Registration Dossier

Administrative data

Description of key information

Skin sensitisation

WoE, in silico, QSAR, Danish QSAR database, battery approach: positive

WoE, in silico, ReA analysis, OECD QSAR Toolbox v4.4: positive

WoE, GPMT, TDI (10 %) produced strong contact sensitivity in guinea pigs (prolonged antibody production) (Karol 1981).

WoE, LLNA, TDI (1 %), SI between 3.2 and 13.7 in different mouse strains (Woolhiser 2000)

WoE, LLNA: TDI (0.75 %), SI 1.8 (Van de Briel 2000)

WoE, non-LLNA: MDI; 0.73 mg/kg bw caused sensitisation in 50% of the animals

Respiratory sensitisation

WoE, in silico, QSAR, Danish QSAR database, Leadscope model: positive

WoE, in silico, ReA analysis, OECD QSAR Toolbox v4.4: positive

WoE, guinea pig, inhalation, induction concentration > 0.02 ppm, challenge concentration 0.02 ppm, significant pulmonary response (Aoyama 1994)

WoE, GPMT, pulmonary response in sensitised animals (Karol, 1987)

WoE, MDI, guinea pig, in vivo, clear induction of bronchial hyperreactivity (doses 0.0003-0.3% and 10-100% MDI, respectively), inhalation challenge (concentration: 2.5-3.6 ppm) (Rattray et al. 1994)

WoE, MDI, rats, in vivo, elicitation threshold of 5 mg/m³, Pauluhn and Poole, 2011

Key value for chemical safety assessment

Skin sensitisation

Link to relevant study records

Referenceopen allclose all

Endpoint:
skin sensitisation: in vivo (non-LLNA)
Remarks:
Allergic Contact Dermatitis Guinea pig and human
Type of information:
(Q)SAR
Adequacy of study:
weight of evidence
Study period:
22 June 2020
Reliability:
2 (reliable with restrictions)
Rationale for reliability incl. deficiencies:
results derived from a valid (Q)SAR model and falling into its applicability domain, with adequate and reliable documentation / justification
Justification for type of information:
1. SOFTWARE
Danish QSAR database, OECD QSAR Toolbox v4.4

2. MODEL (incl. version number)
SciQSAR v3.1.00
CASEUltra v1.4.6.6
Leadscope v3.1.1‐10

3. SMILES OR OTHER IDENTIFIERS USED AS INPUT FOR THE MODEL
CC(C)c1cc(C(C)C)c(N=C=O)c(C(C)C)c1N=C=O

4. SCIENTIFIC VALIDITY OF THE (Q)SAR MODEL
please refer to attached QMRFs

5. APPLICABILITY DOMAIN
Please refer to attached QPRF
Guideline:
other: REACH Guidance on QSARs R.6
Principles of method if other than guideline:
- Software tool(s) used including version: Danish QSAR database, OECD QSAR Toolbox v4.4
- Model(s) used: SciQSAR, Leadscope, CaseUltra
- Model description: see field 'Justification for non-standard information', 'Attached justification'
- Justification of QSAR prediction: see field 'Justification for type of information', 'Attached justification'
Justification for non-LLNA method:
Data for these models include results from human epidemiological studies on ACD and results from the Guinea Pig Maximization Test (GPMT). The experimental protocol for GPMT is described in OECD guideline 406.
Specific details on test material used for the study:
CC(C)c1cc(C(C)C)c(N=C=O)c(C(C)C)c1N=C=O
Remarks on result:
positive indication of skin sensitisation
Remarks:
Leadscope model (in applicability domain)
Remarks on result:
positive indication of skin sensitisation
Remarks:
SciQSAR model (in applicability domain)
Remarks on result:
not determinable
Remarks:
CASEUltra model
Key result
Remarks on result:
positive indication of skin sensitisation
Remarks:
battery approach (2 out of three models give positive prediction in applicability domain)

The substance was predicted positive by the models SciQSAR and Leadscope. These predictions were in the applicability domain of these models. The model CaseUltra could not make a prediction. Overall, applying the battery approach, the substance is predicted to be a skin sensitiser (2 out of 3 models make positive predictions within the applicability domain).

Interpretation of results:
Category 1 (skin sensitising) based on GHS criteria
Conclusions:
The substance was predicted positive by the models SciQSAR and Leadscope. These predictions were in the applicability domain of these models. The model CaseUltra could not make a prediction. Overall, applying the battery approach, the substance is predicted to be a skin sensitiser (2 out of 3 models make positive predictions within the applicability domain).
Executive summary:
Skin sensitisation was predicted by the statistical QSAR models SciQSAR, Leadscope and CaseUltra implemented in the Danish QSAR database. The substance was predicted positive by the models SciQSAR and Leadscope. These predictions were in the applicability domain of these models. The model CaseUltra could not make a prediction. Overall, applying the battery approach, the substance is predicted to be a skin sensitiser (2 out of 3 models make positive predictions within the applicability domain). Thus, the prediction is considered to be reliable.
Endpoint:
skin sensitisation: in vivo (LLNA)
Type of information:
(Q)SAR
Adequacy of study:
weight of evidence
Study period:
22 June 2020
Reliability:
2 (reliable with restrictions)
Rationale for reliability incl. deficiencies:
results derived from a valid (Q)SAR model, but not (completely) falling into its applicability domain, with adequate and reliable documentation / justification
Justification for type of information:
1. SOFTWARE
OECD QSAR Toolbox v4.4

2. MODEL (incl. version number)
Read-across analysis (manual approach)

3. SMILES OR OTHER IDENTIFIERS USED AS INPUT FOR THE MODEL
CC(C)c1cc(C(C)C)c(N=C=O)c(C(C)C)c1N=C=O

4. SCIENTIFIC VALIDITY OF THE (Q)SAR MODEL
please refer to attached justification

5. APPLICABILITY DOMAIN
The substance does not fall completely in the applicability domain of the category as the logKow is higher than the one of the other category members. Nevertheless, this is considered not to reduce the reliability of the prediction, because the sensitisation mode-of-action is clearly determined by the isocyanate group. This isocyanate group is the common feature of all category members
Guideline:
other: REACH Guidance on QSARs R.6
Principles of method if other than guideline:
- Software tool(s) used including version: OECD QSAR Toolbox v4.4
- Model(s) used: Read-across analysis (manual approach)
- Model description: see field 'Justification for non-standard information', 'Attached justification'
- Justification of QSAR prediction: see field 'Justification for type of information', 'Attached justification'
Specific details on test material used for the study:
CC(C)c1cc(C(C)C)c(N=C=O)c(C(C)C)c1N=C=O
Remarks on result:
positive indication of skin sensitisation based on QSAR/QSPR prediction
Remarks:
The prediction is based on 14 values, 12 of them (85,7%) equal to predicted value. Prediction confidence is measured by the p-value: 0.00647
Interpretation of results:
Category 1A (indication of significant skin sensitising potential) based on GHS criteria
Conclusions:
The substance was predicted positive in a read-across analysis performed with the OECD QSAR Toolbox v4.4.
Executive summary:

A read-across analysis was performed with the OECD QSAR Toolbox v4.4. Grouping was performed with the profiler Protein binding alerts for skin sensitization according to GHS which identified Isocyanates as alert for skin sensitisation. The prediction was positive. The prediction is based on 14 values, 12 of them (85,7%) equal to predicted value. Prediction confidence is measured by the p-value: 0.00647.

