Skin Sensitization to Linalyl Hydroperoxide: Support for Radical

The sensitizing properties of these molecules have been screened on mice using the local lymph node assay (LLNA) and further evaluated on guinea pigs ...
0 downloads 0 Views 205KB Size
Download PDF
Chem. Res. Toxicol. 1997, 10, 987-993

987

Skin Sensitization to Linalyl Hydroperoxide: Support for Radical Intermediates Michael Bezard,† Ann-Therese Karlberg,‡ Johan Montelius,‡ and Jean-Pierre Lepoittevin*,† Laboratoire de Dermatochimie associe´ au CNRS, Universite´ Louis Pasteur, Clinique Dermatologique, CHU, F-67091 Strasbourg Cedex, France, and Department of Occupational Health, Dermatology Section, National Institute for Working Life, S-17184 Solna, Sweden Received February 4, 1997X

In order to better understand the skin sensitization mechanism of allylic hydroperoxides, linalyl hydroperoxide (1) and several of its potential rearrangement productssepoxylinalool (2), epoxynerol (3), epoxygeraniol (4), and furan (5) and pyran (6) derivativesswere synthesized. The sensitizing properties of these molecules have been screened on mice using the local lymph node assay (LLNA) and further evaluated on guinea pigs using the Freund’s complete adjuvant test (FCAT). Linalyl hydroperoxide (1) and linalyl epoxide (2) were found to be sensitizers, while the other compounds were classified as mild sensitizers or nonsensitizers. In the guinea pigs, no cross-reactions were observed between skin sensitizers 1 and 2. Radical-trapping experiments were carried out on linalyl hydroperoxide (1) using TTBP as trapping agent and Fe3+-TPP as radical inducer. The major reaction taking place is the formation of a furan ring by intramolecular reaction of the oxygen-centered radical with the isoprenyl double bond with the formation of a tertiary radical. Reaction of this intermediate with radicals derived from TTBP gave compounds 10a,b in 25% yield. The second important reaction, accounting for 14%, is taking place on the allylic double bond with the formation of a less stable primary radical which is not trapped by a TTBP-derived radical but by a hydroxy radical to give a mixture of epoxides 3 and 4. These results are in favor of the formation of a carbon-centered reactive radical as intermediate in the skin sensitization to linalyl hydroperoxide.

Introduction Many functional groups have electrophilic properties and are able to react with various nucleophiles on proteins to form covalent bonds. This is probably the major mechanism of antigen formation involved in the allergic contact dermatitis (ACD)1 process (1). Lysine and cysteine are the nucleophilic targets most often cited, but other amino acids containing nucleophilic heteroatoms, for example, histidine, methionine, and tyrosine, can also react with electrophiles. It seems increasingly obvious that each hapten has its own chemical reactivity pattern, even if the effect of this chemical “selectivity” on the processing of hapten-protein conjugates by antigenpresenting cells and the selection and/or activation of T-cells is not known. Thus 2,4-dinitrobenzene (DNCB) which reacts readily with lysine residues (2) and methyl alkanesulfonates which react with histidine residues (3) are both strong sensitizers, while 2,4-dinitrothiocyanatobenzene (DNTB) which reacts with thiol groups (4) has been considered as a tolerogen. In recent years, radical mechanisms have gained increased prominence in the discussion of hapten-protein binding (5). Our previous * Address correspondence to this author. † Clinique Dermatologique. ‡ National Institute for Working Life. X Abstract published in Advance ACS Abstracts, August 15, 1997. 1 Abbreviations: allergic contact dermatitis, ACD; correlation and nuclear Overhauser spectroscopy, CONOESY; 2,4-dinitrochlorobenzene, DNCB; 2,4-dinitrothiocyanatobenzene, DNTB; dimethylformamide, DMF; Freund’s complete adjuvant, FCA; Freund’s complete adjuvant test, FCAT; heteronuclear multiple band correlation, HMBC; heteronuclear multiquantum correlation, HMQC; local lymph node assay, LLNA; m-chloroperbenzoic acid, m-CPBA; nuclear Overhauser effect, NOE; tritertiobutylphenol, TTBP; Fe3+-tetraphenylporphyrine, Fe3+-TPP.

