Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
47, 93–98 (1999) Copyright © 1999 by the Society of Toxicology TOXICOLOGICAL SCIENCES Quantitative Structure-Activity Relationships for Nasal Pungency Thresholds of Volatile Organic Compounds K. M. Hau,* ,1 D. W. Connell,† and B. J. Richardson‡ *Health and Safety Office, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong; †Faculty of Environmental Sciences, Griffith University, Nathan, Queensland 4111, Australia; and ‡Department of Biology and Chemistry, City University of Hong Kong, 83 Tat Chee Avenue, Kowloon, Hong Kong Received November 15, 1997; accepted August 20, 1998 teristics that provide information on the potency of chemicals in causing nasal irritation. Nasal irritation data in terms of RD 50 and RD 0 have been obtained on animals (Alarie et al., 1995; Hansen and Nielsen, 1994a, b; Nielsen and Alarie, 1982; Nielsen et al., 1988; Nielsen and Kristiansen, 1988; Nielsen and Vinggaard, 1988; Nielsen and Yamagiwa, 1989; Silver et al., 1986), and an extensive database has been compiled (Schaper, 1993). It is, however, still uncertain as to how well these animal data correlate with human data. Nasal pungency thresholds have been collected in recent years on humans. Cometto-Muniz and Cain (1990, 1991, 1993, 1994a, b) have investigated the relationships between the human data and a number of physicochemical factors for a range of VOCs. More recently, Abraham et al. (1996) have developed a quantitative structure-activity relationship (QSAR) for the human data using a solvation equation. The levels of VOCs commonly encountered in the nonindustrial work environment (Brown et al., 1994) are often well below the known nasal irritation thresholds of most VOCs. Individual VOCs at typical indoor levels, which may not cause nasal irritation separately, but together can have an additive or even synergistic effect in causing nasal irritation. In order to understand the relationship between VOCs found in the indoor environment and nasal irritation, the combined effect of individual VOCs on nasal irritation needs to be elucidated. The development of QSARs for nasal irritation caused by VOCs is a useful tool in understanding the mechanism involved and thus the possible combined effect of VOCs. Also, QSARs allow the prediction of nasal irritation by compounds for which this characteristic has not been measured. The aim of the present paper is to develop a QSAR for human nasal pungency thresholds, utilizing a model developed originally for odor thresholds (Hau and Connell, 1998). A model was developed for describing the triggering of nasal pungency in humans, based on the partition of volatile organic compounds (VOCs) between the air phase and the biophase. Two partition parameters are used in the model: the water-air partition coefficient and the octanol-water partition coefficient. The model was validated using data from the literature, principally on alcohols, acetates and ketones. The model suggests that all test compounds, regardless of their chemical functional groups, bind to a common receptor site within the hydrophobic interior of the bilayer membrane of the trigeminal nerve endings. There is probably only a slight, non-specific interaction between the VOC molecule and the receptor molecule, whereas this type of non-specific interaction for the detection of odor is much stronger. In practical terms, the suggestion that all VOCs share a common irritation receptor site implies that nasal-pungency thresholds of individual VOCs may be additive. Quantitative structure-activity relationships (QSARs) for nasal-pungency thresholds were also developed from the model, which can be used to predict nasal-pungency thresholds of common VOCs. Although the present model does not offer additional precision over that of M. H. Abraham et al., 1996, Fundam. Appl. Toxicol. 