Neurosteroid biosynthetic pathways changes in prefrontal cortex in Alzheimer\'s disease

Neurobiology of Aging 32 (2011) 1964–1976

Neurosteroid biosynthetic pathways changes in prefrontal cortex in Alzheimer’s disease Sabina Luchetti a,b,∗ , Koen Bossers a , Saskia Van de Bilt a , Vincent Agrapart c , Rafael Ramirez Morales c , Giovanni Vanni Frajese c,d , Dick F. Swaab a a

Netherlands Institute for Neuroscience (NIN), An Institute of the Royal Netherlands Academy of Arts and Sciences, Neuropsychiatric Disorders Lab, Meibergdreef 47, 1105 BA Amsterdam, The Netherlands b Istituto Dermopatico Immacolata (IDI-IRCCS), Santa Lucia Foundation Scientific Institute, Rome, Italy c Internal Medicine Department, Endocrinology Lab, University of Rome “Tor Vergata”, Rome, Italy d Faculty of Motor Sciences, University of Cassino, Cassino, Italy Received 29 July 2009; received in revised form 10 December 2009; accepted 17 December 2009 Available online 31 December 2009

Abstract Expression of the genes for enzymes involved in neurosteroid biosynthesis was studied in human prefrontal cortex (PFC) in the course of Alzheimer’s disease (AD) (n = 49). Quantitative RT-PCR (qPCR) revealed that mRNA levels of diazepam binding inhibitor (DBI), which is involved in the first step of steroidogenesis and in GABAergic transmission, were increased, as were mRNA levels for several neurosteroid biosynthetic enzymes. Aromatase, 17␤-hydroxysteroid dehydrogenase type 1 (HSD17B1) and aldo-keto reductase 1C2 (AKR1C2), were all increased in the late stages of AD. Several GABA-A subunits were significantly reduced in AD. Increased expression of aromatase in the PFC was confirmed by immunohistochemistry and was found to be localized predominantly in astrocytes. These data suggest a role for estrogens and allopregnanolone produced by astrocytes in the PFC in AD, possibly as part of a rescue program. The reduced gene expression of some synaptic and extra-synaptic GABA-A subunits may indicate a deficit of modulation of GABA-A receptors by neuroactive steroids, which may contribute to the neuropsychiatric characteristics of this disease. © 2009 Elsevier Inc. All rights reserved. Keywords: Neurosteroids; GABA-A receptors; Alzheimer’s disease; Postmortem; Prefrontal cortex; Aromatase; Gene expression

1. Introduction The sex steroids, i.e. estrogens, androgens and progesterone, when synthesized and metabolized in the central nervous system (CNS), are known as neurosteroids (Baulieu, 1998). In neural tissue, the enzymes involved in steroidogenesis are present both in glial cells and neurons (Do Rego et al., 2009; Mellon and Vaudry, 2001; Stoffel-Wagner, 2001). A scheme of the sex steroid biosynthesis pathway in the brain and the abbreviations used in the text are shown in Fig. 1. There is substantial evidence suggesting that sex steroids can mediate neuroprotection and influence neuronal survival, ∗ Corresponding author at: Meibergdreef 47, 1105 BA Amsterdam, The Netherlands. Tel.: +31 205665505; fax: +31 205666121. E-mail address: [email protected] (S. Luchetti).

0197-4580/$ – see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.neurobiolaging.2009.12.014

neuronal and glial differentiation and myelination in the CNS by regulating gene expression of neurotrophic factors and anti-inflammatory molecules (Behl, 2002; Bialek et al., 2004; Djebaili et al., 2005; Melcangi et al., 2008; Schumacher et al., 2007). Progesterone, testosterone and estradiol have been shown to have neuroprotective and regenerative effects in in vitro models of neurodegeneration and in animal models of brain injury (Gouras et al., 2000; Schumacher et al., 2003; Vongher and Frye, 1999). On the other hand, some metabolites of pregnenolone, progesterone, testosterone and deoxycorticosterone (DOC) are also regarded as “neuroactive” because of their ability to modulate neurotransmitter activity. Among these, 3␣5␣-tetrahydro progesterone (3␣5␣-THP or allopregnanolone), androstanediol and 3␣5␣-tetrahydro DOC (3␣5␣-THDOC) are positive allosteric modulators of

S. Luchetti et al. / Neurobiology of Aging 32 (2011) 1964–1976

1965

Fig. 1. Sex steroid biosynthetic pathway. Sex steroids as well as many of their metabolites have been shown to have pleiotropic actions in the CNS, all of which could contribute to neurodegenerative processes and repair (Schumacher et al., 2000). All the main steps of sex steroids synthesis are shown, including the initial transport of cholesterol into the mitochondria, mediated by the complex: STAR, TSPO and DBI. Numbers after the enzyme name indicate the isoform type. Abbreviations: AKR1C, aldo-keto reductase 1C; CYP11A1, cytochrome P450scc; CYP17A1, cytochrome P450c17A1; CYP21A2, cytochrome P450c21B; DBI, diazepam binding inhibitor; DOC, deoxycorticosterone; 5␣-DH-DOC, 5␣-dehydro-doxycorticosterone; DHEA, dehydroepiandrosterone; 5␣-DHP, 5␣-dehydroprogesterone; HSD3B, 3␤-hydroxysteroid dehydrogenase; HSD17B, 17␤-hydroxysteroid dehydrogenase; STAR, steroid acute regulator; SRD5A, 5␣-reductase; SULT, sulphotransferase; STS, steroid-sulphatase; 3␣5␣-THDOC, 3␣ 5␣-tetrahydro-deoxycorticosterone; 3␣5␣-THP, 3␣5␣-tetrahydroprogesterone; TSPO, 18 kDa translocator protein.

ionotropic ␥-amino-butyric acid (GABA-A) receptors. In particular, allopregnanolone is considered the most potent allosteric modulator of GABA-A receptors, acting in a benzodiazepine (BDZ)-like manner (Belelli and Lambert, 2005). The modulatory activity of neuroactive steroids on the GABA system is well-established (Belelli and Lambert, 2005). This involves interaction with post-synaptic GABA-A receptors, most commonly containing the ␣1 (GABRA1), ␤2 (GABRB2) and ␥2 (GABRG2) subunits, and extra-synaptic GABA-A receptors commonly containing ␣4 (GABRA4), ␦ (GABRD) or ␧ (GABRE) subunits (Farrant and Nusser, 2005). Neuroactive steroids can also regulate the expression of GABA-A receptor subunit genes in vitro and in vivo (Biggio et al., 2001). As a consequence of these properties, allopregnanolone and the other neuroactive compounds modulate memory processes, anxiety, sleep processes, responses to stressful stimuli and seizure susceptibility and may influence cognitive and neuropsychiatric symptoms such as those seen in AD (Dubrovsky, 2005). While a role for sex steroids in neuroprotection has been demonstrated in animal studies including AD models (Carroll et al., 2007; Ciriza et al., 2004; He et al., 2004), information is lacking about the neurosteroid biosynthetic pathway in the human CNS during neurodegenerative processes in AD, partly because of the difficulty in obtaining suitable human brain tissue. Decreased blood levels of sex steroids with aging have been associated with an increased risk of AD (Cholerton et al., 2002; Pike et al., 2006). Combined with evidence of reduced levels of steroids such as testosterone in human AD brain (Rosario et al., 2004), this raises the possibility that alterations in gene expression of the enzymes which synthe-

size neurosteroids may be involved in the pathology of AD, which may in turn result in reduced neuroprotective actions. Furthermore, the evidence that the GABA system is relatively conserved in AD prefrontal cortex (PFC) compared to other neurostransmitter systems (Francis, 2003; Lowe et al., 1988; Reinikainen et al., 1988) suggests that this system represents an important target of the neurosteroids, especially in late stage AD when the neurodegenerative process is advanced. The goal of the present study is to elucidate the gene expression of the enzymes involved in the synthesis of neurosteroids in the human PFC during the neuropathological progression of AD. Using quantitative RT-PCR (qPCR) we analyzed a list of 37 genes including the key biosynthetic enzymes, the steroid hormone receptors, and the GABA-A receptor subunits on which the neurosteroids exert their modulatory actions in the brain. Immunohistochemistry (IHC) experiments were also performed to confirm the main qPCR findings at the protein level.