Endpoint:
skin sensitisation: in vivo (non-LLNA)
Type of information:
read-across based on grouping of substances (category approach)
Adequacy of study:
weight of evidence
Justification for type of information:
Please refer to Read-across statement in section 13
Reading:
1st reading
Group:
negative control
Remarks on result:
other: not specified
Reading:
1st reading
Group:
test group
Remarks on result:
positive indication of skin sensitisation
Reading:
1st reading
Group:
positive control
Remarks on result:
other: not specified
Interpretation of results:
Category 1 (skin sensitising) based on GHS criteria
Conclusions:
The test item is a skin sensitizer.
Executive summary:

The skin sensitisation potential of several diisocyanates has been known already long ago and has been summarized in ECHA's scientific report on diisocyanates (ECHA, 2019). Thorne et al. (1987) studied the sensitising potential of HDI, TDI and MDI in mice. The results of this study showed that HDI was the most potent one, followed by MDI and TDI. The dose that caused sensitisation in 50% of the animals was 0.73 mg/kg bw for MDI.

Endpoint:
skin sensitisation: in vivo (non-LLNA)
Type of information:
read-across based on grouping of substances (category approach)
Adequacy of study:
weight of evidence
Justification for type of information:
Please refer to Read-across statement in section 13
Reading:
2nd reading
Hours after challenge:
168
Group:
test group
Dose level:
1 %
No. with + reactions:
0
Total no. in group:
5
Clinical observations:
Antibody titres
Reading:
2nd reading
Hours after challenge:
168
Group:
test group
Dose level:
10 %
No. with + reactions:
4
Total no. in group:
6
Clinical observations:
Antibody titres
Reading:
2nd reading
Hours after challenge:
168
Group:
test group
Dose level:
25 %
No. with + reactions:
3
Total no. in group:
4
Clinical observations:
Antibody titres
Reading:
2nd reading
Hours after challenge:
168
Group:
test group
Dose level:
100 %
No. with + reactions:
8
Total no. in group:
8
Clinical observations:
Antibody titres
Group:
negative control
Remarks on result:
no indication of skin sensitisation
Group:
positive control
Remarks on result:
positive indication of skin sensitisation

Skin sensitivity (contact sensitivity) to TDI was apparent by the seventh day following exposure. After fourteen days, animals were additionally evaluated for TDI sensitivity by serologic analysis and by bronchial provocation challenge. Antibodies to TDI were detected using passive cutaneous anaphylaxis, double diffusion in gel, and radiolabeled antigen binding (Farr) assays. Antibodies were specific for TDI and showed little, if any, cross-reactivity with hexamethylene diisocyanate (HDI), diphenylmethane diisocyanate (MDI), ordicyclohexylmethane diisocyanate (hydrogenated MDI). Respiratory hypersensitivity to TDI could be demonstrated by inhalation challenge with 0.005 ppm TDI, or aerosols of either TDI -protein conjugates or conjugates of p-tolyl isocyanate with protein. The development of respiratory hypersensitivity as a result of dermal contact with TDI emphasizes the importance of careful workplace practices and suggests reevaluation of the role of inhalation and/or dermal contact in causing pulmonary sensitization of workers.

Antibody response to TDI

Sera drawn from guinea pigs 14 days following dermal sensitization with TDI were evaluated for anti-TDI antibodies using three independent serologic techniques.

In passive cutaneous anaphylaxis assays, sera from animals exposed to TDI solutions (1 to 100%) were evaluated for anti-TDI antibodies using TDI-GSA conjugate for intravenous challenge.

TABLE 1

Passive Cutaneous Anaphylaxis Titres of Guinea Pigs Topically Sensitized with In­creasing Concentrations of TDI

Concentration TDI used for sensitization

Number of animals tested

Number positive/ total

Antibody titres (a)

1(b)

5

0/5

0, 0, 0, 0, 0

10%

6

4/6

0,0,2,4, 4, 16

25%

4

3/4

0, 32, 32, 128

100%

8

8/8

32, 32, 32, 32, 32

 

 

 

64, 64, 64

100%, 2days (total

 

 

 

amount100/il)

4

4/4

64, 64, 64, 64

(a) Titre is the reciprocal of the highest serum dilution yielding a reaction of at least 5mm diameter. Intravenous challenge was performed with 2.3mg TDI-GSA in0.3mL pbs containing 3mg Evans blue dye. A latent period of 6 hr was allowed between intradermal injection and intravenous challenge.

(b) Total amount 50 µL.

Antibodies were detected in animals sensitized with 10% TDI, 25% TDI, and 100% TDI solutions. Control sera obtained from the same animals prior to sensitization were uniformly negative. Exposure to 1% TDI did not result in antibody production in any of the animals tested. The results suggested a dose-response relationship between the concentration of TDI used for sensitization and both the number of animals responding per group and the titer of antibodies produced.

The duration of the antibody response resulting from single dermal exposure was investigated by examination of serial bleedings from individual animals extending over a 4-month period. Maximum antibody titres were always observed 2-3 weeks following exposure; thereafter, titers declined. Values of anti-TDI antibodies from an individual guinea pig sensitized to 25% TDI are listed in Table 3.

TABLE2

Duration of Passive Cutaneous Anaphylaxis Titres following TDI Sensitization

Days following sensitization

Antibody titre (a)

0

0

14

128(IgE=8)

21

128

28

32

49

32

120

32

(a) Titre as defined in Table2 using a 6-hr latent period. IgE titre determined using 8-day latent period.

The presence of IgE antibodies was investigated by (1) PCA assays in which antigen was injected intravenously 8 days after intradermal administration of antisera, and (2) heating sera at 56°C for 4 hr prior to PCA assay. IgE antibodies were detected in sera obtained 14-21 days following TDI skin application in much lower quantities than IgG antibodies (determined from 6 hour PCA tests). Comparison of 6 hr and 8 day PCA antibody titres is shown in Table 3.

Antibodies to TDI were also detected using the double diffusion technique. Sera drawn from guinea pigs 14 days following topical exposure to 100% TDI contained antibodies which precipitated with TDI-GSA in gel. Specificity of antibodies was apparent from these studies. Precipitation occurred only with antisera and TDI-GSA conjugate. No precipitin bands were evident when sera were diffused against HDI-GSA, MDI-GSA, or hydrogenated MDI-GSA. In addition, no precipitation occurred with GSA, and pre-exposure sera, obtained from the same guinea pigs prior to TDI exposure were negative in immunodiffusion tests.

A final serologic method was used to confirm antibody specificity. The antigen binding capacity of sera was tested in the presence and absence of isocyanate-containing inhibitors. Using the ammonium sulfate precipitation technique, sera from animals sensitized by topical application of 50 µL100% TDI were evaluated for antibodies to TDI. Serial dilutions of sera resulted in decreased antigen binding. Inhibition studies revealed the TDI-specific nature of antibodies induced by topical TDI exposure. In these studies, the amount of125I antigen bound by antiserum (1:10 dilution of pooled anti-TDI sera) in the presence of various isocyanate-protein conjugates was compared with that bound in the presence of protein carrier.

TABLE3

Inhibition of TDI-[125I]HSA Binding to Anti-TDI Sera by Heterologous Isocyanate-Protein Conjugates

 

 

 

Isocyanate hapten

Concentration

isocyanate-protein added

 

TDI

0.1mg

 

 

1mg

 

HDI

1mg

 

MDI

0.1mg

 

 

1mg

 

HMDI

1mg

 

" Average of three determinations.

Unlabeled TDI-HSA inhibited antigen binding by 71% when added at 0.1 mg/ml concentration. By comparison, the other diisocyanate conjugates showed minimal if any inhibition even at concentrations of 1 mg/ml. The absence of immunologic cross-reactivity among these diisocyanates was confirmed by antigen binding assays using antisera raised specifically to other diisocyanates. Sets of four guinea pigs were sensitized by intradermal injection of either HDI, MDI, or HMDI. Sera obtained 14 days following sensitization were individually tested at a 1:5 dilution for ability to bind radiolabeled TDI-HSA.

