The Journal of Toxicological Sciences
Online ISSN : 1880-3989
Print ISSN : 0388-1350
Assessment of the skin sensitizing potential of chemicals, contained in foods and/or cosmetic ingredients, using a modified local lymph node assay with an elicitation phase (LLNA:DAE) method
Kunihiko YamashitaHiroshi MiyazakiShinsuke ShinodaSaori HagiwaraHirohito TakahashiHiroshi Itagaki
Author information

2018 Volume 43 Issue 8 Pages 513-520


We evaluated the skin sensitizing potential of 10 natural organic chemicals, or their derivatives, which are included in foods and/or skin products, using a modified local lymph node assay (LLNA), with an elicitation phase (LLNA:DAE). The following compounds were tested: carminic acid, esculetin, 4-methyl esculetin, coumarin, quercetin, curcumin, naringenin, chlorogenic acid, isoscopoletin, and shikonin. Esculetin, 4-methyl esculetin, isoscopoletin, and shikonin yielded positive results. In particular, shikonin at a very low concentration (0.05%) induced an elicitation response. In conclusion, four of the 10 natural organic chemicals tested had a skin sensitization potential, with shikonin producing serious reaction even at a very low concentration.


Currently, natural organic materials are used in a wide range of nutritive supplement food and/or cosmetics due to various biological activities. As an example, chlorogenic acid is used for its anti-oxidative effects, quercetin for its anti-inflammatory properties and shikonin, the active component of shikon, for its multiple biological activities (Hwang et al., 2014; Li et al., 2016). The wide use of these chemicals increases the likelihood of exposure, either through ingestion or skin contact. The issue of exposure is important to consider based on previous research that has shown the relationship between prolonged exposure to flavoring agents such as cinnammon, via ingestion, and the development of intraoral allergic dermatitis or stomatitis (Torgerson et al., 2007; Feller et al., 2017; Minciullo et al., 2016). Therefore, there is a clinical need to evaluate the skin sensitizing potential of chemicals used in foods and cosmetics.

Several in new in-vivo skin sensitization tests have been developed which have replaced the guinea pig sensitization test (GPMT; Magnusson and Kligman, 1969), achieving sufficient sensitivity without the use of adjuvants and in a shorter test period, while reducing the use of animals. These new tests include the LLNA, described in the Organization for Economic Co-operation and Development (OECD) test guideline (TG) 429 (Dean et al., 2001; Interagency Coordinating Committee on the Validation of Alternative Methods (ICCVAM), 2010; OECD, 2002, 2010b), the LLNA:DA, OECD TG442A (Yamashita et al., 2005; Idehara et al., 2008; Omori et al., 2008; OECD, 2010a) and the LLNA:BrdU-ELISA, OECD TG442B (Takeyoshi et al., 2001; Takeyoshi et al., 2003; OECD, 2010c). Although these LLNAs have the potential to provide an efficient evaluation of the sensitizing potential of chemicals, some issues remain to be resolved. Specifically, the surfactant used in LLNAs, which is a skin irritant, can lead to false positive results via a non-specific lymph node proliferation effect (Basketter et al., 2009). In addition, as these new tests do not include the elicitation process of the sensitizing response, they are not suitable for assessment of the cross-sensitization potential of chemicals. To resolve these limitations, we developed the LLNA:DAE method (Yamashita et al., 2014; 2015a, 2015b; 2016a, 2016b). The LLNA:DAE method provides several advantages over LLNA methods, including being of lower cost (due to the lower number of animal tests used) and not requiring the use of radio-isotopic substances. Specifically with regard to animal use, LLNA testing usually requires 5 animal groups, for each assay, with 4 animals per group. By comparison, LLNA:DAE uses 2 groups of 5 animals (10 versus 20 animals needed for LLNA).

