Comparison of Phototoxicity Sensitivity by the Neutral Red Uptake Method for BALB/c 3T3, HaCaT, and HDFa Cells In Vitro

Phototoxicity is defined as a toxic response elicited by topically or systemically administered photoreactive chemicals after exposure of the body to environmental light.^[1] Photoactive chemical substances are excited by sunlight and converted into cytotoxic products, causing abnormal skin conditions.^[2] In general, symptoms of skin phototoxicity include skin irritation, erythema, itching, edema, and sunburn.^[3] With the introduction of the 3R principle (reduce, reuse, recycle) and promotion of in vitro replacement of cosmetic testing, and compared with traditional rabbit eye phototoxicity assessment,^[4] the in vitro 3T3 neutral red uptake (NRU) phototoxicity test, which is more animal friendly, was proposed by the Organization for Economic Co-operation and Development (OECD) and is listed as a test standard.^[1] In 2012, the European Centre for the Validation of Alternative Methods–European Federation of Pharmaceutical Industries and Associations (ECVAM-EFPIA) workshop evaluated the practical application of the test method. They believe that the 3T3 NRU phototoxicity test can determine 100% of phototoxic substances and has high sensitivity, but positive results on the test often do not translate into a clinical phototoxicity risk.^[5] There is a high frequency of positivity, which means that the method is oversensitive; as a result, the 3T3 NRU phototoxicity assay can lead to a great deal of follow-up work to assess whether there is any risk to humans.^[6]

Keratinocytes represent the main cell types of the epidermis, whereas fibroblasts are mainly distributed in the dermis.^[7] At present, using normal human skin fibroblasts and human epithelial keratinocytes to study UV irradiation, UV protection activities and repair of DNA damage caused by UV light induction have been reported.^[8–10] Keratinocytes are the most human skin-relevant cells to be affected by sunlight stimulation,^[11,12] and several studies have reported that keratinocytes were used to estimate phototoxicity models in vitro and showed effective UV light stimulation evaluation of the test substance.^[13–18] In addition, a report explored the differences in the effect of UVA light on the Nrf2 signaling pathway in human cells.^[19] Thus, choosing skin cells to identify the potential phototoxicity of a test chemical following activation by exposure to UV light need not only be limited in BALB/c 3T3 cells; human cell lines could be used for assessments and to seek ways to optimize the existing overly high sensitivity of the in vitro assay.

In OECD Test No. 432, chlorpromazine hydrochloride (CPZ) and amiodarone hydrochloride (Amiodar) are both evaluated as phototoxic-positive compounds, and L-histidine (L-His) is used as a negative control. Structures are shown in Figure 1.^[1] Between the two positive controls, Amiodar, which has a relatively lower photo-irritation factor (PIF) and the mean photo effect (MPE) values, has few reports on phototoxicity in human skin cell lines. This study focuses on the sensitivity of the three cell lines and makes some suggestions for the selection of in vitro phototoxicity testing cell lines.

This study involved in vitro experiments only and is therefore exempt from ethical review.

CPZ (CAS NO.69–09–0), Amiodar (CAS NO. 19774–82–4), L-His (CAS NO. 71–00–1), neutral red (NR; CAS NO. 553–24–2), and dimethyl sulfoxide (DMSO) were purchased from Sigma Chemical Co. (Sigma-Aldrich). Dulbecco's modified Eagle medium (DMEM), fetal bovine serum (FBS), 0.25% trypsin-EDTA, Earle balanced salt solution (EBSS), and phosphate-buffered saline (PBS) were purchased from Gibco (Thermo Fisher). Fibroblast basal medium (FBM), the fibroblast growth kit–low serum, trypsin-EDTA for primary cells, and trypsin-neutralizing solution were purchased from American Type Culture Collection (ATCC).

The UV light source for phototoxicity testing was an SOL 500 solar simulator (Dr. Hönle UV Technology) with a spectral range (295–3000 nm) corresponding to natural sunlight. The simulator was equipped with an H1 filter transmitting wavelengths of 315–380 nm. The UVA output before each experiment was measured by a UVA meter (Dr. Hönle UV Technology).

CPZ and Amiodar have limited solubility in water; thus, DMSO was selected as the solvent based on OECD Test No. 432 recommendations. The DMSO was then diluted with EBSS to reach a final concentration of 0.5% (vol/vol) in medium. L-His was directly dissolved in EBSS and diluted to the testing concentration. All test materials were freshly prepared before use.

The absorption spectra of all tested compounds were obtained after dissolving them in the appropriate solvent at a concentration of 100 μg/mL. The absorbance was evaluated in the 250–700 nm range by a Multiskan SkyHigh Microplate Reader (Thermo Scientific). Blanks with each solvent were performed to eliminate interferences.

