regard to feeding behavior, which could contribute to explaining the difference in observation made in the two studies.²³

5.2. Bioaccumulation of carbon nanotubes, nZVI, liposomes and PLGA

For carbon nanotubes, liposomes and PLGA some relevant information does exist specific to the bioaccumulation whereas no information has been obtained on nZVI.

Petersen et al. (2008) observed the biota-sediment accumulation factors (BSAFs) for worms (Lumbriculus variegatus) to be almost an order of magnitude lower for SWCNTs and MWCNTs compared to pyrene, which is known to be accumulative.⁷⁷ No indications of any systematic differences were observed between the SWCNTs and MWCNTs. When decreasing the organic carbon content by a factor of 8 a decrease in the BSAF value from 0.51 ± 0.09 to 0.035 ± 0.015 was observed after 14 days of exposure. Petersen et al.⁷⁸ observed that bioaccumulation factors (BAF) of the SWCNTs (99 ± 1% C) and MWCNTs (91.1 ± 0.2% C) by Eisenia foetida were almost 2 orders of magnitude lower than those for pyrene, but a soil dependent difference was observed between SWCNTs and MWCNTs. Bio-uptake of MWCNTs after 14 days was larger than that for SWCNTs in the Chelsea soil and in versa in Ypsilanti soil.

Even though the EEA in their publication “Late Lessons from Early: The Precautionary Principle 1896–2000” mainly focused on bioaccumulation in the environment, accumulation in the human body is relevant as well for a wide range of biomedical applications of nanoparticles and materials including liposomes. In this relation, accumulation is perceived as a positive. One major advantage of nanosized liposomes is their ability to encapsulate the active drug or incorporate the drug in the lipid bilayer to change its physicochemical properties.¹²

The challenge in this regard is to get drugs to accumulate at the injured site.⁷⁹Via i.v. injection of drug-loaded nanosized liposomes a fast clearance by macrophages in obtained whereby drug accumulation is enabled in the liver and spleen.¹¹ A decreased liposomal uptake by macrophages is promoted through stealth liposomes stabilized by polymer coating on the surface. A protective layer is formed over the liposomal surface, which increases circulation time in the bloodstream.⁵¹ Upon tumour growth or inflammation the venous become more permeate, and consequently nano-sized liposomes accumulate at injured sites. Targeting of liposomes to a site can be more specific by conjugation of ligands for specific receptors on the surface of liposomes. This increased delivery of drug into cells expressing certain receptors on the surface, and especially, cells important for cancer progression are extensively studied for a specific delivery using liposomes.⁸⁰ Endocytosis is the main mechanism by which the liposomes enter into the endosomes of the cell. In the endosomes the liposomes are enzymatically cleaved, and the phospholipids are hydrolysed.

Since PLGA nanoparticles and liposomes degrade rather rapidly in mammals, it seems unlikely that uptake rates of PLGA and liposomes from the environment can exceed their rates of degradation.⁸¹ In low temperature environments, degradation rates in poikilothermic animals (e.g. fish) may in addition to the physical hydrolyzation process depend significantly on the immunological properties of the organism. While PLGA nanoparticles might degrade at a lower rate in low temperature environments, it appears unlikely that concentrations of bioavailable PLGA particles can build up to an extent to which bioaccumulation in poikilothermic organisms becomes available. However, this suggestion needs to be experimentally tested under different relevant environmental conditions.

6. Potentially irreversible action

The final warning signal set forward by the EEA² is termed “potentially irreversible action” and hence indicates the possibility of an event (or series of events) that cannot be reverted once it has been initiated. In itself an “irreversible action” does not imply harm, but harm may follow as a consequence of the irreversibility and other intrinsic properties of a material. An example of potentially irreversible action is the manufacture and use of a persistent or bioaccumulative chemical substance. If it later turns out that the chemical has hazardous properties, associated adverse impacts could be irreversible and/or take decades to reverse. In the case of chemicals, examples of hazardous effects could be carcinogenicity, reproductive toxicity, cardiovascular disease, neurotoxicity and endocrine disrupting effects. Another example is the use of a dispersive, not readily degradable chemical (e.g. methyl tert-butyl ether (MTBE)) or the widespread applications of a given technology (e.g. genetically modified organisms). Since use is widespread, any adverse impacts, should they occur, would also be widespread.

6.1. Potentially irreversible action of carbon nanotubes

Since the discovery of carbon nanotubes, concerns have been raised based on the visual similarities with asbestos^82,83 and whether nanotubes might have the same hazardous properties as asbestos.² Since 2004 a series of studies have indicated that some CNTs are able to cause effects that would be classified as irreversible (e.g. Poland et al.⁸⁴ and Smith et al.⁸⁵).

Lam et al.⁸⁶ were the first to report on the pulmonary toxicity of carbon nanotubes after having tested SWCNTs of different purities, i.e. raw and purified nanotubes. Lam et al.⁸⁶ found that all the tested nanotubes induced dose-dependent granulomas and interstitial inflammation in mice. The results by Lam were supported by observations made by Warheit et al.⁸⁷ who also observed pulmonary granulomas in rats after exposure to SWCNT soot (5% Ni and Co).

