Pregnenolone (PREG) and dehydroepiandrosterone (DHEA) have been reported to improve memory in aged rodents. In brain, these neurosteroids are transformed predominantly into 7α-hydroxylated metabolites by the cytochrome P450-7B1 (CYP7B). The biological role of steroid B-ring hydroxylation is unclear. It has been proposed to generate bioactive derivatives that enhance cognition, immune, and other physiological processes. In support, 7α-hydroxylated DHEA increases the immune response in mice with greater potency than the parent steroid. Whether the memory-enhancing effects of PREG in rats is mediated via its 7α-hydroxylated metabolite 7α-hydroxyPREG is not known. We investigated this by treating memory-impaired aged rats (identified by their spatial memory performances in the Morris water maze task compared with young controls) with 7α-hydroxyPREG or PREG administered intracerebroventricularly using osmotic minipumps and then tested the rats during week 2 of steroid treatment in the eight-arm radial-arm version of the water maze (RAWM) that allows repeated assessment of learning. CYP7B bioactivity in hippocampal tissue (percentage conversion of [14C]DHEA to [14C]7α-hydroxyDHEA) was decreased selectively in memory-impaired aged rats compared with both young and memory-intact aged rats. 7α-hydroxyPREG (100 ng/h) but not PREG (100 ng/h) administration to memory-impaired aged rats for 11 d enhanced spatial memory retention (after a 30 min delay between an exposure trial 1 and test trial 2) in the RAWM. These data provide evidence for a biologically active enzyme product 7α-hydroxyPREG and suggests that reduced CYP7B function in the hippocampus of memory-impaired aged rats may, in part, be overcome by administration of 7α-hydroxyPREG.
In addition to being a prime target site of peripheral steroid hormones, the brain itself can also synthesize steroids de novo from cholesterol or by in situ metabolism of blood-borne precursors (Robel and Baulieu, 1995; Vallee et al., 2004). Blood levels of dehydroepiandrosterone (DHEA) in humans decline significantly with age and parallels cognitive decline (Orentreich et al., 1992). Rodents, in contrast, have very little DHEA in brain because of the lack (or undetectable levels) of cyp17α (17α-hydroxylase) that catalyzes the conversion of pregnenolone (PREG) to DHEA in rat and mouse adrenal cortex (Le Goascogne et al., 1991; Mellon and Deschepper, 1993). However, high levels of PREG, the neurosteroid precursor of DHEA, can be detected in the hippocampus of young rats, whereas aged rats have lower levels, which correlate with cognitive deficits (Vallee et al., 1997). PREG administration improves learning and memory in rats and mice (Flood et al., 1992, 1995; Vallee et al., 1997), but the mechanisms involved are unclear.
Local metabolism may influence steroid activity by converting precursor steroid to active or inactive metabolites (Mellon and Deschepper, 1993; Seckl, 1997). We have previously characterized a novel cytochrome P450-7B1 (CYP7B), from a rat hippocampal cDNA library (Stapleton et al., 1995). CYP7B, unusually for a cytochrome P450, is highly expressed in brain, particularly in the hippocampus (Rose et al., 1997). DHEA and PREG are converted to their respective 7α-hydroxylated metabolites by recombinant CYP7B in vitro (Rose et al., 1997). The main product of DHEA or PREG incubation with rat brain extracts is also 7α-hydroxyDHEA or 7α-hydroxyPREG (Akwa et al., 1992), suggesting that the major metabolic route for these neurosteroids in brain is 7α-hydroxylation by CYP7B metabolism. Extrahepatic tissues from homozygous Cyp7b−/− mice (including brain) fail to metabolize DHEA and PREG to their major 7α (and 7β) products (Rose et al., 2001). Thus, outside the liver, CYP7B provides the primary metabolic route for DHEA, PREG, and related 3β-hydroxysteroids. The role played by the 7α-hydroxy metabolites in brain is not clear. Intriguingly, CYP7B may convert precursor steroids to active products. Thus, 7α-hydroxyDHEA, but not DHEA, is neuroprotective in hippocampal cultures, reducing ischemia-induced neuronal damage in vitro (Pringle et al., 2003). Similarly, 7α-hydroxyDHEA, but not DHEA, acts as an immune stimulant in vitro (Lafaye et al., 1999). Given that PREG (and DHEA) have been proposed to be protective against age-associated cognitive decline (Flood and Roberts, 1988; Vallee et al., 2001), a clear understanding of whether local metabolism by CYP7B in the hippocampus regulates its effects on cognition merits investigation. To test this, we exploited the considerable interindividual differences in cognitive abilities found in aged rats (Issa et al., 1990; Gallagher et al., 1993; Yau et al., 1995; Vallee et al., 1997) to first determine whether hippocampal CYP7B bioactivity is related to spatial memory impairments with aging in rats and second whether central administration of the PREG metabolite of CYP7B metabolism, 7α-hydroxyPREG, can improve spatial memory in cognitively impaired aged rats.