31, 71–76, it requires fewer descriptors and offers a physiological basis to the QSAR. Another advantage of the present model is that it also provides a basis for comparison between the olfactory process and nasal pungency. Key Words: human nasal pungency thresholds; volatile organic compounds (VOCs); quantitative structure-activity relationships (QSARs); VOC receptor sites. Common volatile organic compounds (VOCs) in the indoor environment can be a major cause of nasal irritation among building occupants. It is believed by many researchers (Cain and Cometto-Muniz, 1995; Cometto-Muniz and Cain, 1992; Molhave, 1992; WHO, 1989) that nasal irritation caused by VOCs may contribute to the loosely defined sick-building syndrome (SBS). Nasal irritation is mediated through the trigeminal nerve, which inervates the nasal cavity and the surrounding facial area (Nielsen and Hansen, 1993; Silver, 1992). Nasal pungency thresholds of chemicals are useful charac1 METHODS Model development. The model developed was based on the behavior of common indoor VOCs that are relatively non-reactive compounds. These substances tend to interact with the biological system through reversible processes without inflicting permanent damage on biological tissues. The To whom correspondence should be addressed. 93 94 HAU, CONNELL, AND RICHARDSON model can be seen as involving a partition process of the airborne VOC molecules between the air phase, the mucus covering the nasal epithelium, and the cell membrane of the trigeminal nerve endings scattered throughout the epithelium. Once within the biophase (the bilayer cell membrane), the molecule combines with a receptor site that is commonly believed to be part of a protein molecule within the biophase (Nielsen, 1991). The partition process can be represented by the following: Kma Kbm @S#air 43 @S#mucus 43 @S#bio, where [S] air is the concentration of the VOC in the air, [S] mucus is the concentration of the VOC in the mucus, [S] bio is the concentration of the VOC in the biophase, K ma is the partition coefficient between the mucus and the air phase, and K bm is the partition coefficient between the biophase and the mucus. Once the VOC molecule is in the biophase, it can reversibly bind to a relevant chemoreceptor site, thus activating the receptor molecule. This process can be represented as follows: may vary from one type of compound-receptor complex to another. At the pungency threshold of a compound, Equation 1 becomes p5 @S#bio 1 @R# O ¢ ¡ @S2R#, where [R] is the concentration of unactivated receptor sites in the biophase relevant to the VOC, [S 2 R] is the concentration of the VOC-receptor complex in the biophase, and K A is the equilibrium constant of the reversible reaction. It is assumed that all partition processes and reversible reactions with the chemoreceptors reach equilibrium so rapidly that the non-equilibrium state need not be considered. Thus all concentration terms used below are equilibrium concentrations. This means that the following relationships occur: where p is the minimum proportion of the activated receptor sites necessary for the detection of pungency, and [NPT] is the nasal pungency threshold of the compound. It is further hypothesized that VOCs with a common chemical functional group bind to the same receptor site to give a response. Thus for a homologous series, p is a constant for all the members in that series. To make the relationship in Equation 2 useful, the two terms K A and K ba need to be expressed in terms of readily available physicochemical properties of a compound. By definition, Kba 5 Kbm p Kma. Kma 5 Kwa Kbm 5 a K mow Kbm 5 @S#bio/@S#mucus Kba 5 @S#bio/@S#air 5 $@S#bio/@S#mucus% p $@S#mucus/@S#air% 5 Kbm p Kma, where K ba is the partition coefficient between the biophase and the air phase. Therefore K A 5 [S 2 R]/{[S] bio p [R]}. This model is similar to that used in Molhave and Nielsen (1992) and Hansen and Nielsen (1994a), which is based on the classical receptor theory. An important principle of that theory is that the sensory response is related to the proportion of the activated receptor sites. This principle is also used in the present model: KA 5 b K now KA@S#bio@R# @S#bio 5 . @R# 1 KA@S#bio@R# @S#bio 1 1/KA Previously, it was shown that [S] bio equals [S] airK ba, thus: @S#airKba @S#air @S#bio 5 5 . @S#bio 1 1/KA @S#airKba 1 1/KA @S#air 1 $1/KAKba% (5) (6) where b and n are empirical constants for a homologous series. Combining Equations 2 to 6 gives p5 By substituting for [S 2 R] derived from the equation for K A above, then (4) where K ow is the partition coefficient between octanol and water, and a and m are empirical constants for a given receptor environment. The term K A may change from one homologous series to another, and with the lipophilicity of the members within a homologous series. Within a homologous series K A can be expressed as follows: @S 2 R# Proportion of the activated receptors 5 . @R# 1 @S 2 R% 5 (3) where K wa is the partition coefficient between water and air. It is also