Sensitization Potential of Low-monomer Diisocyanate Prepolymers

Transcription

1 Sensitization Potential of Low-monomer Diisocyanate Prepolymers Michael Woolhiser, Amber Stephenson and Darrell Boverhof The Dow Chemical Company, Midland, MI ABSTRACT. From the perspective of environmental, health and safety, diisocyanate prepolymers have had limited concern for toxicity, mostly due to their chemical and physical characteristics. Most prepolymer products contain significant concentrations of residual diisocyanate monomer (e.g., TDI) which drive classification and safe handling procedures. In the case of some prepolymer products, the concentration of residual monomer is less than 0.1% wt. such that classification is not driven by diisocyanate monomer. In these instances, prepolymer products can be considered as novel diisocyanate substances and default diisocyanate hazards need to be considered. As sensitization is viewed as the most prominent health effect of diisocyanate substances, several low residual TDIbased prepolymers were tested for sensitization potential using the mouse Local Lymph Node Assay. The test substances had 2 diisocyanate functional groups, were initiated using different polyol substances, and ranged from g/mole. The lower molecular weight prepolymers (950 and 1800 g/mole) exhibited evidence of sensitization potential, and their potencies demonstrated some correlation with MW but were on the order of 100- fold below that of TDI monomer. The 2950 g/mole prepolymer demonstrated lymph node responses which were equivalent to the residual concentration of TDI and, thus, was not considered to have distinct sensitization potential. These data suggest diisocyanate prepolymers can have sensitization potential and reinforces the need to evaluate/test for this hazard. As diisocyanates are drawing increased attention by regulators around the world, this approach is intended to help understand the potential health effects and classifications for prepolymers. INTRODUCTION. Polyisocyanate prepolymers are discrete molecules which are produced from reaction of a polyol molecule with (excess) diisocyanate monomers. For these diisocyanate prepolymer products, the ability to test and directly interpret toxicity potential for the prepolymers is challenging. However, with residual concentrations of diisocyanate monomers at greater than 0.1% wt., classification and labeling of polyisocyanate prepolymers is driven by the hazards of the diisocyanate monomer, therefore specific testing for the prepolymers has been less of a concern. While diisocyanates monomers carry a number of health warnings related to irritation, inhalation exposures, and possible carcinogenic effects, sensitization potential is considered a predominant health effect due to diisocyanate exposures. From the perspective of environmental, health and safety, diisocyanate prepolymers have had limited concern for toxicity due to their chemical and physical characteristics. Typically, these substances have large molecular weights (e.g., > 1,000 g/mole) and low volatility. In the absence of residual diisocyanate monomers (i.e., < 0.1% wt), prepolymers with low residual monomer have little basis for hazard assessment. However, as these products are considered to be reactive diisocyanate substances, one should consider default hazard classification, including classification as sensitizaters, in the absence of specific evidence to the contrary. As the sensitization potency of diisocyanates is considered to be relatively strong and the principal concern for human health, our laboratory designed experiments to test the sensitization potential of low monomer diisocyanates prepolymer molecules. This work was designed to address the hypothesis that polyisocyanate prepolymers lack sensitization potential. The Local Lymph Node Assay (LLNA) was conducted to assess the potential for dermal contact sensitization by measuring the lymphocyte proliferative responses from auricular lymph nodes following topical application of the test material to the mouse ear. Test materials that elicit a stimulation index (SI) of > 3 (i.e., 3-fold greater

