- Email: [email protected]
2J mice: Skeletal, histological, and endocrinological evidence
EPILEPSY RESEARCH ELSEVIER
Epilepsy Research 20 (1995) 203-211
Chronic administration of sodium valproic acid slows pubertal maturation in inbred DBA/2J mice: Skeletal, histological, and endocrinological evidence Peter J. Snyder a,,, Lori L. Badura b a
Behavioral Epilepsy Program, Department of Psychiatry and Allegheny Neuropsychiatric Institute, Medical College of Pennsylvania and Hahneman University, Pittsburgh, PA, USA b Behavioral Neuroscience Division, Department of Psychology, State University of New York at Buffalo, Buffalo, NY, USA
Received 4 October 1994; accepted 6 October 1994
Abstract Sodium valproic acid (VPA) has been reported to occasionally delay pubertal maturation in children. In the current study, we sought to establish a valid animal model with which to further investigate the neuroendocrinological sequelae of VPA administration. Male and female DBA/2J mice were weaned at 2 weeks of age and administered either VPA (17-20 m g / k g / d a y ) or control solution via drinking water. Animals were weighed and sacrificed via decapitation at 4, 6, or 8 weeks of age. Testes and ovaries were prepared for histological analyses. In addition, the length of the left humerus bone from each animal was obtained as an index of skeletal growth, and trunk blood was assayed for circulating follicle-stimulating hormone (FSH) and prolactin (PRL). For males, testicular weights of the animals receiving VPA were significantly lower than those of control animals at all three sampling ages. No between-group differences were found for body weight at any sampling age, and yet the rate of skeletal maturation (zs indexed by humerus length) was decreased significantly for the VPA-treated males at all three sampling periods. Additionally, while hormone levels did not consistently differ, histological analyses of the gonadal tissue demonstrated signific~tntly decreased rates of spermatogenesis at all sampling points for VPA-treated animals. For females, chronic VPA administration led to a significant reduction in uterine weight at the 4 and 6 week sampling periods, and yet by 8 weeks of age the uterine weights for the two groups did not differ. Histological analyses of the ovarian tissue revealed that both the density of atrial follicles and corpora lutea were significantly less in VPA-treated animals after 4 weeks of treatment, but not after 2 or 6 weeks of treatment. However, these results were not corroborated by differences in circulating FSH or PRL levels. Finally, although body weight did not differ between the two groups at any sampling period, humerus length was significantly less for the VPA-treated females at the 4 week sampling period. These data indicwLe that chronic administration of VPA delays reproductive and skeletal maturation in genetically epilepsy-prone mice. Keywords: Valproic acid; Reproductivematuration;Skeletal maturation;Histological analysis
* Corresponding author. Behavioral Epilepsy Program, Department of Psychiatry and A. N. I., 320 East North Avenue, Pittsburgh, PA 15212, USA. 0920-1211/95/$09.50 ~ 1995 Elsevier Science B.V. All rights reserved SSDI 0920-121 1(94)00080-8
204
P.J. Snyder, L.L. Badura / Epilepsy Research 20 (1995) 203-211
I. Introduction
Valproic acid (VPA; 2-propylpentanoic acid), originally synthesized by Burton in 1882 (cited in [18]), was first shown to have anticonvulsant properties in 1963 [28]. Licensed for use in the United States as an antiepileptic drug (AED) since 1978, it is used primarily for the treatment of seizure disorders involving generalized spike-and-wave discharges [18]. The use of VPA in the treatment of epilepsy has increased dramatically since the late 1970s, and it is currently one of the most frequently prescribed AEDs. In addition to its more commonly known use as an AED, VPA has become increasingly accepted as a mood stabilizing agent for the treatment of a range of affective disorders, including major depression and bipolar disorder [19,27,37]. Despite these widespread clinical applications, however, the precise sites and mechanisms of VPA action remain poorly understood. VPA likely enhances GABA-mediated inhibition by increasing GABA concentrations in the CNS [13]. It has also been suggested to retard rapid neuronal firing, and alter the excitability of neuronal networks [11]. Detailed reviews of the pharmacological bases of VPA's anticonvulsant properties may be found elsewhere [6,13,14,20,25]. VPA has been shown to decrease the secretion of several pituitary hormones, such as LH, FSH, GH, and ACTH ([9,16,21,24], but see [22]), although clinical symptoms associated with decreased pituitary hormone release have not been described until recently [23,34]. Cook et al. [8], for example, described a 12 year-old girl with complex partial seizures who had been treated with VPA for several years. During this period of time the patient suffered an arrest of body growth and secondary sexual development, and yet two months after discontinuing VPA treatment she showed a resumption of both pubertal growth and skeletal maturation. This recent case report raises the possibility of adverse effects of chronic VPA administration, such as disturbances of growth, physical maturation, sexual development and fertility that are associated with the interruption of the normal secretion of pituitary hormones [24]. It might be argued that particular attention should be paid to the possibility of pubertal arrest, as VPA has been used in the treatment of multiple seizure
types in children [5,8]. Still, endocrine abnormalities have also been reported in adult patients following chronic administration, particularly in the form of unresponsiveness to GnRH challenge [7,8,24], transient amenorrhea [26], and both polycystic ovaries and hyperandrogenism [17]. There is little information available regarding the reversibility of these adverse effects following drug discontinuation. Thus, prolonged administration of VPA, if directly associated with continuous disruption of the hypothalamohypophyseal axis, could result in serious health risks. For example, significant declines in GH secretion, if continued throughout the period of skeletal joint fusation, may entirely abolish the skeletal growth spurt in a fashion similar to that seen in patients with hypopituitarism [12]. Prior research has shown that chronic VPA administration has little or no effect on skeletal maturation for the majority of pre-pubertal children placed on this AED, most likely due to the normality of other growth factors such as somatomedin-C [23]. Nonetheless, a deceleration of skeletal growth occasionally occurs in some patients who are placed on VPA [23], and the precise risk factors for this side effect remain unclear. Given recent descriptions of adverse clinical symptoms associated with VPA administration in some children, it seemed reasonable to further investigate VPA's ability to modulate pituitary hormone secretion and subsequent gonadal function. This study evaluates the effects of chronic VPA exposure on pubertal and skeletal maturation in an inbred strain of genetically epilepsy-prone mouse.