2. Materials and methods 2.1. Subjects Postmortem human brain tissue was obtained from The Netherlands Brain Bank, Netherlands Institute for Neuroscience, Amsterdam (NBB). Donors or their next of kin gave written informed consent to the NBB to allow the brain autopsy and to use the material and clinical information for research purposes.

1966

S. Luchetti et al. / Neurobiology of Aging 32 (2011) 1964–1976

Table 1 Patients neuropathological information. Braak stage (n = 7/group)

Age (years) (mean ± SD)

BR 0 BR 1 BR 2 BR 3 BR 4 BR 5 BR 6

70.5 80.3 76.7 85 82.3 74.3 70.3

± ± ± ± ± ± ±

9.6 5.5 7.8 6.35 4.92 6.5 7.85

M/F

PMD hours (mean ± SD)

4/3 3/4 3/4 3/4 3/4 4/3 3/4

6.8 6 7.3 6 5 5.6 4.7

± ± ± ± ± ± ±

1.6 1 1.6 1.7 1.7 1.3 0.76

Brain pH (mean ± SD) 6.7 6.2 6.7 6.7 6.6 6.5 6.7

± ± ± ± ± ± ±

0.3 0.2 0.2 0.3 0.2 0.1 0.2

RIN (mean ± SD) 8.1 8.7 9 7.7 8.2 8.2 7.6

± ± ± ± ± ± ±

1.1 1 0.1 0.9 0.84 0.84 0.47

SD = standard deviation; M/F = male/female patients proportion; PMD = postmortem delay; RIN = RNA integrity number.

Donors were grouped by Braak stage according to neuropathological diagnosis (Braak and Braak, 1991). Based on the distribution of neurofibrillary tangles, 7 patients were chosen for each Braak stage (BR 1–6) and 7 subjects with no tangles (BR 0) and no neurological or psychiatric disease were included as controls. Subsequently patients were grouped based on the cognitive characteristics associated with each Braak stage as follows: BR 0–2 = no cognitive impairment; BR 3–4 = mild cognitive impairment; BR 5–6 = fully developed AD (Bancher et al., 1996). In addition, the same subjects were also divided into three groups based on the amount of amyloid plaques (Braak and Braak, 1991): A (absent or low amount), B (moderate amount), C (high amount). Subjects that were administered with indomentacin, hormone therapies and corticosteroids up to 1 month or less before death were excluded from the study. Alzheimer’s patients were treated with benzodiazepine and opioids, which are routinely administered in the last days of hospitalization, but controls were matched for such treatment to minimize variability of the results. Subjects of each group were matched by age, gender, brain pH and postmortem delay. Neuropathological information is summarized in Table 1. 2.2. RNA isolation and cDNA synthesis Snap-frozen postmortem prefrontal medial gyrus samples were used. From each sample, approximately 30 sections of 50 ␮m were cut using a cryostat, and grey matter was dissected out using scalpels. RNA isolation and cDNA synthesis were performed as described previously (Bossers et al., 2009) using 1 ␮g of Dnase-treated RNA input for cDNA synthesis. RNA purity was determined using a NanoDrop ND-1000 spectrophotometer (Nanodrop Technologies, Wilmington, Delaware). RNA Integrity Number (RIN) was measured on the Agilent 2100 bioanalyzer (Agilent Technologies, Palo Alto, CA). All RNA samples used in this study had RIN values of at least 6.5. A dilution of 1:10 of the total cDNA yield was used for the qPCR experiments. 2.3. Quantitative PCR Sequences of the primer pairs and product sizes are shown in Supplementary Table S1. For the AKR1C gene isoforms

1, 2 and 3, which are highly similar, specific primer pairs were generated. Where possible, primer pairs were designed to span the 3 -most intron to avoid amplification of DNA templates that may be present in trace amounts in the RNA samples. The qPCR reaction contained 10 ␮l of 2× SYBR Green Mastermix (Applied Biosystems, Foster City, CA, USA), 3 ␮l of each primer pair (1 ␮M) and 7 ␮l of template cDNA (equivalent to 7 ng of total RNA) in a 20 ␮l reaction volume. The PCR was performed in a GeneAmp 7300 thermocycler PCR program: 10 min at 95 ◦ C, followed by 40 cycles of 15 s at 95 ◦ C and 1 min at 60 ◦ C. For the AKR1C1 gene, the long product (280 bp) was amplified by increasing the annealing/extension phase to 2 min. The specificity of the amplification was checked by a melting curve analysis and electrophoresis of the products on an 8% polyacrylamide gel. Sterile water, RNA samples without addition of reverse transcriptase in the cDNA synthesis and DNA samples were used as controls. Linearity of each qPCR assay was tested by preparing a series of dilutions of the same stock cDNA in multiple plates. The normalization factor was based on the geometric mean of the following 7 reference genes selected by a geNorm analysis (Vandesompele et al., 2002): AURKAIP1; DHX16; ERBP KLHDC5; ISOC2; SNW1; TM9SF4. The relative absolute amount of target genes was calculated by the formula 1010 × E−ct (E = 10−(1/slope) ) (Pfaffl, 2001). The absolute amount of transcript thus determined was then divided by the normalization factor to obtain the relative mRNA expression of the target gene. 2.4. Immunohistochemistry Immunohistochemistry was performed on formalin fixed paraffin sections (6 ␮m thick) of human prefrontal cortex (PFC) from 3 patients per Braak stage (Braak 0–6). Sections were mounted on SuperfrostPlus slides (Menzel, Germany) and dried overnight at 58 ◦ C followed by 24–36 h at 37 ◦ C. The sections were deparaffinized and rehydrated in an ethanol series and then rinsed in distilled water. After antigen retrieval treatment (0.01 M Citrate Buffer pH 6 microwave for 10 min at 90 ◦ C), sections were allowed to cool for 30 min. Primary antibodies and dilutions used were: (1) rabbit polyclonal anti-aromatase (AROM)(Yague et al., 2006), 1:2000, kindly provided by

S. Luchetti et al. / Neurobiology of Aging 32 (2011) 1964–1976

1967

Prof. L.M. Garcia Segura, Instituto Cajal, Madrid, Spain; (2) rabbit anti-AKR1C (AKR1C1–4) (Lin et al., 2004), 1:500, kindly provided by Prof. T.M. Penning, University of Pennsylvania School of Medicine, USA; (3) rabbit polyclonal anti-steroid 5␣-reductase (SDR5A) (Thigpen et al., 1993), 1:500, kindly provided by Prof. D.W. Russell, University of Texas Southwestern Medical Center, Dallas, USA. For AKR1C a blocking step with TBS/5% milk was performed, while for SDR5A sections were blocked with TBS/0.1% milk. A goat-anti rabbit secondary antibody (1:400, Vector Laboratories) was applied, followed by avidin biotin complex (ABC, 1:800, Vector Laboratories) and colour developed with 0.5 mg/ml 3,3-Diaminobenzidinetetrahydrochloride (Sigma, Zwijndrecht) and ammonium nickel sulphate (2.2 mg/ml). For glial fibrillary acidic protein (GFAP) immunofluorescent double staining, mouse anti-GFAP (Sigma–Aldrich, 1:400) with AROM (1:1000) or AKR1C (1:500) were used. Donkey anti-rabbit (Alexa488-conjugate; 1:400) and donkey anti-mouse (Alexa 594-conjugate; 1:400) were used as secondary antibodies. 2.5. Statistical analysis Statistical analysis was conducted with SPSS (version 16.0, SPSS Incorporation). The differences between the groups were statistically evaluated by one-way ANOVA followed by a Bonferroni post hoc test. Sex differences were evaluated by a two-way ANOVA. Values of p < 0.05 were considered significant.

