Hypothalamic Grafts Induce the Recovery of Lordosis in Female Hamsters with Lesions of the Ventromedial Hypothalamus

Hormones and Behavior 32, 192–200 (1997) Article No. HB971421

Hypothalamic Grafts Induce the Recovery of Lordosis in Female Hamsters with Lesions of the Ventromedial Hypothalamus Marnie B. Bartholomew1 and Owen R. Floody2 Department of Psychology, Bucknell University, Lewisburg, Pennsylvania 17837 Received May 19, 1997; revised June 19, 1997; accepted September 29, 1997

Previous studies have documented the ability of neural grafts to stimulate the recovery of lordosis from neurochemical deficits. However, it was unclear if grafts also could reverse deficits in lordosis caused by lesions at critical points in the neural circuit controlling this response. To address this question, female hamsters were subjected to unilateral lesions of the ventromedial hypothalamus (VMN), a structure well known for its mediation of hormonal effects on lordosis. The effects of these lesions were described by noting the ability of manual stimulation of one flank to reinstate a deteriorating lordosis response. Consistent with past results, unilateral VMN lesions decreased responsiveness to stimulation of just the contralateral flank. Females showing such lateralized decrements then received control treatments or implants into the lesioned area of basal hypothalamic tissue from a neonatal male or female hamster. Approximately 1 month later, tests of lordosis reinstatement by ipsi- or contralateral manual stimuli were repeated. Whereas lateralized decrements in responsiveness persisted in control subjects, implants of tissue from male or female neonates led to reliable improvements in lordosis, reversing the lesion-induced decrease in contralateral responsiveness. The mechanism responsible for this change is unclear, but could involve an elevation in a lordosis-facilitating neuromodulator. Alternatively, it could depend on the reinforcement or replacement of neural circuits for lordosis, possibly including those that connect the two VMNs with each other or with the periaqueductal gray of the midbrain. q 1997 Academic Press

Several reports have documented increases in mating behavior after the introduction into the ventricles or hypothalamus of neural grafts from a fetal or neonatal donor. In most of these studies, baseline levels of behav1 Current address: Duke University Medical Center, Box 2716, Durham, NC 27710. 2 To whom correspondence and reprint requests should be addressed. Fax: (717) 524-3760. E-mail: [email protected].

ior were depressed by factors other than brain lesions, i.e., by advanced age (Huang, Kissane, and Hawrylewicz, 1987), treatment with suboptimal levels of gonadal hormones (Arendash and Gorski, 1982), observation of behaviors typical of the opposite sex (Arendash and Gorski, 1982), or a genetic reduction in levels of gonadotropin-releasing hormone (GnRH; Gibson, Charlton, Perlow, Zimmerman, Davies, and Krieger, 1984; Gibson, Krieger, Charlton, Zimmerman, Silverman, and Perlow, 1984; Gibson, Moscovitz, Kokoris, and Silverman, 1987; Krieger, Perlow, Gibson, Davies, Zimmerman, Ferin, and Charlton, 1982). The behavioral effects of these grafts have been taken to suggest their ability to elevate levels of a hormone or hormone receptor required to activate an intact but quiescent behavioral circuit. Such an interpretation is best supported in the studies of GnRH-deficient mice, in which animals with grafts showed increased gonadotropin levels and effects, suggesting activity in the GnRH-containing fibers that grafts projected into the host’s median eminence (Gibson et al., 1984a,b, 1987; Krieger et al., 1982). These studies strongly suggest the ability of neural grafts to normalize mating behavior by modulating activity in existing brain circuits. However, they do not test the ability of grafts to reverse deficits in sexual behavior by repairing or replacing damaged circuits. This issue has been addressed most successfully in a recent series of studies of male-typical behavior in rats (Paredes, Pin˜a, Ferna´ndez-Ruiz, and Bermu´dez-Rattoni, 1990; Paredes, Pin˜a, and Bermu´dez-Rattoni, 1993). Here, subjects received bilateral lesions of the medial preoptic area (mPOA). The resulting deficits in mating behavior were largely reversed by grafts of fetal hypothalamus. This recovery was attributed to the establishment of functional connections between graft and host, replacing some of the connections destroyed by the mPOA lesions. This interpretation was supported by anatomical evidence of projections from the graft to 0018-506X/97 $25.00 Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.