expected that the chemoreceptors are certain protein molecules inside the lipid bilayer membrane of the trigeminal nerve endings. As the biophase is lipophilic, the term K bm can then be expressed as a function of K ow (Hansch and Leo, 1979; Lyman et al., 1990). A general expression is Kma 5 @S#mucus/@S#air @S 2 R# (2) As the layer of mucus covering the nasal cavity is chemically similar to saline water at about pH 7.2 (Davies and Taylor, 1959), the partition of organic compounds in the mucus layer would be similar to that of water, and thus: KA @R# 1 @S 2 R# @NPT# , @NPT# 1 $1/KAKba% @NPT# . @NPT# 1 $1/ ab KwaK m1n ow % m1n Rearranging, then, [NPT]abK waK ow 5 p/(1 2 p) and [NPT]K wa 5 {p/ab 2(m1n) 2c (1 2 p)} K ow . Thus [NPT]K wa 5 gK ow , where g 5 {p/ab(1 2 p)}, and c 5 m 1 n. Taking the logarithms of both sides: log$@NPT#Kwa% 5 G 2 c log Kow (7) log@NPT# 5 G 2 log Kwa 2 c log Kow (8) or (1) It is hypothesized that a minimum proportion of the available receptor sites need to be occupied by the VOC molecules before pungency can be detected. This minimum proportion corresponds to the nasal pungency threshold, and it where G 5 log {p/ab(1 2 p)}, both G and c being constants for a homologous series. The model, as summarized in Equation 7, predicts that the plot of the logarithm of the product [NPT]K wa against the logarithm of K ow should give a 95 QSARs FOR NASAL PUNGENCY THRESHOLDS linear relationship for compounds with a common functional group binding to a common receptor, and thus at least for a homologous series. Model validation. The model represented by Equation 7 was validated by fitting the Equation with available data on nasal pungency thresholds, together with the other two physicochemical properties K wa and K ow, and evaluating the linear relationship between log {[NPT]K wa} and log K ow. NPT data published in Abraham et al. (1996) were used as this set is the only systematically obtained human data available. Validation was first carried out with three individual homologous series, aliphatic alcohols, acetates, and ketones. Results from validating the homologous series were then further investigated to see if the different series are related to each other. Values in units of part per million (ppm) by volume were used in this study. Data on partition coefficients. Two partition coefficients, K wa and K ow, are needed to validate the model. Values of K wa were either obtained directly from Amoore and Hautala (1983) or derived from the water solubilities, vapor pressures, or Henry’s Law constants in Howard (1989) or Verschueren (1996). Values of K ow were obtained from Howard (1989), Hansch and Leo (1979), BioByte (1997), or Lyman et al. (1990). Treatment of data. Data used for both K wa and K ow are dimensionless ratios based on concentration units of mass per unit volume (or mol/m 3) usually calculated for or determined at 25°C. It is therefore essential that the nasal pungency threshold value in Equation 7 is also based on units of mass per unit volume in order to make Equation 7 physically meaningful. The data of [NPT] from Abraham et al. (1996) are expressed in units of ppm (volume by volume). The ppm unit in the gas phase can be easily converted to mol/m 3 by dividing it by a constant factor under standard temperature and pressure without involving the use of individual molecular weights. It is therefore possible to directly use the concentration unit of ppm (v/v) for all the [NPT] values in validating Equation 7, and the conversion factor is incorporated in the constant term G of the equation. To validate the model expressed in Equations 7 or 8, linear regression analysis was carried out on the logarithm of [NPT]pK wa against the logarithm of K ow for the three homologous series. Equation 7 was used as the basis of the validation, as this equation allows ready evaluation of the linear relationship predicted by the model, and also estimation of the empirical constants G and c. NPT data for other VOCs published in Abraham et al. (1996) were then fitted into the resulting regression equations for further examination. RESULTS The data for the test compounds used for this study are summarized in Table 1. The regression lines for the three homologous series (alcohols, acetates and ketones) are shown in Figure 1. The corresponding linear regression equations are: Alcohols: log{[NPT]Kwa} 5 7.63 2 0.86 log Kow r 5 20.971, n 5 11 TABLE 1 Nasal Pungency Thresholds and Partition