2025 Volume 50 Issue 1 Pages 1-9
The human Cell Line Activation Test (h-CLAT) is an in vitro skin sensitization assay adopted by the OECD as Test Guideline 442E. In the h-CLAT, 2,4-dinitrochlorobenzene (DNCB) is used as a positive control; however, DNCB is considered a poisonous substance under the Poisonous and Deleterious Substances Control Act in Japan since 2014 because of its high acute toxicity. Strict control, handling, and storage are required when using DNCB, which is a burden to the users. Although the use of other suitable positive controls with historical data is accepted by the guideline, to our knowledge, there have been no reports of such substances. Therefore, in this study, we investigated suitable positive controls that can be used in addition to DNCB for the h-CLAT. Three candidates that are not considered poisonous substances, imidazolidinyl urea, hydroxycitronellal, and 2,4-dinitrofluorobenzene, were selected. To determine their suitability as positive controls, the h-CLAT was performed repeatedly for each chemical in two laboratories. For imidazolidinyl urea, the results that failed the positive control criteria were observed in both laboratories, indicating that it was inconclusive for the suitability as a positive control at the tested concentration. In contrast, all experiments with hydroxycitronellal and 2,4-dinitrofluorobenzene met the criteria and resulted in relative fluorescence intensity values of CD86/CD54, which were comparable to those for DNCB. Based on these results, hydroxycitronellal and 2,4-dinitrofluorobenzene may be used as positive controls. This study would provide valuable information for users examining other suitable positive controls in the h-CLAT.
Skin sensitization assessment of ingredients is mandatory for consumer health because repeated topical exposure to skin sensitizers can lead to allergic contact dermatitis, a common inflammatory skin disease (Martin, 2015). Skin sensitization was typically assessed using traditional animal tests, such as the guinea pig maximization test (OECD, 2022) and the murine local lymph node assay (LLNA) (OECD, 2010). However, because of regulatory and ethical demands, new approach methodologies (NAMs), such as in chemico and in vitro methods and in silico predictions have been developed for skin sensitization assessments (Gądarowska et al., 2022; Pistollato et al., 2021; Barthe et al., 2021; Silva and Tamburic, 2022).
Various in chemico and in vitro test methods that address three Key Events (KE) of the skin sensitization Adverse Outcome Pathway have been validated and adopted by the OECD (Gądarowska et al., 2022). For example, the human Cell Line Activation Test (h-CLAT) was adopted by the OECD under Test Guideline No. 442E (TG442E) as one of the in vitro methods that address KE3, the activation of dendritic cells, and quantifies changes in CD86 and CD54 expression on THP-1 cells following exposure to test chemicals (OECD, 2023). The h-CLAT is also included in defined approaches (DAs) combining NAMs, such as 2 out of 3 and integrated testing strategy DAs adopted as OECD Guideline No. 497 (OECD, 2021) and can be used in a next generation risk assessment framework for skin sensitization (Gilmour et al., 2020; Gilmour et al., 2023).
In the h-CLAT, 2,4-dinitrochlorobenzene (DNCB) with high stability and a stable response is used as a positive control (OECD, 2023; Sakaguchi et al., 2010). It is also widely used in human clinical and experimental studies (Newell et al., 2013; Hopkins et al., 2005). However, DNCB is considered a poisonous substance under the Poisonous and Deleterious Substances Control Act in Japan since 2014 (MHLW, 2014) because of its high acute dermal toxicity (NIHS, 2013). Strict control, handling, and storage (e.g., a distinct storage area dedicated to poisonous substances, assignment of a person responsible for handling poisonous substances, and monitoring the stock quantity of poisonous substances) are required for the use of DNCB (MHLW, 2018), which has become a burden to its users. While DNCB is regulated in other countries, such as under the Toxic Substances Control Act, strict control, handling, and storage requirements seen in Japan do not appear to apply elsewhere. However, due to its toxicity, similar regulations could potentially be implemented in other countries in the future.
The use of other suitable positive controls with historical data to derive comparable run acceptance criteria is acceptable in Test Guideline No. 442E (OECD, 2023); however, to the best of our knowledge, there have not been any reports on other suitable positive controls for the h-CLAT.