192

AID

H&B 1421

/

6810$$$$81

12-18-97 06:46:48

habal

AP: H & B

193

Graft-Induced Recovery of Lordosis

the dorsolateral tegmentum, a normal target of mPOA efferents involved in the control of male-typical behavior. Whereas impressive progress has been made in studies using neural grafts to counteract lesion-induced deficits in male-typical reproductive behavior, less progress has been made in comparable studies of lordosis or other female-typical acts. In the most relevant such study, female rats received bilateral lesions of the ventromedial hypothalamus (VMN), some of which were followed by implants of fetal hypothalamus (Jenssen, Gruner, Do¨cke, and Do¨rner, 1988). Animals with surviving grafts tended to show higher lordosis quotients than controls. However, the reported statistical analyses seem inadequate to confirm the reliability of these changes. Another potentially relevant study used chemical lesions to deprive the VMN of serotonergic inputs that normally inhibit lordosis (Luine, Renner, Frankfurt, and Azmitia, 1984). Subsequent grafts of developing raphe tissue accelerated behavioral recovery, apparently by normalizing VMN levels of serotonin. In this case, it seems reasonable to view the lesion and graft as acting to modulate activity in otherwise intact VMN circuits for lordosis (Kimble, 1990). Therefore, like the early studies of suboptimally primed (Arendash and Gorski, 1982) or GnRH-deficient females (Gibson et al., 1984a,b, 1987), this report seems insufficient to determine whether neural grafts can reverse deficits in lordosis such as are prompted by more drastic VMN damage. In beginning to address this issue, we have drawn on early studies of graft effects on behavior (e.g., Bjo¨rklund and Stenevi, 1979; Perlow, Freed, Hoffer, Seigler, Olson, and Wyatt, 1979), which capitalized on unilateral lesions, grafts, and responses to quickly exclude many potential sources of behavioral change through the use of one side of an animal as a control for the other. This strategy obviously requires a response system sensitive to unilateral damage. An appropriate system is provided by the lordosis response of the female hamster, which responds to unilateral damage or stimulation with lateralized changes that are readily quantified by tests using unilateral manual stimulation (Floody, 1989; Floody and Cramer, 1986; Muntz, Rose, and Shults, 1980; Ostrowski, Scouten, and Malsbury, 1981). Accordingly, we have subjected adult female hamsters to unilateral VMN lesions, testing the extent to which the resulting lateralized decrements in lordosis can be reversed by grafts of hypothalamic tissue from hamster neonates. At the same time, we tested the specificity of graft effects by comparing the impacts of grafts from young males and females. Whereas implants from males and females are similar in their abilities to in-

AID

H&B 1421

/

6810$$$$81

12-18-97 06:46:48

crease GnRH levels sufficiently for the improved modulation of endocrine and neural activity in hypogonadal mice (Gibson et al., 1984a,b, 1987; Krieger et al., 1982), it currently is unclear whether they are equivalent in their potentials to reduce lesion-induced dysfunctions in the neural mechanisms controlling female-typical reproductive behavior.

METHODS Animals Complete data were obtained from 12 adult Lak:LVG female hamsters (Mesocricetus auratus) that averaged 132.5 g in weight (SEM Å 6.2) near the start of testing. Like the sexually experienced males that served as stimuli, each subject was housed individually in a 34 1 18 1 18 cm wire-mesh cage in a colony kept at 20 – 257C and on a reversed 14:10 h light:dark cycle. All animals had free access to food and water and received lettuce at least once each week.

Procedures Subjects were observed during natural estrus, in tests of the ability to reverse the eventual deterioration of male-elicited lordosis by light manual stimulation of one flank (Muntz et al., 1980). Such tests permit the assessment of responses to unilateral treatments (Floody, 1989; Muntz et al., 1980) and seem more sensitive to the effects of lesions than measures of lordosis durations to males (Floody, 1993). The responses to manual stimulation were categorized on the basis of their intensity. Full lordosis was defined by immobility, together with an elevation of the tail sufficient to place its tip even with or above the level of the back at a point between the hips. Partial lordosis was distinguished by a tail that was at least horizontal, but with its tip below the criterion for the full response. After 1 min of adaptation to an elevated 41 1 21 cm platform, a female was paired with a stimulus male for the minimal period required for the elicitation of full lordosis (generally 10–30 s, maximum Å 60 s). The male then was carefully removed, and lordosis was maintained for 10 s by light stroking of the fur along the dorsal lumbar midline with a 1-cm-wide camel’s hair brush. At this time, stimulation was withheld until the start of the female’s emergence from lordosis was signaled by an abrupt limb movement, usually accompanied by an equally abrupt lowering of the tail. As quickly as possible, manual stimulation was reinitiated. However, this was confined to one flank (the right or left for