S0893-228x(97)00014-3 CCC: $14.00

studies indicate that radical reactions could be important for haptens containing hydroperoxide groups (6-8). One might expect such molecules to react with totally different amino acids and thus generate different T-epitopes. Understanding of mechanisms leading to the formation of radicals and their subsequent reaction with proteins could open new insights in our understanding of antigen formation. In the course of our investigations on allylic hydroperoxides as potential sources for radical intermediates, we have been interested in linalyl hydroperoxide. This molecule, which could be one of the oxidation product of linalool (9) and thus account for the sensitizing potential of ylang ylang (Cananga odorata Hook.f. & Thoms.) or lavender (Lavendula latifolia (L.) Vill.) oils, offers many possibilities for the formation of radical intermediates of various stabilities and lifetimes. Alkoxy or peroxy radicals, generated from peroxides in the presence of metal traces such as Fe(II) or Fe(III), are known to react with double bonds to form a variety of structures depending on the geometry of the molecule and the stability of the intermediates. Thus, noncyclic allylic hydroperoxides such as fatty acid hydroperoxides form mainly hydroxy epoxides (10, 11) when treated with Fe(II). The presence of two double bonds of different reactivity in linalyl hydroperoxide could lead to the formation of several compounds (Chart 1). We report the synthesis of linalyl hydroperoxide and several of its potential rearrangement products. The sensitizing properties of these molecules have been screened on mice using the local lymph node assay (LLNA) and further evaluated on guinea pigs using the Freund’s complete adjuvant test (FCAT). The rearrangement of linalyl hydroperoxide has been further investi© 1997 American Chemical Society

988 Chem. Res. Toxicol., Vol. 10, No. 9, 1997 Chart 1. Potential Rearrangement Compounds from Linalyl Hydroperoxide (1)

gated using radical-trapping agents and isolation and characterization of intermediates.

Materials and Methods Hazardous Materials: Skin contact with hydroperoxides must be avoided. As sensitizing substances, these compounds must be handled with care. Materials. 1H and 13C NMR spectra were recorded on Bruker AC200 and AM500 MHz spectrometers in CDCl3 unless otherwise specified. Chemical shifts are reported in ppm (δ) with respect to TMS, and CHCl3 was used as internal standard (δ ) 7.27 ppm). Multiplicities are indicated by s (singlet), d (doublet), t (triplet), m (multiplet). Infrared spectra were obtained on a Perkin-Elmer FT-IR 1600 spectrometer; peaks are reported in reciprocal centimeters. Melting points were determined on a Buchi Tottoli 510 apparatus and are uncorrected. The HPLC analyses were performed on a Waters Associates apparatus equipped with a model M680 automated gradient controller and two M510 pumps. The solutions were analyzed and compounds separated on reversed-phase columns (C18, Nucleosil 5 µm, Interchrom, 4.6 mm; C18, RSil 5 µm, Interchrom, 10 mm, respectively), and the effluent absorption was monitored by an M686 tunable absorbance detector at 255 nm. Chemicals. Dried solvents were freshly distilled before use. Tetrahydrofuran and diethyl ether were distilled from sodium benzophenone. Dichloromethane was dried over P2O5 before distillation. All air- or moisture-sensitive reactions were conducted in flame-dried glassware under an atmosphere of dry argon. Chromatographic purifications were conducted on silica gel columns according to the flash chromatography technique. Freund’s complete adjuvant (FCA) was purchased from Difco (Detroit, MI). Olive oil was purchased from Aldrich. All chemicals were purified prior to animal testing and gave satisfactory microanalyses except for hydroperoxide 1 for which a difference of 0.6% was found between calculated and measured values. Linalyl Hydroperoxide (1). To a solution of hydrogen peroxide (100 mL, 35%) and sulfuric acid (2 mL) was added a

