A mini-review on limitations associated with UV filters

2.1.1 Aquatic animals

With both direct and indirect methods of environmental exposure, chemical UV filters can ultimately make their way and accumulate in different bodies of water. Accumulation of UV filters thus exposes various aquatic animals such as fish, cephalopods, and crustaceans, among others to these chemicals. Once in the system of these sea creatures, UV filters are capable of inducing alterations in different metabolic, developmental and reproductive processes (Blüthgen et al., 2014). In addition, these animals make up a portion of the food chain and human diet. Thus, the effects of these UV filters can make their way back to the initial users.

Several studies have been conducted to determine the effects of these UV filters on the aquatic life. In a study done on zebrafish, adult male zebrafish were randomly placed in tanks with 22–383 µg/L of octocrylene. The exposure to this chemical lasted for either 8 or 16 days, with complete tank exchange every 48 h for static-water renewal. In addition, embryos were exposed to 69–925 µg/L of octocrylene for 144 h post fertilization. Embryos were transferred to new octocrylene-containing beakers every 24 h to maintain static water renewal. Total RNA in pooled brain and liver samples of adult zebrafish was determined by microarray analysis, while total RNA of pooled embryos and adult brain, liver, and testes were subjected to RT-qPCR analysis. After 16 days of exposure to octocrylene, 628 instances of altered transcripts in the brain were detected, including but not limited to: transcripts for developmental processes, anatomical structure, positive regulation of cellular processes and organ development. Data from the liver of these fish showed that 136 transcripts were altered including transcripts for xenobiotic metabolic processes, cellular response to xenobiotic stimulus, and urea cycle processes among others. RT-qPCR validated these results. Thus, octocrylene is thought to affect gene transcription in development and metabolic processes (Blüthgen et al., 2014).

In another laboratory study, rainbow trout and Japanese medaka fish were used to determine the estrogenic activity and reproductive effects of oxybenzone. Rainbow trout were exposed to three different concentrations, 10, 100, 1000 µg/L of oxybenzone for 14 days with water renewal occurring every other day. On the last day of the study, the fish were anesthetized, and blood samples were collected. There was a median concentration of 749 µg/L plasma vitellogenin expression, a phospholipoprotein commonly found in female vertebrates. This protein is oftentimes used as a biomarker of environmental estrogen exposure. Based on an unrelated study, immature and mature rainbow trout were found to normally produce less than 10 µg/L of this protein (Copeland et al., 1986). On the other hand, sexually mature female and male Japanese medaka were exposed to 16–620 µg/L oxybenzone for 21 days. The daily number of eggs produced per female was counted and collected for three weeks after oxybenzone exposure; these eggs were observed for hatching. In addition, livers from the fish were excised. The vitellogenin concentration in male medaka was induced, indicating possible feminization, yet was unaltered in females. In addition, it was found that oxybenzone significantly reduced the number of eggs produced by each female fish daily after a week of exposure, yet numbers were able to return to normal in three weeks. Not only were the numbers of eggs depleted, but the percent of eggs that hatched was significantly reduced (Coronado et al., 2008).

Other studies have focused on determining the concentration of UV filters in fish found in lakes and other bodies of water. In a study done in Switzerland, surface water samples were taken throughout the year from different lakes. The peak concentrations of oxybenzone, octocrylene and octinoxate were 35, 5 and 7 ng/L, respectively. White fish, roach and perch from the lakes were also collected. Edible parts of the fish were homogenized and analyzed using accelerated solvent extraction and column extraction. The maximum concentration of UV filters present in these fish was found to be 5 ng/g. However, when accounting for the lost lipid concentrations due to not analyzing the entire fish, it was found that up to 123 ng/g of oxybenzone, 64 ng/g of octinoxate and 25 ng/g of octocrylene were present in the fish. In addition, both untreated and treated samples were obtained from eight municipal waste-water treatment plants. A maximum concentration of 7.8 µg/L of oxybenzone was found in influent wastewater, while 12 µg/L of octocrylene, and 19 µg/L of octinoxate was detected. Values found from effluent wastewater presented a lower concentration for each UV filter (Balmer et al., 2005). In a similar study done in Norway, cod, prawn, crab, perch, whitefish, and burbot were analyzed for the presence of UV filters (Langford et al., 2015). It was found that 80 % of cod livers contained octocrylene, with one fish containing 12 µg/g of this UV filter. In addition, oxybenzone was found in 50 % of cod and prawn samples. Freshwater species such as whitefish contained lower concentrations and detection frequencies of UV filters. Yet, four individual whitefish were found to contain a maximum of 200 ng/g of octinoxate and oxybenzone.