3. Results 3.1. Gene expression changes in neurosteroid biosynthetic pathways One-way ANOVA followed by a Bonferroni post hoc test between the individual Braak stages (0–6) showed no significant differences in gene expression between Braak stages 0–2, between Braak 3 and 4 or between Braak 5 and 6. Subsequent analysis on the patients classified by clinical stages as BR 0–2 (no cognitive impairment n = 21), BR 3–4 (mild cognitive impairment, n = 14) and BR 5–6 (fully developed AD, n = 14), found statistically significant changes in transcript levels for several genes. DBI was found to be increased 1.3-fold in BR 3–4 and 1.5fold in BR 5–6 compared to BR 0–2 (Fig. 2A). AKR1C2 was increased 1.3-fold in BR 3–4 and 1.3-fold in BR 5–6 (Fig. 3A) and HSD17B1 was 2.1-fold increased in BR 5–6 compared to BR 0–2 (Fig. 2B). The enzyme aromatase showed increased mRNA levels in the BR 3–4 group (1.8-fold), and reached 2.4fold higher in BR 5–6 (Fig. 4A). Transcript levels of the ratelimiting enzyme SRD5A1 were 1.2-fold increased in BR 5–6 compared to BR 3–4 although this failed to reach significance (p = 0.051; data not shown).

Fig. 2. DBI and HSD17B1 gene expression changes in AD PFC. Results of qPCR are shown as relative mRNA expression for each gene normalized to the housekeeping gene values as described in Section 2. Group Braak stage is indicated as BR followed by the number (0–6). (A) DBI gene expression was significantly increased in BR 3–4 (* p < 0.05) and in BR 5–6 (*** p < 0.001) compared to BR 0–2. (B) HSD17B1 mRNA expression was increased in BR 5–6 (*** p < 0.001) compared to BR 0–2. (C) When patients classified by amyloid plaque load, HSD17B1 was increased in group C compared to groups A and B (* p < 0.05).

In order to understand the influence of the amyloid plaques on the neurosteroid biosynthetic pathway the data were reanalyzed by classifying the subjects on the basis of the amount of amyloid plaques present. Patients were divided into three groups according to plaque load (A = absent or low,

1968

S. Luchetti et al. / Neurobiology of Aging 32 (2011) 1964–1976

Fig. 3. AKR1C2 gene expression and immunocytochemical changes in AD PFC. (A) Graph shows AKR1C2 significantly increased gene expression in BR 3–4 (** p < 0.01) and BR 5–6 (* p < 0.05). (B–D) Photomicrographs of AKR1C immunostaining in PFC sections (layer 1–2) show increased staining in BR 3 (C) compared to BR 0 (B) and BR 6 (D). Arrows indicate astrocytes where expression is predominantly found. Scale bar = 25 ␮m.

n = 16; B = medium n = 11 and C = high plaque load n = 22). The same genes were found to differ significantly between groups as when subjects were classified by Braak stage. DBI and AKR1C2 mRNA levels were significantly increased by 1.5 and 1.3-fold respectively in group C compared to group A (Supplementary Fig. 1A and B). Furthermore, aromatase was 1.9-fold increased in group C compared to both groups A and B (Fig. 4B) while HSD17B1 was 2.2 and 1.8-fold increased in group C compared to groups A and B (Fig. 2C). 3.2. Changes in gene expression of GABA-A receptor subunits Expression of some of the GABA-A receptor subunits was significantly changed in AD PFC as follows: GABRA1 subunit was 1.5-fold reduced in BR 3–4 and BR 5–6 compared to BR 0–2 (Fig. 5A). GABRA2 was significantly decreased (1.3-fold) in BR 5–6 compared to BR 0–2 (Fig. 5B). GABRA4 was also 1.3-fold reduced both in BR 3–4 and BR 5–6 (Fig. 5C), while GABRD was 1.7-fold reduced in BR 5–6 compared to BR 0–2 (Fig. 5D).

As before, data were also analyzed by amyloid plaque load classification. Similarly to the Braak group analysis, expression of GABRA1 and GABRA4 subunits were significantly reduced (respectively by 1.6 and 1.5-fold) in group C compared to group A (Supplementary Fig. 2A and B). GABRD subunit was 1.4-fold reduced in group C compared to A and to B (Supplementary Fig. 2D). In addition, in the amyloid plaque classification analysis, expression of GAD1 and GABRB2 showed 1.5-fold reduced expression in group C compared to group A (Supplementary Fig. 2C and E). However, in this analysis, GABRA2 was not significantly changed. For genes studied other than those described above no significant changes were found. No significant differences were found in mRNA expression of the enzymes: AKR1C1, AKR1C2, CYP11A1 CYP17A1, CYP21A2, HSD3B1-2, HSD17B2, HSD17B3, HSD17B4, HSD17B6, HSD17B7, HSD17B10, STS, SULT2A1, SULT2B1. SRD5A2 mRNA was not detected in PFC while it was readily detected in a liver positive control. Furthermore, no significant differences were found for STAR, TSPO, the hormone receptors (AR,

S. Luchetti et al. / Neurobiology of Aging 32 (2011) 1964–1976

1969

Fig. 4. Aromatase gene expression and immunocytochemical changes in AD PFC. (A) Graph shows increased mRNA expression of the enzyme aromatase in BR 5–6 compared to BR 0–2 (*** p < 0.001). (B) When patients were classified by amyloid plaque load, aromatase mRNA was increased in group C compared to groups A and B (* p < 0.05). (C–H) Photomicrographs of aromatase immunostaining of the PFC in BR 0 (C and F), BR 3 (D and G) and BR 6 (E and H). In layer 1 (C, D and E) staining is increased in astrocytes. In layers 2–3 (F, G and H) aromatase expression in PFC increases with the progression of the disease and the cellular localization changes: it is found mainly in neurons in BR 3 (G) but more in astrocytes in BR 6 (H). Scale bar = 25 ␮m.

ER1, ER2a, PGR) and GABA receptor subunits (GABRB1, GABRG2, GABRE, GABRQ). No significant gender differences were found for any genes examined. No significant correlations were found with any other factor, including possible confounders like age, brain weight, CSF pH or postmortem delay. 3.3. Aromatase, AKR1C and SRD5A immunohistochemistry in PFC We performed IHC on paraffin sections of PFC in order to determine if the main findings of the qPCR were also evident at the protein level, and to determine the cellular localizations of the affected enzymes and their changes in the course of AD. The results were evaluated with a semi-quantitative analysis in which the intensity of the staining was scored in 5

categories ranging from absence (−) to very intense (++++). The staining intensities of Aromatase and AKR1C in neurons and astrocytes are shown in Supplementary Tables S2 and S3. Aromatase staining was greatly increased in layer 1 astrocytes in BR 5–6 compared to the other stages (Fig. 4C–E). Moderate staining of aromatase was found in neurons at all the Braak stages (Fig. 4F–H). As shown in Fig. 3B–D, AKR1C staining was present predominantly in astrocytes and was increased in BR 3–4 compared to BR 0–2 and BR 5–6 in agreement with the qPCR data. Expression of aromatase and AKR1C in astrocytes was confirmed by double stainings with glial fibrillary acidic protein (GFAP). GFAP-AKR1C (Fig. 6A, C and E) and GFAP-aromatase (Fig. 6B, D and F) colocalization was found both in the reactive astrocytes rich in GFAP and other astrocytes expressing moderate amounts of GFAP. SDR5A

1970

S. Luchetti et al. / Neurobiology of Aging 32 (2011) 1964–1976

moderately stained neurons but no clear increase was found in the BR 5–6 compared to other Braak stages (Supplementary Table S4).