Bezard et al. solution of linalool (1 g, 6.48 mmol) in pentane (30 mL). The reaction mixture was vigorously stirred for 4 days at 0 °C and then extracted with hexane (3 × 100 mL). Combined organic layers were washed with brine (100 mL), dried over MgSO4, filtered, and concentrated under reduced pressure to give the crude hydroperoxide which was purified by chromatography over silica gel (hexane-ether, 15%) to give 143 mg (0.84 mmol, 13% yield) of 1 as a yellow liquid: 1H NMR (CDCl3) δ 1.34 (s, 3H, CH3-C-OOH), 1.55-1.72 (m, 2H, -CH2-C-OOH), 1.61 (bs, 3H, CH3-CdCH-), 1.68 (bs, 3H, CH3-CdCH-), 1.95-2.10 (m, 2H, Me2CdCH-CH2-), 5.12 (tsept, 1H, Me2CdCH-, J ) 7.1, 1.4 Hz), 5.24 (A part of an ABX system, 1H, CHdCH-C-OOH, JAX ) 10.8 Hz, JAB ) 1.2 Hz), 5.25 (B part of an ABX system, 1H, CHdCHC-OOH, JBX ) 17.9 Hz, JAB ) 1.2 Hz), 5.90 (X part of an ABX system, 1H, CH2dCH-C-OOH, JAX ) 10.8 Hz, JBX ) 17.9 Hz), 7.32 (s, 1H, -OOH); 13C NMR (CDCl3) δ 17.6, 21.1, 22.3, 25.6, 37.3, 84.7, 115.6, 124.1, 132.0, 141.0; IR (neat) ν 3412 (O-H). General Epoxidation Procedure. To anhydrous dichloromethane (20 mL) were added Ti(O-i-Pr)4 (193 µL, 0.65 mmol), a solution of anhydrous t-BuO2H in toluene (6.5 mL, 3 M, 19.5 mmol), and the allylic alcohol (2 g, 12.96 mmol). The reaction mixture was stirred at -20 °C for several hours, hydrolyzed with water (5 mL) and brine (5 mL), vigorously stirred at room temperature for 30 min, and extracted with dichloromethane (3 × 80 mL). Combined organic layers were dried over MgSO4, filtered, and concentrated under reduced pressure to give a crude epoxide which was purified by column chromatography over silica gel. Epoxylinalool (2). Starting from linalool and after a reaction time of 24 h, the crude epoxide was purified on column chromatography over silica gel (hexane-AcOEt, 20%) to give 883 mg (5.19 mmol, 40% yield) of 2 as a yellowish liquid (42/58 diastereomeric mixture): 1H NMR (CDCl3) δ 1.19 (s, CH3-COH), 1.31 (s, CH3-C-OH), 1.50-1.75 (m, 2H, -CH2-C-OH), 1.62 (bs, 3H, CH3-CdCH-), 1.68 (bs, 3H, CH3-CdCH-), 1.83 (s, 1H, -OH), 1.98-2.21 (m, 2H, Me2CdCH-CH2-), 2.65-3.00 (m, 3H, CH2-CH-C-OH), 5.05-5.20 (m, 1H, Me2CdCH-); 13C NMR (CDCl3) δ 17.6, 22.0, 22.2, 22.8, 25.6, 26.0, 38.7, 41.2, 43.3, 44.2, 57.6, 57.8, 69.2, 69.3, 124.0, 124.2, 131.8, 131.9; IR (neat) ν 3456 (O-H). Anal. Calcd for C10H18O2: C, 70.55; H, 10.65. Found: C, 70.30; H, 10.71. Epoxynerol (3). Starting from nerol and after a reaction time of 3 h, the crude epoxide was purified on column chromatography over silica gel (hexane-AcOEt, 35%) to give 2.10 g (12.33 mmol, 95% yield) of 3 as a colorless liquid: 1H NMR (CDCl3) δ 1.32 (s, 3H, CH3-C-O-), 1.37-1.75 (m, 2H, -CH2-CH2C-O-), 1.61 (bs, 3H, CH3-CdCH-), 1.67 (bs, 3H, CH3-CdCH-), 1.96-2.17 (m, 2H, Me2CdCH-CH2-), 2.18-2.30 (m, 1H, -OH), 2.95 (X part of an ABX system, 1H, HO-CH2-CH-O-, JBX ) 6.9 Hz, JAX ) 4.2 Hz), 3.63 (B part of an ABX system, 1H, -CH-OH, JAB ) 12.1 Hz, JBX ) 6.9 Hz), 3.80 (A part of an ABX system, 1H, -CH-OH, JAB ) 12.1 Hz, JAX ) 4.2 Hz), 5.09 (tsept, 1H, Me2CdCH-, J ) 7.1, 1.5 Hz); 13C NMR (CDCl3) δ 17.6, 22.2, 24.2, 25.6, 33.1, 61.2, 61.5, 64.3, 123.3, 132.4; IR (neat) ν 3420 (O-H). Anal. Calcd for C10H18O2: C, 70.55; H, 10.65. Found: C, 70.28; H, 10.71. Epoxygeraniol (4). Starting from geraniol and after a reaction time of 3 h, the crude epoxide was purified on column chromatography over silica gel (hexane-AcOEt, 35%) to give 2.16 g (12.69 mmol, 98% yield) of 4 as a colorless liquid: 1H NMR (CDCl3) δ 1.28 (s, 3H, CH3-C-O-), 1.36-1.75 (m, 2H, -CH2CH2-C-O-), 1.60 (bs, 3H, CH3-CdCH-), 1.67 (bs, 3H, CH3-CdCH), 1.94-2.17 (m, 2H, Me2CdCH-CH2-), 2.19-2.41 (m, 1H, -OH), 2.95 (X part of an ABX system, 1H, HO-CH2-CH-O-, JBX ) 6.6 Hz, JAX ) 4.2 Hz), 3.65 (B part of an ABX system, 1H, -CH-OH, JAB ) 12.1 Hz, JBX ) 6.6 Hz), 3.81 (A part of an ABX system, 1H, -CH-OH, JAB ) 12.1 Hz, JAX ) 4.2 Hz), 5.06 (tsept, 1H, Me2CdCH-, J ) 7.2, 1.5 Hz); 13C NMR (CDCl3) δ 16.8, 17.6, 23.7, 25.7, 38.5, 61.2, 61.4, 63.0, 123.4, 132.2; IR (neat) ν 3415 (O-H). Anal. Calcd for C10H18O2: C, 70.55; H, 10.65. Found: C, 70.21; H, 10.75. Synthesis of cis- and trans-2-Methyl-2-vinyl-5-(1′-hydroxy-1′-methylethyl)tetrahydrofuran (5a,b) and cis- and