Coefficients of Test Compounds log NPT a Alcohols Methanol Ethanol Propanol Propan-2-ol Butan-1-ol Butan-2-ol t-Butanol Pentan-1-ol Hexan-1-ol Heptan-1-ol Octan-1-ol Acetates Methyl acetate Ethyl acetate n-propyl acetate n-butyl acetate s-butyl acetate t-butyl acetate n-pentyl acetate n-hexyl acetate n-heptyl acetate n-octyl acetate n-decyl acetate n-dodecyl acetate Ketones Acetone Pentan-2-one Heptan-2-one Nonan-2-one Others pyridine toluene ethylbenzene n-propylbenzene chlorobenzene octyne 4.53 3.908 3.494 4.259 3.199 3.975 4.516 3.205 2.621 2.322 1.988 Kwa b 5600 4800 3600 3000 2800 2400 2000 2060 1520 876 960 5.051 4.828 4.245 3.562 3.597 3.976 3.217 2.803 2.492 1.954 0.699 0.097 210 180 110 91 63 73 25 18.7 12.1 7.2 2.5 0.9 5.116 3.472 2.906 2.53 620 380 180 89 3.105 4.471 4.004 3.172 4.023 4.486 log Kow c 1700 3.7 2.9 3.6 6.4 0.302 20.77 20.31 0.25 0.05 0.88 0.61 0.35 1.56 2.03 2.72 2.97 0.18 0.73 1.23 1.82 1.51 1.38 2.26 2.83 3.36 3.88 4.94 6 20.24 0.91 1.98 3.14 0.65 2.73 3.15 3.57 2.84 3.6 a Log (Nasal pungency thresholds in ppm) obtained from Abraham et al. (1996). b Dimensionless partition coefficients between water and air. c Log (dimensionless partition coefficients between octanol and water). (9) Acetates: log{[NPT]Kwa} 5 7.72 2 1.29 log Kow r 5 20.996, n 5 12 (10) Ketones: log{[NPT]Kwa} 5 7.34 2 1.00 log Kow r 5 20.972, n 5 4. (11) As the slopes and intercepts of Equations 9 –11 are close to one another, the three sets of data were pooled and a linear regression analysis performed on the combined data set yielding the following linear regression equation: Combined set: log{[NPT]Kwa} 5 7.72 2 1.21 log Kow r 5 20.977, n 5 27. (12) As the individual data sets can be pooled without significantly lowering the correlation coefficient, the rest of the data published in Abraham et al. (1996) were also incorporated into the regression analysis. It was found that all data, with the exception of heptan-4-ol for which a reliable value of K wa is not available, published in Abraham et al. (1996), can be fitted to a linear regression line with a high correlation coefficient as reflected in Equation 13. 96 HAU, CONNELL, AND RICHARDSON FIG. 1. Plot of Log {[NPT]K wa} vs Log K ow and the regression lines for the series of alcohols, acetates and ketones in Abraham et al. (1996). NPT refers to the nasal pungency thresholds (in ppm). K wa is the partition coefficient between water and air, while K ow is the partition coefficient between octanol and water. Combined set plus others: log{[NPT]Kwa} 5 7.69 2 1.16 log Kow r 5 20.971, n 5 33. (13) Figure 2 shows the plot of all test data and the regression lines represented by Equations 12 and 13. DISCUSSION Evaluation of the Model and Mechanisms of Action Equations 9 to 11 and their regression lines in Figure 1 show that the logarithm of the product of [NPT] and K wa is linearly correlated with the logarithm of K ow for the three homologous series with high-correlation coefficients. These linear relationships are as predicted by the present model and are well described by Equation 7 which is based on a partition process of the VOC molecules between the air phase and the biophase. The linear relationship for each test series suggests the hypothesis (i.e., that different molecules with a common functional group have a similar partition behavior, bind to the same receptor site, and need to occupy the same minimum proportion (p) of the sites to trigger a response) appears to be valid. While the high correlation coefficients indicate the model described by Equation 7 is reasonable, the slopes of the regression lines give other important information. Franke (1984) described a general physicochemical model for binding of drug molecules to biological receptors, which provides insight into the property of the environment of binding sites. The model, when applied to sensory irritation (Nielsen et al., 1990), suggests that the slope of the regression line should be approximately unity, if the binding site is totally within a hydrophobic environment, or the slope should be approximately 0.5, if the binding site is covered with water so that part of the molecule is bound to the protein while the other portion is surrounded by water. The theory developed by Franke (1984) does not account for a slope greater than unity. The model in the present study, however, predicts that a slope actually greater than unity can occur if K ow influences the partition