habal

AP: H & B

194

Bartholomew and Floody

equal subgroups) as the test female was closely examined for the occurrence of partial and full lordosis responses. After continuing for a maximum of 10 s, stimulation was terminated until the female had been out of lordosis for at least 30 s. At this time, the male was reintroduced to begin a new trial. Testing continued in this way, with stimulation alternating between the flanks until each had been targeted in a total of five trials. The measures extracted from these tests were the frequencies of the full and partial lordosis responses observed to stimulation of each flank (the proportion of the five trials/side that was positive for the response in question). Upon the completion of these tests, each female was anesthetized (85 mg/kg of sodium pentobarbital, ip) and subjected to unilateral lesions of the mediobasal hypothalamus, including the VMN. One goal at this point was to maximize the percentage of subjects showing the large decrements in lordosis (operationally defined as a drop of at least 40% in the frequency of full lordosis responses to stimulation of the flank contralateral to the lesion) required of potential graft recipients. For this reason, each female received two lesions separated by 0.5 mm along the dorsal – ventral axis. In addition, each lesion was relatively large, reflecting the passage, for 25 s, of 8 mA of radiofrequency current (Grass LM4 lesion maker) between the 0.5-mm bared tip of a Teflon-insulated stainless-steel electrode (0.25 mm diameter, Medwire) and a rectal probe. Equal subgroups were selected at random to receive lesions aimed at the right or left VMN. After an average of 8.4 days of recovery (SEM Å 1.0), subjects were retested for lordosis, using procedures identical to those used to establish baseline levels of receptivity. Testing here and in the subsequent phase of the study was conducted without knowledge of a subject’s group or lesioned side. However, all behavioral results eventually were referred to lesion placement, so as to describe levels of responsiveness to manual stimuli applied ipsi- and contralateral to hypothalamic lesions. Animals showing substantial drops in lordosis due to lesions were divided equally among three groups (each N Å 4), balanced with respect to lesion side and earlier performance. The latter was assessed using frequencies of full lordosis, since past work has found this measure to be more sensitive to lesions than the frequency of partial lordosis (Floody, 1993). For two groups, each member received a unilateral implant of an approximately 1-mm3 block of tissue from the mediobasal hypothalamus of a neonatal male (Male group) or female (Female group) hamster. The required surgery was scheduled an average of 3.2 days (SEM Å 1.2) after the completion of postlesion testing and 11.6 days (SEM Å

AID

H&B 1421

/

6810$$$$81

12-18-97 06:46:48

1.1) after the production of lesions. Donors were sexed and decapitated within 12 h of birth. Each brain was rapidly dissected in cold saline under sterile conditions. The distinctive landmarks available on the brain’s ventral surface were used to isolate a block of tissue (between the interpeduncular fossa and the optic chism, within 1 mm of the midline, extending no more than 1 mm into the brain) that should have included the developing VMN. The dissected tissue was gently packed into a sterile length of 22-gauge hypodermic tubing This was lowered into the brain of an anesthetized recipient at the same anterior–posterior and medial–lateral coordinates used in the earlier lesion surgery. A closed length of sterile 26-gauge tubing was used to eject the implant, aiming for a point near the middle of the lesion cavity and dorsal to the level of the now-damaged VMN. No more than 15 min elapsed between the time at which the donor was sacrificed and the completion of the graft’s implantation. Members of the third (Sham) group were subjected to comparable surgery, but using an empty 22-gauge cannula. Following surgery, subjects were allowed to recover for an average of 31.7 days (SEM Å 2.5). Each then experienced two behavioral tests on successive estrus days. Preliminary analysis revealed a reliable decrease in the frequency of full lordosis over these tests. This decline did not interact with group membership and probably represents just a response to repeated, closely spaced tests (Floody, 1993). For these reasons, and because of the use of isolated tests to assess preand postlesion performance, we used the data from each animal’s first postimplant test to describe her behavior at this phase of the study.

Analysis At the end of testing, subjects were deeply anesthetized with sodium pentobarbital and perfused transcardially with 10% phosphate-buffered formalin. Subsequently, the brains were removed, frozen, and sectioned at 50 mm in the frontal plane. Sections stained with formol thionin were examined microscopically to locate and describe the lesions and grafts. A microprojector then was used to trace the critical parts of sections on reference drawings. These tracings later were digitized in preparation for quantitative comparisons (NIH Image) of the relative volumes of lesions and grafts. The behavioral analyses were complicated by the fact that groups sometimes achieved perfect performance (lordosis frequencies of 1.0 by all members), especially during prelesion testing and in response to ipsilateral stimuli. Because of this and the limits on the experimental designs that can be accommodated by individual nonparametric tests, it was necessary to use a combination

habal

AP: H & B

195

Graft-Induced Recovery of Lordosis

of nonparametric and parametric analyses of variance (ANOVAs). Specifically, the comparisons of pre- and postlesion scores required for the assessment of lesion effects relied on Friedman’s two-way ANOVAs (Siegel, 1956). These tests were one-tailed, since there are numerous studies suggesting that VMN lesions disrupt lordosis (e.g., Ostrowski et al., 1981). In contrast, the comparisons of postlesion and postimplant scores that were required to assess responses to implants relied largely on parametric ANOVAs, the results of which were clarified by Tukey (a) tests (Winer, 1971). These analyses were two-tailed, since (a) implant effects on lordosis have not been extensively studied and (b) implants can disrupt, rather than facilitate, recovery from lesions (e.g., Woodruff, Baisden, and Nonneman, 1992).

