Behavioral Neuroscience 1994, Vol. 108, No. 2, 379-394
Copyright 1994 by the American Psychological Association, Inc. 0735-7044/94/S3.00
Expression of Fos-Like Immunoreactivity in the Preoptic Area of Maternally Behaving Virgin and Postpartum Rats
Michael Numan and Marilyn J. Numan
This study uses Fos immunocytochemistry to show that the medial preoptic area and ventral bed nucleus of the stria terminalis are activated in maternally behaving female rats. In Experiment 1, virgin female rats that showed maternal behavior toward pups had more cells in these regions that expressed Fos-like immunoreactivity than did virgin females that were not maternally responsive. In Experiment 2, postpartum rats that were exposed to pups and showed maternal behavior had more Fos-labeled cells in these regions than did postpartum rats exposed to candy. Evidence also indicated that functional modifications in the medial amygdala were related to the changes in Fos expression observed in the preoptic area and ventral bed nucleus of the stria terminalis.
The preoptic area is critical for the normal display of maternal behavior in the rat. Cell-body-specific lesions of the medial or lateral preoptic area, produced by intracerebral injection of an excitotoxic amino acid, disrupt maternal behavior (Numan, Corodimas, Numan, Factor, & Piers, 1988), and injections of estradiol or prolactin into the medial preoptic area stimulate the behavior (Bridges, Numan, Ronsheim, Mann, & Lupini, 1990; Numan, Rosenblatt, & Komisaruk, 1977). In addition, endogenous opiates appear to act on the medial preoptic area to depress maternal behavior (Mann & Bridges, 1992). In a related line of research, investigators have attempted to fit the preoptic area into a larger neural circuitry underlying maternal behavior. An early finding was that knife cuts that sever the lateral connections of the medial preoptic area disrupt maternal behavior (Numan, 1974), suggesting that the lateral efferent projections from the medial preoptic area may be most critical (also see Numan & Callahan, 1980; Numan, McSparren, & Numan, 1990; and Terkel, Bridges, & Sawyer, 1979). Subsequent research has provided strong evidence that a descending projection from the preoptic area to the brainstem is involved (Numan & Numan, 1991; Numan & Smith, 1984). The lateral efferents of the preoptic area project to several brainstem regions (Conrad & Pfaff, 1976; Simerly & Swanson, 1988; Swanson, 1976), but the site or sites that are influenced by the preoptic area in its regulation of maternal responsiveness have not been determined (Numan & Numan, 1991). The rationale of the present study is that one might be better able to locate the brainstem site or sites to which preoptic area efferents project if one first determines the location of the specific neurons within the preoptic area that are involved in
Michael Numan and Marilyn J. Numan, Department of Psychology, Boston College. This research was supported by a grant from the Whitehall Foundation. The assistance of Eric Huerter, Jennifer Lee, Jeff Petruska, and Leigh Vaughan is greatly appreciated. We thank M. Potegal for his advice on the Fos immunocytochemistry procedure. Correspondence concerning this article should be addressed to Michael Numan, Department of Psychology, McGuinn Hall, Boston College, Chestnut Hill, Massachusetts 02167. Electronic mail may be sent to firstname.lastname@example.org.
controlling maternal behavior. Once these neurons are located, one could trace their efferent projections. In the current study, we used Fos immunocytochemistry to locate neurons that may be involved in maternal behavior. Fos is a nuclear protein that serves as a transcriptional factor that can alter the expression of target genes (Morgan & Curran, 1991; Sheng & Greenberg, 1990). Because there is a close correlation between Fos production and 2-deoxyglucose uptake within neurons, it has been suggested that the immunohistochemical detection of Fos can give one a picture of the neurons that are activated as a result of particular forms of neural stimulation (Morgan & Curran, 1991; Sharp, Gonzalez, Sharp, & Sagar, 1989). However, because Fos production can increase in neurons in the absence of concomitant increases in glucose uptake (Morgan & Curran, 1991), it is probably more accurate to view Fos production as a marker for changes in gene transcription within neurons. Fos is located within the nucleus of neurons and therefore can serve as a marker for individual cells that are functionally modified under particular conditions. In this study we examined whether Fos production increases in the preoptic area of maternally behaving female rats, and we describe the anatomical distribution of such Fos-containing neurons. Because the medial amygdala is also involved in maternal behavior and relays chemosensory input to the preoptic area (Fleming, Vaccarino, & Luebke, 1980; Numan, 1988; Numan, Numan, & English, 1993), the distribution of Fos-containing neurons is examined there as well.
The preoptic area is a brain region that contains neurons involved in various functions. As one example, in addition to being involved in maternal behavior, the preoptic area is involved in neuroendocrine regulation, influencing the release of both luteinizing hormone and prolactin from the anterior pituitary (Silverman, 1988; Wiersma & Kastelijn, 1990). Therefore, in examining the possibility that more neurons in the preoptic area are activated in maternal females (as measured by the number of neurons containing Fos-like immunoreactivity) than in nonmaternal females, the type of preparation used was important. In designing this experiment, we took advantage of the following information. The nulliparous estrous-cycling
MICHAEL NUMAN AND MARILYN J. NUMAN Maternal behavior testing commenced at about 9 a.m. on the first day of testing (Day 0). At this time each female was presented with three test pups, 2-8 days old, with one pup being placed in each quadrant outside the female's nest area or sleeping corner. The female was then observed continuously for 15 min for whether it retrieved any of the pups (denned as carrying a pup from one quadrant to another), whether it retrieved all the pups to one quadrant (grouping), and whether it crouched over any of the pups in a nursing posture. After this initial 15-min observation, females were observed briefly at 30, 45, and 60 min after the onset of the test for the occurrence of maternal behaviors. Females were then checked at hourly intervals until 4 p.m. During these spot checks, the locations of the pups and the female subject were recorded, as well as any behavior the female directed toward the pups. Females were allowed to remain with the pups overnight. At 9 a.m. on Day 1 of testing, the pups from the previous day's test were removed, and a freshly nourished group of test pups was presented to each female. The Day 1 test followed the same procedure as the Day 0 test. (None of the females showed any maternal behavior over the first 2 test days.) Beginning with the Day 2 test, this procedure was modified in that pups were not left with females overnight. That is, at 4 p.m. on Day 2 (after the last behavioral observation) the pups were removed from all females. In all other respects, the tests were similar to those on the previous days. Beginning at 9 a.m. on Day 3, and on each subsequent test day, the procedure followed that for Day 2. The reason we did not leave pups with females overnight after the first 48 hr of pup exposure was that we wanted to accurately detect the time of onset of maternal behavior so that we could then allow the females in the maternal group approximately 2 hr of maternal experience. Although the average time of onset of maternal behavior after exposing nulliparous females to pups is 6-7 days, some females may begin to show maternal behavior by the third test day. Maternal behavior was defined as being either partial or complete. Partial maternal behavior was indicated when a female retrieved all pups to one quadrant (retrieving and grouping), and complete maternal behavior was indicated when a female retrieved, grouped, and crouched over the young in a nursing posture. Females in the maternal behavior group were sacrificed by perfusion 2 hr after the onset of either partial or complete maternal behavior. That is, once a female was observed to retrieve and group, a 2-hr clock was started. During these 2 hr, the female was observed briefly at 0 min and again at 60 min, and then continuously during the last 20 min. During these observations, the occurrence of various components of maternal behavior, particularly crouching, was noted. At the end of the 2-hr period, the maternal female was injected with an overdose of sodium pentobarbitol (Nembutal; 100 mg/kg ip). The rat was placed back in the observation cage, and the pups were removed. Once the female was unconscious, a vaginal smear was taken. She was then transported to another room for perfusion. Approximately 1 hr later another virgin female, one that had not yet shown either partial or complete maternal behavior, was injected with Nembutal, a vaginal smear was taken, and the rat was perfused. Vaginal smears were taken to determine whether there were any differences in the estrous cycle of the maternal females and the nonmaternal females at the time of perfusion. Vaginal smears were rated according to the following scale: 0 = diestrus, 1 = diestrus-proestrus or metestrus-diestrus, 2 = proestrus or metestrus, 3 = proestrus-estrus or estrus-metestrus, and 4 = estrus. The cellular patterns associated with these different stages were derived from descriptions found in Zarrow, Yochim, and McCarthy (1964; also see Numan et al., 1977). Perfusion and tissue sectioning. Deeply anesthetized females were perfused intracardially with 0.9% saline for 5 min followed by 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) for 20 min. The brain was removed from the skull and postfixed in fresh fixative overnight at 4 °C, followed by a soak with a 30% solution of sucrose