The BALB/c 3T3 fibroblast cell lines were purchased from the Shanghai Cell Bank of the Chinese Academy of Sciences, and the immortalized human keratinocyte (HaCaT) cell lines were purchased from China Center for Type Culture Collection. BALB/c 3T3 and HaCaT cells were cultured in DMEM supplemented with FBS (10%; vol/vol) in a humidified atmosphere with CO₂ (5%; vol/vol) at 37°C. Normal human primary dermal fibroblasts (HDFas) were purchased from ATCC (PCS-201-012) and cultured in FBM and using the Fibroblast Growth Kit–Low Serum according to the manufacturer's instructions.

The phototoxicities of CPZ, Amiodar, and L-His were determined according to the OECD 432 guideline with minor modifications.^[1] For the experiments, 3T3 cells were seeded into the central 60 wells of 96-well plates (Corning) at a density of 1.5 × 10⁴ cells per well. The outer wells of each plate were filled with 150 μL of PBS. After 24 hours, the cells were washed with EBSS (150 μL/well) and incubated with test solutions/controls prepared in EBSS (100 μL/well) of each test material for 1 hour in the dark at 37°C. CPZ, Amiodar, and L-His were tested up to maximum concentrations of 50 μg/mL, 100 μg/mL, and 1000 μg/mL with 10^1/3 dilutions (eight concentration ranges both with and without UVA irradiation). After incubation of two identical 96-well plates, one was exposed to UVA light (total dose, 5 J/cm²), and the other was covered in lightproof aluminum foil and incubated under a sunlight simulator. During irradiation, the 96-well plates were placed on ice-cold panels (4°C) to limit heating of the solution and thus possible thermal decomposition and excessive evaporation.

Subsequently, the cells were washed with EBSS (150 μL/well) and incubated for 24 hours under standard conditions. On the following day, cells were washed with EBSS (150 μL/well) and incubated in DMEM (100 μL/well) containing 50 μg/mL NR dye without FBS at 37°C for 3 hours. Cells were washed with EBSS (150 μL/well) and blotted to remove any remaining buffer. Specifically, 150 μL of freshly prepared desorb solution (ethanol, water, and acetic acid, mixed in a 50:49:1 ratio) was added per well, and the plate was incubated at room temperature for 10 minutes with gentle shaking. The absorbance of the resulting homogeneous pink solution was measured without a lid at 540 nm (OD₅₄₀) by a Multiskan SkyHigh Microplate Reader. The outer wells of each plate were used as references.

For the cytotoxicity and phototoxicity assays, HaCaT cells were cultured in 96-well plates at a seeding density of 2.0 × 10⁴ cells/well. The cells were cultured in the plate for 24 hours, and the following steps were the same as in the in vitro BALB/c 3T3 NRU phototoxicity test.

For the cytotoxicity and phototoxicity assays, HDFa cells were cultured in 96-well plates at a seeding density of 8.0 × 10³ cells/well. The cells were cultured in the plate for 24 hours, and the following steps were the same as in the in vitro BALB/c 3T3 NRU phototoxicity test.

Phototox version 2.0 software, available from OECD,^[20] was used to calculate PIF and MPE.

If an IC₅₀ (half-maximal inhibitory concentration) in the presence or absence of light (Irr) cannot be calculated, a PIF cannot be determined for the test material.

Test chemical results with a PIF < 2 or an MPE < 0.1 predicted “no phototoxicity.” A PIF ≥ 2 and < 5 or an MPE > 0.1 and < 0.15 predicted “equivocal” phototoxicity, and a PIF ≥ 5 or an MPE > 0.15 predicted “phototoxicity.”

The statistical analysis was performed with SPSS software (version 22.0; IBM). The graphs were generated with GraphPad Prism for Windows software (version 7.00; GraphPad Software, Inc.). Each detected material was tested three times in independent assays.

For all the test compounds, the spectra exhibited absorptions in the 300–500 nm range, which is a primary condition before conducting phototoxicity tests (Fig. 2). The absorption peaks of CPZ and Amiodar occurred at wavelengths emitted by the light source used, whereas L-His did not present an absorption peak in the range. The absorption intensity range of the UV spectrum from high to low was CPZ, Amiodar, and L-His.

A 4°C ice bath was used to ensure that the cells in the illuminated group were not affected by heating from the visible light in the solar simulator, as accumulated heat will cause a significant decrease in cell viability, and the phototoxicity of the test substance cannot be effectively determined. The comparison of cell viability using an ice bath and without an ice bath is not reported here.

Fibroblasts are larger than keratinocytes. HDFa cells have a large, long spindle structure, BALB/c 3T3 cells have a multisynaptic star structure, and HaCaT cells are smaller with a short fusiform structure. It is necessary to ensure that the cell confluence when seeded in a 96-well plate after 24 hours of incubation can reach more than 80% without stacking growth and to ensure that the mean OD₅₄₀ NRU of the solvent controls is > 0.4 according to the requirements of OECD 432. Therefore, an appropriate cell inoculation number is a prerequisite to ensure effective determination of phototoxicity. In the early stage, the optimal inoculation numbers of the three kinds of cells were evaluated. The final inoculation number of HaCaT cells (smallest cell size) was 2 × 10⁴ cells/well, the inoculation number of BALB/c 3T3 cells (medium cell size) was 1.5 × 10⁴ cells/well, and the inoculation number of HDFa cells (largest cell size) was 8 × 10³ cells/well. The cell seeding density was successively decreased, and effective NRU cell viability data were obtained under this condition.

