Brain Research, 99 (1975) 69-83 (© Elsevier Scientific Publishing Company, Amsterdam - Printed in The Netherlands
E F F E C T S OF I N S U L I N I N J E C T I O N O N RESPONSES BULB A N D A M Y G D A L A S I N G L E U N I T S TO O D O R S
69
OF O L F A C T O R Y
DONALD P. CAIN* Department of Psychology, McGill University, Montreal, Quebec H3C 3G1 (Canada) (Accepted May 20th, 1975)
SUMMARY
The effect of insulin injection on transmission of neural activity within the olfactory system of the anesthetized male rat was investigated at the single unit level. It was found that insulin changed the response to odors of approximately 27 ~,] of olfactory bulb units and 21 ~o of amygdala units tested. Many ~f the changes were in the direction of an increase in response magnitude, but there were some reversals in response direction and other complex changes. There was no evidence of a selective facilitation of responses to food odor as compared to non-food odors. Control observations of the response of thalamic somatosensory units to tactual stimulation showed no effects of insulin. These results suggest that hypothalamic hunger mechanisms may normally interact with olfactory mechanisms to augment and otherwise change the response of some olfactory system units to various odors.
INTRODUCTION
Centrifugal control mechanisms are known to be capable of modifying transmission within certain sensory systems beginning at the receptor level. For example, the mammalian muscle spindle is, the avian retina 3°, and the reptilian taste receptor 6 have been shown to be under some degree of centrifugal control from more central structures. Certain relay structures within sensory systems also seem to be subject to centrifugal control2L In its simplest form the centrifugal control mechanism seems to function as a servomechanism governing receptor sensitivity 18, or as a means of preventing the consequences of an animal's own movements from disrupting the operation of its sensory apparatus 44. Evidence is available which suggests that centrifugal control systems may also be involved in higher level functions concerned with motivated behavior. For example, Flynn and coworkers have analyzed the attack behavior of the cat that is elicited by * Present address: Kinsman Laboratory of Neurological Research, Faculty of Medicine, University of British Columbia, Vancouver, B.C.V6T, 1 WS, Canada.
70 electrical stimulation of the hypothalamus, and shown that the stimulation establishes or enlarges sensory fields that are normally operative during the course of any attack ~. a,24 Similar interactions may take place between sensory systems and hormonedependent mechanisms concerned with sexual behavior 2z,23,a6,aT. However, the nature of the interactions at the microphysiological level is almost completely unknown. The possible influence of neural mechanisms concerned with the regulation of food intake on transmission of information within the olfactory system would seem to be a fruitful question for study. The olfactory system has been investigated intensively in both its neuroanatomical and electrophysiological aspects within recent years, and neural mechanisms underlying regulation of food intake have been the subject of interest for some years. Recent neuroanatomical findings link both centripetal and centrifugal olfactory projections with basal forebrain structures implicated in the regulation of food intake 39,4°. In the present study responses of olfactory bulb units to odors before and after injection of insulin, which is known to induce feeding in satiated rats 4, were compared, with the expectation that the response properties might be changed after the injection. Recordings were obtained from units in the mitral cell layer since efferent fibers to the olfactory receptors have not been reported. Recordings were also obtained from units in the amygdala, since this structure is known to receive olfactory input 7,19,38 and input from forebrain structures involved in the regulation of feeding1°. METHODS
Data were obtained from 152 male hooded rats (Canadian Breeding Farms, 285-500 g at the time of surgery) prepared acutely for single unit recording from either the olfactory bulb or amygdala, in an apparatus that allowed for the controlled presentation of odor stimuli. The response properties of a unit to 2-5 food and non-food odorants were determined, followed by subcutaneous control injection of saline, followed by a redetermination of response properties, and injection of a fast-acting form of insulin (Insulin-Toronto, Connaught) that had previously been shown to induce a state behaviorally indistinguishable from that resulting from food deprivation (ref. 4 and Cain, unpublished observations). The optimum dose o f 10 units/kg (ref. 4 and Cain, unpublished observations) was used in all experiments. The response properties of the unit were again determined using the same odorants under identical conditions of presentation: As a control for possible non-specific facilitation of sensory responses by insulin, units in nucleus ventralis posterior were tested with somatosensory stimulation in a paradigm analogous to that outlined above.