Linalyl Hydroperoxide trans-3-Hydroxy-2,2,6-trimethyl-6-vinyltetrahydropyran (6a,b). To linalool (10 g, 64.83 mmol) in anhydrous dichloromethane (25 mL) was added, at 0 °C and over a period of 4 h, a solution of m-CPBA (22.4 g, 71.3 mmol) in dichloromethane (100 mL). The reaction mixture was stirred at 0 °C for 4 h, neutralized with a saturated solution of NaHCO3, and extracted with dichloromethane (3 × 100 mL). Combined organic layers were washed with brine (200 mL), dried over MgSO4, filtered, and concentrated under reduced pressure. The crude mixture was purified by column chromatography over silica gel (hexane-AcOEt, 20%) to give 8.72 g (51.22 mmol, 79% yield) of 5 and 1.67 g (9.81 mmol, 15.1% yield) of 6. 5a: yellow liquid; 1H NMR (CDCl3) δ 1.13 (s, 3H, -CH3), 1.22 (s, 3H, -CH3), 1.31 (s, 3H, -CH3), 1.65-1.98 (m, 4H, -CH2-), 2.08 (bs, 1H, -OH), 3.79-3.92 (m, 1H, -CH-), 4.99 (A part of an ABX system, 1H, CHdCH-C-O-, JAX ) 10.6 Hz, JAB ) 1.4 Hz), 5.18 (B part of an ABX system, 1H, CHdCH-C-O-, JBX ) 17.5 Hz, JAB ) 1.4 Hz), 5.97 (X part of an ABX system, 1H, CH2dCHC-O-, JAX ) 10.6 Hz, JBX ) 17.5 Hz); 13C NMR (CDCl3) δ 24.3, 26.0, 26.5, 27.3, 37.9, 71.2, 82.7, 85.5, 111.5, 144.3; IR (neat) ν 3562, 3459 (O-H). Anal. Calcd for C10H18O2: C, 70.55; H, 10.65. Found: C, 70.27; H, 10.72. 5b: yellow liquid; 1H NMR (CDCl3) δ 1.12 (s, 3H, -CH3), 1.23 (s, 3H, -CH3), 1.31 (s, 3H, -CH3), 1.66-1.98 (m, 4H, -CH2-), 2.16 (bs, 1H, -OH), 3.74-3.85 (m, 1H, -CH-), 5.00 (A part of an ABX system, 1H, CHdCH-C-O-, JAX ) 10.6 Hz, JAB ) 1.6 Hz), 5.19 (B part of an ABX system, 1H, CH)CH-C-O-, JBX ) 17.2 Hz, JAB ) 1.6 Hz), 5.88 (X part of an ABX system, 1H, CH2dCHC-O-, JAX ) 10.6 Hz, JBX ) 17.2 Hz); 13C NMR (CDCl3) δ 24.2, 26.3, 26.8, 27.1, 37.4, 71.1, 83.0, 85.5, 111.3, 143.7; IR (neat) ν 3455 (O-H). Anal. Calcd for C10H18O2: C, 70.55; H, 10.65. Found: C, 70.22; H, 10.71. 6a: white solid, mp ) 66-67 °C; 1H NMR (CDCl3) δ 1.16 (bs, 6H, -CH3), 1.24 (s, 3H, -CH3), 1.42 (d, 1H, -OH, J ) 5.9 Hz), 1.47-1.80 (m, 4H, -CH2-), 3.35-3.52 (m, 1H, -CH-), 4.97 (A part of an ABX system, 1H, CHdCH-C-O-, JAX ) 10.8 Hz, JAB ) 0.8 Hz), 4.98 (B part of an ABX system, 1H, CHdCH-C-O-, JBX ) 18.0 Hz, JAB ) 0.8 Hz), 5.96 (dedoubled X part of an ABX system, 1H, CH2dCH-C-O-, JAX ) 10.8 Hz, JBX ) 18.0 Hz, J ) 1.0 Hz); 13C NMR (CDCl3) δ 20.7, 25.6, 29.5, 31.6, 32.5, 73.4, 74.8, 75.9, 110.6, 146.2; IR (CHCl3) ν 3621, 3410 (O-H). Anal. Calcd for C10H18O2: C, 70.55; H, 10.65. Found: C, 70.51; H, 10.71. 6b: yellow liquid; 1H NMR (CDCl3) δ 1.21 (s, 6H, -CH3), 1.23 (s, 3H, -CH3), 1.62-2.07 (m, 5H, -CH2- + -OH), 3.36-3.48 (m, 1H, -CH-), 4.96 (A part of an ABX system, 1H, CHdCH-C-O-, JAX ) 11.1 Hz, JAB ) 1.0 Hz), 5.01 (B part of an ABX system, 1H, CHdCH-C-O-, JBX ) 17.7 Hz, JAB ) 1.0 Hz), 5.93 (dedoubled X part of an ABX system, 1H, CH2dCH-C-O-, JAX ) 11.1 Hz, JBX ) 17.7 Hz, J ) 0.7 Hz); 13C NMR (CDCl3) δ 24.2, 26.3, 27.2, 27.7, 30.7, 71.2, 73.5, 75.2, 110.5, 146.8; IR (neat) ν 3435 (OH). Anal. Calcd for C10H18O2: C, 70.55; H, 10.65. Found: C, 70.29; H, 10.65. Trapping Experiment. To a solution of linalyl hydroperoxide (1) (600 mg, 3.52 mmol) in degassed dichloromethane (30 mL) were added tritertiobutylphenol (TTBP) (4.62 g, 17.62 mmol) and Fe3+-tetraphenylporphyrine (7 mg, 9.86 µmol). The reaction mixture was stirred for 1.5 h, hydrolyzed with water (30 mL), and extracted with dichloromethane (3 × 80 mL). Combined organic layers were dried over MgSO4, filtered on Celite, and concentrated under reduced pressure. Excess of TTBP was removed by column chromatography over silica gel (hexane-toluene, 10%, and then hexane-dichloromethane, 10%). The mixture was then fractionated by column chromatography on silica gel (hexane-Et2O, 15%, and then hexaneAcOEt, 35%) into polar and nonpolar compounds. Further purification on the nonpolar fraction was carried out using HPLC (C18 reversed-phase, gradient of H2O/CH3CN, 80-95%) to give a nonseparable mixture of diastereomers in a 76/24 ratio (25% yield) identified as trans- and cis-2-methyl-5-[1′-(1′′,3′′,5′′tri-tert-butyl-4′′-oxo-2′′,5′′-cyclohexadienyl)-1′-methylethyl]-2-vinyltetrahydrofuran (10a,b). The 500 MHz 1H NMR and 125 MHz 13C NMR data are listed in Tables 1 and 2. Other data:

Chem. Res. Toxicol., Vol. 10, No. 9, 1997 989 Table 1. 10a (δ, ppm) 0.77 1.09 1.08 1.24

1H

NMR Data of Trapping Products 10a,b

10b (δ, ppm) 0.79 1.07 1.08 1.30

1.24 1.26 1.26 1.59-1.90 4.34-4.42 4.97 4.92 5.16

5.15

5.92

5.88

6.82

6.83

7.34

7.30

Table 2.

13C

multiplicities s, 3H, CH3-C-CH3 s, 3H, CH3-C-CH3 s, 9H, t-Bu s, 3H, CH3-C-CHdCH2 s, 9H, t-Bu s, 9H, t-Bu m, 4H, -CH2-CH2m, 1H, -CH-CH2-CH2A part of an ABX system, 1H, CHdCH-C 10a: JAX ) 10.8 Hz, JAB ) 1.5 Hz 10b: JAX ) 10.7 Hz, JAB ) 1.5 Hz B part of an ABX system, 1H, CHdCH-C 10a: JBX ) 17.4 Hz, JAB ) 1.5 Hz 10b: JBX ) 17.3 Hz, JAB ) 1.5 Hz X part of an ABX system, 1H, CH2dCH-C 10a: JBX ) 17.4 Hz, JAX ) 10.8 Hz 10b: JBX ) 17.3 Hz, JAX ) 10.7 Hz d, 1H, -CHdC-CdO 10a: J ) 3.1 Hz 10b: J ) 3.0 Hz d, 1H, -CHdC-CdO 10a: J ) 3.1 Hz 10b: J ) 3.0 Hz

NMR Data of Trapping Products 10a,b

10a (δ, ppm)

