Possible role of pineal allopregnanolone in Purkinje … et al. PNAS...Possible role of pineal...

11
Possible role of pineal allopregnanolone in Purkinje cell survival Shogo Haraguchi a,b , Sakurako Hara a , Takayoshi Ubuka a , Masatoshi Mita b , and Kazuyoshi Tsutsui a,1 a Laboratory of Integrative Brain Sciences, Department of Biology and Center for Medical Life Science, Waseda University, Tokyo 162-8480, Japan; and b Department of Biology, Faculty of Education, Tokyo Gakugei University, Tokyo 184-8501, Japan Edited by Bruce S. McEwen, The Rockefeller University, New York, NY, and approved November 5, 2012 (received for review June 25, 2012) It is believed that neurosteroids are produced in the brain and other nervous systems. Here, we show that allopregnanolone (ALLO), a neurosteroid, is exceedingly produced in the pineal gland com- pared with the brain and that pineal ALLO acts on the Purkinje cell, a principal cerebellar neuron, to prevent apoptosis in the juvenile quail. We rst demonstrated that the pineal gland is a major organ of neurosteroidogenesis. A series of experiments using molecular and biochemical techniques has further demonstrated that the pineal gland produces a variety of neurosteroids de novo from cholesterol in the juvenile quail. Importantly, ALLO was far more actively produced in the pineal gland than in the brain. Pinealectomy (Px) decreased ALLO concentration in the cerebellum and induced apoptosis of Purkinje cells, whereas administration of ALLO to Px quail chicks prevented apoptosis of Purkinje cells. We further found that Px signicantly increased the number of Purkinje cells that expressed active caspase-3, a key protease in apoptotic pathway, and daily injection of ALLO to Px quail chicks decreased the number of Purkinje cells expressing active caspase-3. These results indicate that the neuroprotective effect of pineal ALLO is associated with the decrease in caspase-3 activity during the early stage of neuronal development. We thus provide evidence that the pineal gland is an important neurosteroidogenic organ and that pineal ALLO may be involved in Purkinje cell survival during development. This is an im- portant function of the pineal gland in the formation of neuronal circuits in the developing cerebellum. steroid | melatonin | HPLC | gas chromatography/mass spectrometry T he cerebellar cortex has been used as an excellent model to study synaptic formation and transmission of neural networks because it forms relatively simple neuronal networks compared with those of other brain regions. The Purkinje cell is a principal cerebellar neuron that integrates the process of memory and learning. It is known that in birds and mammals pinealectomy (Px) induces cell loss in the brain including Purkinje cells during de- velopment (1, 2). This observation suggests that a certain compo- nent(s) in the pineal gland contributes to Purkinje cell survival during development. It is now established that steroids can be synthesized de novo in the central and peripheral nervous systems. Such steroids are called neurosteroids,and de novo neurosteroidogenesis from cholesterol is a conserved property of the vertebrate brain (for reviews, see refs. 37). The Purkinje cell is known as a site of neurosteroidogenesis in the brain (for review, see ref. 8). This cerebellar neuron produces progesterone (PROG) and estradiol- 17β (E2) de novo from cholesterol during neonatal life, when cerebellar neuronal circuit formation occurs. Both PROG and E2 promote dendritic growth, spinogenesis, and synaptogenesis via each cognate nuclear receptor in the developing Purkinje cell (911). Allopregnanolone (ALLO; 3α,5α-tetrahydroprogester- one), a progesterone metabolite, is also synthesized in the cer- ebellum and facilitates Purkinje cell survival in the neonate (12). Until recently, we believed that neurosteroids are produced only in the brain and other nervous systems. On the basis of our previous study of chickens (13), however, the pineal gland may be a major organ in producing neurosteroids de novo from cholesterol. We thus hypothesized that pineal neurosteroid(s) may be involved in Purkinje cell survival during development. In the present study, we rst demonstrated the biosynthetic pathway of neurosteroids in the pineal gland of quail. We found that the pineal gland of quail chicks exceedingly synthesizes ALLO compared with the brain. Px decreased ALLO concentration in the cerebellum and induced apoptosis of Purkinje cells, whereas ad- ministration of ALLO to Px quail chicks prevented apoptosis of Purkinje cells. Administration of ALLO to Px quail chicks also decreased the expression of caspase-3, a key protease in the apo- ptotic pathway, in Purkinje cells. Thus, ALLO rescues developing Purkinje cells in the cerebellum from apoptosis. Our results further point to a role of pineal ALLO in preventing the death of de- veloping Purkinje cells. Results De Novo Pregnenolone Formation from Cholesterol in the Quail Chick Pineal Gland. Because pregnenolone (PREG) formation is the rst step in steroid synthesis (14, 15), we rst investigated whether the pineal gland synthesizes PREG from cholesterol de novo. We used pineal glands of male juvenile quail of posthatch day 7 (P7). Ste- roidogenic acute regulatory protein (StAR; gene name StAR) delivers cholesterol to the mitochondrial cytochrome P450 side- chain cleavage enzyme (P450scc; gene name Cyp11a) that produces PREG. Reverse transcription PCR (RT-PCR) analyses have dem- onstrated the expressions of StAR and Cyp11a mRNAs in the pineal gland (Fig. 1 A and B and Table S1). The amplied cDNA bands from the pineal gland were sequenced, and it was veried that they were authentic fragments of StAR (GenBank accession no. NM204686) and Cyp11a (GenBank accession no. NM001001756). To investigate PREG formation from cholesterol in the quail pineal gland, pineal glands of male chicks at P7 were incubated with tritiated cholesterol as a precursor, and the radioactive me- tabolite was analyzed by reversed-phase HPLC. As shown in Fig. 1C, a single radioactive peak was detected, and it exhibited the same retention time as that of tritiated PREG, a reference stan- dard, under the same chromatographic condition. The radioactive peak corresponding to PREG increased in a time-dependent manner from 0 to 24 h of incubation (Fig. 1C). In addition, 50 μM aminoglutethimide, an inhibitor of P450scc, reduced the ampli- tude of this peak (Fig. 1C). PREG synthesis in the pineal gland was further demonstrated by gas chromatography/mass spectrometry (GC-MS) as described previously (1621). Heptauorobutyrate derivatives of the au- thentic PREG and the metabolite of nonradioactive cholesterol Author contributions: S. Haraguchi and K.T. designed research; S. Haraguchi, S. Hara, and K.T. performed research; S. Haraguchi, T.U., and K.T. analyzed data; and S. Haraguchi, T.U., M.M., and K.T. wrote the paper. The authors declare no conict of interest. This article is a PNAS Direct Submission. Freely available online through the PNAS open access option. 1 To whom correspondence should be addressed. E-mail: [email protected]. This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. 1073/pnas.1210804109/-/DCSupplemental. 2111021115 | PNAS | December 18, 2012 | vol. 109 | no. 51 www.pnas.org/cgi/doi/10.1073/pnas.1210804109

Transcript of Possible role of pineal allopregnanolone in Purkinje … et al. PNAS...Possible role of pineal...

Page 1: Possible role of pineal allopregnanolone in Purkinje … et al. PNAS...Possible role of pineal allopregnanolone in Purkinje cell survival Shogo Haraguchia,b, Sakurako Haraa, Takayoshi

Possible role of pineal allopregnanolone in Purkinjecell survivalShogo Haraguchia,b, Sakurako Haraa, Takayoshi Ubukaa, Masatoshi Mitab, and Kazuyoshi Tsutsuia,1

aLaboratory of Integrative Brain Sciences, Department of Biology and Center for Medical Life Science, Waseda University, Tokyo 162-8480, Japan;and bDepartment of Biology, Faculty of Education, Tokyo Gakugei University, Tokyo 184-8501, Japan

Edited by Bruce S. McEwen, The Rockefeller University, New York, NY, and approved November 5, 2012 (received for review June 25, 2012)

It is believed that neurosteroids are produced in the brain and othernervous systems. Here, we show that allopregnanolone (ALLO),a neurosteroid, is exceedingly produced in the pineal gland com-pared with the brain and that pineal ALLO acts on the Purkinje cell,a principal cerebellar neuron, to prevent apoptosis in the juvenilequail. We first demonstrated that the pineal gland is a major organof neurosteroidogenesis. A series of experiments using molecularand biochemical techniques has further demonstrated that thepineal gland produces a variety of neurosteroids de novo fromcholesterol in the juvenile quail. Importantly, ALLO was far moreactively produced in the pineal gland than in the brain. Pinealectomy(Px) decreased ALLO concentration in the cerebellum and inducedapoptosis of Purkinje cells, whereas administration of ALLO to Pxquail chicks prevented apoptosis of Purkinje cells. We further foundthat Px significantly increased the number of Purkinje cells thatexpressed active caspase-3, a key protease in apoptotic pathway,and daily injection of ALLO to Px quail chicks decreased the numberof Purkinje cells expressing active caspase-3. These results indicatethat the neuroprotective effect of pineal ALLO is associatedwith thedecrease in caspase-3 activity during the early stage of neuronaldevelopment. We thus provide evidence that the pineal gland is animportant neurosteroidogenic organ and that pineal ALLO may beinvolved in Purkinje cell survival during development. This is an im-portant function of the pineal gland in the formation of neuronalcircuits in the developing cerebellum.

steroid | melatonin | HPLC | gas chromatography/mass spectrometry

The cerebellar cortex has been used as an excellent model tostudy synaptic formation and transmission of neural networks

because it forms relatively simple neuronal networks comparedwith those of other brain regions. The Purkinje cell is a principalcerebellar neuron that integrates the process of memory andlearning. It is known that in birds and mammals pinealectomy (Px)induces cell loss in the brain including Purkinje cells during de-velopment (1, 2). This observation suggests that a certain compo-nent(s) in the pineal gland contributes to Purkinje cell survivalduring development.It is now established that steroids can be synthesized de novo

in the central and peripheral nervous systems. Such steroids arecalled “neurosteroids,” and de novo neurosteroidogenesis fromcholesterol is a conserved property of the vertebrate brain (forreviews, see refs. 3–7). The Purkinje cell is known as a site ofneurosteroidogenesis in the brain (for review, see ref. 8). Thiscerebellar neuron produces progesterone (PROG) and estradiol-17β (E2) de novo from cholesterol during neonatal life, whencerebellar neuronal circuit formation occurs. Both PROG andE2 promote dendritic growth, spinogenesis, and synaptogenesisvia each cognate nuclear receptor in the developing Purkinje cell(9–11). Allopregnanolone (ALLO; 3α,5α-tetrahydroprogester-one), a progesterone metabolite, is also synthesized in the cer-ebellum and facilitates Purkinje cell survival in the neonate (12).Until recently, we believed that neurosteroids are produced

only in the brain and other nervous systems. On the basis of ourprevious study of chickens (13), however, the pineal gland may be amajor organ in producing neurosteroids de novo from cholesterol.