Marine bacteria play a vital role in maintaining the marine ecosystems through symbiotic relation with organisms such as coral, seaweed, algae, and sponge. In a recent study the impact of five FDA-approved sunscreen UV filters (benzophenone-3, octocrylene, ethylhexyl methoxycinnamate, 4-methyl benzylidene camphor and homosalate) was observed in 27 marine bacteria sampled from the Mediterranean Sea (Banyuls sur Mer, France). The bacterial population included seven α-Proteobacteria, three Firmicutes, five Bacteroidetes, nine γ-Proteobacteria and three Actinobacteria. Twenty-six percent (7 out of 27) of tested bacterial species showed a sensitivity to either one or more UV filters at a concentration of 1000 μg L⁻¹. Benzophenone-3 inhibited the growth of P. halotolerans from 100 to 4000 μg L⁻¹, with no significant difference between concentrations from 100 to 4000 μg L⁻¹. A similar trend was noticed for A. aurescens with octocrylene and D. maris/P. glucanolyticus with ethylhexyl methoxycinnamate. On the other hand, A. ornithinivorans, H. dabanensis and E. mobile was inhibited in a dose-dependent manner to homosalate and benzophenone-3, with EC50 of 772, 1000 and 364 μg L⁻¹, respectively (Fig. 4). Octinoxate was found to be the most toxic UV filter at 1000 μg L⁻¹ on growing cells, which affected 5 out of 7 sensitive species. Homosalate and ethylhexyl methoxycinnamate were found to impact both gram-positive and gram-negative bacteria, while benzophenone-3 mainly affected gram-negative bacteria. Further, solar radiation modulated the toxicity of UV filters (Lozano et al., 2020). The studies discussed above clearly indicate environmental risks to the aquatic creatures resulting from the extensive use of sun blockers.

Close

Fig. 4

Bacterial growth curves with different concentrations of UV filters. Only species that showed non monotonic response were presented (average ± standard deviation, n = 3). Arthrobacter aurescens (BBCC 172) against EHMC and OC; Algoriphagus ornithinivorans (BBCC 48) against HS; Dietzia maris (BBCC 167) against EHMC; Epibacterium mobile (BBCC 367) against BP3; Halobacillus dabanensis (BBCC 119) against HS and EHMC; Paenibacillus glucanolyticus (BBCC 237) against EHMC and Pelagibacterium halotolerans (BBCC 52) against BP3 and EHMC. benzophenone-3 (BP3), octocrylene (OC) and 4-methylbenzilidene camphor (4-MBC), homosalate (HS), ethylhexyl methoxy cinnamate (EHMC). Reproduced with permission from Lozano et al. (Lozano et al., 2020).

Export to PPT

2.1.2 Coral reefs

Coral reefs are endangered colonies of marine biodiverse animals that include jellyfish, anemones, and coral. The coral comprising these reef secrete calcium carbonate, which is responsible for the foundation and texture of the reef (Raffa et al., 2019). In other words, the chemicals that the coral emit will incorporate into the structure. Algae, more specifically zooxanthellae, is capable of accumulating at these surfaces and protecting the reefs from any threat or damage (Fig. 5) (Raffa et al., 2019). Yet, if these corals are under any stress, such as exposure to xenobiotics like chemical UV filters, the algae leaves and the coral presents a white appearance, which is referred to as “bleaching” (Fig. 6D). It is this image that raised concerns about the harm UV filters bring to these beings.

