Enhanced Sensitivity to Androstenone Following Regular Exposure to Pemenone

Transcription

1 Enhanced Sensitivity to Androstenone Following Regular Exposure to Pemenone David A. Stevens and Robert J. O'Connell Frances L. Hiatt School of Psychology, Clark University, 95 Main St, Worcester, MA 6 and Worcester Foundation for Biomedical Research, Maple Ave, Shrewsbury, MA 545, USA Correspondence to be sent to: R. J. O'Connell, Worcester Foundation for Biomedical Research, Maple Avenue, Shrewsbury, MA 545, USA Abstract The responses elicited by olfactory stimuli may be modified throughout an organism's life. For example, Wysocki et al. found that regular presentations of 5a-androst-6-en-3-one (androstenone) to anosmic subjects resulted in a graded increase in the perceived intensity of this substance in about half of their subjects (Wysocki et al., 989). The increased sensitivity they observed appeared to be specific to the exposed odorant and was presumed to occur only in anosmic subjects. Here, we continue our explorations of the individual differences in olfactory capabilities among subjects initially classified by their ability to detect and identify the odor of the diastereoisomeric ketone, c/s-4-(4'-t-butylcyclo-hexyl)-4-methyl--pentanone (pemenone) which shares with androstenone a pronounced urine-sweaty odor. We asked if regular pemenone exposure enhances the sensitivity of human subjects to pemenone, androstenone, isovaleric acid, or phenylethyl alcohol and, if shifts in threshold occurred, were they specific to particular odorants and classes of subject? Detection thresholds for the four substances were determined before and after 7-8 weeks of regular, biweekly, exposure to pemenone (n = 8 subjects) or a control substance ( subjects). Significant decreases in threshold were seen in the experimental group, relative to the control group, for androstenone, but not for the other compounds evaluated. Neither gender nor initial pemenone osmicity significantly affected the frequency of subjects with enhanced sensitivity. These findings show that a subject's sensitivity to one odorant can be enhanced by exposure to another, chemically distinct compound. Enhancement was not limited to anosmic individuals, but was also observed in some initially osmic and allosmic subjects. Chem. Senses : 43-49, 995. Downloaded from at Pennsylvania State University on March 6, 6 Introduction The ability of olfactory stimuli to alter the anatomical and physiological capabilities (Kaplan et al., 985; Woo and Leon, 987; Woo et al., 987; Guthrie et al., 99; Wang et al., 993) of the olfactory system is potent throughout development and adult life (Brunjes and Frazier, 986; Coopersmith and Leon, 986; Meisami, 989). Thus, the behavioral responses elicited by odorants may be modified or induced by experience, presumably as a consequence of the continuing developmental plasticity of olfactory receptor neurons and the connections they make centrally in the olfactory system. Regular odorant exposure elicits both a direct sensory experience and, in some subjects, a facilitatory effect on the perceived intensity of the exposed odor. For example, regular presentations of amyl acetate (AA) and androstenone (AND) to AND anosmic subjects by Wysocki et al. (989) resulted in a gradual decrease in the threshold Oxford University Press

2 44 D. A. Stevens and R. J. O'Connell for AND in about half of the human subjects without altering their sensitivity to AA or another compound. Regular exposure to AND or isovaleric acid (IVA) also increased the magnitude of the stimulus-induced voltage (EOG) measured across the olfactory epithelium of mouse strains initially selected for a reduced behavioral sensitivity to AND and IVA, respectively (Wang etal., 993). Increased sensitivity to these compounds was not seen following regular odor exposure in other mouse strains selected for a normal behavioral sensitivity to these odorants. Crossinduction of sensitivity to a non-exposed compound (AA) was not observed in any of the mouse strains. These human and animal studies collectively suggest that the shift in sensitivity following odor exposure is a peripheral physiological event, perhaps involving the capabilities of individual olfactory receptor neurons, and that it only occurs for particular odorants in exposed individuals with an initially reduced sensitivity to the exposed compound (Wang et al., 993). Pemenone (PEM) and AND are unrelated chemically (Ohloff et al., 983), but share many of the same perceptual characteristics. For example, the relative sensitivity of human subjects to the two compounds is positively correlated across individuals. Statistically significant relationships are found between the odor quality reports they elicit, their threshold concentrations (Stevens and O'Connell, 99) and the intensity scores elicited by above-threshold concentrations (O'Connell et al., 989). Certain individuals who seem to have a good sense