Although using DNCB as a positive control for h-CLAT is a burden to its users, no studies have been reported regarding suitable alternatives based on our understanding. To address the issue, we evaluated other potential positive controls that are not considered poisonous under the law. Three candidates were selected based on previous reports. To examine the suitability of these candidates as positive controls, the h-CLAT for each was performed repeatedly in two separate laboratories.
The h-CLAT was performed according to OECD Test Guideline No. 442E (OECD, 2023) in two laboratories, including Kao Corporation and Shiseido Co., Ltd. Briefly, the cells were treated with a test chemical for 24 hr and stained with FITC-labeled anti-CD86, anti-CD54, or mouse IgG1 isotype control antibodies. Dead cells were labeled with propidium iodide. Changes in CD86 and CD54 expression were measured by flow cytometry and the relative fluorescence intensity (RFI) of CD86 and CD54 compared with the solvent/vehicle control was calculated for the prediction model. An RFI value was obtained for each chemical per run. If the RFI of CD86 was ≥150% and/or the RFI of CD54 was ≥200% (with cell viability ≥50%), the test chemical was considered positive. If at least two independent runs are positive in two or three independent runs, the final prediction is positive according to OECD Test Guideline No. 442E. However, this study used information on prediction in the specified concentrations per run rather than the final prediction. The estimated concentration of the positive chemical that induces 150% RFI in CD86 expression (EC150) and/or the estimated concentration inducing 200% RFI in CD54 expression (EC200) were calculated using the following formula: EC150 or EC200 = Bconcentration + [(150 or 200 – BRFI)/(ARFI – BRFI) × (Aconcentration – Bconcentration)], where Aconcentration is the lowest concentration with an RFI >150 or 200, Bconcentration is the highest concentration with an RFI <150 or 200, ARFI is the RFI at the lowest concentration with an RFI >150 or 200, and BRFI is the RFI at the highest concentration with an RFI <150 or 200.
ChemicalsImidazolidinyl urea, hydroxycitronellal, and 2,4-dinitrofluorobenzene (DNFB) were selected as candidate positive controls. All three chemicals were purchased from Sigma-Aldrich (St. Louis, MO, USA). The detailed information, including catalog numbers and purity, is listed in Table 1.
;%0A%09%09%09newWindow.document.open();%0A%09%09%09newWindow.document.write('<img src=%22./Graphics/50_1-t1.jpg%22>');%0A%09%09%09newWindow.document.close();%0A%09%09)
A positive control is defined in the guideline as “a replicate containing all components of a test system and treated with a substance known to induce a positive response. To ensure that variability in the positive control response across time can be assessed, the magnitude of the positive response should not be excessive” (OECD, 2023). In addition, acceptance criteria in the h-CLAT regarding the positive control (DNCB) are “RFI values of both CD86 and CD54 should meet the positive criteria (CD86 RFI ≥ 150 and CD54 RFI ≥ 200) and cell viability should be more than 50%” (OECD, 2023). Although there are a few reports on the selection of positive controls for the development of in vitro assays (OECD, 2018; Petersen et al., 2021), there have been no reports, to our knowledge, on the detailed characteristics considered for the selection of a positive control for in vitro assays. The important characteristics of a positive control may be different depending on the assay (Petersen et al., 2021); however, 10 characteristics were mentioned, including the biological mechanism of action, ease of preparation, chemical purity, verifiable physical properties, stability, ability to generate responses spanning the dynamic range of the assay, technical and biological interference, commercial availability, user toxicity, and disposability. These represent goals to consider rather than absolute requirements. This study established detailed characteristics considered for the selection of candidates as positive controls for the h-CLAT. The following primary characteristics were considered: (1) not regarded as a poisonous substance under the law in Japan, (2) positive both in CD86 and CD54 with cell viability >50%, and (3) comparable RFI values with DNCB. Items 2 and 3 are desired to be met at a single concentration between laboratories with high reproducibility. Based on these characteristics, three chemicals, imidazolidinyl urea, hydroxycitronellal, and DNFB, were selected as candidates. The detailed information including historical LLNA EC3 values (the estimated concentration that produces a stimulation