| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Neuroendocrinology |
Monell Chemical Senses Center,3 Philadelphia, Pennsylvania 19104
Department of Dermatology,4 School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania 19104
Department of Animal Biology,5 School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania 19104
Department of Obstetrics and Gynecology,6 School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania 19104
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODS AND MATERIALSRESULTSDISCUSSIONREFERENCES |
|---|
gonadotropin-releasing hormone, luteinizing hormone, menstrual cycle, neuroendocrinology, pheromones
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODS AND MATERIALSRESULTSDISCUSSIONREFERENCES |
|---|
In humans, several studies have indicated that interpersonal relations among women, as well as between men and women, may alter reproductive endocrinology, suggesting the presence of primer pheromone activity. These relations include menstrual synchrony among women, first documented in all-women groups [7] and later replicated in a variety of other situations ([1, 2] for review).
Russell et al. [8] were the first to present evidence that suggested menstrual synchrony could be mediated by axillary secretions. Further studies employing these secretions from women also suggested that extracts of pooled samples, collected from women across the menstrual cycle, could be used to bring a recipient group of women into synchrony with the donors [9]. Extracts of male axillary secretions also appear to affect women with a history of irregular cycle lengths. Lengths of the menstrual cycles of these women showed a significant shift (vs. controls) toward the normal cycle length of 29.5 ± 3 days [10] upon application of male axillary secretion to the upper lip of the recipient women.
Additional support for axillary components from women having primer-pheromone activity is contained within the studies by Stern and McClintock [11] and Shinohara et al. [12]. These studies have focused on female axillary secretions being applied to women recipients. The results of Stern and McClintock [11] suggest that exposing women with normal menstrual cycles to axillary extracts from women in their follicular phase (the days following menses but several days prior to ovulation) shortens the length of the recipient's menstrual cycle by 1.7 ± 0.9 days. Exposing these same women to axillary extracts collected near the time of ovulation of the donors lengthens the recipient's menstrual cycle by 1.4 ± 0.5 days. Interestingly, similar effects on the lengths of the estrous cycles of female rats were noted when the rats were exposed to the follicular and ovulatory odors of other rats [13, 14]. Studies that have sought to provide a measure of primer pheromone activity by looking for a change in the length of the menstrual cycle have received some criticism for statistical and/or methodological errors [15–18]. Other criticisms have been fueled by the intra- and intersubject variability found in normal, consecutive menstrual cycles [19, 20].
Shinohara et al. [12] recently demonstrated changes in frequency of LH pulses in women exposed to female axillary secretions. These investigators demonstrated an average 28% decrease in LH pulse frequency in women receiving follicular phase secretions and a 16% increase in LH pulse interval in women receiving ovulatory phase secretions. The authors suggest that the changes in LH pulsing caused by axillary components are modulating the time of ovulation and changing cycle length; these changes may provide an objective measure of human primer pheromone activity.
Exogenous stimuli, such as a conspecific's chemical signal, appear to alter reproductive function by influencing the hypothalamic GnRH pulse generator. Although GnRH levels are not easily measured in humans, alterations in GnRH output can be inferred by measuring LH pulsing in blood. Changes in pulse frequency, height, and interval occur across a normal menstrual cycle. As women move from the early follicular phase of the menstrual cycle toward ovulation, LH pulses increase in frequency; however, LH pulse amplitude decreases in the midfollicular phase and increases at the time of ovulation. LH pulse frequency and amplitude progressively decrease across the luteal phase as the next menses draws near [21].
We hypothesized that a human male primer pheromone could act to alter menstrual cycle length and timing by altering LH parameters in a fashion analogous to that previously reported for female goats [22] and ewes [23]. In these animals, exposure of estrogen-primed females to either the male goat or his hair decreased the LH pulse interval, bringing on the next pulse after exposure to male stimuli faster than exposure to control conditions [22, 23]. Data presented herein demonstrate that applications of extracts of male axillary secretions cause a significant decrease in the latency to the next LH pulse when compared with application of a control substance. In addition to this primer pheromone effect, women also reported experiencing less tension and being more relaxed during exposure to the male extract, suggesting the presence of a modulator pheromone. These effects may serve as objective measures of primer and modulator pheromone activity that could guide an analytical isolation of the active axillary constituents.