2. Methods
2.1. Animals and housing Male and female inbred mice (DBA/2J strain) were bred in the colony of Dr. V. Denenberg, Department of Biobehavioral Sciences, University of Connecticut (originally bred from a colony founded by progenitors from the Jackson Laboratory, Bar Harbor, Maine). Animals were weaned at 14 days of age and group housed (2-4 per cage) in polypropylene cages and had ad libitum access to food and water. Animals were maintained from birth through-
1o..1.Snyder, L.L. Badura/ Epilepsy Research 20 (1995) 203-211
out experimental use in a controlled photoperiod (12L:12D; lights off at 18.00 h) colony room. Since D B A / 2 J mice typically become reproductively mature at 30-35 days of age (V. Denenberg, personal communication), this genetic strain was selected for the rapid study of anti-reproductive and maturational effects of VPA in the developing animal. The sample size for all groups in this study was eight animals, with the exception of the VPA males sacrificed at 8 weeks ( N = 9), the control males sacrificed at 6 weeks (N = 9), and the control females sacrificed at 8 weeks ( N = 7). No animals died prematurely dutring the course of this study. 2.2. Drug treatment and collection of data
All mice received oral administration of VPA or the control solution vehicle. VPA, a highly water soluble agent with a clearance rate of 8-12 hours [5], was administered via the drinking water to which 1% sucrose had been added to increase palatability. VPA is readily absorbed from the digestive tract and oral administration is considered the preferable mode of delivery in a clinical setting [5]. Under these conditions, individual mice consume roughly 6 - 8 ml fluid/day. Drug administration was calculated to correlate with the dosages used in the clinical administration of VPA as an anti-convulsant (i.e., 15-60 m g / k g / d a y ) [1,2], such that animals received approximately 17 m g / k g / d a y . In addition, drug dosage was increased in two-week steps to account for developmental increases in body weight and tolerance effects. Thus, drinking water contained 60 /zg/ml VPA, weeks 2-4; 90 /zg/ml, weeks 4-6; and 120 /xg/ml, weeks 6-8. At the conclusion of drug administration (4, 6, or 8 weeks of age), all animals were sacrificed via decapitation between 08.00 and 10.00 hours to minimize circadian effects on circulating hormone levels, and trunk blood was collected for determination of FSH and PRL. Sub,;equently, the uteri and ovaries, or testes, were removed and weighed, and the gonads were post-fixed in 10% formalin for 1-3 weeks, followed by cryoprotection in 10% formalin-30% sucrose for 3 - 4 weeks prior to sectioning for histological evaluation of gonadal function (i.e., follicular development or sperm production). Frozen sections (40/zm) of each gonad type were
205
taken on a sliding microtome, mounted on slides, and stained using cresyl violet. For testes, the tissue was evaluated for the presence of spermatogonia, primary spermatocytes, elongated spermatocytes, and sperm using a five-point scale with '0' indicating an absence of the variable (prepubescent) and '4' indicating an abundance of the variable (reproductively mature). All ratings were conducted by two observers blind to the experimental assignment of the animals. The two sets of ratings for each of the four cell types were well correlated (range: R = 0.72, P < 0 . 0 1 to R = 0 . 9 0 , P < 0 . 0 0 1 ; mean R = 0 . 7 7 , P < 0.01), indicating satisfactory inter-rater reliability for these ratings; thus the values obtained by each rater on each measure were averaged for further analysis. For ovarian tissue, the sections were evaluated for the presence of atrial follicles and corpora lutea using a three-point scale with '0' indicating an absence of that variable and '2' indicating an abundance of the variable. Ratings were conducted by two observers blind to the experimental assignment of the animals. The two sets of ratings for both the atrial follicles (R = 0.46, P < 0.001) and corpora lutea (R = 0.66, P < 0.0001) were well correlated, making it permissible to average the ratings from each observer for further statistical analyses. Finally, the left humerus bone was fully exposed and its length (mm) was recorded. The humerus bone was selected for measurement as an index of skeletal growth, as previous evidence from this laboratory (H.P. Mclsaac, unpublished observations) has shown that humerus length varies independently of gonadal developmental status in a different rodent species (Siberian hamsters), thus providing a valid measurement of non-reproductively related growth. 2.3. Blood sampling and radioimmunoassays
Trunk blood, obtained at the time of sacrifice, was allowed to clot for 24 h at 4°C, centrifuged for 20 min at 4500 rpm to extract the serum, and then stored at - 5 0 ° C until determination of hormone levels. Samples were assayed for prolactin (PRL) and follicle-stimulating hormone (FSH) levels via double antibody radioimmunoassay. For PRL, anti, hamster PRL (1:15,000) donated by Dr. Borer was used as the primary antibody with goat anti-rabbit
P.J. Snyder, L.L. Badura / Epilepsy Research 20 (1995) 203-211
206 VPA ~ OVWT
CON ~ / 7 / ~ OVWT
VPA ~ UTWT
CON UTWT 140 130
E 1-
6
120
5
110 100 E 9o
80
4 3
>'< O
2 1
4 WEEKS
6 WEEKS AGE (WEEKS)
~ O ,'n
7O 60 50 40
~ m z w
30 20 10 0
~
separately for each sex using 2 × 2 (drug × age) between-subjects analysis of variance. Follow-up analyses were conducted with analysis of simple main effects and/or Fisher's LSD where appropriate. Ratings of gonadal status were analyzed with the Mann-Whitney U-test for independent samples, and inter-rater reliabilities (see above) were calculated with the Pearson product-moment correlation. All values were considered significant if P < 0.05.
3. Results
8 WEEKS
Fig. 1. Mean ( + SEM) ovarian ( O V W T ) and uterine ( U T W T ) weights (mg) for control and VPA-treated groups at each sampling point in the experiment. (* Significantly greater than the VPA-treated group within sampling time; see text for details.)
gamma globulin (GARGG; 1:12) used as second antibody. Purified hamster PRL obtained from Dr. Parlow and iodinated using the chloromine-T method served as trace, and serial dilutions of this hormone provided the standard. Curves derived for pooled sera from gonadally intact male and lactating female mice are parallel to the curve obtained with this standard (Badura, unpublished observations). Samples (35 /zl) were dispensed in 0.4% bovine albumin-0.1% gelatinized phosphate-buffered saline (GeI-PBS; pH 7.0) to a final volume of 300 /zl. The lower limit of detectability of the assay (90% bound) was 0.91 ng/ml, and the intra-assay coefficients of variance (CVs) at 50% and 65% bound were 11.24 and 10.3, respectively. FSH levels were determined using the NIAMMD rat FSH kit, courtesy of Dr. Raiti, using NIADDKFSH-RP-2 as the reference standard and iodinated NIADDK-rFSH-I6 as the trace. NIADDK-rFSH-S-11 (1:1500) served as primary antibody and GARGG (1:12) as the secondary antibody. Samples (20 /xl) were dispensed in Gel-PBS to a final volume of 300 /xl. The lower limit of detectability of the assay (90% bound) was 1.25 n g / m l and the CVs at 35% and 85% bound were 11.23 and 3.47, respectively.
3.1. Gonadal weights For females, chronic VPA administration led to a significant retardation in the development of uterine weight at the four-week ( F = 5.46(1,14), P = 0.03) and six-week ( F = 16.12(1,14), P = 0.001) sampiing periods (see Fig. 1). With regard to ovarian weights, although no significant differences were found at the four-week sampling period, a trend towards a significant difference was found at 6 weeks of age (4 weeks of VPA treatment), with the VPAtreated females showing reduced ovarian weights in comparison to controls ( F = 3.5(1,14), P = 0.08; see Fig. 1). By 8 weeks of age, the two groups did not differ across both measures of uterine and ovarian weight. For males, VPA administration significantly decreased testicular growth at the four-week ( F = VPA
Gonadal weights, body weights, circulating hormone levels, and humerus lengths were analyzed
CON
220 200 180 E ~" 160 ~140 120 100 8O O.9_ I.- 6O
~- 40 20 0 4 WEEKS
2.4. Statistical analyses
~
6 WEEKS
8 WEEKS
AGE (WEEKS)
Fig. 2. Mean (_+SEM) testicular weights (rag) for control and VPA-treated groups at each sampling point in the experiment. (* Significantly greater than the VPA-treated group within sampling time; see text for details.)
P.J. Snyder, L.L. Badura ~Epilepsy Research 20 (1995) 203-211 I
VPA
~
PRL
CON
~
PRL
VPA FSH
~
I
CON
12 10
VPA
~
PRL
FSH 16
6 -
14
5
207
CON
~
VPA
PRL
FSH
~
CON FSH 70
~
~
6o
12
_z ~<
8~
6
~3 <,
~4
6
30
u.
2
20 u"
4 2 8
4 WEEKS
6 WEEKS
0
8 WEEKS
0 4 WEEKS
AGE (WEEKS)
6 WEEKS
8 WEEKS
AGE (WEEKS)
Fig. 3. Mean ( + SEM)circulating levels of prolactin (PRL) and follicle-stimulating hormone (FSH) for control and VPA-treated females at each sampling point in the experiment.
5.98(1,14), P = 0.03), six-week ( F = 7.73(1,15), P =0.01), and eight-week ( F = 12.24(1,15), P = 0.003) sampling periods, although there were significant within-group increases over time (see Fig. 2).
3.2. Histological analyses For males, chronic VPA administration led to significant decreases in spermatogenesis (as evidenced by observed proliferation of sperm) at the four-week (U = 12.0, two-tailed P = 0.03), six-week (U = 17.0, two-tailed P = 0.045), and eight-week
Fig. 4. Mean (_+ SEM) circulating levels of prolactin (PRL) and follicle-stimulating hormone (FSH) for control and VPA-treated males at each sampling point in the experiment.
(U = 16.0, two-tailed P = 0.046) sampling periods (see Table 1). For ovarian tissue, VPA treatment was associated with a significant reduction at the six-week sampling time of both atrial follicular number (U = 6.5, twotailed P = 0.007) and corpora lutea number (U = 9.5, two-tailed P = 0.02). Significant differences were not found for either measure at the other sampling points, although a trend toward decreased number was observed for atrial follicles at 8 weeks (U = 13.5, two-tailed P = 0.09) (see Table 1).
Table 1 Mean ( + SEM) ratings for gonadal tissue for control and VPA-treatment group " Group
Sex
No. primary spermatocytes
No. elongated spermatocytes
No. sperm
Sex
No. atrial follicles
No. corpora lutea
Four week Controls VPA Tx.
M M
3.44 (0.12) 3.06 (0.19)
1.44 (0.28) 1.50 (0.34)
1.88 (0.30) 1.06 (0.16) b
F F
1.59 (0.17) 1.28 (0.19)
1.38 (0.21) 1.21 (0.25)
Six week Controls VPA Tx.
M M
3.72 (0.09) 3.63 (0.17)
3.28 (0.20) 2.94 (0.19)
3.11 (0.19) 2.63 (0.08) b
F F
1.91 (0.10) 1.19 (0.19) b
1.94 (0.11) 1.25 (0.25) b
Eight week Controls VPA Tx.