index of 3 in LLNA) (OECD, 2021; Roberts et al., 2011) and h-CLAT data (OECD, 2021; Urbisch et al., 2015) on these three chemicals and DNCB is listed in Table 2. All three chemicals are not considered poisonous under Japanese law. To our knowledge, these chemicals have no data that meet the criteria of a poisonous substance, such as high acute toxicity data. Imidazolidinyl urea and hydroxycitronellal were selected from proficiency substances (6 sensitizers) in the h-CLAT (OECD, 2023) and chemicals in a validation study report (ECVAM, 2013a) and a ring trial (Sakaguchi et al., 2010) for the h-CLAT (16 and 14 sensitizers, respectively). Imidazolidinyl urea is a proficiency chemical showing a positive result for both CD86 and CD54 in the h-CLAT with RFI values similar to that of DNCB compared with nickel sulfate, another proficiency chemical exhibiting positive CD86 and CD54 (OECD, 2023; ECVAM, 2013b; Sakaguchi et al., 2010). For hydroxycitronellal, positive results for both CD86 and CD54 were observed in all experiments in all five laboratories in the 1st ring trial for h-CLAT (Sakaguchi et al., 2010). In addition, DNFB with a similar structure, skin sensitization potency (EC3), and h-CLAT data (Urbisch et al., 2015) to DNCB was selected. Based on these findings, we hypothesized that these three chemicals are reproducibly positive for CD86 and CD54 and yield comparable RFI values with DNCB at a single concentration between laboratories (i.e., these three chemicals are suitable as a positive control in the h-CLAT). To test this hypothesis, the h-CLAT was performed for each candidate in two laboratories, and RFI values for each candidate were compared with those of DNCB.
;%0A%09%09%09newWindow.document.open();%0A%09%09%09newWindow.document.write('<img src=%22./Graphics/50_1-t2.jpg%22>');%0A%09%09%09newWindow.document.close();%0A%09%09)
This study performed basically 10 runs, as a study that provides a useful reference for users examining other suitable positive controls in the h-CLAT, for each candidate. The suitability as a positive control was examined based on the following criteria: (1) the acceptance criteria for the positive control were met in all runs, and (2) the RFI values of CD86/CD54 for each candidate were almost within the variation in RFI values for DNCB obtained during the study (i.e., comparable RFI values to DNCB). A candidate that did not meet these criteria was considered inconclusive regarding its suitability in this study.
To examine the suitability of imidazolidinyl urea, hydroxycitronellal, and DNFB as positive controls (i.e., whether these three chemicals show positive both in CD86 and CD54 and comparable RFI values to DNCB at a single dose between laboratories with high reproducibility) and determine the appropriate dose as a positive control in the h-CLAT, the assay was performed for each candidate in two laboratories (Labs A and B). Note that the single appropriate dose is not necessarily the same between laboratories and may be determined in each laboratory as described in the guideline (OECD, 2023). The results are shown in Fig. 1.
Overall h-CLAT data in the examination of the suitability of three candidates as a positive control. h-CLAT was performed for DNCB (n=29-33) (A), Imidazolidinyl urea (n = 8-15) (B), hydroxycitronellal (n = 10) (C), and DNFB (n = 10) (D) in Labs A (green circles) and B (orange circles). An RFI value was obtained for each chemical per run. Ten runs were basically performed for each candidate as described in the Materials and Methods, except for imidazolidinyl urea, where the number of runs varied based on each laboratory’s decision due to some runs failing the acceptance criteria for positive control. Circles indicate values of CD86/CD54 RFI and cell viability from each experiment. Solid lines show the positive criteria of CD86/CD54 (150% for CD86 and 200% for CD54) and cell viability (50%) in the h-CLAT. Dashed lines indicate the variation in RFI values for DNCB to compare RFI values for CD86/CD54 between each candidate and DNCB. Gray and open circles represent results that failed the acceptance criteria for the positive control because of the low RFI value and cell viability, respectively.
DNCB was tested at a concentration of 4 μg/mL alongside the candidates as a positive control. The variation in RFI values was compared between DNCB and each candidate to determine if the candidates produced RFI values comparable to DNCB at the tested concentration. The RFI values for DNCB obtained in this study ranged from 171% to 698% for CD86 and from 234% to 1241% for CD54 (Fig. 1A), which fall within historical RFI ranges for DNCB reported from four laboratories in the validation study (ECVAM, 2013c).