Close

Fig. 5

Zooxanthellae release from hard corals in control and sunscreen addition samples. (A) TEM images of healthy zooxanthellae (intact cell structure and membrane) in control untreated Acropora nubbin, and (B) zooxanthellae damaged by sunscreen treatment: cells appear swollen and vacuolated, without chloroplasts and double the size of the controls; the thylakoids are unpacked and dispersed inside the cells, and cell-membrane integrity is lost (arrowhead). (C) Autofluorescence images showing healthy (red) zooxanthellae in control sample and (D) some healthy (H) and damaged and partially damaged (T, transparent and pale) zooxanthellae released after sunscreen treatment. Scale bars = 2 μm (A, B) and 5 μm (C, D). Reproduced with permission from Environmental Health Perspectives, Danovaro et al. (Danovaro et al., 2008).

Export to PPT

Close

Fig. 6

Bleaching of Acropora spp. nubbins caused by the inorganic filters. Photographs of the corals in the control (unexposed corals to inorganic filters; A and B) and exposed to zinc oxide (C and D), Eusolex T2000 (E and F) and Optisol (G and H) at the start (t0) and at the end (after 48 h) of the experiment. Reproduced with permission from Corinaldesi et al. (Corinaldesi et al., 2018).

Export to PPT

One of the first studies addressing the correlation between coral reef bleaching and the use of UV filters was undertaken between 2003 and 2007. Researchers conducted an in-situ experiment in four coral reef locations throughout the world. Small pieces of Acropora spp., Stylophora pistillata and Millepora complanata were collected along with 50 mL samples of the surrounding seawater. These coral samples were incubated and treated with different sunscreens and individual UV filters to determine bleaching. Colorimetric analysis of digital photographs was used to quantify coral bleaching levels. To determine the transfer of sunscreen components into seawater, sunscreen formulas were applied to the hands of two individuals at a dose recommended by the FDA, and then immersed in 2 L of 24 °C, 0.45-µm filtered seawater for 20 min. As a control, the same experiment was conducted with no sunscreen application to the hands; each experiment was done in triplicate. An HPLC-UV was used on the extracted seawater post-hand immersion. It was determined that about 25 % of the mass of sunscreen ingredients originally applied are released into water over 20 min. In addition, octinoxate and oxybenzone were found to cause complete coral bleaching at low concentrations, while other UV filters such as octocrylene, avobenzone, and octisalate had either minor or no bleaching effects. This bleaching effect was determined to be associated with an increase in viral abundance in the seawater. The viral count was measured from water samples by counting and using epifluorescence microscopy. An increase in viral count indicated that sunscreens induced the lytic cycle in zooxanthellae with latent viral infections (Danovaro et al., 2008).

In another study, Stylophora pistillata samples were collected and placed in artificial seawater to determine the effect of light on sunscreen coral bleaching. These coral samples were then exposed to 2.28 µg/L – 228 mg/L of oxybenzone for four different time periods including eight hours of light or dark, as well as twenty-four hours of darkness and a normal twenty-four-hour cycle. Oxybenzone is of particular interest due to its photo-toxicant properties. In the presence of light, this chemical is known to exacerbate its adverse effects. Chlorophyll fluorescence, mortality, morphology, and planula ciliary movement were observed. Within four hours of exposure, there was a significant change in morphology and a reduction of ciliary movement. In addition, “bleaching” of the coral samples was observed as a result of less zooxanthellae present. Transmission electron microscopy showed that tissue deterioration and cell death from the surface to the center was visible in coral exposed to both light and darkness, yet severity increased with exposure to light. In addition, there was an abundance of vacuolated bodies and no plasma membrane in the cell layer below the ciliated cells. Regression models were used to estimate LC₅₀ values. Coral exposed to eight hours of light had an LC₅₀ value of 3.1 mg/L, whereas for eight hours of dark exposure the LC₅₀ value was 16.8 mg/L. On the other hand, the LC₅₀ value after normal twenty-four-hour exposure was 103.8 µg/L, while twenty-four-hour darkness was 873.4 µg/L. This study concluded that autophagy is the dominant cellular response to exposure to oxybenzone (Downs et al., 2016).

