- Email: [email protected]
Olfaction in utero: Behavioral studies of the mouse fetus
B HAVIOURAL 6 ROCESSES ELXVIER
Behavioural Processes 39 (1997) 53-68
Olfaction in utero: Behavioral studies of the mouse fetus David M. Coppola
*,
Leah C. Milk
Department of Biology, Davidson College, Davidson. NC 28036, USA
Received 22 August 1994; revised 9 July 1996; accepted 12 July 1996
Abstract On the day before birth, mouse fetuses were tested for their behavioral responses to iso-amyl acetate and iso-valeric acid delivered into the nasal cavity in liquid phase. Pilot studies established that most subjects responded to most concentrations used with an increase in total behaviors and number of different behaviors displayed. When iso-amyl acetate was systematically delivered at three concentrations (10-j M, 10m4 M and 10m5 M) subjects showed significant responses to each. In contrast, responses were not evident when saline was the stimulus. The magnitude of the average behavioral response did not decline with decreasing concentration of odorant. In a second experiment, the ability of the fetus to discriminate between iso-amyl acetate and iso-valeric acid was studied. Results establish that the fetuses responded differently to the two odorants. Given the immaturity of the mouse’s accessory olfactory system before birth and the observed responses to concentrations of odorants below the threshold of the trigeminal system, the results suggest that the mouse fetus has a functionally competent main olfactory system including the ability to discriminate between purified odorants. The results are discussed in terms of current models of chemosensory ontogeny. 0 1997 Elsevier Science B.V. Keywords:
Fetus; Behavior; Olfaction; Vomeronasal; Prenatal
1. Introduction in its development. Its competence at birth is necessary for survival (Alberts, 1976) and yet significant developmental events extend well into the postnatal period (see Pedersen et al., 1985 and Brunjes and Frazier, 1986, for reviews). What’s more, the olfactory system, a major conduit of chemical information to the CNS, The
sense
of smell
posses
an interesting
paradox
for
those
interested
Corresponding author. Current address: Department of Neurobiology, Box 3209, Duke University Medical Center, Durham, NC 27710. l
0376-6357/97/$17.00 PII
SO376-6357(96)00044-7
Copyright 0
1997 Elsevier Science B.V. All rights reserved
54
D.M. Copplo.
L.C. Millcw/
Belruciorrrul Procerser 3Y f 1997) 53-6X
shows plasticity throughout life both in its turnover of primary receptor neurons (Graziadei and Monti Graziadei, 1978) and in its functional responses to odors (Wang et al., 1993; Wysocki et al., 1989). A fundamental question of development arises from such an appraisal of our current knowledge and ignorance in this area. How does a sensory system, like the sense of smell, gain functional competence? For example, does it occur incrementally or is it in an all-or-none fashion. To begin to answer this question we must ask how the system’s capabilities change ontologically. Unfortunately, a paucity of data prevents a rigorous comparison of the chemosensory capabilities of the neonate with those of the adult for any species (but see Mistretta and Bradley, 1983). We know that the young mammal relies on olfactory abilities immediately after birth to orient toward and attach to its mother’s nipple (see Pedersen et al., 1985 and references therein). Also, various reports suggest that memories of chemosensory stimuli important for survival (e.g food preference) are inculcated soon after birth (Galef and Henderson, 1972; Galef and Sherry, 1973). Preliminary attempts have been made to examine the sensitivity of the developing animal to odorants (Alberts and May, 1980; Apfelbach et al., 1991). Collectively these findings point to at least some sophistication in the abilities of even a nascent chemosensory system. More recently, owing to changing views of the neural competence of the mammalian fetus as well as technological innovations that made real-time evaluation of fetal behavior possible in vivo, investigators have pushed back the frontiers of sensory developmental studies into the prenatal period (Hepper, 1992). In a spate of elegant studies, investigators have provided convincing evidence that the rat fetus not only detects chemical stimuli in the amniotic fluid but can become classically conditioned to these cues such that their behavior is altered even after birth in response to them (Pedersen and Blass, 1982; Smotherman, 1982; Smotherman and Robinson, 1985 and review by Schaal and Orgeur, 1992). Studies of human premature babies and fetuses in utero suggest too that chemosensory abilities exist in our own species weeks before birth (reviewed by Schaal and Orgeur, 1992). Taken together the results of behavioral studies suggest that the chemosensory abilities of the young, even unborn, mammal are far from inept. Yet, it still is unknown which systems are mediating the chemosensory phenomena observed in the perinatal period. Complicating this problem are the various chemosensory or potential chemosensory systems present in mammals including taste, trigeminal, accessory olfactory and olfactory. Attempts to verify which chemosensory systems are functioning in the perinatal period have been further hampered by the friable condition of the fetus and neonate and methodological problems that make it impossible to eliminate a given system’s 1982; Smotherman and Robinson, 1985; Smotherman and Robinson, participation (Smotherman, 1988; Smotherman and Robinson, 1989; Smotherman and Robinson, 1990). The mouse provides an advantage as a subject of study compared to the rat which heretofore has been the dominate model of mammalian chemosensory development. Unlike the rat, its vomeronasal duct, which is the portal for stimulus access to the accessory olfactory system, is not patent before birth (Coppola and O’Connell, 1989) having a protracted period of postnatal development (Coppola et al., 1993). This fact can be exploited to illuminate the olfactory capabilities of the developing mouse given the appropriate choice of stimuli. The goals of this study were to determine: if the late-term mouse fetus demonstrates behavioral responses to chemical stimulation of the naso-oral cavity with purified odorants; how responses, if present, are altered by varying stimulus concentration including concentrations below the reported detection threshold for the trigeminal or gustatory systems; and if the fetus can discriminate between purified odorants. Our results support the conclusion that even before birth the mammalian olfactory system is highly functional.
D.M. Coppoltr.
L.C. Millor/Behat,Gmd
Processes 39
f 1997)
53-68
55
2. Materials and methods 2.1. Dams Timed pregnant mice of the CD-I strain were used in all experiments (Charles River, Warrington, MA). One hundred and twenty dams were used in the pilot and experimental studies. They were received on the 14th or 15th day of gestation and housed in polycarbonate cages with corn cob bedding. Mouse chow and tap water were always available. All other husbandry practices were in accordance with the Guide for the Care and Use of Laboratory Animals, KHEW Publication No. (NIH) 80-23. On the 18th day of gestation, considering the first day a sperm plug is discovered as day zero, the subject dams underwent a chemomyelotomy procedure prior to the observation of their fetuses. 2.2. Chemomyelotomy Chemomyelotomy, a technique for desensitizing and immobilizing a pregnant subject was pioneered in the 1960’s (Basmajian and Ranney, 1961) and has been used extensively in studies of fetal psychobiology (e.g. Smotherman and Robinson, 1988; Smotherman and Robinson, 1989; Smothetman and Robinson, 1990). This technique humanely desensitizes the mother so that her abdomen can be surgically prepared to allow direct view of the fetus without exposing it to general anesthesia. If the gravid uterus is delivered into a saline bath maintained at 37°C the viability of the fetus and the mother can be maintained throughout the experimental period. In this study the procedures of Smotherman (1982) were followed or modified as noted below. The chemomyelotomy was performed under Metofane anesthesia. After induction, an one cm midline incision was made in the skin overlying the first and second lumbar vertebrae. Following blunt dissection to visualize the intervertebral gap, 10 ,ul of 100% ethyl alcohol was injected into the space surrounding the spinal cord with a 100 ,ul syringe affixed with a 30 ga. hypodermic needle. After the injection the dam was placed on a heating pad and allowed to recover from the anesthesia. Upon establishing a successful chemomyelotomy the dam was anesthetized again to briefly immobilize her. A three cm long midline incision was made in the abdomen through which the uterus was to be delivered. The dam was then comfortably restrained in a supine position on an inclined ramp such that her abdomen was submerged in a bath of physiological saline maintained at 37°C. The dam was allowed to awaken from anesthesia after placement into the bath and after her uterus was exteriorized through the abdominal incision. The dam was maintained in this state for the length of the behavioral observation which was restricted such that it did not begin until the dam had been awake for 15 min and ended within one hour. Upon completion of the behavioral observations the dam and fetuses were killed with an overdose of Metofane. 2.3. Preparation
of fetus
One or two fetuses were selected at random from a given dam. A fetus that was selected for behavioral observations was exteriorized through a small incision made in the uterus and fetal
D.M. Coppolu, L.C. Milk/
56
Behauiourul Processes 39 (1997) 53-68
membranes. Care was taken not to disturb the fetal or maternal blood flow or to compromise the placental-uterine connection which was visually monitored throughout the observation period. In a series of pilot studies, a system was developed for the reliable delivery of chemical stimuli into the fetal nasal cavity. A 30 cm length of tubing (PE, 1 mm i.d.) was affixed with a small amount of cyanoacrylate cement to the external naris of the fetus. This required that the fetus’ snout be gently raised momentarily above the surface of the bath to allow the cement to adhere to the skin around the naris. Three mm from the end of the main tube that was attached to the nares, another piece of similar sized tubing was attached to form a manifold for the sensitivity tests (experiment one) and a second tube was attached for the discrimination tests (experiment two). The opposite end of each piece of tubing was attached to a separate perfusion pump which drove a five cc syringe. Initially, air was passed through the main tube to establish that a complete seal had been made between the tubing and the skin around the nares. Air bubbles coming from the mouth without any leakage from the junction between the tubing and the nares was taken as evidence the experiment could proceed. Air was purged from the tubing with physiological saline which was pumped through the nasal cavity by a perfusion pump at a rate of 28 pl/min throughout the remainder of the study. 2.4. Stimulus parameters The stimuli used in this study consisted of two purified odorants (Sigma) diluted to the desired concentration in physiological saline. Iso-amyl acetate and Iso-valeric acid were chosen for a number of reasons. First, they are odorants that have been widely used in other studies on rodents and therefore data is available that is relevant to the interpretation of the results reported here. Second, both are relatively weak trigeminal irritants, and in the case of iso-amyl acetate there exists a large disparity between olfactory and trigeminal threshold. Third, these two odorants have very different odor qualities for humans and presumably for rodents i.e. iso-amyl acetate is a ‘floral’ odor and iso-valeric acid is a ‘putrid’ odor. Guiding the choice of stimulus delivery parameters were estimates of the volume of the fetal nasal cavity made from morphometric analysis of histological sections through the nasal cavity (unpublished data). These estimates, based on two individuals, suggested that together the left and right sides of the fetal nasal cavity had a volume of approximately five mm 3. Therefore, we chose five ~1 as the volume of stimulus in all experiments so we could be certain that the stimulus was not overly diluted by the fhrid volume in the nasal cavity. The perfusion pumps were programmed such that the nominal flow rate would not change when the stimulus was being added to the stream. While actual flow rates were not measured, inspection of the flow of a colored stream into the saline bath confirmed an apparent smooth and constant delivery of the stimulus. Considering the flow rate of the saline purge stream and the estimated naris volume the nasal cavity was evacuated approximately every five seconds during the entire observation period. 2.5. Behavioral
observations
The behavior of each subject fetus was observed and recorded using the computer program The Observer (Vers. 3.0, Aldus, The Netherlands). Observations were done between 9.00 and 17.00 h (L: D 12: 12, lights on at 6.00 am) by a single individual. Within a given experiment a subject fetus was randomly assigned to a treatment condition. ignorant of the particular treatment condition.