Although a granuloma could be everything from harmless nevus to malignant tumors, their mere presence is a matter of concern, as inflammation can be seen as a forerunner to lung fibrosis. The fact that purified nanotubes produced prominent granulomas in the study by Lam et al.⁸⁶ indicated that nanotubes were the cause for the observed lesions and not the residue metals often present in batches of non-purified carbon nanotubes. This finding has since been confirmed by others.^88,89

In a direct comparison between long and short MWCNTs, Muller et al.⁸⁹ compared 0.7 μm ground and 5.9 μm unground MWCNTs with chrysotile asbestos and ultrafine carbon black. Muller et al.⁸⁹ found that ground MWCNTs produced a similar inflammatory and fibrogenic response as chrysotile asbestos in rats by equal mass dose and a greater response than ultrafine carbon black.⁸⁹

The finding that length of CNTs affect the biological activity has since then been supported by findings by Takagi et al.⁹⁰ and Muller et al.⁹¹ Takagi et al.⁹⁰ found that long >5 μm MWCNT) caused fibrotic peritoneal adhesions and peritoneal tumors after intraperitoneal injection whereas Poland et al.⁸⁴ observed granulomatous lesions similar to those observed in crocidolite asbestos-treated animals. In contrast, Poland et al.⁸⁴ and Muller et al.⁹¹ observed only acute inflammation when animals were exposed to short <1 μm MWCNTs by intraperitoneal injection.⁹² Finally, mesothelial tumors have also been reported in a susceptible strain of mice after intraperitoneal injection of 10–20 μm MWCNTs⁹⁰ but not by short <1 μm MWCNTs.^91,92

Most studies have used intratracheal or intraperitoneal administration. However, questions have been raised about the biological relevance of intratracheal and intraperitoneal instillation as it bypasses upper respiratory tract defenses and does not deposit particles evenly in the lung similar to what would be observed after inhalation.⁹³ During the last couple of years, a number of nose-only inhalation studies have been published that support previous findings such as Ellinger-Ziegelbauer and Pauluhn,⁹⁴ Ma-Hock et al.⁸⁸ and Pauluhn.⁹⁵ For instance, in a 90-day nose-only inhalation toxicity study in MWCNTs, Ma-Hock et al.⁸⁸ observed an increased lung weight, pronounced multifocal granulomatous inflammation, diffuse histiocytic and neutrophilic inflammation and intra-alveolar lipoproteinosis in lung and lung-associated lymph nodes at 0.5 and 2.5 mg m³. The incidence and severity of granulomatous inflammation of the lung and the lung draining lymph nodes were concentration-dependent, which has previously been demonstrated for intratracheally instilled SWCNTs⁸⁶ and MWCNTs.⁸⁹ Exposure via inhalation also revealed inflammation in the nasal cavity, larynx, and trachea, where the particles deposited during the inhalation exposure as well as alveolar lipoproteinosis, which was not observed while using intratracheal or intraperitoneal administration.⁸⁸

6.2. Potentially irreversible action of TiO₂, nZVI, liposomes and PLGA

The potentially hazardous properties of nanoTiO₂ have mainly focused on respiratory tract toxicity and whether nanoTiO₂ is carcinogenic. According to Stone et al.⁶ several authors have shown that TiO₂ nanoparticles (in a size range of about 20–30 nm) are considerably more toxic than their micro-TiO₂ (>100 nm) counterpart (see e.g. Ferin et al.;⁹⁶ Renwick et al.;⁹⁷ Chen et al.⁹⁸). Most of the studies identified used a single dose of particles, administered via intratracheal instillation and toxicity observations include: pulmonary inflammatory response; epithelial damage, increased permeability of the lung epithelium and cytotoxicity, which were measured within the bronchoalveolar lavage fluid (BALF); and morphological alteration within the lung.⁶

In regard to carcinogenicity, epidemiological evidence remains inconclusive, although one study by Boffetta et al.⁹⁹have observed an elevation in lung cancer mortality among male TiO₂ workers in the latter study when compared to the general population. However, there was no indication of an exposure–response relationship in that study. Heinrich et al.¹⁰⁰ described a statistically significant increase observed in lung cancer in rats following chronic inhalation after repeated exposure. Based on this study the National Institute for Occupational Safety and Health (NIOSH) in the United States has determined that exposure to ultrafine TiO₂ should be considered a potentially occupational carcinogen. NIOSH furthermore “concluded that TiO₂ is not a direct-acting carcinogen, but acts through a secondary genotoxicity mechanism that is not specific to TiO₂ but primarily related to particle size and surface area” and surface area was found to be the critical metric for occupational inhalation exposure to TiO₂. Besides respiratory tract toxicity and carcinogenicity, a number of studies have connected nanoTiO₂ to cardiovascular toxicity¹⁰¹ and neurotoxicity.^102,103 Finally, 35 nm titanium dioxide nanoparticles have recently been found to cause smaller uteri and smaller fetuses after being injected intravenously into pregnant mice. Titanium dioxide nanoparticles were found in the placenta, fetal liver and fetal brain.¹⁰⁴

The in situ use of nZVI for environmental remediation involves the direct injection of nZVI into the environment, and therefore definite environmental exposures occur through its intended use. The question of whether the environment around injection sites – and possibly other areas if nZVI migrates in the environment – may be returned to previous states before introduction of nZVI is not yet known. On one hand, if nZVI may be degraded or transformed through abiotic or biotic mechanisms including biological degradation then it is plausible that the introduction of nZVI into the environment may be essentially ‘reversed’ and the original state(s) of the environment may be returned. On the other hand, if degradation of nZVI is not possible or if the introduction of nZVI results in other unforeseen consequences leading to irreversible changes to the ecosystem, then irreversible consequences may follow nZVI introduction. Furthermore, it is plausible that these irreversible changes may be more problematic if nZVI was found to migrate widely in the environment. In addition to natural or biological degradation mechanisms, there may in theory be technical mechanisms to remove nZVI from the environment, although this possibility is not yet feasible.

Since liposomes can be made from naturally occurring substances it is often said that they are nontoxic, biodegradable and non-immunogenic.¹⁷ However a number of side effects have been observed upon i.v. injection of liposomes in both humans and pigs. In several cases i.v. injection of liposomes loaded with doxorubicin cause an immediate hypersensitivity reaction in the patients.¹⁰⁵ Temporary symptoms include signs of cardiopulmonary distress (dyspnea, tachypnea, hypertension/hypotension); chest and back pain are noted to be rather common among sensitive individuals.^106,107 Although an element of concern, these side effects are not classified as potentially irreversible harmful to humans and environment in general as many of the effects seem to be less severe compared to death, cancer, reproductive toxicity, etc. in addition to being temporary and limited to human patients undergoing medical treatment i.v. under doctor's advice.