Materials and Methods
Male Lister Hooded rats (Charles River, Kent, UK) were obtained at 3 months of age and maintained for up to 26 months under conditions of controlled lighting (lights on from 7:00 A.M. to 7:00 P.M.) and temperature (22°C), with ad libitum access to food (CRM diet cubed; Special Diet Services, Essex, UK) and tap water. The health status was monitored weekly by a veterinarian; animals showing overt signs of respiratory distress, infection, or tumors were excluded. All procedures were performed in strict accordance with the United Kingdom Animals (Scientific Procedures) Act (1986).
Morris water maze.
Young (6 months old; n = 12) and aged rats (24 months; n = 48) (500–600 g body weight) were tested for their spatial learning performance in the Morris water maze, as described previously (Yau et al., 1995). The maze consisted of a 1.8-m-diameter circular white fiberglass pool filled to a depth of 35cm with water (25°C) made opaque with the addition of nontoxic white latex paint and was in a room with extramaze cues on the walls around the pool. A circular escape platform (10 cm diameter) was submerged 1.5 cm below the water surface in one of the quadrants of the pool, and this position remained constant throughout testing. All rats were first habituated to the maze with a 60 s free swim in the pool and were then given 16 trials over 4 consecutive days to learn the location of the submerged platform (four trials per day; 120 s maximum trial duration; 30 s intertrial interval). Latencies to locate the hidden platform and swim paths were monitored by a video camera mounted in the ceiling and a computerized tracking system [HVS image analyzer (Kingston, UK) and Acorn Archimedes computer]. On day 5, the rats were given a 60 s retention test of the spatial location (probe test) with the platform absent, and the percentage time in the pool quadrants was recorded. Finally, the rats were given four 60 s trials (visually cued) in which the platform was raised 1 cm above the water level and had a cardboard flag (8 × 5 cm) placed securely on top to render it visible. This tested for visual, motivational, or motor deficits that may have influenced performance.
The cognitive status of the aged rats was defined on the basis of their latencies to find the platform on days 3 and 4 of testing relative to the mean latency of the young rats. Aged impaired (AI) rats were defined as those animals with mean latencies that differed by >2 SDs from that of the young controls. Aged unimpaired (AU) rats were identified as those whose performance (mean latencies over days 3 and 4) was ≤0.5 SD from that of young controls. Aged rats that were intermediate were not further used in the present study. The 7α-hydroxylase activity in the hippocampus from a selection of the cognitively tested rats was determined. Rats were killed in the morning by decapitation, and brains were dissected and stored frozen at −80°C before the assay. To generate sufficient numbers of age-impaired rats to study the effects of central administration of pregnenolone and 7α-hydroxypregnenolone on spatial memory, another cohort of aged rats (n = 27) and young controls (n = 6) were screened for their spatial learning performance in the Morris water maze as above. The brains from the cognitively tested rats not used in the enzyme bioactivity assays or steroid treatment studies were stored at −80°C for future simultaneous quantification of endogenous levels of PREG, DHEA, and their 7α-hydroxylated metabolites.
