Neural correlates of frog calling. Masculinization by androgens

HORMONES

AND

BEHAVIOR

17, 94-102 (1983)

Neural Correlates of Frog Calling. Masculinization by Androgens ROBERT S. SCHMIDT Institute for the Study of Mind, Drugs and Behavior, Loyola University of Chicago Stritch School of Medicine, Maywood, Illinois 60153 Neural correlates of mating calling can be recorded from isolated brainstems of male, Northern leopard frogs after the circuits for this behavior have been triggered by electrical stimulation of the preoptic area. Correlates can be evoked reliably and by a stimulus of low amplitude. However, such correlates can be evoked only rarely from female brainstems, and then only by a much larger stimulus. The sensitivity to triggering in female brainstems can be masculinized by previous treatment of the intact frog with testosterone propionate or dihydrotestosterone, but not by estradiol benzoate. This suggests that the action of the androgens is direct and does not require aromatization to estrogens. Comparisons with other studies suggest that the androgen effect may be mainly on posterior parts of the calling circuits (i.e., call pattern generator or motoneurons), rather than on the preoptic area trigger of the generator.

The males of many species of anurans (toads and frogs) produce a number of vocalizations, including a mating call that attracts conspecific females. Current models of frog calling (Schmidt, 1966, 1976), postulate that the gross, temporal, acoustic patterns of the various calls, as well as of the movements (e.g., by the larynx) producing them, are determined by a pattern generator consisting of the pretrigeminal nuclei of the isthmotrigeminal tegmentum plus more posterior, unidentified, areas of the medulla. The pretrigeminal nucleus may serve also as a sensory correlation center that receives various sensory inputs and determines which, if any, type of call is an appropriate response. In the case of the mating call, the pattern generator can be triggered by the sound of conspecific mating calls-providing androgen levels are sufficiently high. If androgen levels are high enough, mating calling can occur spontaneously, in the absence of any auditory input. The preoptic area, which concentrates androgens (Kelley, Morrell, and Pfaff, 1973, is essential for the triggering of the mating call generator. Indeed, mating calling has been elicited in several species by electrical stimulation, which apparently substitutes for the natural androgen stimulus, of the preoptic area (Schmidt, 1968; Knorr, 94 0018-506x/83 $1.50 Copyright All rights

8 1983 by Academic Press, Inc. of reproduction in any form reserved.

FROG

CALLING

CIRCUIT

ANDROGENIZATION

95

1976; Wada and Gorbman, 1977). This is strictly a triggering action. The preoptic area does not seem to participate in pattern generation. It has been found that the mating calling generator can be triggered by preoptic area stimulation even in a brainstem that has been completely isolated in a tissue bath (Schmidt, 1976). The resulting activity can be monitored by recording from the laryngeal nerves. The temporal patterns of this activity (Fig. l), of the “calling” activity of the laryngeal muscles as monitored by electromyograms (Schmidt, 1974) or by direct visual observation (Schmidt, 1966), and of the sound oscillograms of the calls are closely correlated. Such neural correlates of mating calling have been recorded routinely in this laboratory from the isolated brainstems of mule leopard frogs. However, periodic attempts to do this with female brainstems have been much less successful. Preliminary experiments have suggested that females can be masculinized in this respect (Le., mating calling correlates can be elicited more readily) by pretreatment with androgens. A more formal test, supporting this conclusion, is reported here. METHODS Two shipments (November 4,1981-males and females; April 6, 1982females) of Northern leopard frogs, Rana p. pipiens, from Hazen and Company, Alburg, Vermont were used. These were maintained in the laboratory (Schmidt and Hudson, 1969) until used (March 30-July 11, 1982). Five groups of 10 frogs each were used. Three groups of females received a daily intraperitoneal injection for 9 days of 4 mg of a steroid (estradiol benzoate, dihydrotestosterone, or testosterone propionate, respectively) suspended in I ml of saline. The fourth female group received injections of saline only. During this time, each female was isolated in a separate plastic container for purposes of identification. The fifth group, all untreated males, remained in our maintenance facility until used for recording. The body length range for the females was 78-97 mm (mean 85 mm) and for the males was 77-82 mm (mean 79 mm). No effort was made to adjust the dose of the steroid to the size of the female. Preliminary experiments had shown that 4 mg injections were more than enough to produce an effect and that even higher doses were not toxic. As nearly as possible, frogs were selected for use in such a way as to equalize the size range among groups. Within 5-49 hr (mean 28 hr) following the last injection, a female was prepared for the recording of mating calling correlates. Males were taken directly from our frog colony when needed. Within the limits of scheduling difficulties, the number of complete experiments was kept equal for each group. In preparation for surgery, a frog was placed in ice water for at least