Surgical and recording procedures The animals were maintained on an ad libitum feeding schedule until the time of surgery, at which time they were anesthetized with urethane (1.5 g/kg i.p.) and atropine was administered (0.3 mg/kg s.c.). A tracheotomy was performed and a breathing tube was inserted into the trachea. A second cannula was introduced through the trachea into the back of the nasal cavity; a vacuum applied to the cannula allowed
71 odorized air to be drawn into the nose at a controlled rate. A hypodermic needle attached to a length of polyethylene tubing and a syringe filled with physiological saline was placed subcutaneously in the dorsal thoracic region. At the appropriate time the syringe could be exchanged for one containing insulin, and the insulin so injected. For some experiments another hypodermic needle attached to a syringe was inserted into the peritoneum for injection of 10 ~o dextrose. Glass micropipettes filled with 2.8 M NaC1 were positioned, using a hydraulic microdrive and standard stereotaxic techniques, according to the atlas of De G r o o t 13. Olfactory bulb placements were intended for the mitral cell layer. Input from the micropipette was led via a high-impedance probe to a Grass P511 preamplifier, thence to an oscilloscope, and Schmitt trigger and gating unit with audiomonitor. Recording was single-ended, with the reference electrode clipped to the skin at the wound margin. In general, half-amplitude filters on the preamplifier were set to pass frequencies between 30 Hz and 30 kHz, except during EEG recording at the micropipette tip, when the bandpass was set for 1 Hz-30 kHz. Cortical E E G was recorded and continuously monitored during most experiments from two screws set in the bone over the parietal cortex. Heart rate was monitored throughout most experiments with a tachograph. Permanent records of all experimental sessions were made on magnetic tape using a four-channel A M - F M data tape recorder, and were transformed into permanent polygraph records. Unit activity from selected experimental sessions was photographed with a continuous recording camera. In addition to the olfactory bulb and amygdala placements 24 control experiments were performed with placements in the nucleus ventralis posterior of the thalamus, which is the main thalamic relay for somatosensory input 43. In these experiments somatosensory stimulation was given in a paradigm analogous to the olfactory stimulation experiments to control for the possibility that any observed insulin effects might be due to non-specific facilitation of all sensory systems, or a direct effect of insulin on brain cells. When a stable unit was observed, its somatosensory receptive field was determined with a thin blunt rod. Somatosensory stimulation came from a long-reach camera cable release positioned rigidly over the receptive field. Repeatable stimulation was given by allowing a hinged weight to fall on the fixed end of the cable release.
Odor stimulation For odor stimulation, air from a regulated line was led to a system of glass drying bottles filled with indicating 'Drierite', activated charcoal, and distilled water to dry, clean and rehumidify the air, and passed by the nose of the rat at a known rate, measured with a flow-meter. Odor-saturated air from a syringe could be added to the cleaned air flow at a special junction. Air flow rates through the nose varied from preparation to preparation, but averaged 200 ml/min. The nose air flow rate was monitored with a flowmeter, and during each experiment was kept constant. A water manometer connected to a line leading from the nose tube provided a further check on possible nasal flow rate changes due to constriction or dialatation of the nasal passages resulting from autonomic influences 49. The relative dilutions of odors used ranged from 1/20 to 1/5 of saturation. Stimulation was 2-5 sec in duration. The effec-
72 tiveness of the odors in eliciting unit activity both before and after the various injections was always compared at the same relative dilution and duration. The purified compounds used for odor stimulation were amyl acetate, benzylamine, butyric acid, cineole, and quinoline. All were obtained as high grade liquids, The food mash stimulus consisted of 50 g of rat chow powder, made from chow pellets taken from the rats' daily food supply, mixed with distilled water to form a wet mash. Fresh odorants were placed in individually marked and capped glass flasks prior to each experiment. Odor-saturated air could be transferred from a flask to the cleaned air stream by withdrawing air from the flask through the cap with a syringe and injecting the air through a fitting on the apparatus. Separate syringes were used for each odor.