female rat is not immediately responsive to test pups, but such females can be induced to show maternal behavior (i.e., retrieving, crouching in a nursing posture over the young, and nest building) if they are exposed to pups over a period of 6-7 days (Numan, 1988). This pup-stimulated or pup-sensitized maternal behavior also occurs in ovariectomized or hypophysectomized females, suggesting that the behavior is free from hormonal control (Rosenblatt, 1967). The sensitized virgin female appeared to be the ideal model for studying the relationship between maternal behavior and Fos expression in the nervous system. Such pup-exposed females would be unaffected by the dramatic endocrine changes associated with pregnancy and lactation. Also, there is variability in the time of onset of maternal behavior after pup exposure. This allowed us to expose all females to pups but to form two groups. A maternal group would comprise females that eventually showed maternal behavior, and a nonmaternal group would comprise females yoked to the maternal females. Nonmaternal females would be exposed to pups for equivalent amounts of time when compared with the maternal counterparts, except that they would not yet have shown the behavior. This nonmaternal group appeared to be an excellent control for extraneous stimulation not related to the performance of maternal behavior. Although lesions of the medial preoptic area prevent the occurrence of maternal behavior in virgin females exposed to pups (Numan et al., 1977), lesions of the medial amygdala facilitate such behavior (Fleming et al., 1980; Numan et al., 1993). Therefore, an analysis of Fos production in the medial amygdala in relation to its production in the medial preoptic area seemed likely to provide important information. Method
Subjects and housing. The subjects were 18 nulliparous estrouscycling female rats of the Charles River CD strain (Charles River Breeding Laboratories, Wilmington, MA). The females were approximately 90 days old at the time of behavioral testing. Females were housed with 1 or 2 other nulliparous females in wire mesh cages (40 x 25 x 18 cm) until 72 hr before the onset of behavioral testing, when they were transferred to clear polycarbonate cages (50 x 40 x 20 cm), which contained wood shavings as bedding material, for observation of maternal behavior. The floors of these cages were divided into four equal quadrants by 5-cm-high Plexiglas partitions that prevented pups from crawling from one quadrant of the cage to another. An additional group of females were used as donor mothers that provided test pups. All females were maintained under a 12-hr reversed day-night cycle (lights on at 7:30 p.m.) and under a relatively constant ambient temperature of 21 °C. Food pellets and water were available ad lib. Behavioral testing. The basic design of this experiment was to expose virgin females to pups daily. Once a female began to show maternal behavior, it was allowed 2 hr to interact with the pups, and it was then sacrificed by perfusion. At this time it was paired with a virgin female that was exposed to pups for an equivalent amount of time but had not yet displayed any maternal behavior. In this way the maternal behavior and nonmaternal behavior groups were formed. The yoked maternal and nonmaternal females were perfused within 1 hr of each other, but the maternal female was always perfused first. Females were perfused 2 hr after the onset of maternal behavior because in pilot work we found that Fos expression in the preoptic area was greatest at this point compared with 1 hr and 3 hr after the onset of the behavior.
PREOPTIC FOS AND MATERNAL BEHAVIOR and phosphate-buffered saline (PBS; pH 7.4) at 4 °C. Within 1 week after the perfusion, 50-u.m-thick frozen sections were cut on a tabletop microtome, and 50 u.m of 100 u,m was saved from the level just rostral to the preoptic area through the caudal hypothalamus. The brain sections were collected into cold PBS. Fos immunocytochemistry. An immunoperoxidase procedure was used to detect Fos-like immunoreactivity. After being rinsed three times in cold PBS, brain sections were incubated for 20 min in 0.3% H2O2 to eliminate endogenous peroxidase activity, followed by another three rinses in cold PBS. Sections were then incubated in 10% normal goat serum for 45 min (in PBS with 0.3% Triton-X) followed by an incubation in 2% normal goat serum for 20 min. Brain sections were then incubated in 0.1 |j,g/ml Fos antibody for 48 hr at 4°C (1:1,000 dilution of rabbit polyclonal antibody AB-2 [Oncogene Science, Uniondale, NY] in 2% normal goat serum and 0.1% crystalline-grade bovine serum albumin in PBS). The complex of Fos and anti-Fos within cells was subsequently visualized through the protocol of Vector's Vectastain Elite ABC peroxidase system (Vector Laboratories, Burlingame, CA) followed by incubation in the chromagen 3, 3'-diaminobenzidine (DAB). The peroxidase reaction product was intensified with nickel chloride. The DAB procedure we used followed the protocol of Vector's DAB Substrate Kit for horseradish peroxidase. Sections were then rinsed in PBS and mounted on chrome-alumcoated slides. After the sections were air-dried, they were counterstained with neutral red and coverslipped. To control for interassay variability, the immunocytochemical detection of Fos in a particular pair of brains from a maternal female and its yoked nonmaternal control was carried out on the same day and under the same assay conditions. Two types of specificity controls were used to show that our immunoperoxidase procedure was detecting a Fos-like peptide. In the first, we ran the reaction as indicated in the previous paragraph except that we eliminated the primary Fos antibody from the protocol. In the second control, brain sections were incubated in primary Fos antiserum that had been preabsorbed for 3 hr with Oncogene Fos peptide (10 jxg peptide was added to 1 u.g antibody). Both control procedures completely eliminated Fos-like immunoreactivity from the brain sections. Microscopic analysis. Microscopic analysis mapped the number and location of neurons containing Fos-like immunoreactivity in the preoptic area, bed nucleus of the stria terminalis (BNST), and corticomedial amygdala. At the level of the preoptic area three rostral-to-caudal frontal sections were chosen for analysis from each female. Care was taken to make sure that the sections were selected from comparable levels for females in the maternal and nonmaternal groups. The most rostral section began at the level of the crossing of the anterior commissure, and the most caudal section ended at the level where the suprachiasmatic nuclei first appear. These three sections were similar to Plates 19-21 in Swanson's (1992) atlas of the rat brain. (We did not quantify Fos labeling in more caudal aspects of the preoptic area [equivalent to Swanson's Plates 22 and 23] because Fos labeling was very low at this point.) For the three coronal sections chosen for detailed analysis, Fos labeling was quantified in each of the following regions for each female: medial and lateral preoptic areas, dorsal BNST (that portion of the BNST above the plane passing through the ventral aspect of the anterior commissure), and ventral BNST. To do this analysis, we first drew diagrams of each of the brain sections with the aid of a microprojector. Subsequently, we examined each brain section microscopically under 100 x magnification and plotted the location of cells containing Fos-like immunoreactivity onto the appropriate reconstructed diagram. Fine focusing allowed us to visualize cells throughout the thickness of each section. An ocular reticule facilitated the counting of Fos-labeled cells. Fos-like immunoreactivity usually presented itself as a dark brown-black reaction product. Cells were counted as positive only when the reaction product was restricted to the nucleus of the cell.