We thus hypothesized that pineal neurosteroid(s) may be involvedin Purkinje cell survival during development.In the present study, we first demonstrated the biosynthetic

pathway of neurosteroids in the pineal gland of quail. We foundthat the pineal gland of quail chicks exceedingly synthesizes ALLOcompared with the brain. Px decreasedALLO concentration in thecerebellum and induced apoptosis of Purkinje cells, whereas ad-ministration of ALLO to Px quail chicks prevented apoptosis ofPurkinje cells. Administration of ALLO to Px quail chicks alsodecreased the expression of caspase-3, a key protease in the apo-ptotic pathway, in Purkinje cells. Thus, ALLO rescues developingPurkinje cells in the cerebellum from apoptosis. Our results furtherpoint to a role of pineal ALLO in preventing the death of de-veloping Purkinje cells.

ResultsDe Novo Pregnenolone Formation from Cholesterol in the Quail ChickPineal Gland. Because pregnenolone (PREG) formation is the firststep in steroid synthesis (14, 15), we first investigated whether thepineal gland synthesizes PREG from cholesterol de novo.We usedpineal glands of male juvenile quail of posthatch day 7 (P7). Ste-roidogenic acute regulatory protein (StAR; gene name StAR)delivers cholesterol to the mitochondrial cytochrome P450 side-chain cleavage enzyme (P450scc; gene nameCyp11a) that producesPREG. Reverse transcription PCR (RT-PCR) analyses have dem-onstrated the expressions of StAR andCyp11amRNAs in the pinealgland (Fig. 1 A and B and Table S1). The amplified cDNA bandsfrom the pineal gland were sequenced, and it was verified thatthey were authentic fragments of StAR (GenBank accession no.NM204686) and Cyp11a (GenBank accession no. NM001001756).To investigate PREG formation from cholesterol in the quail

pineal gland, pineal glands of male chicks at P7 were incubatedwith tritiated cholesterol as a precursor, and the radioactive me-tabolite was analyzed by reversed-phase HPLC. As shown in Fig.1C, a single radioactive peak was detected, and it exhibited thesame retention time as that of tritiated PREG, a reference stan-dard, under the same chromatographic condition. The radioactivepeak corresponding to PREG increased in a time-dependentmanner from 0 to 24 h of incubation (Fig. 1C). In addition, 50 μMaminoglutethimide, an inhibitor of P450scc, reduced the ampli-tude of this peak (Fig. 1C).PREG synthesis in the pineal gland was further demonstrated

by gas chromatography/mass spectrometry (GC-MS) as describedpreviously (16–21). Heptafluorobutyrate derivatives of the au-thentic PREG and the metabolite of nonradioactive cholesterol

Author contributions: S. Haraguchi and K.T. designed research; S. Haraguchi, S. Hara, andK.T. performed research; S. Haraguchi, T.U., and K.T. analyzed data; and S. Haraguchi,T.U., M.M., and K.T. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.1To whom correspondence should be addressed. E-mail: [email protected].

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1210804109/-/DCSupplemental.

21110–21115 | PNAS | December 18, 2012 | vol. 109 | no. 51 www.pnas.org/cgi/doi/10.1073/pnas.1210804109

Page 2: Possible role of pineal allopregnanolone in Purkinje … et al. PNAS...Possible role of pineal allopregnanolone in Purkinje cell survival Shogo Haraguchia,b, Sakurako Haraa, Takayoshi

were prepared and applied to GC-MS analysis. On the basisof GC-MS–selected ion monitoring (SIM) analysis [mass/charge(m/z) 298] (16, 17), it was confirmed that the metabolite hada retention time that was identical to PREG (Fig. 1D).Immunohistochemical (IHC) analysis using anti-humanP450scc

antibody was conducted to analyze the cellular localization ofP450scc in the pineal gland. To confirm that this antibody recog-nizes galliformes P450scc protein, we first performedWestern blotanalysis on the extracts of COS-7 cells transfected with chickenCyp11a cDNA. A single immunoreactive band (60 kDa) was de-tected (Fig. 1E). When the pineal gland extracts were analyzed, asingle band (60 kDa) was detected at the same position (Fig. 1E).This band disappeared when the antibody was preadsorbed withchicken P450scc protein (Fig. 1E). Clear P450scc immunoreactivitywas also observed in the cells forming follicular structures in the

quail pineal gland (Fig. 1F). No immunoreactivity was observedwhen anti-P450scc antibody was preadsorbed with chicken P450sccprotein (Fig. 1G).

Neurosteroid Formations from PREG in the Quail Chick Pineal Gland.To investigate the biosynthetic pathway of neurosteroids in thepineal gland, pineal glands of male quail chicks at P7 were used forRT-PCR analyses to demonstrate the expressions of steroidogenicenzymes, such as cytochrome P450 7α-hydroxylase (P4507α; genename Cyp7b), 3β-hydroxysteroid dehydrogenase/Δ5-Δ4-isomerase(3β-HSD; gene name Hsd3b), 5α-reductase (gene name Srd5a),cytochrome P450 17α-hydroxylase/c17,20-lyase (P45017α,lyase;gene name Cyp17), 17β-hydroxysteroid dehydrogenase (17β-HSD;gene nameHsd17b), and cytochrome P450 aromatase (P450arom;gene name Cyp19). RT-PCR analyses showed cDNA bands of

Fig. 1. De novo PREG formation from cholesterol in the pineal gland. (A andB) RT-PCR analyses of StAR (A) and Cyp11a (B) mRNAs in the pineal gland ofmale quail chicks. Total RNA extracted from the tissue was reverse-transcribedwith (+) or without (−) reverse transcriptase (RTase) followed by PCR ampli-fication. (C) HPLC analysis of PREG formation from cholesterol in the pinealgland of male quail chicks. The pineal gland was incubated with [3H]choles-terol and homogenized after different incubation times, and then each ex-tract was subjected to HPLC. The pineal gland with [3H]cholesterol was alsoincubated with aminoglutethimide. The arrowheads indicate the elutionpositions of the substrate cholesterol (open arrowhead) and its metabolitePREG (solid arrowhead). (D) GC-SIM analysis of the metabolite of non-radioactive cholesterol by the pineal gland of male quail chicks. GC-SIM wastraced at m/z 298 for PREG as the metabolite of nonradioactive cholesterol.The arrowhead shows the peak corresponding to authentic PREG. (E) Westernblot analysis of the pineal gland and extracts of COS-7 cells transfected withchicken Cyp11a cDNA with the anti-human P450scc antibody. Anti-humanP450scc antibody was preadsorbed with chicken P450scc protein for control.(F) IHC of P450scc in the pineal gland. Arrowheads indicate immunoreactivecells. (G) IHC using P450scc antibody preadsorbed with a saturating concen-tration of chicken P450scc protein (10 μg/mL). (Scale bars, 10 μm.) Similarresults were obtained in repeated experiments using three different samples.

Fig. 2. Neurosteroid formation from PREG in the pineal gland. (A, D, and G)RT-PCR analyses of steroidogenic enzyme Cyp7b (A), Hsd3b (D), Srd5a (G),and Cyp17 (G) mRNAs in the pineal gland of male quail chicks. Total RNA wasreverse-transcribed with (+) or without (−) RTase followed by PCR amplifi-cation. (B, E, and H) HPLC analyses of neurosteroid formation in the pinealgland of male quail chicks. (B) The pineal gland homogenates were in-cubated with [3H]PREG + NADPH, and the extracts were subjected to HPLC.The pineal gland homogenates with [3H]PREG were also incubated withketoconazole. The arrowheads indicate elution positions of the substratePREG (open arrowhead) and its metabolite 7α- and/or 7β-OH PREG (solid ar-rowhead). (E) The pineal gland homogenates were incubated with [3H]PREG +NAD+, and the extracts were subjected to HPLC. The pineal gland homoge-nates with [3H]PREG were also incubated with trilostane. The arrowheadsindicate elution positions of the substrate PREG (open arrowhead) and itsmetabolite PROG (solid arrowhead). (H) The pineal gland homogenates wereincubated with [3H]PROG + NADPH, and the extracts were subjected to HPLC.The pineal gland homogenates with [3H]PROG were also incubated withketoconazole and finasteride. The arrowheads indicate elution positionsof the substrate PROG (open arrowhead) and its metabolites 5α-DHP, ALLOand/or EPI and AD (solid arrowheads). (C, F, I, and J) GC-SIM analyses of themetabolites of nonradioactive PREG or PROG by the pineal gland of malequail chicks. GC-SIM was traced at m/z 386 for 7α- and 7β-OH PREG (C) and atm/z 510 for PROG (F), metabolites of nonradioactive PREG, or at m/z 514 forALLO and EPI (I) and at m/z 482 for AD (J), metabolites of nonradioactivePROG. The arrowheads show the peaks corresponding to authentic 7α- and7β-OH PREG (C), PROG (F), ALLO and EPI (I), and AD (J). Similar results wereobtained in repeated experiments using three different samples.