Except
as noted below,
the experimenter
was
D.M. Coppnlu, L.C. Millar/Behu~~iourul Processes 39 (1997) 53-68
57
2.5. I. Behavioral categories In a series of preliminary studies the behavioral repertoire of the E 18 mouse fetus was determined under the conditions described above. Nearly all behaviors of the fetus occur as instantaneous point events which allowed the use of frequency counts as the basis for quantitative analysis. Nine behavioral categories were recognized. Six of these were termed ‘basic’ and refer to the region of the body being moved. They include: head, mouth, trunk, front limbs, hind limbs and tail. For the hind limb and front limb categories bilateral and unilateral movements were not distinguished. Three other categories termed ‘complex’ are comprised of behaviors that not only involve simultaneous movements of different regions of the body, but are also stereotyped with respect to the coordination of the movements. The complex categories included: ‘body wiggle’, involving lateral or torsional movements of the trunk, ‘body stretch’, involving dorsal extension of the trunk and head, and ‘wipe’, a movement of one or both front limbs across the snout as if to wipe something from it. 2.5.2. Data reduction and analysis Three variables were derived from the individual behavioral counts that simplified the analysis without compromising the goals of the study. The first, termed ‘behavioral activity’, entailed the summation of all behavioral events during the observation period. The second, termed ‘behavioral complexity,’ included the number of different behavioral categories displayed during the observation period. The third termed ‘behavioral diversity’ was included for completeness since previous authors have favored this index of uncertainty (entropy) which has its origins in information theory (e.g. Smotherman and Robinson, 1986; Paulus and Geyer, 1993). Diversity or ‘H’ is an expression of the overall entropy of behavior expressed in bits per interval. It was calculated as follows: N
H = C P;( -log, i=
Pi),
I
where N equals the number of different behavioral categories and Pi equals the probability that behavior i will occur in the time interval used (see Shannon, 1948 for the complete derivation and justification of this index). The Observer software was used to tally the frequency of each behavioral category across the pre-stimulus and post-stimulus observation periods. These totals were then exported to SYSTAT (Vers. 5.1, Evanston, IL) statistical software for analysis. In experiment one, data were analyzed by repeated measures analysis of variance (ANOVA) using both the univariate and multivariate approach of SYSTAT. The total variance was further partitioned to test a priori hypotheses about certain groups of means. In experiment 2, paired t-tests were used to make a priori comparisons between the results of certain post-stimulus test periods. Since pilot studies established that only increases in behavior occurred with odor stimulation all tests were one-tailed.
3. Experiment 1 Published data on the rat fetus under conditions similar to this study have established that their behavioral output increases when chemical stimuli are delivered to their oral cavity (Smotherman and Robinson, 1985; Smotherman and Robinson, 1988; Smotherman and Robinson, 1989). In a series of
pilot studies (data not shown) we determined that the mouse fetus would react similarly when solutions of iso-amyl acetate or iso-valeric acid were delivered directly into the nares. The purpose of this experiment was to systematically explore the sensitivity of the E 18 mouse fetus to iso-amyl acetate. A two minute pre-stimulus test, during which saline was flushed through the nasal cavity, was followed by three two minute post-stimulus tests that began with odor delivery. The saline pump, which continuously flushed the nasal cavity, was programmed to deliver 28 pl/min except during the time when the second pump delivered the stimulus. During this later period the saline pump delivered at I / 10th the initial rate and the stimulus pump delivered a 5 ~1 bolus at a nominal rate of 25.2 Fl/min. These parameters were chosen to minimize the change in flow rates associated with the addition of the stimulus to the stream. A three-step log dilution series between IO-’ M and lo-’ M iso-amyl acetate was used along with the diluent, physiological saline, as the control. Sixty-one dams yielded between 22 and 24 fetuses for each of the four treatment groups.
4. Results Diverse studies of behavior in the rat and human fetus have demonstrated that their spontaneous motor activity tends to oscillate frequently and irregularly (see Smotherman et al., 1988 and references therein). Fig. 1 illustrates the behavior of two typical subjects from the saline control group. The two subjects shown were somewhat atypical in the greater magnitude of the oscillations but not in the basic pattern of their behavior. The number of behaviors performed were totalled every 20 s throughout the four two minute observation periods. Clearly, the spontaneous behavior of the mouse fetus, like that of the other species studied, appears to oscillate over short intervals. Any measurements of evoked behavior, such as those describe in this study, will have this highly variable spontaneous behavior as a confounding factor. Besides illustrating the oscillatory nature of mouse fetal behavior, Fig. 1 also illustrates for two individuals what was another general trend. Most fetuses showed no detectable increase in the behavioral variables when saline was the stimulus. In fact, while rarely reaching statistical significance (r-tests not shown), the amount, complexity and diversity of behavior in the post-stimulus tests tended to be lower than in the pre-test when saline was the stimulus. The mean changes in behavioral activity engendered by the delivery of odorant or saline are plotted in Fig. 2a as the difference compared to the pre-test. Note that increases in activity were observed after odorant was delivered in most post-stimulus periods for each the three concentrations of the stimulus while a decline in behavioral activity was observed when saline was the stimulus (not significant by r-test). Only in the second and third post-stimulus tests at the highest stimulus concentration did there appear to be a lack of a response to the odorants. The patterns of the individual means were explored statistically in a repeated measures ANOVA with odor concentration as the between-groups factor and test period as the within-groups factor. Both concentration (df = 3, 87; F = 4.6; p < 0.005) and test period (df = 3, 261; F = 3.9; p < 0.01) were significant factors as was their interaction (df = 9, 261; F = I .9; p < 0.05). The pre-stimulus periods were compared to post-stimulus periods combined using the multivariate matrix operations in SYSTAT. The highly significant result further establishes the effect of stimulus delivery in increasing motor activity (df = 4, 87; F = 5.0; p < 0.001).
D.M. Coppolu. L.C. Millur/Beha~wuraI
Processes 39 (1997) 53-68
59
20
Subject
20 1
Pre Test
1st
2nd
3rd
Consecut ive 2 min Intervals Fig. 1. Behavioral
activity
of two fetuses, tallied at 15 s intervals
across four consecutive
saline was added to the saline purge line at the beginning
of each stimulus
the stimulus
which
did not appear to alter spontaneous
activity
is normally
two min test periods. Fifteen
period (1st through
@I of
3rd). Note that the delivery
of
quite irregular.
In Fig. 3 the behavioral activity of two individuals is tallied at 20 second intervals throughout the 8 min test. These results while clearly providing evidence of responses to the stimulus delivery also illustrate the high and variable activity level in the pretest and the variability in the timing and magnitude of the stimulus response. The typical response to the odorant was for the animal to produce a series of sudden head, mouth and trunk movements along with a complex extension of the head and straightening of the torso termed a ‘stretch’. No behavior was found to be unique to a stimulus response, rather the response could be characterized by an increase in the frequency of a given movement and the rapid succession of different movements particularly of the mouth, head, forelimbs and torso. Examination of the results for behavioral complexity revealed a pattern similar to the results for behavioral activity. Fig. 2b, in which mean behavioral complexity is plotted as the difference compared to the pre-stimulus period, reveals an increase in all post-stimulus tests for each of the three stimulus concentrations. Following the pattern of behavioral activity, behavioral complexity decreased
D.M.CoppolaL.C.Millor
60
/ Beho~~iourul
Processes
39
f 1997)
53-68
n
1st 2 min.
q
2nd2min.
q
3rd 2 min.
m
1st 2min.
a
2nd2min.
0
3rd 2 min.
1
Saline
x 10”M
x 1 0-4M
x 10-5M
Stimulus Concentration
-1-I
L Saline
x 10-3M
x 1 0-4M
X 10s5M
Stimulus Concentration Fig. 2. Mean and SEM behavioral responses are plotted after subtracting the pretest activity (before a stimulus was delivered) for three consecutive two min stimulus trials. Five gl of the stimulus solution was added to the saline purge line at the beginning of each stimulus period. The stimuli were saline or one of three dilutions of iso-amyl acetate. (A) Behavioral activity (total number of behaviors) per two min interval is shown. (B) Behavioral complexity (number of different behaviors displayed) per two min interval is shown. Sample sizes ranged from 22 to 24 subjects per stimulus concentration (see text for statistical comparisons)
on average in all post-stimulus tests when saline was the stimulus. These subjective impressions were supported by the statistical analysis. The repeated measures ANOVA, revealed that both concentration (df = 3, 87; F = 4.7; p < 0.004) and test period (df = 3, 261; F = 4.5; p < 0.004) were significant factors as was their interaction (df = 9, 261; F = 3.6; p < 0.0001). The multivariate comparison
D.M. Coppdu. L.C. Millor/Behcrc~iolr~ctl Pro~e.ws
39 (1997) 53-M
61
_”
I
Subject 50
1
lo-4M
1 O-S M )r 16 .E > '4= 2 i? b .-
12
8
zi _c cv 0
Pre Test
1st
Consecutive
2nd
2 min
4 3rd
Intervals
Fig. 3. Behavioral activity of two fetuses, tallied at 1.5 s intervals across four consecutive two min test periods. Five ~1 of iso-amyl acetate (two different concentrations are shown) as added to the saline purge line at the beginning of each stimulus period (1st through 3rd). Note variable rise above spontaneous behavioral activity after each stimulus delivery.