As in the case of liposomes, PLGA has not been associated with adverse effects such carcinogenicity, reproductive toxicity, etc., and PLGA has been approved by the US Food and Drug Administration for drug delivery and other clinical applications.²⁵ A number of cytotoxicity studies have been performed on PLGA and have observed no significant adverse effects on several non-carcinogen cell lines.^108,109 In one of the few in vivo studies on PLGA, Semete et al.¹⁰⁹ did detect translocation of rhodamine- and coumarin-labelled 200 and 350 nm PLGA particles in the brain, heart, kidney, liver, lungs and spleen of Balb/C mice 7 days after oral administration of concentrations between 1 and 100 mg mL⁻¹. However no histopathological or tissue damage was observed even at the highest dose.¹⁰⁹

7. Discussion

7.1. Liposomes, PLGA nZVI, TiO₂, CNTs, and the EEA's five warning signs

A quite diverse pattern is seen when the five EEA “early warning signs” are applied to the five nanomaterials included in this paper, as shown in Table 1.

Table 1 CNTs, TiO₂, nZVI, liposomes and PLGA evaluated in regard to novelty, persistency, readily dispersibility in the environment, bioaccumulation and potentially irreversible action

Nanomaterial	EEA warning signs
	Novelty	Persistency	Readily dispersed in the environment	Bioaccumulation	Potentially irreversible action
a “Yes” for short CNTs, “no” for longer CNTs based on Poland et al. (2008). b “Yes” if the enhanced reactivity due to larger surface area is considered to be a “nano-related property”; “no” if compared to reaction processes for conventional ZVI particles. c “Yes” if the study by Dunne et al. (2000) carried out at non-standardized conditions (5 °C) is used; “no” if the Dunne et al. (2000) is rejected.
CNTs	Yes	Yes	No	Yes	Yes/no^a
TiO2	Yes	Yes	Yes	Yes	Yes
nZVI	Yes/no^b	No	No	Unknown	Unknown
Liposomes	Yes	No	No	No	No
PLGA	Yes	Yes/no^c	Unknown	No	No

---

In the case of CNTs, it was evaluated that they can be categorized as novel, persistent and bioaccumulative. No CNTs seem to be readily dispersible in the environment at least in their non-functionalized form. There are indications in the literature that only short and entangled CNTs fulfil the warning signs of use leading to potentially irreversible action, whereas long CNTs might not¹¹⁰. This is indicated in Table 1 by using “yes/no” to indicate either/or. This apparently inconclusive outcome of the evaluation illustrates the important fact that individual kinds of nanomaterials within the same overall material type may have very different hazard profiles. Thus different outcomes can be anticipated when applying the five EEA warning sign indicators to the same nanomaterial. Keeping this in mind, the five warning signs can be used to do a comparative analysis between different types of the same nanomaterial (short CNTs versus long entangled CNTs) but also between various types of nanomaterials (e.g. CNTs, TiO₂).

TiO₂ was found to fulfil all five of the warning signs. The novelty of TiO₂ is primarily related to the ability of nanoTiO₂ to scatter electro-magnetic radiation, which can be manipulated by changing the size, surface treatment and lattice doping.^22,23 TiO₂ was also found to be persistent per se since it is an inorganic nanomaterial that will not be chemically or biologically transformed under normal environmental conditions or in organisms (including mammals). TiO₂ nanoparticles will most likely agglomerate in the aquatic and terrestrial environment; however since part of the size distribution may still be <100 nm and agglomeration is a reversible process, the evaluation as “persistent” is retained. It was also found that TiO₂ is “readily dispersible in the environment” as nanoforms of TiO₂ are used in a range of products that are dispersive in nature (e.g., sunscreens, cosmetics, paints) and will lead to direct environmental exposure. However, it has been shown that TiO₂ and other metal oxide nanoparticles are absorbed by e.g. bacteria present within water bodies, but it is currently not clear whether that increases or decreases the dispersibility of TiO₂ over time and how it is affected by the surface charge and stabilizing surfactant of TiO₂.⁷⁰ Finally, it has been found that TiO₂ nanoparticles can be transported through porous media.⁶⁷

Significant bioaccumulation of TiO₂ has been observed in carp and in regard to potentially irreversible harm, the nanoforms of TiO₂ have consistently been observed to be more toxic than micro-TiO₂ in regard to respiratory tract toxicity.^6,111 Although epidemiological evidence on lung carcinogenicity remains inconclusive,^99,111 the nanoforms of TiO₂ have been associated with cardiovascular toxicity, neurotoxicity as well as reproductive toxicity.^101,102,104

In contrast to CNT and TiO₂, nZVI fulfilled only one of the five warning signs and even for this one (novelty) it may be questioned whether nano-sized iron particles are in fact novel since the nZVI particles are not reported to gain specific properties due to the nanoscale (besides the ease of injecting them in contamination plumes compared to larger particles). The transformation of nZVI to first larger particles through agglomeration and afterwards oxidation to iron particles is the reason why nZVI in this study are categorized as “not persistent”. It is at present unclear whether nZVI are bioaccumulative and there is no evidence that the use of nZVI could lead to potentially irreversible action. Even in the case that it would later turn out that it does, this would be local as there seems to be no evidence that nZVI are readily dispersive in the environment, at least when evaluating the uncoated nanoparticles.

Being novel was the only warning sign being raised in regard to liposomes and PLGA and considering that indications are that both of these nanomaterials are not persistent and degrade when used it follows that they are not bioaccumulative, are readily dispersive or might lead to potentially irreversible action. For PLGA there is one study⁴⁴ that indicates that PLGA is persistent according to the ECHA's persistency criteria. However, the study was carried out at 5 °C, which is not in compliance with international guidelines and a slower degradation at lower temperatures is to be expected for PLGA as well as other substances.