Other studies have been conducted to determine the presence of organic UV filters in samples of seawater, coral and sediment. In a study conducted off of the South China Sea, >65 % oxybenzone, and 33–58 % of both octocrylene and octyl dimethyl-p-aminobenzoic acid (ODPABA) in coral tissues were detected. In addition, concentrations of all UV filters located in sediment were the lowest, with the highest concentration measured being 17 ng/g. Water concentrations of these chemicals increased roughly 2-fold, while concentrations present in coral samples varied substantially (3–10-fold) (Tsui et al., 2017). In a similar study done at a coral reef site in Japan, researchers noted that the concentration of octocrylene was 8.1 ng/L and oxybenzone 9.0 ng/L, while at beach sites the concentrations of these UV filters reached 79 ng/L and 1340 ng/L respectively (Tashiro and Kameda, 2013). This suggests that the presence of UV filters can distribute 500 m away from locations where humans swim.

Despite sunscreens being shown to cause coral bleaching in these studies, it is crucial to put into perspective the other environmental factors that also contribute to coral bleaching. Many of these studies are inconclusive due to not taking into account the current state of the coral. For example, industrial pollutants, rising sea temperatures and acidification of the ocean due to climate change can also contribute to coral bleaching (Threats, 2022). In addition, there are other studies that have shown that some chemical sunscreen ingredients are not as harmful to these reefs as is believed (Fel et al., 2019).

2.2.1 In vitro/ex-vivo studies

Organic filters have been studied in vitro with cell culture as well as diffusion models using skin samples. The estrogenic activity of chemical UV filters has been a highly discussed topic. In addition, several studies have tested sunscreen absorption in the skin to determine the extent to which UV filters are absorbed. Many studies have concluded that chemical UV filters are not remaining in the epidermis but rather traveling into the dermis. Some UV filters are capable of generating reactive oxygen species in the cytoplasm of keratinocytes.

A study conducted in MCF-7 breast cancer cells tested nine UV filters for estrogenic activity. Eight of the nine UV filters (which included oxybenzone, homosalate, and octinoxate), produced estrogenic effects by increasing proliferation and pS₂ protein induction. The EC₅₀ values of cell proliferation were also determined. The lowest EC₅₀ value of 0.68 µM was from benzophenone-2, while the highest EC₅₀ value was from oxybenzone at 3.73 µM (Schlumpf et al., 2004). In a different study done on MDA-kb2 cells to determine antiandrogenic activity, oxybenzone had an IC₅₀ value of 4.98 µM, while homosalate had a 5.57 µM value (Ma et al., 2003). For reference, the maximum concentration of oxybenzone allowed in formulations is 60 mg/mL (262 mM), while homosalate is 150 mg/mL (571 mM) (Wang et al., 2011).

In a Franz cell study using excised human skin, water in silicone (W/Si) and water in oil (W/O) emulsions were tested and created using different chemical UV filters. The receptor chamber contained bovine serum albumin in phosphate-buffered saline and was maintained at 37 °C to mimic physiological fluid. 2 mg cm⁻² of sunscreen samples were applied to the skin and then covered with Parafilm. Samples of receptor fluid were taken out at seven time points up to twenty-four hours. HPLC analysis indicated that octinoxate resulted in a 1.21 µg mL⁻¹ cm⁻² penetration in the W/O emulsion and a 0.87 µg mL⁻¹ cm⁻² penetration in the W/Si emulsion. The other FDA-approved UV filters studied were not detected from the receptor fluid (Durand et al., 2009). In a different study using Franz cells, oxybenzone, octinoxate and octisalate were formulated into vehicles of O/W emulgels and petroleum jelly. Approximately 2 mg/cm² of the UV filter samples were placed on human skin samples. Four-time points were investigated ending at six hours after application. Concentrations of all three UV filters were undetected in the receptor fluid of these cells until two hours after application. After two hours post UV filter application, 0.14 % and 0.27 % of oxybenzone from the emulgel and petrolatum vehicles respectively was found in the receptor fluid. At six hours post application, 1.05 % and 4.87 % of oxybenzone was detected. The other two UV filters remained undetected (Treffel and Gabard, 1996). These studies provide evidence that UV filters in formulations are capable of reaching the systemic circulation.