4. Discussion 4.1. Neurosteroid biosynthetic pathway changes in AD The purpose of the study was to explore broadly the gene expression changes of biosynthesis of neurosteroids in human AD PCF, in order to identify pathways of genes that might be of importance in the transition from no to mild cognitive deficit and the fully developed stage of the disease. To our knowledge, we are the first to report changes in the pathway of neurosteroid synthesis in the human brain, during the course of AD. The qPCR data analysis of 37 genes showed that the neurosteroid biosynthesis pathway is altered in the PFC of AD patients during the course of the disease. Changes occur already in the early stages (BR 3–4) and become more pronounced in the late stages of AD (BR 5–6). While some caution is necessary in interpreting the results, changes in regulation of mRNAs for neurosteroid synthetic enzymes are indicative of changes in the production of the enzyme themselves, which would have consequences for the balance of synthesis or metabolism of the various neurosteroids. Likewise, the reduction in mRNA for neuromodulation-sensitive GABA receptor subunits suggests changes in production of these subunits, which would be expected to influence neurotransmission in the affected brain areas. However, further experiments are planned in the future to determine relative quantities of the steroids themselves in the brain and this would help with the full interpretation of the changes we observed. First, the increased mRNA expression of DBI that we found in BR 3–4 and BR 5–6 suggests a specific involvement of cholesterol transport in the development of the disease, starting already in the stage of mild cognitive impairment (BR 3–4). DBI promotes the loading of cholesterol into the mitochondrial inner membrane, both directly and via interaction with peripheral benzodiazepine receptor (PBR), recently renamed 18 kDa translocator protein (TSPO) (Papadopoulos et al., 2006). This is a rate-limiting step for the whole neurosteroid biosynthetic pathway (Costa et al., 1994; Papadopoulos et al., 1997). Our findings suggest that in early AD, there is an attempt to increase the biosynthesis of neurosteroids and neuroactive steroids through increased mitochondrial import of cholesterol. Previously, measurements of enzyme levels in cerebrospinal fluid (CSF) showed that DBI protein levels were increased in AD patients with severe cognitive impairment (Ferrarese et al., 1990) although Fig. 5. GABA-A receptor subunit gene expression changes in AD PFC. Expression of some GABA-A receptor subunits was significantly changed in the late stages of the disease: (A) GABRA1 subunit was reduced in BR 3–4 and BR 5–6 compared to BR 0–2 (*** p < 0.001). (B) GABRA2 was

significantly decreased in BR 5–6 compared to BR 0–2 (* p < 0.05). (C) GABRA4 was also reduced both in BR 3–4 and BR 5–6 (** p < 0.01). (D) GABRD was reduced in BR 5–6 compared to BR 0–2 (*** p < 0.001).

S. Luchetti et al. / Neurobiology of Aging 32 (2011) 1964–1976

1971

Fig. 6. Expression of aromatase and AKR1C in astrocytes in AD PFC. Photomicrographs of immunofluorescent staining for GFAP (red) and aromatase or AKR1C (green). (A and B) GFAP staining, (C) AKR1C staining, (D) aromatase staining, (E) GFAP-AKR1C colocalization and (F) GFAP-aromatase colocalization. Scale bar = 25 ␮m (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article).

other studies failed to find any increase (Barbaccia et al., 1986; Ferrero et al., 1988), a discrepancy which may be related to differences in AD cognitive impairment and age range of the subjects. Second, we found increased expression of AKR1C2 mRNA in PFC of AD patients starting from BR 3. In the

brain this enzyme directs the reduction of 5␣-DHP into allopregnanolone and 5␣ DH-DOC to 5␣ TH-DOC (Penning et al., 2003) while it has been suggested that the reverse reaction is preferentially mediated by HSD17B10 (He et al., 2005). As we were unable to find differences in HSD17B10 gene expression, we presume that the increased amount of

1972

S. Luchetti et al. / Neurobiology of Aging 32 (2011) 1964–1976

AKR1C2 directs the reaction towards the synthesis of allopregnanolone and 5␣ TH-DOC. Allopregnanolone levels in the PFC have been evaluated in several studies but the results were not concordant. WeillEngerer et al. (2002) found no differences between BR 4–6 AD and BR 1–3 control brains while Marx et al. (2006) recently found reduced levels of allopregnanolone in PFC of BR 4–5 AD compared to BR 1–3 male patients. Differences in internal standards used for gas chromatography mass spectrometry (GC/MS) measurement, small patient samples and lack of gender matching between groups of the first study may be involved in the discrepancy. In a previous study (Bossers et al., submitted for publication), a large increase in tangles was found in the patients at BR 3 compared to Braak 0, 1 and 2 which suggests BR 3 may be a critical stage in the development of the disease. Increased AKR1C2 expression beginning at BR 3, as we observed, would promote an increase in allopregnanolone synthesis. One possibility is that this is a compensatory mechanism of the PFC to raise the levels of allopregnanolone. Alternatively, the elevated amount of AKR1C2 may lead to the synthesis of other neuroactive steroids such as 3␣5␣-THDOC and 5␣-DHT (see Fig. 1). Future studies analyzing the amount of these compounds would be helpful to clarify this point. Third, we found increased expression of aromatase and HSD17B1 in AD patients during the course of the disease. By qPCR, aromatase was upregulated in BR 5–6 and also in amyloid stages B and C. By IHC we observed that aromatase expression appears to be upregulated in astrocytes in the later stages of AD. Aromatase converts androgens into estrogens and HSD17B1 converts estrone into estradiol. Therefore, in late stages of AD, the balance of the neurosteroid biosynthetic pathways in PFC appears to be shifted towards the synthesis of estradiol specifically in astroglia. Various studies in the literature are consistent with our results and suggest a neuroprotective action for estradiol in AD. Increased aromatase expression in reactive astroglia was found in response to brain injury provoked by neurotoxic and mechanical lesions (Garcia-Ovejero et al., 2005). Additionally, after brain injury, a neuroprotective effect for aromatase has been shown using aromatase deficient mice or mice treated with aromatase inhibitors (Garcia-Segura, 2008). Attenuation of tau hyperphosphorylation, induced by estradiol in HEK293 cells, suggests a preventive role for estradiol on AD-like tau pathology (Liu et al., 2008). In human primary culture of microglia derived from cortex, estradiol administration enhanced uptake of amyloid beta (Li et al., 2000). Aromatase-expressing astroglia were observed after brain injury induced by kainic acid administration or by introducing a cannula in the parenchyma in several rodent brain areas suggesting a potential for astrocytes to produce estradiol in response to injury. Estrogens are released by astrocytes as trophic factors for damaged neurons and also may be involved in compensatory restructuring of injured brain tissue (Garcia-Ovejero et al., 2005). Therefore, increased aromatase expression may be an attempt to

induce estrogen synthesis in PFC, potentially as a part of a rescue program that takes place in brain tissue affected by a neurodegenerative process such as AD. The upregulation of aromatase may be a general reparative response by reactive astrocytes (Garcia-Ovejero et al., 2005; Garcia-Segura, 2008). As these are common to brain injury and AD, the increased aromatase expression may be aimed at promoting neuroprotection, and this mechanism may be common to a variety of conditions where neuronal damage is present. On the other hand, aromatase upregulation was not observed in substantia nigra and caudate nucleus of Parkinson’s disease patients, suggesting that this phenomenon is not present in all neurodegenerative disorders (Luchetti et al., in preparation). 4.2. GABA-A receptor subunit gene expression changes Pregnane steroids (e.g. allopregnanolone and THDOC) are considered potent positive allosteric GABA-A modulators, enhancing GABA-induced Cl− currents. Studies from Biggio and colleagues (Biggio et al., 2006) indicated that their fluctuation may also modulate gene expression of some GABA-A receptor subunits (Concas et al., 1999; Follesa et al., 2005). This would suggest that changes in their levels in the brain may both directly and indirectly affect GABA-A receptor plasticity and consequently GABAergic transmission. To better understand if changes in steroidogenesis in AD affect the GABAergic system, we studied the gene expression of the GABA-A receptor subunits known to be modulated by neuroactive steroids. Studies indicate that all alpha subunits are modulated by pregnane steroids (Belelli et al., 2002). Beta subunits are not modulated while gamma 2 but not gamma 1 subunits are sensitive (Belelli et al., 2002; Herd et al., 2007). Post-synaptic GABA-A receptors mediate phasic inhibition in the CNS (Farrant and Nusser, 2005) and receptors containing ␣1 or ␣2 subunits are the most abundant (Mohler, 2006). In this context, the reduced amount of ␣1 and ␣2 GABA-A subunit expression we found in BR 3–4 and BR 5–6 suggests a loss of GABAergic phasic inhibition in PFC, starting from the pre-clinical phases of the disease. When patients were grouped based on amyloid plaques amount, decreased mRNA levels of GABRB2 and GAD but not of GABRA2 were found in amyloid stage C compare to A. However, reduced mRNA levels for the GABRB2 subunit and GAD, which is responsible for synthesis of GABA, would also suggest a deficit of GABAergic transmission in PFC occurring at the last stages of the disease. Changes in expression of GABA receptor subunits and GAD may partly reflect a loss of GABA-ergic neurons in late stage AD. However, the GABA system is known to be quite robust in AD as GABA concentrations, GABAergic neuron numbers and GAD enzyme activity in PFC were previously found not to be changed (Lowe et al., 1988; Mountjoy, 1984; Reinikainen et al., 1988; Rossor et al., 1982). Furthermore the changes that we found were in certain GABA subunits and