Except for the final EEA warning sign on leading to potentially irreversible action, it is important to realize that a material or process is not inherently dangerous just because it triggers one specific warning sign e.g. novelty or persistency.

7.2. Use of the five warning signs by various stakeholders

Consideration of the five warning signs in regard to one or more nanomaterials and their uses can be useful for different actors and stakeholders throughout the life cycle of the nanomaterial in various ways and with different purposes, as we will illustrate in the following.

In the initial R & D process, nanomaterial researchers and developers can use the warning signs to screen their materials for adverse effects and use this to develop nanomaterials that are “safe by design” from the outset whereas companies that develop and produce nanomaterials can use the warning signs to communicate what they know about the nanospecific risks of the material. Independent parties and regulators can use the warning signs to validate claims about one or more nanomaterials being benign to humans and the environment even in the case of a lack of knowledge. Finally, regulators can use the warning signs to evaluate and prioritize research efforts as well as to trigger default regulatory actions such as bans, moratoriums, etc., depending on how many and which of the warning signs are fulfilled.

7.2.1. Development of nanomaterials that is “safe by design”. Nanomaterial researchers and developers can use the warning signs to screen their materials for adverse effects and use this to develop nanomaterials that are safe by design from the outset.^111,112 If a given nanomaterial has one or more warning signs, developers could try to work deliberately on designing these properties out. For instance, by changing the number of functional groups on C₆₀, the solubility of C₆₀ and hence the dispersibility can be altered. Similarly nZVI could be engineered with coatings that could reduce potential (eco)toxicity or better control environmental migration.

7.2.2. Communication of what is known about NMs by producers. Companies that develop and produce nanomaterials can use the warning signs to communicate what they know about the nanospecific risks of their material in a structured and clear manner. For instance, a company that uses nanoTiO₂ as a UV-filter in sunscreens could provide the following text box along with their product (see Table 2).

Table 2 Examples of how a company could communicate what is known about the use of nanoTiO₂ in sunscreens

NM	Novelty	Persistency	Readily dispersed	Bioaccumulation	Potentially irreversible action
a Scatter electro-magnetic radiation. b Inorganic. c Disperse use, transported through porous media. d Significant bioaccumulation has been observed in carp, enhances the bioaccumulation of Cd and As. e Respiratory tract toxicant inconclusive evidence on lung carcinogenicity remains inconclusive, associated with cardiovascular toxicity, neurotoxicity and as well as reproductive toxicity.
TiO2	Yes^a	Yes^b	Yes^c	Yes^d	Yes^e

---

For liposomes in a medical product, the text box would look as follows (see Table 3).

Table 3 Example of how a company could communicate what is known about the use of liposomes in a medical product

NM	Novelty	Persistency	Readily dispersed	Bioaccumulation	Potentially irreversible action
a ∼100 nm liposomes more easily penetrate biological barriers and the cell membrane. b Not dispersed. Consist of biodegradable phospholipids and degrade in the human body. c Degrade before they are excreted. d Not bioaccumulative. In the endosomes of the cell, the liposomes are enzymatically cleaved, and the phospholipids are hydrolysed. e No potentially irreversible effect observed. Acute adverse allergic-type reactions have been reported to occur as well, within 5–10 min of liposome infusion. Reproducible physiological signs of shock and significant pulmonary hypertension have been observed in pigs after intravenous administration.
Liposomes	Yes^a	No^b	Unknown^c	No^d	No^e

---

7.2.3. Research prioritization by governments and funding agencies. More generically, an analysis of one or more nanomaterials in regard to the warning signs can be used to complete a comparative analysis and rank research efforts in relation to each other depending on the number of outstanding issues noted with an “?” in Table 2. A straightforward approach would be to rank the different types of nanomaterials by the number of yet to be clarified warning signs assuming that all the warning signs are equally important. This would lead to nZVI; liposomes and PLGA being ranked first and hence prioritized as the highest whereas CNT and TiO₂ would be ranked 2^nd and 3^rd. (See Table 4).

Table 4 Ranking of nanomaterials in regard to research priorities depending on the number of outstanding issues noted with an “?”

Res. prior.	NM	Novelty	Persistency	Readily dispersed	Bioaccumulation	Potentially irreversible action
1.	nZVI	Yes	Yes	No	Unknown	No
2.	Liposomes	X	No	Unknown	No	No
3.	PLGA	X	No	Unknown	No	No
4.	CNTs	Yes	Yes	No	Yes	Yes/no
5.	TiO2	Yes	Yes	Yes	Yes	Yes

---

It might be worth noting that some combinations of warning signs are more worrisome than others e.g. persistent and potentially irreversible action might be more worrisome that other combinations of two.

This application of the five warnings signs to guide research prioritize is similar to the approach suggested by Tervonen et al.¹¹³ using MultiCriteria Decision Analysis. Tervonen et al.¹¹³ set forward a number of criteria, both in terms of nanoparticle properties as well bioavailability, bioaccumulation and toxic potential. Quantitative criteria were either measured or based on expert judgments whereas qualitative criteria were established in terms of ordinal classes: 1 was the most favorable (least risk) value class, while 5 the least favorable (highest risk). Weight bounds were then assigned to the criteria by the authors e.g. toxic potential 0.3–0.5, bioavailability and bioaccumulation potentials 0.02–0.08 and the rest of the criteria were assigned weight bounds of 0.05–0.15. A cutting level within the range of 0.65–0.85 was then used to define the minimum sum of weights for the criteria that must be in concordance with the outranking relation to hold.

In regard to the application of the five EEA warning signs, we do not suggest attempting to assign a numeric score on qualitative criteria such as novelty, potentially irreversible harm, etc. In this case, we find that assigning a value of e.g. “2” to a given nanomaterials in regard to e.g. novelty lowers the overall level of transparency as there must be a large level of expert subjectivity when assigning quantitative scores between 1 and 5 to properties that are qualitative in nature.