between the pre-stimulus and post-stimulus periods was also significant (df = 4, 87; F = 7.8; p < 0.0001). Results from the analysis of behavioral diversity were similar to the previous variables. Means and SEM for the different concentrations and test periods are provided in Table 1. The repeated measures ANOVA, revealed that both concentration (df = 3, 87; F = 5.0; p < 0.003) and test period (df = 3, 261; F = 5.7; p < 0.001) were significant factors as was their interaction (df = 9, 261; F = 2.9; p < 0.0002). The multivariate comparison between the pre-stimulus and post-stimulus periods was also significant (df = 4, 87; F = 8.0; p < 0.0001). No obvious pattern was evident across the repeated post-stimulus tests for any of the three behavioral variables. For example, only at the highest stimulus concentration did there seem to be a diminution in response with repeated stimulations as one might expect if sensory adaptation was
Table I Means f SEM behavioral diversity (entropy) of the mouse fetus in response to four different iso-amyl acetate concentrations in four consecutive two minute test periods. See text for derivation of variable and results of statistical analysis Test period pretest
Concentration IO-’ IO-’ 10-s Saline
M M M
0.935 + 0.067
0.982+ 0.070 0.834 f 0.069 0.850 & 0.070
I st 2 min
2nd 2 min
3rd 2 min
1.086 f I ,143f I.064 k 0.750 +
0.989 + 0.070 1.164~0.073 1.150+0.072 0.793 * 0.073
0.964 + 0.063 1.138kO.066 I .04 1f 0.064 0.759 &-0.066
0.077 0.080 0.078 0.080
operating. Besides the possibility of adaptation there was no obvious a priori hypothesis concerning the repeated stimulations and therefor these difference were not systematically analyzed. A question left unanswered by the omnibus tests described thus far is whether the different concentrations of odorant engendered different magnitudes of response. This question was addressed using the multivariate matrix operations of SYSTAT to partitions the variance components of the overall test. First, the saline stimulus condition was compared to the three odorant concentrations in a multivariate test. As expected, the results for behavioral activity (df = 4, 84; F = 4.4; p < 0.003), behavioral complexity (df = 4, 84; F = 9.0; p < 0.0001) and behavioral diversity (df = 4, 84; F = 6.2.0; p < 0.0001) were all highly significant demonstrating a strong affect of odorant in increasing behavioral output. Next, having established that the iso-amyl acetate but not saline significantly increased motor activity, the effects of the three concentrations of odorants were compared in a multivariate test. The magnitude of the behavioral response to the different odor concentrations was not significantly different for any of the three variables including behavioral activity (df = 8, 168; F = 1.3; p > 0.051, behavioral complexity (df = 4, 84; F = I. 1; p > 0.05) and behavioral diversity (df = 4, 84; F = 1.4; p > 0.05).
5. Experiment
2
The purpose of this experiment was to determine if the E 18 mouse fetus could discriminate between iso-amyl acetate and iso-valeric acid, each at a concentration of IO-’ M. A habituation paradigm was employed in which repeated stimulations with one odorant were expected to lead to a diminution in response after which the second odorant would be delivered. The protocol was planned after considering the results of a series of pilot studies in which multiple stimulations of iso-amyl acetate where followed by stimulation with iso-valeric acid. The series of behavioral observations and stimulus deliveries were as follows: There were eight one minute observation periods; the first was a pre-test with a continuous flow of saline; the second through sixth were the habituation trials, each of which began with the delivery of 5 ~1 of iso-amyl acetate; the seventh and eighth periods were the discrimination trials which began with the delivery of 5 ~1 of iso-valeric acid. Twelve dams yielded sixteen fetuses for this part of the study. Analysis of the variable diversity was not done for this experiment as it seemed redundant based on the outcome of experiment one.
D.M. Coppolu. L.C. Millur/
Pwcesses 39 (1997) .(3-6X
Beh~iowul
63
6. Results The repeated deliveries of iso-amyl acetate did not result in a detectable diminution of average response as would be expected if habituation were occurring. In fact, though not significantly different. the second and fourth stimulus delivery engendered more behavior on average than the first. pco.02
f 12
A
p< 0.05
c
7
7
,
Q) $
5 =8
.-b iii
0
Amy1 Acetate
n
lsovaleric
Acid
10
8I-
6,-
T
c
l%
Stimulus
Tr
4-
S r”
2
o-
i Consecutive
1
min Intervals
F
p< 0.003
DC 0.0005
7 Stimulus
6
0
Amy1 Acetate
n
lsovaleric Acid
Consecutive 1 min Intervals Fig. 4. Mean and SEM behavioral responses are plotted after subtracting the pretest for seven consecutive one min stimulus trials. Five ~1 of the stimulus solution was added to the saline purge line at the beginning of each stimulus period. The stimulus for the first five trials was iso-amyl acetate and was iso-valeric acid for the last two trials. Both odorants were delivered at 10-j M concentration. (A) Behavioral activity (total number of behaviors) per two min interval is shown. (B) Behavioral complexity (number of different behaviors displayed) per two min interval is shown. Brackets give statistical comparisons between the last iso-amyl acetate trial and the iso-valeric acid trials. N = 16.
However, to establish discrimination, the critical comparison is between the last delivery of iso-amyl acetate and iso-valeric acid. Each of the five post-stimulus tests with iso-amyl acetate caused a significant increase in behavioral activity compared to the pre-test (Fig. 4a). The two trials with iso-valeric acid confirmed that, despite the lack of habituation, this odorant caused a greater response than iso-amyl acetate. Both the first and second delivery of iso-valeric acid caused a significant increase in behavioral activity compared to the pre-test period (I= 4.59, df = 15, p < 0.0002 and r=4.21, df= 15, p < 0.0007) and importantly the mean responses for these tests were significantly different from the response engendered by the previous (5th) delivery of iso-amyl acetate ( f = 1.9 1, df = 15, p < 0.05 and t = 2.18, df = 15, p < 0.02). The data for behavioral complexity agree with those of behavioral activity (Fig. 4b). Probably owing to the smaller sample size in this study, compared to experiment 1, only the 3rd and 4th delivery of iso-amyl acetate engendered a significant response by this measure, however, all the post-stimulus delivery periods showed the same pattern of increased behavior compared to the pre-test. Both the first and second delivery of iso-valeric acid caused a mean response that was significantly different from the pre-test (t = 4.65, df = 15, p < 0.0003 and r = 2.18, df = 15, p < 0.02). Additional confirmation of the fetuses’ ability to discriminate the odorants was provided by the significant differences found between both iso-valeric acid tests and the previous (5th) iso-amyl acetate test (I = 3,91, df = 15, p < 0.0005 and t = 3.24, df = 15, p < 0.003).
7. Discussion The results of experiments one agree with those reported for the rat fetus which have established an increase in behavioral activity after the delivery of chemical stimuli to the naso-oral region (Smotherman, 1982; Smotherman and Robinson, 1985; Smotherman and Robinson, 1988; Smotherman and Robinson, 1989; Smotherman and Robinson, 1990). The current study was, to our knowledge, the first to examine the influence on the fetus of different concentrations of purified odorants delivered directly into the nasal cavity. The results of the first experiment clearly demonstrate that transient increases in behavioral activity and complexity occurred after iso-amyl acetate was added to a stream of physiological saline which continuously purged the fetal nasal cavity. The three concentrations of odorants used were equally effective in eliciting behavior upon their initial delivery, however, the two lowest concentrations appeared to be more reliable in triggering behavioral increases when delivered a second or third time. The loss of response after the first delivery at the highest concentration of iso-amyl acetate was originally thought to be a habituation or accommodation phenomenon but the results of the discrimination studies (see below) which employed the same concentration, argue against this interpretation. In any case, if there was a trend, it was for the lowest concentration to be more reliable than the highest in eliciting a response. The results of experiment two demonstrate that the late-term mouse fetus can discriminate between purified odorants delivered into the nasal cavity. Despite the failure of the protocol used to habituate the animals to repeated stimulation with iso-amyl acetate an inherent and unexpected difference in response to iso-valeric acid was revealed. On average, the animals showed a higher increase in behavioral activity and complexity when stimulated with iso-valeric acid than with isoamyl acetate when both were delivered at a concentration of lo-’ M. However, since the two odorants doubtless have different intensity profiles, the results are consistent with discrimination based on either odor
D.M. Coppol~. L.C. Millctr / Behuriourul
Processes 39 (I 997) 53-68
65
intensity or quality. Similar results have been reported for the rat fetus which demonstrated discrimination between solutions of lemon and mint extracts infused into the mouth (Smotherman and Robinson, 1992). The ability of the rat fetus to detect chemical stimuli before birth (Smotherman, 1982; Smotherman and Robinson, 1985) and to become conditioned to chemicals in the womb such that their prenatal or postnatal behavior is affected (Hepper, 1988; Smotherman, 1982; Smotherman and Robinson, 1985) has fueled considerable speculation about this phenomena’s adaptive significance (e.g. Hepper, 1987; Schaal and Orgeur, 1992; Smotherman and Robinson, 1988). For example, it has been suggested that attraction to food substances in the maternal diet, attraction to maternal pheromones, or even recognition of kin might all be imprinted during gestation. In theory, this could occur through fetal detection of the appropriate chemical stimuli if it gained access to fetal chemosensory systems through the maternal circulation. Olfaction, including participation by the accessory olfactory system, has been the presumed modality involved as it would appear to be the only chemosensory system with the functional capacity in the adult to mediate the behaviors that have been considered. Despite the potential importance of such an adaptation, there is little direct evidence that the olfactory system is functioning prenatally. A recent study demonstrated that isolating the olfactory bulbs from more posterior brain structures causes a diminution in response to chemical stimulation in the mt fetus (Smotherman and Robinson, 1990). Other studies of fetal behavior, not designed to address the question of sensory modality per se, have typically confounded stimulation of the different chemosensory systems of the naso-oral cavity by using very high concentrations or unpurified substances (eg. Smotherman and Robinson, 1985; Smotherman and Robinson, 