S. Luchetti et al. / Neurobiology of Aging 32 (2011) 1964–1976

not others, indicating specific changes in regulation of these subunits rather than a general downregulation or reduction due to cell loss. Interestingly, the ␦ subunit, found in the extra-synaptic receptors, increases the sensitivity of the GABA-A receptor to neuroactive steroids (Belelli et al., 2002; Stell et al., 2003). In combination with ␣4 and ␤2, it is implicated in the generation of tonic inhibition in rat and mouse thalamo-cortical neurons (Farrant and Nusser, 2005). In addition, its deletion in transgenic mice, if associated with ␣4 loss, reduces the tonic receptor activation in granule cells of the dentate gyrus by 3␣5␣-THDOC, leading to increased neuronal excitability (Stell et al., 2003). From this perspective the downregulation of GABA-A receptor subunits ␣4 and ␦ found during the last stages of AD would suggest that a reduced tonic activity may occur in PFC. Consistent with this, in a transgenic mouse model of Alzheimer’s (Busche et al., 2008), a proportion of neurons near to plaques were hyperactive, as revealed by imaging of calcium transients. This hyperactivity was reduced by diazepam suggesting impairment of GABA-ergic inhibition around the plaques. In support of the hypothesis that abnormalities of GABAergic transmission occur in the course of AD is the increased mRNA expression of DBI that we found in BR 3–4 and BR 5–6. DBI may inhibit GABAergic transmission by acting at the extracellular allosteric modulatory BZD binding site on GABA-A receptor alpha subunits (Costa et al., 1994; Papadopoulos et al., 1997). It is possible that DBI may participate in GABAergic modulation in PFC starting from mild cognitive impairment stages of AD. Changes in GABAergic activity may induce imbalances in inhibitory neurotransmission in the PFC and may play an important role in the cognition and behavioral alterations such as agitation, dysphoria, anxiety, euphoria, apathy, depression, disinhibition, irritability, and aberrant motor behavior occurring in AD (Birzniece et al., 2006; Lanctot et al., 2007). In this context, activation of synthesis of neuroactive steroids such as allopregnanolone and 3␣5␣-THDOC by the enzyme AKR1C2 may be part of a compensatory reaction to increase GABAergic activity at both the post-synaptic and extra-synaptic level in the course of AD. 4.3. Clinical relevance There is considerable evidence for beneficial effects of neurosteroids on neuronal viability and learning and memory processes from animal studies. Such effects were found after administration of estradiol in aging animals, while in studies using injury paradigms estradiol treatment increased neuronal survival and recovery (Melcangi et al., 2008; Schumacher et al., 2003). Progesterone, 5␣-DHP and allopregnanolone have also been shown in animal experiments to have neuroprotective effects on the aging brain (Melcangi et al., 2008; Schumacher et al., 2003), after traumatic or ischemic CNS injury (Djebaili et al., 2005; He et al., 2004;

1973

Sayeed et al., 2006, 2007; Thomas et al., 1999), in demyelinating diseases (Garay et al., 2007), during motoneuron degeneration (Gonzalez Deniselle et al., 2007) and to stimulate myelination by oligodendrocytes (Labombarda et al., 2006). In addition, testosterone was found to reduce the production of amyloidogenic ␤-amyloid peptide in vitro (Gouras et al., 2000). In AD triple transgenic mice studies, androgen depletion accelerates the development of AD-like neuropathology, increasing the accumulation of ␤-amyloid, suggesting that a similar mechanism may underlie the increased risk for AD in men with low testosterone (Rosario et al., 2006). Furthermore, estradiol combined with progesterone affect the expression of beta amyloid and tau proteins respectively, suggesting a potential of these compounds to regulate different aspects of the neuropathological process (Carroll et al., 2007). Therefore, several pre-clinical studies have shown strong evidence for a role for sex steroids as therapeutic agents to prevent neurodegenerative disease and promote successful aging (Schumacher et al., 2003). In humans, reduced blood or brain levels of estrogens in women after the menopause have been associated with a decline of cognitive performances in AD (Manly et al., 2000; Rosario et al., 2009). Clinical studies of hormone replacement or estrogen replacement therapy have had, however, mixed results with the largest study being inconclusive (Mulnard, 2005). In our experiments, aromatase and HSD17B1 start to be upregulated in the PFC from BR 3–4 suggesting that estrogens are already synthesized in the brain in the stage of mild cognitive impairment. This supports the idea that administration of estrogens may be beneficial before these stages, probably during the peri-menopause when the treatment would have highest efficacy (Garcia-Ovejero et al., 2005; Genazzani et al., 2007; Lethaby et al., 2008). Few studies have also shown reduced levels of testosterone in the aging brain as well as in AD male patients (Rosario et al., 2009; Weill-Engerer et al., 2002). Our data also suggest a possible role for testosterone metabolites such as 5␣-DHT or androstanediol in AD, as indicated by the increase in AKR1C2 gene expression. Testosterone represents a potential treatment for AD in men due not only to its own neuroprotective properties but potentially also through its aromatization into estrogens. Testosterone replacement therapy improved cognitive scores in healthy men (Cherrier et al., 2005; Leichtnam et al., 2006), but had minimal effects on cognition in the development of the disease (Hogervorst, 2008; Lu et al., 2006). Therefore, as with HRT in women, testosterone replacement therapy in aging men may be more beneficial if provided before the pathological process is too advanced. Increasing attention has been given to progesterone and allopregnanolone as anti-apoptotic, neuroprotective and neuroregenerative agents (Wang et al., 2007). Indeed, our finding of increased expression of AKR1C2 and unchanged HSD17B10 suggests a role of allopregnanolone synthesis

1974

S. Luchetti et al. / Neurobiology of Aging 32 (2011) 1964–1976

in AD as a neuroprotective or neurotransmitter modulating factor. However, administration of allopregnanolone or its precursors progesterone or medroxyprogesterone had deleterious effects (Kask et al., 2008; Shumaker et al., 2003; van Wingen et al., 2007). Compared to the in vivo and in vitro studies from Wang et al. (2007), which were focused on looking at the effects on neurogenesis 24 h post-injection, all of these studies used different paradigms. In the study from Kask et al. (2008) healthy women were treated intravenously with allopregnanolone immediately before cognitive testing, while in the study of van Wingen et al. (2007), 400 mg progesterone was orally administered 1–3 h before cognitive testing and MRI evaluation. In the clinical trial of Shumaker et al. (2003), postmenopausal women were chronically administered with medroxyprogesterone, a synthetic progestin, which is metabolized differently to progesterone in the brain (Pluchino et al., 2006) and has not shown neuroprotective effects on the nervous system (Nilsen and Brinton, 2003). The exact steroid compound used, the dosage, the treatment regimen and the outcome measures are likely to be important variables in determining the effects of neurosteroid treatment, and careful optimization is likely to be required to achieve positive results. Further clinical studies in AD are necessary to elucidate the potential of progesterone and allopregnanolone as a therapeutic agent in neurodegenerative diseases. On the other hand, the therapeutic potential of neurosteroids in the human nervous system is highlighted by the fact that progesterone has been shown to be potentially beneficial in two phase II clinical trials to treat moderate to severe traumatic brain injury (TBI) (Wright et al., 2007; Xiao et al., 2008). These positive results, confirming the numerous preclinical findings of neuroprotective properties in the injured CNS (Stein et al., 2008), make this a promising therapeutic strategy in neurodegenerative diseases too. In summary these data provide insight into the regulation of neurosteroid and neuroactive steroid synthesis and their modulatory activity on GABA-A function in the course of AD. Our findings of increased gene expression of some key enzymes suggest a biosynthetic pathway favoring the synthesis of estradiol and allopregnanolone. These compounds exert either a neuroprotective action or have modulatory effects on GABA-A receptors, suggesting a compensatory mechanism which may partly protect the PFC from development of the disease. In addition, the reduced gene expression of some synaptic and extra-synaptic GABA-A subunits suggests a deficit of modulation of GABA-A receptors by neuroactive steroids, which may play a role in the cognitive impairment and neuropsychiatric characteristics of this disease.