TABLE4

Cross-Reactivity of Anti-Diisocyanate Antibodies

Antisera

Percentage TDI-[125I]HSA precipitated"

TDI

71

HDI

2

MDI

10

HMDI

0

" For the assay100 µLantiserum(1:5dilution) was incubated with25 µLTDI-[125I]HSA(13.3ng,100,000cpm) for2hr at37°C.Globulin-bound TDI-[125I]HSA was precipitated using saturated ammonium sulfate. Results indicate the average binding obtained with four antisera to each diisocyanate.

Anti-HDI and anti-HMDI sera failed to bind TDI-[125I]HSA. Antisera to MDI showed a small degree of reactivity with the labelled antigen. All sera had been found to contain antibodies to the respective immunizing diisocyanate by PCA assay.

Pulmonary Sensitivity

The possible presence of respiratory tract reactivity as a consequence of contact sensitization to TDI was investigated by bronchial provocation challenge of animals. Two weeks after dermal exposure to TDI, animals were challenged by inhalation of TDI vapour and/or aerosols of protein conjugates of toluene (mono) isocyanate and toluene diisocyanate. Toluene (mono)isocyanate was used for antigen preparation to avoid cross-linking polymerization of protein. Such reactions had been observed with TDI and proteins (Karolet al,1979b). Bronchial sensitivity was assessed by measurement of the increase in respiratory rate which occurs when immunologically sensitized animals are challenged with aerosolized antigens.

Animals challenged with aerosols of conjugate antigens consisted of eight guinea pigs sensitized by single topical exposure to 100% TDI (Group IV) and four sensitized by two topical applications of 100% TDI (Group V).

Challenge of animals with aerosols containing TDI-protein resulted in elevated respiratory rates in 4 of 12 animals while challenge with TMI-protein aerosol resulted in respiratory rate increases in 5 of 12 guinea pigs. These reactions were "immediate" and occurred during antigen challenge. From previous work in this laboratory using more then 300 guinea pigs, it has been determined that the mean respiratory rate increase during aerosol challenge in control non sensitized guinea pigs is 11% ± 12 (SD) (Karolet al.,1978, 1979a). Because of this background response, only respiratory rate changes greater than 3 SD from the mean (i.e., 47) were considered indicative of a pulmonary reaction. Using this criterion, the respiratory rate changes of the remaining guinea pigs were not considered indicative of bronchial sensitivity. The hapten-specific nature of the respiratory reactions was apparent from the absence of response when animals were challenged with protein carriers alone.

Three sets of guinea pigs were challenged by exposure to 0.005 ppmTDI. The first set consisted of four guinea pigs sensitized by topical application of 50 µL25% TDI. One of the four guinea pigs responded to challenge with a respiratory rate increase of 49%. The other three did not demonstrate reactivity. The second setconsisted of four guinea pigs sensitized by a single application of 50 µL100% TDI (Group IV). None of these animals responded to challenge with TDI vapour. The third set was composed of four animals from Group V sensitized by two applications of 100% TDI. One of these four animals responded to challenge with0.005 ppm TDI with a 61% increase in respiratory rate; a second animal in this group responded with a 44% increase in respiratory rate. The first guinea pig had also given evidence of sensitivity in challenges with the TDI-conjugates (75% increase) and TMI-conjugate (47%increase).

Interpretation of results:
Category 1 (skin sensitising) based on GHS criteria
Conclusions:
The publication is well documented and meets basic scientific principles. Therefore it is considered to be of high quality (Klimisch 2). The controls were valid and the reporting is well accomplished. The test material did induce skin and respiratory sensitisation.
Executive summary:

The potential of the substance toluene diisocyanate was investigated in an guinea pig maximisation test in guinea pigs by Karol, et al.,1981. The test substance was solved in olive oil and applied to skin epi- and percutaneous for topical induction and epicutanous open for challenge purposes. Topical induction was conducted in group I with a total dose of 50 µL 1% TDI, in group II with a total dose of 50 µL 10% TDI, in group III with a total dose of 50 µL 25% TDI, in group IV with a total dose of 50 µL 100% TDI and in group V with a total dose of 100 µL 100% TDI (applied as two 25 µL applications on Day 1 and two 25 µL applications to new dorsal sites on Day 3). Intradermal induction was conducted with injections of 50 µL 100% TDI into each of two dorsal sites.

Topical challenge was performed with the application of 25 µL 0.1% TDI onto clean depilated dorsal sites. These concentrations of TDI did not cause dermal irritation in control guinea pigs. Skin sensitivity (contact sensitivity) to TDI was apparent by the seventh day following exposure. After fourteen days, animals were additionally evaluated for TDI sensitivity by serologic analysis and by bronchial provocation challenge. Antibodies to TDI were detected using passive cutaneous anaphylaxis, double diffusion in gel, and radiolabeled antigen binding (Farr) assays.

In the study reported here, single dermal exposures to TDI produced strong contact sensitivity in guinea pigs. Dermal sensitivity was highly specific for TDI; no cross-reaction was observed upon challenge of animals with HDI. The single dermal exposure to TDI also resulted in prolonged antibody production. The magnitude of the antibody response in the current study was dependent upon the concentration of TDI used for sensitisation. Fifty microliters of 10% TDI was sufficient to stimulate antibody production which was still demonstrable 3 months later. Antibodies were specific for TDI and showed little, if any, cross-reactivity with hexamethylene diisocyanate (HDI), diphenylmethane diisocyanate (MDI), ordicyclohexylmethane diisocyanate (hydrogenated MDI).

Dermal exposure to 100% TDI resulted in pulmonary sensitivity in a percentage of animals. Although all animals produced cytophilic antibodies to TDI as a result of dermal TDI exposure, lung reactions were observed in only some of the animals. Respiratory hypersensitivity to TDI could be demonstrated by inhalation challenge with 0.005 ppm TDI, or aerosols of either TDI -protein conjugates or conjugates of p-tolyl isocyanate with protein. Bronchial challenge with 0.005 ppm TDI vapour elicited respiratory reactions in 2 of 12 animals. However, inhalation challenge with a TDI-protein conjugate elicited reactions in 4 of 12 animals, and challenge with a TMI-protein conjugate elicited respiratory reactions in 5 of the 12 animals. All the respiratory reactions were "immediate" and occurred either during challenge or within 30 min following challenge. In challenges using isocyanate-protein conjugates, hapten-specificity of the pulmonary responses was demonstrated by the failure of aerosolised proteins to evoke pulmonary reactions in any of the animals. TDI-specific pulmonary reactions were elicited more effectively when the TDI (or TMI) was inhaled as a hapten-protein conjugate rather than as TDI vapour. This occurrence may be related to several factors such as solubility, hapten concentration, or size of hapten.

This report of immunologic sensitivity resulting from dermal exposure of animals has importance for setting an appropriate TLV for TDI. Repeated inhalation of TDI as well as dermal contact with TDI each resulted in pulmonary sensitisation to TDI. In the case of dermal contact, 1 drop (50 µL) 100% TDI applied to the skin in the absence of any adjuvant caused antibody production in 100% of animals and pulmonary sensitisation in approximately 30 -40 % of animals.