The NRU phototoxicity test results of BALB/c 3T3 cells (Table 1) showed that CPZ had phototoxicity, Amiodar had potential phototoxicity, and L-His had no phototoxicity. The PIF values of these three substances met the reference range given by OECD 432. The MPE value was slightly lower than the range in OECD432, but the judgment results were still consistent, indicating that after the optimization of the preceding experimental conditions, the evaluation results of this study were accurate. The neutral red phototoxicity test results of HaCaT cells are shown in Table 1. Among them, the test results of CPZ indicated phototoxicity, and the PIF value was higher than the PIF measured by the BALB/c 3T3 model (28.69 ± 16.69 > 14.41 ± 1.76). Amiodar evaluated according to the PIF value was potentially phototoxic, and that evaluated according to the MPE value was nonphototoxic. The L-His test results indicated no phototoxicity. Compared with BALB/c 3T3 cells, HaCaT cells showed essentially the same ability to identify substances with strong phototoxicity and no phototoxicity, and their detection sensitivity for potentially phototoxic substances was slightly lower than that of BALB/c 3T3 cells. The comparison with human primary fibroblasts was closer to the actual situation of human skin. The NRU phototoxicity test results of HDFa cells (Table 1) revealed that CPZ was still phototoxic, whereas the Amiodar and L-His test results were nonphototoxic. There was a certain difference in the sensitivity of human primary fibroblasts and mouse fibroblast lines to phototoxicity. When the test substance was tested using BALB/c 3T3 cells, the determination result of the test substance was PIF ≥ 2 and < 5 or MPE > 0.1, and in the range of 0.1 and < 0.15, it may be judged as nonphototoxic in the HDFa phototoxicity model. As shown in these in vitro experiments, human fibroblasts are less sensitive to phototoxicity than mouse fibroblast cell lines. The phototoxicity intensities of the three kinds of cells on CPZ, Amiodar, and L-His exhibited the same trend, which was CPZ > Amiodar > L-His.

The effects of different tested substances at the same dose on cell viability and morphology were compared in the three kinds of cells (Fig. 3). The phototoxic effects of CPZ on the three kinds of cells were consistent, and 1 μg/mL CPZ completely damaged cells under UVA irradiation. Amiodar had a stronger effect on the viability of BALB/c 3T3 cells, and the cell morphology was obviously contracted. The effects of Amiodar on HaCaT and HDFA cells were similar, as the number of cells was reduced, but the original cell morphology was maintained. L-His did not induce phototoxicity in the three different cell lines.

Studies have shown that when UV rays in sunlight directly contact human skin, they will cause acute or chronic skin reactions and have a certain impact on the skin's immune response.^[21] UV light triggers skin cell apoptosis through DNA damage, death receptor activation, and reactive oxygen generation, thereby triggering phototoxicity.^[22] The development of in vitro phototoxicity assessment methods started as early as the 1980s.^[23,24] After multiple verifications, the OECD issued an in vitro phototoxicity evaluation standard in 2004.^[1] However, because of differences in sensitivity caused by differences in species, although BALB/c 3T3 cells have been regarded an accurate judgment for phototoxic substances, the assay has always been oversensitive.^[25] Therefore, researchers are constantly exploring, with the aim of developing a more effective in vitro phototoxicity assessment model.

In recent years, some reports have used HaCaT cells to explore the photodamage mechanism of pyrene derivatives under UV irradiation.^[10] HaCaT cells are recommended for in vitro phototoxicity testing.^[13] The results of this study also indicate that the evaluation results of HaCaT cells are closer to the results of the in vitro phototoxicity test of human primary fibroblasts. Of course, exploring the phototoxicity of benzophenones on 3T3 fibroblasts and HaCaT keratinocytes has also been reported.^[16] The anti-UV damage effect of strawberry extract as a cosmetic raw material in HDFs was reported in 2017.^[26] There have been reports of using other cell lines to assess phototoxicity in vitro, such as the L-929^[27] and THP-1 cells were compared with 3T3 cells in phototoxicity tests.^[28] Using a variety of cells to evaluate the phototoxicity of target substances can obtain more comprehensive evaluation results. Moreover, determining the simultaneous phototoxicity of fibroblasts and keratinocytes can more closely reflect the actual situation of the human body.

According to the results obtained, we conclude that the sensitivity of mouse fibroblasts and human fibroblasts to the determination of phototoxic substances is different. The sensitivity of human fibroblasts is lower than that of mouse fibroblasts, and the phototoxicity results of HaCaT cells are closer to those of human fibroblast cells and can also effectively identify strongly phototoxic substances. It is recommended to use HaCaT cells for in vitro phototoxicity determinations of substances. In the future, further research on phototoxicity models will be carried out in conjunction with 3D skin models.