1988; Smotherman and Robinson, 1989). Anatomical studies, while not definitive, generally support the notion of precocious olfactory competence (see Pedersen et al., 198.5 and Brunjes and Frazier, 1986, for reviews). The only electrophysiological study available, reported electro-olfactograms and single units responses from receptor cells of the prenatal rat (Gesteland et al., 1982). In this study, the olfactory cells responded to relatively low concentrations of purified odorants but did not gain specificity for odorants until the day of birth. The studies reported here provide among the first behavioral evidence that the mammalian fetus has a functioning main olfactory system. Involvement of the accessory olfactory system can be excluded in the mouse since the duct which connects the lumen of this organ to the nasal cavity is not patent before birth (Coppola and O’Connell, 1989; Coppola et al., 1993). Participation of the trigeminal system can also be excluded on the grounds that responses were manifest to concentrations of iso-amyl acetate below the threshold of the trigeminal system for this odor. Tucker (1963). in his now classic studies using recordings from the trigeminal nerve, showed that its threshold for iso-amyl acetate across two different species, tortoise and rabbit, fall in the same narrow range (10-l and IO-‘.” of vapor saturation) well above the lowest concentration used in this study. Subsequently, Silver and Moulton (1982) found strikingly similar thresholds in the rat to this compound. Clearly, rabbits, rats and tortoises constitute a diverse group both phylogenetically and biologically and yet their trigeminal responses are quite similar. Thus, we conclude that our across species comparison is justified. Concerning potential disparities related to the delivery of stimuli in a fluid medium, Tucker ( 1963) also found, using recordings from the olfactory nerve, that the medium (water or air) in which odorant was delivered had little effect on response profiles. Lastly, to exclude trigeminal involvement we rely on comparisons between fetuses and adult animals. However, for our argument to be invalid on these grounds the fetus would have to have a lower trigeminal threshold than the adult. To the
66
D.M. Coppoltr. L.C. MiNur / Behuviotmd
Proce.w.~ 39 (I 997) 53-68
contrary, a gradual increase in chemosensory sensitivity during development would appear to be the rule although clearly more research in this area is needed (Alberts and May, 1980; Apfelbach et al., 1991). It is unknown to what extent the trigeminal or other chemosensory channels contributed to the discrimination by the fetus of iso-amyl acetate and iso-valeric acid. The concentrations used in this part of the study were presumably high enough to stimulate the trigeminal system. While neither odorant is a particularly strong trigeminal irritant there is some evidence that iso-valeric acid is more potent in this regard than iso-amyl acetate (Doty, 1975). Taken together the results from this study challenge some of the prevailing views of chemosensory development. Pedersen et al. (1985) proffered one of the most well delineated models after consideration of the exiting literature and their own findings using the 2-deoxyglucose technique in the rat (Greer et al., 1982; Pedersen et al., 1983; Teicher et al., 1980 see also Alberts, 1976). They suggested that the accessory olfactory system might mediate chemoreception before birth, while the modified-glomerular complex, a specialized region of the olfactory bulb, and a subset of olfactory receptor neurons might mediate chemoreception in the early postnatal period eventually giving way to the main olfactory system which would mediate the reception of airborne chemicals in adulthood. Our findings make this theory untenable in the mouse and thus not generalizable for macrosmatic vertebrates. Despite the lack of a functional accessory olfactory system, the mouse fetus can detect purified odorants at relatively low concentrations and can discriminate between different odorants. It remains possible that some subset of the main olfactory apparatus, like the modified-glomerular complex, might be mediating these processes but there is no indication from our results that the late-term fetus has a functionally incomplete or inept olfactory system. In fact, we found no diminution of response even at the lowest stimulus concentration. Alberts and May ( 1980) found, in a study of odor-induced polypnea in rats from 1 to 17 days of age, that sensitivity to amyl acetate increased with age. Decreasing thresholds for amyl acetate were also found in older rats as a function of increasing age (Apfelbach et al., 1991). Therefore, it might be predicted that the fetus has a lower sensitivity to odorants than adults. A growing literature is at odds with the notion that sensory experience commences at birth in mammals (Hepper, 1992; Schaal and Orgeur, 1992 and references therein). However, despite considerable speculation and circumstantial evidence, it is still unknown whether prenatal chemosensory experience alters postnatal behavior so as to increase survival. The results of this study provide evidence that the main olfactory system or some part of it is functioning in the late-term mouse fetus. Thus, it is possible that this system, which dominates chemoreception of airborne cues in the postnatal period, may begin to acquire experiences with odors in utero. If this process occurs as a natural consequence of fetal development, it certainly will require a reevaluation of what constitutes an ‘innate’ response to chemical stimuli. Much additional work is needed to determine the pervasiveness of this phenomenon and its significance in shaping odor-mediated behaviors that are critical for survival. Acknowledgements The authors would like to gratefully acknowledge the support of the Whitehall Foundation and the NSF (grant BNS 911305 I). Special thanks to Alan Mackay-Sim and his colleagues at Griffith University for extending D.C. their hospitality during the preparation of this manuscript.
67
References Alberts, J.R., 1976. Olfactory contributions to behavioral development in rodents. In: R.L. Doty (Editor), Mammalian Olfaction, Reproductive Processes and Behavior. Academic Press, New York, pp. 67-95. Alberts, J.R. and May, B., 1980. Ontogeny of olfaction: Development of the rat’s sensitivity to urine and amyl acetate. Physiol. Behav., 24: 965-970. Apfelbach, R., Russ, D. and Slotnick, B.M., 1991. Ontogenetic changes in odor sensitivity, olfactory receptor area and olfactory receptor density in the rat. Chem. Sens., 16: 209-218. Basmajian, J.V. and Ranney, D.A., 1961. Chemomyelotomy: substitute for general anesthesia on experimental surgery. J. App. Physiol., 16: 386-392. Brunjes, P.C. and Frazier, L.L., 1986. Maturation and plasticity in the olfactory system of vertebrates. Brain Res. Bull., 11: 1-45. Coppola, D.M., Budde, J. and Millar, L., 1993. The Vomeronasal duct has a protracted postnatal development in the mouse. J. Morphol., 218: 59-64. Coppola, D.M. and O’Connell, R.J., 1989. Stimulus access to olfactory vomeronasal receptors in utero. Neurosci. Lett., 106: 241-248. Doty, R.L., 1975. Intranasal trigeminal detection of chemical vapors by humans. Physiol. Behav., 14: 855-859. Galef, B.G. and Henderson, P.W., 1972. Mother’s milk: A determinant of the feeding preferences of weaning rat pups. J. Comp. Physiol. Psych., 78: 213-219. Galef, B.C. and Sherry, F.D., 1973. Mother”s milk: A medium for transmission of cues reflecting the flavors of mother’s diet. J. Comp. Physiol. Psych., 83: 374-378. Gesteland, R.C., Yancey, R.A. and Farbman, A.I., 1982. Development of olfactory receptor neurons selectivity in the rat fetus. Neuroscience, 7: 3 127-3 136. Graziadei, P.P.C. and Monti Graziadei, G.A., 1978. Continuous nerve cell renewal in the olfactory system. In: M. Jacobsen (Editor), Handbook of Sensory Physiology, Vol. IX. Springer-Verlag, New York, pp. 98-112. Greer, CA., Stewart, W.B., Teicher, M.H. and Shepherd, G.M., 1982. Functional localization of the olfactory bulb and a unique glomerular complex in the neonatal rat. J. Neurosci., 2: 1744-1759. Hepper, P.G., 1987. The amniotic fluid, An important priming role in kin recognition. Anim. Behav., 35: 1343-1346. Hepper, P.G., 1988. Adaptive fetal learning: Prenatal exposure to garlic affects postnatal preferences. Anim. Behav., 36: 935-936. Hepper, P.G., 1992. The quest for the developmental origins of behavior. Q. J. Exp. Psych., 44 B: 161-163. Mistretta, C.M. and Bradley, R.M., 1983. Neural basis of developing taste sensation, Response changes in fetal, postnatal and adult sheep. J. Comp. Neurol., 215: 199-210. Paulus, M.P. and Geyer, M.A., 1993. Three independent factors characterize spontaneous rat motor activity. Behav. Brain. Res., 53: 1 I-20. Pedersen, P.E. and Blass, E.M., 1982. Prenatal and postnatal determinants of the first suckling episode in the albino rat. Dev. Psychobiol., 15: 349-356. Pedersen, P.E., Stewart W.B., Greer, C.A. and Shepherd, G.M., 1985. Early development of olfactory function. In: E.M. Blass (Editor), Handbook of Behavioral Neurology, Vol. 8. Plenum, New York, pp. 163-203. Pedersen, P.E., Stewart, W.B., Cheer, CA. and Shepherd, G.M., 1983. Evidence for olfactory function in utero. Science, 22 1: 478-480. Schaal, B. and Orgeur, P., 1992. Olfaction in Utero: Can the rodent model be generalized. Q. J. Exp. Psych., 44 B: 245-278. Shannon, C.E., 1948. A mathematical theory of communication. Bell Sys. Tech. J., 27: 379-423 and 623-656. Silver, W. and Moulton, D., 1982. Chemosensitivity of rat nasal trigeminal receptors. Physiol. Behav., 28: 927-931. Smotherman, W.P., 1982. Odor aversion learning by the rat fetus. Physiol. Behav., 29: 769-771. Smotherman, W.P. and Robinson, S.R., 1985. The rat fetus in its environment, behavioral adjustments to novel, familiar, aversive and conditioned stimuli presented in utero. Behav. Neurosci., 99: 521-530. Smotherman, W.P. and Robinson, S.R., 1986. Environmental determinants of behavior in the rat fetus. Anim. Behav., 34: 1859-1873. Smotherman, W.P. and Robinson, S.R., 1988. Behavior of rat fetuses following chemical or tactile stimulation. Behav. Neurosci.. 102: 24-34.
6X
D.M. Copplo,
Smotherman,
L.C. Mil/ur/
Behucwurtrl Prncesses 39 f 1997) 53-68
W.P. and Robinson,
S.R., 1989. Rat fetuses respond to chemical
W.P. and Robinson,
S.R.,
stimuli
in gas phase. Physiol.
Behav.,
47:
863-868. Smotherman,
tactile stimulation. Smotherman,
W.P. and Robinson,
Smotherman,
W.P., Robinson,
Comp. Psych., Teicher,
M.H.,
glomerular
1990. Olfactory
bulb transectton
alters fetal behavior
after chemosensory
but not
Dev. Brain Res., 57: 175-180. S.R.. 1992. Habituation
S.R. and Robertson,
in the rat fetus. Q. J. Exp. Psych., 44 B: 215-230.
S.S., 1988. Cyclic
motor activity
in the fetal rat (Rams
J.
noregicus).
102: 78-82.
Stewart, region
W.B.,
Kauer,
in the developing
J.S. and Shepherd, rat olfactory
bulb
G.M.,
1980. Suckling
revealed
pheromone
by the 2-deoxyglucose
stimulation method.
of a modified Brain
Res.,
194:
530-535. Tucker,
D.. 1963. Olfactory,
and Taste I. Pergamon Wang, H., Wysocki, Wysocki,
vomeronasal
C.J. and Gold, G.H.,
C.J., Dorries,
and trigeminal
Press, New York,
1993. Induction
K. and Beauchamp,
receptor responses to odorants.