Individual hippocampi from young (n = 6) and aged rats defined as cognitively impaired and unimpaired (n = 6 per group) from the water maze testing were homogenized in 3 vol of ice-cold PBS (containing 1 mm EDTA, 20% glycerol, and protease inhibitors) and centrifuged at 4000 × g for 5 min at 4°C. The supernatant was removed, and protein concentration was determined, aliquoted, and stored at −80°C until use. Ethanol solutions of [14C]DHEA (53.8 mCi/mmol, 0.02 mCi/ml) were made (1:10 dilution) and dried down in glass tubes. Incubations were performed in a total volume of 200 μl constituting of W buffer (0.1 m KPO4 and 1 mm EDTA, pH 7.4) and hippocampal extract (0.5 mg protein). The tubes were vortexed and placed in a shaking water bath at 37°C, and 25 μl of 8 mm NADPH was added (to give final concentration of 1 mm). After incubating for 30 min, the reactions were stopped by placing the tubes on ice and extracting the steroids (2× 500 μl ethyl acetate). The organic phase was dried down under compressed air at 60°C and was resuspended in 30 μl of ethyl acetate. The reaction constituents were applied to silica thin-layer chromatography (TLC) plates (Merck Biosciences, Nottingham, UK) and developed in solvent solution (ethyl acetate, n-hexane, and acetic acid, 16:8:1). Reaction substrates and products were visualized and quantified by a phosphoimager.
Treatment and behavioral testing schedule.
The performance of old rats in spatial learning tasks can be significantly influenced by previous exposure to that task, even months earlier (van Groen et al., 2002). To avoid this, the animals were retested during intracerebroventricular neurosteroid treatment in a spatial working memory task using the radial arm water maze (RAWM) because it has been reported that age-related memory impairments are lost with repeated testing in the water maze, whereas previous experience in the water maze did not have any effect on learning of the radial maze (Dellu et al., 1997). This is a cognitively more demanding spatial learning task, so the age-impaired rats assigned for treatment and young controls were first trained in the RAWM for 5 d (with no delays between trials; see below) before the start of neurosteroid treatment. One week after stereotaxic implantation of brain cannulas and osmotic minipumps, the rats were tested in the RAWM (during week 2 of osmotic minipump drug delivery) and were given five trials per day with no delays for 3 d. Then on the next 2 consecutive days, they were each given a 30 min delay retention test between trials 1 and 2. At the end of the osmotic minipump action 2 d later, all rats were killed in the morning (9:00–10:00 A.M.) by decapitation, trunk blood was collected in EDTA-coated tubes for corticosterone measurements, and brains were removed and stored frozen (−80°C).
Surgery and steroid treatment.
Animals were anesthetized with halothane and subjected to stereotaxic implantation of chronic indwelling stainless steel cannulas (Alzet brain infusion cannula; Alza, Palo Alto, CA) into the lateral ventricle (coordinates were as follows: anteroposterior, 1.0 mm posterior to bregma; lateral, 1.5 mm; dorsoventral, 4.5 mm ventral to dura). The cannula was secured to the skull with dental acrylic, and the external part of the intracerebroventricular cannula was connected with polyethylene tubing to an Alzet osmotic minipump (2 weeks, 0.5 μl/h; model 2002; Alza) placed subcutaneously behind the neck. Steroids were dissolved in artificial CSF with 10% hydroxypropyl betacyclodextrine (a polymer that is nontoxic and devoid of intrinsic effects, used to dissolve hydrophobic compounds in aqueous media). A selection of the aged rat cohort with significantly impaired spatial learning compared with young controls (as assessed in the Morris water maze task) were randomly allocated to treatment with PREG (100 ng/h, n = 6; Sigma, Poole, UK), 7α-hydroxyPREG (100 ng/h, n = 7; Steraloids, Newport, RI), or vehicle (0.5 μl/h, n = 8). A young control vehicle group (0.5 μl/h, n = 6) was also included. All rats were singly housed after surgery. The dose of PREG chosen was based on a previous study that showed significantly increased memory performance with 100 ng/h PREG-sulfate (S) given intracerebroventricularly in mice (Ladurelle et al., 2000) and acute doses used in other studies (Flood and Roberts, 1988; Flood et al., 1992). The estimated rate of CSF production in the nervous system of mice is ∼20 μl/h, whereas in rats it was ∼520 μl/h (Ladurelle et al., 2000), so the highest effective dose was chosen because this will be effectively diluted by the greater volume of CSF in the rat. No other study has tested the effects of the 7α-hydroxylated metabolites on memory performance and the endogenous levels in brain are not known, so the same dose as the parent steroid was used for comparison.
Radial arm water maze.