7.2.4. Precautionary measures implemented by regulators. In a MultiCriteria Mapping exercise among US stakeholders, Hansen¹¹⁴ found that a wide range of policy options were and are available to political decision-makers such as for instance initiation of more nanorisk research, development of safety guidance, implementation of a moratorium as well as banning one or more nanomaterials. Regulators might use the five warning signs to evaluate a given nanomaterial and decide on implementation of various measures depending on the number of warning signs that are fulfilled e.g. 5 out of 5 = ban, 4 out of 5 = moratorium, 3 out of 5 = pre-authorized uses only, etc. 2 out of 5 = limited and controlled uses, 1 out of 5 = implement risk research (see Table 5).

Table 5 Use of the five warning signs to decide on implementation of various measures depending on the number of warning signs that are fulfilled

NM	Novelty	Persistency	Readily dispersed	Bioaccumulation	Potentially irreversible action	Default regulatory action
TiO2	Yes	Yes	Yes	Yes	Yes	Ban
CNTs	Yes	Yes	No	Yes	Yes	Moratorium
nZVI	Yes	Yes	No	Unknown	No	Controlled uses only
Liposomes	Yes	No	Unknown	No	No	Invest in risk research
PLGA	Yes	No	Unknown	No	No	Invest in risk research

---

There is currently a substantial risk of “paralysis by analysis” in the field of risk assessment and management and regulation of nanomaterials¹¹⁵ and calls for more research are often made. It could seem somewhat too simplistic to base complex regulatory decision on such simple decision-making rules and principles that again would trigger the default implementation of precautionary measures. However, having such default decision-making rules and principles has a number of advantages. First of all, it would make it clear as to which criteria and principles are those on which the decisions to implement precautionary measures are based and it would also initiate discussions in cases where regulators decided to deviate from the default decision making processes and they would have to explain why they decided not to implement precautionary measures, something regulators hardly ever have to do currently. It would furthermore be possible to overcome the limitations of a case-by-case risk assessment of nanomaterials and expose any attempt by various stakeholders to delay the implementation of risk reducing measures.

8. Conclusion

In cases of ignorance, uncertainty and risk, the European Environment Agency (EEA) in 2001 recommended paying particular attention to important warning signs such as novelty, persistency, whether materials are readily dispersed in the environment, whether they bioaccumulate or lead to potentially irreversible action. After having applied these warning signs to five well-known nanomaterials (titanium dioxide, carbon nanotubes, liposomes, poly(lactic-co-glycolic acid) and nanoscale zero-valent iron), we found that only nanoTiO₂ fulfils all the five criteria. Depending on the length of the nanotubes, carbon nanotubes fulfils 3 or 4 criteria whereas liposomes, poly(lactic-co-glycolic acid), and nanoscale zero-valent iron fulfil only one criteria. Based on this analysis, we conclude that the five warning signs can easily be used to screen nanomaterials for harmful properties and we furthermore show that these warning signs can be used by different stakeholders to design benign nanomaterials, communicate what is known about risks related to nanomaterials and decide on whether to implement precautionary regulatory measures.