M M
3.88 (0.09) 3.78 (0.09)
3.75 (0.20) 3.67 (0.15)
3.69 (0.14) 2.83 (0.35) b
F F
1.75 (0.28) 1.34 (0.14) ¢
1.64 (0.23) 1.09 (0.28)
a Ratings based on ordinal scales for analyses of histological preparations of gonadal tissue, as described in Results Section. b Ratings for VPA-treated group are significantly decreased vs. those for control group, at P < 0.05. c Ratings for VPA-treated group exhibit trend toward decrease vs. those for control group, at P = 0.09.
208
P.J. Snyder, L.L. Badura /Epilepsy Research 20 (1995) 203-211 I
VPA MALE
~
VPA ~ FEMALE
CON MALE
. . . . .
CON FEMALE
3.4. Rate of skeletal growth: humerus bone length
14
12 E E ~10
m
6
w ~
4
i 2
4 WEEKS
6 WEEKS AGE (WEEKS)
8 WEEKS
For the females, humerus length was significantly smaller for VPA-treated animals at the four-week sampling time ( F = 11.30(1,14), P = 0 . 0 0 5 ) , approached a significant difference at the six-week sampling time ( F = 3.88(1,14), P = 0.07), and did not differ between groups at the eight-week sampling time. For the males, chronic VPA administration led to significantly decreased humerus bone lengths at the four-week ( F = 15.75(1,14), P = 0 . 0 0 1 ) , sixweek ( F = 5.62(1,15), P = 0.03), and eight-week ( F = 8.82(1,15), P = 0.01) sampling periods (see Fig. 5).
Fig. 5. Mean (_ SEM) humerus lengths (mm) for both male and female control (CON) and VPA-treated mice at each sampling point in the experiment. (* Significantly greater lengths than VPA-treated group within samplingtime, see text for details.)
3.5. Total body weight
3.3. Radioimmunoassay of PRL and FSH release
Body weights, for both males and females, were not significantly affected by chronic VPA administration at any time point (see Fig. 6).
VPA administration did not have a significant effect upon circulating FSH or PRL levels in females, although there was a trend toward depression of FSH at the six-week sampling time ( F = 3.91(1,14), P = 0.06; see Fig. 3). VPA administration did not have a significant effect upon circulating FSH or PRL levels in males at any sampling time (see Fig. 4).
I
VPA MALE
~
CON MALE
~
VPA ~ FEMALE
CON FEMALE
30
L9 I.U >-
~DIO
@ 4 WEEKS
6 WEEKS AGE (WEEKS)
8 WEEKS
Fig. 6. Mean ( + SEM) body weights(g) for both male and female control (CON) and VPA-treated mice at each sampling point in the experiment.
4. Discussion This study was designed to investigate the neuroendocrinological bases for the role that chronic administration of VPA might play in attenuating pubertal growth in children. An animal model for this research was selected, as it afforded the ability to study the effects of VPA on testicular and ovarian maturation. Our findings indicate that, in female mice, chronic VPA administration leads to significant retardation in the development of uterine weight early in development, but by 8 weeks of age, the VPA-treated groups did not differ from controls on measures of uterine and ovarian weights. These findings were mirrored by the results obtained for rating of the ovarian histology, with the greatest decrease in atrial follicular number and corpora lutea number seen at 6 weeks of age. Similar trends were observed for the effects of VPA treatment on humerus length, with females showing significant differences early in development, and no noticeable differences at 8 weeks. However, VPA significantly decreased the rate of testicular and skeletal growth, in males, at all three time points. These combined results (for both males
P,I. Snyder, L.L. Badura // Epilepsy Research 20 (1995) 203-211
and females) indicate that growth was slowed, but not prevented, in the VPA-treated animals. Interestingly, more consistent differences with regard to gonadal weight and skeletal growth were found in males rather than in females. This observation may be clinically relevant, as any effects of VPA on delayed pubertal maturation in young boys may be less readily apparent than in girls. Amenorrhea and slowed development of secondary sex characteristics in females are arguably more noticeable than is delayed testicular development in boys. Although significant differences in circulating FSH and PRL were not observed among groups in the current study, the data for FSH do suggest a delayed 'surge' in circulating levels in males treated with VPA. That is, FSH appeared to peak in control animals at 4 weeks of age, where~s this peak seemed to occur in VPA-treated animals at 6 weeks of age (see Fig. 4). This 'surge' of FSH is considered to be a marker of pubertal onset in many species. Regardless of potential sex differences in the slowing of physical growth and maturation by VPA's action on the hypothalamic-pituitary-gonadal axis, our results in males show that VPA does not prevent or halt pubertal onset, but instead may act to slow the rate of normal maturational changes. This study does not address the idiosyncratic, very rapid rate of murine metabolism of VPA [30], and nor does it address the inconsistent rates of delivery of the drug via drinking water in animals who are nocturnal, and thus more likely to drink during the lights-off phase. Our method of delivery likely resulted in a palLtern of blood plasma levels of VPA which mimics a 'TID' (thrice daily) or 'QID' (four times daily) drug administration pattern in humans, with a long break overnight [31]. It is not like the more constant blood plasma levels of VPA achieved by q6 hours ,or q8 hours dosing utilized for institutionalized patients or those with severe childhood epilepsy. Thus, the fact that we found significant differences betwe,en our VPA-treated and control groups, across measures of gonadal development and skeletal maturation, attests to the robust nature of the effect of VPA on the hypothalamic-hypophyseal axis--that is, despite a non-constant-rate method of drug delivery. However, our inability to detect group differences in PRL and FSH levels might be directly related to our chosen method of
209
drug delivery. Future efforts, relying on an osmotic pump delivery (subcutaneous implantation), leading to a constant rate of application [31], should re-examine group differences in PRL and FSH secretion, as well as evaluate potential effects on circulating GH. Although estrous cyclicity was not directly monitored in this study, it is unlikely that proestrous-related variations in PRL and FSH contributed to the experimental variability as these samples were obtained in the early morning hours, and surges of these anterior pituitary hormones in rodents do not occur until late afternoon [3]. Further studies are in progress that will include the systematic investigation of the side effects of VPA at different dose levels, with the hope that such work might lead to refinements in therapeutic dosing strategies and the identification of potential warning signals of endocrinological dysfunction during treatment. Information regarding the neurochemical basis of VPA action within the hypothalamo-hypophyseal system might lead to the development of agents which block the negative actions of VPA on hormone release without diminishing its anticonvulsant properties. One possibility may be a disruption of the cholinergic system that regulates GH secretion during nocturnal sleep. Legido et al. [22] have provided evidence to support this possibility in a study of 35 epileptic children (25 of whom were placed on VPA for at least 6 months) and 26 neurologically normal children. They found that neither carbamazepine nor VPA affect the somatostatinergic, GABAergic, or dopaminergic systems that regulate PRL, TSH, and gonadotropin secretion. However, the chronic administration of either of the two AEDs was found to significantly reduce GH release during nocturnal sleep. As described above, chronic VPA administration may lead to global decreases in pituitary hormone secretion, with abnormal growth and development in some human patients [8,10,24,26], and yet it is unclear where along the hypothalamic-pituitary axis VPA might act to diminish pituitary activity. The hypothalamus contains a multitude of peptidergic stimulating and inhibiting systems which influence the release of a variety of protein hormones by the pituitary, and each system possesses unique anatomical substrates and morphological characteristics (see [32,33] for reviews). In addition, these hypothalamic
210
P.J. Snyder, L.L. Badura / Epilepsy Research 20 (1995) 203-211
neuroendocrine systems are further modulated by a plexus of neuronal inputs, both of peptidergic and classical neurotransmitter origin (see [36] for a review). Our understanding of this multi-integrated hypothalamo-hypophyseal axis is still incomplete. For example, it is widely accepted that dopamine of tuberoinfundibular origin represents the major inhibitory influence upon PRL secretion from the anterior pituitary (see [4,29] for reviews); however, the hypothalamic releasing systems which comprise important stimulatory inputs for this system are still subject to debate (see [15,35] for reviews). Further work will also center on elucidating the sites and mechanisms of VPA action along the hypothalamic-pituitary axis. Preliminary data from this laboratory have indicated that anterior pituitary tissue harvested from animals that have been chronically treated with VPA shows alterations in basal FSH release in a perfusion culture system that parallel the trends in the group differences reported above for circulating levels (Badura and Snyder, unpublished data, 1994). Thus, although many of the endocrinological analyses in the present study yielded non-significant differences in the predicted direction, the pilot data obtained from the tissue culture study provide further support for effects of VPA on the hypothalamic-pituitary axis, possibly directly on the pituitary itself. We recognize that our comparisons of the endocrinological data in the present report suffered from low statistical power. Sample sizes will be adjusted accordingly, in future studies, to provide increased statistical power for the analyses of variables that are prone to substantial individual differences in pubertal animals. Acknowledgements The authors wish to express their appreciation to Dr. Victor Denenberg for providing the mice and colony space for this study, and to Aaron Wilson for technical support. Also, we are grateful to Dr. Borer and Dr. Parlow for their donations of reagents for use in the PRL RIA, and Dr. Riati for reagents used in the FSH RIA. This project was completed while P.J.S. was a Wilder Penfield Research Fellow of the Epilepsy Foundation of America, and it was funded in part by NIH institutional grant NS07324 and NSF grant IBN9224804 funds to L.L.B.
