|Home | About | Journals | Submit | Contact Us | Français|
Seasonally breeding animals use a combination of photic (i.e., day length) and non-photic (e.g., food availability, temperature) cues to regulate their reproduction. How these environmental cues are integrated is not understood. To assess the potential role of two candidate neuropeptides, kisspeptin and RFamide-related peptide-3 (RFRP), we monitored regional changes in their gene expression in a seasonally breeding mammal exposed to moderate changes in photoperiod and food availability. Adult male Siberian hamsters (Phodopus sungorus) were housed in a long (16 h light/day; 16L) or intermediate (13.5L) photoperiod and fed ad libitum or a progressive food restriction schedule (FR; reduced to 80% of ad libitum) for 11 weeks. Gonadal regression occurred only in FR hamsters housed in 13.5L. Immunohistochemistry was used to identify diencephalic populations of kisspeptin- and RFRP-immunoreactive cells, and quantitative PCR was used to measure gene expression in adjacent coronal brain sections. Photoperiod but not food availability altered RFRP mRNA expression in the dorsomedial sections, whereas food availability but not photoperiod altered Kiss1 expression in the arcuate sections; intermediate photoperiods elevated RFRP expression, and food restriction suppressed Kiss1 expression. Regional- and neuropeptide-specific activity of RFamides may provide a mechanism for integration of multi-modal environmental information in the seasonal control of reproduction.
In environments where robust seasonal cycles of temperature and food availability prevail, seasonal cues regulate the phase of the geophysical cycle during which critical events in mammalian reproduction occur (puberty, ovulation, breeding, parturition, weaning) (Paul et al., 2008, Prendergast et al., 2009). In Siberian hamsters (Phodopus sungorus), as in other small photoperiodic rodents, summer photoperiods ≥14 h light/day (14L) stimulate, and winter photoperiods <12L inhibit, testicular development, gametogenesis, ovarian function, and reproductive behavior (Hoffmann, 1982, Park et al., 2004, Schlatt et al., 1995). The light-dark cycle entrains a circadian rhythm of nocturnal pineal melatonin secretion (Illnerová, 1991), the duration of which acts on pituitary, thalamic, and hypothalamic melatonin-sensitive target tissues to convey photoperiodic information to the reproductive axis (Badura & Goldman, 1992, Bartness et al., 1993, Carter & Goldman, 1983a, b, Glass & Lynch, 1982, Morgan & Hazlerigg, 2008).
Recent work has suggested that photoperiodic regulation of two hypothalamic RFamides (Arg-Phe-NH2), kisspeptin and RFamide-related peptide-3 (RFRP), may figure prominently in the transduction of melatonin signals to the HPG axis. Kisspeptin stimulates (Gottsch et al., 2004, Greives et al., 2007, Irwig et al., 2004, Messager et al., 2005), and RFRP inhibits (Kriegsfeld et al., 2006) LH secretion in mammals. In photoperiodic rodents, short photoperiods that trigger gonadal regression yield decreases in Kiss1 and RFRP mRNA and their respective protein expression in the anteroventral periventricular nucleus (AVPV; for kisspeptin) and mediobasal hypothalamus including the dorsomedial hypothalamus (DMH; for RFRP) (Greives et al., 2007, Mason et al., 2007, Revel et al., 2008, Revel et al., 2006). In Siberian hamsters, kisspeptin peptide and mRNA expression increase in the arcuate nucleus (ARC) after transfer to short photoperiods (Greives et al., 2007, Mason et al., 2007, Simonneaux et al., 2009); in Syrian hamsters, however, transfer to short photoperiods decreases kisspeptin expression in this nucleus (Revel et al., 2006). Kisspeptin and RFRP may act in concert to stimulate or inhibit the HPG axis during photoperiod-induced reproductive transitions (Kauffman et al., 2007, Smith et al., 2008).