Another Franz cell study was done using full-thickness porcine ear skin and a sunscreen formulated using oxybenzone. The full-thickness skin sheets were gently shaved three times with pressure intensity similar to that of shaving human skin. A finite dose of the sunscreen was spread uniformly onto the skin. A single 1.0 mg/cm² application of sunscreen was evaluated in a six-hour study using previously shaved skin. In a second experiment, sunscreen was reapplied three hours after the first exposure in a six-hour study. Lastly, a single application of sunscreen was applied to freshly shaved skin in a six-hour study. Application to the freshly shaved skin resulted in the highest concentration detected in the receptor fluid compared to the other two trials indicating that sunscreen application right after skin disruption increases UV filter permeation. In addition, reapplying the sunscreen did not result in a 2-fold increase in ‘systemic’ exposure, but rather more than a 2-fold increase in unabsorbed oxybenzone (Hojerová et al., 2017).

Ethylhexyl methoxycinnamate was determined to exhibit clearance from rat and mouse hepatocytes rapidly, with a half-life of less than or equal to 3.16 min. On the other hand, human hepatocyte exposure presented a longer elimination half-life of less than or equal to 48 min (Fennell et al., 2018).

2.2.2 In vivo studies

Since sunscreens are widely used all over the world, in vivo studies have been carried out with the use of human subjects. Some of the simplest tests include applying sunscreen as one would normally do and sampling blood or urine at different time intervals. Since some UV filters are capable of causing visible skin irritation, this type of data has also been collected. Several pre-clinical experiments have been conducted to provide a more mechanistic understanding of the effects of chemical UV filters.

Shaved piglets were used to determine the transdermal penetration of oxybenzone after topical application. Once these piglets were shaved to an area of 150 cm², one gram of product was applied on the shaved area. Scotch tape was applied and pressed for ten seconds, then peeled off to provide skin tape stripping results. This was done 2, 12, and 48 h after product application. In addition, blood samples were collected at eight different time points and urine samples were obtained after 48 h. HPLC-UV was used to determine the concentration of oxybenzone and its metabolites from the samples collected. Concentrations of oxybenzone in the skin were the highest immediately after application, and then decreased with time. The applied dose was mostly contained within the upper level of the stratum corneum after 2 h following administration. The permeation decreased after 12 h yet oxybenzone was mostly distributed between the upper and lower layers of the stratum corneum as well as the surface of the viable epidermis. Plasma concentrations revealed that 12–20 µg/mL of oxybenzone were present within two hours and remained in the plasma up to 48 h. Oxybenzone was also detected in all urine samples during the study (Kasichayanula et al., 2007).

In a study conducted on humans, commercially available sunscreens were given for application. Individuals were told to apply the product to the entire forearm. The formulation was to remain untouched for twelve hours and then removed with soap and water. Each volunteer was told to collect their urine before application as well as 48 h after application. Other drugs, exercise, alcohol, caffeine, and sunlight were to be avoided during the entirety of the study. Urine samples were then analyzed with HPLC-UV. It was determined that roughly 1–2 % of the applied UV filter was absorbed into the systemic circulation over 10 h (Hayden et al., 1997). Similar studies have reported similar conclusions (Lassen et al., 2013, Meeker et al., 2013).

Some UV filters have been linked to endocrine disruption. As a result, a study determined estrogenic activity of chemical UV filters in immature Long Evans rats using the uterotrophic assay. These rats were given UV filters dissolved in olive oil on postnatal days 21, 22, 23 by oral gavage. The doses of UV filters given did not present overt toxicity in the rats. However, six of the nine UV filters tested showed uterotrophic activity by increasing the uterine weight of these rats. Oxybenzone and octinoxate are two of the six UV filters demonstrating this ability. It is hypothesized that the activity of these UV filters is due to their affinity for estrogen receptor beta (ERβ) (Schlumpf et al., 2004). The dose of one UV filter given to the rats when extrapolated to a 74.6 kg individual revealed that the cumulative dose would be achieved after sunscreen application to the entire body every day for 34.6 years(Wang et al., 2011).