Notes and references

1. B. Wynne, Global Environ Change, 1992, 2, 111–127.
2. EEA, Late Lessons From Early Warnings: The Precautionary Principle 1896–2000, European Environmental Agency, Copenhagen, 2001.
3. K. Grieger, S. F. Hansen and A. Baun, Nanotoxicology, 2009, 3(3), 1–12.
4. RCEP, Twenty-Seventh Report Novel Materials in the Environment: The Case of Nanotechnology Royal Commission on Environmental Pollution, The Stationery Office, Norwich, 2008.
5. S. F. Hansen, Regulation and Risk Assessment of Nanomaterials – Too Little, Too Late?, Technical University of Denmark, Kgs. Lyngby, 2009.
6. V. Stone, S. Hankin, R. Aitken, K. Aschberger, A. Baun, F. Christensen, T. Fernandes, S. F. Hansen, N. B. Hartmann, G. Hutchinson, H. Johnston, C. Micheletti, S. Peters, B. Ross, B. Sokull-Kluettgen, D. Stark and L. Tran, Engineered Nanoparticles: Review of Health and Environmental Safety (ENRHES), Project Final Report, European Commission, Brussels, 2009.
7. K. Grieger, I. Linkov, S. F. Hansen and A. Baun, Nanotoxicology, 2012, 6(2), 196–212.
8. Nanoscale Science Engineering and Technology Subcommittee, The Committee on Technology, National Science and technology Committee, The National Nanotechnology Initiative Strategy Plan December 2004, National Nanotechnology Coordination Office, Arlington, VA, 2004.
9. A. S. T. M. Int'l, E 2456-06 Terminology for Nanotechnology, ASTM International, Pennsylvania, 2006.
10. Royal Society & the Royal Academy of Engineering, Nanoscience and Nanotechnologies: Opportunities and Uncertainties, Royal Society, London, 2004.
11. V. P. Torchilin, Nat. Rev. Drug Discov, 2005, 4(2), 145–160.
12. V. P. Torchilin, Eur. J. Pharm. Biopharm., 2009, 71(3), 431–444.
13. H. Devalapally, A. Chakilam and M. M. Amiji, J. Pharm. Sci., 2006, 96(10), 2547–2565.
14. M. L. Immordino, F. Dosio and L. Cattel, Int. J. Nanomed., 2006, 1, 297–315.
15. B. Maherani, E. R. Arab-Tehrany, M. Mozafari, C. Gaiani and M. Linder, Current Nanoscience, 2011, 7(3), 436–452.
16. N. Ø. Knudsen, L. Jorgensen, J. Hansen, C. Vermehren, S. Frokjaer and C. Foged, Int. J. Pharm., 2011, 416(211), 478–485.
17. D. D. Lasic, Chapter 10 Applications of Liposomes, in Handbook of Biological Physics, ed. R. Lipowsky and E. Sackmann, Elsevier Science B.V, 1995, pp. 491–519.
18. S. Iijima, Nature, 1991, 354(6348), 56–58.
19. M.-F. Yu, O. Lourie, M. J. Dyer, K. Moloni, T. F. Kelly and R. S. Ruoff, Science, 2000, 287(5453), 637–640.
20. PEN 2011, Consumer Products, http://www.nanotechproject.org/inventories/consumer/, accessed 9 August 2011.
21. K. Grieger, A. Fjordbøge, N. B. Hartmann, E. Eriksson, P. L. Bjerg and A. Baun, J. Cont. Hydro., 2011, 118(3–4), 165–183.
22. Environmental Defense, DuPont, Nano Risk Framework, Environmental Defense and DuPont, Washington, D.C., 2007.
23. US EPA, External Review Draft Nanomaterial Case Studies: Nanoscale Titanium Dioxide in Water Treatment and in Topical Sunscreen, November 2010 EPA/600/R-09/057F National Center for Environmental Assessment Office of Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC., 2010.
24. F. Esmaeili, M. H. Ghahremani, B. Esmaeili, M. R. Khoshayand, F. Atyabi and R. Dinarvand, Int. J. Pharm., 2008, 349, 249–255.
25. J.-M. Lü, X. Wang, C. Marin-Muller, H. Wang, P. H. Lin, Q. Yao and C. Chen, Expert Rev. Mol. Diagn., 2009, 9, 325–341.
26. R. C. Mundargi, V. R. Babu, V. Rangaswamy, P. Patel and T. M. Aminabhavi, J. Controlled Release, 2008, 125, 193–209.
27. J. Siepmann and F. Siepmann, Prog. Colloid Polym. Sci., 2006, 133, 15–21.
28. D. K. Gilding and A. M. Reed, Polymer, 1979, 20, 1459–1464.
29. V. Manolova, A. Flace, M. Bauer, K. Schwarz, P. Saudan and M. F. Bachmann, Eur. J. Immunol., 2008, 38, 1404–1413.
30. L. J. Cruz, P. J. Tacken, R. Fokkink, B. Joosten, M. C. Stuart, F. Albericio, R. Torensma and C. G. Figdor, J. Controlled Release, 2010, 144, 118–126.
31. C. J. van Leeuwen and J. L. M. Hermens, Risk Assessment of Chemicals – An Introduction, Kluwer Academics, Dordrecht, The Netherlands, 1995.
32. ECHA 2008. Guidance on information requirements and chemical safety assessment Chapter R.11: PBT Assessment, http://guidance.echa.europa.eu/docs/guidance_document/information_requirements_r11_en.pdf?vers=20_08_08, accessed 17 June 2011.
33. J. Diamond, H. Latimer, K. Munkittrick, K. Kidd, K. Thornton and S. Bartell, Diagnostic Tools to Evaluate Impact of Trace Organic Compounds, Prioritization Framework for Trace Organic Compounds, Water Environment Research Foundation, IWA Publishing, London, UK, 2010.
34. S. F. Hansen, M. P. Krayer von Krauss and J. A. Tickner, Risk Analysis, 2007, 27(1), 255–269.
35. K. A. Athanasiou, G. G. Niederauer and C. M. Agrawal, Biomaterials, 1996, 17, 93–102.
36. K. Avgoustakis, A. Beletsi, Z. Panagi, P. Klepetsanis, E. Livaniou, G. Evangelatos and D. S. Ithakissios, Int. J. Pharm., 2003, 259, 115–127.