References [1] Axnaout, M.A., Garthwaite, T.L., Martinson, D.R. and Hagen, T.C., Vasoactive intestinal polypeptide is synthesized in anterior pituitary tissue, Endocrinology, 119 (1986) 20522057. [2] Badura, L.L. and Goldman, B.D., Prolactin-dependent seasonal changes in pelage: Role of the pineal gland and dopamine, J. Exp. Zool., 261 (1992) 27-33. [3] Badura, L.L., Sisk, C.L. and Nunez, A.A., Effects of photoperiod and hypothalamic knife cuts on the timing of FSH surges in hamsters, Brain Res. Bull., 26 (1991) 313-316. [4] Ben-Jonathan, N., Arbogast, L.A. and Hyde, J.F., Neuroendocrine regulation of prolactin release, Prog. Neurobiol., 33 (1989) 399-447. [5] Bruni, J. and Wilder, B.J., Valproic acid: Review of a new antiepileptic drug, Arch Neurol., 36 (1979) 393-398. [6] Chapman, A , Keane, P.E, Meldrum, B.S., Simiand, L and Vernieres, J.C, Mechanism of anticonvulsant action of valproate, Prog. Neurobiol., 19 (1982) 315-359. [7] Conran, M.LC., Kearney, PJ., Callaghen, M.N., Murphy, D. and Goggin, T., Hypothalamic pituitary function testing on children receiving carbamazepine or sodium valproate, Epilepsia, 26 (1985) 585-588. [8] Cook, J.S., Bale Jr., J.F. and Hoffman, R.P. Pubertal arrest associated with valproic acid therapy, Pediatr. Neurol., 8 (1992) 229-231. [9] Crawford, P., Belchetz, P. and Chadwick, D., Growth hormone response of diazepam, clonidine, and glucagon in patients with epilepsy, Epilepsy Res., 3 (1989) 63-69. [10] Egger, J. and Brett, E.M., Effects of sodium valproate in 100 children with special reference to weight, Br. Med. J., 283 (1981) 577-581. [11] Faingold, C.L. and Browning, R.A., Mechanisms of anticonvulsant drug action, Eur. J. Pediatr., 146 (1987) 8-14. [12] Frohman, LA., The anterior pituitary. In: J.B. Wyngarden, L.H. Smith Jr. and J.C. Bennett (Eds.), Textbook of Medicine (19th Edition), Vol. 2, W.B. Saunders, Philadelphia, 1992, chapter 213, pp. 1228-1231. [13] Godin, Y., Heiner, L., Mark, J. and Mandel, P., Effects of di-n-propylacetate, an anticonvulsive compound, on GABA metabolism, J. Neurochem., 16 (1969) 869-873. [14] Hackman, J.C., Grayson, V. and Davidoff, R.A, The presynaptic effects of valproic acid in the isolated frog spinal cord, Brain Res., 220 (1981) 269-285. [15] Hyde, J.F. and Ben-Jonathan, N., Characterization of prolactin-releasing factor in the rat posterior pituitary, Endocrinology, 122 (1988) 2533-2539. [16] Invitti, C., Danesi, L., Dubini, A. and Cavagnini, F., Neuroendocrine effects of chronic administration of sodium valproate in epileptic patients, Acta Endocrinol., 118 (1986) 381-388. [17] lsojarvi, J.I.T, Laatikainen, T.J., Pakarinen, A.J., Juntunen, K.T.S. and Myllyla, V.V., Polycystic ovaries and hyperandrogenism in women taking valproate for epilepsy, N. EngL J. Med., 329 (1993) 1383-1388. [18] Johnston, D., Valproic acid: Update on its mechanisms of action, Epilepsia, 25, Suppl. 1 (1984) S1-$4.
P.J. Snyder, L.L. Badura / Epilepsy Research 20 (1995) 203-211 [19] Kemp, L.I., Sodium valproate as an antidepressant, Br. J. Psychiatry, 160 (199:2) 121-123. [20] Kerwin, R.W., Olpe, H.R. and Schmutz, M., The effect of sodium-n-dipropyl acetate on GABA-dependent inhibition in the rat cortex and substantia nigra in relation to its anticonvulsant activity, Br. J. Pharmacol., 71 (1980) 545-551. [21] Kritzler, R., Vining, E. and Plotnick, L., Sodium valproate and corticotropin suppression in the child treated for seizures, J. Pediatr., 102 (19813) 142-143. [22] Legido, A., Lopez, M.J., Garagorri, J., Orden, I., Baselga, C., Bueno, M. and Grover, W.D., Hypothalamo-pituitary function in epileptic children treated with carbamazepine or valproate, Epilepsia, 33, Suppl. 3 (1992) 9. [23] Legido, A., Lopez, ~'.J., Garagorri, J., Orden, I., Esteva, F., Baselga, C., Bueno, M. and Grover, W.D., Growth and growth hormone secretion in epileptic children treated with valproate, Epilepsia, 33, Suppl. 3 (1992) 9. [24] Lundberg, B., Nergardh, A., Ritzen, E. and Samuelson, K., Influence of valproic acid on the gonadotropin releasing hormone test in puberty, Acta Pediatr. Scand., 75 (1986) 787-792. [25] MacDonald, R.L. and Bergey, G.K., Valproic acid augments GABA mediated postsynaptic inhibition in cultured mammalian neurons, Brain Res., 170 (1979) 558-562. [26] Margraf, J.W. and Dreifuss, F.E., Amenorrhea following intiation of therapy with valproic acid, Neurology, 31 (1981) 159. [27] McElroy, S.L., Keck, P.E. Jr., Pope, H.G., Jr. and Hudson, J.l., Valproate in the treatment of bipolar disorder: Literature review and clinical guidelines, J. Clin. Psychopharmacol., 12, Suppl. 1 (1992) S,~.2-$52. [28] Meunier, H., Carraz, G., Meunier, Y., Eymard, P. and Aimard, M., Proprietes pharmacodynamique de l'acide n-dipropylacetique, Therapie, 18 (1963) 435-438.
211
[29] Moore, K.E. and Demarest, K.T., Tuberoinfundibular and tuberohypophyseal dopaminergic neurons. In: W.F. Ganong and L Matrini (Eds.), Frontiers in Neuroendocrinology, Raven Press, New York, 1982, pp. 161-190. [30] Nau, H. and L/Sscher, W., Valproic acid and active unsaturated metabolite (2-en): Transfer to mouse liver following human therapeutic doses, Biopharm. Drug Dispos., 6 (1985) 1-8. [31] Nau, H. and Zierer, R., Pharmacokinetics of valproic acid and metabolites in mouse plasma and brain following constant-rate application of the drug and its unsaturated metabolite with an osmotic delivery system, Biopharm. Drug Dispos., 3 (1982) 317-328. [32] Page, R.B., The anatomy of the hypothalamo-hypophyseal complex. In: E. Knobil and J. Neill (Eds.), The Physiology of Reproduction, Raven Press, New York, 1988, pp. 1161-1233. [33] Sagar, S.M. and Martin, J.B., Hypothalamohypophysiotropic peptide systems. In: V.B. Mountcastle, F.E. Bloom and S.E. Geiger (Eds.), Handbook of Physiology, Waverly Press, Baltimore, 1986, pp. 413-462. [34] Schmidt, D., Adverse effects of valproate, Epilepsia, 25 (1984) $44-$49. [35] Shin, S.H., Papas, S. and Obansawin, M.C., Current status of the rat prolactin-releasing factor, Can. J. PhysioL Pharmacol., 65 (1987) 2036-2043. [36] Swanson, L.W., Organization of mammalian neuroendocrine system. In: V.B. Mountcastle, F.E. Bloom and S.E. Geiger (Eds.), Handbook of Physiology, Waverly Press, Baltimore, 1986, pp. 317-363. [37] Wilder, B.J., Pharmacokinetics of valproate and carbamezapine, J. Clin. Psychopharm., 12 (1992) $64-$68.