Endpoint:
skin sensitisation: in vivo (LLNA)
Type of information:
read-across based on grouping of substances (category approach)
Adequacy of study:
weight of evidence
Justification for type of information:
Please refer to Read-across statement in section 13
Parameter:
SI
Value:
1.8
Test group / Remarks:
0.75 % TDI
Parameter:
SI
Test group / Remarks:
only one concentration tested
Remarks on result:
not measured/tested

Lymphocyte proliferation. Mice were treated by topical application on both ears with DNCB, TMA, TDI, PA, or AOO as vehicle control for three consecutive days. In the first study, Day 7 after the first application was found to be optimal for detecting IL-4 production. Thus, 7 days after the first application the local lymphnode were excised. For all compounds, topical application strongly increased the lymphnode weight relative to the vehicle control (p < 0.001; data not shown). DNCB (p < 0.01), TMA (p < 0.01), and TDI (p < 0.001) induced a stronger increase than did PA (data not shown). For all compounds, application strongly increased the LN cell number relative to the vehicle control (p < 0.001; data not shown). TMA (p < 0.05) and TDI (p < 0.001) induced a stronger increase than did PA, while TDI induced a stronger increase than did DNCB (p < 0.001) and TMA (p < 0.01; data not shown). After culturing the LN cells with [3H]thymidine for 24 h, 3H-incorporation was measured. For all compounds, application strongly induced lymphocyte proliferation relative to the vehicle control. Application of TMA and PA resulted in a higher lymphocyte proliferation than DNCB. Per two LN (per animal), similar results were obtained. The magnitude of the response compared to the vehicle control, however, was increased. Application of DNCB and PA resulted in a similar lymphocyte proliferation, while application of TMA and TDI resulted in a 1.8-fold higher proliferation.

Cytokine production. For all compounds, topical application strongly induced IFN-g production compared to the vehicle control. The induction by PA was about threefold lower than by the other allergens. Per two LN (per animal), similar results were obtained. The magnitude of the response compared to the vehicle control, however, was increased. The induction by PA was about fourfold lower than by the other allergens.

For all compounds, application strongly induced IL-4 production compared to the vehicle control. However, the induction by DNCB was about 15- to 30-fold lower than by the other allergens. The induction by TMA was stronger than by TDI and PA. Per two LN (per animal), similar results were obtained. The magnitude of the response compared to the vehicle control, however, was increased. Again, the induction by DNCB was about 15- to 30-fold lower than by the other allergens. Induction by TMA and TDI was stronger than by PA.

In the second study, in an attempt to obtain similar proliferative responses for DNCB and TMA, the TMA concentration was reduced to 10%. Moreover, 10% TMA has been shown to induce a response equivalent to 1% DNCB (Dearman et al. 1996b). The concentrations used for TDI (0.75%) and PA (25%) were similar to those used by Dearman et al. (1996a and 1992a, respectively). Topical application of DNCB and PA resulted in a similar lymphocyte proliferation, while application of TMA and TDI resulted in a 1.8-fold higher proliferation. A clear distinction between the LLNA performed in our studies and by Kimber et al. (1995) is that in our experiments [3H]thymidine labelling is done ex vivo, whereas Kimber et al. use in vivo labelling.

Interpretation of results:
Category 1 (skin sensitising) based on GHS criteria
Conclusions:
The publication is based on experiments according to the local lymph node assay, and was conducted to not only identify allergens but also mark them as either a contact or a respiratory allergen and is considered to be of high quality (reliability Klimisch 2). The validity criteria of the test system are fulfilled. The test material did induce sensitisation and should be classified accordingly.
Executive summary:

Van de Briel and co-workers investigated whether the LLNA was sufficient to not only identify allergens but also mark them as either a contact or a respiratory allergen (Van de Briel et al., 2000). It has been shown that contact allergens preferentially induce a T-helper 1 (TH1) response, whereas respiratory allergens preferentially induce a T-helper 2 (TH2) response. These responses can be discriminated on the basis of cytokine production, such as IFN-g, which is produced by TH1 cells, and IL-4, which is produced by TH2 cells. To this end, IFN-g and IL-4 mRNA expression as well as the production of these cytokines was measured at various time points after primary sensitisation with DNCB and TMA. In a second study, production of IFN-g and IL-4 was measured after primary sensitization with DNCB and TMA, as well as the respiratory sensitizers toluene-2,4-diisocyanate (TDI) and phthalic anhydride (PA); these two were assayed 7 days after the first application. Topical application of DNCB and TMA for three consecutive days does not induce IFN-g mRNA expression relative to the vehicle control, and similarly induces IL-4 mRNA expression. While application of DNCB and PA results in a similar proliferative response, PA induces 16-fold more IL-4 (and fourfold less IFN-g). Furthermore, while application of TMA and TDI results in a 1.8-fold larger proliferative response compared to DNCB, 15- to 30-fold more IL-4 is produced (and similar amounts of IFN-g). Thus, TMA, TDI, and PA can be discriminated from DNCB on the basis of IL-4 production within the LLNA. IFN-g production was similar for DNCB, TMA, and TDI, and fourfold lower for PA, while IL-4 production was similar for TMA, TDI, and PA, and 24-fold lower for DNCB.

In summary, both studies showed induction of IL-4 production by respiratory allergens, with little or no induction by the contact allergen, holding promise for the possibility of identifying allergens (based on lymphocyte proliferation) and respiratory allergens within the LLNA by measuring IL-4 production 7 days after the first application.

Endpoint:
skin sensitisation: in vivo (LLNA)
Type of information:
read-across based on grouping of substances (category approach)
Adequacy of study:
weight of evidence
Justification for type of information:
Please refer to Read-across statement in section 13
Positive control results:
HCA (5, 25 and 50%) exposures elicited lymph node proliferation which was dose responsive within each strain tested (PB0.01). The highest SI following HCA exposure occurred in DBA:2 mice (50% HCA induced a SI of 12.0 and 4148 dpm:2 nodes). DBA:2 is the only strain which showed both SI\3 and statistically significance at all three concentrations of HCA.
According to dpm values, the high to low rank order of the strains following 50% HCA was as follows: SJL\CBA]BALB:c\C57BL:6\ DBA:2\B6C3F1.
Alternatively, the rank order of strains as determined using SI was DBA:2\ BALB:c\CBA\B6C3F1\C57BL:6\SJL.
Parameter:
SI
Value:
13.7
Test group / Remarks:
1 % TDI in CBA mice
Parameter:
SI
Value:
9.4
Test group / Remarks:
1 % TDI in DBA/2 mice
Parameter:
SI
Value:
9.8
Test group / Remarks:
1 % TDI in BALB/c mice
Parameter:
SI
Value:
12.8
Test group / Remarks:
1 % TDI in B6C3F1 mice
Parameter:
SI
Value:
3.2
Test group / Remarks:
1 % TDI in C57BL/6 mice
Parameter:
SI
Value:
4.9
Test group / Remarks:
1 % TDI in SJL mice

Body weights were recorded at the termination of the study and no statistical differences between groups were noted.

The LLNA response of each strain to acetone (vehicle), HCA, TDI and DNFB are shown in Table 1.

Table 1
Local lymph node assay (LLNA) response (dpm/2 nodes)a
Groups DBA/2 BALB/c CBA B6C3F1 C57BL/6 SJL
Acetone 345 +/- 54 574 +/- 113  694 +/- 92 410 +/- 25 887 +/- 128 2111 +/- 307 ***
5%HCA 1028 +/- 135 b** 970 +/- 169 1766 +/- 265 436 +/- 236 1267 +/- 281 2899 +/- 463
25%HCA 2389 +/- 454 b** 2883 +/- 332 b** 2878 +/- 774 b 1175 +/- 79 2895 +/- 497 b* 4967 +/- 1339
50% HCA 4148 +/- 295 b** 6279 +/- 713 b** 6541 +/- 2225b* 3263 +/- 190 b** 4517 +/- 873 b** 8764 +/- 650 b**
1%TDI 3250 +/- 574 ** 5644 +/- 796 ** 9526 +/- 1678 ** 5231 +/- 674 ** 2810 +/- 299 ** 10 410 +/- 1072 **
0.15% DNFB 11 160 +/- 1468 ** 20 280 +/- 1294 ** 24 480 +/- 3779 ** 12 680 +/- 985 ** 25 870 +/- 2103 ** 16 790 +/- 1398 **
aLymph node cell proliferation following exposure to a-hexylcinnamaldehyde (HCA), toluene diisocyanate (TDI) or 2,4-dinitrofluorobenzene (DNFB). Data are expressed as group means+standard error, n= 4 or 5 mice/group.
bLLNA response which was 3-fold greater than acetone controls.
*P<0.05;
**P<0.01 indicate statistical significance from strain acetone control using a Dunnett's test.
***Statistical difference among strain acetone controls using Tukey's multiple comparison test (P<0.01).