Non-photic environmental cues, such as food availability, ambient temperature, and social interactions, also contribute to seasonal timekeeping, but their underlying neurobiology is not understood (reviewed in Ball, 1993, Paul et al., 2008). Recently, a model was developed whereby non-photic seasonal regulation can be studied under static photoperiods. Siberian hamsters housed in an intermediate day length (13.5 h light/day; 13.5L) from birth undergo gonadal growth. After this gonadal development occurs, mild food restriction or increased population density can trigger substantial reproductive inhibition in hamsters housed in 13.5L, but these non-photic cues are without effect on hamsters housed in a long photoperiod (16L; Paul et al., 2009). This model permits investigation into candidate neural substrates responsible for the integration of photoperiodic and non-photic cues for the reproductive axis (e.g., kisspeptin and RFRP). For example, if kisspeptin is one such site of environmental cue integration, then intermediate photoperiods and mild food restriction, neither of which elicits gonadal responses alone, should each induce partial changes in Kiss1 mRNA (e.g., moderate decreases in Kiss1 mRNA in the AVPV); when 13.5L and mild food restriction are presented in combination— and reproductive responses occur— larger Kiss1 mRNA responses would be expected. Alternatively, photoperiod and food restriction may be integrated via distinct kisspeptin neuronal populations: transfer to intermediate photoperiods may decrease Kiss1 mRNA in the AVPV and mild food restriction may decrease Kiss1 mRNA in the ARC. Similar predictions could be made for RFRP or for coordinated changes in both kisspeptin and RFRP signals among the AVPV, ARC, and DMH. On the other hand, if the integration of photoperiod and food restriction occurs upstream from the RFamides, then mRNA patterns in each nucleus should simply mirror gonadal responses: altered RFamide mRNA expression should only occur in hamsters challenged with both intermediate photoperiods and food restriction.
Here we describe an experiment that examined changes in Kiss1 and RFRP gene expression in response to modest changes in photoperiod and food availability, which combined, were sufficient to induce gonadal regression. The results suggest that these two RFamides respond to photic and non-photic cues in a peptide- and regionally-specific manner and may work in concert to integrate seasonal cues via distinct neuronal pathways.
Adult (>4 months) male Siberian hamsters (Phodopus sungorus) were obtained from 13.5:10.5 h light:dark cycle (13.5L) or 16L (lights off at 1230 h CST for both photoperiods) breeding colonies maintained at the University of Chicago and housed singly from weaning in polypropylene cages (28 × 17 × 12 cm) with wood shavings (Harlan Sani-Chips, Harlan Inc., Indianapolis, IN, USA) under their same natal photoperiods. Ambient temperature of the experimental rooms was 20 ± 0.5°C and relative humidity was maintained at 53 ± 2%. Food (Teklad Rodent Diet 8604, Harlan Inc.) and filtered tap water were provided ad libitum, except where otherwise indicated. The data reported here are derived from a subset of 23 animals in a recent study that described the efficacy of intermediate day lengths in unmasking morphological reproductive and hormonal responses to food restriction (Paul et al., 2009). All procedures conformed to the USDA Guidelines for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Chicago.
The length and width of the left testis were measured (±0.1 mm) through the abdominal skin using calipers while hamsters were under light isoflurane anesthesia. The size of the testis was estimated as the product of the length × width2 (estimated testis volume; ETV), which is correlated with testis weight, circulating testosterone, and spermatogenesis (Gorman, 1995, Schlatt et al., 1995). Body masses (±0.1 g) were also recorded at the time of reproductive measurements.
Food restriction (FR) was accomplished by progressive reductions from mass-specific ad libitum intake (0.13 g food / g body mass; determined in a pilot study; data not shown) administered to each hamster in a single daily ration shortly after the onset of darkness (16L-FR: n=6; 13.5L-FR: n=6). From weeks 0 through 6, FR hamsters were initially provided with 90% of ad libitum daily intake (0.117 g/g), and thereafter (weeks 6 though 11) received 80% of ad libitum intake (0.104 g/g). This pattern of food restriction was designed to more closely simulate the progressive late-summer decrease in food availability that occurs in this species’ natural environment (Weiner, 1987). Control hamsters had free access to food throughout the study (AdLib; 16L-AdLib: n=5; 13.5L-AdLib: n=6).