Epilepsy Research 20 (1995) 203-211
Chronic administration of sodium valproic acid slows pubertal maturation in inbred DBA/2J mice: Skeletal, histological, and endocrinological evidence Peter J. Snyder a,,, Lori L. Badura b a
Behavioral Epilepsy Program, Department of Psychiatry and Allegheny Neuropsychiatric Institute, Medical College of Pennsylvania and Hahneman University, Pittsburgh, PA, USA b Behavioral Neuroscience Division, Department of Psychology, State University of New York at Buffalo, Buffalo, NY, USA
Received 4 October 1994; accepted 6 October 1994
Abstract Sodium valproic acid (VPA) has been reported to occasionally delay pubertal maturation in children. In the current study, we sought to establish a valid animal model with which to further investigate the neuroendocrinological sequelae of VPA administration. Male and female DBA/2J mice were weaned at 2 weeks of age and administered either VPA (17-20 m g / k g / d a y ) or control solution via drinking water. Animals were weighed and sacrificed via decapitation at 4, 6, or 8 weeks of age. Testes and ovaries were prepared for histological analyses. In addition, the length of the left humerus bone from each animal was obtained as an index of skeletal growth, and trunk blood was assayed for circulating follicle-stimulating hormone (FSH) and prolactin (PRL). For males, testicular weights of the animals receiving VPA were significantly lower than those of control animals at all three sampling ages. No between-group differences were found for body weight at any sampling age, and yet the rate of skeletal maturation (zs indexed by humerus length) was decreased significantly for the VPA-treated males at all three sampling periods. Additionally, while hormone levels did not consistently differ, histological analyses of the gonadal tissue demonstrated signific~tntly decreased rates of spermatogenesis at all sampling points for VPA-treated animals. For females, chronic VPA administration led to a significant reduction in uterine weight at the 4 and 6 week sampling periods, and yet by 8 weeks of age the uterine weights for the two groups did not differ. Histological analyses of the ovarian tissue revealed that both the density of atrial follicles and corpora lutea were significantly less in VPA-treated animals after 4 weeks of treatment, but not after 2 or 6 weeks of treatment. However, these results were not corroborated by differences in circulating FSH or PRL levels. Finally, although body weight did not differ between the two groups at any sampling period, humerus length was significantly less for the VPA-treated females at the 4 week sampling period. These data indicwLe that chronic administration of VPA delays reproductive and skeletal maturation in genetically epilepsy-prone mice. Keywords: Valproic acid; Reproductivematuration;Skeletal maturation;Histological analysis
* Corresponding author. Behavioral Epilepsy Program, Department of Psychiatry and A. N. I., 320 East North Avenue, Pittsburgh, PA 15212, USA. 0920-1211/95/$09.50 ~ 1995 Elsevier Science B.V. All rights reserved SSDI 0920-121 1(94)00080-8
204
P.J. Snyder, L.L. Badura / Epilepsy Research 20 (1995) 203-211
I. Introduction
Valproic acid (VPA; 2-propylpentanoic acid), originally synthesized by Burton in 1882 (cited in [18]), was first shown to have anticonvulsant properties in 1963 [28]. Licensed for use in the United States as an antiepileptic drug (AED) since 1978, it is used primarily for the treatment of seizure disorders involving generalized spike-and-wave discharges [18]. The use of VPA in the treatment of epilepsy has increased dramatically since the late 1970s, and it is currently one of the most frequently prescribed AEDs. In addition to its more commonly known use as an AED, VPA has become increasingly accepted as a mood stabilizing agent for the treatment of a range of affective disorders, including major depression and bipolar disorder [19,27,37]. Despite these widespread clinical applications, however, the precise sites and mechanisms of VPA action remain poorly understood. VPA likely enhances GABA-mediated inhibition by increasing GABA concentrations in the CNS [13]. It has also been suggested to retard rapid neuronal firing, and alter the excitability of neuronal networks [11]. Detailed reviews of the pharmacological bases of VPA's anticonvulsant properties may be found elsewhere [6,13,14,20,25]. VPA has been shown to decrease the secretion of several pituitary hormones, such as LH, FSH, GH, and ACTH ([9,16,21,24], but see [22]), although clinical symptoms associated with decreased pituitary hormone release have not been described until recently [23,34]. Cook et al. [8], for example, described a 12 year-old girl with complex partial seizures who had been treated with VPA for several years. During this period of time the patient suffered an arrest of body growth and secondary sexual development, and yet two months after discontinuing VPA treatment she showed a resumption of both pubertal growth and skeletal maturation. This recent case report raises the possibility of adverse effects of chronic VPA administration, such as disturbances of growth, physical maturation, sexual development and fertility that are associated with the interruption of the normal secretion of pituitary hormones [24]. It might be argued that particular attention should be paid to the possibility of pubertal arrest, as VPA has been used in the treatment of multiple seizure
types in children [5,8]. Still, endocrine abnormalities have also been reported in adult patients following chronic administration, particularly in the form of unresponsiveness to GnRH challenge [7,8,24], transient amenorrhea [26], and both polycystic ovaries and hyperandrogenism [17]. There is little information available regarding the reversibility of these adverse effects following drug discontinuation. Thus, prolonged administration of VPA, if directly associated with continuous disruption of the hypothalamohypophyseal axis, could result in serious health risks. For example, significant declines in GH secretion, if continued throughout the period of skeletal joint fusation, may entirely abolish the skeletal growth spurt in a fashion similar to that seen in patients with hypopituitarism [12]. Prior research has shown that chronic VPA administration has little or no effect on skeletal maturation for the majority of pre-pubertal children placed on this AED, most likely due to the normality of other growth factors such as somatomedin-C [23]. Nonetheless, a deceleration of skeletal growth occasionally occurs in some patients who are placed on VPA [23], and the precise risk factors for this side effect remain unclear. Given recent descriptions of adverse clinical symptoms associated with VPA administration in some children, it seemed reasonable to further investigate VPA's ability to modulate pituitary hormone secretion and subsequent gonadal function. This study evaluates the effects of chronic VPA exposure on pubertal and skeletal maturation in an inbred strain of genetically epilepsy-prone mouse.