In: Y. Zotterman
(Editor),
Olfaction
pp. 45-69. G.K.,
anosmic people. Proc. Natl. Acad. Sci. U.S.A.,
of olfactory
1989. Ability 86: 7976-7978.
receptor sensitivity
to perceive
in mice. Science, 260: 998-1000.
androstenone
can be acquired
by ostensibly
Behavioural Processes 39 (1997) 53-68
Olfaction in utero: Behavioral studies of the mouse fetus David M. Coppola
*,
Leah C. Milk
Department of Biology, Davidson College, Davidson. NC 28036, USA
Received 22 August 1994; revised 9 July 1996; accepted 12 July 1996
Abstract On the day before birth, mouse fetuses were tested for their behavioral responses to iso-amyl acetate and iso-valeric acid delivered into the nasal cavity in liquid phase. Pilot studies established that most subjects responded to most concentrations used with an increase in total behaviors and number of different behaviors displayed. When iso-amyl acetate was systematically delivered at three concentrations (10-j M, 10m4 M and 10m5 M) subjects showed significant responses to each. In contrast, responses were not evident when saline was the stimulus. The magnitude of the average behavioral response did not decline with decreasing concentration of odorant. In a second experiment, the ability of the fetus to discriminate between iso-amyl acetate and iso-valeric acid was studied. Results establish that the fetuses responded differently to the two odorants. Given the immaturity of the mouse’s accessory olfactory system before birth and the observed responses to concentrations of odorants below the threshold of the trigeminal system, the results suggest that the mouse fetus has a functionally competent main olfactory system including the ability to discriminate between purified odorants. The results are discussed in terms of current models of chemosensory ontogeny. 0 1997 Elsevier Science B.V. Keywords:
Fetus; Behavior; Olfaction; Vomeronasal; Prenatal
1. Introduction in its development. Its competence at birth is necessary for survival (Alberts, 1976) and yet significant developmental events extend well into the postnatal period (see Pedersen et al., 1985 and Brunjes and Frazier, 1986, for reviews). What’s more, the olfactory system, a major conduit of chemical information to the CNS, The
sense
of smell
posses
an interesting
paradox
for
those
interested
Corresponding author. Current address: Department of Neurobiology, Box 3209, Duke University Medical Center, Durham, NC 27710. l
0376-6357/97/$17.00 PII
SO376-6357(96)00044-7
Copyright 0
1997 Elsevier Science B.V. All rights reserved
54
D.M. Copplo.
L.C. Millcw/
Belruciorrrul Procerser 3Y f 1997) 53-6X
shows plasticity throughout life both in its turnover of primary receptor neurons (Graziadei and Monti Graziadei, 1978) and in its functional responses to odors (Wang et al., 1993; Wysocki et al., 1989). A fundamental question of development arises from such an appraisal of our current knowledge and ignorance in this area. How does a sensory system, like the sense of smell, gain functional competence? For example, does it occur incrementally or is it in an all-or-none fashion. To begin to answer this question we must ask how the system’s capabilities change ontologically. Unfortunately, a paucity of data prevents a rigorous comparison of the chemosensory capabilities of the neonate with those of the adult for any species (but see Mistretta and Bradley, 1983). We know that the young mammal relies on olfactory abilities immediately after birth to orient toward and attach to its mother’s nipple (see Pedersen et al., 1985 and references therein). Also, various reports suggest that memories of chemosensory stimuli important for survival (e.g food preference) are inculcated soon after birth (Galef and Henderson, 1972; Galef and Sherry, 1973). Preliminary attempts have been made to examine the sensitivity of the developing animal to odorants (Alberts and May, 1980; Apfelbach et al., 1991). Collectively these findings point to at least some sophistication in the abilities of even a nascent chemosensory system. More recently, owing to changing views of the neural competence of the mammalian fetus as well as technological innovations that made real-time evaluation of fetal behavior possible in vivo, investigators have pushed back the frontiers of sensory developmental studies into the prenatal period (Hepper, 1992). In a spate of elegant studies, investigators have provided convincing evidence that the rat fetus not only detects chemical stimuli in the amniotic fluid but can become classically conditioned to these cues such that their behavior is altered even after birth in response to them (Pedersen and Blass, 1982; Smotherman, 1982; Smotherman and Robinson, 1985 and review by Schaal and Orgeur, 1992). Studies of human premature babies and fetuses in utero suggest too that chemosensory abilities exist in our own species weeks before birth (reviewed by Schaal and Orgeur, 1992). Taken together the results of behavioral studies suggest that the chemosensory abilities of the young, even unborn, mammal are far from inept. Yet, it still is unknown which systems are mediating the chemosensory phenomena observed in the perinatal period. Complicating this problem are the various chemosensory or potential chemosensory systems present in mammals including taste, trigeminal, accessory olfactory and olfactory. Attempts to verify which chemosensory systems are functioning in the perinatal period have been further hampered by the friable condition of the fetus and neonate and methodological problems that make it impossible to eliminate a given system’s 1982; Smotherman and Robinson, 1985; Smotherman and Robinson, participation (Smotherman, 1988; Smotherman and Robinson, 1989; Smotherman and Robinson, 1990). The mouse provides an advantage as a subject of study compared to the rat which heretofore has been the dominate model of mammalian chemosensory development. Unlike the rat, its vomeronasal duct, which is the portal for stimulus access to the accessory olfactory system, is not patent before birth (Coppola and O’Connell, 1989) having a protracted period of postnatal development (Coppola et al., 1993). This fact can be exploited to illuminate the olfactory capabilities of the developing mouse given the appropriate choice of stimuli. The goals of this study were to determine: if the late-term mouse fetus demonstrates behavioral responses to chemical stimulation of the naso-oral cavity with purified odorants; how responses, if present, are altered by varying stimulus concentration including concentrations below the reported detection threshold for the trigeminal or gustatory systems; and if the fetus can discriminate between purified odorants. Our results support the conclusion that even before birth the mammalian olfactory system is highly functional.
D.M. Coppoltr.
L.C. Millor/Behat,Gmd
Processes 39
f 1997)
53-68
55
2. Materials and methods 2.1. Dams Timed pregnant mice of the CD-I strain were used in all experiments (Charles River, Warrington, MA). One hundred and twenty dams were used in the pilot and experimental studies. They were received on the 14th or 15th day of gestation and housed in polycarbonate cages with corn cob bedding. Mouse chow and tap water were always available. All other husbandry practices were in accordance with the Guide for the Care and Use of Laboratory Animals, KHEW Publication No. (NIH) 80-23. On the 18th day of gestation, considering the first day a sperm plug is discovered as day zero, the subject dams underwent a chemomyelotomy procedure prior to the observation of their fetuses. 2.2. Chemomyelotomy Chemomyelotomy, a technique for desensitizing and immobilizing a pregnant subject was pioneered in the 1960’s (Basmajian and Ranney, 1961) and has been used extensively in studies of fetal psychobiology (e.g. Smotherman and Robinson, 1988; Smotherman and Robinson, 1989; Smothetman and Robinson, 1990). This technique humanely desensitizes the mother so that her abdomen can be surgically prepared to allow direct view of the fetus without exposing it to general anesthesia. If the gravid uterus is delivered into a saline bath maintained at 37°C the viability of the fetus and the mother can be maintained throughout the experimental period. In this study the procedures of Smotherman (1982) were followed or modified as noted below. The chemomyelotomy was performed under Metofane anesthesia. After induction, an one cm midline incision was made in the skin overlying the first and second lumbar vertebrae. Following blunt dissection to visualize the intervertebral gap, 10 ,ul of 100% ethyl alcohol was injected into the space surrounding the spinal cord with a 100 ,ul syringe affixed with a 30 ga. hypodermic needle. After the injection the dam was placed on a heating pad and allowed to recover from the anesthesia. Upon establishing a successful chemomyelotomy the dam was anesthetized again to briefly immobilize her. A three cm long midline incision was made in the abdomen through which the uterus was to be delivered. The dam was then comfortably restrained in a supine position on an inclined ramp such that her abdomen was submerged in a bath of physiological saline maintained at 37°C. The dam was allowed to awaken from anesthesia after placement into the bath and after her uterus was exteriorized through the abdominal incision. The dam was maintained in this state for the length of the behavioral observation which was restricted such that it did not begin until the dam had been awake for 15 min and ended within one hour. Upon completion of the behavioral observations the dam and fetuses were killed with an overdose of Metofane. 2.3. Preparation
of fetus
One or two fetuses were selected at random from a given dam. A fetus that was selected for behavioral observations was exteriorized through a small incision made in the uterus and fetal
D.M. Coppolu, L.C. Milk/
56
Behauiourul Processes 39 (1997) 53-68
membranes. Care was taken not to disturb the fetal or maternal blood flow or to compromise the placental-uterine connection which was visually monitored throughout the observation period. In a series of pilot studies, a system was developed for the reliable delivery of chemical stimuli into the fetal nasal cavity. A 30 cm length of tubing (PE, 1 mm i.d.) was affixed with a small amount of cyanoacrylate cement to the external naris of the fetus. This required that the fetus’ snout be gently raised momentarily above the surface of the bath to allow the cement to adhere to the skin around the naris. Three mm from the end of the main tube that was attached to the nares, another piece of similar sized tubing was attached to form a manifold for the sensitivity tests (experiment one) and a second tube was attached for the discrimination tests (experiment two). The opposite end of each piece of tubing was attached to a separate perfusion pump which drove a five cc syringe. Initially, air was passed through the main tube to establish that a complete seal had been made between the tubing and the skin around the nares. Air bubbles coming from the mouth without any leakage from the junction between the tubing and the nares was taken as evidence the experiment could proceed. Air was purged from the tubing with physiological saline which was pumped through the nasal cavity by a perfusion pump at a rate of 28 pl/min throughout the remainder of the study. 2.4. Stimulus parameters The stimuli used in this study consisted of two purified odorants (Sigma) diluted to the desired concentration in physiological saline. Iso-amyl acetate and Iso-valeric acid were chosen for a number of reasons. First, they are odorants that have been widely used in other studies on rodents and therefore data is available that is relevant to the interpretation of the results reported here. Second, both are relatively weak trigeminal irritants, and in the case of iso-amyl acetate there exists a large disparity between olfactory and trigeminal threshold. Third, these two odorants have very different odor qualities for humans and presumably for rodents i.e. iso-amyl acetate is a ‘floral’ odor and iso-valeric acid is a ‘putrid’ odor. Guiding the choice of stimulus delivery parameters were estimates of the volume of the fetal nasal cavity made from morphometric analysis of histological sections through the nasal cavity (unpublished data). These estimates, based on two individuals, suggested that together the left and right sides of the fetal nasal cavity had a volume of approximately five mm 3. Therefore, we chose five ~1 as the volume of stimulus in all experiments so we could be certain that the stimulus was not overly diluted by the fhrid volume in the nasal cavity. The perfusion pumps were programmed such that the nominal flow rate would not change when the stimulus was being added to the stream. While actual flow rates were not measured, inspection of the flow of a colored stream into the saline bath confirmed an apparent smooth and constant delivery of the stimulus. Considering the flow rate of the saline purge stream and the estimated naris volume the nasal cavity was evacuated approximately every five seconds during the entire observation period. 2.5. Behavioral
observations
The behavior of each subject fetus was observed and recorded using the computer program The Observer (Vers. 3.0, Aldus, The Netherlands). Observations were done between 9.00 and 17.00 h (L: D 12: 12, lights on at 6.00 am) by a single individual. Within a given experiment a subject fetus was randomly assigned to a treatment condition. ignorant of the particular treatment condition.