The RAWM consisted of the 1.8 m circular pool with eight white polypropylene inserts (54 cm high) within the tank that divide the pool into eight swim paths (59-cm-long arms) radiating out of an open central area (62 cm diameter). This spatially more complex task forces the rat to swim in either the central open area or one of the arms and takes advantage of the variable spatial complexity of the radial arm maze and the efficient learning of the water maze (Diamond et al., 1999). A hidden platform (10 cm diameter) was located in one of the arms 1.5 cm below the water surface at least 8 cm from the tank wall and remained in the same position within 1 d but was in a different arm across days. Each trial began with the rat placed directly in the entrance of one of the arms with no hidden platform (“start arm”) with the rat facing the central area. Start locations were randomized. Young and aged rats were given five trials per day (60 s maximum duration) to locate the platform. If the rat did not find the platform in 60 s, it was guided to the platform by the experimenter. The rat remained on the platform for 30 s before the next trial was initiated. The latency to find the platform was recorded, and the number of errors made per trial was noted. An error was committed when a rat entered an arm that did not contain the platform or if a rat entered the correct arm but did not find the platform. For the memory retention trial, a 30 min delay period was used after the first trial. During this time, the rat was dried in a towel and placed in its cage under a warming lamp. The rest of the trials continued with no delay. The 30 min delay was chosen because a longer delay of 60 min impaired memory retention in young rats in our pilot studies (data not shown).
Plasma corticosterone levels were measured by radioimmunoassay modified for microtiter plate scintillation proximity assay (Amersham Biosciences, Little Chalfont, UK) with a highly specific antiserum (Dr. C. Kenyon, University of Edinburgh, Edinburgh, UK) and [3H]corticosterone (Amersham Biosciences).
Data were analyzed using one- or two-way ANOVA with the age and treatment as between-subject factors. Scheffé's tests were applied for post hoc analysis. For spatial learning in the Morris water maze, the repeated within-subject factor was the day, and, for spatial working memory in the RAWM, the trial was the within-subject factor. All data are expressed as group means ± SEM. Significance was defined as p < 0.05 or p < 0.01.
Water maze spatial learning impairments only in a subgroup of aged rats
Across the 4 d of training, escape latencies decreased (F(3,189) = 59.9; p < 0.001) in both young and aged rats as they showed learning of the submerged platform location. The aged rat cohort exhibited significantly higher mean escape latencies than the young controls (F(1,63) = 16.5; p < 0.05) on each day (Fig. 1A). Of the 48 aged rats tested, 25% were classed as cognitively AI and 25% AU (Fig. 1B). The first trial latencies on day 1 of spatial training were not significantly different between young and aged rats (p = 0.19; young, 70.3 ± 11.5 s; AI, 96.7 ± 10.3 s; AU, 92.3 ± 10.7 s). In contrast, the mean escape latencies (block of four trials) were significantly different between AI and AU rats (p < 0.001) but not between AU and young controls for each of days 2–4 (Fig. 1B). The day after the last acquisition trial, each rat was given a 60 s probe test with the submerged platform absent from the pool. AI animals spent significantly less time in the target quadrant of the pool compared with either AU or young rats (F(2,38) = 8.3; p < 0.01), whereas AU did not significantly differ from young controls (Fig. 1C). Swim speeds were similar in young and aged rats (p = 0.07; young, 20.7 ± 0.6 cm/s; AU, 18.3 ± 0.5 cm/s; AI, 19.8 ± 0.9 cm/s), suggesting no motor deficit to explain differences in latency in AI rats. Furthermore, there was no significant difference in latencies to locate the visible platform between AI (6.8 ± 0.8 s) and AU (5.7 ± 0.3 s) or young (6.3 ± 0.5 s), indicating that the impaired spatial memory performance of the AI animals is not a consequence of poor vision or inability to swim during this task.
CYP7B bioactivity is selectively decreased in the hippocampus of cognitively impaired aged rats
The 7α-hydroxylase activity of CYP7B in the hippocampus was measured by incubating hippocampal microsomes from young, AI, and AU rats with [14C]DHEA as substrate and then identification of the [14C]7α-hydroxyDHEA metabolite produced based on the retention factor identical to that of the reference steroid on TLC plates. DHEA was chosen rather than PREG because this has been found previously to be the preferred substrate in rat brain microsomes (Akwa et al., 1992; Rose et al., 1997). CYP7B bioactivity was decreased (F(2,12) = 13; p < 0.01) selectively in AI compared with young (34% reduction) and AU (45% reduction) rats, whereas AU rats showed similar CYP7B bioactivity to young controls (Fig. 2).