"< tc -d
111 Z 0. DC
Days of Pup Exposure
Figure 1. Cumulative percentage of rats showing maternal behavior on each test day for females in the maternal behavior group. Similar procedures were used for mapping Fos-labeled cells in the corticomedial amygdala. Three rostral-to-caudal frontal sections were analyzed for each female at the level of the anterior medial nucleus of the amygdala (MA; Plates 25-27 in Swanson, 1992), and three additional sections were sampled through the posterior MA (Plates 28-29 in Swanson, 1992). We separated the MA into anterior and posterior parts because the posterior part (particularly its dorsal region) provides the major input to the medial preoptic area (Simerly & Swanson, 1986). For the sections containing the anterior MA, Fos-labeled cells were counted in the following regions for all females: anterior MA, bed nucleus of the accessory olfactory tract (BAOT; this region appeared on only one of the three sections), and the anterior cortical nucleus. For the sections containing posterior MA, Foslabeled cells were counted in the ventral and dorsal regions of the posterior MA. Our analysis of Fos-labeled cells in the BNST, as just described, did not include a major part of the principal nucleus of the BNST, because this subnucleus is largest at caudal levels of the preoptic area. However, we wanted to analyze this area because it receives input from the MA (Krettek & Price, 1978) and, in turn, projects strongly to the medial preoptic area (Simerly & Swanson, 1986). To do this analysis, we counted the number of Fos-labeled cells in three rostral-tocaudal frontal sections through the principal BNST [Plates 21-23 in Swanson, 1992). Finally, we also wanted to quantitatively determine the presence of Fos-containing cells in a brain region usually not associated with maternal behavior. We selected the ventromedial nucleus of the hypothalamus (VMN), which is an area integral to the control of sexual receptivity (Pfaff & Schwartz-Giblin, 1988). All brain sections on which the VMN appeared were microscopically examined at lOOx magnification, and the total number of Fos-positive cells in the VMN was counted. It is important to note that the VMN is capable of expressing Fos-like immunoreactivity and that such expression increases during sexual stimulation (Dudley, Rajendren, & Moss, 1992; Pfaus, Kleopoulos, Mobbs, Gibbs, & Pfaff, 1992).
Results Figure 1 shows the cumulative percentage of females in the maternal behavior group showing maternal behavior on each of the test days. As can be seen, no female displayed maternal behavior over the first 48 hr, and this was followed by a steady
MICHAEL NUMAN AND MARILYN J. NUMAN note that medial preoptic area labeling dropped off dramatically in the most caudal section (this section is caudal to the most posterior section shown in Figure 2, and this level was not used for the quantitative analysis presented in Table 1) and that the principal BNST was only sparsely labeled. Recall that 3 rats in the maternal behavior group showed complete maternal behavior whereas the other 6 only retrieved and grouped their pups. Because of the small number of rats showing complete maternal behavior, we did not statistically analyze the differences in Fos labeling between these females and those that showed partial maternal behavior. However, the females showing complete maternal behavior had more Fos-labeled cells in the medial and lateral preoptic areas and the ventral BNST than did females that only retrieved and grouped, and this difference was particularly obvious for the medial preoptic area and the ventral BNST. Also, the anterior paraventricular nucleus of the hypothalamus was well labeled with cells containing Fos-like immunoreactivity in the virgins that showed complete maternal behavior (see Figure 5). This nucleus remained unlabeled in the virgins showing partial maternal behavior. (All of these observations were made without knowledge of the behavior shown by the females involved.) We computed Pearson product-moment correlation coefficients between number of days of pup exposure and Fos labeling in the medial preoptic area. For both the maternal group (r = .27) and the nonmaternal group (r = .05) the correlation between these variables was not significant. A similar lack of correlation was obtained when labeling in the ventral BNST was examined. For the lateral preoptic area there was not a significant correlation between time with pups and Fos labeling when the maternal group was considered, but for the nonmaternal group this correlation was significant, r(l) = .605, p < .05 (one-tailed test). Table 2 presents the number of cells in the VMN that contained Fos-like immunoreactivity in the maternal and nonmaternal groups. The two groups did not differ in the number of sections saved through the VMN, nor did they differ in the average number of Fos-labeled VMN cells per 50-u.m section (we chose to analyze Fos labeling on all sections that contained the VMN because labeling was sparse in comparison with the preoptic region). Table 3 presents a quantitative analysis of the number of Fos-labeled cells in the corticomedial amygdala and principal BNST of maternal and nonmaternal females, and Figure 6 maps the distribution of such cells in the amygdala. The maternal females had more Fos-labeled cells in the anterior cortical nucleus, t(16) = 1.70, p = .05 (one-tailed test), and posterodorsal MA, t(15) = 1.78,p = .05 (one-tailed test). A Mann-Whitney U test on vaginal smear cytology immediately before perfusion indicated that the two groups did not differ in estrous cycle state. The mean (±SEu) vaginal smear rating for the maternal group was 2.2 ± 0.52, and that for the nonmaternal group was 2.2 ± 0.40 (mdn = 2).