Except
as noted below,
the experimenter
was
D.M. Coppnlu, L.C. Millar/Behu~~iourul Processes 39 (1997) 53-68
57
2.5. I. Behavioral categories In a series of preliminary studies the behavioral repertoire of the E 18 mouse fetus was determined under the conditions described above. Nearly all behaviors of the fetus occur as instantaneous point events which allowed the use of frequency counts as the basis for quantitative analysis. Nine behavioral categories were recognized. Six of these were termed ‘basic’ and refer to the region of the body being moved. They include: head, mouth, trunk, front limbs, hind limbs and tail. For the hind limb and front limb categories bilateral and unilateral movements were not distinguished. Three other categories termed ‘complex’ are comprised of behaviors that not only involve simultaneous movements of different regions of the body, but are also stereotyped with respect to the coordination of the movements. The complex categories included: ‘body wiggle’, involving lateral or torsional movements of the trunk, ‘body stretch’, involving dorsal extension of the trunk and head, and ‘wipe’, a movement of one or both front limbs across the snout as if to wipe something from it. 2.5.2. Data reduction and analysis Three variables were derived from the individual behavioral counts that simplified the analysis without compromising the goals of the study. The first, termed ‘behavioral activity’, entailed the summation of all behavioral events during the observation period. The second, termed ‘behavioral complexity,’ included the number of different behavioral categories displayed during the observation period. The third termed ‘behavioral diversity’ was included for completeness since previous authors have favored this index of uncertainty (entropy) which has its origins in information theory (e.g. Smotherman and Robinson, 1986; Paulus and Geyer, 1993). Diversity or ‘H’ is an expression of the overall entropy of behavior expressed in bits per interval. It was calculated as follows: N
H = C P;( -log, i=
Pi),
I
where N equals the number of different behavioral categories and Pi equals the probability that behavior i will occur in the time interval used (see Shannon, 1948 for the complete derivation and justification of this index). The Observer software was used to tally the frequency of each behavioral category across the pre-stimulus and post-stimulus observation periods. These totals were then exported to SYSTAT (Vers. 5.1, Evanston, IL) statistical software for analysis. In experiment one, data were analyzed by repeated measures analysis of variance (ANOVA) using both the univariate and multivariate approach of SYSTAT. The total variance was further partitioned to test a priori hypotheses about certain groups of means. In experiment 2, paired t-tests were used to make a priori comparisons between the results of certain post-stimulus test periods. Since pilot studies established that only increases in behavior occurred with odor stimulation all tests were one-tailed.
3. Experiment 1 Published data on the rat fetus under conditions similar to this study have established that their behavioral output increases when chemical stimuli are delivered to their oral cavity (Smotherman and Robinson, 1985; Smotherman and Robinson, 1988; Smotherman and Robinson, 1989). In a series of
pilot studies (data not shown) we determined that the mouse fetus would react similarly when solutions of iso-amyl acetate or iso-valeric acid were delivered directly into the nares. The purpose of this experiment was to systematically explore the sensitivity of the E 18 mouse fetus to iso-amyl acetate. A two minute pre-stimulus test, during which saline was flushed through the nasal cavity, was followed by three two minute post-stimulus tests that began with odor delivery. The saline pump, which continuously flushed the nasal cavity, was programmed to deliver 28 pl/min except during the time when the second pump delivered the stimulus. During this later period the saline pump delivered at I / 10th the initial rate and the stimulus pump delivered a 5 ~1 bolus at a nominal rate of 25.2 Fl/min. These parameters were chosen to minimize the change in flow rates associated with the addition of the stimulus to the stream. A three-step log dilution series between IO-’ M and lo-’ M iso-amyl acetate was used along with the diluent, physiological saline, as the control. Sixty-one dams yielded between 22 and 24 fetuses for each of the four treatment groups.
4. Results Diverse studies of behavior in the rat and human fetus have demonstrated that their spontaneous motor activity tends to oscillate frequently and irregularly (see Smotherman et al., 1988 and references therein). Fig. 1 illustrates the behavior of two typical subjects from the saline control group. The two subjects shown were somewhat atypical in the greater magnitude of the oscillations but not in the basic pattern of their behavior. The number of behaviors performed were totalled every 20 s throughout the four two minute observation periods. Clearly, the spontaneous behavior of the mouse fetus, like that of the other species studied, appears to oscillate over short intervals. Any measurements of evoked behavior, such as those describe in this study, will have this highly variable spontaneous behavior as a confounding factor. Besides illustrating the oscillatory nature of mouse fetal behavior, Fig. 1 also illustrates for two individuals what was another general trend. Most fetuses showed no detectable increase in the behavioral variables when saline was the stimulus. In fact, while rarely reaching statistical significance (r-tests not shown), the amount, complexity and diversity of behavior in the post-stimulus tests tended to be lower than in the pre-test when saline was the stimulus. The mean changes in behavioral activity engendered by the delivery of odorant or saline are plotted in Fig. 2a as the difference compared to the pre-test. Note that increases in activity were observed after odorant was delivered in most post-stimulus periods for each the three concentrations of the stimulus while a decline in behavioral activity was observed when saline was the stimulus (not significant by r-test). Only in the second and third post-stimulus tests at the highest stimulus concentration did there appear to be a lack of a response to the odorants. The patterns of the individual means were explored statistically in a repeated measures ANOVA with odor concentration as the between-groups factor and test period as the within-groups factor. Both concentration (df = 3, 87; F = 4.6; p < 0.005) and test period (df = 3, 261; F = 3.9; p < 0.01) were significant factors as was their interaction (df = 9, 261; F = I .9; p < 0.05). The pre-stimulus periods were compared to post-stimulus periods combined using the multivariate matrix operations in SYSTAT. The highly significant result further establishes the effect of stimulus delivery in increasing motor activity (df = 4, 87; F = 5.0; p < 0.001).
D.M. Coppolu. L.C. Millur/Beha~wuraI
Processes 39 (1997) 53-68
59
20
Subject
20 1
Pre Test
1st
2nd
3rd
Consecut ive 2 min Intervals Fig. 1. Behavioral
activity
of two fetuses, tallied at 15 s intervals
across four consecutive
saline was added to the saline purge line at the beginning
of each stimulus
the stimulus
which
did not appear to alter spontaneous
activity
is normally
two min test periods. Fifteen
period (1st through
@I of
3rd). Note that the delivery
of
quite irregular.
In Fig. 3 the behavioral activity of two individuals is tallied at 20 second intervals throughout the 8 min test. These results while clearly providing evidence of responses to the stimulus delivery also illustrate the high and variable activity level in the pretest and the variability in the timing and magnitude of the stimulus response. The typical response to the odorant was for the animal to produce a series of sudden head, mouth and trunk movements along with a complex extension of the head and straightening of the torso termed a ‘stretch’. No behavior was found to be unique to a stimulus response, rather the response could be characterized by an increase in the frequency of a given movement and the rapid succession of different movements particularly of the mouth, head, forelimbs and torso. Examination of the results for behavioral complexity revealed a pattern similar to the results for behavioral activity. Fig. 2b, in which mean behavioral complexity is plotted as the difference compared to the pre-stimulus period, reveals an increase in all post-stimulus tests for each of the three stimulus concentrations. Following the pattern of behavioral activity, behavioral complexity decreased
D.M.CoppolaL.C.Millor
60
/ Beho~~iourul
Processes
39
f 1997)
53-68
n
1st 2 min.
q
2nd2min.
q
3rd 2 min.
m
1st 2min.
a
2nd2min.
0
3rd 2 min.
1
Saline
x 10”M
x 1 0-4M
x 10-5M
Stimulus Concentration
-1-I
L Saline
x 10-3M
x 1 0-4M
X 10s5M
Stimulus Concentration Fig. 2. Mean and SEM behavioral responses are plotted after subtracting the pretest activity (before a stimulus was delivered) for three consecutive two min stimulus trials. Five gl of the stimulus solution was added to the saline purge line at the beginning of each stimulus period. The stimuli were saline or one of three dilutions of iso-amyl acetate. (A) Behavioral activity (total number of behaviors) per two min interval is shown. (B) Behavioral complexity (number of different behaviors displayed) per two min interval is shown. Sample sizes ranged from 22 to 24 subjects per stimulus concentration (see text for statistical comparisons)
on average in all post-stimulus tests when saline was the stimulus. These subjective impressions were supported by the statistical analysis. The repeated measures ANOVA, revealed that both concentration (df = 3, 87; F = 4.7; p < 0.004) and test period (df = 3, 261; F = 4.5; p < 0.004) were significant factors as was their interaction (df = 9, 261; F = 3.6; p < 0.0001). The multivariate comparison
D.M. Coppdu. L.C. Millor/Behcrc~iolr~ctl Pro~e.ws
39 (1997) 53-M
61
_”
I
Subject 50
1
lo-4M
1 O-S M )r 16 .E > '4= 2 i? b .-
12
8
zi _c cv 0
Pre Test
1st
Consecutive
2nd
2 min
4 3rd
Intervals
Fig. 3. Behavioral activity of two fetuses, tallied at 1.5 s intervals across four consecutive two min test periods. Five ~1 of iso-amyl acetate (two different concentrations are shown) as added to the saline purge line at the beginning of each stimulus period (1st through 3rd). Note variable rise above spontaneous behavioral activity after each stimulus delivery.