%

ROBERT

S. SCHMIDT

30 min. The brainstem was then removed under hypothermia and placed in a tissue bath with chilled saline aerated with 95% 02/5% CO*. The anterior end of the brainstem was split at the midline back into the infundibular foramen, and the two halves were retracted laterally. This rotated the two halves of the preoptic nuclei such that their medial (ventricular) surfaces were upward and thus exposed to direct visual observation for electrode placement. The laryngeal nerves were placed on electrodes, in mineral oil-filled channels, for recording of correlates. During recordings, the saline was maintained at 23”C, and recording of correlates was not attempted until the saline had been at this temperature for at least 60 min. The pH of the aerated saline at this temperature was about 7.4. The tissue bath has been described in detail (Schmidt, 1976). However, three recent improvements are worth mentioning. The use of Tefloncoated, rather than the original enamel-coated, wire for the laryngeal nerve electrodes has largely eliminated the occasional problem of shorting of electrodes. In the original bath, the platform holding the specimen consisted of the modified bottom of a plastic box. This box was taped to the top of a similar box to complete the bath. Leakage at the tape was a persisent problem. The present bath is identical, except that the box with the specimen platform is inverted and nested within a slightly larger and deeper box, thus making leakage of saline impossible. Also, the saline has been changed slightly (Schmidt, 1980). Published descriptions of amphibian salines differ greatly in pH, osmolarity, and composition. It seems that for most short-term experiments, amphibian tissues can endure a wide variety of salines. The one used here was designed on the basis of preliminary studies of long-term viability. (Correlates have been recorded as long as 48 hr following removal of the brainstem.) The mating calling circuits were triggered by stimulation of the ventral portion of the ventricular surface of the anterior preoptic nucleus (Schmidt, 1974). The concentric stimulating electrode consisted of a 250~pm silver wire core projecting about 500 pm from the end of a hypodermic tubing sheath, both insulated to their tips. The stimulus consisted of a train (maximum duration, 10 set; stimulus stopped as soon as a correlate was detected) of biphasic pulses, each phase 70 Hz and 0.5 msec. These circuits were monitored by recording the activity of the laryngeal nerves on a light-beam osciilograph, after conventional amplification. Efforts to evoke “calling” began by stimulating one side at 100 PA. If a response was obtained, additional stimulations were made (3-10 set following termination of the preceding correlate) in an effort to obtain three consecutive responses. This was then repeated, with successive lowerings of the amplitude by 10 PA, until the lowest effective stimulus was found. In the absence of a response, the stimulus was increased until “calling” occurred, or until 2000 PA had been reached. At least 3

FROG CALLING

CIRCUIT

ANDROGENIZATION

97

min intervened between an amplitude change and resumption of stimulation. This procedure was then repeated on the opposite side. The smallest effective stimulus, from either side, was used for further interpretation. Two types of responses were considered: first, “calling” (a single train of conspicuous pulses followed by a burst of activity) and second, repeatable “calling” (a positive response to at least three consecutive stimuli. RESULTS