Histological and statistical procedures After each experiment the brains were fixed by perfusion with normal saline and 10 ~ formalin with the micropipette in place at the bottom of the penetration track. For the olfactory bulb penetrations it was necessary after the recording session, but before perfusion, to advance the microdrive until the micropipette penetrated nearly the full depth of the olfactory bulb to adequately visualize the micropipette track during subsequent histological analysis. Serial sections were cut at 30/~m on a cryotome and stained with cresyl violet. The location of each cell recorded from was reconstructed by visual inspection of the micropipette track through a microscope with the aid of depth records at which each cell was encountered taken during the experiments. Counts of unit discharges during presentation of olfactory and somatosensory stimuli were obtained using a Computer of Average Transients (CAT 1000) accessory counting circuit. A repeated measures analysis of variance was performed on the baseline data in order to obtain an estimate of within-subjects variability to be used in determining the significance of changes in responses to stimuli after insulin. For this purpose the responses to amyl acetate (in spikes/sec) of units tested to that odor 3 or more times before injection of insulin were tabulated separately for the olfactory bulb and amygdala. A separate baseline tabulation of responses to somatosensory stimulation by nucleus ventralis posterior units was also made. Data from the first 3 baseline presentations were used in the analysis, giving 28 subjects in the olfactory bulb group, 26 subjects in the amygdala group, and 22 subjects in the nucleus ventratis posterior group. The data were normalized using the transformation: X' = x/X x/X ÷ 1. The interaction of subjects by baseline stimulus presentations gives an estimate of within-subjects variability, and the square root of the corresponding mean square represents the standard error of an observation. A unit was considered to have changed its response to a stimulus if the frequency of discharge to the stimulus after injection of insulin was outside the range of X ± 1.96 (S.E.), where X -- mean baseline response to the stimulus, and S.E. ~- standard error of an observation. This analysis was appropriate for units which showed simple changes after insulin, such as responses after insulin where there were none before insulin, or potentiation of a pre-existent response. However, results from a number of experiments were striking in that there was a repatterning of discharge to the olfactory
73
~ 0
r~
I
E F F E C T S OF I N S U L I N I N J E C T I O N O N RESPONSES BULB A N D A M Y G D A L A S I N G L E U N I T S TO O D O R S
69
OF O L F A C T O R Y
DONALD P. CAIN* Department of Psychology, McGill University, Montreal, Quebec H3C 3G1 (Canada) (Accepted May 20th, 1975)
SUMMARY
The effect of insulin injection on transmission of neural activity within the olfactory system of the anesthetized male rat was investigated at the single unit level. It was found that insulin changed the response to odors of approximately 27 ~,] of olfactory bulb units and 21 ~o of amygdala units tested. Many ~f the changes were in the direction of an increase in response magnitude, but there were some reversals in response direction and other complex changes. There was no evidence of a selective facilitation of responses to food odor as compared to non-food odors. Control observations of the response of thalamic somatosensory units to tactual stimulation showed no effects of insulin. These results suggest that hypothalamic hunger mechanisms may normally interact with olfactory mechanisms to augment and otherwise change the response of some olfactory system units to various odors.
INTRODUCTION
Centrifugal control mechanisms are known to be capable of modifying transmission within certain sensory systems beginning at the receptor level. For example, the mammalian muscle spindle is, the avian retina 3°, and the reptilian taste receptor 6 have been shown to be under some degree of centrifugal control from more central structures. Certain relay structures within sensory systems also seem to be subject to centrifugal control2L In its simplest form the centrifugal control mechanism seems to function as a servomechanism governing receptor sensitivity 18, or as a means of preventing the consequences of an animal's own movements from disrupting the operation of its sensory apparatus 44. Evidence is available which suggests that centrifugal control systems may also be involved in higher level functions concerned with motivated behavior. For example, Flynn and coworkers have analyzed the attack behavior of the cat that is elicited by * Present address: Kinsman Laboratory of Neurological Research, Faculty of Medicine, University of British Columbia, Vancouver, B.C.V6T, 1 WS, Canada.