Spatial memory and RAWM performance of rats before intracerebroventricular administration of PREG or 7α-hydroxyPREG
The individual aged rat treatment groups did not differ in their spatial learning (escape latency across days; repeated-measures ANOVA, F(2,18) = 0.08; p = 0.92) (Fig. 3A) before neurosteroid administration. All three aged rat treatment groups showed impaired spatial learning compared with young controls (p < 0.05) (Fig. 3A). Spatial working memory, tested in the RAWM, did not differ between aged rat treatment groups (F(2,18) = 0.46; p = 0.64). All three aged rat treatment groups show significantly impaired spatial working memory (within-days mean trials over 3 consecutive days) compared with young controls (p < 0.01) (Fig. 3B).
RAWM working memory of aged rats after intracerebroventricular administration of PREG or 7α-hydroxyPREG
Vehicle-control aged rats showed impaired short-term working memory (within-days mean trials over days 8 and 9 of treatment) compared with young vehicle controls (repeated-measures ANOVA, F(1,12) = 7.2; p < 0.05) (Fig. 4A). Neither PREG (F(1,12) = 0.48; p = 0.5) nor 7α-hydroxyPREG (F(1,13) = 0.33; p = 0.57) significantly affected short-term spatial learning in the aged rats (Fig. 4A). Trial 1 latencies (F(3,23) = 0.85; p = 0.48) and errors (F(3,23) = 1.1; p = 0.38) did not differ between the treatment groups. On days 10 and 11 of treatment, young vehicle controls show retention of the platform location after a 30 min delay, as indicated by a decrease in trial 2 latency compared with trial 1 (p < 0.001, paired t test) (Fig. 4B). In contrast, aged vehicle rats showed impaired spatial memory retention (trials 1 and 2 latencies did not differ; p = 0.64, paired t test) (Fig. 4B). Aged vehicle controls were impaired in both escape latencies (F(1,12) = 16.5; p < 0.01) and errors (F(1,12) = 28.4; p < 0.001) compared with young vehicle controls. 7α-hydroxyPREG improved spatial memory retention in the AI rats compared with aged vehicle controls in both escape latencies (F(1,13) = 7.6; p < 0.05) and errors (F(1,13) = 8.2; p < 0.05), and this was not significantly different from young vehicle controls (Fig. 4B). In contrast, PREG had no significant effect on spatial memory retention in AI rats compared with aged vehicle controls in both escape latencies (F(1,12) = 2.6; p = 0.13) and errors (F(1,12) = 4.1; p = 0.07) (Fig. 4B).
Swim speeds and plasma corticosterone levels
The average swim speeds was not influenced by treatment (F(3,23) = 1.6; p = 0.2), but all groups did show significantly increased swim speeds across the five trials (F(4,92) = 22.9; p < 0.001) (Table 1). Plasma corticosterone levels measured from trunk blood was not significantly affected by treatment (young, 9.59 ± 1.9 μg/dl; old, vehicle, 13.3 ± 1.8 μg/dl; old, PREG, 13.4 ± 5.6 μg/dl; old, 7α-hydroxyPREG, 7.6 ± 1.6 μg/dl).
Although tissue-specific metabolism of glucocorticoids by 11β-hydroxysteroid dehydrogenase type 1 (11β-HSD1) affects hippocampal-dependent learning and memory in aged mice (Yau et al., 2001), little is known about any analogous importance of local metabolism of neurosteroids such as DHEA and PREG. The major ex vivo metabolic route for PREG and DHEA in brain is 7α-hydroxylation by CYP7B (Stapleton et al., 1995), a reaction that is conserved across mammalian species, including primates and humans (Akwa et al., 1992; Doostzadeh and Morfin, 1996; Weill-Engerer et al., 2003; Yau et al., 2003). Here we describe, for the first time, (1) a significant and selective decrease in hippocampal CYP7B bioactivity in aged 24-month-old rats with impaired spatial memory performance in the water maze but not in age-matched cognitively intact rats and (2) treatment with the CYP7B product 7α-hydroxyPREG to cognitively impaired aged rats enhances spatial memory retention after a 30 min delay in the radial arm water maze, whereas PREG, the parent steroid, had no significant effect at the dose tested.