Haraguchi et al. PNAS | December 18, 2012 | vol. 109 | no. 51 | 21111

NEU

ROSC

IENCE

Page 3: Possible role of pineal allopregnanolone in Purkinje … et al. PNAS...Possible role of pineal allopregnanolone in Purkinje cell survival Shogo Haraguchia,b, Sakurako Haraa, Takayoshi

Cyp7b, Hsd3b, Srd5a, Cyp17, Hsd17b, and Cyp19 (Fig. 2 A, D, andG; Fig. S1 A and D and Table S1). The amplified cDNA bandsin the pineal gland were sequenced, and it was verified that theywere authentic fragments of Cyp7b (GenBank accession no.AB329632), Hsd3b (GenBank accession no. FJ607242), Srd5a(GenBank accession no. XM001235446), Cyp17 (GenBank acces-sion no. AB281617),Hsd17b (GenBank accession no. NM204943),and Cyp19 (GenBank accession no. AF533667) cDNAs.To demonstrate neurosteroid formation in the quail pineal gland,

pineal gland homogenates of male chicks at P7 were incubated withtritiated PREG as a precursor, and radioactive metabolites wereanalyzed by reversed-phase HPLC as described previously (18–21).The radioactive metabolites corresponding to 7α- and/or 7β-hydroxypregnenolone (7α- and/or 7β-OH PREG; Fig. 2B) andPROG (Fig. 2E) increased in a time-dependent manner. Thesemetabolites were reduced by the treatment of ketoconazole, aninhibitor of cytochrome P450s (Cyps) (Fig. 2B), or by trilostane, aninhibitor of 3β-HSD (Fig. 2E). ALLO and/or epipregnanolone(EPI; 3β,5β-tetrahydroprogesterone), 5α-dihydroprogesterone (5α-DHP), and androstenedione (AD) were produced from the pre-cursor PROG (Fig. 2H). Ketoconazole and finasteride, an inhibitorof 5α-reductase, reduced the productions of these metabolites (Fig.2H). Testosterone (T) was produced from the precursor AD (Fig.S1B). 5α- and/or 5β-Dihydrotestosterone (5α- and/or 5β-DHT) andE2 were produced from the precursor T (Fig. S1E), and ketoco-nazole and finasteride reduced these products (Fig. S1E). Isoforms,such as 7α- and 7β-OH PREG; ALLO and EPI; and 5α- and 5β-DHT were not separated by the retention times in HPLC.

Neurosteroid formation in the pineal gland was further con-firmed by GC-SIM analysis as described previously (16–21).Derivatives of the authentic 7α- and 7β-OH PREG, PROG,ALLO, EPI, AD, T, 5α- and 5β-DHT, E2, and the metabolites ofnonradioactive steroids produced by the pineal gland were appliedtoGC-SIManalysis (m/z 386 for 7α- and 7β-OHPREG,m/z 510 forPROG, m/z 514 for ALLO and EPI, m/z 482 for AD, m/z 680 forT,m/z 486 for 5α- and 5β-DHT, andm/z 664 for E2). The isoforms7α- and 7β-OH PREG (Fig. 2C); ALLO and EPI (Fig. 2I); and 5α-and 5β-DHT (Fig. S1F) had different retention times in GC-MS,respectively, unlike HPLC. The neurosteroids produced in thepineal gland were thus identified as 7α- and 7β-OH PREG (Fig.2C), PROG (Fig. 2F), ALLO and EPI (Fig. 2I), AD (Fig. 2J),T (Fig. S1C), 5α- and 5β-DHT (Fig. S1F), and E2 (Fig. S1G).

ALLO and 7α-OH PREG Are Abundantly Synthesized and Released fromthe Quail Chick Pineal Gland. To identify major neurosteroids syn-thesized in the pineal gland, the pineal glands were cultured inmedium 199 with [3H]PREG as a precursor. PREG was convertedprimarily to 7α- and/or 7β-OH PREG shown in blue and to ALLOand/or EPI shown in red by the pineal gland (Fig. 3 A and B).We then compared the syntheses of these major neurosteroids

by HPLC and the expressions of their steroidogenic enzymemRNAs by real-time PCR in the pineal gland among both sexes ofadult and juvenile quail. The synthesis of 7α- and/or 7β-OHPREGand the expression ofCyp7bmRNAwere detected in both sexes ofadults and juvenile quail, but there was a clear age difference ineach parameter (Fig. 3C). 7α- and/or 7β-OH PREG synthesis andCyp7b mRNA expression were greater in juveniles than in adults

Fig. 3. ALLO and 7α-OH PREG are abundantly synthesized and released from the quail chick pineal gland. (A) Neurosteroids synthesized from PREG in thepineal gland of male quail chicks. The pineal glands were incubated with the substrate [3H]PREG, and the extracts were subjected to HPLC. The arrows in-dicate elution positions of the substrate PREG (open arrow) and its metabolites (solid arrows). (B) Comparison of the amount of neurosteroids synthesizedfrom PREG in the pineal gland of male quail chicks by HPLC (n = 8). **P < 0.01 vs. 5α-DHP, AD, or T; †††P < 0.001 vs. PROG, 5α- and/or 5β-DHT, or E2. (C)Comparisons of 7α- and/or 7β-OH PREG synthesis and Cyp7b mRNA expression in the pineal gland of adults and chicks of both sexes (n = 8). **P < 0.01 vs.adult. (D) Comparisons of 7α- and/or 7β-OH PREG synthesis and Cyp7b mRNA expression among the pineal gland, cerebellum, and diencephalon of quailchicks of both sexes (n = 8). **P < 0.01 vs. cerebellum or diencephalon. (E) Comparisons of 7α- and 7β-OH PREG releases from the pineal gland, cerebellum,and diencephalon of male quail chicks by GC-MS (n = 6). ***P < 0.001 vs. cerebellum or diencephalon. (F) Comparisons of ALLO and/or EPI synthesis and Srd5amRNA expression in the pineal gland of adults and chicks of both sexes (n = 8). **P < 0.01 vs. adult. (G) Comparisons of ALLO and/or EPI synthesis and Srd5amRNA expression among the pineal gland, cerebellum, and diencephalon of quail chicks of both sexes (n = 8). ***P < 0.001 vs. cerebellum or diencephalon.(H) Comparisons of ALLO and EPI releases from the pineal gland, cerebellum, and diencephalon of male quail chicks by GC-MS (n = 6). **P < 0.01 vs. cerebellumor diencephalon. (I) Identified biosynthetic pathways of neurosteroids in the pineal gland. Each column and vertical line in B–H represent the mean ± SEM.

21112 | www.pnas.org/cgi/doi/10.1073/pnas.1210804109 Haraguchi et al.

Page 4: Possible role of pineal allopregnanolone in Purkinje … et al. PNAS...Possible role of pineal allopregnanolone in Purkinje cell survival Shogo Haraguchia,b, Sakurako Haraa, Takayoshi

in both sexes (Fig. 3C). ALLO and/or EPI synthesis and Srd5amRNA expression were also greater in juveniles than in adults inboth sexes (Fig. 3F). The syntheses of these major neurosteroidsand the expressions of their steroidogenic enzyme mRNAs in thepineal gland were compared with those of different brain regions.7α- and/or 7β-OH PREG synthesis and Cyp7b mRNA expressionwere greater in the pineal gland than in the cerebellum anddiencephalon (Fig. 3D). ALLO and/or EPI synthesis and Srd5amRNAexpression were also greater in the pineal gland than in thecerebellum and diencephalon (Fig. 3G).To investigate neurosteroid release from the pineal gland, the

pineal glands were cultured in medium 199, and major neuro-steroids weremeasured byGC-MS. Significant amounts of 7α-OHPREG and ALLO were released from the pineal gland into theculture medium, unlike 7β-OH PREG and EPI (Fig. 3 E and H).In sum, 7α-OH PREG and ALLO were major products secretedby the pineal gland (Fig. 3I).

Pineal ALLO Saves Purkinje Cells from Cell Death in Px Quail Chicks.To investigate whether ALLO and 7α-OH PREG, major pinealneurosteroids, or melatonin, a major hormone of the pineal gland,are involved in Purkinje cell survival in the male juvenile quail, Pxwas performed at P2, and 7α-OH PREG and ALLO were injecteddaily, or quail chicks were s.c. implanted with a silastic plate con-taining melatonin (22) from P2 to P7. Px at P2 significantly de-creased the number of Purkinje cells in lobes I–VIII at P21 (Fig. 4A and B and Fig. S2A). Px decreased ALLO concentration in thecerebellum at P7 compared with control (Fig. 4C). Daily injectionof ALLO to Px quail chicks from P2 to P7 improved Purkinje cellsurvival in lobes II–VII at P21 (Fig. 4 A and B and Fig. S2A) andincreased ALLO concentration in the cerebellum at P7 comparedwith Px quail chicks (Fig. 4C). On the contrary, daily injection of7α-OH PREG or melatonin to Px quail chicks from P2 to P7 didnot increase Purkinje cell survival (Fig. 4 A and B and Fig. S2A).The effect of ALLO and 7α-OH PREG on Purkinje dendritic

length in lobes IV and IX were investigated among the fivegroups (control, Px, Px + 7α-OH PREG, Px + ALLO, and Px +melatonin) because the number of Purkinje cells in lobe IV wasmost vulnerable to Px and Px did not affect Purkinje cell num-bers in lobe IX (Fig. 4 A and B and Fig. S2A). In contrast to thenumber of Purkinje cells, there was no significant difference inthe maximal dendritic length of Purkinje cells among the fivegroups in lobes IV and IX (Fig. 4B and Figs. S2A and S3).To investigate whether melatonin influences the synthesis of

ALLO and/or EPI in the cerebellum and circulating ALLO inthe male juvenile quail, Px was performed at P2 and the quailchicks were s.c. implanted with a silastic plate containing mela-tonin or vehicle (22). Px or melatonin administration did notinfluence the synthesis of ALLO and/or EPI in the cerebellumand circulating ALLO in the male juvenile quail (Fig. S4).

Pineal ALLO Reaches the Adjacent Cerebellar Purkinje Cells byDiffusion. To better understand the mechanism of how pinealALLO reaches the adjacent cerebellar Purkinje cells, Px was per-formed at P2, and ALLO concentration was measured in therostral (lobes II–V) and caudal (lobes VI–IX) cerebellum of quailchicks. Px decreased ALLO concentration only in the rostral cer-ebellum and not in the caudal cerebellum 12 h after Px (Fig. 4E).To further investigate whether pineal ALLO reaches the cerebel-lum by diffusion in the juvenile quail at P2, [3H]ALLOwas injectedclose to the pineal lumen. [3H]ALLO content was significantlyhigher in the rostral cerebellum than in the caudal cerebellum 12and 24 h after [3H]ALLO injection (Fig. 4F).