70 electrical stimulation of the hypothalamus, and shown that the stimulation establishes or enlarges sensory fields that are normally operative during the course of any attack ~. a,24 Similar interactions may take place between sensory systems and hormonedependent mechanisms concerned with sexual behavior 2z,23,a6,aT. However, the nature of the interactions at the microphysiological level is almost completely unknown. The possible influence of neural mechanisms concerned with the regulation of food intake on transmission of information within the olfactory system would seem to be a fruitful question for study. The olfactory system has been investigated intensively in both its neuroanatomical and electrophysiological aspects within recent years, and neural mechanisms underlying regulation of food intake have been the subject of interest for some years. Recent neuroanatomical findings link both centripetal and centrifugal olfactory projections with basal forebrain structures implicated in the regulation of food intake 39,4°. In the present study responses of olfactory bulb units to odors before and after injection of insulin, which is known to induce feeding in satiated rats 4, were compared, with the expectation that the response properties might be changed after the injection. Recordings were obtained from units in the mitral cell layer since efferent fibers to the olfactory receptors have not been reported. Recordings were also obtained from units in the amygdala, since this structure is known to receive olfactory input 7,19,38 and input from forebrain structures involved in the regulation of feeding1°. METHODS
Data were obtained from 152 male hooded rats (Canadian Breeding Farms, 285-500 g at the time of surgery) prepared acutely for single unit recording from either the olfactory bulb or amygdala, in an apparatus that allowed for the controlled presentation of odor stimuli. The response properties of a unit to 2-5 food and non-food odorants were determined, followed by subcutaneous control injection of saline, followed by a redetermination of response properties, and injection of a fast-acting form of insulin (Insulin-Toronto, Connaught) that had previously been shown to induce a state behaviorally indistinguishable from that resulting from food deprivation (ref. 4 and Cain, unpublished observations). The optimum dose o f 10 units/kg (ref. 4 and Cain, unpublished observations) was used in all experiments. The response properties of the unit were again determined using the same odorants under identical conditions of presentation: As a control for possible non-specific facilitation of sensory responses by insulin, units in nucleus ventralis posterior were tested with somatosensory stimulation in a paradigm analogous to that outlined above.
Surgical and recording procedures The animals were maintained on an ad libitum feeding schedule until the time of surgery, at which time they were anesthetized with urethane (1.5 g/kg i.p.) and atropine was administered (0.3 mg/kg s.c.). A tracheotomy was performed and a breathing tube was inserted into the trachea. A second cannula was introduced through the trachea into the back of the nasal cavity; a vacuum applied to the cannula allowed
71 odorized air to be drawn into the nose at a controlled rate. A hypodermic needle attached to a length of polyethylene tubing and a syringe filled with physiological saline was placed subcutaneously in the dorsal thoracic region. At the appropriate time the syringe could be exchanged for one containing insulin, and the insulin so injected. For some experiments another hypodermic needle attached to a syringe was inserted into the peritoneum for injection of 10 ~o dextrose. Glass micropipettes filled with 2.8 M NaC1 were positioned, using a hydraulic microdrive and standard stereotaxic techniques, according to the atlas of De G r o o t 13. Olfactory bulb placements were intended for the mitral cell layer. Input from the micropipette was led via a high-impedance probe to a Grass P511 preamplifier, thence to an oscilloscope, and Schmitt trigger and gating unit with audiomonitor. Recording was single-ended, with the reference electrode clipped to the skin at the wound margin. In general, half-amplitude filters on the preamplifier were set to pass frequencies between 30 Hz and 30 kHz, except during EEG recording at the micropipette tip, when the bandpass was set for 1 Hz-30 kHz. Cortical E E G was recorded and continuously monitored during most experiments from two screws set in the bone over the parietal cortex. Heart rate was monitored throughout most experiments with a tachograph. Permanent records of all experimental sessions were made on magnetic tape using a four-channel A M - F M data tape recorder, and were transformed into permanent polygraph records. Unit activity from selected experimental sessions was photographed with a continuous recording camera. In addition to the olfactory bulb and amygdala placements 24 control experiments were performed with placements in the nucleus ventralis posterior of the thalamus, which is the main thalamic relay for somatosensory input 43. In these experiments somatosensory stimulation was given in a paradigm analogous to the olfactory stimulation experiments to control for the possibility that any observed insulin effects might be due to non-specific facilitation of all sensory systems, or a direct effect of insulin on brain cells. When a stable unit was observed, its somatosensory receptive field was determined with a thin blunt rod. Somatosensory stimulation came from a long-reach camera cable release positioned rigidly over the receptive field. Repeatable stimulation was given by allowing a hinged weight to fall on the fixed end of the cable release.