In a recent clinical study done by the FDA, researchers randomized 1–4 sunscreen products of different dosage forms containing avobenzone, oxybenzone, octocrylene, homosalate, octinoxate, and octisalate between participants. Sunscreen samples were applied to ∼ 75 % of the body surface area once on day one and then after every-two hours for four-time points over the next three days. Thirty-four blood samples were collected from each participant over the course of 21 days. Individuals were told not shower before the first dose of the day or after the first blood sample collection. UV filters formulated into lotions showed the highest plasma concentrations, while the pump sprays resulted in the lowest plasma concentrations. Oxybenzone regardless of the formula had the highest plasma concentration followed by homosalate. Overall, these compounds reached the systemic circulation in concentrations>0.5 ng/ml (FDA threshold) irrespective of formulation type (Matta et al., 2020). Despite their findings, the FDA suggests that further evaluations need to be done in order to determine what these results mean, and in the meantime, people should not stop using sunscreens.

When testing the metabolism and disposition of ethylhexyl methoxycinnamate in rats and mice, it was determined that two metabolites associated with the hydrolysis of the ester, 2-ethylhexanol and 2-ethylhexanoic acid, are potential developmental and reproductive toxicants. Rodents were given this UV filter by oral gavage or dermally in concentrations ranging from 0.1 to 10 %. This UV filter was primarily excreted in the urine 72 h after oral administration. It was also determined that within this same time period, 34–42 % and 54–62 % of the UV filter dose was absorbed in rats and mice respectively (Fennell et al., 2018).

Oxybenzone was placed in female Harlan Sprague Dawley rat feed to achieve dietary exposure to 3,000–30,000 ppm, which is within 4-fold of the human dose. Blood was collected from randomly selected animals on postnatal days 28 and 56 and assessed through LC-MS/MS. It was determined that free concentrations of the metabolite, 2,4-dihydroxybenzophenone, were greater than free concentrations of the parent compound. Total concentrations were also greater than free concentrations indicating extensive conjugation. Analyte concentrations were also found to be dose and time-dependent (Mutlu et al., 2017).

In a study determining the association between oxybenzone with serum total testosterone levels in children and adolescent participants found that oxybenzone was associated with significantly lower serum total testosterone levels in male adolescents. Participants were interviewed, examined, and provided blood and urine samples. Serum samples were analyzed by isotope dilution liquid chromatography tandem mass spectrometry and urine samples by NCEH/DLS. There was no significant association between oxybenzone and testosterone levels in children of both genders. On the other hand, in female adolescents, total testosterone levels were significantly higher for females with 8.7–23.7 ng/mL oxybenzone exposure versus those with less than 8.7 ng/mL. However, adolescent females with higher than 23.7 ng/mL of oxybenzone exposure showed no significant associations (Scinicariello and Buser, 2016).

It is important to note that despite the findings of all of these studies, there remains an insignificant amount of data about the short term and long-term effects of such ingredients. Studies have looked at the absorption of these actives but little to no studies have shown the half-life, elimination, or clearance in humans. In addition, animal studies are difficult to translate to humans in many cases. Therefore, the accuracy of exposure as well as systemic processes may not be entirely realistic. More studies need to be conducted with repeatable results to fully understand the effects of these ingredients.

3.1.1 In vitro studies

Nanoparticulate systems are frequently used to deliver physical UV filters. Manufacturers use these nanoparticulate systems to avoid the white cast that is disliked by consumers. Yet, due to their size, these nanoparticles are believed to penetrate through the stratum corneum and make their way into the systemic circulation—however, this has still not been proven. In addition, these two chemicals are photocatalysts that are commonly used in photovoltaics (Gilbert et al., 2013). Therefore, with exposure to the UV light, the electrons present in the chemicals become excited and create electron hole pairs, which are capable of inducing reactive oxygen species (Wang et al., 2016).

In a study evaluating the effect of particle size (micro vs nano) in titanium dioxide and zinc oxide particles, cell viability was determined in human astrocytoma U87 cells and human HFF-1 fibroblasts. Methods included an MTT assay, lactate dehydrogenase (LDH) release to determine necrosis, and Annexin V-FITC staining for detecting apoptosis. In U87 cells, exposed to > 1 µg/mL of titanium dioxide, survival was concentration- and time-dependent. Nanoparticles were more potent at lower concentrations, yet the IC₅₀ values of both sizes were relatively the same. Determination of LDH release showed that treatment with both microparticles and nanoparticles of titanium dioxide for forty-eight hours only slightly increased LDH levels when introduced at a concentration above 50 µg/mL, which suggests minimal necrotic damage. Annexin-V staining revealed that apoptotic cells progressively increased with increasing titanium dioxide concentrations. Zinc oxide was less potent than titanium dioxide. It was noted that at concentrations above 20 µg/mL, zinc oxide decreased cell survival to five percent less than the control (Lai et al., 2008). This study suggested that physical UV filters are capable of inducing cell death and damage irrespective of particle size.