2. Methods
2.1. Animals and housing Male and female inbred mice (DBA/2J strain) were bred in the colony of Dr. V. Denenberg, Department of Biobehavioral Sciences, University of Connecticut (originally bred from a colony founded by progenitors from the Jackson Laboratory, Bar Harbor, Maine). Animals were weaned at 14 days of age and group housed (2-4 per cage) in polypropylene cages and had ad libitum access to food and water. Animals were maintained from birth through-
1o..1.Snyder, L.L. Badura/ Epilepsy Research 20 (1995) 203-211
out experimental use in a controlled photoperiod (12L:12D; lights off at 18.00 h) colony room. Since D B A / 2 J mice typically become reproductively mature at 30-35 days of age (V. Denenberg, personal communication), this genetic strain was selected for the rapid study of anti-reproductive and maturational effects of VPA in the developing animal. The sample size for all groups in this study was eight animals, with the exception of the VPA males sacrificed at 8 weeks ( N = 9), the control males sacrificed at 6 weeks (N = 9), and the control females sacrificed at 8 weeks ( N = 7). No animals died prematurely dutring the course of this study. 2.2. Drug treatment and collection of data
All mice received oral administration of VPA or the control solution vehicle. VPA, a highly water soluble agent with a clearance rate of 8-12 hours [5], was administered via the drinking water to which 1% sucrose had been added to increase palatability. VPA is readily absorbed from the digestive tract and oral administration is considered the preferable mode of delivery in a clinical setting [5]. Under these conditions, individual mice consume roughly 6 - 8 ml fluid/day. Drug administration was calculated to correlate with the dosages used in the clinical administration of VPA as an anti-convulsant (i.e., 15-60 m g / k g / d a y ) [1,2], such that animals received approximately 17 m g / k g / d a y . In addition, drug dosage was increased in two-week steps to account for developmental increases in body weight and tolerance effects. Thus, drinking water contained 60 /zg/ml VPA, weeks 2-4; 90 /zg/ml, weeks 4-6; and 120 /xg/ml, weeks 6-8. At the conclusion of drug administration (4, 6, or 8 weeks of age), all animals were sacrificed via decapitation between 08.00 and 10.00 hours to minimize circadian effects on circulating hormone levels, and trunk blood was collected for determination of FSH and PRL. Sub,;equently, the uteri and ovaries, or testes, were removed and weighed, and the gonads were post-fixed in 10% formalin for 1-3 weeks, followed by cryoprotection in 10% formalin-30% sucrose for 3 - 4 weeks prior to sectioning for histological evaluation of gonadal function (i.e., follicular development or sperm production). Frozen sections (40/zm) of each gonad type were
205
taken on a sliding microtome, mounted on slides, and stained using cresyl violet. For testes, the tissue was evaluated for the presence of spermatogonia, primary spermatocytes, elongated spermatocytes, and sperm using a five-point scale with '0' indicating an absence of the variable (prepubescent) and '4' indicating an abundance of the variable (reproductively mature). All ratings were conducted by two observers blind to the experimental assignment of the animals. The two sets of ratings for each of the four cell types were well correlated (range: R = 0.72, P < 0 . 0 1 to R = 0 . 9 0 , P < 0 . 0 0 1 ; mean R = 0 . 7 7 , P < 0.01), indicating satisfactory inter-rater reliability for these ratings; thus the values obtained by each rater on each measure were averaged for further analysis. For ovarian tissue, the sections were evaluated for the presence of atrial follicles and corpora lutea using a three-point scale with '0' indicating an absence of that variable and '2' indicating an abundance of the variable. Ratings were conducted by two observers blind to the experimental assignment of the animals. The two sets of ratings for both the atrial follicles (R = 0.46, P < 0.001) and corpora lutea (R = 0.66, P < 0.0001) were well correlated, making it permissible to average the ratings from each observer for further statistical analyses. Finally, the left humerus bone was fully exposed and its length (mm) was recorded. The humerus bone was selected for measurement as an index of skeletal growth, as previous evidence from this laboratory (H.P. Mclsaac, unpublished observations) has shown that humerus length varies independently of gonadal developmental status in a different rodent species (Siberian hamsters), thus providing a valid measurement of non-reproductively related growth. 2.3. Blood sampling and radioimmunoassays
Trunk blood, obtained at the time of sacrifice, was allowed to clot for 24 h at 4°C, centrifuged for 20 min at 4500 rpm to extract the serum, and then stored at - 5 0 ° C until determination of hormone levels. Samples were assayed for prolactin (PRL) and follicle-stimulating hormone (FSH) levels via double antibody radioimmunoassay. For PRL, anti, hamster PRL (1:15,000) donated by Dr. Borer was used as the primary antibody with goat anti-rabbit
P.J. Snyder, L.L. Badura / Epilepsy Research 20 (1995) 203-211
206 VPA ~ OVWT
CON ~ / 7 / ~ OVWT
VPA ~ UTWT
CON UTWT 140 130
E 1-
6
120
5
110 100 E 9o
80
4 3
>'< O
2 1
4 WEEKS
6 WEEKS AGE (WEEKS)
~ O ,'n
7O 60 50 40
~ m z w
30 20 10 0
~
separately for each sex using 2 × 2 (drug × age) between-subjects analysis of variance. Follow-up analyses were conducted with analysis of simple main effects and/or Fisher's LSD where appropriate. Ratings of gonadal status were analyzed with the Mann-Whitney U-test for independent samples, and inter-rater reliabilities (see above) were calculated with the Pearson product-moment correlation. All values were considered significant if P < 0.05.
3. Results
8 WEEKS
Fig. 1. Mean ( + SEM) ovarian ( O V W T ) and uterine ( U T W T ) weights (mg) for control and VPA-treated groups at each sampling point in the experiment. (* Significantly greater than the VPA-treated group within sampling time; see text for details.)
gamma globulin (GARGG; 1:12) used as second antibody. Purified hamster PRL obtained from Dr. Parlow and iodinated using the chloromine-T method served as trace, and serial dilutions of this hormone provided the standard. Curves derived for pooled sera from gonadally intact male and lactating female mice are parallel to the curve obtained with this standard (Badura, unpublished observations). Samples (35 /zl) were dispensed in 0.4% bovine albumin-0.1% gelatinized phosphate-buffered saline (GeI-PBS; pH 7.0) to a final volume of 300 /zl. The lower limit of detectability of the assay (90% bound) was 0.91 ng/ml, and the intra-assay coefficients of variance (CVs) at 50% and 65% bound were 11.24 and 10.3, respectively. FSH levels were determined using the NIAMMD rat FSH kit, courtesy of Dr. Raiti, using NIADDKFSH-RP-2 as the reference standard and iodinated NIADDK-rFSH-I6 as the trace. NIADDK-rFSH-S-11 (1:1500) served as primary antibody and GARGG (1:12) as the secondary antibody. Samples (20 /xl) were dispensed in Gel-PBS to a final volume of 300 /xl. The lower limit of detectability of the assay (90% bound) was 1.25 n g / m l and the CVs at 35% and 85% bound were 11.23 and 3.47, respectively.
3.1. Gonadal weights For females, chronic VPA administration led to a significant retardation in the development of uterine weight at the four-week ( F = 5.46(1,14), P = 0.03) and six-week ( F = 16.12(1,14), P = 0.001) sampiing periods (see Fig. 1). With regard to ovarian weights, although no significant differences were found at the four-week sampling period, a trend towards a significant difference was found at 6 weeks of age (4 weeks of VPA treatment), with the VPAtreated females showing reduced ovarian weights in comparison to controls ( F = 3.5(1,14), P = 0.08; see Fig. 1). By 8 weeks of age, the two groups did not differ across both measures of uterine and ovarian weight. For males, VPA administration significantly decreased testicular growth at the four-week ( F = VPA
Gonadal weights, body weights, circulating hormone levels, and humerus lengths were analyzed
CON
220 200 180 E ~" 160 ~140 120 100 8O O.9_ I.- 6O
~- 40 20 0 4 WEEKS
2.4. Statistical analyses
~
6 WEEKS
8 WEEKS
AGE (WEEKS)
Fig. 2. Mean (_+SEM) testicular weights (rag) for control and VPA-treated groups at each sampling point in the experiment. (* Significantly greater than the VPA-treated group within sampling time; see text for details.)
P.J. Snyder, L.L. Badura ~Epilepsy Research 20 (1995) 203-211 I
VPA
~
PRL
CON
~
PRL
VPA FSH
~
I
CON
12 10
VPA
~
PRL
FSH 16
6 -
14
5
207
CON
~
VPA
PRL
FSH
~
CON FSH 70
~
~
6o
12
_z ~<
8~
6
~3 <,
~4
6
30
u.
2
20 u"
4 2 8
4 WEEKS
6 WEEKS
0
8 WEEKS
0 4 WEEKS
AGE (WEEKS)
6 WEEKS
8 WEEKS
AGE (WEEKS)
Fig. 3. Mean ( + SEM)circulating levels of prolactin (PRL) and follicle-stimulating hormone (FSH) for control and VPA-treated females at each sampling point in the experiment.
5.98(1,14), P = 0.03), six-week ( F = 7.73(1,15), P =0.01), and eight-week ( F = 12.24(1,15), P = 0.003) sampling periods, although there were significant within-group increases over time (see Fig. 2).
3.2. Histological analyses For males, chronic VPA administration led to significant decreases in spermatogenesis (as evidenced by observed proliferation of sperm) at the four-week (U = 12.0, two-tailed P = 0.03), six-week (U = 17.0, two-tailed P = 0.045), and eight-week
Fig. 4. Mean (_+ SEM) circulating levels of prolactin (PRL) and follicle-stimulating hormone (FSH) for control and VPA-treated males at each sampling point in the experiment.
(U = 16.0, two-tailed P = 0.046) sampling periods (see Table 1). For ovarian tissue, VPA treatment was associated with a significant reduction at the six-week sampling time of both atrial follicular number (U = 6.5, twotailed P = 0.007) and corpora lutea number (U = 9.5, two-tailed P = 0.02). Significant differences were not found for either measure at the other sampling points, although a trend toward decreased number was observed for atrial follicles at 8 weeks (U = 13.5, two-tailed P = 0.09) (see Table 1).
Table 1 Mean ( + SEM) ratings for gonadal tissue for control and VPA-treatment group " Group
Sex
No. primary spermatocytes
No. elongated spermatocytes
No. sperm
Sex
No. atrial follicles
No. corpora lutea
Four week Controls VPA Tx.
M M
3.44 (0.12) 3.06 (0.19)
1.44 (0.28) 1.50 (0.34)
1.88 (0.30) 1.06 (0.16) b
F F
1.59 (0.17) 1.28 (0.19)
1.38 (0.21) 1.21 (0.25)
Six week Controls VPA Tx.
M M
3.72 (0.09) 3.63 (0.17)
3.28 (0.20) 2.94 (0.19)
3.11 (0.19) 2.63 (0.08) b
F F
1.91 (0.10) 1.19 (0.19) b
1.94 (0.11) 1.25 (0.25) b
Eight week Controls VPA Tx.