10b (δ, ppm) 186.3

147.3

147.1 146.2

145.4 144.7 144.5 111.2 82.9 82.8

145.6 144.9 144.7 110.8 82.7 82.4 52.5

46.0 39.4 36.5

45.9 39.5 37.5 35.2

30.1

30.0 29.5 29.6

28.6 27.1 24.4 20.3

28.9 26.9 24.8 20.5

signals CdO -CHdC-CdO -CHdC-CdO -CHdC-CdO -CHdC-CdO CH2dCH-C CH2dCH-C -CH-CH2-CH2CH3-C-CHdCH2 (CH3)3C-C-CHdC CH3-C-CH3 (CH3)3C-C-CHdC -CH-CH2-CH2(CH3)3C-CdCH(CH3)3C-C-CHdC (CH3)3C-CdCH(CH3)3C-CdCH-CH-CH2-CH2CH3-C-CHdCH2 CH3-C-CH3 CH3-C-CH3

IR (CHCl3) ν 1652, 1626 (CdO); UV (CH3CN) λ 250 nm ( ) 9450). Anal. Calcd for C28H46O2: C, 81.10; H, 11.18. Found: C, 81.10; H, 11.27. The polar fraction was found to contain linalool 7 (4%) and epoxides 3 (3%) and 4 (11%) characterized by comparison with authentical samples. Yields correspond to pure isolated compounds, and the material balance is mainly composed of minor unidentified nonpolar compounds and some polymeric material. Sensitization Experiments in Mice. The LLNA was carried out as essentially recommended by Kimber and Basketter (12). Female mice (CBA/Ca strain, 6-10 weeks old), in groups of four, received 25 µL of the test chemical dissolved in dimethylformamide (DMF) on the dorsum of both ears for 3 consecutive days. Test solutions were made fresh each day and applied within 30 min. Compounds 1-6 were tested in three different concentrations: 1%, 3%, and 9% (w/w), respectively. Control mice were treated with an equal volume of DMF alone. Five days after the first treatment, all mice were injected intravenously through the tail vein with 20 µCi of [3H]thymidine (specific activity 2 Ci/mmol; Amersham International, Amersham, U.K.) in 250 µL of phosphate-buffered saline. After 5 h, the mice were sacrificed, the draining auricular lymph nodes were excised and pooled for each group, and a single-cell suspension of lymph node cells was prepared. After washing and precipitation with trichloroacetic acid, thymidine incorporation was determined by β-scintillation counting (13). Results

990 Chem. Res. Toxicol., Vol. 10, No. 9, 1997 Scheme 1. Preparation of Linalyl Hydroperoxide (1)

are expressed as mean dpm (decompositions/minute)/lymph node for each experimental group and as stimulation index (SI), i.e., test group value/control group value. Sensitization Experiments in Guinea Pigs. The sensitization experiments were performed in outbred female albino Dunkin-Hartley guinea pigs obtained from AB Sahlins Fo¨rso¨ksdjursfarm (Malmo¨, Sweden). The animals were kept on a standard diet from EWOS AB (So¨derta¨lje, Sweden), and their average weight was 300-350 g. The experiments were carried out according to the FCAT (14). Closed challenge was performed according to previous experience (15). For induction the animals received intradermal injections of 0.1 mL of the test substance in FCA/water emulsion in the upper back on days 0, 6, and 10. The controls received FCA/water emulsion only. Challenge testing was performed on day 21. The test material (15 µL of each dose) was applied on the shaved flanks for 24 h, using Finn chambers (alumina chambers, 7 mm i.d.). The reactions were assessed at 48 and 72 h after application. The minimum criterion for a positive reaction was a confluent erythema. The animals were randomized into five groups. Group A (n ) 10) was induced with the hydroperoxide 1, group B (n ) 9) with epoxide 2, group C (n ) 10) with epoxide 3, group D (n ) 10) with epoxide 4. The fifth group (n ) 9) was a sham-treated control group. To determine the minimum irritating concentration for induction, three other animals were injected with epoxide 4 in concentrations of 10%, 3.3%, and 1.1% in olive oil. The concentration for inductions was determined to be 4%, which corresponds to 0.24 M. The same concentration was used for all test substances to get comparable results. Challenge testing was performed in all groups with all substances at the same time using a blind randomized application system. Before testing the maximum nonirritating concentration was determined. One intradermal injection of FCA was given to each of three guinea pigs, which were tested with epoxides 2 and 4 in concentrations of 9%, 3.3%, 1.1%, and 0.33% in olive oil. No irritation was seen, and 3.3% in olive oil and concentrations below were chosen for all substances. A vehicle control with olive oil was applied on all animals.