A typical male, mating calling correlate is shown in Fig. 1A. Similar “calls” were found in all of the female groups (e.g., Fig. 1B). In the saline and estradiol benzoate groups, the vocal phase sometimes had irregular pulsing and was not followed by an inspiratory phase (Fig. IC). Both types could be found in the same frog (Figs. lB, 1C). The most obvious differences between groups were in various measures of sensitivity of the calling circuits to triggering (Fig. 2). (Because of the difficulty of evoking “calling” in the saline and estradiol benzoate groups

i

i

: I

I

I I

FIG. 1. Oscillograms of neural correlates of mating calling recorded from the isolated brainstem of Northern leopard frogs. Upper trace is from left, and center trace from right, laryngeal nerve. Lower trace is stimulus monitor. Vertical time lines at IOO-msec intervals. (A). Untreated male. Note that the recording consists of the vocal phase, a train of pulses which is the correlate of the laryngeal movements producing the trilled sound of the mating call. followed immediately by the inspiratory phase, a burst of activity which is the correlate of the tonic glottal opening associated with return of air to the lungs. (B). Saline-injected female. Correlate essentially the same as that of the male. (Cl. The correlate immediately preceding the one shown in B. Pulsing is irregular and inspiratory phase absent.

98

ROBERT S. SCHMIDT

9

I 0

S

EP

Stimulus to evoke repeotablo tolling” Mean f S.E.M. ( AI A) 300+0 300*5274 Ran a (aA) (200-400) (380-300)

DHT

TP

156.75232.26

141.11+35.18

55.0024.28

(70-300)

(70-400)

(30-60)

FIG. 2. Effects, compared to normal males (M), of injecting females with saline (S), estradiol benzoate (EB), dihydrotestosterone (DHT), or testosterone propionate (TP) on number of brainstems showing “calling,” repeatable “calling,” or repeatable “calling” evokable by a stimulus of less than 95 PA, and on the mean and range of the minimum stimulus evoking repeatable “calling.”

and, therefore, the small number of “calls” available for study, it did not seem practical to attempt a detailed analysis of the characteristics of the correlate oscillograms.) Number of frogs “calling” (i.e., in which at least one “call” could be evoked). Both the saline (x2 = 5.95, P = 0.015) and estradiol benzoate (x2 = 4.27, P = 0.04) groups were significantly less than the male group but not different from each other (x2 = 0, P > 0.99). The testosterone propionate group was significantly greater than the saline group (x2 = 5.95, P = 0.015) but not different from the male group (x2 = 0, P > 0.99). The dihydrotestosterone group was almost significantly greater than the saline group (x2 = 3.52, P = 0.06) but not different from the male group (x2 = 0, P > 0.99). The dihydrotestosterone group was not significantly less than the testosterone propionate group (x2 = 0, P > 0.99). Thus, testosterone propionate clearly, and dihydrotestosterone possibly, masculinized the female brainstems, while estradiol benzoate did not. Number of frogs showing repeatable “calling” (i.e., at least three consecutive “calls” could be evoked). Both the saline (x2 = 10.21, P < 0.001) and the estradiol benzoate (x2 = 7.91, P = 0.005) groups were significantly less than the male group but not different from each other (x2 = 0, P > 0.99). The testosterone propionate group was significantly greater than the saline group (x2 = 7.27, P = 0.007) but not different

FROG

CALLING

CIRCUIT

ANDROGENIZATION

99

from the male group (x2 = 0, P > 0.99). The dihydrotestosterone group was significantly greater than the saline group (x2 = 5.00, P = 0.03) but did not differ from the male group (x2 = 0.55, P = 0.45). The dihydrotestosterone group was not significantly less than the testosterone propionate group (x2 = 0, P > 0.99). Thus, both androgens clearly masculinized the female brainstems, while estradiol benzoate did not. Number of frogs showing repeatable “calling” that could be evoked with a stimulus of less than 95 PA. Both the saline and estradiol benzoate groups were significantly less than the male group (x2 = 16.2, P < 0.001) but not different from each other (x2 = 0, P > 0.99). The testosterone