of smell for most odors, lack the ability to perceive the aroma characteristic of particular compounds. This deficiency, expressed- by an elevated threshold, is often accompanied by a pronounced difference, when compared to osmics, in the perceived quality of liminal odor concentrations (O'Connell, 99). We classify those subjects with elevated thresholds and changed quality reports as specific allosmics to distinguish them from osmics, hyposmics and true anosmics who are unable to detect the highest possible concentration of a substance (Amoore et al., 976). Specific anosmias, allosmias and hyposmias (Amoore, 966; Amoore et al., 976; O'Connell et al, 994) exist for both PEM and AND. The distribution of specific defects for these substances are positively correlated across subjects (O'Connell etal., 989). Pemenone allosmics usually report mild floral, fruity or vegetable odor qualities for both PEM and AND, whereas osmics report strong sweaty, urinous or rancid odors. Finally, brief exposures to PEM cross-adapt the response to AND, but not the responses to IVA or phenylethyl alcohol (PEA), respectively, compounds which share the sweaty and floral odor quality labels sometimes elicited by PEM and AND in PEM allosmics (O'Connell et ai, 994). Collectively, these observations support a multiple profile model of odor quality encoding (Polak, 973), and suggest that PEM and AND interact with some of the same perceptual channels. They also suggest that PEM encoding may engage a greater number of perceptual channels, as it elicits a greater number of alternative odor quality labels across subjects than does AND. If, as we suspect, AND and PEM share some of the same odor quality encoding mechanisms, they might well share an ability to induce increased sensitivity in subjects through regular exposure. Moreover, PEM exposure might induce increased sensitivity only in those initially anosmic or allosmic to PEM, or it may induce an increase in all classes of subject, regardless of initial capabilities. The effects of regular PEM exposure on the thresholds and quality reports elicited by PEM, AND, IVA and PEA were determined by contrasting the effects of PEM exposure on these measures with those observed following regular exposure to a control odorant (mineral oil). Materials and methods Subjects The subjects were 4 students enrolled in one of four undergraduate psychology courses meeting twice a week at Clark University. Volunteers in two of the courses ( females and 6 males) were assigned to the experimental condition, and those in the other two (7 females and 5 males) were assigned to the control condition. All of the subjects were entered in a $ lottery as an honorarium for their services. Stimuli A binary dilution series was prepared in mineral oil for each stimulus. The starting concentration and the number of binary dilution steps for each odorant were: AND, 5.4 mm, steps; IVA, 5 mm, 9 steps; PEM, 4 mm, 5 steps; and PEA, 5 mm, steps. The stimuli were presented on polyester swabs (Falcon Swube no. 78), each holding, on average, 5 \i\. Sets of stimuli were stored at 4 C and warmed to room temperature before each testing session. During threshold testing, the sample sets were evaluated in Downloaded from at Pennsylvania State University on March 6, 6

3 Petnenone Enhances Androstenone Sensitivity I 45 vented exhaust hoods in the laboratory to minimize background odor in the experimental area. Procedure Detection thresholds for AND, IVA, PEA and PEM were determined in the laboratory using the method of limits with a single ascending series of concentrations (Amoore et ai, 968; Stevens and O'Connell, 99) before and after at least 5 bi-weekly classroom exposures to pemenone ( mm). To determine thresholds, the subjects were presented with a series of stimulus sets, each having five Swubes. Within each set, three of the Swubes contained a swab saturated with the diluent only, and two contained a swab saturated with a single concentration of an odorant. The position of the odorized swabs in each set of five was randomized. The initial concentration selected for presentation was one expected to be below threshold for the subject. Subjects were cautioned against touching the swabs to any part of their face and then required to pick the two swabs from the set that had an odor and to state the perceived odor quality from a standard list of odor descriptors (Stevens and O'Connell, 99). If the subject chose the two odorized swabs in the first set, a lower concentration was selected to begin the threshold determination. Sets with increasing concentrations of odorant were then given until the subject correctly identified the two odorized swabs in two successive sets. The highest dilution (i.e. lowest concentration) of the two correctly determined dilution steps was scored as the subject's threshold. If the subject failed to correctly identify the odorized swabs in the highest available concentration the next dilution