The draining lymph node proliferation following acetone exposure varied across strains.

All six mouse strains responded with lymph node proliferation, and stimulation indexes greater than three, following administration of HCA, TDI or DNFB (Tables 1 and 2).

Table 2
Stimulation indexes (SI) following [3H]thymidine uptake in draining lymph nodesa
Dose DBA/2 BALB/c CBA B6C3F1 C57BL/6 SJL
5% HCA 3.0 1.7 2.5 1.0 1.4 1.4
25% HCA 6.9 5.0 4.1 2.9 3.3 2.4
50% HCA 12.0 10.9 9.4 8.0 5.1 4.1
1.0% TDI 9.4 9.8 13.7 12.8 3.2 4.9
0.15% DNFB 32.4 35.3 35.3 30.9 29.2 8.0
aLocal lymph node assay (LLNA) stimulation indexes following exposure to a-hexylcinnamaldehyde (HCA), toluene diisocyanate (TDI) or 2,4-dinitrofluorobenzene (DNFB). Stimulation indexes for each chemical exposure was determined by dividing the mean dpm response towards a chemical by the mean dpm response following acetone applications.n=4 or 5 mice per group.

Following 1.0% TDI exposure CBA (9526 dpm: 2 nodes) and B6C3F1 (5231 dpm:2 nodes) mice experienced the highest SI values of 13.7 and 12.8, respectively. DBA:2 and BALB:c mice demonstrated similar SI values of 9.4 and 9.8 following proliferative responses of 3250 and 5644 dpm:2 nodes. The lowest SI values occurred in C57BL:6 and SJL mice. While C57BL:6 mice demonstrated a SI of 3.2 and 2810 dpm:2 nodes following 1% TDI, SJL mice experienced a SI of 4.9 in spite of the highest proliferative response to TDI (10 410 dpm:2 nodes). TDI dose groups within each strain were statistically different from the respective strain acetone controls (PB0.01). The rank order of strains following TDI exposures were different than those following HCA or DNFB exposures; based on dpm values, the highest to lowest order was SJL\CBA\BALB:c]B6C3F1\DBA\ C57BL:6. The rank order based on SI values was CBA\B6C3F1\BALB: c]DBA:2\SJL\ C57BL:6.

Interpretation of results:
Category 1 (skin sensitising) based on GHS criteria
Conclusions:
The publication is based on experiments performed according to the ICCVAM Peer Review Panel report (NIEHS, 1999), Local lymph node assay, and was conducted in 6 different mouse strains and considered to be of high quality (reliability Klimisch 2). The validity criteria of the test system are fulfilled. The test material did induce sensitisation and should be classified accordingly.
Executive summary:

The potential of toluene diisocyanate to induce skin sensitivity was investigated in different mouse strains via the local lymph node assay to assess the impact of choice of mouse strain (Woolhiser et al., 2000). The objective of these studies was to begin to assess the response of chemical sensitizers in the LLNA across six strains of female mice (C57BL:6, SJL:J, BALB:c, B6C3F1, DBA:2 and CBA). Draining lymph node proliferation was evaluated following exposures to three concentrations of a-hexylcinnamaldehyde (HCA), a moderate contact sensitizer which is one of the chemicals recommended by the OECD as a positive control for the LLNA. In addition, the strong contact sensitizer 2,4-dinitrofluorobenzene and the potent IgE-mediated sensitizer toluene diisocyanate were evaluated at single, moderate concentrations as positive controls for T cell mediated and IgE-mediated responses. The six murine strains demonstrated varying levels of lymph node proliferation following exposure to three chemical sensitizers. These studies suggest that the specific combination of strain and antigen may be more important than a strain’s Th1:Th2 predominance.

DBA:2, B6C3F1, BALB:c and CBA mice had essentially equal levels of lymph node proliferation following exposure to the three chemicals. While C57BL:6 mice gave similar results as CBA mice following DNFB and HCA administration, the LLNA response to TDI was considerably lower. SJL mice provided low stimulation indexes (SI) values for all three chemicals evaluated. Regardless of the level of LLNA response, all six mouse strains identified the sensitisation potential of HCA, TDI or DNFB. Based on these studies, DBA:2, B6C3F1 and BALB:c mice are good choices for continued evaluation as additional mouse strains for use in the LLNA.

Although C57BL:6 mice are reported to be predisposed to Th1 immune responses (Särnstrand et al., 1999) and as such had the lowest response to the IgE inducing chemical TDI, SJL mice (low IgE responders) gave the highest dpm response to TDI. The other four strains demonstrated high proliferative responses to all three chemicals irrespective of any possible Th1:Th2 predominance.

Regardless of the lower SI values, SJL mice still identified all three chemicals as sensitizers according to 3 -fold SI values, statistical significance and dose responsive proliferation. This initial set of studies highlights the importance of mouse strain when developing assay models. While CBA mice were verified as a good selection for the LLNA, the data from these studies suggest DBA:2, B6C3F1 and BALB:c mice are essentially equal to CBA mice when evaluating the sensitizing potential of these three chemicals. While C57BL:6 mice gave similar results as CBA mice following treatment with DNFB and HCA, the LLNA response to TDI was considerably lower than those of DBA:2, B6C3F1, BALB:c and CBA mice. Based on this single series of studies, DBA:2, B6C3F1 and BALB:c mice appear to be reasonable alternative strains for evaluation for use in the LLNA.

Endpoint conclusion
Endpoint conclusion:
adverse effect observed (sensitising)
Additional information:

Skin sensitisation

in silico

Skin sensitisation was predicted by the statistical QSAR models SciQSAR, Leadscope and CaseUltra implemented in the Danish QSAR database. The substance was predicted positive by the models SciQSAR and Leadscope. These predictions were in the applicability domain of these models. The model CaseUltra could not make a prediction. Overall, applying the battery approach, the substance is predicted to be a skin sensitiser (2 out of 3 models make positive predictions within the applicability domain). The prediction is considered to be reliable.

In addition, a read-across analysis was performed with the OECD QSAR Toolbox v4.4. Grouping was performed with the profiler Protein binding alerts for skin sensitization according to GHS which identified Isocyanates as alert for skin sensitisation. The prediction was positive. The prediction is based on 14 values, 12 of them (85,7%) equal to predicted value. Prediction confidence is measured by the p-value: 0.00647.

in vivo (read-across)

The potential of the substance toluene diisocyanate was investigated in a guinea pig maximisation test in guinea pigs by Karol, et al. (1981). The test substance was solved in olive oil and applied to skin epi- and percutaneous for topical induction and epicutaneous open or via injection for challenge purposes. Topical induction was conducted in group I with a total dose of 50 µL 1% TDI, in group II with a total dose of 50 µL 10% TDI, in group III with a total dose of 50 µL 25% TDI, in group IV with a total dose of 50 µL 100% TDI and in group V with a total dose of 100 µL 100% TDI (applied as two 25 µL applications on day 1 and two 25 µL applications to new dorsal sites on Day 3). Intradermal induction was conducted with injections of 50 µL 100% TDI into each of two dorsal sites.

Topical challenge was performed with the application of 25 µL 0.1% TDI onto clean depilated dorsal sites. These concentrations of TDI did not cause dermal irritation in control guinea pigs. Skin sensitivity (contact sensitivity) to TDI was apparent by the seventh day following exposure. After 14 days, animals were additionally evaluated for TDI sensitivity by serologic analysis and by bronchial provocation challenge. Antibodies to TDI were detected using passive cutaneous anaphylaxis, double diffusion in gel, and radiolabeled antigen binding (Farr) assays.