process as well as the binding equilibrium constant K A. The dependence of K A on K ow is a result of enhanced binding of the VOC molecule to the receptor site because of the interaction between the hydrophobic portion of the VOC molecule and the receptor protein. The slope of the regression line is thus a measure of the hydrophobicity of the environment surrounding the receptor site plus the hydrophobic interaction between the VOC molecule and the receptor molecule. The slopes for the three series are close to unity, with that for acetates slightly above unity. The near-unity values can be explained by one of the following two possibilities. The first one is that the relevant receptor sites for the test series are close to the interface between the lipid bi-layer membrane and the mucus layer, so that the biophase in which binding occurs is more polar than octanol, and thus the empirical constant m in Equation 5 is significantly smaller than 1. However, the hydrophobic interaction between the VOC molecule and the receptor molecule is such that K A in Equation 6 is positively related to K ow. The two terms m and n combine to give a slope which could be close to or slightly above unity. The second possibility is that the receptor site is inside the hydrophobic environment so that m is near unity, but the hydrophobic interaction between the receptor molecule and the VOC molecule is weak, so that K A is not strongly related to K ow. The three regression lines in Figure 1 are close and almost parallel to each other. When all the data from the three homologous series are pooled, the linear relationship is still ex- FIG. 2. Plot of Log {[NPT]K wa} vs Log K ow and the regression lines for combined data of alcohols, acetates and ketones, and for the entire set of compounds (combined set plus others) in Abraham et al. (1996). NPT refers to the nasal pungency thresholds (in ppm). K wa is the partition coefficient between water and air, while K ow is the partition coefficient between octanol and water. QSARs FOR NASAL PUNGENCY THRESHOLDS tremely good with a high correlation coefficient as reflected in Equation 12. This suggests that VOCs belonging to the 3 series may bind to the same receptor site or closely related receptor sites. Further incorporation of other unrelated VOCs in Abraham et al. (1996) into the regression analysis did not substantially change the outcome as represented in Equation 13. The very good linear relationship obtained suggests that all test compounds bind to either the same receptor site or closely related receptor sites regardless of their functionality. An overall slope of 1.16 for the regression line of the combined set, plus others having different functional groups, indicates that the common receptor site is probably within the hydrophobic interior of the lipid bilayer membrane of the nerve ending, and that there is a slight, non-specific hydrophobic interaction between the VOC molecule and the receptor molecule so that the combined influence of K ow is greater than a single order. Development of QSARs The model in the present study provides some understanding of the partition process and the receptor environment involved in triggering nasal irritation, and it also allows the development of quantitative structure-activity relationships (QSARs) for nasal irritation caused by VOCs. While the QSAR developed by Abraham et al. (1996) is based on a solvation equation utilizing four descriptors, much simpler QSARs based on the present model can be developed by rearranging Equations 9 to 11, and 13 as follows: Alcohols: log[NPT] 5 7.63 2 log Kwa 2 0.86 log Kow (14) Acetates: log[NPT] 5 7.72 2 log Kwa 2 1.29 log Kow (15) Ketones: log[NPT] 5 7.34 2 log Kwa 2 1.00 log Kow (16) All functional groups: log[NPT] 5 7.69 2 log Kwa 2 1.16 log Kow. (17) In order to examine how well the QSARs in Equations 14 –17 predict the nasal pungency thresholds of the test compounds, the predicted NPTs for the test compounds were worked out using the QSARs and compared with their corresponding observed values. Good linear relationships were obtained between the predicted and the observed values. The correlation coefficients for the linear relationships are 0.96, 0.98, 0.95, and 0.92, respectively, for the four QSARs represented by Equations 14 –17. For VOCs belonging to the alcohols, acetates and ketones families, the nasal pungency thresholds can, therefore, be accurately predicted using QSARs represented