increase in the number of females showing sensitized maternal behavior. The median latency to onset of maternal behavior was 6 days, and the mean was 6.7 days (SEM = 1.15). This finding fits well with previously observed sensitization latencies for this strain of rat (see Numan et al., 1993; Rosenblatt, 1967). Because each female in the nonmaternal group was yoked to a maternal female, the median and the mean number of days of pup exposure for nonmaternal females was the same as for the maternal females. Concerning the degree of maternal responsiveness of the females in the maternal group, 3 of the 9 rats were observed to show complete maternal behavior (retrieving, grouping, and crouching), whereas the remaining 6 females were observed to only retrieve and group. Females in the nonmaternal group briefly sniffed and sometimes licked the test pups each morning after their presentation, but the females tended to ignore the pups for most of the remainder of the day. Therefore, although these females were exposed to pups for equivalent amounts of time as were their maternal counterparts, the nature and degree of stimulation they received from the pups was different. Figure 2 shows the distribution of cells with Fos-like immunoreactivity at the three levels of the preoptic area that we used in our quantitative analysis for a representative maternal and nonmaternal female. Fos-labeled cells were observed in both rats, but the maternal female had more labeled cells, particularly in the more posterior and lateral areas of the medial preoptic area. Figures 3 and 4 are photomicrographs of Fos-labeled cells at the level of the preoptic area in a maternal and nonmaternal female. Table 1 presents the mean number of Fos-labeled cells in the preoptic area and BNST for each of the two groups. The averages were obtained after first summing the number of labeled cells across the three preoptic sections for each female. Statistical analysis through the use of t tests indicated that in comparison with the nonmaternal group, the maternal group had significantly more labeled cells in the medial preoptic area, ?(16) = 2.52, p < .05; lateral preoptic area, t(\6) = 2.35,p < .05; and the ventral BNST, f(16) = 1.99, p < .05, one-tailed test. The two groups did not differ in the number of labeled cells found in the dorsal BNST. The medial preoptic area and the BNST comprise several subnuclei (Ju & Swanson, 1989; Simerly, Gorski, & Swanson, 1986; Simerly & Swanson, 1988). Although Figure 2 gives a quantitative description of Fos-labeled cells over three rostralto-caudal sections, it does not indicate the subnuclear distribution. Figure 5 presents this information for a representative maternal and nonmaternal female. In the rostromedial preoptic area, the medial preoptic nucleus (MPN) and the anterodorsal preoptic nucleus were lightly labeled in both rats. The nonmaternal female had more Fos-labeled cells in the periventricular region at this level. In the more posterior parts of the preoptic area, large differences between the 2 females emerged (see the middle panel in Figure 5). In the maternal female the MPN was intensely labeled in its ventral, medial, and dorsolateral parts (the central part of the MPN was unlabeled), as was the adjoining dorsolateral portion of the medial preoptic area. At this level the ventral BNST was heavily labeled in the maternal female, particularly in the ventral and magnocellular subnuclei of this region (see Ju & Swanson, 1989). Finally,
The first experiment demonstrated an increase in Fos labeling in the preoptic area of maternally behaving sensitized virgin female rats. A question that arises is whether this
PREOPTIC FOS AND MATERNAL BEHAVIOR
Figure 2. The distribution of cells labeled with Fos-like immunoreactivity at the level of the preoptic area in 1 female from the maternal behavior (MB) group and 1 from the nonmaternal behavior (NMB) group. For each female the series of three frontal sections are from anterior to posterior. Each dot represents five labeled cells. The area analyzed in all sections is represented in the lower right section by dashed lines. AC = anterior commissure; cp = caudate-putamen; dst = dorsal bed nucleus of the stria terminalis; F = fornix; gp = globus pallidus; Ip = lateral preoptic area; Is = lateral septum; mp = medial preoptic area; OC = optic chiasm; pc = piriform cortex; si = substantia innominata; vst = ventral bed nucleus of the stria terminalis. labeling is limited to the virgin condition, the initial onset of maternal behavior, or a combination of the two (recall that the maternal rats of Experiment 1 were perfused about 2 hr after the onset of maternal behavior). Because the preoptic area is essential for all conditions and phases of maternal behavior in the rat (Numan, 1988), one might expect to see Fos production, which is a presumed correlate of neural activation, in the preoptic area whenever maternal behavior occurs.
MICHAEL NUMAN AND MARILYN J. NUMAN
Figure 3. Photomicrographs showing Fos-labeled cells in the preoptic area and bed nucleus of the stria terminalis of a virgin female rat that showed maternal behavior. Top: Low magnification; bar = 500 ^m. Bottom: High magnification of the left preoptic area shown in top panel (third ventricle is on the right); bar =150 ^.m. Sections were counterstained with neutral red.
PREOPTIC FOS AND MATERNAL BEHAVIOR
Figure 4. Photomicrographs showing the relative absence of Fos-labeled cells in the preoptic area and bed nucleus of the stria terminalis of a virgin female rat that did not show maternal behavior. Top: Low magnification; bar = 500 (j,m. Bottom: High magnification of the right preoptic area shown in top panel (third ventricle is on the left); bar = 150 u,m. Sections were counterstained with neutral red.
MICHAEL NUMAN AND MARILYN J. NUMAN butter cups). Each female was observed continuously for 15 min and then briefly at 30,60,90, and 120 min after the onset of the test. During these observations, we recorded whether the females in the maternal group retrieved and nursed their young and whether females in the nonmaternal group sniffed, licked, or ate the candy and whether the candy was hoarded (carried from one quadrant of the cage to another). At the end of the 2-hr test, each female was anesthetized with an
Number of Fos-Labeled Cells in Three Frontal Sections for Female Rats Showing Maternal or Nonmaternal Behavior
Preoptic area Behavior Maternal Medial 582.90* 93.86 294.60 65.34 Lateral 118.40* 31.60 36.40 14.66 BNST Ventral 175.60** 57.20 Dorsal 35.80 5.14 32.00 14.09
Note, n = 9 rats per group. BNST = bed nucleus of the stria terminalis. *p < .05, significantly different from the nonmaternal group. **p < .05 (one-tailed test), significantly different from the nonmaternal group. We used a one-tailed t test because the direction of the effect was predicted on the basis of the possible involvement of this region in maternal behavior.
With these questions in mind, we conducted Experiment 2 to examine whether Fos production increases in the preoptic region of maternally behaving postpartum lactating female rats. Method
The conditions of this experiment followed those of the previous one except for the following details. Subjects. The subjects were 14 lactating female rats of the same strain as used in Experiment 1. Eight of the rats were multiparous (they had raised a previous litter), and the remaining 6 females were primiparous. The females gave birth in opaque plastic cages (38 x 30 x 18 cm) that contained wood shavings as bedding material. At between 7 and 11 days postpartum, each female was transferred to the polycarbonate cage described in Experiment 1 for observation of maternal behavior, and her litter was reduced to 8 pups. Only rats that had raised healthy litters (as indicated by pup appearance) were used as subjects in this experiment. Behavioral testing. After being transferred to the observation cages, females were left undisturbed for 24 hr. They were then given a pretest for maternal behavior. For this test, each female's pups were removed from the cage, and 6 of them were replaced in the quadrant of the cage diagonally opposite the female's nest area. Each female was then observed continuously for 15 min, and we noted whether the mother retrieved all pups to the nest area and whether the female displayed any nursing behavior (crouching over the pups to give access ,to nipples). All 14 rats retrieved their pups to the nest area, and 12 of these females adopted nursing postures over their young. At the end of this 15-min observation, all pups were removed from the cages, and the females were left undisturbed for 3 days. Seventy-two hours after the pretest for maternal behavior (on Days 11-15 postpartum), each female was presented with stimulus items in the quadrant of the cage diagonally opposite the nest area, and a 2-hr test commenced. At this time, we split the rats into a maternal behavior group (« = 7) and a nonmaternal behavior group (n = 1). Each female in the maternal group was paired with a female in the nonmaternal group and the pair received the 2-hr test on the same day. Members of each pair were on the same postpartum day when they were tested, and they had the same previous breeding history (multiparous or primiparous). What distinguished the members of each pair was the stimulus items placed in their cages. Females in the maternal behavior group received six healthy pups (7-10 days old), whereas females in the nonmaternal behavior group received six pieces of candy (chocolate-covered peanut
Figure 5. Distribution of cells labeled with Fos-like immunoreactivity in the various subnuclei of the preoptic area and bed nucleus of the stria terminalis for a virgin female rat in the maternal behavior group (MB; left), and for one in the nonmaternal behavior group (NMB; right). Each dot represents five labeled cells. AC = anterior commissure; ADP = anterodorsal preoptic nucleus; AV = anteroventral preoptic nucleus; BST = bed nucleus of the stria terminalis; F = fornix; GP = globus pallidus; MPN = medial preoptic nucleus; OC = optic chiasm; Pr = principal subnucleus of the BST; PS = parastrial nucleus; PV = anteroventral periventricular nucleus; PVH = paraventricular nucleus of the hypothalamus.