Moreover, LLNA:DAE measures the skin sensitizing potential of a chemical using lymph node weight and, therefore, does not require special experimental equipment, such as a liquid scintillation counter. The experimental period of LLNA:DAE is, however, longer than for the original LLNA as it evaluates the elicitation phase of skin sensitizing, whereas LLNA evaluates only the induction phase of skin sensitization. Of note is the potential of LLNA:DAE testing to discriminate true skin sensitization from skin irritation, as well as to evaluate the cross-sensitizing potential between chemicals. Thus, the LLNA:DAE method might provide a useful alternative to GPMT. Therefore, our aim in this study was to evaluate the skin sensitizing potential of natural organic chemicals, and their derivatives, which are contained in foods and/or cosmetics using the LLNA:DAE method, and to compare our results to other well-known skin sensitizing chemicals.



The following 10 chemicals were tested in our study, with relevant information summarized in Table 1: carminic acid, esculetin, 4-Methyl esculetin (purchased from Tokyo Kasei, Tokyo Chemical Industry Co. Ltd., Tokyo, Japan); coumarin, quercetin, curcumin, naringenin, chlorogenic acid acetone and dimethyl sulfoxide (DMSO) (purchased from Wako, Wako Pure Chemical Industries, Ltd., Osaka, Japan); isoscopoletin (purchased from Funakoshi, Funakoshi Co. Ltd., Tokyo, Japan); and shikonin (purchased from Kishida, Kishida Chemical Co. Ltd., Osaka, Japan). Olive oil used for testing was purchased from Yoshida Pharmaceutical Co. Ltd., Japan.

Table 1. Chemicals, structure, sources, purity, and vehicle for primary LLNA:DAE.


Female CBA/J mice (Charles River Japan Inc., Kanagawa, Japan) were used for all experiments. All animals were 8 weeks old at the time of purchase, with the LLNA:DAE studies initiated at 9 weeks of age. The animal holding rooms were maintained at constant temperature (23 ± 2ºC) and humidity (50 ± 15%), with the rooms ventilated at a rate of 12 cycles per hour. All procedures were performed according to the institutional guidelines for animal care of the Drug Safety Testing Center Co., Ltd. Guidelines for the ethical use of animals in research.

Evaluation of the skin sensitization potential of chemicals

Dose finding tests

The following concentrations of the 10 different chemicals were applied to the dorsum of the right ear in each animal on days 1, 2 and 3, and subsequently applied to both ears on day 10: 25.0%, 12.5%, 6.25%, and 3.13% concentrations of curcumin, naringenin and chlorogenic acid, respectively, diluted in AOO; 25.0%, 12.5%, 6.25%, and 3.13% concentrations of carminic acid and quercetin, respectively, diluted in DMSO; 5%, 2.5%, 1.3%, and 0.63% concentration of shikonin diluted in AOO; and 5%, 2.5%, 1.3%, and 0.63% of coumarin, esculetin, 4-methyl esculetin, and isoscopoletin, respectively, diluted in DMSO. The highest concentrations were defined by either the solubility test or based on previous reports (Masamoto, 2001). One mouse was used for each chemical. The ear erythema scores were measured from day 1 through day 12. Ear thickness was measured on days 1, 3, 5, 10, and 12. Body weight was recorded on days 1 and 12. The ear lymph node was excised and weighed on day 12.

Primary LLNA:DAE study

The concentration for each chemical used for the LLNA:DAE tests, selected based on the dose finding tests, were as follows: 25.0% of curcumin, naringenin and chlorogenic acid, dissolved in AOO; 25.0% of quercetin and carminic acid, dissolved in DMSO; 5% of coumarin, esculetin and 4-methyl esculetin, dissolved in DMSO; 10% isoscopoletin, dissolved in DMSO; and 0.05% of shikonin, dissolved in DMSO. Again, only one chemical was tested in each mouse, with the chemical applied to the dorsum of the right ear on days 1, 2 and 3, and subsequently re-applied to the dorsum of both ears on day 10. Each chemical was tested in 5 mice. As a control, the chemical was applied to the dorsum of the left ear of 5 mice (for each chemical) on day 10. The lymph nodes of each ear were excised and weighed on day 12.

Estimation of skin sensitizing potency

To estimate the skin sensitizing potency of tested chemicals, we calculated the increase in lymph node weight of the left ear per 1% dose (1% increase value) using the following formula:

1% increase value = (Left ear lymph node weight from test group -Left ear lymph node weight from control group) / testing dose (%)

Statistical analysis

A chemical was considered to have a positive skin sensitizing potential if the lower limit of the 95% confidence interval (95%CI) was greater than 0. The 95%CI was calculated for the mean difference between the average weights of the lymph nodes of the left ears in the control group versus the test groups.