Neuroprotective Effect of Pineal ALLO Is Associated with theDecrease in Caspase-3 Activity in Purkinje Cells. Finally, we in-vestigated the factor that mediates the neuroprotective effect ofpineal ALLO in Purkinje cells. It is well known that caspase-3

plays an important role in Purkinje cell death in vertebrates (23,24). Caspase-3 is a crucial mediator of apoptosis (23), including inbirds (24, 25). Accordingly, the expression of caspase-3 and thefragmentation of nuclear DNA were analyzed in Px quail chicksand Px plus ALLO-administrated quail chicks from P3 to P7.Px significantly increased the number of Purkinje cells that

expressed active caspase-3 in lobe IV of P3 and P5 quail chickscompared with control (Fig. 5 A and B). In contrast, daily in-jection of ALLO to Px chicks decreased the number of Purkinjecells expressing active caspase-3 at P3 and P5 (Fig. 5 A and B).On the other hand, the effect of Px or ALLO administration onactive caspase-3 expression was not observed at P7 (Fig. 5 A andB). Px or ALLO administration had no effect on active caspase-3expression in lobe IX (Fig. 5 C and D).DNA fragmentation was further investigated by terminal

deoxynucleotidyl transferase-mediated dUTP nick-end labeling(TUNEL) in Purkinje cells in lobe IV. There was no effect of Pxor ALLO administration on the number of TUNEL-positivePurkinje cells at P3 (Fig. 5 E and F). However, Px significantlyincreased the number of TUNEL-positive Purkinje cells com-pared with control at P5 and P7 (Fig. 5 E and F), and daily in-jection of ALLO to Px quail chicks decreased the number of

Fig. 4. Pineal ALLO saves Purkinje cells from cell death in Px quail chicks.Cerebella of control male quail chicks, Px at P2 male quail chicks, and Px at P2male quail chicks treated with daily injection of 7α-OH PREG (30 ng/5 μL) orALLO (30 ng/5 μL), or Px at P2 male quail chicks s.c. implanted with a silasticplate containing melatonin (10 mg/plate) from P2 to P7 were analyzed. (A)Number of Purkinje cells in the cerebellar lobes I–X at P21. Purkinje cellnumber is presented as the percentage of control (n = 12). *P < 0.05 or **P <0.01 vs. control; †P < 0.05 or ††P < 0.01 Px plus ALLO vs. Px. (B) Morphology ofPurkinje cells in the cerebellar lobes IV and IX at P21. M, molecular layer; P,Purkinje cell layer. (Scale bars, 40 μm.) (C) Effects of Px and daily injection ofALLO on ALLO concentration in the cerebellum of male quail chicks at P7(n = 12). **P < 0.01 vs. control; ††P < 0.01 Px plus ALLO vs. Px. (D) Compar-isons of ALLO concentration among the pineal gland, rostral cerebellum,and caudal cerebellum in P2 male quail chicks (n = 7). **P < 0.01 vs. rostralcerebellum or caudal cerebellum. (E) Effects of Px on ALLO concentrations inthe rostral and caudal cerebellum of male quail chicks (n = 7). *P < 0.05 vs.0 h. (F) Contents of [3H]ALLO in the rostral and caudal cerebellum after in-jection of [3H]ALLO close to the pineal lumen of male quail chicks at P2. [3H]ALLO (20 pmol) was injected close to the pineal lumen, and the content of[3H]ALLO in the rostral and caudal cerebellum was counted in a liquidscintillation counter (n = 7). ★★★P < 0.001 vs. 0 h; ☆☆P < 0.01 or ☆☆☆P < 0.001vs. 0 h; †P < 0.05 rostral cerebellum vs. caudal cerebellum. Each column andvertical line in A and C–F represent the mean ± SEM.

Haraguchi et al. PNAS | December 18, 2012 | vol. 109 | no. 51 | 21113

NEU

ROSC

IENCE

Page 5: Possible role of pineal allopregnanolone in Purkinje … et al. PNAS...Possible role of pineal allopregnanolone in Purkinje cell survival Shogo Haraguchia,b, Sakurako Haraa, Takayoshi

TUNEL-positive Purkinje cells compared with Px quail chicks atP5 and P7 (Fig. 5 E and F).

DiscussionA series of experiments using molecular and biochemical techni-ques has demonstrated that the pineal gland is a major neuro-steroidogenic organ that produces a variety of neurosteroids denovo from cholesterol in the juvenile quail, a demonstration of denovo neurosteroidogenesis in the pineal gland in a vertebrate class.Importantly, ALLOand 7α-OHPREGwere exceedingly producedin the pineal gland compared with the brain. These major pinealneurosteroids were abundantly released from the pineal gland (Fig.3E andH; Fig. S2). These results suggest that pineal ALLO and/or7α-OH PREG play important roles in the juvenile quail.It has been reported that Px induces Purkinje cell loss in the

developing cerebellum of chicks (2). In this study, we hypothesizedthat pineal ALLO and/or 7α-OH PREG may facilitate Purkinjecell survival in the juvenile quail. To test this hypothesis, we con-ducted a series of experiments using Px quail chicks. Px decreasedthe number of Purkinje cells in the cerebellum of quail chicks atP21, suggesting that pineal ALLO and/or 7α-OH PREG may beinvolved in the survival of Purkinje cells in the cerebellum duringdevelopment. Administration of ALLO but not of 7α-OH PREGto Px quail chicks from P2 to P7 facilitated Purkinje cell survival.Accordingly, ALLO secreted by the pineal glandmay contribute toPurkinje cell survival during development. Although 7α-OHPREG did not facilitate Purkinje cell survival, this neurosteroidenhances locomotor activities of quail (21) and chickens (13).Neuronal cell death is an essential feature of developing

nervous systems and neurodegenerative diseases (26–28). MostPurkinje cells in cerebellar organotypic culture die when sampledfrom 1- to 5-d-old mice, whereas they survive when sampledbefore or after these ages (29, 30). This critical period correlateswith a time window when Purkinje cells are engaged in intensesynaptogenesis and dendritic remodeling. Our present studysuggests that ALLO secreted by the pineal gland is involved inPurkinje cell survival during the critical period of intense syn-aptogenesis and dendritic remodeling.Px did not decrease the number of Purkinje cells in the pos-

terior lobes IX and X, unlike the anterior lobes I–VIII. It is wellknown that cerebellar defects are preferentially localized to theanterior (lobes I–V) and central (lobes VI and VII) lobes (31–33).

It was shown in the study of Niemann-Pick type C (NP-C) micethat ALLO is synthesized in the cerebellum and involved in Pur-kinje and granule cell survival in the developing cerebellum (12).According to Langmade et al. (32), Purkinje cell number was re-duced in npc1−/− mice, a model of NP-C disease, compared withWT mice. The greatest loss of Purkinje cells occurred in lobes I–IV, and the loss was not apparent in lobes IX and X (32). Thus,NP-C disease primarily affects the anterior lobes of the cerebellum,which were also affected by Px and ALLO treatment in this study.We found that the pineal gland is an important source of ALLO inthe cerebellum (Fig. 4 C–F and Fig. S2). The present and previousstudies (12, 32) suggest that pineal ALLO and cerebellar ALLOare involved in Purkinje cell survival during development (Fig. S2).It is known that in birds and mammals Px induces cell loss in

the brain, including loss of Purkinje cells during development (1,2). Although the neuroprotective action of pineal melatonin isknown in birds and mammals (34, 35), it has also been reportedthat melatonin does not fully ameliorate Purkinje cell loss duringdevelopment (36). These observations suggest that certain othercomponent(s) in the pineal gland may contribute to Purkinje cellsurvival during development. In this study, pineal melatonin didnot facilitate Purkinje cell survival during development (Fig. 4 Aand B) and did not affect cerebellar ALLO and/or EPI synthesisand circulating ALLO level in the male juvenile quail (Fig. S4).These results suggest that pineal ALLO but not melatonin actsas an important component of the pineal gland for Purkinje cellsurvival during development.Pineal ALLO concentration (more than 200 pmol/g tissue in

the intact pineal gland) was much higher than ALLO concen-tration in the cerebellum (around 40 pmol/g tissue in the intactcerebellum) (Fig. 4D). These concentrations of ALLO werephysiologically relevant compared with ALLO concentration (10–100 pmol/g tissue in the brain) in other vertebrates (37). Px de-creased ALLO concentration only in the rostral cerebellum 12 hafter Px (Fig. 4E). [3H]ALLO concentration increased in therostral cerebellum significantly more than in the caudal cerebel-lum 12 and 24 h after [3H]ALLO injection close to the pineallumen (Fig. 4F). These results suggest that pineal ALLO reachesthe adjacent cerebellar Purkinje cells by diffusion. In addition, theeffect of pineal ALLO on the prevention of Purkinje cell deathseems to be restricted to the anterior (lobes I–V) and central(lobes VI and VII) lobes of the cerebellum, which might also be

Fig. 5. Neuroprotective effect of pineal ALLO is associated with the decrease in caspase-3 activity in Purkinje cells. (A and C) Purkinje cells expressing activecaspase-3 in lobes IV (A) and IX (C) of control male quail chicks, Px male quail chicks, and Px plus ALLO male quail chicks at P3, P5, and P7. (B and D)Comparison of the number of Purkinje cells expressing active caspase-3 in lobes IV (B) and IX (D) among control male quail chicks, Px male quail chicks, and Pxplus ALLO male quail chicks at P3, P5, and P7 (n = 12). **P < 0.01 vs. control; †P < 0.05 or ††P < 0.01 Px plus ALLO vs. Px. (E) TUNEL-positive Purkinje cells in lobeIV of control male quail chicks, Px male quail chicks, and Px plus ALLO male quail chicks at P3, P5, and P7. (F) Comparison of the number of TUNEL-positivePurkinje cells in lobe IV among control male quail chicks, Px male quail chicks, and Px plus ALLO male quail chicks at P3, P5, and P7 (n = 12). *P < 0.05 vs.control; †P < 0.05 Px plus ALLO vs. Px. Each column and vertical line in B, D, and F represent the mean ± SEM.

21114 | www.pnas.org/cgi/doi/10.1073/pnas.1210804109 Haraguchi et al.