2 proliferation than control animals) should be considered positive for dermal sensitization potential. Interpolation between two test concentrations eliciting SI values above and below three (3) can be used to estimate the test material concentration that results in a 3-fold increase in proliferation (EC3) and thus provides the ability to compare relative sensitization potency between test materials. The testing approach was first evaluated using an experimentally prepared TDI-based prepolymer to determine the feasibility of this testing model for such substances and to help define the appropriate controls and analytical assessments required for complete interpretation. Subsequently, the contact sensitization potentials of two additional prepolymer substances with varying molecular weight and polyol chemistry were tested to see if these factors might impact the LLNA sensitization responses and subsequent conclusions on the sensitization potential of polyisocyanate prepolymers. MATERIALS AND METHODS. The LLNA was conducted according to test guidelines from the U.S. EPA (Health Effects Test Guidelines, OPPTS Skin Sensitization) and the OECD Guideline for Testing of Chemicals (Guideline 429: Skin Sensitization: Local Lymph Node Assay). Animal Welfare. Female BALB/c mice (Charles River, Portage, MI) were used because of their availability and vast array of reliable data in this laboratory. Each animal was evaluated by a laboratory veterinarian or a trained animal/toxicology technician under the direct supervision of a laboratory veterinarian to determine the general health status and acceptability for study purposes upon arrival at the laboratory (fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International - AAALAC International). In accordance with the U.S. Department of Agriculture animal welfare regulations, 9 CFR, Subchapter A, Parts 1-4, the animal care and use activities required for conduct of this study were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC). Test Materials. Three TDI-based prepolymer substances were tested: TDI-PTMEG (MW = 2950 g/mole), TDI- Adipate (MW = 1800 g/mole), and TDI-PTMEG (MW = 950 g/mole), where PTMEG = polytetramethylene glycol. All three prepolymer materials were prepared such that residual TDI monomer would be less than 0.01% wt. Vehicle and Positive Controls. Acetone: olive oil (4:1, vol:vol; AOO), the preferred solvent in LLNA guidelines was selected based upon maximum miscibility while maintaining a solution suitable for application. TDI (80:20 ratio of 2,4- and 2,6- isomers; TDI 80 Type 1, The Dow Chemical Company) at a concentration of 0.1% was used as a positive control in this study. TDI is considered a potent sensitizer and has reported LLNA EC3 values in the range of 0.02% % following a standard, guideline protocol (Hilton et al., 1995; Selgrade et al., 2006). A TDI residual control group (mixture of 2,4- and 2,6-TDI representative of analytical measurements) was also included to test the contribution of residual TDI in the prepolymer test material to the LLNA response. Additional reference materials, including the polyols used in the prepolymer synthesis were included in the study design to determine the potential contribution of these reactants to alter the assay response or background signal.. Dose Preparation and Analysis. The prepolymer test substances were dissolved at a maximum concentration of 50% wt, which provided a homogenous solution which could be administered using a disposable pipette. Mid (20% wt) and Low (5% wt) doses of prepolymer samples were also prepared to provide dose response information in the LLNA. All dosing solutions were prepared in an acetone:olive oil vehicle that was previously dried over anhydrous sodium sulfate. A single dose solution was prepared and divided among separate dosing vials (argon-purged septum vials) for each of the three dosing days. The concentrations of the test and reference materials (polyols) in the dose solutions were based on their gravimetric preparation only and were not further analytically verified. The concentrations of any residual TDI occurring in the TDI-prepolymer samples were analytically assessed in order to determine the appropriate concentration and isomer ratio, at which the TDI residual control dosing solutions were prepared. Analytical Methods. The concentrations of residual 2,4- and 2,6-toluenediisocyanate (TDI) isomers in the test and reference substances were determined using a derivatization procedure coupled with HPLC-MS analysis. The