The cytotoxicity of these nanoparticles was studied on different cell lines. In human MSTO-211H mesothelioma cells and type 3 T3 Swiss Albino mice fibroblasts, varying concentrations of up to 30 ppm of zinc oxide and titanium dioxide were added in a three-day treatment. An MTT assay as well as a fluorometric DNA assay were used to quantify cell damage and death. At 15 ppm exposure of zinc oxide, virtually all cells died in both cell lines (Brunner et al., 2006). In another study conducted in Neuro-2a neuroblastoma cells, flow cytometry revealed that zinc oxide induced apoptosis in the cells (Jeng and Swanson, 2006). Apoptosis was determined to be caused through the p53-p38 pathway in a study done on human dermal fibroblasts (Meyer et al., 2011).

Another study hypothesized that nanoparticle delivery to the vascular endothelium after acute exposure is capable of inducing inflammation and thus contributing to the development of cardiovascular disease. In order to determine the correlation between inflammation of the aorta and exposure to metal oxide nanoparticles, researchers incubated human aortic endothelial cells (HAEC) with concentrations of up to 50 µg/mL of zinc oxide for four hours. A trypan blue assay was conducted after the incubation period as well as 24 h later. By the end of the incubation period, a 50 µg/mL concentration of zinc oxide nanoparticles resulted in approximately 50 % cell loss, while at 10 µg/mL 20 % cell loss was observed. mRNA levels of inflammatory markers, ICAM-1, IL-8, and MCP-1, were increased with higher concentrations of zinc oxide as well. It was found that at one- and two-hour time points zinc oxide significantly increased inflammatory response (Gojova et al., 2007). However, these findings only show a possible risk, as data has still not proven that these filters enter into circulation.

3.1.2 Ex vivo studies

Results between in vitro and in vivo tests do not always agree. In a study using intact in vivo human skin xenografts, nanoparticles of titanium dioxide in an emulsion were placed for 24 h using an occlusive bandage. After the treatment period, biopsies were taken, and ion microscopy was used to evaluate how far the particles penetrated through the skin layers. Titanium dioxide signal peaks were detected around elements composing the stratum corneum, signifying that these nanoparticles do not reach the deeper cell layers in the skin (Kiss et al., 2008). In a separate study, quantum dot nanoparticles, used for detection purposes, were applied to SKH-1 mice with and without ultraviolet radiation exposure—used to induce skin damage. Exposure to UV rays increased the penetration of these nanoparticles relative to conditions of no UV exposure. UV radiation made it possible for these nanoparticles to travel between corneocytes and into the dermis (Mortensen et al., 2008). Therefore, skin health plays a role in the ability of nanoparticles, such as titanium dioxide or zinc oxide, to penetrate into the skin layers.

3.2.1 Aquatic animals

200 male Nile tilapias held in aquariums were exposed to either 1 or 2 mg/L of zinc oxide nanoparticles alone or in conjunction with 500 mg/kg of a vitamin C and E mixture. Muscles of twenty fish of each group were collected on days 7 and 15 of treatment. It was determined that high levels of malondialdehyde and decreases in antioxidant enzymes (glutathione peroxidase, glutathione reductase, and glutathione-S-transferase), as well as their mRNA expression levels resulted from exposure to zinc oxide nanoparticles alone. In the groups exposed to vitamins C and E, the activities of the antioxidant enzymes were shown to return to normal. This provides evidence of zinc oxide nanoparticles’ ability to induce oxidative stress (Abdelazim et al., 2018).

In another study using fish, white suckers (Castostomus commersonii) were exposed to 1.0 mg/L of nano zinc oxide. Researchers found evidence of oxidative and cellular stress in gill tissue. In addition, gill Na⁺/K⁺-ATPase activity was also threefold higher than untreated fish. It was also found that there were increases in caspase 3/7 activity, heat shock protein expression, and malondialdehyde levels. These fish also exhibited a decrease in heart rate by 35 % with no changes in tissue energy stores or resting oxygen consumption. The authors conclude that tissue damage or cellular stress activates neuroepithelial cells and triggers a hypoxic response in nano zinc oxide exposed fish (Bessemer et al., 2015).