Results and Discussion Synthesis. The most classical method for preparation of tertiary allylic hydroperoxides is the quenching with H2O2 of a carbocation generated either from a hydroxyl function under acidic conditions (16) or from halides treated with silver triflate (17). Linalool was thus treated under several acidic conditions such as H2SO4, or trifluoroacetic acid in a biphasic system (aqueous H2O2, pentane), or Amberlyst 15 in anhydrous H2O2 in Et2O (18). In all cases the reaction was found to be nonselective probably due to the presence of an electron rich terminal double bond. The linalyl hydroperoxide has never been isolated in more than 16% yield (Scheme 1), but the same reaction carried out on a derivative lacking the terminal insaturation gave the hydroperoxide in high yields. The photoreaction of dihydromyrcene with dyephotosensitized oxygen, which could also potentially lead to the expected hydroperoxide, was known to be nonspecific (19) and therefore not tried. Epoxides 2-4 were prepared according to the Sharpless epoxidation method (20) starting from linalool, nerol, and geraniol, respectively (Scheme 2). Allylic alcohols

Bezard et al. Scheme 2. Preparation of Hydroxy Epoxides 2-4

Scheme 3. Preparation of Furan and Pyran Derivatives 5a,b and 6a,b

were treated with an anhydrous solution of tert-butyl hydroperoxide in dichloromethane in the presence of titanium isopropoxide as catalyst to give the corresponding epoxides in high yields (over 96%) except for the epoxylinalool which was obtained in 40% yield. This latter result can probably be explained by the steric hindrance of the tertiary alcohol. Analytical data were in accordance with the one expected from the literature. Furan and pyran derivatives were prepared according to the literature (21) by epoxidation of linalool with m-CPBA at 0 °C for 4 h (Scheme 3). Furan derivatives 5a,b were predominantly formed (79% yield) together with pyran derivatives 6a,b (15% yield). Diastereomers were separated to have NMR reference datasassigned from the literature (21)sfor the identification of radicaltrapping compounds. Sensitization Experiments According to LLNA. The LLNA was used as a first screening test of the allergenic activity of hydroperoxide 1, epoxides 2-4, furan 5, and pyran 6. Results are shown in Table 3. The hydroperoxide 1 and epoxide 2 fulfill the present criteria for a chemical to be classified as a sensitizer in the LLNA, i.e., a dose dependent increase in proliferation and at least one test concentration inducing a 3-fold or greater increase (SI-value g 3) in isotope incorporation compared with vehicle-treated controls (12). Epoxides 3 and 4 as well as furan 5 and pyran 6 failed to induce any substantial proliferation in the concentration interval

Linalyl Hydroperoxide

Chem. Res. Toxicol., Vol. 10, No. 9, 1997 991

Table 3. Cell Proliferation Induced by Hydroperoxide 1, Epoxides 2-4, and Compounds 5 and 6 in the LLNAa

chemical hydroperoxide 1

epoxide 2

epoxide 3

epoxide 4

furan 5

pyran 6

concn (w/w, %)

[3H]thymidine incorpn (dpm/node)

control 1.0 3.0 9.0 control 1.0 3.0 9.0 control 1.0 3.0 9.0 control 1.0 3.0 9.0 control 1.0 3.0 9.0 control 1.0 3.0 9.0

215 467 2976 3641 215 296 385 699 341 298 476 451 341 365 348 368 132 139 189 279 132 177 222 190

stimulation index 2.2 13.8 16.9 1.4 1.8 3.2 0.9 1.4 1.3 1.1 1.0 1.1 1.1 1.4 2.1 1.3 1.7 1.4

a Groups of mice (n ) 4) received 25 µL of the test chemical dissolved in vehicle (DMF) in the concentrations indicated, on the dorsum of both ears daily for 3 consecutive days. Control animals were treated in the same way with the vehicle alone. All mice were injected intravenously 5 days after the first treatment, with 250 µL of PBS containing 20 µCi of [3H]tymidine. Five hours later, draining auricular lymph nodes were excised and pooled for each group, and a single-cell suspension of lymph node cells was prepared. The thymidine incorporation was measured by β-scintillation counting. The increase in thymidine incorporation relative to vehicle-treated controls was derived for each experimental group and recorded as a stimulation index.