between the pre-stimulus and post-stimulus periods was also significant (df = 4, 87; F = 7.8; p < 0.0001). Results from the analysis of behavioral diversity were similar to the previous variables. Means and SEM for the different concentrations and test periods are provided in Table 1. The repeated measures ANOVA, revealed that both concentration (df = 3, 87; F = 5.0; p < 0.003) and test period (df = 3, 261; F = 5.7; p < 0.001) were significant factors as was their interaction (df = 9, 261; F = 2.9; p < 0.0002). The multivariate comparison between the pre-stimulus and post-stimulus periods was also significant (df = 4, 87; F = 8.0; p < 0.0001). No obvious pattern was evident across the repeated post-stimulus tests for any of the three behavioral variables. For example, only at the highest stimulus concentration did there seem to be a diminution in response with repeated stimulations as one might expect if sensory adaptation was
Table I Means f SEM behavioral diversity (entropy) of the mouse fetus in response to four different iso-amyl acetate concentrations in four consecutive two minute test periods. See text for derivation of variable and results of statistical analysis Test period pretest
Concentration IO-’ IO-’ 10-s Saline
M M M
0.935 + 0.067
0.982+ 0.070 0.834 f 0.069 0.850 & 0.070
I st 2 min
2nd 2 min
3rd 2 min
1.086 f I ,143f I.064 k 0.750 +
0.989 + 0.070 1.164~0.073 1.150+0.072 0.793 * 0.073
0.964 + 0.063 1.138kO.066 I .04 1f 0.064 0.759 &-0.066
0.077 0.080 0.078 0.080
operating. Besides the possibility of adaptation there was no obvious a priori hypothesis concerning the repeated stimulations and therefor these difference were not systematically analyzed. A question left unanswered by the omnibus tests described thus far is whether the different concentrations of odorant engendered different magnitudes of response. This question was addressed using the multivariate matrix operations of SYSTAT to partitions the variance components of the overall test. First, the saline stimulus condition was compared to the three odorant concentrations in a multivariate test. As expected, the results for behavioral activity (df = 4, 84; F = 4.4; p < 0.003), behavioral complexity (df = 4, 84; F = 9.0; p < 0.0001) and behavioral diversity (df = 4, 84; F = 6.2.0; p < 0.0001) were all highly significant demonstrating a strong affect of odorant in increasing behavioral output. Next, having established that the iso-amyl acetate but not saline significantly increased motor activity, the effects of the three concentrations of odorants were compared in a multivariate test. The magnitude of the behavioral response to the different odor concentrations was not significantly different for any of the three variables including behavioral activity (df = 8, 168; F = 1.3; p > 0.051, behavioral complexity (df = 4, 84; F = I. 1; p > 0.05) and behavioral diversity (df = 4, 84; F = 1.4; p > 0.05).
5. Experiment
2
The purpose of this experiment was to determine if the E 18 mouse fetus could discriminate between iso-amyl acetate and iso-valeric acid, each at a concentration of IO-’ M. A habituation paradigm was employed in which repeated stimulations with one odorant were expected to lead to a diminution in response after which the second odorant would be delivered. The protocol was planned after considering the results of a series of pilot studies in which multiple stimulations of iso-amyl acetate where followed by stimulation with iso-valeric acid. The series of behavioral observations and stimulus deliveries were as follows: There were eight one minute observation periods; the first was a pre-test with a continuous flow of saline; the second through sixth were the habituation trials, each of which began with the delivery of 5 ~1 of iso-amyl acetate; the seventh and eighth periods were the discrimination trials which began with the delivery of 5 ~1 of iso-valeric acid. Twelve dams yielded sixteen fetuses for this part of the study. Analysis of the variable diversity was not done for this experiment as it seemed redundant based on the outcome of experiment one.
D.M. Coppolu. L.C. Millur/
Pwcesses 39 (1997) .(3-6X
Beh~iowul
63
6. Results The repeated deliveries of iso-amyl acetate did not result in a detectable diminution of average response as would be expected if habituation were occurring. In fact, though not significantly different. the second and fourth stimulus delivery engendered more behavior on average than the first. pco.02
f 12
A
p< 0.05
c
7
7
,
Q) $
5 =8
.-b iii
0
Amy1 Acetate
n
lsovaleric
Acid
10
8I-
6,-
T
c
l%
Stimulus
Tr
4-
S r”
2
o-
i Consecutive
1
min Intervals
F
p< 0.003
DC 0.0005
7 Stimulus
6
0
Amy1 Acetate
n
lsovaleric Acid
Consecutive 1 min Intervals Fig. 4. Mean and SEM behavioral responses are plotted after subtracting the pretest for seven consecutive one min stimulus trials. Five ~1 of the stimulus solution was added to the saline purge line at the beginning of each stimulus period. The stimulus for the first five trials was iso-amyl acetate and was iso-valeric acid for the last two trials. Both odorants were delivered at 10-j M concentration. (A) Behavioral activity (total number of behaviors) per two min interval is shown. (B) Behavioral complexity (number of different behaviors displayed) per two min interval is shown. Brackets give statistical comparisons between the last iso-amyl acetate trial and the iso-valeric acid trials. N = 16.
However, to establish discrimination, the critical comparison is between the last delivery of iso-amyl acetate and iso-valeric acid. Each of the five post-stimulus tests with iso-amyl acetate caused a significant increase in behavioral activity compared to the pre-test (Fig. 4a). The two trials with iso-valeric acid confirmed that, despite the lack of habituation, this odorant caused a greater response than iso-amyl acetate. Both the first and second delivery of iso-valeric acid caused a significant increase in behavioral activity compared to the pre-test period (I= 4.59, df = 15, p < 0.0002 and r=4.21, df= 15, p < 0.0007) and importantly the mean responses for these tests were significantly different from the response engendered by the previous (5th) delivery of iso-amyl acetate ( f = 1.9 1, df = 15, p < 0.05 and t = 2.18, df = 15, p < 0.02). The data for behavioral complexity agree with those of behavioral activity (Fig. 4b). Probably owing to the smaller sample size in this study, compared to experiment 1, only the 3rd and 4th delivery of iso-amyl acetate engendered a significant response by this measure, however, all the post-stimulus delivery periods showed the same pattern of increased behavior compared to the pre-test. Both the first and second delivery of iso-valeric acid caused a mean response that was significantly different from the pre-test (t = 4.65, df = 15, p < 0.0003 and r = 2.18, df = 15, p < 0.02). Additional confirmation of the fetuses’ ability to discriminate the odorants was provided by the significant differences found between both iso-valeric acid tests and the previous (5th) iso-amyl acetate test (I = 3,91, df = 15, p < 0.0005 and t = 3.24, df = 15, p < 0.003).
7. Discussion The results of experiments one agree with those reported for the rat fetus which have established an increase in behavioral activity after the delivery of chemical stimuli to the naso-oral region (Smotherman, 1982; Smotherman and Robinson, 1985; Smotherman and Robinson, 1988; Smotherman and Robinson, 1989; Smotherman and Robinson, 1990). The current study was, to our knowledge, the first to examine the influence on the fetus of different concentrations of purified odorants delivered directly into the nasal cavity. The results of the first experiment clearly demonstrate that transient increases in behavioral activity and complexity occurred after iso-amyl acetate was added to a stream of physiological saline which continuously purged the fetal nasal cavity. The three concentrations of odorants used were equally effective in eliciting behavior upon their initial delivery, however, the two lowest concentrations appeared to be more reliable in triggering behavioral increases when delivered a second or third time. The loss of response after the first delivery at the highest concentration of iso-amyl acetate was originally thought to be a habituation or accommodation phenomenon but the results of the discrimination studies (see below) which employed the same concentration, argue against this interpretation. In any case, if there was a trend, it was for the lowest concentration to be more reliable than the highest in eliciting a response. The results of experiment two demonstrate that the late-term mouse fetus can discriminate between purified odorants delivered into the nasal cavity. Despite the failure of the protocol used to habituate the animals to repeated stimulation with iso-amyl acetate an inherent and unexpected difference in response to iso-valeric acid was revealed. On average, the animals showed a higher increase in behavioral activity and complexity when stimulated with iso-valeric acid than with isoamyl acetate when both were delivered at a concentration of lo-’ M. However, since the two odorants doubtless have different intensity profiles, the results are consistent with discrimination based on either odor
D.M. Coppol~. L.C. Millctr / Behuriourul
Processes 39 (I 997) 53-68
65
intensity or quality. Similar results have been reported for the rat fetus which demonstrated discrimination between solutions of lemon and mint extracts infused into the mouth (Smotherman and Robinson, 1992). The ability of the rat fetus to detect chemical stimuli before birth (Smotherman, 1982; Smotherman and Robinson, 1985) and to become conditioned to chemicals in the womb such that their prenatal or postnatal behavior is affected (Hepper, 1988; Smotherman, 1982; Smotherman and Robinson, 1985) has fueled considerable speculation about this phenomena’s adaptive significance (e.g. Hepper, 1987; Schaal and Orgeur, 1992; Smotherman and Robinson, 1988). For example, it has been suggested that attraction to food substances in the maternal diet, attraction to maternal pheromones, or even recognition of kin might all be imprinted during gestation. In theory, this could occur through fetal detection of the appropriate chemical stimuli if it gained access to fetal chemosensory systems through the maternal circulation. Olfaction, including participation by the accessory olfactory system, has been the presumed modality involved as it would appear to be the only chemosensory system with the functional capacity in the adult to mediate the behaviors that have been considered. Despite the potential importance of such an adaptation, there is little direct evidence that the olfactory system is functioning prenatally. A recent study demonstrated that isolating the olfactory bulbs from more posterior brain structures causes a diminution in response to chemical stimulation in the mt fetus (Smotherman and Robinson, 1990). Other studies of fetal behavior, not designed to address the question of sensory modality per se, have typically confounded stimulation of the different chemosensory systems of the naso-oral cavity by using very high concentrations or unpurified substances (eg. Smotherman and Robinson, 1985; Smotherman and Robinson, 1988; Smotherman and Robinson, 1989). Anatomical studies, while not definitive, generally support the notion of precocious olfactory