|
METHODS AND MATERIALS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODS AND MATERIALSRESULTSDISCUSSIONREFERENCES |
|---|
Subjects were given a gynecologic exam and Pap smear; each had a full history taken. Each subject met the following criteria: 1) was between 21 and 45 yr of age (median and average ages: 27 yr and 28.2 ± 5.7 yr [SD], respectively); 2) had regular menstrual cycles of 25 to 35 days duration in the previous year (by self-report) and must not have skipped any cycle; 3) had stable exercise and sleep patterns; 4) had not taken any hormonal medication in the 3 mo preceding the study; 5) had a normal physical exam with no acne, hirsutism, or galactorrhea, and had body weight within 15% of normal according to the Metropolitan Life tables (Metropolitan Life Insurance Co., 1960); 6) had normal thyroid function tests; 7) had serum prolactin levels <15 ng/ml; and 8) was heterosexual. All subjects displayed normal olfactory ability and no general anosmia as evidenced by their ability to perceive dilute solutions of trimethylamine (0.25%) and 3-methyl-2-hexenoic acid (0.1%).
Once admitted to the study, subjects participated for three menstrual cycles discussed below as baseline, test, and final cycle.
Baseline Cycle
Beginning with the first or second day of menses, subjects recorded their basal body temperature on a standard calendar every morning. Morning urine samples were tested for LH daily, from Day 10 until the LH surge (using a commercially available kit; OvuQuick One Step, Henry Schein, Melville, NY), to determine the time of ovulation. Approximately 7 days following ovulation, serum progesterone levels were measured; a level >4 ng/ml was required for subjects to be included in the study.
Test Cycle
Basal body temperature, ovulation detection, and luteal phase progesterone measures were recorded as in the baseline cycle. However, within the first 7 days of the test cycle onset (the first day of menses is Day 1 of the menstrual cycle), subjects were admitted to the General Clinical Research Center (GCRC) at the Hospital of the University of Pennsylvania for a 12 h period: 0800–2000 h. During this time, subjects received 0.5 ml of control (ethanol) or male extract applied to the nasal region/upper lip every 2 h using a cotton pad. As described below, both extract and control were fragranced to be indistinguishable. Blood samples (
5 ml) were drawn through a heparin lock every 10 min. The nurses who drew the blood and applied the stimuli as well as the subjects were blind to the nature of the applied materials.
Since we had no a priori knowledge regarding potential lasting effects of male extracts upon LH pulsing parameters, half of the women were assigned to receive the control treatment during the first 6 h (0800–1400 h) of their stay in the GCRC and the male extract during the final 6 h (1400–2000 h); the other half received the treatments in reverse order.
Final Cycle
As a check on any changes in menstrual cycle length, patients were followed for an additional cycle by basal body temperature and urine LH measurements just as they were during their baseline cycle. Luteal phase progesterone levels were measured 7–10 days after the urinary LH indicated ovulation had occurred.
Male Extract and Control Solutions
The extract was created by using axillary secretions from four to six male donors. Secretions were collected on cotton pads as previously described [10, 24]. Three extracts were employed in the course of this study. Male donors were all healthy and ranged from 22–45 yr of age. All refrained from using deodorants/antiperspirants and washed/bathed using only plain soap with little to no fragrance (Ivory Soap, Proctor and Gamble, Cincinnati, OH) for a total of 4 wk. One week into this protocol, axillae were swabbed and axillary odor was assessed. Axillary swabs were used to generate bacteria cultures and determine the type and number of endogenous cutaneous bacteria. Following this, subjects collected axillary secretions on cotton pads three times per week. All pads were frozen (-10°C) until extracted. Male donors had no contact with and were not familiar with the female recipients.
We had no knowledge about the structural nature of the active axillary constituents; consequently, we arbitrarily chose to standardize all extracts to 0.5 µg/ml of 3-methyl-2-hexenoic acid (3M2H), the major analytical odorant found in the male axillary extracts [24]. This was done in the following manner: half of each extract was concentrated to 10 ml, and a 5 µl portion was analyzed by gas chromatography/mass spectrometry, as per previous studies [24], to determine the concentration of 3M2H. After quantifying the concentration of 3M2H, portions of the concentrated extract were added to the more dilute extract such that the concentration of 3M2H in the extract to be applied to subjects was 0.5 µg/µl.
The extract was then scented with a deodorant fragrance supplied by the Mennen division of Colgate-Palmolive (Piscataway, NJ) to a level of 0.005%. A similar volume of ethanol was used to dilute an aliquot of fragrance to 0.005% for use as an odor-matched control.