Disclosure statement None of the authors have reported biomedical financial interests or potential conflicts of interest in this work.

Acknowledgments Authors would like to thank, R. Balesar, B. Fisser, A. Sluiter, U. Unmehopa, J.J. Van Heerikhuize for technical assistance, Dr. M. Hofman for statistical advises, Dr. M.R.J. Mason and Dr. W. Kamphuis for critical comments of the paper. S.L. was supported by funds from the Istituto Dermopatico Immacolata (IDI-IRCCS to Prof. G. Frajese) and the Santa Lucia Foundation (Santa Lucia-IRCCS to Prof G. Bernardi), Rome, Italy. Experimental materials were funded by the Netherlands Institute for Neuroscience to D.F.S. We thank the Netherlands Brain Bank for providing the human material. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.neurobiolaging. 2009.12.014. References Bancher, C., Jellinger, K., Lassmann, H., Fischer, P., Leblhuber, F., 1996. Correlations between mental state and quantitative neuropathology in the Vienna Longitudinal Study on Dementia. Eur. Arch. Psychiatry Clin. Neurosci. 246 (3), 137–146. Barbaccia, M.L., Costa, E., Ferrero, P., Guidotti, A., Roy, A., Sunderland, T., Pickar, D., Paul, S.M., Goodwin, F.K., 1986. Diazepam-binding inhibitor. A brain neuropeptide present in human spinal fluid: studies in depression, schizophrenia, and Alzheimer’s disease. Arch. Gen. Psychiatry 43 (12), 1143–1147. Baulieu, E.E., 1998. Neurosteroids: a novel function of the brain. Psychoneuroendocrinology 23 (8), 963–987. Behl, C., 2002. Oestrogen as a neuroprotective hormone. Nat. Rev. Neurosci. 3 (6), 433–442. Belelli, D., Lambert, J.J., 2005. Neurosteroids: endogenous regulators of the GABA(A) receptor. Nat. Rev. Neurosci. 6 (7), 565–575. Belelli, D., Casula, A., Ling, A., Lambert, J.J., 2002. The influence of subunit composition on the interaction of neurosteroids with GABA(A) receptors. Neuropharmacology 43 (4), 651–661. Bialek, M., Zaremba, P., Borowicz, K.K., Czuczwar, S.J., 2004. Neuroprotective role of testosterone in the nervous system. Pol. J. Pharmacol. 56 (5), 509–518. Biggio, G., Follesa, P., Sanna, E., Purdy, R.H., Concas, A., 2001. GABAAreceptor plasticity during long-term exposure to and withdrawal from progesterone. Int. Rev. Neurobiol. (46), 207–241. Biggio, F., Gorini, G., Caria, S., Murru, L., Mostallino, M.C., Sanna, E., Follesa, P., 2006. Plastic neuronal changes in GABA(A) receptor gene expression induced by progesterone metabolites: in vitro molecular and functional studies. Pharmacol. Biochem. Behav. 84 (4), 545–554. Birzniece, V., Backstrom, T., Johansson, I.M., Lindblad, C., Lundgren, P., Lofgren, M., Olsson, T., Ragagnin, G., Taube, M., Turkmen, S., Wahlstrom, G., Wang, M.D., Wihlback, A.C., Zhu, D., 2006. Neuroactive steroid effects on cognitive functions with a focus on the serotonin and GABA systems. Brain Res. Rev. 51 (2), 212–239. Bossers, K., Meerhoff, G., Balesar, R., van Dongen, J.W., Kruse, C.G., Swaab, D.F., Verhaagen, J., 2009. Analysis of gene expression in Parkinson’s disease: possible involvement of neurotrophic support and axon guidance in dopaminergic cell death. Brain Pathol. 19 (1), 91–107. Braak, H., Braak, E., 1991. Neuropathological stageing of Alzheimer-related changes. Acta Neuropathol. 82 (4), 239–259.

S. Luchetti et al. / Neurobiology of Aging 32 (2011) 1964–1976 Busche, M.A., Eichhoff, G., Adelsberger, H., Abramowski, D., Wiederhold, K.H., Haass, C., Staufenbiel, M., Konnerth, A., Garaschuk, O., 2008. Clusters of hyperactive neurons near amyloid plaques in a mouse model of Alzheimer’s disease. Science 321 (5896), 1686–1689. Carroll, J.C., Rosario, E.R., Chang, L., Stanczyk, F.Z., Oddo, S., LaFerla, F.M., Pike, C.J., 2007. Progesterone and estrogen regulate Alzheimerlike neuropathology in female 3xTg-AD mice. J. Neurosci. 27 (48), 13357–13365. Cherrier, M.M., Matsumoto, A.M., Amory, J.K., Asthana, S., Bremner, W., Peskind, E.R., Raskind, M.A., Craft, S., 2005. Testosterone improves spatial memory in men with Alzheimer disease and mild cognitive impairment. Neurology 64 (12), 2063–2068. Cholerton, B., Gleason, C.E., Baker, L.D., Asthana, S., 2002. Estrogen and Alzheimer’s disease: the story so far. Drugs Aging 19 (6), 405–427. Ciriza, I., Azcoitia, I., Garcia-Segura, L.M., 2004. Reduced progesterone metabolites protect rat hippocampal neurones from kainic acid excitotoxicity in vivo. J. Neuroendocrinol. 16 (1), 58–63. Concas, A., Follesa, P., Barbaccia, M.L., Purdy, R.H., Biggio, G., 1999. Physiological modulation of GABA(A) receptor plasticity by progesterone metabolites. Eur. J. Pharmacol. 375 (1–3), 225–235. Costa, E., Cheney, D.L., Grayson, D.R., Korneyev, A., Longone, P., Pani, L., Romeo, E., Zivkovich, E., Guidotti, A., 1994. Pharmacology of neurosteroid biosynthesis. Role of the mitochondrial DBI receptor (MDR) complex. Ann. N. Y. Acad. Sci. 746, 223–242. Djebaili, M., Guo, Q., Pettus, E.H., Hoffman, S.W., Stein, D.G., 2005. The neurosteroids progesterone and allopregnanolone reduce cell death, gliosis, and functional deficits after traumatic brain injury in rats. J. Neurotrauma 22 (1), 106–118. Do Rego, J.L., Seong, J.Y., Burel, D., Leprince, J., Luu-The, V., Tsutsui, K., Tonon, M.C., Pelletier, G., Vaudry, H., 2009. Neurosteroid biosynthesis: enzymatic pathways and neuroendocrine regulation by neurotransmitters and neuropeptides. Front Neuroendocrinol.. Dubrovsky, B.O., 2005. Steroids, neuroactive steroids and neurosteroids in psychopathology. Prog. Neuropsychopharmacol. Biol. Psychiatry 29 (2), 169–192. Farrant, M., Nusser, Z., 2005. Variations on an inhibitory theme: phasic and tonic activation of GABA(A) receptors. Nat. Rev. Neurosci. 6 (3), 215–229. Ferrarese, C., Appollonio, I., Frigo, M., Meregalli, S., Piolti, R., Tamma, F., Frattola, L., 1990. Cerebrospinal fluid levels of diazepam-binding inhibitor in neurodegenerative disorders with dementia. Neurology 40 (4), 632–635. Ferrero, P., Benna, P., Costa, P., Tarenzi, L., Baggio, G., Bergamasco, B., Bergamini, L., 1988. Diazepam binding inhibitor-like immunoreactivity (DBI-LI) in human CSF. Correlations with neurological disorders. J. Neurol. Sci. 87 (2–3), 327–349. Follesa, P., Mostallino, M.C., Biggio, F., Gorini, G., Caria, S., Busonero, F., Murru, L., Mura, M.L., Sanna, E., Biggio, G., 2005. Distinct patterns of expression and regulation of GABA receptors containing the delta subunit in cerebellar granule and hippocampal neurons. J. Neurochem. 94 (3), 659–671. Francis, P.T., 2003. Glutamatergic systems in Alzheimer’s disease. Int. J. Geriatr. Psychiatry 18 (Suppl. 1), S15–S21. Garay, L., Deniselle, M.C., Lima, A., Roig, P., De Nicola, A.F., 2007. Effects of progesterone in the spinal cord of a mouse model of multiple sclerosis. J. Steroid Biochem. Mol. Biol. 107 (3–5), 228–237. Garcia-Ovejero, D., Azcoitia, I., DonCarlos, L.L., Melcangi, R.C., GarciaSegura, L.M., 2005. Glia-neuron crosstalk in the neuroprotective mechanisms of sex steroid hormones. Brain Res. Brain Res. Rev. 48 (2), 273–286. Garcia-Segura, L.M., 2008. Aromatase in the brain: not just for reproduction anymore. J. Neuroendocrinol. 20 (6), 705–712. Genazzani, A.R., Pluchino, N., Luisi, S., Luisi, M., 2007. Estrogen, cognition and female ageing. Hum. Reprod. Update 13 (2), 175–187. Gonzalez Deniselle, M.C., Garay, L., Gonzalez, S., Saravia, F., Labombarda, F., Guennoun, R., Schumacher, M., De Nicola, A.F., 2007. Progesterone modulates brain-derived neurotrophic factor and choline acetyltrans-