RESULTS Lesions Each subject in this study received a VMN lesion between its first and its second behavioral test. To describe the resulting damage, all animals were ranked on the basis of an estimate of lesion volume derived from computerized analyses of tracings of the areas of damage evident on each brain section. Sketches of lesions approximating the median, first, and third quartiles of this distribution are reproduced in Fig. 1. These suggest that each animal suffered extensive damage to the targeted VMN. Analyses of the volumes of VMN remaining on the lesioned and unlesioned sides (see below) suggest that an average of 86% (SEM Å 5%) of the targeted VMN was destroyed. At the same time, damage typically extended beyond the VMN, especially into more dorsal areas including the dorsomedial hypothalamus. This clearly reflects the vertically stacked pair of lesions received by each subject, possibly in combination with incidental damage caused by the cannula used to deliver sham or real implants. Estimates of lesion volume also were used to test the existence of group differences in the extent of damage that could underlie differential responses to lesions or implants, a danger that possibly is accentuated in designs with small groups. To guard against the possibility that distortions in the tissue available for histological analysis (e.g., reflecting large lesions and a long period of survival postlesion) would compromise analyses of lesion size, the extent of damage was assessed in three ways. First, as indicated above, the full extent of the damage in each animal was traced from each relevant section. Once digitized and measured, the resulting areas were combined with section thickness to estimate

AID

H&B 1421

/

6810$$$$81

12-18-97 06:46:48

lesion volume. Second, an estimate focusing on damage within the VMN began with tracings of the VMN tissue on the lesioned and control sides of each section. From these, it was possible to estimate the volumes of the intact VMN and of the VMN tissue spared by each lesion. The difference between these approximates the extent of VMN damage. Finally, in the analysis that should be most resistant to the effects of tissue distortion, we defined a block of tissue roughly corresponding to the ventral third of the brain stem through the anterior – posterior extent of the VMN. The volume of this block was determined for the lesioned and unlesioned hemispheres of each subject. Any difference between these must reflect tissue loss due to lesion or implant surgery. Not surprisingly, the estimates of damage provided by these methods varied, across mean lesion volumes of 0.19 – 0.92 mm3 (SEMs Å 0.02 – 0.13). Nevertheless, the application of ANOVA to these data failed to reveal any group difference in extent of damage (across these three analyses, F(2, 9) ° 2.78, P ¢ 0.11). This suggests that any group differences in behavior do not reflect differences in the amount or pattern (e.g., extent within the VMN) of lesion damage.

Implants Methods similar to those applied to lesions were used to describe the grafts received by members of the Male and Female groups. The results suggest that implants persisted in all of these animals, but that most were small (mean volume Å 0.022 mm3, SEM Å 0.007) and concentrated dorsally in the lesion cavity, above the level of the damaged VMN (Figs. 1 and 2). Even considering the imprecision of our estimates of the volumes of the blocks of tissue prepared for implantation, it is clear that only a small fraction of each implant typically survived. However, this level of survival is not uncommon (e.g., Gibson et al., 1984b) and does not exclude the possibility of disproportionately large behavioral effects (Gibson et al., 1984b, 1987; Kimble, 1990; Paredes et al., 1993). As in the case of the lesions, we were concerned that behavioral differences between these small groups could result from differences in graft volume. For this reason, and to test for differences in the survival of grafts from male and female donors, ANOVA was used to compare graft volumes in the two groups with implants. This analysis failed to reveal any suggestion of a reliable difference in volume (F(1, 6) Å 1.07, P Å 0.34). This suggests that donor sex did not affect graft survival.

habal

AP: H & B

196

Bartholomew and Floody

FIG. 1. (A – E) Tracings of sections at intervals of roughly 250 mm through the VMN of a control subject. These sketches describe the lesions (irregular areas enclosed by solid or dashed lines) and implants (crosshatched areas) of the three subjects with implants that best represent all animals in terms of lesion volume, implant volume, and extent of VMN damage. To mimimize the extent of overlap of lesions or implants, the one lesion depicted by dashed lines is represented to the left of the midline in A – C, but to the right of the midline in D – E. aha, anterior hypothalamic area; cc, corpus callosum; dmn, dorsomedial nucleus of the hypothalamus; f, fornix; fi, fimbria; mt, mammillothalamic tract; pvn, paraventricular nucleus; ot, optic tract; sm, stria medullaris; v, ventricle; vmn, ventromedial nucleus of the hypothalamus. The asterisk in D indicates the source of the photomicrograph reproduced in Fig. 2.