number was assigned as the 'threshold' to distinguish these subjects from the occasional subject that first correctly identified the odorized swabs at the highest available concentration. These latter subjects were assigned a threshold value equal to that dilution step. A 3-s intertrial interval was imposed between the evaluation of individual stimulus sets. Thresholds were determined for each of the four odorants, with the order of odorant presentation determined randomly for each subject. The detection threshold procedure was performed again, or more days later, after which the odor exposure phase of the study was begun. Subjects in the control group sniffed swabs saturated with mineral oil whereas those in the experimental group sniffed swabs containing mm of PEM in mineral oil. Individuals were exposed at the beginning of their regular bi-weekly class meeting for 7 or 8 weeks. Each exposure consisted of two or three full sniffs of the PEM or control swab. The control and experimental groups met in different classrooms with at least days between each exposure. After the exposure phase, two determinations of threshold and reports of odor quality for the four odorants were again obtained in the laboratory from all subjects, again on separate days. Results All of the subjects reported thresholds for two or more of the test odorants at concentrations within the range expected for those with normal olfactory function. The subjects were then classified by the odor quality reports generated during the initial PEM threshold trials. Individuals were classified as: PEM osmics if they used a putrid odor descriptor (e.g. rancid, urine, sweaty or fecal) for PEM; allosmic if other quality reports (e.g. floral, vegetable or woody) were used; and anosmic if a 'no odor' descriptor was provided, or if the subject failed to correctly identify the highest available concentration of PEM. Of the 8 subjects in the PEM exposure condition, five were osmic to PEM, were allosmic, and one was anosmic. Of the subjects in the control condition, were osmic to PEM, eight were allosmic and three were anosmic. As determined previously (Stevens and O'Connell, 99), there was a regular relationship between odor quality classification and a subjects' threshold for PEM. The median initial PEM threshold for the osmic, allosmic and anosmic groups were.6, 59.5 and 8.8 mm, respectively. A Kruskal-Wallace test (Siegel, 956) showed significant differences in the PEM thresholds of these groups (H = 4, P<). For each subject, the median of the threshold dilution steps found on the two threshold testing days was determined for each of the four odorants before and after the exposure phase. Then, separately for each odorant, the difference in the median threshold of subjects before and after exposure was tested for statistical significance by comparing the experimental condition (PEM exposure) and the control condition (blank exposure) by Mann-Whitney tests (Siegel, 956). These tests showed a significant decrease in threshold for AND in the experimental group following exposure to PEM relative to the control group, but no significant differences for the other three odorants. The median threshold dilution-step differences for the test compounds and their probabilities are listed in Table. Ten of the 8 subjects in the experimental condition showed a decrease in threshold (difference score > ) for AND versus 4 of in the control group. The frequency distribution of these threshold Downloaded from at Pennsylvania State University on March 6, 6

4 46 I D. A. Stevens and R. J. O'ConneU Table Median threshold difference scores* PEM AND IVA PEA Control -.5 Exp..5. z P *A positive value indicates an increase in sensitivity (i.e. a decreased threshold) decreases across treatments was significantly different, as evaluated by chi square (x = 6.8, P<). Thus, both the magnitude of the shift in AND threshold and the number of subjects with an increased sensitivity were enhanced by PEM exposure. Figure shows simple dot charts (Cleveland, 985) of the difference scores produced by each subject, organized by exposure condition, and indicating both the gender and initial PEM osmicity classification of each subject. Mann- Whitney tests of these data failed to reveal significant - i-- « e a- T I 6 - D O - 9 PEUENONE EXPOSURE FIMINONE CD ANDROSTTNONK cm BLANK EXFOSUBB o i- 6- ISOVALXRIC ACID o PEMENONE EXPOSURE CA CD CO PHENYLETHYL ALCOHOL BLANK EXPOSURE Downloaded from at Pennsylvania State University on March 6, D a cm CD D FEMBNONE EXPOSURE BLANK EXPOSUKE PEMENOME EXPOSURE J OE> -fficcccft m cm* BLANK EXPOSURE Figure The differences in the median threshold dilution step for each subject before and after PEM exposure are plotted in separate dot charts for each of the test odorants. Positive scores indicate increased sensitivity (decreased threshold concentrations). The pre-exposure PEM osmicity dassifaction and sex of each subject are also indicated.