Mood Ratings
At 1 and 4 h into each 6-h exposure session, subjects were asked to provide ratings on a 7-point, Likert-type, categorical scale on the following: energetic, sensuous, tense, tired, calm, sexy, anxious, fatigued, relaxed, and active (see [25] for a full discussion of this type of scaling method). Subjects were allowed to self-interpret these adjectives. Horizontal scales, each on a separate sheet of paper, were composed from 1, labeled "I am not at all...," to 7, which was labeled "I am extremely...." Adjectives were scored by circling a numeral one at a time. After completing each scale, the forms were collected by the nurse so that subjects could not refer to previous ratings.
LH Levels and Pulse Detection
Each blood sample from our subjects was analyzed for LH using a standard immunoradiometric assay [26] by the Chemistry Core Lab of the GCRC of the Hospital of the University of Pennsylvania. After collection, blood samples were allowed to stand on ice for 30 min and then centrifuged. Serum was separated and frozen at -20°C until analysis. The minimum detectable level of LH was 0.1 mIU/ml of blood. Analyses of serum samples from groups of three to six subjects were run at different times during the course of this study. A coefficient of variation for the assay was calculated for each "batch" of samples by analysis of a pooled serum sample doped with standards. The coefficient of variation ranged from 2.8% to 12.3% during the course of this study.
The location, amplitude, and frequency of the LH pulses in the data were determined using the pulse detection software Cluster Analysis of Urban et al. [27]. With this program, significant pulses were detected using 2 x 1 cluster sizes (2 test samples for a nadir and 1 for a peak) and t-values of 2 x 2 to minimize false positives and negatives in the pulse detection. In addition to these parameters, we employed 0.75, 1.0, and 1.5 mIU/ml as minimum peak sizes for an LH pulse. All analyses yielded similar results. The results using 1.5 mIU/ml as a minimum peak size for a pulse are presented here.
Data Analysis
The mean LH levels, LH pulse amplitude, LH pulse frequency, and latency to next LH pulse in the control and extract periods of the follicular phase were compared using repeated measures analysis of variance (ANOVA) for each variable independently. Sequence of applications formed the independent groups factor; all others were repeated measures. Latency to the next LH pulse after an application also was analyzed with the nonparametric Wilcoxon signed ranks test. Possible diurnal effects were evaluated with paired-samples t-tests while ignoring type and order of stimulus application.
Categorical mood ratings were analyzed with the nonparametric sign test or the Wilcoxon signed ranks test. Prior to analyses of mood ratings, the two ratings in each phase (male extract and control) were compared with the sign test to determine whether there were any significant differences between ratings within each experimental condition. There were none; therefore, the two ratings in each phase of the experiment were averaged for subsequent analysis.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODS AND MATERIALSRESULTSDISCUSSIONREFERENCES |
|---|
Post-hoc interviews with all subjects demonstrated that none correctly guessed the true nature of the stimuli that were being delivered to their nasal region. All subjects perceived "alcohol" and/or "fragrance." None guessed that sweat/axillary-secretion extracts had been applied when it was revealed to them at the end of the interview. No changes in menstrual cycle length attributable to our 6-hr application of extract were noted.
Figure 1 presents examples of the pulse patterns seen in our female recipients. Relative to the average latency to the next pulse during control applications (59 ± 5 min), male axillary extracts shortened the time to the next LH pulse by an average of 20% (to 47 ± 5 min; see Fig. 2). Since our study was a repeated measures design, the error bars in Figure 2 can be misleading. Hence, average latency difference scores were calculated by subtracting the average latency to the next pulse during control applications from the average latency to the next pulse during application of male extract. These difference scores are seen in Figure 3.
|
|
|
The reduction in time between exposure to the extract and the onset of the next LH pulse was most effective following the first two applications (see Fig. 4) and suggests that there may be a refractory period following repeated exposures to male stimuli. Relative to the control condition, male extract did not affect the average number of LH pulses, the average LH pulse amplitude, or the basal level of LH (see Table 1).
|
|
We also noted significantly greater average LH pulse heights in the second 6-h interval (afternoon to early evening [1400–2000 h]; 4.51 mIU/ml) than in the first interval (morning; 3.85 mIU/ml; t(17) = 5.18, P < 0.001) and significantly higher basal LH levels in the afternoon (3.10 mIU/ml) than in the morning (2.80 mIU/ml; t(17) = 3.44, P < 0.005), both of which appear to agree with published findings that LH levels in the afternoon are elevated relative to the morning [28]. The average number of pulses in the morning (4.3) was not different from the number in the afternoon (4.2). We also examined the possibility that a diurnal influence might be responsible for the decreased latency to the next LH pulse that was observed subsequent to application of male extract. We ignored the type of stimulus application and compared the latencies with the next LH pulse after an application in the first 6 h (0800–1400 h) to the latencies noted during the next 6 h (1400–2000 h). The results of the ANOVA indicated that there was no significant difference between the morning (latency = 55 ± 5 min) and afternoon (latency = 51 ± 4 min) sessions (F1,17 = 1.75, P > 0.20).