Lordosis The behavioral analyses were intended to answer two questions: (1) Did VMN lesions affect lordosis? and (2) Were any lesion effects reversed by implants of hypothalamic tissue from male or female neonates? The first of these was addressed by a series of Friedman’s ANOVAs, each comparing pre- and postlesion frequencies of full or partial lordosis for a specific group and side of stimulation. For all groups and both forms of lordosis, responses to ipsilateral stimuli were similar across tests, suggesting the resistance of these measures to changes incidental to the surgery required for the placement of lesions and implants. Conversely, each

AID

H&B 1421

/

6810$$$$81

12-18-97 06:46:48

group showed a reliable postlesion decrease in the frequency of full lordosis responses to contralateral stimulation (each P Å 0.0227, Fig. 3). In the group scheduled to receive implants from female neonates, this was accompanied by a similar decrease in the frequency of partial lordosis responses to contralateral stimuli (P Å 0.0227, Table 1). Responses to implants were assessed using several parametric ANOVAs on relevant subsets of the data, a strategy required by the fact that even the postlesion and postimplant data included some cells with perfect scores and no variability. The first of these compared the three groups in terms of full lordosis responses to contralateral stimulation during postlesion and postim-

habal

AP: H & B

197

Graft-Induced Recovery of Lordosis

appeared in the other groups immediately postlesion apparently were reversed by the grafts that distinguish these groups from the controls. Finally, postlesion and postimplant partial lordosis scores were subjected to corresponding analyses using nonparametric ANOVAs. These failed to reveal significant increases in lordosis frequency postimplant, a finding which is consistent with the relatively small lesion effect on this form of lordosis.

DISCUSSION

FIG. 2. Photomicrograph (original magnification, 1200) of the border (dark line running diagonally, from top right to bottom left) between a lesion cavity containing an implant (to the right of the border) and the adjacent unlesioned tissue (on left). The subject and region depicted are indicated by an asterisk in Fig. 1D.

plant tests. This revealed a reliable Group 1 Phase interaction (F(2, 9) Å 6.53, P Å 0.0177). Further analysis confirmed that animals with implants from males or females showed reliable increases in responsiveness to contralateral stimuli (Ps õ 0.01, Tukey tests, Fig. 3). Controls showed no such change. A second ANOVA compared full lordosis frequencies for both sides and all groups, but just during postimplant tests. This revealed a highly reliable interaction of Group and Side of stimulation (F(2, 9) Å 17.76, P Å 0.0008). Further analysis confirmed that this reflects the fact that Sham animals continued the pattern established postlesion of showing lordosis in response to ipsilateral, but not contralateral, manual stimulation (P õ 0.01, Tukey test, Fig. 3). The similar effects that had

AID

H&B 1421

/

6810$$$$81

12-18-97 06:46:48

Unilateral lesions of the VMN produced very reliable decrements in the incidence of high-quality (full) lordosis responses to stimulation of the contralateral flank. This effect is consistent with those of direct applications of estrogen (Floody, 1989) and cuts targeting VMN efferents (Ostrowski et al., 1981). These effects presumably reflect the dependence of lordosis on the VMN, together with the crossing of most VMN projections to the periaqueductal gray (PAG), the area thought to integrate VMN and other influences on lordosis (e.g., Pfaff, 1980). Despite its size and consistency, this effect did not erase the capacity for lordosis, even to contralateral stimuli. This is suggested by the fact that most (9 of 12) animals showed some full lordosis responses postlesion. In addition, it is supported by the resistance to lesion-induced change shown by lower quality (partial) lordosis. Conceivably, these responses to contralateral stimulation reflect the difficulty of confining manual stimulation to one side of the midline. However, even bilateral lesions of the PAG leave hamsters with the capacity for some lordosis responses (Floody and O’Donohue, 1980). This suggests that the partial sparing of lordosis responses to contralateral stimuli reflects a partial sparing of central mechanisms for this response. Possibly relevant connections include the contralateral projections of spared parts of the lesioned VMN, along with the ipsilateral projections of the unlesioned VMN. The partial sparing of mechanisms for lordosis could play a permissive role in any recovery of this response. However, it was not sufficient for recovery, since control animals showed no reliable change in responsiveness to contralateral stimulation over a postlesion period exceeding 5 weeks. This is very different from the pattern observed in animals with implants of basal hypothalamic tissue from neonatal hamsters. The increase in contralateral responsiveness shown by these animals suggests a powerful facilitation of recovery by these implants.