5 Pemenonc Enhances Androstenone Sensitivity I 47 differences in the median threshold difference scores in groups of subjects classified by gender, either within exposure conditions, or overall; Ps >5. To determine if the effects of exposure differed with respect to a subject's initial osmicity for PEM, the osmic and allosmic classes were combined to increase within-group N, and then compared with anosmics independently within control and experimental conditions for each of the four odorants. Mann- Whitney tests again showed no classification effects within the exposure conditions for any of the odorants; Ps >5 in every instance. Quality shifts As before, a standard set of semantic equivalents was used to categorize the odor quality labels employed by the subjects (Stevens and O'Connell, 99). In this way, all of the odor quality reports were subsumed into a standard set of semantic equivalents (putrid, floral-fruity, wood-vegetable, none) which were then tabulated and converted to percentages for each of the four odor compounds. We then computed a difference score by subtracting the pre-exposure percentage from the corresponding post-exposure percentage so that a positive score indicated that a particular odorant elicited that label more frequently after exposure. Only small (<5%) difference scores were observed in the odor quality labels elicited by the individual odors in the control group. Of the 5 subjects in the exposure group initially classified as anosmic for AND, nearly half reported some quality for AND after regular exposure to PEM. In contrast, only one of the 5 AND anosmic control subjects provided an odor quality report after exposure (x = 6.4, P<). Other than this shift from the anosmic condition, there was no other evidence that PEM exposure affected the quality reports elicited by the individual test odorants. Most allosmic and osmic subjects reported the same quality after exposure as before exposure. Table lists the number of observed shifts in quality and the subjects' initial osmic classification, as a function of experimental treatment for both PEM and AND. Discussion In several regards these results are similar to those reported by Wysocki et al. (989), in that increases in sensitivity to AND were observed, but not all of the subjects chronically exposed to an odorant showed alterations in their AND threshold. Here, slightly more than half of the subjects in the experimental group ( of 8) showed an increase in Table Frequency of shifts in quality report Classification PEM osmic allosmic anosmic AND osmic allosmic anosmic Group Exposure Shift 7 No shift 4 8 Control Shift 3 No shift sensitivity to AND. Of these, six were initially classified as allosmic to PEM and four were osmic. Eight of these same subjects were also classed as anosmic to AND, along with one allosmic and one osmic. Wysocki et al. (989) found that 4 of their subjects, all initially classified as anosmic, showed some increase in sensitivity to AND following chronic exposure to AND, with induction defined as a decrease in threshold of three or more binary dilution steps. Wang et al. (993) failed to find a net increase in AND sensitivity in osmic mice following chronic exposure to AND, using the magnitude of the electro-olfactogram as a metric. These studies suggest that increased sensitivity does not occur in osmic subjects. In the present study, induction was defined by a decrease in threshold of at least one binary dilution step. Collectively, these results suggest that individuals are not equally affected by chronic exposure and, for some, an increase in sensitivity may be impossible. Unfortunately, logic prevents a strong test of this latter, null possibility. However, our data do show that one's initial sensitivity to an odorant does not preclude increases in sensitivity as a result of chronic exposure. A much larger study will be required to determine if the amount of induction is graded and proportional to ones initial sensitivity. An interesting difference from the Wysocki et al. (989) study is the observation of cross-induction between odorants without significant self-induction of sensitivity. Chronic exposure to PEM produced a significant increase in the sensitivity to AND in some subjects, but no significant change in the thresholds of the other test odorants, PEM, IVA and PEA. Although Wang et al. (993) suggested that Downloaded from at Pennsylvania State University on March 6, 6