3 concentrations of 2,4- and 2,6-TDI derivatives were determined using external standard calibration, with calibration standards prepared from 1,000 µg/ml certified calibration solutions of the 2,4-TDI (99.9% pure) and 2,6-TDI (99.2% pure) in DMSO (Supelco, Bellefonte, PA, USA). For derivatization, samples of the test and reference material dose solutions (100 µl) were combined with 900 µl of a 4 mm 1-(2-pyridyl)piperazine (Aldrich Chemical, 99.5+%) solution in dry acetonitrile. The resulting mixtures were vortexed for approx. 10 min. and then centrifuged at approx. 2,000 r.p.m. for approx. 10 min. to separate oil and aqueous phases. The resulting supernatants were drawn off and diluted 1:1 (vol:vol) with dry acetonitrile, and the resulting samples were analyzed by HPLC-MS. Dermal Sensitization via the Local Lymph Node Assay. Dosing solutions were applied to the ears (25 µl/ear) of mice (five mice/group) for three consecutive days (days 1-3). Ears were inspected prior to application of the test material solutions, and erythema was evaluated on days 2, 3, and 6. On day 6, all mice received a 250 μl intravenous injection (i.v.) via the lateral tail vein containing 20 μci of 3 H-thymidine (specific activity 2 Ci/mmol; Perkin Elmer product number NET027Z005MC) diluted in phosphate-buffered saline (PBS). Approximately five hours post administration, the mice were euthanized via CO 2 asphyxiation and both auricular lymph nodes located at the bifurcation of the jugular veins were excised and placed in PBS. A single cell suspension of the auricular lymph nodes from each mouse was prepared by gentle mechanical disaggregation using a tissue homogenizer (Stomacher 80 Lab System, Seward Ltd., London, United Kingdom). The cells were washed twice and were suspended in 5% trichloroacetic acid (TCA) for approximately hours. The suspended precipitates were centrifuged (200 x g for 10 minutes) and the supernatant removed. The pellet from each mouse was reconstituted in 5% TCA and subsequently transferred to a scintillation vial containing Aquasol-2 scintillation cocktail (Packard Instrument Company, Meridan, Connecticut). The radioactivity in each precipitate was measured using a scintillation counter and reported as disintegrations per minute (dpm) per mouse. A mean dpm value + SD (standard deviation) was calculated for each experimental group and then the mean SI + SD was calculated for each experimental group using the mean dpm value from the vehicle control mice as the denominator. When appropriate, the EC3 value is determined by interpolating between two values one above and one below the SI value of 3. EC3 = XL + [(3-YL)/(Yh-YL)](Xh-XL): Where, YL = SI value below 3 XL = chemical concentration that elicits YL Yh = SI value above 3 Xh = chemical concentration that elicits Yh Based on the EC3 values derived from the LLNA, it has been proposed that contact allergens can be categorized as weak, moderate, strong, or extreme (Loveless, 2010). RESULTS AND DISCUSSION. Analysis of the experimental TDI-PTMEG (2950 g/mole) revealed that the sample contained % residual TDI of which 3% was 2,4-TDI and 97% was 2,6-TDI (3:97, wt:wt). Therefore, a control group of mice was included that was dosed using a sample of 3:97 TDI at this residual concentration. Furthermore, some additional mice were treated using 80:20 TDI in an attempt to assess any differences in LLNA response relative to TDI isomer ratios. PTMEG that had not been reacted with TDI was also included as a control, as was a methanol-reacted TDI-PTMEG (MeOH-capped TDI-PTMEG), to fully assess the influences of the prepolymer backbone on the LLNA response. Substances with similar chemistry, (e.g., polyols, long chain fatty acid-like alcohols), have been known to cause elevated SI responses in the LLNA but are generally not regarded to be sensitizers (Kreiling et al., 2008; Basketter et al., 2009). A sample of 80:20 TDI (80% 2,4-TDI and 20% 2,6- TDI) at a concentration of 0.1% was included as a positive control for the LLNA. Mice dosed with 0.1% 80:20 TDI exhibited an average stimulation index (SI) of 6.6 relative to the vehicle control mice (Figure 1), which confirmed the proper conduct of the LLNA and is consistent with the reported EC3 value for TDI of 0.03% (Selgrade et al., 2006). Examination of the response to the unreacted PTMEG revealed a SI that was