Titanium dioxide nanoparticles were tested for their genotoxic and cytotoxic potential on RTG-2 rainbow trout cells. These properties were tested either in combination with UVA radiation or up to a maximum of 50 µg/mL titanium dioxide alone. Cytotoxicity in particular was determined using the neutral red retention assay, while genotoxicity was measured using cytokinesis-blocked micronucleus assay, single cell gel electrophoresis, or the comet assay. DNA damage was not observed over 48-hour exposure in the absence of UV radiation, yet lysosomal integrity was reduced. However, with 3 kJ/m² of UV exposure, DNA strand breaks and oxidative damage were observed (Vevers and Jha, 2008).

The toxicity of titanium dioxide nanoparticles in the blood, muscle, liver, and brain of Prochilodus lineatus, a Neotropical detritivorous fish, was tested in another study. Titanium dioxide exposure to juvenile fish ranged from 0 to 50 mg/L for either 48 h or 14 days. It was determined that this UV filter acutely decreased white blood cells and increased monocytes. There was no genetic damage observed to red blood cells, yet with subchronic exposure, red blood cells, white blood cells, and lymphocytes decreased indicating this UV filter’s potential to affect the immune system. Increased glutathione-S-transferase and glutathione content with a decrease in superoxide dismutase activity was observed in the liver revealing an alteration in metabolic processes. Accumulation of titanium dioxide was also seen in the muscle and brain causing decreased acetylcholinesterase activity and potential neurotoxicity (Carmo et al., 2019).

3.2.2 Coral reefs

Inorganic UV filters have also been discovered to affect coral reefs. Coral nubbins belonging to Acropora spp. were collected from different colonies along the reef area of Vavvaru Island (Lhaviyani Atoll, Maldives). These nubbins were placed in experimental mesocosms close to the sampling site and acclimatized for 48 h and then washed with virus-free seawater. A sample set of these corals were exposed to roughly 12 % concentrations of inorganic UV filters in the form of uncoated zinc oxide nanoparticles and two modified forms of titanium dioxide (Eusolex® T2000 and Optisol™). It was determined that the uncoated zinc oxide nanoparticles induced rapid and severe coral bleaching resulting from the alteration of the symbiotic relationship between the coral and present zooxanthellae. This UV filter was also found to affect dinoflagellates and stimulate microbial enrichment. Both titanium dioxide forms were not found to cause bleaching and resulted in only minimal alteration in the symbiotic relationship between the coral and zooxanthellae (Corinaldesi et al., 2018) (Fig. 6).

In a study determining the effect of nano zinc oxide on the physical state of the lipid membrane of the coral Seriatopora caliendrum, researchers found that this UV filter caused alterations in the glycerophosphocholine (GPC) profile of the coral’s lipid membrane. Coral colonies were collected from Kenting National Park, Taiwan and maintained in an aquarium near the collection location. The zooxanthellates in the corals were identified as belonging to clade C. The coral was exposed to three nano zinc oxide levels ranging from 50 to 200 µg/L for 24 h. After exposure, the coral was examined using scanning electron microscopy, and stalk positions and tips were collected for lipid extraction. Images showed that nano zinc oxide particles and aggregates accumulated on the gastrodermal surface of the coral. In addition, lipid extraction analysis showed that there were increases in lyso-GPCs and decreases in arachidonic acid-possessing GPCs as a result from the exposure (Tang et al., 2017).

The coral species Acropora yongei was exposed to three individual sunscreen brands—two of which containing only inorganic UV filters—for six hours a day, five days a week for five weeks. Following exposure, five weeks of no sunscreen addition was used as a recovery period. Coral bleaching was observed on almost all of the coral exposed to all sunscreens. However, since stress was similar during nighttime and daytime, the authors conclude that the zooxanthellae were not the source of stress-induced bleaching. This suggests that the UV filters generate oxidative stress once they enter the coral tissue, as was proven from fluorescence data (Romero et al., 2020).