Women were less tense (exact, two-tailed, significance = 0.012) and more relaxed (exact, two-tailed, significance = 0.022) during exposure to the male extract relative to the control stimulus (Fig. 5). No other effects on mood were noted. Mood ratings also were reanalyzed to determine whether there was a diurnal effect: there was none. The effects on moods appear to be specific to the male extract: a woman nurse was in close contact with the subject throughout the experiment, yet the mood effect was noted only during application of the male extract.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODS AND MATERIALSRESULTSDISCUSSIONREFERENCES |
|---|
This study presents a potential mechanism by which male axillary secretions may alter menstrual cycle length and timing. The change in LH pulsing is predominantly seen during the first two applications of extract and suggests that the male stimulus may have only a transient effect upon the GnRH pulse generator. In addition to the noted endocrine change, which may be accurately described as a primer pheromone effect, the mood data, demonstrating a lessening of tension and increased feeling of relaxation, suggest the presence of a modulator pheromone. Whether these two effects are mediated by the same or different compounds, or sets of compounds, is unknown. No other behavioral effects were noted as a result of extract application.
All subjects perceived that alcohol or fragrance was being applied to them during their stay in the GCRC; none guessed the true nature of the stimuli or the delivery sequence. This suggests that conscious awareness of body odor is not necessary to elicit the primer and modulator pheromonal responses we observed.
LH pulse maxima (heights) during extract application were significantly greater relative to control applications in the afternoon and early evening. Further, LH pulse maxima during afternoon control applications were greater than both morning control and morning extract maxima. These results suggest a diurnal shift in LH pulse maxima. LH has been reported to increase to maximal levels during the 1100–1700 h time-frame [28]; this time encompasses 3 h in our morning/afternoon session (0800–1400 h) and 3 h in our afternoon/early evening session (1400–2000 h). Alternatively, male extract may have a "lingering effect," which is reflected as higher levels of LH during afternoon applications of the control stimulus. Male extract may "uncouple" the controlling effect that endogenous opiates have upon the GnRH pulse generator. In this regard, its action appears similar to a small, acute dose of an opiate antagonist such as naloxone. In the follicular phase of the human menstrual cycle [31], as well as in ovoids [32], chronic administration of opiate antagonists raises tonic LH levels. As reported above, we may have observed this in subjects receiving male extract; however, we cannot say with certainty that a diurnal effect is not responsible for the elevated LH. Future studies should attempt to resolve this question by employing the same 12-h protocol reported herein and exposing subjects to the control stimulus every 2 h for the entire 12 h.
Other studies that have examined the effects of male and female axillary secretions upon the menstrual cycle have relied upon changes in cycle length as a measure of activity. The results reported here present a potential neuroendocrine measure of extract activity and a mechanism by which axillary secretion components may change cycle length. A periodic acceleration of LH pulses might be expected to occur during intimate heterosexual encounters as part of a stable heterosexual relationship. This pattern may ultimately result in a greater frequency of pulses and entrain the pulse generator as well as stimulate ovarian follicles with a subsequent increase in tonic estrogen levels; this increase has been demonstrated in domestic ruminants [33, 34] and is suggested in data from humans where plasma estrogen levels in women having regular, weekly heterosexual intercourse are higher than in women having infrequent or no heterosexual intercourse [35].