Page 6: Possible role of pineal allopregnanolone in Purkinje … et al. PNAS...Possible role of pineal allopregnanolone in Purkinje cell survival Shogo Haraguchia,b, Sakurako Haraa, Takayoshi

taken as evidence that ALLO acts simply by diffusion (Fig. S2).Nevertheless, we cannot exclude the possibility of an anatomicallink (e.g., blood vessels) from the pineal gland to the cerebellum.The intracellular signaling pathway exerting a neuroprotective

effect of ALLO in the brain was poorly understood, although theinvolvement of the GABAA receptor was suggested (12, 38). Ourresults suggested that pineal ALLO exerts antiapoptotic effectsin Purkinje cells by suppressing active caspase-3 expression in theearly stage of neuronal development. Px significantly increasedthe number of Purkinje cells that expressed active caspase-3 inlobe IV of P3 and P5 quail chicks, and daily injection of ALLO toPx quail chicks decreased the number of Purkinje cells expressingactive caspase-3 at these ages. The increase in active caspase-3expression by Px at P3 and P5 may have led to DNA fragmen-tation at P5 and P7. Apoptosis following DNA fragmentation isone of the mechanisms controlling neuronal cell number in thebrain. The massive loss of neurons in many regions of the de-veloping brain, including Purkinje cells, provides quantitative ad-justment of populations of interconnecting neurons (39). It istherefore considered that pineal ALLO may play important rolesin the prevention of Purkinje cell death and the formation ofneuronal circuits in the developing cerebellum.

Materials and MethodsJapanese quail, Coturnix japonica, at various ages were used in this study.Quail were incubated under daily photoperiods of 12-h light/12-h darkness

cycles with the light provided by white fluorescent lamps. Pineal glands wereisolated from the light-exposed animals at zeitgeber time 6. To analyze theexpressions of steroidogenic enzyme mRNAs, RT-PCR analyses were con-ducted according to our previous methods (10, 11, 13). To assess neuro-steroid formation in the quail pineal gland, conversions of substrate steroidswere measured biochemically by HPLC and GC-MS according to our previousmethods (18–21). To investigate whether pineal neurosteroids or melatoninare involved in Purkinje cell survival during development, Px at P2 and ad-ministration of ALLO, 7α-OH PREG, or melatonin from P2 to P7 were con-ducted using juvenile quail. Px and sham operation on P2 chicks wereperformed as described previously (22). Quantification of Purkinje cells wasperformed by counting the number of calbindin-immunoreactive cell bodiesin the Purkinje cell layer in each lobe. ALLO concentration after Px wasmeasured by GC-MS. Diffusion of [3H]ALLO administrated close to the pineallumen was measured by liquid scintillation counter. Furthermore, the factorthat mediates the neuroprotective effect of pineal ALLO in Purkinje cells wasinvestigated. Parasagittal cerebellar sections of chicks at P3, P5, and P7 wereanalyzed by IHC with an antibody against cleaved caspase-3, a key proteasein the apoptotic pathway (23–25), and by TUNEL to detect apoptotic cells asdescribed previously (22). Details are described in SI Materials and Methods.

ACKNOWLEDGMENTS. We thank Dr. Y. Fukada (University of Tokyo) for hisvaluable discussion. We also thank S. Ichimaru, M. Kusaka, and H. Takeda(Waseda University) for their technical assistance. This work was supportedin part by Grants-in-Aid for Scientific Research 22132004 and 22227002 (toK.T.) and by Grant-in-Aid for Japan Society for the Promotion of ScienceFellows 11J10683 (to S. Haraguchi) from the Ministry of Education, Culture,and Science (Tokyo).

1. Kilic E, Hermann DM, Isenmann S, Bähr M (2002) Effects of pinealectomy and mela-tonin on the retrograde degeneration of retinal ganglion cells in a novel model ofintraorbital optic nerve transection in mice. J Pineal Res 32(2):106–111.

2. Tunç AT, et al. (2006) Neonatal pinealectomy induces Purkinje cell loss in the cere-bellum of the chick: A stereological study. Brain Res 1067(1):95–102.

3. Baulieu EE (1997) Neurosteroids: Of the nervous system, by the nervous system, forthe nervous system. Recent Prog Horm Res 52:1–32.

4. Compagnone NA, Mellon SH (2000) Neurosteroids: Biosynthesis and function of thesenovel neuromodulators. Front Neuroendocrinol 21(1):1–56.

5. Do Rego JL, et al. (2009) Neurosteroid biosynthesis: Enzymatic pathways and neuro-endocrine regulation by neurotransmitters and neuropeptides. Front Neuroendocrinol30(3):259–301.

6. Mellon SH, Vaudry H (2001) Biosynthesis of neurosteroids and regulation of theirsynthesis. Int Rev Neurobiol 46:33–78.

7. Tsutsui K, Mellon SH (2006) Neurosteroids in the brain neuron: Biosynthesis, actionand medicinal impact on neurodegenerative disease. Central Nerv Syst Agents MedChem 6(1):73–82.

8. Tsutsui K (2008) Neurosteroids in the Purkinje cell: Biosynthesis, mode of action andfunctional significance. Mol Neurobiol 37(2–3):116–125.

9. Sakamoto H, Ukena K, Tsutsui K (2001) Effects of progesterone synthesized de novo inthe developing Purkinje cell on its dendritic growth and synaptogenesis. J Neurosci 21(16):6221–6232.

10. Sakamoto H, Mezaki Y, Shikimi H, Ukena K, Tsutsui K (2003) Dendritic growth andspine formation in response to estrogen in the developing Purkinje cell. Endocrinol-ogy 144(10):4466–4477.

11. Sasahara K, et al. (2007) Mode of action and functional significance of estrogen-in-ducing dendritic growth, spinogenesis, and synaptogenesis in the developing Pur-kinje cell. J Neurosci 27(28):7408–7417.

12. Griffin LD, Gong W, Verot L, Mellon SH (2004) Niemann-Pick type C disease involves dis-rupted neurosteroidogenesis and responds to allopregnanolone. Nat Med 10(7):704–711.

13. Hatori M, et al. (2011) Light-dependent and circadian clock-regulated activation ofsterol regulatory element-binding protein, X-box-binding protein 1, and heat shockfactor pathways. Proc Natl Acad Sci USA 108(12):4864–4869.

14. Tsutsui K, Yamazaki T (1995) Avian neurosteroids. I. Pregnenolone biosynthesis in thequail brain. Brain Res 678(1–2):1–9.

15. Zuber MX, Mason JI, Simpson ER, Waterman MR (1988) Simultaneous transfection ofCOS-1 cells with mitochondrial and microsomal steroid hydroxylases: Incorporation of asteroidogenic pathway into nonsteroidogenic cells. Proc Natl Acad Sci USA 85(3):699–703.

16. Liere P, et al. (2000) Validation of an analytical procedure to measure trace amountsof neurosteroids in brain tissue by gas chromatography-mass spectrometry. J Chro-matogr B Biomed Sci Appl 739(2):301–312.

17. Meffre D, et al. (2007) Steroid profiling in brain and plasma of male and pseudo-pregnant female rats after traumatic brain injury: Analysis by gas chromatography/mass spectrometry. Endocrinology 148(5):2505–2517.

18. Haraguchi S, Koyama T, Hasunuma I, Vaudry H, Tsutsui K (2010) Prolactin increasesthe synthesis of 7α-hydroxypregnenolone, a key factor for induction of locomotoractivity, in breeding male newts. Endocrinology 151(5):2211–2222.

19. Haraguchi S, et al. (2012) Acute stress increases the synthesis of 7α-hydroxypreg-nenolone, a new key neurosteroid stimulating locomotor activity, through cortico-sterone action in newts. Endocrinology 153(2):794–805.

20. Matsunaga M, Ukena K, Baulieu EE, Tsutsui K (2004) 7α-Hydroxypregnenolone acts asa neuronal activator to stimulate locomotor activity of breeding newts by means ofthe dopaminergic system. Proc Natl Acad Sci USA 101(49):17282–17287.

21. Tsutsui K, et al. (2008) 7α-Hydroxypregnenolone mediates melatonin action un-derlying diurnal locomotor rhythms. J Neurosci 28(9):2158–2167.

22. Ubuka T, Bentley GE, Ukena K, Wingfield JC, Tsutsui K (2005) Melatonin induces theexpression of gonadotropin-inhibitory hormone in the avian brain. Proc Natl Acad SciUSA 102(8):3052–3057.

23. Puig B, Ferrer I (2001) Cell death signaling in the cerebellum in Creutzfeldt-Jakobdisease. Acta Neuropathol 102(3):207–215.

24. Olkowski AA, et al. (2008) A study on pathogenesis of sudden death syndrome inbroiler chickens. Res Vet Sci 85(1):131–140.

25. Matsunaga E, et al. (2004) RGM and its receptor neogenin regulate neuronal survival.Nat Cell Biol 6(8):749–755.

26. Cheng XS, et al. (2011) Neuronal apoptosis in the developing cerebellum. Anat HistolEmbryol 40(1):21–27.

27. Eccles JC (1970) Neurogenesis and morphogenesis in the cerebellar cortex. Proc NatlAcad Sci USA 66(2):294–301.

28. Thomaidou D, Mione MC, Cavanagh JF, Parnavelas JG (1997) Apoptosis and its re-lation to the cell cycle in the developing cerebral cortex. J Neurosci 17(3):1075–1085.

29. Dusart I, Airaksinen MS, Sotelo C (1997) Purkinje cell survival and axonal regenerationare age dependent: An in vitro study. J Neurosci 17(10):3710–3726.

30. Ghoumari AM, Wehrlé R, Bernard O, Sotelo C, Dusart I (2000) Implication of Bcl-2 andCaspase-3 in age-related Purkinje cell death in murine organotypic culture: An in vitromodel to study apoptosis. Eur J Neurosci 12(8):2935–2949.

31. Hess EJ, Wilson MC (1991) Tottering and leaner mutations perturb transient de-velopmental expression of tyrosine hydroxylase in embryologically distinct Purkinjecells. Neuron 6(1):123–132.

32. Langmade SJ, et al. (2006) Pregnane X receptor (PXR) activation: A mechanism forneuroprotection in a mouse model of Niemann-Pick C disease. Proc Natl Acad Sci USA103(37):13807–13812.

33. Levitt P, Noebels JL (1981) Mutant mouse tottering: Selective increase of locus ceruleusaxons in a defined single-locus mutation. Proc Natl Acad Sci USA 78(7):4630–4634.

34. Cagnoli CM, Atabay C, Kharlamova E, Manev H (1995) Melatonin protects neuronsfrom singlet oxygen-induced apoptosis. J Pineal Res 18(4):222–226.