6 48 D. A. Stevens and R. J. O'Connell increased sensitivity due to exposure is odorant-specific (an expectation that is difficult to evaluate, given the constellation of odors that could be affected), it seems clear from this study that increased sensitivity as the result of exposure is not odorant-specific for AND, at least to the extent that another chemically unrelated compound (PEM) is effective. The ability of PEM exposure to produce cross-induction without apparent self-induction is consistent with our view of the multiple profile model of odor quality coding (Polak, 973). Previously, we had suggested that PEM activates more receptor channels than does AND, that they both activate at least one of these channels in common and that the perception of a 'urinous' odor quality by a subject is normally the result of the collective activation of several of these receptor channels. These suggestions are based on our earlier findings, that PEM stimulation elicits a greater variety of odor quality reports across individuals than does AND, that there is a positive correlation between the occurrence of PEM and AND osmias, allosmias and anosmias in a sample of subjects, and the observation that PEM is a strong adaptor of AND sensitivity (O'Connell et al., 989; O'Connell, 99; Stevens and O'Connell, 99; O'Connell and Stevens, 994; O'Connell et al., 994). The results presented here, together with Wysocki et al. 's earlier findings (989) that AND induction was apparently specific (of the compounds tested, only the sensitivity to AND was affected by exposure), suggest that at least one of the receptor channels shared by PEM and AND is inducible. Moreover, the presence of cross-induction and the lack of significant self-induction (i.e. a significant increase in PEM sensitivity ACKNOWLEDGEMENTS following PEM exposure), suggests that the shared inducible receptor channel is a major component in a subject's sensitivity to AND, but only one of several receptor channels that contribute to the coding of PEM. This is also indicated by the small, but non-significant, increase in PEM sensitivity after exposure (Table ). Although IVA elicits quality reports that are similar to those elicited by PEM and AND in osmics, its coding seems not to involve the PEM inducible receptor channel. Increased sensitivity to odorants, over time, has been observed by many investigators (Engen, 96; Doty et al., 98; Rabin and Cain, 986; Wysocki et al., 989). In several of these studies, the shift in threshold observed was attributed to the effects of repeated testing or regular exposure to a variety of odorants. There are substantial differences among all of these studies with respect to the compounds evaluated, the training and instruction techniques employed, and the temporal distribution and total amount of testing performed. Although this makes comparisons among studies, as to mechanism, difficult, shifts in sensitivity within a test session have been observed after hundreds of closely spaced trials (Doty et al., 98) or were found to generalize from one odor quality to another (Rabin and Cain, 986). In contrast, the brevity of our PEM odor exposure treatment and the specificity of its effect on the threshold of AND suggests that the psychological phenomena related to repeated testing described by others to account for increased sensitivity are not responsible for the decreased thresholds reported here. The authors would like to thank Dr G. Ohloff, FTRMENICH SA, Geneva, Switzerland for providing the sample of PEM, and the students and staff of our respective institutions for their assistance. Supported by NIDCD grants DC3 and DCOO37 to RJO. Portions of these data were presented at the Association for Chemoreception Sciences 5th Annual Meeting, April 993. Downloaded from at Pennsylvania State University on March 6, 6 REFERENCES Amoore, J.E. (966) Specific anosmias and primary odors. In Tanyolac, N.N. (ed.), Theories of Odor and Odor Measurement. Robert College Research Center, Bebek, pp Amoore, J.E., Venstrom, D. and Davis, A.R (968) Measurement of specific anosmia. Percept. Motor Skills, 6, Amoore, J.E., Forrester, L.J. and Pelosi, P. (976) Specific anosmia to isobutyradehyde: the malty primary odor. Chem. Senses,, 7-5. Brunjes, P.C. and Frazier, L.L. (986) Maturation and plasticity in the olfactory system of vertebrates. Brain Res., 396, -45. Cleveland, W S. (985) The Elements of Graphing Data. Wadsworth Advanced Books and Software, Monterey. Coopersmith, R. and Leon, M. (986) Enhanced neural response by adult rats to odors experienced early in life. Brain Res., 37, 4^43.