4 elevated relative to the vehicle control (SI=2.85) but which did not meet the criteria for a positive sensitization response. The response in this group was highly variable with three responding and two non-responding animals when compared to that of the vehicle. The response to the MeOH-capped TDI-PTMEG was similar to that of the PTMEG alone with an SI of 1.78 and a mixture of responding and non-responding mice (data not shown). This group is likely to further represent the variability of the response to PTMEG alone, as the reaction with methanol removes any reactive isocyanate functionalities. Examination of the response in the 0.016% 3:97 TDI residual control group revealed a SI of 2.7 relative to the vehicle control group, which is consistent with the reported EC3 value of 0.03% wt. as 80:20 TDI (Selgrade et al., 2006). A similar response was noted for 0.016% 80:20 TDI control, which yielded an SI value of 1.49 relative to vehicle control mice (data not shown). The responses between these two groups is indicative of the biological variability in response to TDI at these low concentrations and also suggests that this response is not highly isomerspecific at these lower concentrations. Examination of the response using 2950 g/mole TDI-PTMEG dissolved to 50% wt. revealed a SI of 3.32 relative to the vehicle control. Although this exceeds the threshold for a positive response in the LLNA assay, the response in this group should be interpreted relative to the responses in the corresponding control groups. The objective of this study was to determine if the TDI-PTMEG demonstrates sensitization potential which is separate and distinct from any free residual TDI. Comparison of the response in the TDI-PTMEG group to that in the 0.016% TDI residual control group reveals that the 2950 g/mole TDI-PTMEG response does not differ significantly from that which can be attributed to the residual TDI. Thus, the authors conclude that the experimental 2950 g/mole TDI-PTMEG prepolymer did not possess sensitization potential beyond that which could be attributed to the free residual TDI. Subsequent to this experiment which successfully demonstrated the feasibility of testing a pure, diisocyanate prepolymer using the mouse LLNA, two additional TDI-prepolymer substances were evaluated (TDI-PTMEG 950 g/mole or TDI-Adipate 1800 g/mole). The TDI residual concentration for these substances was determined to be 0.018% wt. and 0.025% wt., respectively. Female mice/group received 5% wt., 20% wt. or 50% wt. of TDI-PTMEG 950 g/mole or TDI-Adipate 1800 g/mole on days 1-3. Additional groups of mice were included as controls including the vehicle control (4:1 AOO), the 0.1% 80:20 TDI positive control, the polyol controls and the respective TDI residual controls (0.018% wt. and 0.025% wt. TDI for TDI-PTMEG 950 g/mole and TDI-Adipate 1800 g/mole, respectively) As with the initial study, proper conduct of the LLNA was confirmed via a positive response using 0.1% TDI, which elicited a SI of 6.2 in comparison to vehicle-treated mice. TDI-PTMEG 950 g/mole at concentrations of 5%, 20%, or 50% elicited proliferative responses with stimulation indices that were respectively, 5.5, 6.4, or 7.5 in comparison to vehicle-treated mice (Figure 2). The concentration that would cause a 3-fold increase in proliferation (EC3) was calculated to be 0.1%, which is consistent with a strong dermal sensitization potential as described by an expert ECETOC panel (Loveless et al, 2010). Similarly, the TDI-Adipate 1800 g/mole prepolymer elicited proliferative responses with stimulation indices that were respectively 2.7, 4.9, and 5.2 in comparison to vehicle-treated mice (Figure 3). However, the concentration determined to cause a 3-fold increase in proliferation (EC3) was calculated to be 7.0%, which is consistent with a moderate dermal sensitization potential (Loveless et al, 2010). No notable responses were observed for the respective control groups. In this experiment, the TDI residual concentration at 0.018% wt elicited a stimulation index of 1.0 in comparison to the vehicle-treated mice, while the TDI residual concentration of 0.025% elicited a stimulation index of 1.2. Stimulation indices for the respective polyol controls were also near vehicle (AOO) responses and were distinctive from the TDI-prepolymer data. This experiment was conducted twice and demonstrated consistent results for these TDI-prepolymers. These data indicate that these particular TDI-prepolmers demonstrated sensitization potential which was not attributable to any residual TDI concentration in the preparation.