competence (see Pedersen et al., 198.5 and Brunjes and Frazier, 1986, for reviews). The only electrophysiological study available, reported electro-olfactograms and single units responses from receptor cells of the prenatal rat (Gesteland et al., 1982). In this study, the olfactory cells responded to relatively low concentrations of purified odorants but did not gain specificity for odorants until the day of birth. The studies reported here provide among the first behavioral evidence that the mammalian fetus has a functioning main olfactory system. Involvement of the accessory olfactory system can be excluded in the mouse since the duct which connects the lumen of this organ to the nasal cavity is not patent before birth (Coppola and O’Connell, 1989; Coppola et al., 1993). Participation of the trigeminal system can also be excluded on the grounds that responses were manifest to concentrations of iso-amyl acetate below the threshold of the trigeminal system for this odor. Tucker (1963). in his now classic studies using recordings from the trigeminal nerve, showed that its threshold for iso-amyl acetate across two different species, tortoise and rabbit, fall in the same narrow range (10-l and IO-‘.” of vapor saturation) well above the lowest concentration used in this study. Subsequently, Silver and Moulton (1982) found strikingly similar thresholds in the rat to this compound. Clearly, rabbits, rats and tortoises constitute a diverse group both phylogenetically and biologically and yet their trigeminal responses are quite similar. Thus, we conclude that our across species comparison is justified. Concerning potential disparities related to the delivery of stimuli in a fluid medium, Tucker ( 1963) also found, using recordings from the olfactory nerve, that the medium (water or air) in which odorant was delivered had little effect on response profiles. Lastly, to exclude trigeminal involvement we rely on comparisons between fetuses and adult animals. However, for our argument to be invalid on these grounds the fetus would have to have a lower trigeminal threshold than the adult. To the
66
D.M. Coppoltr. L.C. MiNur / Behuviotmd
Proce.w.~ 39 (I 997) 53-68
contrary, a gradual increase in chemosensory sensitivity during development would appear to be the rule although clearly more research in this area is needed (Alberts and May, 1980; Apfelbach et al., 1991). It is unknown to what extent the trigeminal or other chemosensory channels contributed to the discrimination by the fetus of iso-amyl acetate and iso-valeric acid. The concentrations used in this part of the study were presumably high enough to stimulate the trigeminal system. While neither odorant is a particularly strong trigeminal irritant there is some evidence that iso-valeric acid is more potent in this regard than iso-amyl acetate (Doty, 1975). Taken together the results from this study challenge some of the prevailing views of chemosensory development. Pedersen et al. (1985) proffered one of the most well delineated models after consideration of the exiting literature and their own findings using the 2-deoxyglucose technique in the rat (Greer et al., 1982; Pedersen et al., 1983; Teicher et al., 1980 see also Alberts, 1976). They suggested that the accessory olfactory system might mediate chemoreception before birth, while the modified-glomerular complex, a specialized region of the olfactory bulb, and a subset of olfactory receptor neurons might mediate chemoreception in the early postnatal period eventually giving way to the main olfactory system which would mediate the reception of airborne chemicals in adulthood. Our findings make this theory untenable in the mouse and thus not generalizable for macrosmatic vertebrates. Despite the lack of a functional accessory olfactory system, the mouse fetus can detect purified odorants at relatively low concentrations and can discriminate between different odorants. It remains possible that some subset of the main olfactory apparatus, like the modified-glomerular complex, might be mediating these processes but there is no indication from our results that the late-term fetus has a functionally incomplete or inept olfactory system. In fact, we found no diminution of response even at the lowest stimulus concentration. Alberts and May ( 1980) found, in a study of odor-induced polypnea in rats from 1 to 17 days of age, that sensitivity to amyl acetate increased with age. Decreasing thresholds for amyl acetate were also found in older rats as a function of increasing age (Apfelbach et al., 1991). Therefore, it might be predicted that the fetus has a lower sensitivity to odorants than adults. A growing literature is at odds with the notion that sensory experience commences at birth in mammals (Hepper, 1992; Schaal and Orgeur, 1992 and references therein). However, despite considerable speculation and circumstantial evidence, it is still unknown whether prenatal chemosensory experience alters postnatal behavior so as to increase survival. The results of this study provide evidence that the main olfactory system or some part of it is functioning in the late-term mouse fetus. Thus, it is possible that this system, which dominates chemoreception of airborne cues in the postnatal period, may begin to acquire experiences with odors in utero. If this process occurs as a natural consequence of fetal development, it certainly will require a reevaluation of what constitutes an ‘innate’ response to chemical stimuli. Much additional work is needed to determine the pervasiveness of this phenomenon and its significance in shaping odor-mediated behaviors that are critical for survival. Acknowledgements The authors would like to gratefully acknowledge the support of the Whitehall Foundation and the NSF (grant BNS 911305 I). Special thanks to Alan Mackay-Sim and his colleagues at Griffith University for extending D.C. their hospitality during the preparation of this manuscript.
67
References Alberts, J.R., 1976. Olfactory contributions to behavioral development in rodents. In: R.L. Doty (Editor), Mammalian Olfaction, Reproductive Processes and Behavior. Academic Press, New York, pp. 67-95. Alberts, J.R. and May, B., 1980. Ontogeny of olfaction: Development of the rat’s sensitivity to urine and amyl acetate. Physiol. Behav., 24: 965-970. Apfelbach, R., Russ, D. and Slotnick, B.M., 1991. Ontogenetic changes in odor sensitivity, olfactory receptor area and olfactory receptor density in the rat. Chem. Sens., 16: 209-218. Basmajian, J.V. and Ranney, D.A., 1961. Chemomyelotomy: substitute for general anesthesia on experimental surgery. J. App. Physiol., 16: 386-392. Brunjes, P.C. and Frazier, L.L., 1986. Maturation and plasticity in the olfactory system of vertebrates. Brain Res. Bull., 11: 1-45. Coppola, D.M., Budde, J. and Millar, L., 1993. The Vomeronasal duct has a protracted postnatal development in the mouse. J. Morphol., 218: 59-64. Coppola, D.M. and O’Connell, R.J., 1989. Stimulus access to olfactory vomeronasal receptors in utero. Neurosci. Lett., 106: 241-248. Doty, R.L., 1975. Intranasal trigeminal detection of chemical vapors by humans. Physiol. Behav., 14: 855-859. Galef, B.G. and Henderson, P.W., 1972. Mother’s milk: A determinant of the feeding preferences of weaning rat pups. J. Comp. Physiol. Psych., 78: 213-219. Galef, B.C. and Sherry, F.D., 1973. Mother”s milk: A medium for transmission of cues reflecting the flavors of mother’s diet. J. Comp. Physiol. Psych., 83: 374-378. Gesteland, R.C., Yancey, R.A. and Farbman, A.I., 1982. Development of olfactory receptor neurons selectivity in the rat fetus. Neuroscience, 7: 3 127-3 136. Graziadei, P.P.C. and Monti Graziadei, G.A., 1978. Continuous nerve cell renewal in the olfactory system. In: M. Jacobsen (Editor), Handbook of Sensory Physiology, Vol. IX. Springer-Verlag, New York, pp. 98-112. Greer, CA., Stewart, W.B., Teicher, M.H. and Shepherd, G.M., 1982. Functional localization of the olfactory bulb and a unique glomerular complex in the neonatal rat. J. Neurosci., 2: 1744-1759. Hepper, P.G., 1987. The amniotic fluid, An important priming role in kin recognition. Anim. Behav., 35: 1343-1346. Hepper, P.G., 1988. Adaptive fetal learning: Prenatal exposure to garlic affects postnatal preferences. Anim. Behav., 36: 935-936. Hepper, P.G., 1992. The quest for the developmental origins of behavior. Q. J. Exp. Psych., 44 B: 161-163. Mistretta, C.M. and Bradley, R.M., 1983. Neural basis of developing taste sensation, Response changes in fetal, postnatal and adult sheep. J. Comp. Neurol., 215: 199-210. Paulus, M.P. and Geyer, M.A., 1993. Three independent factors characterize spontaneous rat motor activity. Behav. Brain. Res., 53: 1 I-20. Pedersen, P.E. and Blass, E.M., 1982. Prenatal and postnatal determinants of the first suckling episode in the albino rat. Dev. Psychobiol., 15: 349-356. Pedersen, P.E., Stewart W.B., Greer, C.A. and Shepherd, G.M., 1985. Early development of olfactory function. In: E.M. Blass (Editor), Handbook of Behavioral Neurology, Vol. 8. Plenum, New York, pp. 163-203. Pedersen, P.E., Stewart, W.B., Cheer, CA. and Shepherd, G.M., 1983. Evidence for olfactory function in utero. Science, 22 1: 478-480. Schaal, B. and Orgeur, P., 1992. Olfaction in Utero: Can the rodent model be generalized. Q. J. Exp. Psych., 44 B: 245-278. Shannon, C.E., 1948. A mathematical theory of communication. Bell Sys. Tech. J., 27: 379-423 and 623-656. Silver, W. and Moulton, D., 1982. Chemosensitivity of rat nasal trigeminal receptors. Physiol. Behav., 28: 927-931. Smotherman, W.P., 1982. Odor aversion learning by the rat fetus. Physiol. Behav., 29: 769-771. Smotherman, W.P. and Robinson, S.R., 1985. The rat fetus in its environment, behavioral adjustments to novel, familiar, aversive and conditioned stimuli presented in utero. Behav. Neurosci., 99: 521-530. Smotherman, W.P. and Robinson, S.R., 1986. Environmental determinants of behavior in the rat fetus. Anim. Behav., 34: 1859-1873. Smotherman, W.P. and Robinson, S.R., 1988. Behavior of rat fetuses following chemical or tactile stimulation. Behav. Neurosci.. 102: 24-34.
6X
D.M. Copplo,
Smotherman,
L.C. Mil/ur/
Behucwurtrl Prncesses 39 f 1997) 53-68
W.P. and Robinson,
S.R., 1989. Rat fetuses respond to chemical
W.P. and Robinson,
S.R.,
stimuli
in gas phase. Physiol.
Behav.,
47:
863-868. Smotherman,
tactile stimulation. Smotherman,
W.P. and Robinson,
Smotherman,
W.P., Robinson,
Comp. Psych., Teicher,
M.H.,
glomerular
1990. Olfactory
bulb transectton
alters fetal behavior
after chemosensory
but not
Dev. Brain Res., 57: 175-180. S.R.. 1992. Habituation
S.R. and Robertson,
in the rat fetus. Q. J. Exp. Psych., 44 B: 215-230.
S.S., 1988. Cyclic
motor activity
in the fetal rat (Rams
J.
noregicus).
102: 78-82.
Stewart, region
W.B.,
Kauer,
in the developing
J.S. and Shepherd, rat olfactory
bulb
G.M.,
1980. Suckling
revealed
pheromone
by the 2-deoxyglucose
stimulation method.
of a modified Brain
Res.,
194:
530-535. Tucker,
D.. 1963. Olfactory,
and Taste I. Pergamon Wang, H., Wysocki, Wysocki,
vomeronasal
C.J. and Gold, G.H.,
C.J., Dorries,
and trigeminal
Press, New York,
1993. Induction
K. and Beauchamp,
receptor responses to odorants.
In: Y. Zotterman
(Editor),
Olfaction
pp. 45-69. G.K.,
anosmic people. Proc. Natl. Acad. Sci. U.S.A.,
of olfactory
1989. Ability 86: 7976-7978.
receptor sensitivity
to perceive
in mice. Science, 260: 998-1000.
androstenone
can be acquired
by ostensibly