Given the opposite photoperiodic responses of kisspeptin peptide and mRNA expression in the AVPV and ARC of Siberian hamsters (Greives et al., 2007, Mason et al., 2007, Simonneaux et al., 2009), measures of whole hypothalamic Kiss1 expression is problematic. To achieve both quantitative and regionally specific RFamide data, a method combining immunohistochemistry (IHC) and quantitative PCR (qPCR) was used. One of two alternating sets of 30 μm coronal sections, rostral to caudal, was processed using IHC for RFamide immunoreactivity (RFamide-ir) in the anteroventral periventricular nucleus (AVPV), the arcuate nucleus (ARC), and the dorsomedial nucleus of the hypothalamus (DMH). The remaining set of sections was reserved for qPCR. Total RNA was extracted from the fixed sections adjacent to those that exhibited immunoreactivity in the AVPV, ARC, and DMH (3 sections per brain region). Sections were pooled within each brain region to generate 3 tissue samples for each hamster, 1 containing the AVPV, 1 containing the ARC, and 1 containing the DMH. Immunopositive perikarya were restricted to the AVPV, ARC, and DMH nuclei (Fig. 1A-C; cf. Gottsch et al., 2004). We further quantified Kiss1 and RFRP mRNA levels within the hypothalamus, thalamus, and cortex of a separate set of hamsters to validate claims that the changes in gene expression were in fact of hypothalamic origin.
FR and AdLib treatments continued through the day of sacrifice. At the midpoint of the light phase, hamsters were deeply anesthetized with an overdose of sodium pentobarbital (15 mg/animal, i.p.) and were perfused transcardially with 40 ml of 0.9% saline, followed by 40 ml of 4% paraformaldehyde in phosphate buffered saline (pH 7.4). Brains were postfixed for 2 days in fixative before cryoprotection with a mixture of buffered fixative and 30% sucrose for 2–3 days, and freezing at −80°C until sectioned.
Brains were sliced coronally at 30 μm on a freezing sliding microtome and free-floating sections were stored in cryoprotectant (Watson et al., 1986) at −20°C until processed for RFamide-ir. For each animal, alternating sections, rostral to caudal, were stained for using standard ABC immunocytochemistry, following the methods of Kramer et al. (2006). RFamide-ir cells were labeled using a rabbit anti-human kisspeptin antiserum diluted at 1:5000 (T-4771; Peninsula Laboratories Inc, Bachem, San Carlos, CA) raised against the following amino acids Tyr-Asn-Trp-Asn-Ser-Phe-Gly-Leu-Arg-Phe-NH2, corresponding to amino acids 4-13. Sections were examined under bright field illumination on a Nikon Eclipse 80i microscope.
In the Siberian hamster hypothalamus, the T-4771 antibody cross-reacts with kisspeptin and RFRP, detecting cells in the AVPV, the ARC, and the DMH. Pre-adsorption procedures have indicated that the immunoreactive cells in the AVPV and ARC express kisspeptin, whereas those in the DMH express RFRP (Greives et al., 2007). Nonetheless, concerns remain regarding the specificity of this antibody (Goodman et al., 2007, Mikkelsen & Simonneaux, 2009), and recent studies have indicated that a diffuse population of kisspeptin cells may also exist in the DMH, at least in the laboratory mouse (Clarkson et al., 2009). Given these concerns and the fact that IHC is only semi-quantitative by nature, we restricted our use of this staining for localization of RFamide expression within the target nuclei, allowing selection of sections containing the AVPV, ARC, or DMH for qPCR.
Extractions were performed according the manufacturer’s protocol (RNeasy FFPE kit, Qiagen, Valencia, CA, USA) with one exception: paraffin removal steps were skipped. Extracted RNA was suspended in 30μl RNase-free water and RNA concentration and quality were determined by spectrophotometer. All RNA samples were stored at −70° C until further analysis. cDNA was created via reverse transcription of 2μg of RNA from each sample with MMLV Reverse Transcriptase enzyme (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s protocol.
To design primers and a probe for quantitative PCR with high specificity for this species, a portion of each gene of interest was sequenced. To sequence portions of these genes, semi-quantitative PCR was conducted on 1μl of pooled Siberian hamster hypothalamic cDNA with Taq DNA Polymerase enzyme (Invitrogen) according to the manufacturer’s protocol in a thermocycler for 40 cycles (Bio-Rad). Degenerate primers were designed based on conserved regions among multiple species with known gene sequences (GenBank) using PrimerExpress software (Applied Biosystems, Foster City, CA, USA). PCR gene product amplification was visualized on 2% TAE-agarose gels containing ethidium bromide using a CCD camera. To verify amplification of the correct gene, PCR products were purified (Centricon-100, Millipore, Billerica, MA, USA) and directly sequenced at the University of Chicago Cancer Research Center DNA Sequencing Facility. The resulting amplicon sequences for Siberian hamster Kiss1 and RFRP were >90% homologous to published sequences for Mus musculus Kiss1 and RFRP. Sequencing information was entered in the GENBANK database: RFRP (Accession # EU365871) and Kiss1 (Accession # EU365872).