In the study reported here, single dermal exposures to TDI produced strong contact sensitivity in guinea pigs. Dermal sensitivity was highly specific for TDI; no cross-reaction was observed upon challenge of animals with HDI. The single dermal exposure to TDI also resulted in prolonged antibody production. The magnitude of the antibody response in the current study was dependent upon the concentration of TDI used for sensitisation. Fifty microliters of 10% TDI was sufficient to stimulate antibody production which was still demonstrable 3 months later. Antibodies were specific for TDI and showed little, if any, cross-reactivity with hexamethylene diisocyanate (HDI), diphenylmethane diisocyanate (MDI), or dicyclohexylmethane diisocyanate (hydrogenated MDI).

Dermal exposure to 100% TDI resulted in pulmonary sensitivity in a percentage of animals. Although all animals produced cytophilic antibodies to TDI as a result of dermal TDI exposure, lung reactions were observed in only some of the animals. Respiratory hypersensitivity to TDI could be demonstrated by inhalation challenge with 0.005 ppm TDI, or aerosols of either TDI -protein conjugates or conjugates of p-tolyl isocyanate with protein. Bronchial challenge with 0.005 ppm TDI vapour elicited respiratory reactions in 2 of 12 animals. However, inhalation challenge with a TDI-protein conjugate elicited reactions in 4 of 12 animals, and challenge with a TMI-protein conjugate elicited respiratory reactions in 5 of the 12 animals. All the respiratory reactions were "immediate" and occurred either during challenge or within 30 min following challenge. In challenges using isocyanate-protein conjugates, hapten-specificity of the pulmonary responses was demonstrated by the failure of aerosolised proteins to evoke pulmonary reactions in any of the animals. TDI-specific pulmonary reactions were elicited more effectively when the TDI (or TMI) was inhaled as a hapten-protein conjugate rather than as TDI vapour. This occurrence may be related to several factors such as solubility, hapten concentration, or size of hapten.

This report of immunologic sensitivity resulting from dermal exposure of animals has importance for setting an appropriate TLV for TDI. Repeated inhalation of TDI as well as dermal contact with TDI each resulted in pulmonary sensitisation to TDI. In the case of dermal contact, 1 drop (50 µL) 100%TDI applied to the skin in the absence of any adjuvant caused antibody production in 100% of animals and pulmonary sensitisation in approximately 30 -40 % of animals.

The potential of toluene diisocyanate to induce skin sensitivity was investigated in different mouse strains via the local lymph node assay to assess the impact of choice of mouse strain (Woolhiser et al., 2000). The objective of these studies was to begin to assess the response of chemical sensitizers in the LLNA across six strains of female mice (C57BL:6, SJL:J, BALB:c, B6C3F1, DBA:2 and CBA). Draining lymph node proliferation was evaluated following exposures to three concentrations of a-hexylcinnamaldehyde (HCA), a moderate contact sensitizer which is one of the chemicals recommended by the OECD as a positive control for the LLNA. In addition, the strong contact sensitizer 2,4-dinitrofluorobenzene and the potent IgE-mediated sensitizer toluene diisocyanate were evaluated at single, moderate concentrations as positive controls for T cell mediated and IgE-mediated responses. The six murine strains demonstrated varying levels of lymph node proliferation following exposure to three chemical sensitizers. These studies suggest that the specific combination of strain and antigen may be more important than a strain’s Th1:Th2 predominance.

DBA:2, B6C3F1, BALB:c and CBA mice had essentially equal levels of lymph node proliferation following exposure to the three chemicals. While C57BL:6 mice gave similar results as CBA mice following DNFB and HCA administration, the LLNA response to TDI was considerably lower. SJL mice provided low stimulation indexes (SI) values for all three chemicals evaluated. Regardless of the level of LLNA response, all six mouse strains identified the sensitisation potential of HCA, TDI or DNFB. Based on these studies, DBA:2, B6C3F1 and BALB:c mice are good choices for continued evaluation as additional mouse strains for use in the LLNA.

Although C57BL:6 mice are reported to be predisposed to Th1 immune responses (Särnstrand et al., 1999) and as such had the lowest response to the IgE inducing chemical TDI, SJL mice (low IgE responders) gave the highest dpm response to TDI. The other four strains demonstrated high proliferative responses to all three chemicals irrespective of any possible Th1:Th2 predominance.

Regardless of the lower SI values, SJL mice still identified all three chemicals as sensitizers according to 3 -fold SI values, statistical significance and dose responsive proliferation. This initial set of studies highlights the importance of mouse strain when developing assay models. While CBA mice were verified as a good selection for the LLNA, the data from these studies suggest DBA:2, B6C3F1 and BALB:c mice are essentially equal to CBA mice when evaluating the sensitising potential of these three chemicals. While C57BL:6 mice gave similar results as CBA mice following treatment with DNFB and HCA, the LLNA response to TDI was considerably lower than those of DBA:2, B6C3F1, BALB:c and CBA mice. Based on this single series of studies, DBA:2, B6C3F1 and BALB:c mice appear to be reasonable alternative strains for evaluation for use in the LLNA.

Van de Briel and co-workers investigated whether the local lymph node assay (LLNA) was sufficient to not only identify allergens but also mark them as either a contact or a respiratory allergen (van de Briel et al., 2000). It has been shown that contact allergens preferentially induce a T-helper 1 (TH1) response, whereas respiratory allergens preferentially induce a T-helper 2 (TH2) response. These responses can be discriminated on the basis of cytokine production, such as IFN-g, which is produced by TH1 cells, and IL-4, which is produced by TH2 cells. To this end, IFN-g and IL-4 mRNA expression as well as the production of these cytokines was measured at various time points after primary sensitisation with DNCB and TMA. In a second study, production of IFN-g and IL-4 was measured after primary sensitisation with DNCB and TMA, as well as the respiratory sensitizers toluene-2,4-diisocyanate (TDI) and phthalic anhydride (PA); these two were assayed 7 days after the first application. Topical application of DNCB and TMA for three consecutive days does not induce IFN-g mRNA expression relative to the vehicle control, and similarly induces IL-4 mRNA expression. While application of DNCB and PA results in a similar proliferative response, PA induces 16-fold more IL-4 (and fourfold less IFN-g). Furthermore, while application of TMA and TDI results in a 1.8-fold larger proliferative response compared to DNCB, 15- to 30-fold more IL-4 is produced (and similar amounts of IFN-g). Thus, TMA, TDI, and PA can be discriminated from DNCB on the basis of IL-4 production within the LLNA. IFN-g production was similar for DNCB, TMA, and TDI, and fourfold lower for PA, while IL-4 production was similar for TMA, TDI, and PA, and 24-fold lower for DNCB.

In summary, both studies showed induction of IL-4 production by respiratory allergens, with little or no induction by the contact allergen, holding promise for the possibility of identifying allergens (based on lymphocyte proliferation) and respiratory allergens within the LLNA by measuring IL-4 production 7 days after the first application.

The skin sensitisation potential of several diisocyanates has been known already long ago and has been summarized in ECHA's scientific report on diisocyanates (ECHA, 2019). Thorne et al. (1987) studied the sensitising potential of HDI, TDI and MDI in mice. The results of this study showed that HDI was the most potent one, followed by MDI and TDI. The dose that caused sensitisation in 50% of the animals was 0.73 mg/kg bw for MDI.