We found large interindividual differences in the spatial memory performances of aged rats in the Morris water maze, consistent with previous reports (Issa et al., 1990; Gallagher et al., 1993; Yau et al., 1995). Hippocampal CYP7B bioactivity was reduced selectively in aged-impaired, but not in aged unimpaired, rats that maintained levels similar to young controls. Decreased hippocampal CYP7B bioactivity is therefore not an inevitable consequence of aging. A previous study also reported somewhat lower 7α-hydroxylation of PREG and DHEA by mouse brain microsomes with age (Doostzadeh and Morfin, 1996), albeit the oldest mice tested were only 10 months old, less than half their lifespan, and these were compared with 2-month-old juveniles [CYP7B expression in hippocampus shows a complex ontogeny (Bean et al., 2001)]. Hippocampal PREG levels decline with age and correlate with cognitive impairments in rats (Vallee et al., 1997), suggesting that both the substrate for CYP7B and the enzyme itself may decline in a subset of aged rodents with cognitive deficits. Interestingly, CYP7B mRNA levels are markedly reduced in surviving hippocampal neurons of Alzheimer's disease postmortem brain compared with age-matched controls without CNS disease (Yau et al., 2003). It is unknown whether CYP7B bioactivity decreases in parallel with its mRNA expression or whether this occurs in cognitive impairment without frank dementia in aging humans. Nonetheless, data from rodents and humans are consistent with an association between reduced hippocampal CYP7B function and cognitive decline with aging.
Although recent studies have reported anti-glucocorticoid, immunomodulatory, and antiapoptotic effects of 7α-hydroxylated DHEA (Loria and Padgett, 1998; Hampl et al., 2000; Morfin and Starka, 2001; Pringle et al., 2003) and 7α-hyroxyPREG (Morfin and Courchay, 1994), the effect of the CYP7B metabolites on memory have not been examined. Although CYP7B bioactivity was assayed with [14C]DHEA as the preferred substrate based on previous findings (Akwa et al., 1992; Rose et al., 1997), we chose to examine the effects of PREG and its 7α-hydroxylated metabolite on memory in the aged-impaired rats because this is the more abundant neurosteroid in rodent brain. DHEA is mostly a human steroid with very low levels detected in rodent brain; one recent study detected 0.27 ng/g DHEA in adult rat brain (Ebner et al., 2006). PREG, at least at the dose tested, which has been shown previously to enhance memory retention in young rodents (Ladurelle et al., 2000) and transiently improve memory deficits in cognitively impaired aged rats (Vallee et al., 1997), did not significantly improve spatial memory retention in aged impaired rats (after a 30 min delay in RAWM), although there was a trend for improved memory (decreased errors and latency after delay). It may be that a higher dose is required for effects on memory in aged animals. In contrast, 7α-hydroxyPREG at the same dose clearly enhanced spatial memory retention after the delay, and this was not attributable to nonspecific effects on swim speeds, which did not change with treatment. Future studies should examine lower doses of 7α-hydroxyPREG as well as 7α-hydroxyDHEA by intrahippocampal infusion to confirm the primary site of action of the 7α-hydroxylated metabolites. Although hippocampal levels of 7α-hydroxyPREG and PREG in the treated rats remain to be determined, our data nevertheless show that 7α-hydroxyPREG is an active metabolite, at least in enhancing spatial memory retention. Because the cognitively impaired aged rats show a 30–40% decline in CYP7B bioactivity, some of the PREG administered will still get converted to the 7α-hydroxylated metabolite, which may explain the trend for PREG to show some improvement in memory and why the effects of 7α-hydroxyPREG in the RAWM was not significantly better than PREG-treated aged rats. One key study would be to treat Cyp7b knock-out mice (Rose et al., 2001) with PREG or DHEA; the memory-enhancing effects reported previously in mice (Flood et al., 1988, 1995) should be absent in the Cyp7b knock-out mice if these steroids require activation by 7α-hydroxylation.
Consistent with the 7α-hydroxy steroids being active, peripheral injection of both 7α-hydroxyPREG and 7α-hydroxyDHEA have been reported to increase immune responses in mice (Morfin and Courchay, 1994). In addition, 7α-hydroxyDHEA can be further converted to 7-oxo-DHEA by 11β-HSD1 (Robinzon et al., 2003), itself widespread in the CNS (Moisan et al., 1990), and 7-oxo-DHEA reverses scopolamine-induced memory impairments in young mice (Shi et al., 2000). Indeed, if 7α-hydroxylated metabolites are more active than the parent steroid, then this may in part explain why oral DHEA replacement in the elderly has not successfully improved memory (Wolf et al., 1997; Huppert and Van Niekerk, 2001), whereas DHEA administration is effective at improving memory recollection and mood in healthy young subjects (Alhaj et al., 2006), presumably because they maintain greater hippocampal CYP7B activity. It may be that the 7α-hydroxylated metabolite rather than the parent steroid would be more effective therapy in the elderly.