35. Espinar A, et al. (2000) Neuroprotection by melatonin from glutamate-induced ex-citotoxicity during development of the cerebellum in the chick embryo. J Pineal Res28(2):81–88.

36. Edwards RB, Manzana EJ, Chen WJ (2002) Melatonin (an antioxidant) does notameliorate alcohol-induced Purkinje cell loss in the developing cerebellum. AlcoholClin Exp Res 26(7):1003–1009.

37. Higashi T, Nagura Y, Shimada K, Toyo’oka T (2009) Studies on neurosteroids XXVI.Fluoxetine-evoked changes in rat brain and serum levels of neuroactive androgen,5α-androstane-3α,17β-diol. Biol Pharm Bull 32(9):1636–1638.

38. Hosie AM, Wilkins ME, da Silva HM, Smart TG (2006) Endogenous neurosteroidsregulate GABAA receptors through two discrete transmembrane sites. Nature444(7118):486–489.

39. Oppenheim RW, Prevette D, Yin QW, Collins F, MacDonald J (1991) Control of em-bryonic motoneuron survival in vivo by ciliary neurotrophic factor. Science 251(5001):1616–1618.

Haraguchi et al. PNAS | December 18, 2012 | vol. 109 | no. 51 | 21115

NEU

ROSC

IENCE

Page 7: Possible role of pineal allopregnanolone in Purkinje … et al. PNAS...Possible role of pineal allopregnanolone in Purkinje cell survival Shogo Haraguchia,b, Sakurako Haraa, Takayoshi

Supporting InformationHaraguchi et al. 10.1073/pnas.1210804109SI Materials and MethodsAnimals. Japanese quail, Coturnix japonica, at various ages wereused in this study. Quail were incubated under daily photoperiodsof 12-h light/12-h darkness cycles with the light provided by whitefluorescent lamps. Pineal glands were isolated from the light-ex-posed animals at zeitgeber time 6. The experimental protocol wasapproved in accordance with the Guide for the Care and Use ofLaboratory Animals of Waseda University.

PCR Protocol for Sex Determination. Sexes of juvenile quail weredetermined by PCR to detect genomic size differences of thechromo-helicase–DNA-binding protein (CHD) betweenW (male)and Z (female) chromosomes (1). To isolate genomic DNA, braintissue was lysed in 100 μL of digestion buffer [10mMTris·HCl (pH8.0)/150 mM NaCl/10 mM EDTA/0.1% SDS] with 2 μg of Pro-teinase K at 65 °C overnight, extracted with phenol/chloroform/isoamyl alcohol (25:24:1), and precipitated in ethanol. The primers(Table S1) bind both CHD-W and CHD-Z. PCR was performed at94 °C for 5 min for 1 cycle, followed by 94 °C for 1 min, 54 °C for1 min, and 72 °C for 1 min for 32 cycles. Males had one band at∼1,000 bp fromCHD-W, and females had two bands, one at∼1,000bp from CHD-W and the other at 267 bp from CHD-Z.

RT-PCR Analyses of Neurosteroidogenic Enzyme mRNAs. Total RNAwas extracted from the quail pineal gland with Sepazol-RNA ISuper (Nacalai Tesque) and reverse-transcribed (2–4). All PCRamplifications (for StAR, Cyp11a, Cyp7b, Hsd3b, Srd5a, Cyp17,Hsd17b, and Cyp19) were performed in a reaction mixture con-taining Bio Taq polymerase (Bioline), 0.2 mM dNTP, and 0.5%dimethyl sulfoxide using a thermal cycler (3, 4). Forward primersand reverse primers (Table S1) were designed according to thenucleotide sequence of quail and chicken steroidogenic enzymemRNAs. β-Actin, a housekeeping gene, was used as control. ThePCR condition was the following: 1 cycle of 5 min at 94 °C, 30cycles of 15 s at 94 °C, 15 s at 60 °C, 15 s at 72 °C, and, finally, 1cycle of 10 min at 72 °C.

Biochemical Analysis of Pregnenolone Formation from Cholesterol.To assess pregnenolone (PREG) formation from cholesterol inthe quail pineal gland, conversion of [3H]cholesterol (specific ac-tivity, 53.0 Ci/mmol; PerkinElmer) to [3H]PREG was measuredbiochemically using organ-cultured pineal glands. Biochemicalanalysis in this study was performed as described previously (3, 5,6). In brief, five pineal glands were cultured in medium 199 (me-dium 199 supplemented with 10 mMHepes-NaOH at pH 7.4, 100U/mL penicillin, and 100 μg/mL streptomycin; Invitrogen) con-taining 210 nM [3H]cholesterol, 2% (vol/vol) propylene glycol, 1%protease inhibitor mixture (Nacalai Tesque) for 0, 12, or 24 h at37 °C. The homogenates were incubated in a water-saturated at-mosphere (80% O2, 5% CO2) to maintain the pH at 7.4. Afterincubation, steroids were extracted by ethyl acetate and subjectedto HPLC analysis by using reversed-phase column, Capcell PakC18 MG (Shiseido). The HPLC was performed with an isocraticcondition of acetonitrile/isopropanol (60:40, vol/vol) at a flow rateof 0.3 mL/min. The eluate was counted in a flow scintillation an-alyzer (Radiomatic 525TR; PerkinElmer). Reference standardsof tritiated cholesterol and PREG were chromatographed to de-tect their elution positions. To confirm the involvement of P450sccin the formation of PREG, five pineal glands and [3H]cholesterolwere cultured with aminoglutethimide (Sigma), an inhibitor ofP450scc, at a final concentration of 50 μM (3).

Biochemical Analysis of Neurosteroids Formed from PREG. To assessneurosteroid formation from PREG in the quail pineal gland,conversions of substrate steroids [PREG, progesterone (PROG),androstenedione (AD), or testosterone (T)] were measured bio-chemicallyusingpinealglandhomogenates.Biochemicalanalysis inthis study was performed as described previously (4, 7–13). In brief,10 mg of pineal gland homogenates were incubated in PBS con-taining 70 nM [3H]PREG (specific activity, 22.9 Ci/mmol; Perki-nElmer) and NADPH or NAD+, 70 nM [3H]PROG (specificactivity, 96.6 Ci/mmol; PerkinElmer) and NADPH, 70 nM [3H]AD(specific activity, 98.2 Ci/mmol; PerkinElmer) and NADPH, or 70nM [3H]T (specific activity, 70 Ci/mmol; PerkinElmer) andNADPH for 0, 30, or 60 min at 37 °C. After incubation, neuro-steroids were extracted by ethyl acetate and subjected to HPLCanalysis using a reversed-phase column, LiChrospher 100 RP-18(Kanto). Tritiated steroids [7α-hydroxypregnenolone (7α-OHPREG), PROG, 5α-dihydroprogesterone (5α-DHP), allopreg-nanolone (ALLO), AD, T, 5α-dihydrotestosterone (5α-DHT), andestradiol-17β (E2)] were chromatographed as standards to detecttheir elution positions. The isoforms 7α- and 7β-OHPREG;ALLOand EPI; and 5α- and 5β-DHT had the same retention times inHPLC.To confirm the involvement of steroidogenic enzymes in theformation of neurosteroids, tritiated steroids were cultured with 50μM ketoconazole, an inhibitor of cytochrome P450s (Cyps) (12)(Sigma); 50 μM trilostane, an inhibitor of 3β-HSD (9) (Sigma); or50 μM finasteride, an inhibitor of 5α-reductase (14) (Sigma). All ofthe tritiated steroids were purchased from PerkinElmer.

Identification of Neurosteroids by Gas Chromatography/MassSpectrometry. Neurosteroids, produced in the quail pineal gland,were identified by gas chromatography/mass spectrometry (GC-MS) (GCMS-QP5000; Shimadzu) using a CP-Sil 5CB capillarycolumn (Varian) as described previously (8, 9, 12, 13, 15, 16). Inbrief, 400 mg of pineal gland homogenates were incubated in PBScontaining nonradioactive steroids for 60 min. The obtainedsamples were homogenized in an aliquot of methanol/H2O [75:25(vol/vol); 1 mL] on ice. The samples were passed through the C-18cartridge column, and the neurosteroid fractions were eluted withmethanol and dried. For the identification of neurosteroids in thepineal gland, heptafluorobutyrate (for PREG, PROG, ALLO,EPI, AD, T, 5α- and 5β-DHT, and E2) or trimethylsilyl ether (for7α- and 7β-OH PREG) derivatives of the neurosteroids wereprepared before GC-MS by treating the dried samples with hep-tafluorobutyric anhydride (Wako Pure Chemical) or bis(trime-thylsilyl)trifluoroacetamide (Wako Pure Chemical) for 30 min atroom temperature. For the identification of neurosteroids, GC-SIM analysis was conducted atm/z 386 for 7α- and 7β-OH PREG,m/z 510 for PROG, m/z 514 for ALLO and EPI, m/z 482 for AD,m/z 680 for T, m/z 486 for 5α- and 5β-DHT, and m/z 664 for E2.Importantly, the isoforms 7α- and 7β-OH PREG; ALLO and EPI;and 5α- and 5β-DHT had different retention times in GC-MS (7α-OH PREG at 18.3 min, 7β-OH PREG at 19.1 min; ALLO at 15.3min; EPI at 15.8 min; 5α-DHT at 14.8 min, and 5β-DHT at 15.9min), respectively, unlike HPLC.