by Equations 14 –16. For other VOCs, the NPT can also be predicted as a first approximation using the generic QSAR represented by Equation 17. However, the applicability of the QSAR represented by Equation 17 to a wide range of VOCs requires further investigation. Comparison can also be made between the QSARs in the 97 present study and those for odor thresholds (Hau and Connell, 1998). The slopes of the regression lines for odor thresholds of alcohols, acetates and ketones are 1.65, 1.65, and 1.88 respectively. The degree of dependence on K ow is significantly higher for odor thresholds when compared to the current work on pungency. This higher degree of dependence is either related to the more hydrophobic receptor environment or due to the higher level of hydrophobic interaction between the VOC molecule and the olfactory receptor protein. This dependence also means that the sensitivity of humans to odor increases sharply as the size of the VOC molecule increases. The phenomenon can be seen as an adaptive feature in humans to compensate for the much lower volatility of larger molecules, as there is still a practical need for the human nose to detect odors of some relatively non-volatile compounds. On the other hand, there does not appear to be a need for humans to maintain a high sensitivity to irritation by large molecules. As the molecular size of a VOC increases, so odor becomes the dominant chemical sense, while nasal pungency diminishes in relative importance. Responses to Mixtures The suggestion that all test compounds (and possibly other relatively non-reactive VOCs) may bind to a common receptor site has an important implication in the understanding of nasal pungency thresholds for mixtures. With a common receptor site, nasal pungency thresholds for the individual components in a mixture may be additive, i.e., sub-threshold concentrations of individual compounds may become irritating in a mixture. Since sick-building syndrome (SBS) is believed to be associated with the sensory perception of mixtures of VOCs, various approaches to setting standards for VOCs have been proposed (Molhave and Nielsen, 1992) including the use of the concept of total volatile organic compounds (TVOC). Molhave and Nielsen (1992) have discussed the prospect of setting a nasal irritation based standard for VOC mixtures. This approach appears to be logical, as hundreds of VOCs may be present in the indoor environment which, though all may be below their respective nasal pungency thresholds, can add up to exert an irritation effect. However, this suggestion suffers the limitation that the concentrations of common VOCs in the indoor environment are usually in ppb levels, even in the so-called “Sick Buildings,” which are at least three orders of magnitude below their nasal pungency thresholds. It is therefore doubtful if the total concentrations of VOCs would be high enough to exert a combined irritation effect in those “Sick Buildings.” Instead, odor thresholds of common VOCs are closer to the usual concentrations in the indoor environment. It is therefore suggested that an odor-based VOC standard (Hau and Connell, 1998) would be more appropriate than an irritation-based standard. 98 HAU, CONNELL, AND RICHARDSON REFERENCES Abraham, M. H., Andonian-Haftvan, J., Cometto-Muniz, J. E., and Cain, W. S. (1996). An analysis of nasal irritation thresholds using a new solvation equation. Fundam. Appl. Toxicol. 31, 71–76. Alarie, Y., Nielsen, G. D., Andonian-Haftvan, J., and Abraham, M. H. (1995). Physicochemical properties of non-reactive volatile organic chemicals to estimate RD50: Alternative to animal studies. Toxicol. Appl. Pharmacol. 134, 92–99. Amoore, J. E., and Hautala, E. (1983). Odor as an aid to chemical safety: odor thresholds compared with threshold limit values and volatilities for 214 industrial chemicals in air and water dilution. J. Appl. Toxicol. 3, 272–230. BioByte Corp. (1997). Medchem Software 3.6. BioByte Corp., Claremont, CA. Brown, S. K., Sim, M. R., Abramson, M. J., and Gray, C. N. (1994). Concentrations of volatile organic compounds in indoor air—a review. Indoor Air 4, 123–134. Cain, W. S., and Cometto-Muniz, J. E. (1995). Irritation and odor as indicators of indoor pollution. Occupational Medicine: State of the Art Reviews 10, 133–145. Cometto-Muniz, J. E., and Cain, W. S. (1990). Thresholds for odor and nasal pungency. Physiol. Behav. 