Respiratory sensitisation

Link to relevant study records

Referenceopen allclose all

Endpoint:
respiratory sensitisation: in vivo
Type of information:
(Q)SAR
Adequacy of study:
weight of evidence
Study period:
22 June 2020
Reliability:
2 (reliable with restrictions)
Rationale for reliability incl. deficiencies:
results derived from a valid (Q)SAR model, but not (completely) falling into its applicability domain, with adequate and reliable documentation / justification
Justification for type of information:
1. SOFTWARE
OECD QSAR Toolbox v4.4

2. MODEL (incl. version number)
Read-across analysis (manual approach)

3. SMILES OR OTHER IDENTIFIERS USED AS INPUT FOR THE MODEL
CC(C)c1cc(C(C)C)c(N=C=O)c(C(C)C)c1N=C=O

4. SCIENTIFIC VALIDITY OF THE (Q)SAR MODEL
please refer to attached justification

5. APPLICABILITY DOMAIN
The substance does not fall completely in the applicability domain of the category as the logKow is higher than the one of the other category members. Nevertheless, this is considered not to reduce the reliability of the prediction, because the sensitisation mode-of-action is clearly determined by the isocyanate group. This isocyanate group is the common feature of all category members
Guideline:
other: REACH Guidance on QSARs R.6
Principles of method if other than guideline:
- Software tool(s) used including version: OECD QSAR Toolbox v4.4
- Model(s) used: Read-across analysis (manual approach)
- Model description: see field 'Justification for non-standard information', 'Attached justification'
- Justification of QSAR prediction: see field 'Justification for type of information', 'Attached justification'
Specific details on test material used for the study:
CC(C)c1cc(C(C)C)c(N=C=O)c(C(C)C)c1N=C=O
Results:
The prediction of the read-across analysis was positive. The prediction is based on 3 values, 3 of them (100%) equal to predicted value. Prediction confidence is measured by the p-value: 0.125
Interpretation of results:
Category 1 (respiratory sensitising) based on GHS criteria
Conclusions:
The substance was predicted positive for respiratory sensitisation in a read-across analysis performed with the OECD QSAR Toolbox v4.4.
Executive summary:

A read-across analysis was performed with the OECD QSAR Toolbox v4.4. Grouping was performed with the profiler US EPA New Chemical Categories in the group Isocyanates. The prediction was positive. The prediction is based on 3 values, 3 of them (100%) equal to predicted value. Prediction confidence is measured by the p-value: 0.125

Endpoint:
respiratory sensitisation: in vivo
Type of information:
(Q)SAR
Adequacy of study:
weight of evidence
Study period:
22 June 2020
Reliability:
2 (reliable with restrictions)
Rationale for reliability incl. deficiencies:
results derived from a valid (Q)SAR model and falling into its applicability domain, with adequate and reliable documentation / justification
Justification for type of information:
1. SOFTWARE
Danish QSAR database, OECD QSAR Toolbox v4.4

2. MODEL (incl. version number)
Leadscope Enterprise model for respiratory sensitisation in humans, Danish QSAR Group at DTU Food. (Leadscope v3.1.1‐10)

3. SMILES OR OTHER IDENTIFIERS USED AS INPUT FOR THE MODEL
CC(C)c1cc(C(C)C)c(N=C=O)c(C(C)C)c1N=C=O

4. SCIENTIFIC VALIDITY OF THE (Q)SAR MODEL
please refer to attached QMRFs

5. APPLICABILITY DOMAIN
please refer to attached QPRF
Guideline:
other: REACH Guidance on QSARs R.6
Principles of method if other than guideline:
- Software tool(s) used including version: Danish QSAR database, OECD QSAR Toolbox v4.4
- Model(s) used: Leadscope Enterprise model for respiratory sensitisation in humans, Danish QSAR Group at DTU Food. (Leadscope v3.1.1‐10)
- Model description: see field 'Justification for non-standard information', 'Attached justification'
- Justification of QSAR prediction: see field 'Justification for type of information', 'Attached justification'
Specific details on test material used for the study:
CC(C)c1cc(C(C)C)c(N=C=O)c(C(C)C)c1N=C=O
Results:
The prediction is positive. It falls within the applicability domain of the model. Thus, it is considered to be reliable.
Interpretation of results:
Category 1 (respiratory sensitising) based on GHS criteria
Conclusions:
The substance is predicted positive for respiratory sensitisation by the statistical QSAR model Leadscope.
Executive summary:

Respiratory sensitisation was predicted by the Leadscope model (Danish QSAR database). The prediction is positive. It falls within the applicability domain of the model. Thus, the prediction is considered to be reliable.

Endpoint:
respiratory sensitisation: in vivo
Type of information:
read-across based on grouping of substances (category approach)
Adequacy of study:
weight of evidence
Justification for type of information:
Please refer to Read-across statement in section 13
Results:
a more clear induction of bronchial hyperreactivity in guinea pigs by a single intradermal or epidermal application of MDI followed by inhalation challenge 21 days later than by exposure to MDI by inhalation only.
Interpretation of results:
Category 1 (respiratory sensitising) based on GHS criteria
Conclusions:
The test item is a respiratory sensitiser.
Executive summary:

The respiratory sensitisation potential of diisocyanates is well established. Diisocyanates like MDI have also been shown to cause respiratory hypersensitivity upon dermal exposure according to the studies summarized in the ECHA's Scientific Report on Diisocyanates (ECHA, 2019). In the study of Rattray et al. (1994), Guinea pigs were exposed to MDI by i.d. injection, by topical application or by inhalation. Their data indicate that the route of exposure influences markedly the effectiveness of sensitization to respiratory allergens such as MDI and that skin contact may be an important cause of occupational respiratory allergy.

The study indicated a more clear induction of bronchial hyperreactivity in guinea pigs by a single intradermal or epidermal application of MDI (doses 0.0003-0.3% and 10-100% MDI, respectively) followed by inhalation challenge 21 days later at concentrations of 25.9-36.5 mg/m3 (2.5-3.6 ppm; corresponding to 8.7-12 mg/m³ NCO) than by exposure to MDI by inhalation only [19.4-23.7 mg/m³ (1.9-2.3 ppm; corresponding to 6.5-8.0 mg/m³ NCO) 3 h/day for 5 consecutive days, and challenge 21 days after the first exposure at 34.6-44.1 mg/m³ (3.4-4.3 ppm; corresponding to 12-15 mg/m³ NCO)].

Endpoint:
respiratory sensitisation: in vivo
Type of information:
read-across based on grouping of substances (category approach)
Adequacy of study:
weight of evidence
Justification for type of information:
Please refer to Read-across statement in section 13
Results:
Elicitation threshold of 5 mg/m³
The potential to cause sensitisation was slightly higher for high-dose, short-term exposure than equal cumulative exposure during a longer exposure period.
Interpretation of results:
Category 1 (respiratory sensitising) based on GHS criteria
Conclusions:
The substance is a respiratory sensitiser.
Executive summary:

Several studies investigating the effects of different diisocyanates in in vivo asthma models have  been  published and have been summarized in ECHA's scientific report on diisocyanates (ECHA, 2019).

Pauluhn and Poole (2011) presented a dose-dependent increase in respiratory rate and bronchioalveolar lavage parameters in rats exposed to MDI. The authors identified an elicitation threshold of 5 mg/m³. The potential to cause sensitisation was slightly higher for high-dose, short-term exposure than equal cumulative exposure during a longer exposure period.

Endpoint conclusion
Endpoint conclusion:
adverse effect observed (sensitising)
Additional information:

in silico

Respiratory sensitisation was predicted by the Leadscope model (Danish QSAR database). The prediction is positive. It falls within the applicability domain of the model. Thus, the prediction is considered to be reliable.