37. L. R. Beck, V. Z. Pope, C. E. Flowers, D. R. Cowsar, T. R. Tice, D. U. Lewis, R. L. Dunn, A. B. Moore and R. M. Gilley, Poly (DL-lactide-co-glycolide)/norethisterone microcapsules: an injectable biodegradable contraceptive, Biol. Reprod., 1983, 28, 186–195.
38. M. K. Hassan, K. A. Mauritz, R. F. Storey and J. S. Wiggins, Polymer, 2006, 47, 1960–1969.
39. J. Panyam, M. M. Dali, S. K. Sahoo, W. Ma, S. S. Chakravarthi, G. L. Amidon, R. J. Levy and V. Labhasetwara, J. Controlled Release, 2003, 92, 173–187.
40. J. Panyam and V. Labhasetwar, Adv. Drug Delivery Rev., 2003, 55, 329–347.
41. Y.-Y. Yang, T.-S. Chung and N. Ping Ng, Biomaterials, 2001, 22, 231–241.
42. H. Yushu and S. Venkatraman, J. Appl. Polym. Sci., 2006, 101, 3053–3061.
43. J. M. Anderson and M. S. Shive, Adv. Drug Delivery Rev., 1997, 28, 5–24.
44. M. Dunne, O. I. Corrigan and Z. Ramtoola, Biomaterials, 2000, 21, 1659–1668.
45. R. Gerlach, A. B. Cunningham and F. Caccavo Jr, Environ. Sci. Technol., 2000, 34(12), 2461–2464.
46. G. V. Lowry and E. Casman, in Nanotechnology: Risks and Benefits, ed. I. Linkov and J. Steevens, Springer, Dordrecht, 2009, p. 125.
47. H. Kim, T. Phenrat, R. D. Tilton and G. V. Lowry, Environ. Sci. Technol., 2009, 43(10), 3824–3830.
48. N. Ø. Knudsen, S. Rønholt, R. Djønne Salte, L. Jorgensen, T. Thormann, L. Hollesen Basse, J. Hansen, S. Frokjaer and C. Foged, Eur. J. Pharm. Biopharm., 2012 , in press.
49. D. Needham, T. McIntosch and D. Lasic, Biochim. Biophys. Acta, 1992, 1108, 40–48.
50. D. D. Lasic, Trends Biotechnol., 1998, 16, 307–321.
51. M. Woodle, Chem. Phys. Lipids, 1993, 64, 249–262.
52. B. L. Allen, P. D. Kichambare, P. Gou, I. I. Vlasova, A. A. Kapralov, N. Konduru, V. E. Kagan and A. Star, Nano Lett., 2008, 8, 3899–3903.
53. A. P. Roberts, A. S. Mount, B. Seda, J. Souther, R. Qiao, S. J. Lin, P. C. Ke, A. M. Rao and S. J. Klaine, Environ. Sci. Technol., 2007, 41(8), 3025–3029.
54. C. M. Sayes, J. D. Fortner, W. Guo, D. Lyon, A. M. Boyd, K. D. Ausman, Y. J. Tao, B. Sitharaman, L. J. Wilson, J. B. Hughes, J. L. West and V. L. Colvin, Nano Lett., 2004, 4(10), 1881–1887.
55. J. D. Fortner, D. Y. Lyon, C. M. Sayes, A. M. Boyd, J. C. Falkner, E. M. Hotze, L. B. Alemany, Y. J. Tao, W. Guo, K. D. Ausman, V. L. Colvin and J. B. Hughes, Environ. Sci. Technol., 2005, 39(11), 4307–4316.
56. A. Baun, S. N. Sørensen, R. F. Rasmussen and N. I. Bloch, Aquat. Toxicol., 2008, 86(3), 379–387.
57. Y. Zhang, Y. Chen, P. Westerhoff, K. Hristovski and J. C. Crittenden, Water Res., 2008, 42(8–9), 2204–2212.
58. Scientific Committee on Emerging and Newly Identified Health Risks (SCENIHR), Risk Assessment of Products of Nanotechnologies, European Commission health and consumer Protection Directorate – general, Directorate C – Public Health and Risk Assessment, C7-Risk Assessment, Brussels, 2009.
59. N. Saleh, H. Kim, T. Phenrat, G. V. Lowry, R. D. Tilton and K. Matyjaszewski, Environ. Sci. Technol., 2008, 42(9), 3349–3355.
60. P. G. Tratnyek and R. L. Johnson, Nano Today, 2006, 1, 44–48.
61. X. Li, D. W. Elliott and W. Zhang, Crit. Rev. Solid State Mater. Sci., 2006, 31, 11–22.
62. K. Sellers, C. Mackay, L. L. Bergeson, S. R. Clough, M. Hoyt, J. Chen, K. Henry and J. Hamblen, Nanotechnology and the Environment, Taylor & Francis Group, Boca Raton, 2009.
63. H. Hyung, J. D. Fortner, J. B. Hughes and J. H. Kim, Environ. Sci. Technol., 2007, 41, 179–184.
64. A. J. Kennedy, J. C. Gunter, M. A. Chappell, J. D. Goss, M. S. Hull, R. A. Kirgan and J. A. Steevens, Environ. Toxicol. Chem., 2009, 28(9), 1930–1938.
65. B. Han and M. N. Karim, Scanning, 2009, 30(2), 213–220.
66. H. Hyung and J.-H. Kim, Environ. Sci. Technol., 2008, 42(12), 4416–4421.
67. H. F. Lecoanet and M. R. Wiesner, Environ. Sci. Technol., 2004, 38(16), 4377–4382.
68. D. P. Jaisi and M. Elimelech, Environ. Sci. Technol., 2008, 43(24), 9161–9166.
69. R. Kaegi, A. Ulrich, B. Sinnet, R. Vonbank, A. Wichser, S. Zulegg, H. Simmler, S. Brunner, H. Vonmont, M. Burkhardt and M. Boller, Environ. Pollut., 2008, 156(2), 233–239.
70. L. K. Limbach, R. Bereiter, E. Muller, R. Krebs, R. Gälli and W. J. Stark, Environ. Sci. Technol., 2008, 42, 5828–5833.
71. J. Kuzma, Livest. Sci., 2010, 130, 14–24.
72. J. Kuzma, J. Romanchek and A. Kokotovich, Risk Analysis, 2008, 4, 1081–1098.
73. K. N. Nielsen, B. N. Fredriksen and A. I. Myhr, Nanoethics, 2011, 5, 57–71.
74. A. R. Petosa, D. P. Jaisi, R. v. Quevedo, M. Melimelech and N. Tufenkji, Environ. Sci. Technol., 2010, 44, 6532–6549.
75. X. Z. Zhang, H. W. Sun and Z. Y. Zhang, Huanjing Kexue, 2006, 27(8), 1631–1635.
76. G. Federici, B. J. Shaw and R. R. Handy, Aquat. Toxicol., 2007, 84, 415–430.
77. L. P. Burkhard and M. T. Lukasewycz, Environ. Toxicol. Chem., 2000, 19, 1427–1429.
78. E. J. Petersen, Q. Huang and W. J. Weber, Jr., Environ. Health Perspect., 2008, 116, 496–500.