Odor stimulation For odor stimulation, air from a regulated line was led to a system of glass drying bottles filled with indicating 'Drierite', activated charcoal, and distilled water to dry, clean and rehumidify the air, and passed by the nose of the rat at a known rate, measured with a flow-meter. Odor-saturated air from a syringe could be added to the cleaned air flow at a special junction. Air flow rates through the nose varied from preparation to preparation, but averaged 200 ml/min. The nose air flow rate was monitored with a flowmeter, and during each experiment was kept constant. A water manometer connected to a line leading from the nose tube provided a further check on possible nasal flow rate changes due to constriction or dialatation of the nasal passages resulting from autonomic influences 49. The relative dilutions of odors used ranged from 1/20 to 1/5 of saturation. Stimulation was 2-5 sec in duration. The effec-
72 tiveness of the odors in eliciting unit activity both before and after the various injections was always compared at the same relative dilution and duration. The purified compounds used for odor stimulation were amyl acetate, benzylamine, butyric acid, cineole, and quinoline. All were obtained as high grade liquids, The food mash stimulus consisted of 50 g of rat chow powder, made from chow pellets taken from the rats' daily food supply, mixed with distilled water to form a wet mash. Fresh odorants were placed in individually marked and capped glass flasks prior to each experiment. Odor-saturated air could be transferred from a flask to the cleaned air stream by withdrawing air from the flask through the cap with a syringe and injecting the air through a fitting on the apparatus. Separate syringes were used for each odor.
Histological and statistical procedures After each experiment the brains were fixed by perfusion with normal saline and 10 ~ formalin with the micropipette in place at the bottom of the penetration track. For the olfactory bulb penetrations it was necessary after the recording session, but before perfusion, to advance the microdrive until the micropipette penetrated nearly the full depth of the olfactory bulb to adequately visualize the micropipette track during subsequent histological analysis. Serial sections were cut at 30/~m on a cryotome and stained with cresyl violet. The location of each cell recorded from was reconstructed by visual inspection of the micropipette track through a microscope with the aid of depth records at which each cell was encountered taken during the experiments. Counts of unit discharges during presentation of olfactory and somatosensory stimuli were obtained using a Computer of Average Transients (CAT 1000) accessory counting circuit. A repeated measures analysis of variance was performed on the baseline data in order to obtain an estimate of within-subjects variability to be used in determining the significance of changes in responses to stimuli after insulin. For this purpose the responses to amyl acetate (in spikes/sec) of units tested to that odor 3 or more times before injection of insulin were tabulated separately for the olfactory bulb and amygdala. A separate baseline tabulation of responses to somatosensory stimulation by nucleus ventralis posterior units was also made. Data from the first 3 baseline presentations were used in the analysis, giving 28 subjects in the olfactory bulb group, 26 subjects in the amygdala group, and 22 subjects in the nucleus ventratis posterior group. The data were normalized using the transformation: X' = x/X x/X ÷ 1. The interaction of subjects by baseline stimulus presentations gives an estimate of within-subjects variability, and the square root of the corresponding mean square represents the standard error of an observation. A unit was considered to have changed its response to a stimulus if the frequency of discharge to the stimulus after injection of insulin was outside the range of X ± 1.96 (S.E.), where X -- mean baseline response to the stimulus, and S.E. ~- standard error of an observation. This analysis was appropriate for units which showed simple changes after insulin, such as responses after insulin where there were none before insulin, or potentiation of a pre-existent response. However, results from a number of experiments were striking in that there was a repatterning of discharge to the olfactory
73
~ 0
r~
I