Performance of h-CLAT for imidazolidinyl ureaTo determine the test concentration of imidazolidinyl urea, a dose finding assay was newly performed using eight doses in each laboratory (Fig. 2), because imidazolidinyl urea has a relatively narrow concentration range for positive responses compared to DNCB, hydroxycitronellal and DNFB (Supplementary Fig. 1). The eight doses were based on the CV75 (concentration yielding 75% cell viability, which is used for dose setting (1.2 × CV75, 1 × CV75, 1/1.2 × CV75, 1/1.22 × CV75, 1/1.23 × CV75, 1/1.24 × CV75, 1/1.25 × CV75, 1/1.26 × CV75) in the h-CLAT as specified in OECD Test Guideline No. 442E (OECD, 2023)) or results in the validation study (ECVAM, 2013b). The dose ranges that produced positive results in the three experiments were confirmed by a dose finding assay in each laboratory, and two test concentrations near those doses were selected to identify a more suitable dose that meets the criteria for suitability as a positive control. Positive results were obtained in the three experiments at approximately 35 μg/mL and 45 μg/mL in Labs A and B, respectively. Based on the results from a dose finding assay, the test concentrations for imidazolidinyl urea were set at 32.5 and 35 μg/mL in Lab A, and 45 and 50 μg/mL in Lab B. By performing the h-CLAT repeatedly at the test concentrations in each laboratory, each test concentration failed the acceptance criteria for positive control (Fig. 1B). In Lab A, 3 of 8 experiments at 32.5 μg/mL were negative for CD86 and/or CD54 and 2 of 11 experiments at 35 μg/mL were negative for CD86 and/or CD54. In Lab B, 1 of 15 experiments at 45 μg/mL was negative for CD54, and 2 of 15 experiments at 50 μg/mL did not yield cell viability >50%. According to the suitability criteria outlined in the Materials and Methods, these limited results in this study did not conclude that imidazolidinyl urea at the tested concentrations is suitable as a positive control for the h-CLAT.
Dose-response curves of CD86/CD54 RFI values and cell viability in the h-CLAT for imidazolidinyl urea. To determine the test concentrations for evaluating imidazolidinyl urea as a positive control in the h-CLAT, three experiments (blue, pink, and green lines) with eight test doses were performed in each laboratory. Circles and diamonds indicate RFI values of CD86/CD54 and cell viability, respectively. Solid and dashed lines show the positive criteria of CD86/CD54 (150% for CD86 and 200% for CD54) and cell viability (50%), respectively.
The test concentrations for hydroxycitronellal were determined as 50, 100, and 200 μg/mL in both Labs A and B, because historical data (Nukada et al., 2012) showed positive results at a relatively wide range of concentration, approximately 50–200 μg/mL (Supplementary Fig. 1). All of the experiments met the acceptance criteria and showed RFI values of CD86/CD54 comparable with those for 4 μg/mL of DNCB, particularly at 50 μg/mL of hydroxycitronellal (183%–542% for CD86 and 215%–853% for CD54) (Fig. 1A, C). Based on the criteria for suitability as a positive control in this study, hydroxycitronellal is considered a suitable positive control for the h-CLAT.
Performance of h-CLAT for DNFBDNFB has a similar structure, skin sensitization potency (EC3), and h-CLAT data to DNCB (Table 2). Therefore, 4 μg/mL, a typical concentration of DNCB as a positive control in the h-CLAT, was used as the test concentration for DNFB in Labs A and B. All experiments met the acceptance criteria for the positive control and showed RFI values of CD86/CD54 (233%–668% for CD86 and 364%–939% for CD54) comparable with those for 4 μg/mL of DNCB (Fig. 1A, D). Based on the criteria for suitability as a positive control outlined in this study, these results indicate that DNFB is also considered suitable as a positive control for the h-CLAT.
To examine other suitable positive controls in addition to DNCB for the h-CLAT, we selected three candidates and showed that hydroxycitronellal and DNFB were positive in all experiments with RFI values of CD86/CD54 comparable with those of DNCB at a single concentration in two laboratories (Fig. 1). These results indicate that hydroxycitronellal and DNFB may be used as positive controls in the h-CLAT. It will be helpful for users to incorporate hydroxycitronellal or DNFB as positive controls while maintaining the historical data in the h-CLAT. To the best of our knowledge, this is the first report on other suitable positive controls in the h-CLAT.