Evaluation of the skin sensitization potential and potency of chemicals

Dose finding tests

In all tested conditions, except for treatment with shikonin, an increase in lymph node weight (LNW) of the left ear > 4 mg was not observed. Coumarin, esculetin, 4-methyl esculetin, carminic acid, isoscopoletin, and shikonin increased the LNW of the right ear > 4 mg at, at least one of the test concentrations. These results are indicative of an immunological and/or inflammatory lymph node proliferation. In contrast, the application of quercetin, curcumin, naringenin, and chlorogenic did not produce a remarkable increase in the LNW of either ear. The dose finding tests did not reveal excess skin erythema and systemic toxicity, except with the application of shikonin, which produced a severe erythema at all tested doses. Based on these results, we selected the following concentrations for the LLNA:DAE study: 25% quercetin, curcumin, naringenin and chlorogenic, 10% of isoscopoletin, 5% of coumarin, esculetin, 4-methyl esculetin, carminic acid and 0.05% shikonin

Primary LLNA:DAE study

The results for the LLNA:DAE studies are shown in Fig. 1. Positive skin sensitization results were observed for esculetin, 4-methyl esculetin, isoscopoletin, and shikonin. The other six chemicals (coumarin, carminic acid, quercetin, naringenin, chlorogenic acid, and curcumin) did not induce an elicitation response.

Fig. 1

The results of LLNA:DAE for 10 natural organic chemicals. The black and white columns indicate the values for the right and left ear, respectively. The Y-axis indicates the lymph node weight (LNW). Values are expressed as the mean ± standard deviation (S.D., n = 5). * represents a statistically significant difference in the average lymph node weights for the left ear between the control and test group.

Skin sensitizing potency

The calculated 1% increase in values of the 10 tested chemicals and our previous data are summarized in Table 2. Shikonin was classified as an extreme skin sensitizing chemical, with esculetin and 4-methyl esculetin classified as strong skin sensitizing chemicals and isoscopoletin as a weak-to-moderate skin sensitizing chemical.