habal

AP: H & B

198

Bartholomew and Floody

FIG. 3. Mean (and SEM) frequencies of full lordosis in tests before surgery (Prelesion), shortly after unilateral VMN lesions (Postlesion), and 1 month after control surgery (Sham group) or the implantation of tissue from a neonatal female (Female group) or male (Male group). Each group included four subjects. Lordosis frequencies were determined separately for the flanks Ipsi- and Contralateral to the lesion. No group difference appeared Prelesion (when responsiveness was high to both stimuli) or Postlesion (when each group showed a significant decrease in responsiveness to Contralateral stimuli). Accordingly, the figure was simplified by collapsing the groups (summarizing the data for All animals) in describing Pre- and Postlesion behavior. The group differences that emerged Postimplant are coded by asterisks: * subjects in the Female and Male groups showed Postimplant frequencies to Contralateral stimulation that exceeded those observed Postlesion, Ps õ 0.01, Tukey tests; ** Sham animals (but not those in the other groups) showed more frequent responses to Ipsilateral than to Contralateral stimuli, P õ 0.01, Tukey test.

deficiency in GnRH, not by brain damage (Gibson et al., 1984a,b, 1987). In addition, it is consistent with the bipotentiality of hamster neonates (Clemens and Witcher, 1985). Grafts from males and females that are older and more committed to a sex-typical pattern of

The effectiveness of implants did not seem to depend on the sex of their donors. This extends previous observations documenting similar responses to grafts from male or female fetuses or neonates, but in recipients in which the initial levels of behavior were limited by a TABLE 1 Mean Frequencies of Partial Lordosis (SEMs in Parentheses)

Group and side of stimulation Control Test phase Prelesion Postlesion Postimplant

Female

Male

Ipsi

Contra

Ipsi

Contra

Ipsi

Contra

1.00 (0) 1.00 (0) 1.00 (0)

1.00 (0) 0.80 (0.07) 0.85 (0.08)

1.00 (0) 1.00 (0) 1.00 (0)

1.00 (0) 0.35 (0.18) 1.00 (0)

1.00 (0) 1.00 (0) 0.85 (0.13)

1.00 (0) 0.75 (0.13) 1.00 (0)

Note. Instances of partial lordosis reinstatement were observed to manual stimulation of the flanks ipsilateral (Ipsi) and contralateral (Contra) to the VMN designated for treatment. Reinstatement frequencies were observed before unilateral VMN lesions (Prelesion), shortly after such treatments (Postlesion), and approximately 1 month after a control treatment (Control group) or the introduction into the lesion cavity of basal hypothalamic tissue from a neonatal female (Female group) or male (Male group) donor. Each group included four subjects. Analysis by Friedman’s ANOVAs documented a significant decline in responsiveness to contralateral stimuli immediately postlesion in members of the Female group (P Å 0.0227).

AID

H&B 1421

/

6810$$$$81

12-18-97 06:46:48

habal

AP: H & B

199

Graft-Induced Recovery of Lordosis

hypothalamic development might differ in their ability to promote the recovery of lordosis. However, grafts seem to sharply decline in effectiveness with increases in donor age, so that any such difference might be accompanied by such generally low levels of recovery to be of limited interest (Gibson et al., 1984a,b, 1987). While the facilitation of lordosis by our implants seems clear, it is necessary to consider one respect in which the understanding of this effect is limited by methodological factors. Surprisingly, perhaps, the most relevant limitation seems unrelated to our group sizes, which clearly did not prevent the emergence of very consistent effects. Instead, concern more properly revolves around the use of animals in natural estrus, a condition of optimal hormonal priming that permitted maximal levels of responsiveness to control (ipsilateral) stimulation. Because of this ceiling on levels of responsiveness and recovery, it would be premature to conclude that our implants completely eliminated the lesion-induced deficit in responsiveness to contralateral stimulation. Freed of this constraint, perhaps the level of responsiveness to ipsilateral stimuli would have remained above that to contralateral stimulation even after the full extent of implant-induced recovery. For the same reason, it would be premature to conclude that implants from male and females are equivalent in effectiveness. What is clear is that our implants uniformly occasioned very significant recoveries in lordosis. However, further testing, under controlled levels of hormonal priming, will be required for the full assessment of these effects. Further experimentation also will be necessary to clarify the related issues of the specificity of the responses to our implants (the extent to which implants of extrahypothalamic tissue would cause similar changes) and the mechanism responsible for these effects. Of the many alternatives (e.g, Kimble, 1990), the simplest mechanism would entail the graft’s production and release of a neuromodulator that facilitates lordosis (e.g., Whitman and Albers, 1995). It might be argued that such an explanation is inappropriate here, since any widely circulating facilitator of lordosis would cause postimplant increases in responsiveness to ipsilateral, as well as contralateral, stimuli. However, our failure to note such an improvement could reflect the ceiling effect discussed above. This mechanism also might be faulted because of the extensive damage suffered by the VMN, itself a potential target for neuromodulatory effects on lordosis. The force of this objection is reduced by the fact that VMN lesions typically were incomplete. In addition, neuromodulatory effects on lordosis have been suggested for forebrain sites in addition to the VMN (Whitman and Albers, 1995), raising the possibility that elevated levels of a lordosis-facilitating neuromodulator could compensate for VMN dysfunction by acting at one or more fully intact sites.