M M
3.88 (0.09) 3.78 (0.09)
3.75 (0.20) 3.67 (0.15)
3.69 (0.14) 2.83 (0.35) b
F F
1.75 (0.28) 1.34 (0.14) ¢
1.64 (0.23) 1.09 (0.28)
a Ratings based on ordinal scales for analyses of histological preparations of gonadal tissue, as described in Results Section. b Ratings for VPA-treated group are significantly decreased vs. those for control group, at P < 0.05. c Ratings for VPA-treated group exhibit trend toward decrease vs. those for control group, at P = 0.09.
208
P.J. Snyder, L.L. Badura /Epilepsy Research 20 (1995) 203-211 I
VPA MALE
~
VPA ~ FEMALE
CON MALE
. . . . .
CON FEMALE
3.4. Rate of skeletal growth: humerus bone length
14
12 E E ~10
m
6
w ~
4
i 2
4 WEEKS
6 WEEKS AGE (WEEKS)
8 WEEKS
For the females, humerus length was significantly smaller for VPA-treated animals at the four-week sampling time ( F = 11.30(1,14), P = 0 . 0 0 5 ) , approached a significant difference at the six-week sampling time ( F = 3.88(1,14), P = 0.07), and did not differ between groups at the eight-week sampling time. For the males, chronic VPA administration led to significantly decreased humerus bone lengths at the four-week ( F = 15.75(1,14), P = 0 . 0 0 1 ) , sixweek ( F = 5.62(1,15), P = 0.03), and eight-week ( F = 8.82(1,15), P = 0.01) sampling periods (see Fig. 5).
Fig. 5. Mean (_ SEM) humerus lengths (mm) for both male and female control (CON) and VPA-treated mice at each sampling point in the experiment. (* Significantly greater lengths than VPA-treated group within samplingtime, see text for details.)
3.5. Total body weight
3.3. Radioimmunoassay of PRL and FSH release
Body weights, for both males and females, were not significantly affected by chronic VPA administration at any time point (see Fig. 6).
VPA administration did not have a significant effect upon circulating FSH or PRL levels in females, although there was a trend toward depression of FSH at the six-week sampling time ( F = 3.91(1,14), P = 0.06; see Fig. 3). VPA administration did not have a significant effect upon circulating FSH or PRL levels in males at any sampling time (see Fig. 4).
I
VPA MALE
~
CON MALE
~
VPA ~ FEMALE
CON FEMALE
30
L9 I.U >-
~DIO
@ 4 WEEKS
6 WEEKS AGE (WEEKS)
8 WEEKS
Fig. 6. Mean ( + SEM) body weights(g) for both male and female control (CON) and VPA-treated mice at each sampling point in the experiment.
4. Discussion This study was designed to investigate the neuroendocrinological bases for the role that chronic administration of VPA might play in attenuating pubertal growth in children. An animal model for this research was selected, as it afforded the ability to study the effects of VPA on testicular and ovarian maturation. Our findings indicate that, in female mice, chronic VPA administration leads to significant retardation in the development of uterine weight early in development, but by 8 weeks of age, the VPA-treated groups did not differ from controls on measures of uterine and ovarian weights. These findings were mirrored by the results obtained for rating of the ovarian histology, with the greatest decrease in atrial follicular number and corpora lutea number seen at 6 weeks of age. Similar trends were observed for the effects of VPA treatment on humerus length, with females showing significant differences early in development, and no noticeable differences at 8 weeks. However, VPA significantly decreased the rate of testicular and skeletal growth, in males, at all three time points. These combined results (for both males
P,I. Snyder, L.L. Badura // Epilepsy Research 20 (1995) 203-211
and females) indicate that growth was slowed, but not prevented, in the VPA-treated animals. Interestingly, more consistent differences with regard to gonadal weight and skeletal growth were found in males rather than in females. This observation may be clinically relevant, as any effects of VPA on delayed pubertal maturation in young boys may be less readily apparent than in girls. Amenorrhea and slowed development of secondary sex characteristics in females are arguably more noticeable than is delayed testicular development in boys. Although significant differences in circulating FSH and PRL were not observed among groups in the current study, the data for FSH do suggest a delayed 'surge' in circulating levels in males treated with VPA. That is, FSH appeared to peak in control animals at 4 weeks of age, where~s this peak seemed to occur in VPA-treated animals at 6 weeks of age (see Fig. 4). This 'surge' of FSH is considered to be a marker of pubertal onset in many species. Regardless of potential sex differences in the slowing of physical growth and maturation by VPA's action on the hypothalamic-pituitary-gonadal axis, our results in males show that VPA does not prevent or halt pubertal onset, but instead may act to slow the rate of normal maturational changes. This study does not address the idiosyncratic, very rapid rate of murine metabolism of VPA [30], and nor does it address the inconsistent rates of delivery of the drug via drinking water in animals who are nocturnal, and thus more likely to drink during the lights-off phase. Our method of delivery likely resulted in a palLtern of blood plasma levels of VPA which mimics a 'TID' (thrice daily) or 'QID' (four times daily) drug administration pattern in humans, with a long break overnight [31]. It is not like the more constant blood plasma levels of VPA achieved by q6 hours ,or q8 hours dosing utilized for institutionalized patients or those with severe childhood epilepsy. Thus, the fact that we found significant differences betwe,en our VPA-treated and control groups, across measures of gonadal development and skeletal maturation, attests to the robust nature of the effect of VPA on the hypothalamic-hypophyseal axis--that is, despite a non-constant-rate method of drug delivery. However, our inability to detect group differences in PRL and FSH levels might be directly related to our chosen method of
209
drug delivery. Future efforts, relying on an osmotic pump delivery (subcutaneous implantation), leading to a constant rate of application [31], should re-examine group differences in PRL and FSH secretion, as well as evaluate potential effects on circulating GH. Although estrous cyclicity was not directly monitored in this study, it is unlikely that proestrous-related variations in PRL and FSH contributed to the experimental variability as these samples were obtained in the early morning hours, and surges of these anterior pituitary hormones in rodents do not occur until late afternoon [3]. Further studies are in progress that will include the systematic investigation of the side effects of VPA at different dose levels, with the hope that such work might lead to refinements in therapeutic dosing strategies and the identification of potential warning signals of endocrinological dysfunction during treatment. Information regarding the neurochemical basis of VPA action within the hypothalamo-hypophyseal system might lead to the development of agents which block the negative actions of VPA on hormone release without diminishing its anticonvulsant properties. One possibility may be a disruption of the cholinergic system that regulates GH secretion during nocturnal sleep. Legido et al. [22] have provided evidence to support this possibility in a study of 35 epileptic children (25 of whom were placed on VPA for at least 6 months) and 26 neurologically normal children. They found that neither carbamazepine nor VPA affect the somatostatinergic, GABAergic, or dopaminergic systems that regulate PRL, TSH, and gonadotropin secretion. However, the chronic administration of either of the two AEDs was found to significantly reduce GH release during nocturnal sleep. As described above, chronic VPA administration may lead to global decreases in pituitary hormone secretion, with abnormal growth and development in some human patients [8,10,24,26], and yet it is unclear where along the hypothalamic-pituitary axis VPA might act to diminish pituitary activity. The hypothalamus contains a multitude of peptidergic stimulating and inhibiting systems which influence the release of a variety of protein hormones by the pituitary, and each system possesses unique anatomical substrates and morphological characteristics (see [32,33] for reviews). In addition, these hypothalamic
210
P.J. Snyder, L.L. Badura / Epilepsy Research 20 (1995) 203-211
neuroendocrine systems are further modulated by a plexus of neuronal inputs, both of peptidergic and classical neurotransmitter origin (see [36] for a review). Our understanding of this multi-integrated hypothalamo-hypophyseal axis is still incomplete. For example, it is widely accepted that dopamine of tuberoinfundibular origin represents the major inhibitory influence upon PRL secretion from the anterior pituitary (see [4,29] for reviews); however, the hypothalamic releasing systems which comprise important stimulatory inputs for this system are still subject to debate (see [15,35] for reviews). Further work will also center on elucidating the sites and mechanisms of VPA action along the hypothalamic-pituitary axis. Preliminary data from this laboratory have indicated that anterior pituitary tissue harvested from animals that have been chronically treated with VPA shows alterations in basal FSH release in a perfusion culture system that parallel the trends in the group differences reported above for circulating levels (Badura and Snyder, unpublished data, 1994). Thus, although many of the endocrinological analyses in the present study yielded non-significant differences in the predicted direction, the pilot data obtained from the tissue culture study provide further support for effects of VPA on the hypothalamic-pituitary axis, possibly directly on the pituitary itself. We recognize that our comparisons of the endocrinological data in the present report suffered from low statistical power. Sample sizes will be adjusted accordingly, in future studies, to provide increased statistical power for the analyses of variables that are prone to substantial individual differences in pubertal animals. Acknowledgements The authors wish to express their appreciation to Dr. Victor Denenberg for providing the mice and colony space for this study, and to Aaron Wilson for technical support. Also, we are grateful to Dr. Borer and Dr. Parlow for their donations of reagents for use in the PRL RIA, and Dr. Riati for reagents used in the FSH RIA. This project was completed while P.J.S. was a Wilder Penfield Research Fellow of the Epilepsy Foundation of America, and it was funded in part by NIH institutional grant NS07324 and NSF grant IBN9224804 funds to L.L.B.