1975

ferase in degenerating Wobbler motoneurons. Exp. Neurol. 203 (2), 406–414. Gouras, G.K., Xu, H., Gross, R.S., Greenfield, J.P., Hai, B., Wang, R., Greengard, P., 2000. Testosterone reduces neuronal secretion of Alzheimer’s beta-amyloid peptides. Proc. Natl. Acad. Sci. U.S.A. 97 (3), 1202– 1205. He, J., Hoffman, S.W., Stein, D.G., 2004. Allopregnanolone, a progesterone metabolite, enhances behavioral recovery and decreases neuronal loss after traumatic brain injury. Restor. Neurol. Neurosci. 22 (1), 19–31. He, X.Y., Wegiel, J., Yang, Y.Z., Pullarkat, R., Schulz, H., Yang, S.Y., 2005. Type 10 17beta-hydroxysteroid dehydrogenase catalyzing the oxidation of steroid modulators of gamma-aminobutyric acid type A receptors. Mol. Cell Endocrinol. 229 (1–2), 111–117. Herd, M.B., Belelli, D., Lambert, J.J., 2007. Neurosteroid modulation of synaptic and extrasynaptic GABA(A) receptors. Pharmacol. Ther. 116 (1), 20–34. Hogervorst, E., 2008. Testosterone supplementation did not prevent cognitive decline or increase bone mineral density in older men. Evid. Based Med. 13 (3), 71. Kask, K., Backstrom, T., Nilsson, L.G., Sundstrom-Poromaa, I., 2008. Allopregnanolone impairs episodic memory in healthy women. Psychopharmacology (Berl) 199 (2), 161–168. Labombarda, F., Gonzalez, S., Gonzalez Deniselle, M.C., Garay, L., Guennoun, R., Schumacher, M., De Nicola, A.F., 2006. Progesterone increases the expression of myelin basic protein and the number of cells showing NG2 immunostaining in the lesioned spinal cord. J. Neurotrauma 23 (2), 181–192. Lanctot, K.L., Herrmann, N., Rothenburg, L., Eryavec, G., 2007. Behavioral correlates of GABAergic disruption in Alzheimer’s disease. Int. Psychogeriatr. 19 (1), 151–158. Leichtnam, M.L., Rolland, H., Wuthrich, P., Guy, R.H., 2006. Testosterone hormone replacement therapy: state-of-the-art and emerging technologies. Pharm. Res. 23 (6), 1117–1132. Lethaby, A., Hogervorst, E., Richards, M., Yesufu, A., Yaffe, K., 2008. Hormone replacement therapy for cognitive function in postmenopausal women. Cochrane Database Syst. Rev. (1), CD003122. Li, R., Shen, Y., Yang, L.B., Lue, L.F., Finch, C., Rogers, J., 2000. Estrogen enhances uptake of amyloid beta-protein by microglia derived from the human cortex. J. Neurochem. 75 (4), 1447–1454. Lin, H.K., Steckelbroeck, S., Fung, K.M., Jones, A.N., Penning, T.M., 2004. Characterization of a monoclonal antibody for human aldo-keto reductase AKR1C3 (type 2 3alpha-hydroxysteroid dehydrogenase/type 5 17beta-hydroxysteroid dehydrogenase); immunohistochemical detection in breast and prostate. Steroids 69 (13–14), 795–801. Liu, X.A., Zhu, L.Q., Zhang, Q., Shi, H.R., Wang, S.H., Wang, Q., Wang, J.Z., 2008. Estradiol attenuates tau hyperphosphorylation induced by upregulation of protein kinase-A. Neurochem. Res. 33 (9), 1811– 1820. Lowe, S.L., Francis, P.T., Procter, A.W., Palmer, A.M., Davison, A.N., Bowen, D.M., 1988. Gamma-aminobutyric acid concentration in brain tissue at two stages of Alzheimer’s disease. Brain 111 (Pt 4), 785– 799. Lu, P.H., Masterman, D.A., Mulnard, R., Cotman, C., Miller, B., Yaffe, K., Reback, E., Porter, V., Swerdloff, R., Cummings, J.L., 2006. Effects of testosterone on cognition and mood in male patients with mild Alzheimer disease and healthy elderly men. Arch. Neurol. 63 (2), 177–185. Manly, J.J., Merchant, C.A., Jacobs, D.M., Small, S.A., Bell, K., Ferin, M., Mayeux, R., 2000. Endogenous estrogen levels and Alzheimer’s disease among postmenopausal women. Neurology 54 (4), 833–837. Marx, C.E., Trost, W.T., Shampine, L.J., Stevens, R.D., Hulette, C.M., Steffens, D.C., Ervin, J.F., Butterfield, M.I., Blazer, D.G., Massing, M.W., Lieberman, J.A., 2006. The neurosteroid allopregnanolone is reduced in prefrontal cortex in Alzheimer’s disease. Biol. Psychiatry 60 (12), 1287–1294. Melcangi, R.C., Garcia-Segura, L.M., Mensah-Nyagan, A.G., 2008. Neuroactive steroids: state of the art and new perspectives. Cell Mol. Life Sci. 65 (5), 777–797.