7 Pemenone Enhances Androstenone Sensitivity I 49 Doty, R.L., Snyder, P.J., Huggins, G.R and Lowry, L D. (98) Endocrine, cardiovascular, and psychological correlated of olfactory sensitivity changes during the human menstrual cycle. J. Comp. Physiol. Psychol., 95, Engen, T. (96) Effect of practice and instruction on olfactory thresholds. Percept. Motor Skills,, Guthrie, K.M., Wilson, D.A. and Leon, M. (99) Early unilateral deprivation modifies olfactory bulb function J. Neurosci.,, Kaplan, M.S., McNelly, N.A. and Hinds, J W. (985) Population dynamics of adult-formed granule neurons of the rat olfactory bulb. J. Comp. Neurol., 39, 7-5. Meisami, E. (989) A proposed relationship between increases in the number of olfactory receptor neurons, convergence ratio and sensitivity in the developing rat. Dev. Brain Res., 46, 9-9 O'Connell, R.J. (99) Specific anosmias: implications for the physiological mechanisms of quality discrimination. In Lawless, H.T. and Klein, B.P. (eds), Sensory Science Theory and Applications in Foods. Marcell Dekker, New York, pp 5-5. O'Connell, R.J. and Stevens, D.A. (994) Individual differences in the quality perceptions of specific anosmics. Aroma<hology Rev., 3, -. O'Connell, R.J., Stevens, D.A., Akers, R.P., Coppola, D.M. and Grant, A.J. (989) Individual differences in the quantitative and qualitative responses of human subjects to various odors. Chem. Senses, 4, O'Connell, R.J., Stevens, D.A. and Zogby, L.M. (994) Individual differences in the perceived intensity and quality of specific odors following self- and cross-adaptation. Chem. Senses, 9, Ohloff, G., Giersch, W., Thommen, W. and Willhalm, B. (983) Conformationally controlled odor perception in 'steroid-type' scent molecules. Helvet. Chim. Acta, 66, Polak, E.H. (973) Multiple profile-multiple receptor site model for vertebrate olfaction. J.Theor. Biol., 4, Rabin, M.D. and Cain, W.S. (986) Determinants of measured olfactory sensitivity. Percept. Psychophys., 39, Siegel, S. (956) Nonparametric Statistics for the Behavioral Sciences. McGraw-Hill Book Company, New York. Stevens, D.A. and O'Connell, R.J. (99) Individual differences in thresholds and quality reports of human subjects to various odors. Chem. Senses, 6, Wang, H.-W., Wysocki, C.J. and Gold, G.H. (993) Induction of olfactory receptor sensitivity in mice. Science, 6, Woo, C.C. and Leon, M. (987) Sensitive period for neural and behavioral response development to learned odors. Brain Res., 433, Woo, C C, Coopersmith, R and Leon, M. (987) Localized changes in olfactory bulb morphology associated with early olfactory learning. J. Comp. Neurol., 63, 3-5. Wysocki, C J, Domes, K.M. and Beauchamp, G.K. (989) Ability to perceive androstenone can be acquired by ostensibly anosmic people. Proc. Nat. Acad. Sci. USA, 86, Received on September 3, 994; accepted on March 7, 995 Downloaded from at Pennsylvania State University on March 6, 6