The recent study by Shinohara et al. [12] describes a reduction in LH pulse interval in women exposed to the follicular phase secretions of women donors and an increase in the LH pulse interval in women exposed to ovulatory phase secretions. These results suggest that female axillary secretions affect other women; however, the effects are different from those of male axillary secretions used in our study. Aside from the first pulse after application, we did not observe a significant increase or decrease in pulse intervals across the time we monitored LH. There are, however, significant methodological differences between the studies. The application schedule and protocol of Shinohara et al. [12] differed from ours. These investigators applied their stimuli every 30 min; we applied every 2 h. In addition, they did not use each subject as her own control as we did. In their study, women were randomly assigned to receive isopropanol (control), follicular phase secretions, or ovulatory phase secretions. The latter two were presented on axillary pads moistened with isopropanol swiped on the upper lip; we used an ethanol extract of axillary pads. Despite these differences, these results support and strengthen the impact of our data: 1) axillary secretions from women affect the LH pulses of other women, and 2) the changes seen in LH pulses are of the same magnitude as those seen by us. Shinohara et al. [12] reported that follicular phase secretions decrease LH pulse interval by 28%, and ovulatory phase secretions increase pulse intervals by 16%; we noted a decrease in the latency to the next LH peak by 20%. Shinohara et al. [12] did not examine this parameter. Based on earlier studies by the same authors [36, 37], they propose androstenol (5
-androst-16-en-3-ol) as the axillary constituent that is responsible for the change in LH. One of these earlier studies reported that women who exhibited menstrual synchrony had lower olfactory thresholds to androstenol [36]. In another study [37], women exposed to 2.5 mM androstenol in isopropanol (applied above the upper lip hourly for 4 h) experienced an increase in their LH interpulse interval vs. controls. The level of androstenol applied by these investigators to their subjects (300 µl of the 2.5 mM solution; personal communication to G.P.) far exceeds endogenous levels, reported to be 4–10 ng in male axillae [38, 39]. Furthermore, at the level used, the odor of androstenol would have been obvious to the recipients; no control odor was used to exclude the possibility that the observed results were spurious as the result of a novel odor.
Studies by Jacob and McClintock [5] utilized putative "human pheromones" reported to "stimulate the (human) vomeronasal organ" (VNO) ([40, 41]; however, see [42] for a review of experimental evidence for and against the existence of a functional, human VNO). These compounds are steroids, but only one, 
4,16-androstadienone, has been reported in the axillae in picomolar quantities [38], which are far lower than the levels used by Jacob and McClintock [5] (9 nmol/recipient; the other compound, estra-1,3,5(10)tetraen-3-yl acetate, is a synthetic compound never reported in the axillae). Jacob and McClintock [5] used such high levels (1000 times greater than endogenous concentration) because they wished to increase the compound's concentration near the pits of the "vomeronasal organ." Previous studies from our laboratory have not revealed the presence of this steroid in axillary extracts produced by the extraction techniques we employed in the current study [39]. Consequently, although Jacob and McClintock [5] reported modulator pheromone activity for androstadienone when used at nonphysiological concentrations, we do not speculate on it or any other compound's role in our study.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
2 Correspondence: George Preti, Monell Chemical Senses Center, 3500 Market St., Philadelphia, PA 19104. FAX: 215 898 2084; preti{at}monell.org
Received: 19 June 2002.
First decision: 11 July 2002.
Accepted: 8 January 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODS AND MATERIALSRESULTSDISCUSSIONREFERENCES |
|---|
-Androst-16-en-3-o1. Chem Senses 2000 25:407-411-Androst-16-en-3-o1 on the pulsatile secretion of luteinizing hormone in human female. Chem Senses 2000 25:465-467This article has been cited by other articles:
|
|
 |
|
  M. F Bhutta Sex and the nose: human pheromonal responses J R Soc Med, June 1, 2007; 100(6): 268 - 274. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Wyart, W. W. Webster, J. H. Chen, S. R. Wilson, A. McClary, R. M. Khan, and N. Sobel Smelling a Single Component of Male Sweat Alters Levels of Cortisol in Women J. Neurosci., February 7, 2007; 27(6): 1261 - 1265. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. J. Jechura, M. M. Mahoney, C. D. Stimpson, and T. M. Lee Odor-specific effects on reentrainment following phase advances in the diurnal rodent, Octodon degus Am J Physiol Regulatory Integrative Comp Physiol, December 1, 2006; 291(6): R1808 - R1816. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Chen, A. Katdare, and N. Lucas Chemosignals of Fear Enhance Cognitive Performance in Humans Chem Senses, June 1, 2006; 31(5): 415 - 423. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Jacob, N. A. Spencer, S. B. Bullivant, S. A. Sellergren, J. A. Mennella, and M. K. McClintock Effects of breastfeeding chemosignals on the human menstrual cycle Hum. Reprod., February 1, 2004; 19(2): 422 - 429. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