In addition, a read-across analysis was performed with the OECD QSAR Toolbox v4.4. Grouping was performed with the profiler US EPA New Chemical Categories in the group Isocyanates. The prediction was positive. The prediction is based on 3 values, 3 of them (100%) equal to predicted value. Prediction confidence is measured by the p-value: 0.125

in vivo (read-across)

Aoyama and co-workers conducted an animal exposure experiment (Aoyama et al., 1994), in which animals were sensitized and challenged by inhalation of free TDI, so simulating a workplace exposure situation to compare toluene diisocyanate (TDI) concentrations which resulted in antibody production with those which elicited pulmonary responses and secondly, to provide an experimental basis for setting up OELs of chemical sensitizers. Therefore groups of guinea pigs were exposed to TDI from 0.02 to 1.0 ppm for 3 h/day on 5 consecutive days. Three weeks later the animals were challenged with 0.02 ppm of free TDI for 15 min. TDI specific antibodies and pulmonary responses were evaluated. Specific antibody production showed a linear correlation to TDI concentration at induction. An induction level of 0.02 ppm TDI failed to stimulate antibody production in animals. Most of the animals exposed to TDI levels above 0.2 ppm displayed significant pulmonary responses, but no correlation was found between TDI concentration at induction and the intensity of pulmonary response upon challenge to free TDI. These results indicated that there was a threshold concentration of 0.02 ppm TDI for antibody production and for the development of pulmonary response. Although it failed to demonstrate sensitizing potency, 0.02 ppm free TDI did elicit a pulmonary response in some of the animals sensitized to TDI at levels ranging from 0.2 to 1.0 ppm. It was also found that exposure to TDI at a level lower than its threshold concentration for sensitisation may elicit a response in previously sensitized individuals. The intensity of pulmonary response was not notably influenced by TDI concentration at induction levels above 0.2 ppm. This result may be related to several factors such as solubility, hapten concentration, or size of hapten. This finding is also of great importance in extrapolating from the results of an animal model to an industrial situation. However, free TDI was able to elicit immediate-onset reactions, within 60 min following the challenge. No close association was seen between antibody production and the experience of pulmonary response. High antibody production did not stick to the development of pulmonary response. On the other hand, several animals with low antibody titres displayed positive pulmonary responses. This discrepancy might be derived from differences in the ability to release histamine due to the different distribution of specific antibodies in the lungs of animals.

Karol and co-workers investigated the potential of toluene diisocyanate to induce respiratory sensitisation in guinea pigs (Karol, et al., 1981). The test substance was solved in olive oil and applied to skin epi- and percutaneous for topical induction and epicutaneous open for challenge purposes. Topical induction was conducted in group I with a total dose of 50 µL 1% TDI, in group II with a total dose of 50 µL 10% TDI, in group III with a total dose of 50 µL 25% TDI, in group IV with a total dose of 50 µL 100% TDI and in group V with a total dose of 100 µL 100% TDI (applied as two 25 µL applications on Day 1 and two 25 µL applications to new dorsal sites on Day 3). Intradermal induction was conducted with injections of 50 µL 100% TDI into each of two dorsal sites. Topical challenge was performed with the application of 25 µL 0.1% TDI onto clean depilated dorsal sites. These concentrations of TDI did not cause dermal irritation in control guinea pigs. Skin sensitivity (contact sensitivity) to TDI was apparent by the seventh day following exposure. After fourteen days, animals were additionally evaluated for TDI sensitivity by serologic analysis and by bronchial provocation challenge. Antibodies to TDI were detected using passive cutaneous anaphylaxis, double diffusion in gel, and radiolabeled antigen binding (Farr) assays.

Dermal exposure to 100% TDI resulted in pulmonary sensitivity in a percentage of animals. Although all animals produced cytophilic antibodies to TDI as a result of dermal TDI exposure, lung reactions were observed in only some of the animals. Respiratory hypersensitivity to TDI could be demonstrated by inhalation challenge with 0.005 ppm TDI, or aerosols of either TDI -protein conjugates or conjugates of p-tolyl isocyanate with protein. Bronchial challenge with 0.005 ppm TDI vapour elicited respiratory reactions in 2 of 12 animals. However, inhalation challenge with a TDI-protein conjugate elicited reactions in 4 of 12 animals, and challenge with a TMI-protein conjugate elicited respiratory reactions in 5 of the 12 animals. All the respiratory reactions were "immediate" and occurred either during challenge or within 30 min following challenge. In challenges using isocyanate-protein conjugates, hapten-specificity of the pulmonary responses was demonstrated by the failure of aerosolised proteins to evoke pulmonary reactions in any of the animals. TDI-specific pulmonary reactions were elicited more effectively when the TDI (or TMI) was inhaled as a hapten-protein conjugate rather than as TDI vapour. This occurrence may be related to several factors such as solubility, hapten concentration, or size of hapten.

This report of immunologic sensitivity resulting from dermal exposure of animals has importance for setting an appropriate TLV for TDI. Repeated inhalation of TDI as well as dermal contact with TDI each resulted in pulmonary sensitisation to TDI. In the case of dermal contact, 1 drop (50 µL) 100% TDI applied to the skin in the absence of any adjuvant caused antibody production in 100% of animals and pulmonary sensitisation in approximately 30 -40 %of animals.

In the study reported here, single dermal exposures to TDI produced strong contact sensitivity in guinea pigs. Dermal sensitivity was highly specific for TDI; no cross-reaction was observed upon challenge of animals with HDI. The single dermal exposure to TDI also resulted in prolonged antibody production. Antibodies had previously been found to result from dermal contact with chemicals in guinea pigs (Chase, 1948) and in mice (Thomas et al., 1976). The magnitude of the antibody response in the current study was dependent upon the concentration of TDI used for sensitization. Fifty microliters of 10% TDI was sufficient to stimulate antibody production which was still demonstrable 3 months later. The specificity of the humoral antibody response to TDI was also highly specific for TDI. In gel diffusion studies, precipitin lines were observed only with sera and TDI conjugates; no reactions were seen with HDI, MDI, or HMDI conjugate antigens. Similarly, in the more sensitive radiolabeled antigen-binding assay, antibodies appeared highly specific for TDI.

 

The respiratory sensitisation potential of diisocyanates is well established. Diisocyanates like MDI have also been shown to cause respiratory hypersensitivity upon dermal exposure according to the studies summarized in the ECHA's Scientific Report on Diisocyanates (ECHA, 2019). In the study of Rattray et al. (1994), Guinea pigs were exposed to MDI by i.d. injection, by topical application or by inhalation. Their data indicate that the route of exposure influences markedly the effectiveness of sensitization to respiratory allergens such as MDI and that skin contact may be an important cause of occupational respiratory allergy.

The study indicated a more clear induction of bronchial hyperreactivity in guinea pigs by a single intradermal or epidermal application of MDI (doses 0.0003-0.3% and 10-100% MDI, respectively) followed by inhalation challenge 21 days later at concentrations of 25.9-36.5 mg/m³ (2.5-3.6 ppm; corresponding to 8.7-12 mg/m³ NCO) than by exposure to MDI by inhalation only [19.4-23.7 mg/m³ (1.9-2.3 ppm; corresponding to 6.5-8.0 mg/m³ NCO) 3 h/day for 5 consecutive days, and challenge 21 days after the first exposure at 34.6-44.1 mg/m³ (3.4-4.3 ppm; corresponding to 12-15 mg/m³ NCO)].

Several studies investigating the effects of different diisocyanates in in vivo asthma models have  been  published and have been summarized in ECHA's scientific report on diisocyanates (ECHA, 2019).

Pauluhn and Poole (2011) presented a dose-dependent increase in respiratory rate and bronchioalveolar lavage parameters in rats exposed to MDI. The authors identified an elicitation threshold of 5 mg/m³. The potential to cause sensitisation was slightly higher for high-dose, short-term exposure than equal cumulative exposure during a longer exposure period.

Justification for classification or non-classification

Skin sensitisation:

The test material does meet the criteria for classification and will require labelling for skin sensitisation (Category 1) in accordance with Regulation (EC) No 1272/2008.

Respiratory sensitisation:

The test material does meet the criteria for classification and will require labelling for respiratory sensitisation (Category 1) in accordance with Regulation (EC) No 1272/2008.