Table 2. Skin-sensitizing potency of 40 chemicals.
Chemical T.D.
T.G left
C.C left
% increase
EC3 b
Category b
Shikonin 0.05a 4.7 a 3.1 a 1.6 31.60
2,4-Dinitrofluorobenzene 0.25 a 7.3 a 4.2 a 3.1 12.56 0.03 Extreme
2,4-Dinitrochlorobenzene 0.25 a 5.1 a 2.7 a 2.4 9.44 0.05 Extreme
p-Benzoquinone 1 a 5.3 a 3.3 a 2.0 2.00 0.0099 Extreme
Glutaraldehyde 1.56 a 4.1 a 2.4 a 1.7 1.09 0.1 Strong
Hydroquinone 1 a 3.3 a 2.3 a 1.0 1.04 0.11 Strong
Cobalt chloride 5 a 5.2 a 2.7 a 2.5 0.49 0.6 Strong
p-Phenylenediamine 2.5 a 3.3 a 2.4 a 0.9 0.37 0.16 Strong
4-Methylesculetin 5 3.7 2.2 1.5 0.30
Propyl gallate 6.25 a 3.9 a 2.2 a 1.7 0.27 0.32 Strong
Esculetin 5 3.5 2.3 1.2 0.24
Cinnamic aldehyde 12.5 a 4.6 a 2.2 a 2.4 0.19 3.0 Moderate
Isoeugenol 10 a 3.7 a 2.5 a 1.2 0.12 1.2 Modelate
Citral 25 a 5.1 a 2.7 a 2.5 0.10 13 Weak
Abietic acid 25 a 4.4 a 2.4 a 2.0 0.08 15 Weak
2-Mercaptobenzothiazole 25 a 4.5 a 2.8 a 1.7 0.07 1.7 Modelate
Isospopoletin 10 3.1 2.3 0.7 0.07
Cinnamic alcohol 25 a 3.5 a 1.9 a 1.5 0.06 21 Weak
Nickel sulphate 25 a 3.6 a 2.3 a 1.3 0.05 4.8 Moderate
Resorcinol 25 a 3.2 a 2.2 a 1.0 0.04 5.7 Moderate
Hexylcinnamic aldehyde 25 a 4.0 a 2.9 a 1.1 0.04 11 Weak
Eugenol 50 a 3.8 a 2.0 a 1.8 0.04 13 Weak
Phenyl benzoate 10 a 3.0 a 2.1 a 0.8 0.08 17 Weak
Hydroxycitronellal 50 a 4.1 a 2.5 a 1.6 0.03 33 Weak
Imidazolidinyl urea 40 a 3.3 a 2.2 a 1.2 0.03 24 Weak
Methyl methacrylate 100 a 4.4 a 2.3 a 2.0 0.02 90 Weak
Ethylene glycol dimethacrylate 40 a 3.1 a 2.3 a 0.8 0.02 28 Weak
Curcumin 25 2.2 2.2 0.1 0.00
Coumarin 5 2.7 2.9 -0.2 -0.04 - N.C.
Carminic acid 25 2.3 2.4 -0.1 0.00
Quercetin 25 2.6 2.7 0.0 0.00
Naringenin 25 2.2 2.3 -0.1 -0.01
Chlorogenic acid 25 2.5 2.6 -0.1 0.00
Sulfanilamide 50 a 2.0 a 2.0 a 0.0 0.00 - N.C.
Glycerol 100 a 2.1 a 2.1 a 0.0 0.00 - N.C.
Isopropanol 50 a 2.3 a 2.3 a 0.0 0.00 - N.C.
Lactic acid 25 a 2.0 a 2.2 a -0.3 -0.01 - N.C.
Methyl salicylate 50 a 2.5 a 2.3 a 0.2 0.00 - N.C.
Salicylic acid 50 a 2.2 a 2.2 a 0.0 0.00 - N.C.
n-Hexane 100 a 3.0 a 2.7 a 0.3 0.00 - N.C.

Abbreviations: N.C.: not classified; -: no; not calculated; T.D: Testing dose; T.G left: Testing Group of left ear; C.G left: Control Group of left ear; Chemical names written in bold were tested in this study. aThe data of lymph node weight were taken from Yamashita et al. (2015a). bThe EC3 value and potency category were taken from Nukada et al. (2013).


Several concentrated bio-active chemicals, such as plant-derived additives, are used as supplemental ingredients in food and also in cosmetics. Tremblay and Avon (2008) reported on an association between the ingestion of cinnamalaldehyde-containing foods and the development of allergic contact stomatitis, highlighting the importance of evaluating the skin sensitization potential of chemicals that are used in food and cosmetic production, even if those chemicals are taken from food traditionaly.

Our LLNA:DAE results are compared to previously reported values in Table 3. Among the 10 chemicals we tested, 4 produced a positive skin sensitization response. In agreement with our findings, Nagaoka Hamano et al. (2007) reported that highly purified carminic acid, as used in this study, produced negative results when tested using the local popliteal lymph node assay (PLNA) as an antigenicity test for low molecular weight chemicals. Furthermore, Ikarashi et al. (2014) reported that although carminic acid did induce lymph node proliferation in LLNA:DA, this reaction did not reach the positive threshold criterion for skin sensitization. In this study, lymph node proliferation of the right ear, the induction site for carminic acid, was observed, but with no increase in the LNW of the left ear, the challenge site. Thus, the lymph node proliferation of right ear was judged to be non-specific proliferation by skin irritation. Highly purified coumarin was also reported to produce a negative response using LLNA (Vocanson et al., 2006, 2007). Specifically in LLNA study, when tested at concentrations of 10%, 25% and 50%, the stimulation indices were 1.9, 1.8, and 2.4, respectively, which were indicative of induced lymph node proliferation, but did not reach the cutoff criterion index of 3.0 for skin sensitization. In our LLNA:DAE study, coumarin did induce a clear lymph node proliferation in the right ear in the testing group, up to a LNW of 4.2 mg, but without inducing an elicitation response. Moreover, coumarin also induced a slight lymph node proliferation in the left ears of mice in the testing and control groups, indicating that coumarin has the potential to induce non-allergic lymph node proliferation.