After confirmation of gene products, primers and probes for quantitative PCR were designed using PrimerExpress. Primers and probes were synthesized as follows, with probes labeled with 6-FAM and MGB (non-fluorescent quencher) at the 5′ and 3′ ends, respectively: RFRP forward 5′-GCCCCTGCCAACAAAGTG-3′, RFRP reverse 5′-CAGGGTCCTCCCAAATCTCA-3′, RFRP probe 5′-CCCACTCAGCAGCCA-3′; Kiss1 forward 5′-AACTCATCAATGCCTGGGAAA-3′, Kiss1 reverse 5′-GCTCGCAGTCCTCCAGGTT-3′, Kiss1 probe 5′-CGGTGCGCAGAGAG-3′. A TaqMan 18S Ribosomal RNA primer and probe set (labeled with VIC; Applied Biosystems) was used as the control gene for relative quantification. Amplification was performed on an ABI 7900HT Sequencing Detection System by using Taqman® Universal PCR Master Mix. The universal two-step RT-PCR cycling conditions used were: 50° C for 2 min, 95° C for 10 min, followed by 40 cycles of 95° C for 15 sec and 60° C for 1 min. Relative gene expression of individual samples run in duplicate was calculated by comparison to relative standard curve consisting of serial dilutions of pooled P. sungorus hypothalamic cDNA (1:1, 1:10, 1:100, 1:1000, 1:10,000) followed by normalization to 18S rRNA gene expression. RNA quality for each sample was assessed via 260/280 ratio using a spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA) before reverse transcription. All samples had acceptable ratios between 1.8 and 2.0 (mean: 1.98; range: 1.81-2.0).
Hamsters (n=3) fed ad libitum were perfused in a manner identical to that described above for experimental animals. Brains were removed, post-fixed, frozen, and sectioned. Free-floating sections from each of three rostro-caudal brain regions (AVPV: bregma −0.3 mm; DMH: bregma −2.5 mm; ARC: bregma −3.3 mm) (Paxinos & Watson, 1998) were then dissected into cortex (including limbic structures), hypothalamus, and thalamus under a 4X dissecting microscope at 4°C (Fig. 1D-F). Between 10-12 microdissections from each brain region were pooled and expression levels of Kiss1 and RFRP were determined using qPCR as described above.
Data are expressed as means ± SEM for each group. Repeated measures and factorial ANOVAs were used to measure variation within and between experimental groups, and differences between means were assessed by least significant difference tests and t-tests, where warranted by a significant F statistic. Significance level was set at P≤0.05 for all tests. Analyses were performed using Statview 5.0.1 for the PC (SAS Institute Inc., Cary, NC, USA).
FR treatments inhibited body mass gains in both photoperiods (F=9.9; P=0.005; Fig. 2A); the magnitude of body mass inhibition tended to be greater in FR hamsters housed in 13.5L relative to those in 16L but fell short of significance (P<0.10). FR treatments and photoperiod interacted to affect testis size (F=24.4; P<0.0001); as in the parent population (Paul et al., 2009), significant gonadal regression only occurred in 13.5L hamsters (P<0.0001; Fig. 2B,C).
qPCR analysis of microdissections permitted evaluation of cortical, thalamic, and hypothalamic contributions to Kiss1 and RFRP gene expression obtained from whole coronal sections at each of the 3 rostro-caudal levels investigated (Kiss1: AVPV and ARC; RFRP: DMH; Fig. 3). At the level of AVPV, most of the Kiss1 signal (83%) was derived from hypothalamic cells; the remainder originated from cortex and thalamus. At the level of the ARC, the hypothalamus was the largest source of Kiss1 gene expression (45%), but a substantial Kiss1 signal (37%) was detected in cortex. At the level of the DMH, over 99% of the RFRP signal was hypothalamic in origin (Fig. 3).