5 CONCLUSIONS. These experiments set out to evaluate whether low-residual, diisocyanate prepolymers could be successfully tested for sensitization potential. Diisocyanates are somewhat difficult to handle during experimentation as they are very reactive and readily hydrolyze under aqueous conditions. In the case of diisocyanate prepolymer products, one also needs to be aware of any residual diisocyanate monomer, particularly when evaluating sensitization potential, as diisocyanates are generally considered to be relatively strong sensitizers and small amounts, (e.g., 0.02% % wt,) may impact results. With this in mind, care was taken to use dry the dosing solvents and sensitive analytical analyses were preformed to identify remaining concentrations of unreacted, residual TDI monomer in the prepolymer samples. The additions of the concurrent residual controls were necessary in order to fully interpret the data. This series of experiments indicate low-residual diisocyanate prepolymers can accurately be tested using the mouse LLNA and indicate that diisocyanate prepolymers can possess sensitization potential which appears to be distinct from that which can be attributed to any residual TDI. This interpretation is largely consistent with a published report in which tests conducted using a low residual TDI prepolymer demonstrated sensitizing potential for the skin (Bayer Material Science, 2008). As a result, those investigators indicated,.all LF-TDI prepolymers with a free TDI content of less than 0.1 percent are labeled with the risk phrase 43 ("May cause sensitization by skin contact") While it has been shown experimentally using mice that diisocyanate prepolymers can possess sensitization potential, the data presented here also demonstrate the likelihood that there are some prepolymers which would not require such classification and labeling. The analysis for TDI-PTMEG 2950 g/mole resulted in data where the prepolymer was not interpreted to possess sensitization potential based on comparison to concurrent control mice treated with residual concentrations of TDI. Table 1 presents a summary of our LLNA responses and some key characteristics of the TDI-prepolymers tested. It can be seen that the relative potencies for the prepolymers varied by over two orders of magnitude, and the potency for 80:20 TDI is another order of magnitude more potent. These initial data suggest sensitization potency is a function of molecular weight and % NCO functionality, both of which are intuitively plausible. Using such knowledge, one can begin to evaluate additional prepolymers with known characteristics to make informed assessments of sensitization potential. In the absence of such data, these results suggest the need to include warnings of sensitization for low-residual TDI prepolymer products. Diisocyanates are under increasing focus by agencies around the world (e.g., U.S. EPA Chemical Action Plans) as chemicals that may pose a heightened risk to workers and the public. Proper classification and labeling is essential to mitigate the risk of health effects amongst users of diisocyanates, including prepolymers, to meet the principles of Responsible Care. Acknowledgements. The authors thank Lindsay Sosinski for her expertise in conducting these complicated LLNA experiments, as well as Robert West and Steven Erskine for their detailed and thorough dose preparations and analytical analyses. REFRENCES Bayer Material Science, Urethanes Technology International, Vol. 25 No. 2 (April/May). Basketter D, Ball N, Cagen S, Carrillo JC, Certa H, Eigler D, Garcia C, Esch H, Graham C, Haux C, Kreiling R, Mehling A Application of a weight of evidence approach to assessing discordant sensitisation datasets: implications for REACH. Regul Toxicol Pharmacol. Oct;55(1):90-6. Hilton, J., Dearman, R.J., Basketter, D.A., Kimber, I., Identification of chemical respiratory allergens: doseresponse relationships in the mouse IgE test. Toxicol. Methods. 5, (1), Kreiling R, Hollnagel HM, Hareng L, Eigler D, Lee MS, Griem P, Dreessen B, Kleber M, Albrecht A, Garcia C, Wendel A Comparison of the skin sensitizing potential of unsaturated compounds as assessed by the murine

6 local lymph node assay (LLNA) and the guinea pig maximization test (GPMT). Food Chem Toxicol. Jun;46(6): Loveless, S.E., Api A.M., Crevel R.W., Debruyne E., Gamer A., Jowsey I.R., Kern P., Kimber I., Lea L., Lloyd P., Mehmood Z., Steiling W., Veenstra G., Woolhiser M., and Hennes C Potency values from the local lymph node assay: application to classification, labelling and risk assessment. Regul. Toxicol. Pharmacol. Feb 56(1); Selgrade M.K., Boykin, E.H., Haykal-Coates, N., Woolhiser, M.R., Wiescinski, C., Andrews, D.L., Farraj, A.K., Doerfler, D.L. and Gavett, S.H Inconsistencies between Cytokine Profiles, Antibody Responses, and Respiratory Hyperresponsiveness Following Dermal Exposure to Isocyanates. Tox. Sci. 94(1): TABLE 1: Summary of LLNA responses and select physical/chemical characteristics Prepolymer MW (g/mole) % wt NCO Residual TDI (% wt) EC3 (potency) TDI monomer 174 ~48% -- ~0.02% TDI-PTMEG 950 ~9% % 0.11% TDI-Adipate 1800 ~5% 0.025% 7% TDI-PTMEG 2950 ~3% 0.016% Negative (50%)

7 FIGURE 1: LLNA results using TDI-PTMEG 2950 g/mole LLNA Stimulation Index Vehicle 50% wt TDI-PTMEG Polyol control TDI residual (0.016%) TDI pos control (0.1%) FIGURE 2: LLNA using TDI-PTMEG 950 g/mole Stimulation Index Vehicle 5% 20% 50% Polyol TDI residual (0.018%) TDI pos control (0.1%)

8 FIGURE 3: LLNA using TDI-Adipate 1800 g/mole Stimulation Index Vehicle 5% 20% 50% Polyol TDI residual (0.025%) TDI pos control (0.1%)