Table 3. Summary of skin sensitization test results for chemicals.
Name Test method and results
P or N Potency
Coumarin N Non N a N(L) b
Esculetin P Strong P a n.d.
4-Methylesculetin P Strong P a n.d.
Isoscopoletin P Weak/Moderate P a n.d.
Carminic acid N Non n.d. N(L,P) c.d
Shikonin P Extreme n.d. n.d.
Quercetin N Non n.d. n.d.
Naringenin N Non n.d n.d.
Chlorogenic acid N Non n.d. N(P) e
Curcumin N Non n.d. n.d.

Abbreviations: P: positive; N: negative; N(L): negative results of LLNA ; N(P): negative results of PLNA; N(P,L): negative results of PLNA and LLNA; n.d.; no available data.; GPT: guinea pig test. a The results of GPT were cited from Masamoto (2001). b The result of LLNA was cited from Vocanson et al. (2006). c,dThe results of LLNA:DA and PLNA were cited from c Ikarashi et al. (2014) and d Nagaoka Hamano et al. (2007). e The result of PLNA was cited from Liu et al. (2010).

The calculated 1% increase in values of the 10 tested chemicals and our previous data (Yamashita et al., 2015a) are summarized in Table 2. A positive reaction was identified for esculetin, 4-methylesculetin, and isoscopoletin, with 1% increase value of 0.24, 0.30, and 0.07, respectively. Therefore, 4-methylesculetin and esculetin should be recognized as having a strong skin sensitizing potency, with isoscopoletin having a weak-to-modelate skin sensitizing potency. These results indicate that the skin sensitization potency of esculetin and 4-methylesculetin are higher than those of cinnnamic aldehyde which has been shown to cause intraoral allergic dermatitis or stomatitis.

Shikonin is the active ingredient in Lithospermum erythrorhizon and induces an elicitation response at a very low concentration of 0.05%, with a 1% increase value of 31.6. This value is higher than the 1% increase value of 12.0 for DNCB, which is a known extreme skin sensitizing chemical. These results are indicative that shikonin may have extreme skin sensitizing potency. Despite the severity of the skin sensitizing potency of shikonin, we identified only one previous study on allergic responses to shikonin-containing medicine, Shiunko (Natsuaki et al., 2000). The use of shikonin in cosmetics has been approved by the Japanese Cosmetic Industry Association (JCIA) and, therefore, numerous commercial products containing shikonin may have been classified as being safe for use after careful risk assessment (JCIA, 2008). Shikonin also provides several pharmacological effects that have been used in health care, including anti-tumor and anti-inflammatory effects (Fu et al., 2013; Liao et al., 2017). Interestingly, some of these reports have described an anti-allergic activity of shikonin (Lee et al., 2010; Wang et al., 2014). In addition, Chen et al. (2012) reported that application of shikonin to the skin has the potential to regulate the activation of normal T cells and, therefore, the expression and secretion of RANTES and RANTE as typical CC chemokines, as well as leading to skin dendritic cell migration to lymph node. These biological activities may have influenced our test results. Globally, these findings underline the importance of accurate assessments of the bio-activity (as well as the safety and pharmacological activity) of natural organic chemicals prior to using them in foods and cosmetics.

Other chemicals, quercetin, curcumin, naringenin and chlorogenic acid, produced a negative response on LLNA:DAE testing. Furthermore, negative responses for chlorogenic acid have previously been reported using PLNA (Liu et al., 2010). These findings indicate that these fore chemicals do not have a skin sensitizing potential.

In conclusion, four of the 10 natural organic chemicals tested (esculetin, 4-methyl esculetin, isoscopoletin, and shikonin) produced positive skin sensitizing when tested using LLNA:DAE, and shikonin, esculetin and 4-methyl esculetin have remarkable skin sensitizing potency. The other chemicals tested did not have skin sensitizing potential.

Conflict of interest

The authors declare that there is no conflict of interest.

© 2018 The Japanese Society of Toxicology