PREOPTIC FOS AND MATERNAL BEHAVIOR
Number of Sections Through the Ventromedial Nucleus (VMN) of the Hypothalamus Analyzed for the Presence of Fos-Labeled Cells and Number of Fos-Labeled Cells per 50-tun Section Counted in the VMN Group Maternal behavior (« = 9)
11.20 ± 0.32 12.70 ± 2.62
Analysis No. of VMN sections with Fos cells No. of Fos cells per 50-(j,m VMN section
Nonmaternal behavior (n = 8)
10.60 ± 0.32 12.80 ± 3.57
Note. Values are means plus or minus standard errors of the mean.
overdose of Nembutal, and a vaginal smear was taken. Each female was then perfused as described in Experiment 1. Immunocytochemistry and microscopic analysis. The Fos immunocytochemical reaction was performed as described for Experiment 1. Each member of any maternal-nonmaternal pair was histochemically processed on the same days (to control for day-to-day variability in the processing procedure). As in Experiment 1, our quantitative analysis of Fos labeling focused on the preoptic area, BNST, and corticomedial amygdala and was carried out as described in Experiment 1. Finally, in this experiment we also counted the number of cells in the piriform cortex that contained Fos-like immunoreactivity. The three sections that we chose for analysis of the piriform cortex were the same three sections that we chose for analysis of the preoptic area.
Results All females in the maternal behavior group showed strong maternal behavior on the 2-hr test that was given 72 hr after they were separated from their litters. Six of the 7 rats in this group were observed to both retrieve and nurse their young, while the 7th female was observed only to retrieve. All females in the nonmaternal group were observed to sniff, lick, bite, and eat the candies on their 2-hr test. In addition, 5 of 7 of these females hoarded the candy. Table 3
The two groups did not differ in vaginal cytology. The vaginal smears of 6 of the 7 rats in each of the groups showed they were in diestrus. Figure 7 shows the distribution of Fos-labeled cells from a representative frontal section through the middle preoptic area of a female in the maternal and nonmaternal groups. More cells were labeled in the medial preoptic area and the ventral BNST of the maternal female, but Fos labeling in the piriform cortex was somewhat more intense in the nonmaternal female. The subnuclear distribution of Fos-labeled cells in the medial preoptic area and BNST of the maternal postpartum females was similar to that shown for the maternal female in Figure 5. For the nonmaternal postpartum females, Foslabeled cells were primarily restricted to periventricular regions of the preoptic area and to ventral parts of the medial preoptic area. Table 4 shows the quantitative analysis of Fos-labeled preoptic and BNST cells summed for all females. Also shown are data for the piriform cortex. Statistical analysis indicated that the rats that engaged in maternal behavior had more Fos-labeled cells in the medial preoptic area, t(l2) = 2.24, p < .05, and ventral BNST, t(12) = 1.82, p < .05 (one-tailed test), than did rats in the nonmaternal group. The two groups did not differ in the number of cells containing Fos-like immunoreactivity in the dorsal BNST, lateral preoptic area, and piriform cortex. Figures 8 and 9 are photomicrographs of Fos-labeled cells in the preoptic region of maternal and nonmaternal females. Table 5 shows the number of Fos-labeled cells in the corticomedial amygdala and principal BNST of the maternal and nonmaternal rats. The only significant difference was that the maternal postpartum females had more labeled cells in the posterodorsal MA than did the nonmaternal females, t(12) = 1.12,p = .05 (one-tailed test). The difference in Fos labeling in the BAOT approached significance on a one-tailed test (p = .06). In contrast to Experiment 1, the maternal and nonmaternal females both showed strong Fos labeling in the anterior cortical nucleus, and they did not differ from one another in this regard. Also, for both the maternal and nonmaternal groups, the
Number of Fos-Labeled Cells in Three Frontal Sections for Female Rats Showing Maternal or Nonmaternal Behavior Medial amygdaloid nucleus Behavior Maternal M SEM Nonmaternal M SEM ACo 107.89* 24.88 60.00 13.25 Anterior 84.56 19.33 68.56 17.54 Posterior3 90.33 17.61 52.75 19.74 Posteroventral
Posterodorsal 58.78* 11.41 30.75 10.72
7.00 1.37 7.56 1.99
PBNST 70.89 15.30 57.25 18.16
Note, n = 9 rats per group; only 8 rats were used in the nonmaternal group for the analysis of the posterior, ventral, and dorsal medial amygdaloid nucleus and PBNST because of the loss of brain sections during histological processing. ACo = anterior cortical amygdaloid nucleus; BAOT = bed nucleus of the accessory olfactory tract; PBNST = principal nucleus of the bed nucleus of the stria terminalis. "Posteroventral and posterodorsal areas. *p < .05, significantly different from nonmaternal behavior group (one-tailed / test). We used a one-tailed test because the direction of the effect was predicted on the basis of a sensory activation hypothesis (see the Discussion section).
MICHAEL NUMAN AND MARILYN J. NUMAN
Ilis^ ..< a,
G o* w -2^ 5
™T <D <j 3
IE" -, "
l » l | l l
~ ft 3 S -O —
§ .o c ,, SP E ss g 3 3 a |T .2 | **"* o •- « « « .c a « .B J- .a 13
"2i -2 .5 S •§ « -3 - II -3 S
SPS'R £ | = 5 •> -o E £ u _o o^
« -3 - ll % 2 ^
^i. £ — Ti ^ ^2 c ^^N n rt
cd jr H c
i2 ^ g 3 8
o.v£- I P .2 Z -S
•;; ^ I ^ E 'Is = £ 1 -i •§ H "
S « 0-ijS H «3 >> w o ^r c
! "5 •! s ^ .o .g
MI " e ^- =
« <f -?, rs •o ^
u <u S c
E E "o c C f S « CQ 0 §o 3
c a. ,o 3
1 O <u
*U 3 ^ 3
o 13 .c
i! ca J2 o, a.