AID

H&B 1421

/

6810$$$$81

12-18-97 06:46:48

The net effect of these considerations is to leave open the possibility that our implants affected behavior by increasing levels of a lordosis-facilitating neuromodulator, a mechanism similar to those implicated in previous studies documenting effects of implants on lordosis (Gibson et al., 1984a,b, 1987; Luine et al., 1984). However, other mechanisms also merit consideration, some of which depend on the potential of grafts to build upon the spared capacity and circuits for lordosis that were suggested here by the persistence of partial lordosis responses following lesions. For example, each VMN (e.g., that targeted by our lesions) sends some fibers to the contralateral VMN (Conrad and Pfaff, 1976). In turn, the latter gives rise to a major projection to the contralateral PAG (the PAG contralateral to it and ipsilateral to the VMN in which this circuit originated), along with a smaller projection to the ipsilateral midbrain. Both sets of VMN projections to the PAG seem potentially relevant to lordosis. This is clearest for the crossed connections, which are held responsible for the lateralized effects on lordosis of manipulations of the VMN or its efferents (Floody, 1989; Ostrowski et al., 1981). However, even the uncrossed projections seem to affect lordosis, since hamsters with a small implant of estradiol just lateral to one VMN show longer lordosis responses when stimulated on the ipsilateral flank than when receiving no stimulation (Floody, 1989). These considerations suggest that neural implants could affect lordosis by strengthening or replacing relatively short projections from the lesioned to the intact VMN. Were the effects of this change to be distributed in proportion to the efferents of the unlesioned VMN, any increase in responsiveness to ipsilateral stimuli (stimuli ipsilateral to the intact VMN and contralateral to the lesion, i.e., the stimuli affected most by lesions and implants) would be accompanied by an increase in responsiveness to stimulation of the opposite (control) flank. Though this was not observed, its absence could reflect the ceiling our methods placed on levels of responsiveness. Alternatively, new inputs to the intact VMN could selectively increase the responses most disrupted by lesions, i.e., those controlled largely by projections from the intact VMN to the ipsilateral PAG. In this case, the resulting recovery of function would be specific, affecting responses to one flank only. As these examples suggest, grafts could affect lordosis by acting locally, on connections contained within the hypothalamus. However, they could instead act to strengthen or replace connections between the lesioned VMN and the PAG or some other downstream element in the system of structures controlling lordosis. Such actions are consistent with the mechanism suggested for the restoration of maletypical behavior by mPOA grafts, which tentatively is attributed to the ability of grafted cells to establish connections with the dorsolateral tegmentum, a normal mediator

habal

AP: H & B

200

Bartholomew and Floody

of mPOA influences on sexual behavior (Paredes et al., 1990, 1993). The additional work required to evaluate these mechanisms could capitalize on the methods that served so efficiently here to document the capacity of implants to reverse lesion-induced decrements in lordosis. For example, useful information on the extent of implant-induced recovery, the relative effectiveness of implants from male and female donors, and the dependence of recovery on a highly diffusible neuromodulator could be provided by extending the present methods to include observations under graded estrogen doses. Similarly, instances of recovery that depend on VMN tissue that was unintentionally spared by lesions possibly could be identified on the basis of responses to lesions differing in placement and size. Finally, while anatomical data ultimately are needed to document the dependence of recovery on implant-induced changes in neural connections, useful information on the relative importance of new local (hypothalamic) and remote (extrahypothalamic) connections would be provided by the comparison of implant effects on the deficits in lordosis induced by unilateral and bilateral VMN lesions.

ACKNOWLEDGMENTS We thank Debra Cook-Balducci and Matthew Bellace for help preparing the figures.

REFERENCES Arendash, G. W., and Gorski, R. A. (1982). Enhancement of sexual behavior in female rats by neonatal transplantation of brain tissue from males. Science 217, 1276 – 1278. Bjo¨rklund, A., and Stenevi, U. (1979). Reconstruction of the nigrostriatal dopamine pathway by intracerebral nigral transplants. Brain Res. 177, 555 – 560. Clemens, L. G., and Witcher, J. A. (1985). Sexual differentiation and development. In H. I. Siegel (Ed.), The Hamster: Reproduction and Behavior, pp. 155 – 171. Plenum, New York. Conrad, L. C. A., and Pfaff, D. W. (1976). Efferents from medial basal forebrain and hypothalamus in the rat. II. An autoradiographic study of the anterior hypothalamus. J. Comp. Neurol. 169, 221 – 262. Floody, O. R. (1989). Lateralized effects on hamster lordosis of unilateral hormonal and somatosensory stimuli. Brain Res. Bull. 22, 745–749. Floody, O. R. (1993). Cuts between the septum and preoptic area increase ultrasound production, lordosis, and body weight in female hamsters. Physiol. Behav. 54, 383 – 392. Floody, O. R., and Cramer, R. R. (1986). Effects of tegmental lesions on lordosis and body weight in female hamsters. Brain Res. Bull. 17, 59 – 66. Floody, O. R., and O’Donohue, T. L. (1980). Lesions of the mesence-