Western Blot Analysis with P450scc Antibody. To confirm that themouse anti-human P450scc antibody (Abcam) recognizes galliformP450scc, Western blot analysis was performed as described pre-viously (8). Quail and chicken are both galliforms with predictedhigh interspecific DNA sequence conservation (17). ChickenP450scc and quail P450scc have almost identical structure (96.3%identity in the amino acid sequences). The full-length ORF of the

Haraguchi et al. www.pnas.org/cgi/content/short/1210804109 1 of 5

Page 8: Possible role of pineal allopregnanolone in Purkinje … et al. PNAS...Possible role of pineal allopregnanolone in Purkinje cell survival Shogo Haraguchia,b, Sakurako Haraa, Takayoshi

chicken Cyp11a was amplified from chicken pineal gland cDNAusing the chicken Cyp11a primers (Table S1) and subcloned intothe mammalian expression vector pcDNA3.1/V5-His-TOPO (In-vitrogen). Positive colonies were selected and subcultured, and theplasmidDNAswere purifiedby theWizard plus SVminiprepsDNApurification system (Promega). COS-7 cells were supplied from theRiken Cell Bank andmaintained inDMEM (Sigma) supplementedwith 10% (vol/vol) FBS, penicillin (50 U/mL), streptomycin (50μg/mL), and Hepes (10 mM, pH 7.4). Transfection was performedwith the TransFast transfection reagent (Promega) as describedpreviously (8, 13). After transfection, the cells were harvested,centrifuged (10,000 × g for 5 min at 4 °C), and stored at –80 °C.Electrophoresis of proteins (3 μg each) derived from the quail pi-neal gland and extracts of COS-7 cells transfected with chickenCyp11a cDNA was performed in 12.5% polyacrylamide gels. Aftertransferring to a polyvinylidene fluoride membrane, the membranewas incubated with anti-humanP450scc antibody at 1:1,000 dilutionat 4 °C overnight and then with goat anti-mouse IgG-horseradish-peroxidase conjugate diluted at 1:1,000 for 1 h.The immunoreactiveband was detected in the pineal gland or extracts of COS-7 cellstransfected with chicken Cyp11a cDNA by using ECL primeWestern blotting detection system (GEHealthcare). To confirm thespecificity of the immunoreaction, the primary antibody was pre-adsorbed with chicken P450scc protein (10 μg/mL).

Immunohistochemistry of P450scc. Immunohistochemical localiza-tion of P450scc was performed as described previously (8, 9). Inbrief, quail chicks at P7 were terminated by decapitation. Thebrains were fixed in 4% (vol/vol) paraformaldehyde solutionovernight, and after formic acid decalcification they were soakedin a refrigerated sucrose solution [30% (vol/vol) sucrose in 0.1 Mphosphate buffer]. Whole brains were frozen in optimal cuttingtemperature (OCT) compound (Miles) and sectioned transverselyat a 20-μm thickness on a cryostat at –20 °C. After blocking non-specific binding with 5% (vol/vol) normal goat serum and 1%BSAin PBS containing 0.5% Triton X-100, the sections were immersedovernight at 4 °C in 1:100 dilution of mouse anti-human P450sccantibody (Abcam). The sections were then incubated for 60 minwith Alexa Fluor 555 anti-mouse IgG (Invitrogen) at a dilution of1:1,000 and examined with a fluorescence microscope (Leica).

Quantification of Steroidogenic Enzyme mRNA Expressions. To com-pare the expressions ofCyp7b andSrd5amRNAs in thequail pinealgland, cerebellum, and diencephalon of both sexes at posthatchday 7 (P7) and P90, real-time PCR was conducted by using theStepOnePlus system (Applied Biosystems) as described previously(8). The oligonucleotide primers used in real-time PCR are listedin Table S1. β-Actinwas used as the internal standard. The reactionmixture contained SYBR Green Real-Time PCR Mix (Toyobo),400 nMeach of forward and reverse primers, and 30 ng of cDNA ina final volume of 20 μL. PCR was run with a standard cyclingprogram of 95 °C for 3 min, 40 cycles of 95 °C, 15 s; 60 °C, 15 s; and72 °C, 15 s. An external standard curvewas generated by a serial 10-fold dilution of cDNA obtained from the diencephalon, which hadbeen purified, and its concentration was measured. To confirm thespecificity of the amplification, the PCRproducts were subjected toa melting curve analysis and gel electrophoresis. The results werenormalized to the expression of β-actin using the StepOnePlus 2.0software (Applied Biosystems).

Quantification of Neurosteroid Syntheses.To compare neurosteroidsyntheses in the pineal gland, cerebellum, and diencephalon ofboth sexes at P7 and P90, quail were terminated and tissues weredissected. Each homogenate containing 20 mg of the tissue wasincubated separately with tritiated PREG or PROG for 30 min at37 °C. After incubation, the extracted steroids were subjected toHPLC analysis as described previously (4, 7–13).

Quantification of Neurosteroid Release. The quail pineal gland,cerebellum, and diencephalon were isolated from the light-ex-posed animals and cultured in 500 μL of medium 199 in 12-wellplates (each well containing 5 mg of tissue) at 37 °C under 5%CO2/80% O2. After 6 h of culture in the light, neurosteroids se-creted into the medium were extracted by ethyl acetate andsubjected to GC-SIM analysis to measure the concentrations of7α-OH PREG, 7β-OH PREG, ALLO, and EPI as describedpreviously (2, 8, 9, 12, 13, 15, 16).

Quantification of the Number and the Dendritic Length of PurkinjeCells. To investigate whether neurosteroids are involved in Pur-kinje cell survival in quail, pinealectomy (Px) at P2 and daily in-jection of ALLO or 7α-OH PREG were conducted usingposthatched male quail chicks. Px and sham operation for P2quail chicks were performed under nembutal anesthesia (40 mg/kg) as described previously (18). After the surgery, ALLO (30 ng/5 μL) or 7α-OH PREG (30 ng/5 μL) dissolved in sesame oil wasinjected into the pineal gland region of the quail chick brain onceper day, for 6 d, during P2–P7. For daily injection of neuro-steroids, a 5-μL volume of each reagent was injected at coor-dinates 1.0 mm lateral to the pineal gland region and 1 mm deepvertically to the surface of the skull as described in our previousstudy (19–21). The injection sites were confirmed at autopsy, andthe samples were discarded if the injection site was not at thecorrect position. We injected 30 ng of 7α-OH PREG or ALLOonce a day because ALLO and 7α-OH PREG released from thepineal gland were estimated to be around 4–6 ng/6 h of incubation.Control animals were treated with an equal volume of vehicle(sesame oil). Px quail chicks were s.c. implanted with a silastic(silicone type) plate containing melatonin (10 mg per plate) orvehicle as described in our previous study (18). The injection sitewas determined by visually inspecting the brains of the quail in-jected with 5 μL of 0.15%methylene blue dissolved in saline. Afterdecapitation under deep anesthesia, cerebella of control, Px,ALLO, 7α-OHPREG, ormelatonin-treated quail chicks at P3, P5,P7, and P21 were dissected. Cerebellar sections from 4% (vol/vol)paraformaldehyde-fixed brains were processed and stained withanticalbindin D-28k (Swant Swiss antibodies) as described pre-viously (19–21). The number of calbindin-immunoreactive cellbodies of Purkinje cells was counted in the cerebellar lobes I–X atP21. The length of the molecular layer in the parasagittal sectionwas evaluated as the maximal Purkinje dendritic length as de-scribed previously (19–21).

Quantification of ALLO Concentration in the Cerebellum. The con-centration of ALLO in the whole cerebellum, rostral cerebellum,or caudal cerebellum of quail chicks was measured by GC-SIManalysis. To measure the concentration of ALLO, cerebella ofcontrol, Px, and ALLO-treated quail chicks were dissected afterdecapitation under deep anesthesia. Cerebella were subdividedinto two regions: rostral cerebellum or caudal cerebellum. Eachcerebellar tissue was combined to the total weight of ∼500 mgand applied to GC-SIM analysis as described previously (2, 8, 9,12, 13, 15, 16).

Quantification of ALLO and/or EPI Synthesis. To investigate whethermelatonin influences the synthesis of cerebellar ALLO and/or EPIin the male juvenile quail, Px was performed at P2, and the quailchicks were s.c. implanted with a silastic plate containing mela-tonin (10 mg per plate) or vehicle (18). After decapitation underdeep anesthesia, cerebella of control, Px, or melatonin-treatedquail chicks at P3 were dissected and applied to HPLC analysis asdescribed above.

Quantification of ALLO Concentration in Serum. Trunk blood wascollected into heparinized glass tubes and centrifuged at 1,800 × gfor 20 min at 4 °C. Individual plasma of quail chicks at P3 was

Haraguchi et al. www.pnas.org/cgi/content/short/1210804109 2 of 5

Page 9: Possible role of pineal allopregnanolone in Purkinje … et al. PNAS...Possible role of pineal allopregnanolone in Purkinje cell survival Shogo Haraguchia,b, Sakurako Haraa, Takayoshi

stored at –20 °C. Plasma samples (1 mL each) were homogenizedin an aliquot of methanol/H2O [75:25 (vol/vol); 9 mL] on ice.The samples were passed through the C-18 cartridge column,and the neurosteroid fractions were eluted with methanol, dried,and applied to GC-SIM analysis as described above.

Quantification of [3H]ALLO Content in the Rostral and CaudalCerebellum. To better understand the mechanism of how pinealALLO affects the adjacent cerebellar Purkinje cells, 20 pmol [3H]ALLO (specific activity, 52.4 Ci/mmol; PerkinElmer) was injectedclose to the pineal lumen at P2 quail chicks. The content of [3H]ALLO in the rostral and caudal cerebellum of quail chicks wascounted in a liquid scintillation counter.

Assessment of Neuronal Cell Death. Parasagittal cerebellar sectionsof quail chicks at P3, P5, and P7 were examined by immunos-taining with an antibody against cleaved caspase-3 (Asp175) (CellSignaling Technology) to detect active-type caspase-3 (22) and byterminal deoxynucleotidyl transferase-mediated dUTP nick-endlabeling (Roche) to detect apoptotic cells as described previously(23). The usefulness of the antibody against cleaved caspase-3was validated in chickens as described previously (24).

Statistical Analysis.Results are expressed as the mean ± SEM. Thesignificance of differences between the groups was evaluated byone-way ANOVA followed by the Tukey–Kramer test. The dif-ferences were considered significant if P < 0.05.

1. Wada K, et al. (2006) A molecular neuroethological approach for identifying andcharacterizing a cascade of behaviorally regulated genes. Proc Natl Acad Sci USA103(41):15212–15217.

2. Hatori M, et al. (2011) Light-dependent and circadian clock-regulated activation ofsterol regulatory element-binding protein, X-box-binding protein 1, and heat shockfactor pathways. Proc Natl Acad Sci USA 108(12):4864–4869.

3. Takase M, Haraguchi S, Hasunuma I, Kikuyama S, Tsutsui K (2011) Expression ofcytochrome P450 side-chain cleavage enzyme mRNA and production of pregnenolonein the brain of the red-bellied newt Cynops pyrrhogaster. Gen Comp Endocrinol170(3):468–474.