O "o c 'C 3 0)
PREOPTIC FOS AND MATERNAL BEHAVIOR
degree of Fos labeling did not differ between females that were primiparous and those that were multiparous. Discussion In this study we used Fos immunocytochemistry as a marker for a change in neural function, and our results support the conclusions of lesion and hormone implant studies, that the preoptic area occupies a central position in the neural control of maternal behavior. In Experiment 1, nulliparous female rats that were induced to show maternal behavior through pup exposure had more Fos-labeled cells in the preoptic area than did nulliparous counterparts that were exposed to pups for equivalent amounts of time but had not displayed any maternal behavior. In Experiment 2, postpartum females that were exposed to pups after a 3-day separation had more Fos-labeled neurons in the preoptic area than did postpartum control females exposed to candy. Our results have been extended by the recent work of Fleming, Rusak, and Suh (1992), who found an increase in Fos expression in neurons from the preoptic area of recently parturient females. The differential Fos labeling in the preoptic areas of maternal and nonmaternal females of Experiments 1 and 2 was not caused by differences in ovarian state, because vaginal smear cytology did not distinguish the maternal from nonmaternal females. Also supporting the view that a particular ovarian state was not related to Fos production is the similar pattern of Fos-labeled cells within the medial preoptic area and the ventral BNST of both maternal virgins and maternal postpartum females. The virgins were estrous cycling and therefore exhibited the full range of vaginal cytology, whereas the postpartum females tended to remain in lactational diestrus. Some component of the observed increase in Fos labeling in the maternal females may be related to prolactin release, because the preoptic area is involved in the neuroendocrine regulation of this anterior pituitary hormone (Wiersma & Kastelijn, 1990) and because pup-related stimuli can influence prolactin release (Numan, 1988). Arguing against this possibility, however, is evidence suggesting that the secretion of prolactin is not increased in recently maternal virgins (see Numan, 1988, for a review). Both experiments showed that the medial preoptic area (particularly the MPN subnucleus) and the ventral BNST had a higher number of Fos-labeled cells in the maternal females. Experiment 1, but not Experiment 2, showed a higher number of Fos-labeled cells in the lateral preoptic area of the maternal group, and the reason for this difference is not clear. It should be noted that excitotoxic amino acid lesions of the lateral preoptic area disrupt the maternal behavior of postpartum females, indicating the importance of this region for established maternal behavior (Numan et al., 1988). We quantitatively analyzed the occurrence of Fos-labeled cells in several control regions. First, the number of dorsal BNST cells containing Fos-like immunoreactivity did not differentiate maternal from nonmaternal females. This result shows that there was not a general activation of all neurons during the occurrence of maternal behavior and that some anatomical specificity existed. Second, in Experiment 1 we showed that there was no difference in Fos labeling in the
Figure 7. The distribution of cells labeled with Fos-like immunoreactivity on a single frontal section through the preoptic area of a postpartum female rat that was exposed to pups and showed maternal behavior (top) and from a postpartum female rat that was exposed to candy and therefore did not show maternal behavior (bottom). Each dot represents five labeled cells. In addition to analyzing the distribution of Fos-labeled cells in the preoptic region delineated in Figure 2, we also analyzed the number of such cells in the piriform cortex, cp = caudate-putamen; dst = dorsal bed nucleus of the stria terminalis; F = fornix; gp = globus pallidus; Ip = lateral preoptic area; mp = medial preoptic area; pc = piriform cortex; si = substantia innominata; vst = ventral bed nucleus of the stria terminalis.
VMN of maternal and nonmaternal females. This is important because the VMN has been related to sexual rather than maternal behavior (Pfaff & Schwartz-Giblin, 1988). Because the VMN is capable of expressing Fos-like immunoreactivity during sexual encounters (Dudley et al., 1992; Pfaus et al., 1992), the results indicate that behavioral, as well as anatomical, specificity exists. Because the VMN and the dorsal BNST showed only low levels of Fos-labeled cells, we felt that it would be important to show that a brain region can contain high levels of Fos-like immunoreactivity in both maternal and nonmaternal females. This was the case for the piriform cortex in Experiment 2. In that experiment females were presented with pups or candy, and all females spent time sniffing at the stimulus items (the piriform cortex is a major target site for main olfactory bulb efferents; Scalia & Winans, 1975). These findings are important because they suggest that the maternal and nonmaternal females of Experiment 2 were equally
MICHAEL NUMAN AND MARILYN J. NUMAN
Table 4 Number of Fos-Labeled Cells in Three Frontal Sections in Postpartum Rats Exposed to Pups (Maternal Behavior Group) or to Candy (Nonmaternal Behavior Group)
Preoptic area Behavior Maternal M SEM Nonmaternal Medial 604.00* 175.14 171.57 79.56 Lateral 32.14 9.84 31.57 19.83 Ventral 189.29** 59.01 69.00 29.45 BNST Dorsal 23.86 9.98 78.00 32.10 Piriform cortex 180.86 52.56 342.43 124.01
Note, n = 7 rats per group. BNST = bed nucleus of the stria terminalis. *p < .05, significantly different from the nonmaternal group ((test). **p < .05, significantly different from the nonmaternal group (one-tailed t test). We used a one-tailed test because the direction of the effect was predicted on the basis of the possible involvement of this region in maternal behavior.
aroused by the stimulus items placed in their cages (also see discussion of anterior cortical nucleus of the amygdala that follows). These results indicate that an increase in Fos expression in the medial preoptic area and the ventral BNST is associated with maternal behavior. How can such information be used to delineate the neural substrate of maternal behavior? Our anatomical analysis focused on the preoptic area and adjoining areas because our major long-term objective is to locate candidates for the output neurons essential for the performance of maternal behavior. In examining Figures 2, 5, and 7, one observes Fos-labeled cells over a rather broad preoptic area and ventral BNST region. Some of these cells may represent the critical output neurons, whereas others may represent neurons responding primarily to pup-related sensory input and still others may be local interneurons. If we could determine more exactly the location of the output neurons, we could then trace their efferent connections with anterograde tracers. It has been suggested that Fos labeling in the preoptic area is primarily a correlate of intense sensory stimulation (Baum & Everitt, 1992). However, other studies have shown that Fos labeling can also serve as a marker for neurons whose activation is associated with motor output and behavioral performance (Castro-Alamancos, Borrell, & Garcia-Segura, 1992; Wan, Liang, Moret, Wiesendanger, & Rouiller, 1992). It is this latter approach that coincides with our objectives. We are aware of the possibility that some of the neurons involved in maternal behavior may not have been detected with our Fos immunocytochemical procedure. First, not all activated neurons show a Fos response (Morgan & Curran, 1991; Wan et ah, 1992). Second, we perfused our animals at only one time point: after 2 hr of maternal responding. Because the maximal Fos response may occur before or after this point for some neurons (Bullitt, Lee, Light, & Willcockson, 1992), we may have missed detecting these cells. Note, however, that in pilot studies we perfused animals 1, 2, or 3 hr after a maternal experience, and the maximal Fos response in the medial preoptic area and ventral BNST occurred at 2 hr. Third, the inhibition of certain neural systems may be critical for the occurrence of maternal behavior, but it is not yet clear whether inhibited neurons express the Fos response (Sharp et al., 1989; also, see the
following discussion on the amygdala). Given these caveats, the fact that the medial preoptic area and ventral BNST of maternal females display a robust Fos response allows for the possibility that a subset of these cells represents the critical efferent neurons. In this context, an important finding of this study is the data that suggest that the ventral BNST, which is a junctional region between the dorsolateral part of the medial preoptic area and the dorsomedial part of the lateral preoptic area, may be important for maternal behavior. Several other findings have pointed to the possible involvement of this region in maternal behavior: (a) Coronal knife cuts through the lateral hypothalamus disrupt maternal behavior and sever axons of the ventral BNST neurons (Numan, Morrell, & Pfaff, 1985), (b) coronal knife cuts posterior to the ventral tegmental area in the midbrain disrupt maternal behavior and sever ventral BNST axons (Numan & Numan, 1991), and (c) ventral BNST neurons bind estrogen and project to the brainstem (Fahrbach, Morrell, & Pfaff, 1986). It would be interesting to explore the effects on maternal behavior of excitotoxic amino acid lesions localized to this specific region (cf. Numan et al., 1990). An approach toward a more accurate determination of the output neurons essential for maternal behavior would be to subtract out any of the Fos-labeled neurons that are activated simply by pup-related sensory input and would be so activated even in the absence of maternal performance. We had initially thought that the nonmaternal nulliparous females of Experiment 1, which were exposed to pups for equivalent amounts of time as were the maternal females, formed the perfect control group in this regard. However, the nonmaternal females tended to ignore the pups, whereas the maternal females cared for them. Clearly, some of the differences in the preoptic area and BNST Fos responses between the two groups might have been solely attributed to differences in sensory input. A dominant form of sensory input in maternal females is primary olfactory and vomeronasal input from pups (Fleming, Vaccarino, Tambosso, & Chee, 1979). To what extent was such input driving the medial preoptic area and ventral BNST Fos response? Would the Fos response of nonmaternal females be increased if they could have in some way been forced to receive increased olfactory input from pups? These questions formed the rationale for our quantitative analysis of Fos labeling in the
PREOPTIC FOS AND MATERNAL BEHAVIOR
Figure 8. Photomicrographs showing Fos-labeled cells in the preoptic area and bed nucleus of the stria terminalis of a postpartum rat in the maternal behavior group. Top: Low magnification; bar = 500 i^m. Bottom: High magnification of the right preoptic area shown in top panel (third ventricle is on the left); bar = 150 (j,m. Sections were counterstained with neutral red.