propionate group was not significantly greater than the saline group (x2 = 2.81, P = 0.09) but was significantly less than the male group (x2 = 5.95, P = 0.015). The dihydrotestosterone group was not significantly greater than the saline group (x2 = 0, P > 0.99) but was significantly less than the male group (x2 = 12.93, P < 0.001). The dihydrotestosterone group was not significantly less than the testosterone propionate group (x2 = 1.07, P = 0.30). Thus none of the steroids masculinized the female brainstems. Mean of the minimum stimulus needed to evoke repeatable “calling” (Mann-Whitney U test). Both the saline (P = 0.025) and estradiol benzoate (P = 0.005) groups were significantly greater than the male group but not different from each other (P > 0.1). The testosterone propionate group did not differ from the saline group (P = 0.1) but was significantly greater than the male group (P < 0.001). The dihydrotestosterone group did not differ from the saline group (P = 0.1) but was significantly greater than the male group (P < 0.001). The dihydrotestosterone group was not significantly greater than the testosterone propionate group (P > 0. I).

Thus, none of the steroids masculinized

the female brainstems.

Range of minimum stimulus needed to evoke repeatable “calling” and the lowest minimum stimulus. Three groupings could be seen: the saline

plus estradiol benzoate groups (about 300 PA minimum stimulus), the two androgen groups (about 150 PA), and the males (about 55 PA). There was very little overlap between groups. DISCUSSION

Mating calling correlates were elicited readily in the males (Fig. 2). Not only could “calling” be evoked in each frog, but this could be done repeatedly and at a stimulus of only 80 PA or less. By all criteria used, the mating calling circuits of the saline-injected females were significantly much less sensitive to triggering. There was no evidence that estradiol benzoate treatment masculinized the mating calling circuits of the females. On the other hand, testosterone propionate clearly masculinized the calling circuits by two of the four criteria treated statistically. The lack

100

ROBERT

S. SCHMIDT

of statistical significance in the case of the number of frogs giving repeatable “calling” evoked by a stimulus of less than 95 PA (P = 0.09) or of the mean minimum stimulus needed to evoke repeatable “calling” (P = 0.1) was probably due to the small number of animals used. The masculinizing effects of dihydrotestosterone resembled those of testosterone propionate, but were probably less. Although there was no statistically significant difference between the two androgens with the number of animals used here, it will be noted in Fig. 2 that (with the exception of the smallest and largest minimum stimulus evoking repeatable “calling”) dihydrotestosterone consistently showed lower measures of masculinization. Figure 2 also shows that, except for the effect of testosterone propionate on the number of “calling” frogs, both androgens consistently showed incomplete masculinization. By several criteria, this difference from the males was statistically significant. Such a difference is not surprising, considering the many differences between any experimental application of steroids and the natural hormonal mechanisms of control of the calling circuits. To what extent more complete masculinization could be obtained by experimental manipulation remains to be determined. The masculinizing effect of dihydrotestosterone (a nonaromatizable androgen), but not of estradiol benzoate, suggests that the action of the androgens was direct, rather than through aromatization to estrogens. Similar steroid effects have been found on the pretrigeminal nucleus (Schmidt, 1974) of the American toad (Schmidt, 1982), i.e., treatment with either testosterone propionate or dihydrotestosterone, but not with estradiol benzoate, leads to masculinization (an increase in number of large, succinic dehydrogenase staining neurons) of the female nucleus. (The pretrigeminal nucleus, just anterior to the trigeminal motor nucleus, is involved in calling.) Autoradiographic studies of Xenopus, the South African clawed frog (Kelley, 1980, 1981 et al., 1975; Morrell, Kelley, and Pfaff, 1975), have shown concentration of radioactivity in the preoptic area following application of labeled testosterone or estradiol, but not of dihydrotestosterone, suggesting that androgen action in this structure involves atomatization to estrogens. On the other hand, these studies show concentration of radioactivity in the pretrigeminal nucleus and laryngeal motorneurons following application of labeled testosterone or dihydrotestosterone, but not of estradiol, suggesting that androgen action in these structures is direct. These data suggest that the masculinizing effects of androgens on the female mating calling circuits studied here may be due mainly to a direct action of androgens on the more posterior parts of the circuits (i.e., the call pattern generator and/or motoneurons), rather than to an indirect action (through aromatization to estrogens) on the preoptic area trigger of this generator. It is interesting that essentially normal, male-like mating calling correlates