The mechanism of action of 7α-hydroxyPREG on memory is not clear. PREG can act via nongenomic mechanisms by binding to the GABAA and NMDA receptors to enhance neuronal excitability (Paul and Purdy, 1992; Sliwinski et al., 2004). Whether 7α-hydroxyPREG also interacts with these membrane receptors and with greater potency has yet to be determined. Anti-glucocorticoid effects of 7α-hydroxyDHEA have been shown, for example in counteracting dexamethasone-induced thymic and T cell involution (May et al., 1990). 7α-HydroxyDHEA may also be a substrate for 11β-HSD1 in hepatocytes (Robinzon et al., 2003), suggesting the possibility that it may interfere with 11β-HSD1 regenerating active glucocorticoids from inert 11-keto forms (Muller et al., 2004). Indeed, 11β-HSD1 null mice lack local intracellular glucocorticoid regeneration and are protected from the decline in cognitive function seen in wild-type mice (Yau et al., 2001). Whether 7α-hydroxyPREG has similar effects to 7α-hydroxyDHEA in its neuroprotective (Pringle et al., 2003) and anti-glucocorticoid (May et al., 1990) actions merits additional investigation. If 7α-hydroxyPREG had anti-glucocorticoid effects, this could be partly responsible for the enhanced memory retention in the aged impaired rats because elevated glucocorticoid levels correlates with impaired learning in the water maze (Issa et al., 1990; Yau et al., 1995). Although plasma corticosterone levels were not significantly altered by 7α-hydroxyPREG or PREG treatment, this does not rule out changes in hippocampal tissue corticosterone levels, which can be very different from circulating levels because of the presence 11β-HSD1 (Yau et al., 2001). Future work to measure the hippocampal levels of corticosterone and 7α-hydroxyPREG as well as other neurosteroids in the cognitively tested aged rats (and young controls) may help elucidate the molecular basis for the enhanced memory in the age-impaired rat after 7α-hydroxyPREG treatment. The identification and simultaneous quantification of PREG and DHEA and their 7α-hydroxylated metabolites in individual rat hippocampal samples (to allow levels to be correlated with spatial memory) is technically very challenging. Using the LC-MS-MS (a triple quadrupole mass spectrometer with associated HPLC) and after converting the steroids to oxime derivatives to improve the limit of detection, we were able to identify and quantify PREG (∼3 ng/g tissue) and DHEA (∼1 ng/g tissue) concentrations but not 7α-hydroxyPREG within the same individual control whole rat brain (N. Homer, J. Yau, J. Seckl, and R. Andrew, unpublished preliminary data). A more sensitive method to detect the 7α-hydroxyPREG levels will require the development of a radioimmunoassay. Intracerebroventricular administration of PREG-S to rats has been reported previously to enhance acetylcholine release in basolateral amygdala, cortex, and hippocampus (Mayo et al., 2003). If this is attributable in part to the conversion to the 7α-hydroxylated metabolite rather than PREG per se, then this could be one possible explanation to the memory-enhancing effects of 7α-hydroxyPREG in aged impaired rats. Whatever the mechanism, we have shown that 7α-hydroxyPREG is an active metabolite that can enhance memory retention in aged impaired rats that may have reduced CYP7B activity. Direct administration of the 7α-hydroxylated steroid product of CYP7B may be a novel therapeutic approach to improve cognitive function with aging in humans.
This work was supported by a Biotechnology and Biological Sciences Research Council Science into Ageing grant (J.L.W.Y., J.R.S.) and an Alzheimer's Research Trust Carter Fellowship (J.L.W.Y.). We thank Drs. D. M. Diamond and W. B. Rowe for their helpful discussions on the radial arm water maze protocol.
- Correspondence should be addressed to Dr. Joyce L. W. Yau, Endocrinology Unit, The Queen's Medical Research Institute, University of Edinburgh, 47 Little France Crescent, Edinburgh EH16 4TJ, UK.