4. Matsunaga M, Ukena K, Tsutsui K (2001) Expression and localization of cytochromeP450 17α-hydroxylase/c17,20-lyase in the avian brain. Brain Res 899(1–2):112–122.

5. Tsutsui K, Yamazaki T (1995) Avian neurosteroids. I. Pregnenolone biosynthesis in thequail brain. Brain Res 678(1–2):1–9.

6. Zuber MX, Mason JI, Simpson ER, Waterman MR (1988) Simultaneous transfection ofCOS-1 cells with mitochondrial and microsomal steroid hydroxylases: Incorporation ofa steroidogenic pathway into nonsteroidogenic cells. Proc Natl Acad Sci USA 85(3):699–703.

7. Doi M, et al. (2010) Salt-sensitive hypertension in circadian clock-deficient Cry-nullmice involves dysregulated adrenal Hsd3b6. Nat Med 16(1):67–74.

8. Haraguchi S, Koyama T, Hasunuma I, Vaudry H, Tsutsui K (2010) Prolactin increasesthe synthesis of 7α-hydroxypregnenolone, a key factor for induction of locomotoractivity, in breeding male newts. Endocrinology 151(5):2211–2222.

9. Haraguchi S, et al. (2012) Acute stress increases the synthesis of 7α-hydroxypregnenolone,a new key neurosteroid stimulating locomotor activity, through corticosterone actionin newts. Endocrinology 153(2):794–805.

10. Matsunaga M, Ukena K, Tsutsui K (2002) Androgen biosynthesis in the quail brain.Brain Res 948(1–2):180–185.

11. Matsunaga M, Okuhara K, Ukena K, Tsutsui K (2004) Identification of 3β,5β-tetrahydroprogesterone, a progesterone metabolite, and its stimulatory action onpreoptic neurons in the avian brain. Brain Res 1007(1–2):160–166.

12. Matsunaga M, Ukena K, Baulieu EE, Tsutsui K (2004) 7α-Hydroxypregnenolone acts asa neuronal activator to stimulate locomotor activity of breeding newts by means ofthe dopaminergic system. Proc Natl Acad Sci USA 101(49):17282–17287.

13. Tsutsui K, et al. (2008) 7α-Hydroxypregnenolone mediates melatonin actionunderlying diurnal locomotor rhythms. J Neurosci 28(9):2158–2167.

14. Concas A, et al. (1998) Role of brain allopregnanolone in the plasticity ofγ-aminobutyric acid type A receptor in rat brain during pregnancy and after delivery.Proc Natl Acad Sci USA 95(22):13284–13289.

15. Liere P, et al. (2000) Validation of an analytical procedure to measure trace amountsof neurosteroids in brain tissue by gas chromatography-mass spectrometry. JChromatogr B Biomed Sci Appl 739(2):301–312.

16. Meffre D, et al. (2007) Steroid profiling in brain and plasma of male andpseudopregnant female rats after traumatic brain injury: Analysis by gaschromatography/mass spectrometry. Endocrinology 148(5):2505–2517.

17. Nakao N, et al. (2008) Thyrotrophin in the pars tuberalis triggers photoperiodicresponse. Nature 452(7185):317–322.

18. Ubuka T, Bentley GE, Ukena K, Wingfield JC, Tsutsui K (2005) Melatonin induces theexpression of gonadotropin-inhibitory hormone in the avian brain. Proc Natl Acad SciUSA 102(8):3052–3057.

19. Sakamoto H, Ukena K, Tsutsui K (2001) Effects of progesterone synthesized de novoin the developing Purkinje cell on its dendritic growth and synaptogenesis. J Neurosci21(16):6221–6232.

20. Sakamoto H, Mezaki Y, Shikimi H, Ukena K, Tsutsui K (2003) Dendritic growth andspine formation in response to estrogen in the developing Purkinje cell.Endocrinology 144(10):4466–4477.

21. Sasahara K, et al. (2007) Mode of action and functional significance of estrogen-inducing dendritic growth, spinogenesis, and synaptogenesis in the developingPurkinje cell. J Neurosci 27(28):7408–7417.

22. Karnes HE, Kaiser CL, Durham D (2009) Deafferentation-induced caspase-3 activationand DNA fragmentation in chick cochlear nucleus neurons. Neuroscience 159(2):804–818.

23. Ubuka T, Ukena K, Sharp PJ, Bentley GE, Tsutsui K (2006) Gonadotropin-inhibitoryhormone inhibits gonadal development and maintenance by decreasinggonadotropin synthesis and release in male quail. Endocrinology 147(3):1187–1194.

24. Matsunaga E, et al. (2004) RGM and its receptor neogenin regulate neuronal survival.Nat Cell Biol 6(8):749–755.

Haraguchi et al. www.pnas.org/cgi/content/short/1210804109 3 of 5

Page 10: Possible role of pineal allopregnanolone in Purkinje … et al. PNAS...Possible role of pineal allopregnanolone in Purkinje cell survival Shogo Haraguchia,b, Sakurako Haraa, Takayoshi

Fig. S1. Neurosteroid formation from AD or T in the pineal gland. (A and D) RT-PCR analyses of steroidogenic enzyme Hsd17b (A) and Cyp19 (D) mRNAs in thepineal gland of male quail chicks. Total RNA was reverse-transcribed with (+) or without (−) RTase followed by PCR amplification. (B and E) HPLC analyses ofneurosteroid formation in the pineal gland of male quail chicks. (B) The pineal gland homogenates were incubated with [3H]AD + NADPH, and the extractswere subjected to HPLC. The arrowheads indicate elution positions of the substrate AD (open arrowhead) and its metabolite T (solid arrowhead). (E) The pinealgland homogenates were incubated with [3H]T + NADPH, and the extracts were subjected to HPLC. The pineal gland homogenates with [3H]T were alsoincubated with ketoconazole and finasteride. The arrowheads indicate elution positions of the substrate T (open arrowhead) and its metabolite AD, 5α- and/or5β-DHT, and E2 (solid arrowheads). (C, F, and G) GC-SIM analyses of the metabolites of nonradioactive AD or T in the pineal gland of male quail chicks. GC-SIMwas traced at m/z 680 for T (C), metabolite of nonradioactive AD, or at m/z 486 for 5α- and 5β-DHT (F) and at m/z 664 for E2 (G), metabolites of nonradioactiveT. The arrowheads show the peaks corresponding to authentic T (C), 5α- and 5β-DHT (F), and E2 (G). Similar results were obtained in repeated experimentsusing three different samples.

Fig. S2. A schematic model depicting the possible action of pineal ALLO on Purkinje cell survival in the developing cerebellum. (A) The location of the pinealgland in the quail chick brain. The pineal gland exists adjacent to the cerebellum. The square in the left bottom picture is magnified. (B) A schematic model ofthe neuroprotective action of pineal ALLO on Purkinje cells during development. ALLO is exceedingly produced in the pineal gland compared with the brain,may affect the adjacent cerebellar Purkinje cells by diffusion, and saves Purkinje cells from apoptosis in the juvenile quail. Secreted pineal ALLO inhibits theexpression of active caspase-3 that may facilitate apoptosis of Purkinje cells in the cerebellum. Cerebellar ALLO may also act on Purkinje cells together withpineal ALLO to prevent Purkinje cell death (see text for details).

Haraguchi et al. www.pnas.org/cgi/content/short/1210804109 4 of 5

Page 11: Possible role of pineal allopregnanolone in Purkinje … et al. PNAS...Possible role of pineal allopregnanolone in Purkinje cell survival Shogo Haraguchia,b, Sakurako Haraa, Takayoshi

Fig. S3. Effect of pineal ALLO and 7α-OH PREG or melatonin on the maximal dendritic length of Purkinje cells. Cerebella of control male quail chicks, Px at P2male quail chicks, and Px at P2 male quail chicks treated with daily injection of 7α-OH PREG (30 ng/5 μL) or ALLO (30 ng/5 μL), or Px at P2 male quail chicks s.c.implanted with a silastic plate containing melatonin (10 mg/plate) from P2 to P7 were analyzed. (A and B) Dendritic length of Purkinje cells in cerebellar lobesIV (A) and IX (B) (n = 12). There was no statistical difference (P > 0.05). Each column and vertical line in A and B represent the mean ± SEM.

Fig. S4. Effect of pineal melatonin on the synthesis of ALLO and/or EPI synthesis and the level of circulating ALLO. Cerebella of control male quail chicks, Px atP2 male quail chicks, and Px at P2 male quail chicks s.c. implanted with a silastic plate containing melatonin from P2 to P3 were analyzed. (A) ALLO and/or EPIsynthesis in the cerebellum at P3 (n = 7). (B) Circulating ALLO concentration of serum at P3 (n = 7). There was no statistical difference (P > 0.05). Each columnand vertical line in A and B represent the mean ± SEM.

Table S1. Primers for PCR analyses

Primer Forward primer sequence (5′→3′) Reverse primer sequence (5′→3′)

CHD TTGCCAAGGATGAGAAACTG TCTTCTCCTCCTACTGTGTT

chicken Cyp11a ATGCTCTCCAGGGCTGCACC TCACTCCTGGGGCTGGAGGG

StAR AATCACTCAGCATCCTCGG GGACCTGGTTGATGATGGTC

Cyp11a TGCAGGTTGGTCTCTACGC CTCCAGGATGTGCATGAGG

Cyp7b ATGAACATTCGCATCAGCC TCATCTCATTCATTGCGAGG

Hsd3b AAGGCAGATGGCCAGATGTT TGATGCGTCTGGCTTTCTGT

Srd5a GTGCACCGACATCCGATTTA CGATGGCAAAACCAAACCAT

Cyp17 TGAGTACTCCATCCCCAAGG CTCCTCGGGTTTATCCCACT

Hsd17b GCTGTTTACCCACCTTTGGA TTTTGAAATCCTCCCAGTGC

Cyp19 CATTCTCATCTGCGTTCTGTTTCT AGGCGTTACCTACTCCCATCC

β-Actin TTGTGATGGACTCTGGTGATG TTCTCTCTCGGCTGTGGTG

Haraguchi et al. www.pnas.org/cgi/content/short/1210804109 5 of 5