48, 719 –725. Cometto-Muniz, J. E., and Cain, W. S. (1991). Nasal pungency, odor, and eye irritation thresholds for homologous acetates. Pharmacol. Biochem. Behav. 39, 983–989. Cometto-Muniz, J. E., and Cain, W. S. (1992). Sensory irritation. Relation to indoor air pollution. Ann. NY. Acad. Sci. 641, 152–163. Cometto-Muniz, J. E., and Cain, W. S. (1993). Efficacy of volatile organic compounds in evoking nasal pungency and odor. Arch. Environ. Health 48, 309 –314. Cometto-Muniz, J. E., and Cain, W. S. (1994a). Perception of odor and nasal pungency from homologous series of volatile organic compounds. Indoor Air 4, 140 –145. Cometto-Muniz, J. E., and Cain, W. S. (1994b). Sensory reactions of nasal pungency and odor to volatile organic compounds: the alkylbenzenes. Am. Ind. Hyg. Assoc. J. 55, 811– 817. Davis, J. T., and Taylor, F. H. (1959). The role of adsorption and molecular morphology in olfaction: the calculation of olfactory thresholds. Biol. Bull. 117, 222–238. Franke, R. (1984) Theoretical drug design methods. Elsevier, Amsterdam. Hansch, C., and Leo, A. (1979) Substituent constants for correlation analysis in chemistry and biology. John Wiley and Sons, New York. Hansen, L. F., and Nielsen, G. D. (1994a). Sensory irritation and pulmonary irritation of n-methyl ketones: receptor activation mechanisms and relationships with threshold limit values. Arch. Toxicol. 68, 193–202. Hansen, L. F., and Nielsen, G. D. (1994b). Sensory irritation, pulmonary irritation and structure-activity relationships of alcohols. Toxicol. 88, 81–99. Hau, K. M., and Connell, D. W. (1998). Quantitative Structure-Activity Relationships (QSARs) for odor thresholds of volatile organic compounds (VOCs). Indoor Air 8, 23–33. Howard, P. H. (1989). Handbook of Environmental Fate and Exposure Data for Organic Chemicals. Lewis Publishers, Chelsea. Lyman, W. J., Reehl, W. F., and Rosenblatt, D. H. (1990). Handbook of Chemical Property Estimation Methods. American Chemical Society, Washington, DC. Molhave, L. (1992). Volatile organic compounds and the sick building syndrome. In Environmental toxicants: Human exposure and their health effects (M. Lippmann, Ed.), pp. 633– 646. Van Nostrand Reinhold, New York. Molhave, L., and Nielsen, G. D. (1992). Interpretation and limitations of the concept “Total volatile organic compounds” (TVOC) as an indicator of human responses to exposures of volatile organic compounds (VOC) in indoor air. Indoor Air 2, 65–77. Nielsen, G. D. (1991). Mechanisms of activation of the sensory irritant receptor by airborne chemicals. CRC Crit. Rev. Toxicol. 21, 183–208. Nielsen, G. D., and Alarie, Y. (1982). Sensory irritation, and respiratory stimulation by airborne benzene and alkylbenzenes: prediction of safe industrial exposure levels and correlation with their thermodynamic properties. Toxicol. Appl. Pharmacol. 65, 459 – 477. Nielsen, G. D., and Hansen, L. F. (1993). Sensory irritation of the upper respiratory tract. Pharmacol. Toxicol. 72(Suppl. 3), 32–35. Nielsen G. D., and Kristiansen, U. (1988). Activation of the sensory irritant receptor by C7-C11 n-alkanes. Arch. Toxicol. 61, 419 – 425. Nielsen, G. D., Kristiansen, U., Hansen, L., and Alarie, Y. (1988). Irritation of the upper airways from mixtures of cumene and n-propanol. Arch. Toxicol. 62, 209 –215. Nielsen, G. D., Thomsen, E. S., and Alarie, Y. (1990). Sensory irritant receptor compartment properties. Acta. Pharm. Nord. 1, 31– 44. Nielsen, G. D., and Vinggaard, A. M. (1988). Sensory irritation and pulmonary irritation of C3-C7 n-alkylamines: mechanisms of receptor activation. Pharmacol. Toxicol. 63, 293–304. Nielsen, G. D., and Yamagiwa, M. (1989). Structure-activity relationships of airway irritating aliphatic amines. Receptor activation mechanisms and predicted industrial exposure limits. Chem. Biol. Interact. 71, 223–244. Schaper, M. (1993). Development of a database for sensory irritants and its use in establishing occupational exposure limits. Am. Ind. Hyg. Assoc. J. 54, 488 –544. Silver, W. L. (1992). Neural and pharmacological basis for nasal irritation. Ann. NY. Acad. Sci. 641, 152–163. Silver, W. L., Mason, J. R., Adams, M. A., and Smeraski, C. A. (1986). Nasal trigeminal chemoreception: Responses to n-aliphatic alcohols. Brain Res. 376, 221–229. Verschueren, K. (1996). Handbook of Environmental Data on Organic Chemicals. Van Nostrand Reinhold, New York. WHO (1989). Indoor air quality: organic pollutants. World Health Organization Regional Office for Europe, Copenhagen.