MICHAEL NUMAN AND MARILYN J. NUMAN
Figure 9. Photomicrographs showing the relative absence of Fos-labeled cells in the preoptic area and bed nucleus of the stria terrainalis of a postpartum rat in the nonmaternal behavior group. Top: Low magnification; bar = 500 p.m. Bottom: High magnification of the right preoptic area shown in top panel (third ventricle is on the left); bar = 150 \im. Sections were counterstained with neutral red.
PREOPTIC FOS AND MATERNAL BEHAVIOR
Table 5 Number of FOS-Labeled Cells in Three Frontal Sections in Postpartum Rats Exposed to Pups (Maternal Behavior Group) or to Candy (Nonmaternal Behavior Group)
Medial amygdaloid nucleus Behavior Maternal M SEM Nonmaternal ACo 93.86 18.52 145.71 54.92 Anterior 72.29 13.28
Posterior3 125.29 30.69
Posteroventral 53.29 15.40 47.00 20.41
Posterodorsal 72.00** 18.39
24.14 6.81 11.71 3.51
PBNST 57.57 18.64 23.00 12.97
Note, n = 1 rats per group. ACo = anterior cortical amygdaloid nucleus; BAOT = bed nucleus of the accessory olfactory tract; PBNST = principal nucleus of the bed nucleus of the stria terminalis. *p < .05, significantly different from the nonmaternal group (one-tailed t test). We used a one-tailed test because the direction of the effect was predicted on the basis of a sensory activation hypothesis (see the Discussion section).
corticomedial amygdala and principal BNST. The MA and BAOT are recipients of input from the accessory olfactory bulb, which relays vomeronasal input, and the anterior cortical nucleus is a recipient of input from the main olfactory bulb (Scalia & Winans, 1975). Because MA projects to the MPN and to what we have called the ventral BNST (Krettek & Price, 1978; Simerly & Swanson, 1986), it is certainly possible that vomeronasal input was, to some degree, driving the Fos response in the MPN and ventral BNST. This sensory activation hypothesis is supported by our finding that Fos labeling in the posterodorsal MA was significantly higher in the maternal virgins and the maternal postpartum females in comparison with their nonmaternal counterparts. Importantly, within MA it is the posterodorsal area that projects directly to the MPN (Simerly & Swanson, 1986). It is worth emphasizing, however, that our results do not inform us as to whether the Fos-labeled cells in this region are the ones that project to the MPN and ventral BNST. Another route by which vomeronasal input may reach the MPN is via the principal BNST (Krettek & Price, 1978; Simerly & Swanson, 1986). Our results do not support the hypothesis that increased activation of the MPN is due to increased input from the principal BNST because Fos labeling in the principal BNST did not differentiate maternal from nonmaternal females. Although the anterior cortical nucleus contained more Foslabeled cells in maternal than in nonmaternal virgins, in the postpartum females this region was not differentially labeled. The maternal virgins had just become maternal and their anterior cortical nuclei were probably being activated by intense primary olfactory stimuli from pups, whereas the nonmaternal virgins were ignoring the pups and were therefore not receiving novel olfactory input. In contrast, the postpartum maternal and nonmaternal females were both interacting with a salient olfactory stimulus (pups or candy). Therefore, the Fos response in the anterior cortical nucleus probably represents activation of the primary olfactory system by a salient olfactory stimulus. This view is supported by our piriform cortex findings discussed previously. Because Experiment 2 did not indicate a difference in Fos labeling in the anterior cortical nucleus, and because that region does not project directly to the ventral BNST or MPN (Krettek & Price, 1978; Simerly & Swanson, 1986), it is unlikely that primary olfactory input to the anterior
cortical nucleus was driving the Fos response in the MPN and ventral BNST in either of our two experiments. Although vomeronasal input to MPN and ventral BNST via the posterodorsal MA may have been driving some component of the Fos response in these regions, an important question is the magnitude of this influence. Because females from which the olfactory bulb has been removed show normal maternal behavior (Numan, 1988), it would be important to determine the distribution of Fos-labeled cells in the medial preoptic area and BNST of such females. An approach such as this may allow us to subtract out from the total distribution of Foslabeled neurons in the preoptic area and ventral BNST of maternal females those neurons that are primarily involved in olfactory processing; this would yield a smaller population of cells that is more closely related to maternal performance. In contrast to the view that the increased number of Foslabeled cells in the posterodorsal MA in maternal females is simply reflecting vomeronasal input, other hypotheses should be considered. First, because the MPN projects to the posterodorsal MA (Simerly & Swanson, 1988), it is possible that MPN input to the region was causing the Fos response in the latter structure, rather than the reverse (cf. Baum & Everitt, 1992). Second, because MA lesions facilitate maternal behavior (Fleming et al., 1980; Numan et al., 1993), it is possible that an important functional change must occur in the posterodorsal MA in order for maternal behavior to express itself. More specifically, the Fos response in the posterodorsal MA may be more related to maternal performance than to vomeronasal activation.
Baum, M. J., & Everitt, B. J. (1992). Increased expression of c-fos in the medial preoptic area after mating in male rats: Role of afferent inputs from the medial amygdala and midbrain central tegmental field. Neuroscience, 50, 627-646. Bridges, R. S., Numan, M., Ronsheim, P. M., Mann, P. E., & Lupini, C. E. (1990). Central prolactin infusions stimulate maternal behavior in steroid-treated, nulliparous female rats. Proceedings of the National Academy of Sciences (USA), 87, 8003-8007. Bullitt, E., Lee, C. L., Light, A. R., & Willcockson, H. (1992). The effect of stimulus duration on noxious-stimulus induced c-fos expression in the rodent spinal cord. Brain Research, 580, 172-179.