tested. It has been proposed that the lowest concentration required to induce a stimulation index of 3 or greater may be used to rank the skin-sensitizing potential of chemicals (12, 22). Accordingly, ranking the derivatives for skin-sensitizing potency reveals the following order in decreasing potency: hydroperoxide 1 >>> epoxide 2 > epoxide 3 ) epoxide 4 ) furan 5 ) pyran 6. However, it cannot be ruled out that different irritating properties of the compounds influence the results in the LLNA (13). Sensitization Experiments According to FCAT. To further study the sensitizing potential of the hydroperoxide 1 and the different epoxides 2-4, which are chemically the most potent reactive molecules, experiments in guinea pigs were performed. The results (Table 4) confirmed the findings of the LLNA: with 10/10 positive animals at a challenge concentration of 3.3% and 6/10 at a challenge concentration of 1.1%, linalyl hydroperoxide (1) should be considered as a strong sensitizer. The same conclusion also applies to epoxide 2 with 9/9 positive animals at a challenge concentration of 3% and 7/9 at a challenge concentration of 1.1%. This is in accordance with the sensitizing capacity seen for other hydroperoxides and epoxides such as the hydroperoxide (23) and epoxides (7) of abietic acid as well as limonene hydroperoxide (8) and limonene oxide (24). Epoxide 4 gave 5/10 positive animals at a challenge concentration of 3%, and this epoxide should be considered as a mild sensitizer. At the concentration tested no irritation was seen to any of the derivatives and no sensitizing activity was detected for epoxide 3. Epoxide 4 was not detected as a sensitizer in the LLNA. This may be due to the fact

that the highest concentration tested in the LLNA was 9%, and the possibility exists that positive reactions will occur at higher concentrations. However, this was not tested due to the limited amount available of the compounds and that the outcome would not influence the conclusions made. Cross-challenge experiments were also performed, and it is interesting to note that no cross-reactions were observed between animals sensitized to hydroperoxide 1 and epoxide 2, which clearly indicates the formation of two different epitopes. This is in accordance with our previous studies (8) where no cross-reactivity was observed between limonene hydroperoxide and limonene oxide. However, cross-reactivity was found in the study on 15-hydroperoxyabietic acid and epoxides of abietic acid (7). In the present study some statistically significant reactions were observed in animals sensitized to epoxide 4 when cross-challenged with either hydroperoxide 1 (6/ 10 at 72 h) and epoxide 2 (5/10 at 72 h). Chemical-Trapping Experiments. Radical-trapping experiments were carried out as reported in the literature (10) using TTBP as trapping agent and Fe3+TPP as radical inducer. The excess of TTBP was removed by chromatography over silica gel, and reaction compounds were split in two fractions (polar and nonpolar). The nonpolar fraction was further purified by semipreparative HPLC on a C18 reversed-phase column and structures were established by a combination of 1H and 13C NMR, 1H-13C correlation (HMQC and HMBC), and CONOESY experiments. A major peak was detected at 255 nm, corresponding to a nonseparable mixture of two diastereomers, 10a (19%) and 10b (6%), identified as trapped trans- and cis-furan derivatives, respectively (Scheme 4). The attribution of methyl groups, which was essential for the structure determination, was done using long-range 13C-1H correlation experiments, and the stereochemistry determination was based on observations of nuclear Overhauser effects (NOE). Thus a NOE between H-5 (4.39 ppm) and the methyl group at C-2 (1.30 ppm) allowed us to attribute the cis configuration to compound 10b. Interestingly, the trapping was found to occur through the para carbon-centered radical derived from TTBP and not through the oxygen-centered radical, leading to the formation of cyclohexadienone derivatives (IR data). The para linkage was evidenced by 13C-1H long-range analysis, both H-2′′ and H-6′′ protons being coupled with the dimethylated C-1′ carbon. The 4J coupling constant of 3 Hz between H-2′′ and H-6′′ protons, which could be surprising for such a para-substituted compound, is indeed in accordance with what is reported in the literature (25, 26) for analogous molecules. The polar fraction was found to be a mixture of linalool (7) (4%) and epoxides 3 (3%) and 4 (11%) (each compound was identified by comparison with the reference materials). Mechanistic Interpretations. The reaction of hydroperoxide 1 with Fe3+-TPP in the presence of TTBP leads to the formation of an alkoxy radical intermediate which is potentially able to abstract one hydrogen to give back the linalool (7) (4%) or further react with either the allylic or the isoprenyl double bond. The so-formed carbon-centered radicals might then further react and rearrange or be trapped by either the hydroxy radical generated during the “Fenton-like” reaction or one of the radicals derived from TTBP. Results of our trapping experiments suggest that the major reaction taking place is the formation of a furan ring by intramolecular

992 Chem. Res. Toxicol., Vol. 10, No. 9, 1997

Bezard et al.

Table 4. Sensitization Experiments in Guinea Pigs According to FCATa challenge material (% in olive oil) hydroperoxide 1 guinea pigs

epoxide 2

time (h)

3

1

0.3

48 72

10 10