1976

S. Luchetti et al. / Neurobiology of Aging 32 (2011) 1964–1976

Mellon, S.H., Vaudry, H., 2001. Biosynthesis of neurosteroids and regulation of their synthesis. Int. Rev. Neurobiol. (46), 33–78. Mohler, H., 2006. GABA(A) receptor diversity and pharmacology. Cell Tissue Res. 326 (2), 505–516. Mountjoy, C.Q., 1984. Correlation of cortical cholinergic and GABA deficits with quantitative neuropathological findings in senile dementia. Brain 107 (2), 507–518. Mulnard, R.A., 2005. Methodological issues in estrogen treatment trials for Alzheimer’s disease. Ann. N. Y. Acad. Sci. 1052, 173–181. Nilsen, J., Brinton, R.D., 2003. Divergent impact of progesterone and medroxyprogesterone acetate (Provera) on nuclear mitogen-activated protein kinase signaling. Proc. Natl. Acad. Sci. U.S.A. 100 (18), 10506–10511. Papadopoulos, V., Amri, H., Boujrad, N., Cascio, C., Culty, M., Garnier, M., Hardwick, M., Li, H., Vidic, B., Brown, A.S., Reversa, J.L., Bernassau, J.M., Drieu, K., 1997. Peripheral benzodiazepine receptor in cholesterol transport and steroidogenesis. Steroids 62 (1), 21–28. Papadopoulos, V., Baraldi, M., Guilarte, T.R., Knudsen, T.B., Lacapere, J.J., Lindemann, P., Norenberg, M.D., Nutt, D., Weizman, A., Zhang, M.R., Gavish, M., 2006. Translocator protein (18 kDa): new nomenclature for the peripheral-type benzodiazepine receptor based on its structure and molecular function. Trends Pharmacol. Sci. 27 (8), 402–409. Penning, T.M., Jin, Y., Heredia, V.V., Lewis, M., 2003. Structure–function relationships in 3alpha-hydroxysteroid dehydrogenases: a comparison of the rat and human isoforms. J. Steroid Biochem. Mol. Biol. 85 (2–5), 247–255. Pfaffl, M.W., 2001. A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res. 29 (9), e45. Pike, C.J., Rosario, E.R., Nguyen, T.V., 2006. Androgens, aging, and Alzheimer’s disease. Endocrine 29 (2), 233–241. Pluchino, N., Luisi, M., Lenzi, E., Centofanti, M., Begliuomini, S., Freschi, L., Ninni, F., Genazzani, A.R., 2006. Progesterone and progestins: effects on brain, allopregnanolone and beta-endorphin. J. Steroid Biochem. Mol. Biol. 102 (1–5), 205–213. Reinikainen, K.J., Paljarvi, L., Huuskonen, M., Soininen, H., Laakso, M., Riekkinen, P.J., 1988. A post-mortem study of noradrenergic, serotonergic and GABAergic neurons in Alzheimer’s disease. J. Neurol. Sci. 84 (1), 101–116. Rosario, E.R., Chang, L., Stanczyk, F.Z., Pike, C.J., 2004. Age-related testosterone depletion and the development of Alzheimer disease. JAMA 292 (12), 1431–1432. Rosario, E.R., Carroll, J.C., Oddo, S., LaFerla, F.M., Pike, C.J., 2006. Androgens regulate the development of neuropathology in a triple transgenic mouse model of Alzheimer’s disease. J. Neurosci. 26 (51), 13384–13389. Rosario, E.R., Chang, L., Head, E.H., Stanczyk, F.Z., Pike, C.J., 2009. Brain levels of sex steroid hormones in men and women during normal aging and in Alzheimer’s disease. Neurobiol. Aging. Rossor, M.N., Garrett, N.J., Johnson, A.L., Mountjoy, C.Q., Roth, M., Iversen, L.L., 1982. A post-mortem study of the cholinergic and GABA systems in senile dementia. Brain 105 (Pt 2), 313–330. Sayeed, I., Guo, Q., Hoffman, S.W., Stein, D.G., 2006. Allopregnanolone, a progesterone metabolite, is more effective than progesterone in reducing cortical infarct volume after transient middle cerebral artery occlusion. Ann. Emerg. Med. 47 (4), 381–389. Sayeed, I., Wali, B., Stein, D.G., 2007. Progesterone inhibits ischemic brain injury in a rat model of permanent middle cerebral artery occlusion. Restor. Neurol. Neurosci. 25 (2), 151–159. Schumacher, M., Akwa, Y., Guennoun, R., Robert, F., Labombarda, F., Desarnaud, F., Robel, P., De Nicola, A.F., Baulieu, E.E., 2000. Steroid synthesis and metabolism in the nervous system: trophic and protective effects. J. Neurocytol. 29 (5–6), 307–326.

Schumacher, M., Weill-Engerer, S., Liere, P., Robert, F., Franklin, R.J., Garcia-Segura, L.M., Lambert, J.J., Mayo, W., Melcangi, R.C., Parducz, A., Suter, U., Carelli, C., Baulieu, E.E., Akwa, Y., 2003. Steroid hormones and neurosteroids in normal and pathological aging of the nervous system. Prog. Neurobiol. 71 (1), 3–29. Schumacher, M., Guennoun, R., Stein, D.G., De Nicola, A.F., 2007. Progesterone: therapeutic opportunities for neuroprotection and myelin repair. Pharmacol. Ther. 116 (1), 77–106. Shumaker, S.A., Legault, C., Rapp, S.R., Thal, L., Wallace, R.B., Ockene, J.K., Hendrix, S.L., Jones III, B.N., Assaf, A.R., Jackson, R.D., Kotchen, J.M., Wassertheil-Smoller, S., Wactawski-Wende, J., 2003. Estrogen plus progestin and the incidence of dementia and mild cognitive impairment in postmenopausal women: the Women’s Health Initiative Memory Study: a randomized controlled trial. JAMA 289 (20), 2651– 2662. Stein, D.G., Wright, D.W., Kellermann, A.L., 2008. Does progesterone have neuroprotective properties? Ann. Emerg. Med. 51 (2), 164–172. Stell, B.M., Brickley, S.G., Tang, C.Y., Farrant, M., Mody, I., 2003. Neuroactive steroids reduce neuronal excitability by selectively enhancing tonic inhibition mediated by delta subunit-containing GABAA receptors. Proc. Natl. Acad. Sci. U.S.A. 100 (24), 14439–14444. Stoffel-Wagner, B., 2001. Neurosteroid metabolism in the human brain. Eur. J. Endocrinol. 145 (6), 669–679. Thigpen, A.E., Silver, R.I., Guileyardo, J.M., Casey, M.L., McConnell, J.D., Russell, D.W., 1993. Tissue distribution and ontogeny of steroid 5 alphareductase isozyme expression. J. Clin. Invest. 92 (2), 903–910. Thomas, A.J., Nockels, R.P., Pan, H.Q., Shaffrey, C.I., Chopp, M., 1999. Progesterone is neuroprotective after acute experimental spinal cord trauma in rats. Spine (Phila Pa 1976) 24 (20), 2134–2138. van Wingen, G., van Broekhoven, F., Verkes, R.J., Petersson, K.M., Backstrom, T., Buitelaar, J., Fernandez, G., 2007. How progesterone impairs memory for biologically salient stimuli in healthy young women. J. Neurosci. 27 (42), 11416–11423. Vandesompele, J., De Preter, K., Pattyn, F., Poppe, B., Van, R.N., De Paepe, A., Speleman, F., 2002. Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol. 3 (7), RESEARCH0034. Vongher, J.M., Frye, C.A., 1999. Progesterone in conjunction with estradiol has neuroprotective effects in an animal model of neurodegeneration. Pharmacol. Biochem. Behav. 64 (4), 777–785. Wang, J.M., Irwin, R.W., Liu, L., Chen, S., Brinton, R.D., 2007. Regeneration in a degenerating brain: potential of allopregnanolone as a neuroregenerative agent. Curr. Alzheimer Res. 4 (5), 510–517. Weill-Engerer, S., David, J.P., Sazdovitch, V., Liere, P., Eychenne, B., Pianos, A., Schumacher, M., Delacourte, A., Baulieu, E.E., Akwa, Y., 2002. Neurosteroid quantification in human brain regions: comparison between Alzheimer’s and nondemented patients. J. Clin. Endocrinol. Metab. 87 (11), 5138–5143. Wright, D.W., Kellermann, A.L., Hertzberg, V.S., Clark, P.L., Frankel, M., Goldstein, F.C., Salomone, J.P., Dent, L.L., Harris, O.A., Ander, D.S., Lowery, D.W., Patel, M.M., Denson, D.D., Gordon, A.B., Wald, M.M., Gupta, S., Hoffman, S.W., Stein, D.G., 2007. ProTECT: a randomized clinical trial of progesterone for acute traumatic brain injury. Ann. Emerg. Med. 49 (4), 391–402. Xiao, G., Wei, J., Yan, W., Wang, W., Lu, Z., 2008. Improved outcomes from the administration of progesterone for patients with acute severe traumatic brain injury: a randomized controlled trial. Crit. Care 12 (2), R61. Yague, J.G., Munoz, A., de Monasterio-Schrader, P., Defelipe, J., GarciaSegura, L.M., Azcoitia, I., 2006. Aromatase expression in the human temporal cortex. Neuroscience 138 (2), 389–401.