References [1] Axnaout, M.A., Garthwaite, T.L., Martinson, D.R. and Hagen, T.C., Vasoactive intestinal polypeptide is synthesized in anterior pituitary tissue, Endocrinology, 119 (1986) 20522057. [2] Badura, L.L. and Goldman, B.D., Prolactin-dependent seasonal changes in pelage: Role of the pineal gland and dopamine, J. Exp. Zool., 261 (1992) 27-33. [3] Badura, L.L., Sisk, C.L. and Nunez, A.A., Effects of photoperiod and hypothalamic knife cuts on the timing of FSH surges in hamsters, Brain Res. Bull., 26 (1991) 313-316. [4] Ben-Jonathan, N., Arbogast, L.A. and Hyde, J.F., Neuroendocrine regulation of prolactin release, Prog. Neurobiol., 33 (1989) 399-447. [5] Bruni, J. and Wilder, B.J., Valproic acid: Review of a new antiepileptic drug, Arch Neurol., 36 (1979) 393-398. [6] Chapman, A , Keane, P.E, Meldrum, B.S., Simiand, L and Vernieres, J.C, Mechanism of anticonvulsant action of valproate, Prog. Neurobiol., 19 (1982) 315-359. [7] Conran, M.LC., Kearney, PJ., Callaghen, M.N., Murphy, D. and Goggin, T., Hypothalamic pituitary function testing on children receiving carbamazepine or sodium valproate, Epilepsia, 26 (1985) 585-588. [8] Cook, J.S., Bale Jr., J.F. and Hoffman, R.P. Pubertal arrest associated with valproic acid therapy, Pediatr. Neurol., 8 (1992) 229-231. [9] Crawford, P., Belchetz, P. and Chadwick, D., Growth hormone response of diazepam, clonidine, and glucagon in patients with epilepsy, Epilepsy Res., 3 (1989) 63-69. [10] Egger, J. and Brett, E.M., Effects of sodium valproate in 100 children with special reference to weight, Br. Med. J., 283 (1981) 577-581. [11] Faingold, C.L. and Browning, R.A., Mechanisms of anticonvulsant drug action, Eur. J. Pediatr., 146 (1987) 8-14. [12] Frohman, LA., The anterior pituitary. In: J.B. Wyngarden, L.H. Smith Jr. and J.C. Bennett (Eds.), Textbook of Medicine (19th Edition), Vol. 2, W.B. Saunders, Philadelphia, 1992, chapter 213, pp. 1228-1231. [13] Godin, Y., Heiner, L., Mark, J. and Mandel, P., Effects of di-n-propylacetate, an anticonvulsive compound, on GABA metabolism, J. Neurochem., 16 (1969) 869-873. [14] Hackman, J.C., Grayson, V. and Davidoff, R.A, The presynaptic effects of valproic acid in the isolated frog spinal cord, Brain Res., 220 (1981) 269-285. [15] Hyde, J.F. and Ben-Jonathan, N., Characterization of prolactin-releasing factor in the rat posterior pituitary, Endocrinology, 122 (1988) 2533-2539. [16] Invitti, C., Danesi, L., Dubini, A. and Cavagnini, F., Neuroendocrine effects of chronic administration of sodium valproate in epileptic patients, Acta Endocrinol., 118 (1986) 381-388. [17] lsojarvi, J.I.T, Laatikainen, T.J., Pakarinen, A.J., Juntunen, K.T.S. and Myllyla, V.V., Polycystic ovaries and hyperandrogenism in women taking valproate for epilepsy, N. EngL J. Med., 329 (1993) 1383-1388. [18] Johnston, D., Valproic acid: Update on its mechanisms of action, Epilepsia, 25, Suppl. 1 (1984) S1-$4.
P.J. Snyder, L.L. Badura / Epilepsy Research 20 (1995) 203-211 [19] Kemp, L.I., Sodium valproate as an antidepressant, Br. J. Psychiatry, 160 (199:2) 121-123. [20] Kerwin, R.W., Olpe, H.R. and Schmutz, M., The effect of sodium-n-dipropyl acetate on GABA-dependent inhibition in the rat cortex and substantia nigra in relation to its anticonvulsant activity, Br. J. Pharmacol., 71 (1980) 545-551. [21] Kritzler, R., Vining, E. and Plotnick, L., Sodium valproate and corticotropin suppression in the child treated for seizures, J. Pediatr., 102 (19813) 142-143. [22] Legido, A., Lopez, M.J., Garagorri, J., Orden, I., Baselga, C., Bueno, M. and Grover, W.D., Hypothalamo-pituitary function in epileptic children treated with carbamazepine or valproate, Epilepsia, 33, Suppl. 3 (1992) 9. [23] Legido, A., Lopez, ~'.J., Garagorri, J., Orden, I., Esteva, F., Baselga, C., Bueno, M. and Grover, W.D., Growth and growth hormone secretion in epileptic children treated with valproate, Epilepsia, 33, Suppl. 3 (1992) 9. [24] Lundberg, B., Nergardh, A., Ritzen, E. and Samuelson, K., Influence of valproic acid on the gonadotropin releasing hormone test in puberty, Acta Pediatr. Scand., 75 (1986) 787-792. [25] MacDonald, R.L. and Bergey, G.K., Valproic acid augments GABA mediated postsynaptic inhibition in cultured mammalian neurons, Brain Res., 170 (1979) 558-562. [26] Margraf, J.W. and Dreifuss, F.E., Amenorrhea following intiation of therapy with valproic acid, Neurology, 31 (1981) 159. [27] McElroy, S.L., Keck, P.E. Jr., Pope, H.G., Jr. and Hudson, J.l., Valproate in the treatment of bipolar disorder: Literature review and clinical guidelines, J. Clin. Psychopharmacol., 12, Suppl. 1 (1992) S,~.2-$52. [28] Meunier, H., Carraz, G., Meunier, Y., Eymard, P. and Aimard, M., Proprietes pharmacodynamique de l'acide n-dipropylacetique, Therapie, 18 (1963) 435-438.
211
[29] Moore, K.E. and Demarest, K.T., Tuberoinfundibular and tuberohypophyseal dopaminergic neurons. In: W.F. Ganong and L Matrini (Eds.), Frontiers in Neuroendocrinology, Raven Press, New York, 1982, pp. 161-190. [30] Nau, H. and L/Sscher, W., Valproic acid and active unsaturated metabolite (2-en): Transfer to mouse liver following human therapeutic doses, Biopharm. Drug Dispos., 6 (1985) 1-8. [31] Nau, H. and Zierer, R., Pharmacokinetics of valproic acid and metabolites in mouse plasma and brain following constant-rate application of the drug and its unsaturated metabolite with an osmotic delivery system, Biopharm. Drug Dispos., 3 (1982) 317-328. [32] Page, R.B., The anatomy of the hypothalamo-hypophyseal complex. In: E. Knobil and J. Neill (Eds.), The Physiology of Reproduction, Raven Press, New York, 1988, pp. 1161-1233. [33] Sagar, S.M. and Martin, J.B., Hypothalamohypophysiotropic peptide systems. In: V.B. Mountcastle, F.E. Bloom and S.E. Geiger (Eds.), Handbook of Physiology, Waverly Press, Baltimore, 1986, pp. 413-462. [34] Schmidt, D., Adverse effects of valproate, Epilepsia, 25 (1984) $44-$49. [35] Shin, S.H., Papas, S. and Obansawin, M.C., Current status of the rat prolactin-releasing factor, Can. J. PhysioL Pharmacol., 65 (1987) 2036-2043. [36] Swanson, L.W., Organization of mammalian neuroendocrine system. In: V.B. Mountcastle, F.E. Bloom and S.E. Geiger (Eds.), Handbook of Physiology, Waverly Press, Baltimore, 1986, pp. 317-363. [37] Wilder, B.J., Pharmacokinetics of valproate and carbamezapine, J. Clin. Psychopharm., 12 (1992) $64-$68.