Although DNCB is a suitable positive control in the h-CLAT, it has been considered a poisonous substance under the Poisonous and Deleterious Substances Control Act in Japan since 2014 (MHLW, 2014). Strict control, handling, and storage are required for the use of DNCB (MHLW, 2018). The use of other suitable positive controls with historical data to derive comparable run acceptance criteria is accepted by OECD Test Guideline No. 442E (OECD, 2023); however, there have been no reports, to our knowledge, on other suitable positive controls in the h-CLAT. In this study, by selecting candidates of other suitable positive controls and generating h-CLAT data for these candidates, it was suggested that hydroxycitronellal and DNFB may be used as a positive control. The use of hydroxycitronellal or DNFB would be helpful for users because they are not regarded as poisonous substances under Japanese law, which does not require the strict control, handling, and storage procedures of DNCB. Given the potential for DNCB to face similar regulations in other countries due to its high toxicity, this study may also provide valuable insights for those regions.
The results of this study indicate that among the three candidates, hydroxycitronellal and DNFB may be used as positive controls in the h-CLAT, whereas imidazolidinyl urea was not considered suitable based on the study’s criteria because the results failed the acceptance criteria at each test concentration in Labs A and B. Imidazolidinyl urea showed a relatively steep dose-response curve for cell viability and RFI values of CD86/CD54 (ECVAM, 2013b) (Supplementary Fig. 1 and Fig. 2). Therefore, because of the steep response, the cell viability and RFI values of imidazolidinyl urea at each test concentration were not reproducible between the experiments compared with hydroxycitronellal, DNFB, and DNCB (Fig. 1). It is important to note that despite the data presented here, imidazolidinyl urea has been predicted as positive with high reproducibility using eight test doses according to the h-CLAT prediction model in OECD Test Guideline No. 442E, and it serves as a useful proficiency chemical in the h-CLAT (OECD, 2023; ECVAM, 2013b).
This study offered increased flexibility in the selection of positive controls used in the h-CLAT. While both hydroxycitronellal and DNFB are considered suitable positive controls in addition to DNCB, there are some differences between these chemicals. Hydroxycitronellal is an aldehyde and is unstable under forced degradation conditions, such as exposure to air and light, which results in oxidization (Avonto et al., 2018; Avonto et al., 2019). Consistent with these findings, the reduction of responses (RFI values) for hydroxycitronellal in the h-CLAT was observed at approximately six months after the repeated use of a stock of hydroxycitronellal during testing (data not shown). Thus, proper storage and/or the use of fresh samples is necessary when using hydroxycitronellal as a positive control in the h-CLAT. In contrast, DNFB exhibits high toxicity (e.g., a powerful irritant and strong contact allergen) (NLM, 2013), although it is not considered a poisonous substance under Japanese law and there is no data for DNFB that meets the criteria of a poisonous substance such as high acute dermal toxicity as DNCB (NIHS, 2013). Considering the 10 characteristics of a positive control (the biological mechanism of action, ease of preparation, chemical purity, verifiable physical properties, stability, ability to generate responses spanning the dynamic range of the assay, technical and biological interference, commercial availability, user toxicity, and disposability) described by Petersen et al. (Petersen et al., 2021), hydroxycitronellal may be good from the standpoint of user toxicity, whereas DNFB may be good with respect to stability. When selecting a positive control in the h-CLAT, these chemical characteristics should be considered by each user.
Despite the suggestion of additional positive controls in the h-CLAT, the data presented in the present study were obtained from only two laboratories and at limited test concentrations, number of experiments, and time as a study that provides a useful reference for users examining other suitable positive controls in the h-CLAT. Therefore, when hydroxycitronellal or DNFB is used as a positive control in the h-CLAT, users should determine an appropriate positive control concentration in each laboratory and monitor the results of the positive control over time to assess its performance and variability in the positive control response.
In conclusion, the results of this study suggest that hydroxycitronellal and DNFB are considered suitable positive controls in the h-CLAT in addition to DNCB. The use of these chemicals in the h-CLAT while maintaining the historical data over time will be helpful to users because hydroxycitronellal and DNFB are not considered poisonous substances under Japanese law. Thus, they do not require strict control, handling, and storage like DNCB. This study would provide valuable information for users examining other suitable positive controls in the h-CLAT.
Conflict of interestThe authors declare that there is no conflict of interest.