AID

H&B 1421

/

6810$$$$81

12-18-97 06:46:48

phalic central gray depress ultrasound production and lordosis by female hamsters. Physiol. Behav. 24, 79 – 85. Gibson, M. J., Charlton, H. M., Perlow, M. J., Zimmerman, E. A., Davies, T. F., and Krieger, D. T. (1984a). Preoptic area brain grafts in hypogonadal (hpg) female mice abolish effects of congenital hypothalamic gonadotropin-releasing hormone (GnRH) deficiency. Endocrinology 114, 1938 – 1940. Gibson, M. J., Krieger, D. T., Charlton, H. M., Zimmerman, E. A., Silverman, A.-J., and Perlow, M. J. (1984b). Mating and pregnancy can occur in genetically hypogonadal mice with preoptic area brain grafts. Science 225, 949 – 951. Gibson, M. J., Moscovitz, H. C., Kokoris, G. J., and Silverman, A.-J. (1987). Female sexual behavior in hypogonadal mice with GnRHcontaining brain grafts. Horm. Behav. 21, 211 – 222. Huang, H. H., Kissane, J. Q., and Hawrylewicz, E. J. (1987). Restoration of sexual function and fertility by fetal hypothalamic transplant in impotent aged male rats. Neurobiol. Aging 8, 465 – 472. Jenssen, C., Gruner, B., Do¨cke, F., and Do¨rner, G. (1988). Partial compensation of sexual receptivity deficits in female rats with bilateral lesions of the hypothalamic ventromedial nucleus by transplants of fetal mediobasal hypothalamic tissue. Exp. Clin. Endocrinol. 91, 287–300. Kimble, D. P. (1990). Functional effects of neural grafting in the mammalian central nervous system. Psychol. Bull. 108, 462 – 479. Krieger, D. T., Perlow, M. J., Gibson, M. J., Davies, T. F., Zimmerman, E. A., Ferin, M., and Charlton, H. M. (1982). Brain grafts reverse hypogonadism of gonadotropin releasing hormone deficiency. Nature 298, 468 – 471. Luine, V. N., Renner, K. J., Frankfurt, M., and Azmitia, E. C. (1984). Facilitated sexual behavior reversed and serotonin restored by raphe nuclei transplanted into denervated hypothalamus. Science 226, 1436 – 1439. Muntz, J. A., Rose, J. D., and Shults, R. C. (1980). Disruption of lordosis by dorsal midbrain lesions in the golden hamster. Brain Res. Bull. 5, 359 – 364. Ostrowski, N. L., Scouten, C. W., and Malsbury, C. W. (1981). Reduced lordosis response following unilateral hypothalamic knife cuts. Physiol. Behav. 27, 323 – 329. Paredes, R. G., Pin˜a, A. L., Ferna´ndez-Ruiz, J., and Bermu´dez-Rattoni, F. (1990). Fetal brain transplants induce recovery of male sexual behavior in medial preoptic area-lesioned rats. Brain Res. 523, 331–336. Paredes, R. G., Pin˜a, A. L., and Bermu´dez-Rattoni, F. (1993). Hypothalamic but not cortical grafts induce recovery of sexual behavior and connectivity in medial preoptic area-lesioned rats. Brain Res. 620, 351 – 355. Perlow, M. J., Freed, W. J., Hoffer, B. J., Seigler, A., Olson, L., and Wyatt, R. J. (1979). Brain grafts reduce motor abnormalities produced by destruction of nigrostriatal dopamine system. Science 204, 643 – 647. Pfaff, D. W. (1980). Estrogens and Brain Function. Springer-Verlag, New York. Siegel, S. (1956). Nonparametric Statistics for the Behavioral Sciences. McGraw-Hill, New York. Winer, B. J. (1971). Statistical Principles in Experimental Design. McGraw – Hill, New York. Whitman, D. C., and Albers, H. E. (1995). Role of oxytocin in the hypothalamic regulation of sexual receptivity in hamsters. Brain Res. 680, 73 – 79. Woodruff, M. L., Baisden, R. H., and Nonneman, A. J. (1992). Effects of transplantation of fetal hippocampal or hindbrain tissue into the brains of adult rats with hippocampal lesions on water maze acquisition. Behav. Neurosci. 106, 39 – 50.

habal

AP: H & B