MICHAEL NUMAN AND MARILYN J. NUMAN preoptic area and onset of maternal behavior in the rat. Journal of Comparative and Physiological Psychology, 91, 146-164. Numan, M., & Smith, H. G. (1984). Maternal behavior in rats: Evidence for the involvement of preoptic projections to the ventral tegmental area. Behavioral Neuroscience, 98, 712-727. PfafT, D. W., & Schwartz-Giblin, S. (1988). Cellular mechanisms of female reproductive behaviors. In E. Knobil & J. D. Neill (Eds.), The physiology of reproduction (pp. 1487-1568). New York: Raven Press. Pfaus, J. G., Kleopoulos, S. P., Mobbs, C. V., Gibbs, R. B., & Pfaff, D. W. (1992). Fos and Jun expression in the female rat forebrain following hormone treatment and sexual stimulation. Society for Neuroscience Abstracts, 18, 892. Rosenblatt, J. S. (1967). Nonhormonal basis of maternal behavior in the rat. Science, 156, 1512-1514. Scalia, F., & Winans, S. S. (1975). The differential projections of the olfactory bulb and accessory olfactory bulb in mammals. Journal of Comparative Neurology, 161, 31-55. Sharp, F. R., Gonzalez, M. F., Sharp, J. W., & Sagar, S. M. (1989). C-fos expression and (14C)2-deoxyglucose uptake in the caudal cerebellum of the rat during motor/sensory cortex stimulation. Journal of Comparative Neurology, 284, 621-636. Sheng, M., & Greenberg, M. E. (1990). The regulation and function of c-fos and other immediate early genes in the nervous system. Neuron, 4, 477-485. Silverman, A. (1988). The gonadotropin-releasing hormone (GnRH) neural systems: Immunocytochemistry. In E. Knobil & J. D. Neill (Eds.), The physiology of reproduction (pp. 1283-1304). New York: Raven Press. Simerly, R. B., Gorski, R. A., & Swanson, L. W. (1986). Neurotransmitter specificity of cells and fibers in the medial preoptic nucleus: An immunohistochemical study in the rat. Journal of Comparative Neurology, 246, 343-363. Simerly, R. B., & Swanson, L. W. (1986). The organization of neural inputs to the medial preoptic nucleus of the rat. Journal of Comparative Neurology, 246, 312-342. Simerly, R. B., & Swanson, L. W. (1988). Projections of the medial preoptic nucleus: A Phaseolus vulgaris leucoagglutinin anterograde tract-tracing study in the rat. Journal of Comparative Neurology, 270, 209-242. Swanson, L. W. (1976). An autoradiographic study of the efferent connections of the preoptic region in the rat. Journal of Comparative Neurology, 167, 227-256. Swanson, L. W. (1992). Brain maps: Structure of the rat brain. Amsterdam: Elsevier. Terkel, J., Bridges, R. S., & Sawyer, C. H. (1979). Effects of transecting the lateral neural connections of the medial preoptic area on maternal behavior in the rat: Nest building, pup retrieval, and prolactin secretion. Brain Research, 169, 369-380. Wan, X. S. T., Liang, F., Moret, V., Wiesendanger, M., & Rouiller, E. M. (1992). Mapping of motor pathways in rats: c-Fos induction by intracortical microstimulation of the motor cortex correlated with efferent connectivity of the site of cortical stimulation. Neuroscience, 49, 749-761. Wiersma, J., & Kastelijn, J. (1990). Electrophysiological evidence for a key control function of the medial preoptic area in the regulation of prolactin secretion in cycling, pregnant, and lactating rats. Neuroendocrinology, 51, 162-167. Zarrow, M. X., Yochim, J. M., & McCarthy, J. L. (1964). Experimental endocrinology. San Diego, CA: Academic Press.
Castro-Alamancos, M. A., Borrell, J., & Garcia-Segura, L. M. (1992). Performance in an escape task induces Fos-like immunoreactivity in a specific area of the motor cortex of the rat. Neuroscience, 49, 157-162. Conrad, L. C., & Pfaff, D. W. (1976). Efferents from the medial basal forebrain and hypothalamus in the rat: I. An autoradiographic study of the medial preoptic area. Journal of Comparative Neurology, 169, 185-220. Dudley, C. A., Ranjendren, G., & Moss, R. L. (1992). Induction of FOS immunoreactivity in central olfactory structures of the female rat following exposure to conspecific males. Molecular and Cellular Neurosciences, 3, 360-369. Fahrbach, S. E., Morrell, J. I., & Pfaff, D. W. (1986). Identification of medial preoptic neurons that concentrate estradiol and project to the midbrain in the rat. Journal of Comparative Neurology, 247, 364-382. Fleming, A. S., Rusak, B., & Suh, E. J. (1992). Fos-like immunoreactivity in brain of primiparous rats after mother-litter interactions. Society for Neuroscience Abstracts, 18, 875. Fleming, A. S., Vaccarino, F., & Luebke, C. (1980). Amygdaloid inhibition of maternal behavior in the nulliparous female rat. Physiology & Behavior, 25, 731-743. Fleming, A. S., Vaccarino, F., Tambosso, L., & Chee, P. (1979). Vomeronasal and olfactory system modulation of maternal behavior in the rat. Science, 203, 372-374. Ju, G., & Swanson, L. W. (1989). Studies on the cellular architecture of the bed nuclei of the stria terminalis in the rat: I. Cytoarchitecture. Journal of Comparative Neurology, 280, 587-602. Krettek, J. E., & Price, J. L. (1978). Amygdaloid projections to subcortical structures within the basal forebrain and brainstem in the rat and cat. Journal of Comparative Neurology, 178, 225-254. Mann, P. E., & Bridges, R. S. (1992). Neural and endocrine sensitivities to opioids decline as a function of multiparity in the rat. Brain Research, 580, 241-248. Morgan, J. I., & Curran, T. (1991). Stimulus-transcription coupling in the nervous system: Involvement of the inducible proto-oncogenes fos and jun. Annual Review of Neuroscience, 14, 421-451. Numan, M. (1974). Medial preoptic area and maternal behavior in the female rat. Journal of Comparative and Physiological Psychology, 87, 746-759. Numan, M. (1988). Maternal behavior. In E. Knobil & J. D. Neill (Eds.), The physiology of reproduction (pp. 1569-1645). New York: Raven Press. Numan, M., & Callahan, E. C. (1980). The connections of the medial preoptic region and maternal behavior in the rat. Physiology & Behavior, 25, 653-665. Numan, M., Corodimas, K. P., Numan, M. J., Factor, E. M., & Piers, W. P. (1988). Axon-sparing lesions of the preoptic region and substantia innominata disrupt maternal behavior in rats. Behavioral Neuroscience, 102, 381-396. Numan, M., McSparren, J., & Numan, M. J. (1990). Dorsolateral connections of the medial preoptic area and maternal behavior in rats. Behavioral Neuroscience, 104, 964-979. Numan, M., Morrell, J. I., & PfafT, D. W. (1985). Anatomical identification of neurons in selected brain regions associated with maternal behavior deficits induced by knife cuts of the lateral hypothalamus. Journal of Comparative Neurology, 237, 552-564. Numan, M., & Numan, M. J. (1991). Preoptic-brainstem connections and maternal behavior in rats. Behavioral Neuroscience, 105, 10131029. Numan, M., Numan, M. J., & English, J. B. (1993). Excitotoxic amino acid injections into the medial amygdala facilitate maternal behavior in virgin female rats. Hormones and Behavior, 27, 56-81. Numan, M., Rosenblatt, J. S., & Komisaruk, B. R. (1977). Medial
Received June 15, 1993 Revision received September 27, 1993 Accepted September 27, 1993 •