FROG CALLING

CIRCUIT

ANDRGGENIZATION

101

could be found even in some of the saline- and estradiol benzoate-treated females (Fig. 1B). The occasional “abnormal” correlate (Fig. 1C) is most easily interpreted as resulting from inadequate triggering of the mating call pattern generator, especially since both normal and “abnormal” correlates could be found in the same animal. It thus appears that females already have complete mating calling circuits and that the effect of the androgens was to increase the sensitivity of these circuits to triggering by preoptic area input. This is consistent with the finding that female green tree frogs (Hyla cinerea) will give mating calling movements in response to acoustic stimulation by conspecific mating calls, if their body cavities have been packed first with frog testes and anterior pituitaries (Schmidt, 1966). One mechanism of the masculinization of the female calling circuits might be an androgen-induced increase in the dendritic trees of the pretrigeminal nucleus neurons, thus providing an enlarged target for the preoptic area triggering input. As noted above, androgenization of female toads increases the size of the pretrigeminal neurons (Schmidt, 1982) and others have shown induction of dendritic growth by androgens (e.g., DeVoogd and Nottebohm, 1981). ACKNOWLEDGMENTS This work was supported by NINCDS Grant NS-06673. The electronic facilities were set up and maintained by Mr. Wayne Hudson.

REFERENCES DeVoogd, T., and Nottebohm, F. (1981). Gonadal hormones induce dendritic growth in the adult avian brain. Science 214, 202-204. Kelley, D. B. (1980). Auditory and vocal nuclei in the frog brain concentrate sex hormones. Science

207,

553-555.

Kelley, D. B. (1981). Locations of androgen-concentrating cells in the brain of Xenopus laevis: Autoradiography with ‘H-dihydrotestosterone. J. Camp. Neural. 199, 221-231. Kelley, D. B., Morrell, J. I., and Pfaff, D. W. (1975). Autoradiographic localization of hormone-concentrating cells in the brain of an Amphibian, Xenopus laevis. I. Testosterone. J. Comp. Neurol. 164, 47-62. Knorr, A. (1976). Central control of mating call production and spawning in the green frog Hyla savignyi (Audouin): Results of electrical stimulation of the brain. Behav. Proc. 1, 295-317. Morrell, J. I., Kelley. D. B., and Pfaff, D. W. (1975). Autoradiographic localization of hormone-concentrating cells in the brain of an amphibian, Xenopus laevis. II. Estradiol. J. Comp.

Neurol.

164,

63-78.

Schmidt, R. S. (1966). Central mechanisms of frog calling. Behaviour 26, 251-285. Schmidt, R. S. (1968). Preoptic activation of frog mating behavior. Behaviour 30, 239257. Schmidt, R. S. (1974). Neural correlates of frog calling. Trigeminal tegmentum. J. Camp. Physiol.

92,

229-254.

Schmidt, R. S. (1976). Neural correlates of frog calling. Isolated brainstem. Physiol. 108, 99-l 13.

J. Comp.

102

ROBERT S. SCHMIDT

Schmidt, R. S. (1982). Masculinization of toad pretrigeminal nucleus by androgens. Brain Res. 244, 190-192. Schmidt, R. S., and Hudsbn, W. R. (1969). Maintenance of adult anurans. Lnb. Anim. Care 19, 617-620. Wada, M., and Gorbman, A. (1977). Mate calling induced by electrical stimulation in freely moving leopard frogs, Ram pipiens. Hot-m. Behav. 9, 141-149.