79. G. J. Charrois and T. M. Allen, Biochim. Biophys. Acta, 2003, 1609(1), 102–108.
80. R. M. Schiffelers and G. Storm, Int. J. Pharm., 2008, 364(2), 258–264.
81. B. N. Fredriksen and J. Grip, PLGA/PLA micro- and nanoparticle formulations serve as antigen depots and induce elevated humoral responses in immunized Atlantic salmon (Salmo salar L), Vaccine, 2012, 30(3), 656–667.
82. A. Huczko, H. Lange, E. Calko, H. Grubek-Jaworska and P. Droszcz, Fullerene Sci. Technol., 2001, 9(2), 251–254.
83. A. Huczko, H. Lange, M. Bystrzejewski, P. Baranowski, H. Grubek-Jaworska, P. Nejman, T. Przybylowski, K. Czuminska, J. Glapinski, D. R. M. Walton and H. W. Kroto, Fullerenes, Nanotubes, Carbon Nanostruct., 2005, 13, 141–145.
84. C. A. Poland, R. Duffin, I. Kinloch, A. Maynard, W. A. H. Wallace, A. Seaton, V. Stone, S. Brown, W. Macnee and K. Donaldson, Nat. Nanotechnol., 2008, 3, 423–428.
85. C. J. Smith, B. J. Shaw and R. D. Handy, Aquat. Toxicol., 2007, 82(2), 94–109.
86. C. W. Lam, J. T. James, R. McCluskey and R. I. Hunter, Toxicol. Sci., 2004, 77, 126–134.
87. D. B. Warheit, B. R. Laurence, K. L. Reed, D. H. Roach, G. A. M. Reynolds and T. R. Webb, Toxicol. Sci., 2004, 77(1), 117–125.
88. L. Ma-Hock, S. Treumann, V. Strauss, S. Brill, F. Luizi, M. Mertler, K. Wiench, A. O. Gamer, B. Ravenzwaay and R. Landsiedel, Toxicol. Sci., 2009, 112(2), 468–481.
89. J. Muller, F. Huaux, N. Moreau, P. Misson, J. F. Heilier, M. Delos, M. Arras, A. Fonseca, J. B. Nagy and D. Lison, Toxicol. Appl. Pharmacol., 2005, 207(3), 221–231.
90. A. Takagi, A. Hirose, T. Nishimura, N. Fukumori, A. Ogata and N. Ohashi, J. Toxicol. Sci., 2008, 33(1), 105–116.
91. J. Muller, M. Delos, N. Panin, V. Rabolli, F. Huaux and D. Lison, Toxicol. Sci., 2009, 110(2), 442–448.
92. NIOSH 2011. NIOSH Current Intelligence Bulletin Occupational Exposure to Carbon Nanotubes and Nanofibers. Department Of Health And Human Services Centers for Disease Control and Prevention, National Institute for Occupational Safety and Health, Washington, D.C.
93. M. Oiser and G. Obedörster, Fundam. Appl. Toxicol., 1997, 40, 220–227.
94. H. Ellinger-Ziegelbauer and J. Pauluhn, Toxicology, 2009, 266, 16–29.
95. J. Pauluhn, Toxicol. Sci., 2010, 113(1), 226–242.
96. J. Ferin, G. Oberdorster and D. P. Penney, Am. J. Respir. Cell Mol. Biol., 1992, 6(5), 535–542.
97. L. C. Renwick, D. Brown, A. Clouter and K. Donaldson, Occup. Environ. Med., 2004, 61, 442–447.
98. H.-W. Chen, S.-F. Su, C.-T. Chien, W.-H. Lin, S.-L. Yu, C.-C. Chou, J. J. W. Chen and P.-C. Yang, FASEB J., 2006, 20, E17312.
99. P. Boffetta, A. Soutar, J. W. Cherrie, F. Granath, A. Andersen, A. Anttila, M. Blettner, V. Gaborieau, S. J. Klug, S. Langard, D. Luce, F. Merletti, B. Miller, D. Mirabelli, E. Pukkala, H.-O. Adami and E. Weiderpass, Can. Cau. Cont., 2004, 15(7), 697–706.
100. U. Heinrich, R. Fuhst, S. Rittinghausen, O. Creutzenberg, B. Bellmann, W. Koch and K. Levsen, Inhalation Toxicol., 1995, 7(4), 533–556.
101. M. Helfenstein, M. Miragoli, S. Rohr, L. Muller, P. Wick, M. Mohr, P. Gehr and B. Rothen-Rutishauser, Toxicology, 2008, 253, 70–78.
102. T. C. Long, N. Saleh, R. D. Tilton, G. V. Lowry and B. Veronesi, Environ. Sci. Technol., 2006, 40, 4346–4352.
103. T. C. Long, J. Tajuba, P. Sama, N. Saleh, C. Swartz, J. Parker, S. Hester, G. V. Lowry and B. Veronesi, Environ. Health Perspect., 2007, 115, 1631–1637.
104. K. Yamashita, Y. Yoshioka, K. Higashisaka, K. Mimura, Y. Morishita, M. Nozaki, T. Yoshida, T. Ogura, N. Nabeshi, K. Nagano, Y. Abe, H. Kamada, Y. Monobe, T. Imazawa, H. Aoshima, K. Shishido, Y. Kawai, T. Mayumi, S. Tsunoda, N. Itoh, T. Yoshikawa, I. Yanagihara, S. Saito and Y. Tsutsumi, Nat. Nanotechnol., 2011, 6, 321–328.
105. A. Chanan-Khan, J. Szebeni, S. Savay, L. Liebes, N. M. Rafique, C. R. Alving and F. M. Muggia, Ann. Oncol., 2003, 14, 1430–1437.
106. J. Szebeni, Toxicolology, 2005, 216(2–3), 106–121.
107. S. M. Moghimi and I. Hamad, J. Liposome Res., 2008, 18(3), 195–209.
108. R. de Lima, A. do Espirito Santo Pereira, R. Martins Porto and L. Fernandes Fraceto, J. Polym. Environ., 2011, 19(1), 196–202.
109. B. Semete, L. Booysen, Y. Lemmer, L. Kalombo, L. Katata, J. Verschoor and H. S. Swai, Nanomed. Nanotechnol. Biol. Med., 2011, 6, 662–671.
110. NIOSH, Current Intelligence Bulletin 63 Occupational Exposure to Titanium Dioxide, Department of Health and Human Services Centers for Disease Control and Prevention National Institute for Occupational Safety and Health, Washington, D.C., 2011.
111. R. S. Boethling, E. Sommer and D. DiFiore, Chem. Rev., 2007, 107, 2207–2227.
112. B. M. C. Kelty, Nanoethics, 2009, 3, 79–96.
113. T. Tervonen, J. Figueira, J. Steevens, J. Kim and I. Linkov, J. Nanopart. Res., 2009, 11, 757–766.
114. S. F. Hansen, J. Nanopart. Res., 2010, 12(6), 1959–1970.
115. S. F. Hansen, A. Maynard, A. Baun and J. A. Tickner, Nat. Nanotechnol., 2008, 3(8), 444–447.

---