In sections containing the ARC, Kiss1 gene expression was significantly decreased in FR relative to AdLib hamsters (F=7.9; P=0.01; Fig. 4A) but was not affected by photoperiod (F=1.1; P>0.3); the interaction between these factors was not significant (F=0.2; P>0.8). In 16L hamsters, FR significantly inhibited Kiss1 expression at the level of the ARC (P<0.05); decreases in 13.5L FR hamsters fell just short of significance (P<0.06). At the level of the AVPV, neither photoperiod (F=0.2; P>0.6), nor FR (F=0.9; P>0.3) significantly affected Kiss1 gene expression (Fig. 4B); the interaction was also not significant (F=0.2; P>0.6).
In sections containing the DMH, RFRP expression was significantly greater in 13.5L relative to 16L hamsters (F=9.1; P<0.01), independent of food manipulations (F<0.1; P>0.9; Fig. 4C); again, there was no significant interaction between photoperiod and feeding (F<0.1; P>0.9). In both AdLib and FR hamsters, RFRP expression was ~40-fold higher in 13.5L relative to 16L hamsters (Fisher’s PLSD; P<0.05, both comparisons).
Intermediate day lengths render the reproductive axis of Siberian hamsters more responsive to non-photic, exteroceptive cues such that testicular regression occurs only in hamsters exposed to both food restriction and intermediate day lengths (Paul et al., 2009). The present work extended this finding to investigate neuroendocrine mechanisms of photic and non-photic cue integration.
At the level of the ARC, food restriction induced a ~5-fold decrease in Kiss1 expression but no main effect of photoperiod was detected. Thus, modest reductions in food availability decrease Kiss1 mRNA at the level of the ARC, either by downregulating Kiss1 transcription or by increasing post-transcriptional processing (e.g., translation and degradation). These data are consistent with reports in other non-photoperiodic species (Castellano et al., 2005, Luque et al., 2007, Smith et al., 2006). Given the well-established role of the ARC in regulating energy balance (Hill et al., 2008, Morgan et al., 2006), ARC kisspeptin neurons are ideally situated to monitor the metabolic state of the organism. In sections containing the AVPV, neither photoperiod nor food restriction affected Kiss1 mRNA expression. At the level of the DMH, food restriction did not alter RFRP gene expression; however, exposure to intermediate photoperiods increased RFRP mRNA in the DMH ~40-fold.
The current data provide novel insights into environmental influences on the RFamide system. Non-photic (reduced food availability) and photic (intermediate day lengths) stimuli inadequate to impact reproductive physiology, nonetheless alter Kiss1 and RFRP gene expression, respectively. Remarkably, in neither the ARC nor the DMH were any additive effects of food restriction or photoperiod evident. Gonadal responses were only associated with changes in both ARC Kiss1 and DMH RFRP. Such a pattern of responses is consistent with a role for the RFamide system as an integrator of multimodal seasonal cues (Kriegsfeld, 2006). Cue integration is a complex process and undoubtedly involves other neuropeptides as well. Nonetheless, a simple model in which environmental cues act on distinct GnRH input pathways, including kisspeptin and RFRP, is compatible with the present data.
Microdissections (cortex, thalamus, hypothalamus) of coronal brain sections indicated that at the level of the AVPV and the DMH the overwhelming majority of the Kiss1 (>80%) and RFRP (>99%) mRNA, respectively, originated in the hypothalamus. Given the highly localized patterns of kisspeptin-ir and RFRP-ir in adjacent sections, we are confident that the vast majority of the Kiss1 and RFRP mRNA obtained from whole coronal sections originated in the AVPV and DMH, and that the differences in mRNA expression between treatment groups reflect differences within these nuclei. Some caution is warranted, however, in interpreting data from the coronal sections at the rostro-caudal level of the ARC: although the strongest Kiss1 signal originated in the hypothalamus (45%), a substantial amount of Kiss1 mRNA was also expressed in cortex. Limbic structures (medial amygdala and hippocampus) are the likely source of this cortical Kiss1 (37% of total; cf. Arai, 2009, Gottsch et al., 2004). Although the efferent targets of these populations of kisspeptin neurons have not been identified, numerous amygdalo-hypothalamic projections exist that govern reproductive responses (e.g., pheromone-processing circuits; Baum, 2009). Amygdala Kiss1 may also participate in the regulation of gonadotrophin production by photoperiod or energetic cues. Future studies may attempt more precise neuroanatomical localization via in situ hybridization; however, in common with immunocytochemistry, such an approach is only semiquantitative. The approach we report here describes novel quantitative measurement of hypothalamic gene expression, in conjunction with neuroanatomical localization on adjacent sections, which may prove useful in future studies of brain gene expression.
A recent report indicated decreased RFRP mRNA in the mediobasal hypothalamus including the DMH following 10 weeks of exposure to a short photoperiod (Revel et al., 2008), but several methodological differences, most notably the use of intermediate photoperiods in the present report (cf. categorically short photoperiods in Revel et al., 2008), preclude direct reconciliation of these two studies. However, dynamic changes in the magnitude of RFRP restraint on GnRH neurons (Gibson et al., 2008, Kriegsfeld et al., 2006) may occur over time during photoperiod-induced gonadal regression. For example, an increase in RFRP expression may be required to initiate gonadal regression, but less critical for the maintenance of gonadal involution once regression is completed. Such a mechanism would be analogous to the dynamic stimulatory neuroendocrine events that evolve during photoperiod-induced gonadal growth in mammals and birds: robust, transient increases in GnRH expression and gonadotropin secretion initiate testicular development, but decline markedly thereafter, and are less critical to the maintenance of the long-day reproductive phenotype, once achieved (Bernard et al., 1999, Yellon & Goldman, 1984). Measurement of RFRP expression at several time points during the course of seasonal gonadal growth and regression should permit evaluation of this conjecture.
Gonadal steroids provide feedback regulation of kisspeptin and RFRP (Greives et al., 2008, Kriegsfeld et al., 2006, Smith et al., 2005a, Smith et al., 2005b). Because gonad-intact hamsters were used in the present study, gonadal steroid-dependent effects on Kiss1 and RFRP in the present report cannot be definitively excluded. However, it is unlikely that circulating steroid concentrations account for the broad patterns of altered gene expression for the following reasons. Testosterone concentrations would be expected to be lowest in the group that underwent gonadal regression. However, FR caused decreases in Kiss1 mRNA at the level of the ARC in hamsters maintained in both 16L and 13.5L, whereas gonadal regression occurred only in FR hamsters maintained in 13.5L. Similarly, exposure to 13.5L caused increases in DMH RFRP mRNA in both ad libitum and FR hamsters, despite gonadal regression only occurring in the latter group. In addition, circulating testosterone concentrations of ad libitum fed Siberian hamsters maintained in 13.5L or transferred to 16L do not differ (Paul et al., 2009), yet in the present investigation, RFRP mRNA was elevated in 16L hamsters compared to 13.5L hamsters. Lastly, the direction of these changes in Kiss1 and RFRP expression are opposite of what would be predicted by decreased gonadal steroids (Kriegsfeld et al., 2006, Smith et al., 2005b). Collectively, these arguments suggest that gene expression responses to photoperiod and food reported here are unlikely to be driven by altered gonadal steroid production. This conclusion is consistent with by several other studies which have demonstrated gonadal steroid-independent modulation of RFamide mRNA levels (Greives et al., 2008, Revel et al., 2008, Revel et al., 2006).
In nature, seasonal adaptations are regulated by both day length and non-photic cues. Early-spring and late-summer intermediate-duration day lengths usher an interval of heightened reproductive responsiveness to non-photic cues that ultimately govern the precise timing of seasonal reproductive transitions (Paul et al., 2009). The present results point to two hypothalamic RFamide peptides, kisspeptin and RFRP, in the mediation of photic and non-photic cues on the reproductive system, and suggest that they serve different roles. Intermediate day lengths increase DMH RFRP gene expression, and decreased food availability suppresses Kiss1 mRNA at the level of the ARC. Taken together, the data suggest regional- and neuropeptide-specific recruitment of the RFamide system in the integration of photic and non-photic cues for the control of reproduction.
The authors wish to thank Sean Bradley, August Kampf-Lassin, and Curtis Wilkerson for technical assistance. This work was supported by Grant AI-67406 from the NIAID (BJP), a Social Sciences Divisional Research Fund of the University of Chicago (BJP), Grant NS-58135 from the NINDS (MJP).