Estrogen Differentially Regulates Estrogen and Nerve Growth Factor Receptor Mrnas in Adult Sensory Neurons

The Journal of Neuroscience, February 1994, 74(2): 459-471

Estrogen Differentially Regulates Estrogen and Nerve Growth Factor Receptor mRNAs in Adult Sensory Neurons

Farida Sohrabji, Rajesh C. Miranda, and C. Dominique Toran-Allerand Department of Anatomy and Cell Biology and Center for Reproductive Sciences, Columbia University College of Physicians and Surgeons, New York, New York 10032

We have previously shown that neurons in the basal fore- Exposure to estrogenprofoundly influences developing neural brain colocalize the neurotrophin receptor p7PGFR and es- cytoarchitecture and organization, as well as the expressionand trogen receptors. The present study was designed to ex- maintenance of neuroendocrine and behavioral functions and amine (1) if neural neurotrophin targets respond to estrogen systemsin adulthood. Growth- (neurite-) promoting effects of as a general phenotypic feature and (2) if NGF receptor estrogenwere first demonstratedin organotypic explant cultures mRNAs are regulated by estrogen, using a prototypical tar- of the developing rodent hypothalamus, prcoptic area. and ce- get of NGF, the dorsal root ganglion (DRG) (sensory) neuron. rebral cortex (Toran-Allerand, 1976, 1980, 1984: Toran-Aller- We demonstrate, for the first time, the presence of estrogen and ct al., 1980, 1983) and subsequentlyconfirmed in the de- receptor mRNA and protein (binding sites) in adult female veloping brain in sitar(Nishizuka and Arai. I98 I; Hammer and rat DRG. Moreover, estrogen receptor mRNA expression, Jacobson, 1984; Stanley and Fink, 1986; Stanley et al., 1986: while present in DRG neurons from both the ovariectomized Lustig et al., I99 1) or as transplantsin the eye (Nishiyuka and (OVX; estrogen deficient) and intact female rat, was down- Arai, 1982) or brain (Matsumoto et al., 1988) of adult hosts,as regulated, as in the adult CNS, during proestrus (high estro- well asin estrogentarget regionsofthe deafferented(Matsumoto gen levels) and in OVX animals replaced with proestrus levels and Arai, I98 I) or steroid-deprived (Frankfurt et al., 1990)adult of estrogen, as compared to OVX controls. In contrast, al- CNS. though the mRNAs for the NGF receptors p7!iNGFR and frkA Estrogenregulates the transcription ofa variety ofcytoskeletal were also expressed in DRG neurons from OVX and intact proteins (Guo and Gorski, 1988; Lustig ct al., I99 1). steroid, animals, expression of both NGF receptor mRNAs was up- peptide, and neurotransmitter receptors(Dohanich et al., 1982; regulatedin sensory neurons during proestrus, as compared MacLusky and McEwen, 1978; DeKloet et al., 1986) as well to the OVX condition. Estrogen replacement, on the other as hormones and neuropeptides (Wilcox and Roberts, 1985; hand, resulted in a transient downregulafionof ~75~~~~ mRNA Roman0 et al., 1988) although the mechanismof estrogenac- and a time-dependent upregulation of trkA mRNA. Estrogen tion on many of its target genesis poorly understood. The es- regulation of NGF receptor mRNA in adult peripheral neural trogen receptor is a member of the superfamily of steroid/thy- targets of the neurotrophins supports the hypothesis that roid hormone/vitamin D,/retinoic acid receptors capable of estrogen may regulate neuronal sensitivity to neurotrophins activating genesby directly binding DNA sitescontaining hor- such as NGF and may be an important mediator of neuro- mone-specific regulatory elements (Evans, 1988; Beato. 1989; trophin actions in normal neural function and following neural O’Malley, 1990). An estrogenresponse element (ERE) has been trauma. identified in several estrogen-responsivegenes including vitel- [Key words: estrogen receptors, trkA mRNA, p7PGFR mRNA, logenin (Klein-Hitpass et al., 1986). c&s (Weisz and Rosales, dorsal root ganglia, proestrus, ovariectomy, estrogen re- 1990) prolactin (Watermann et al., 1988), and &luteinizing placement] hormone (Shupnik and Rosenzwcig, I99 1), suggestingthat, in some instances,steroid effects may be mediated through direct activation of relevant genes.Alternatively, estrogen-inducible genesmay be regulated through intermediate stepsvia inter- Received Apr. 9. 1993; revised July 16, 1993; accepted July 29, 1993. actions with, or secondaryactivation of, endogenoustranscrip- We thank Dr. Lois F. Parada (Frederick Cancer Research and Development Center) for the generous gift ofthe pDM97 plasmid. Drs. Michel Ferin and Sharon tion-regulating growth factors or their receptors, such as the Wardlaw (Columbia University) for assistance with the plasma estrogen assay, neurotrophin family of NGF-related peptides and their recep- Dr. Nell J. MacLusky (University of Toronto) for guidance with the estrogen- binding assay, and Wayne D. L. Bentham and Lucinda Lcung for skilled technical tors. Analysis of the identified promoter region (Sehgalet al., assistance at various stages of this project. We also thank Drs. Lloyd A. Greene, 1988)ofthe ~75”“~~ (Toran-Allerand et al., 1992a)and 5’-flank- Thomas M. Jessell. and Sharon L. Wardlaw (Columbia University) for construc- ing region of the t&A genesindicates the presenceof sequences tive comments and critical review of the manuscript. This work was supported in part by grants from NIH (AC-08099). NSF(BNS 87-00400and IBN 9l-09375), with a high degreeof homology to the putative vitellogenin and the American Health Assistance Foundation, Pilot grant from the Center for c&s EREs(Fig. 1), consistent with the hypothesisthat. in some Alzheimer’s Disease Research in New York (AC-08702-03), an ADAMHA Re- instances,estrogen may regulatesteroid-inducible genesthrough search Scientist Award (MH-00192) to C.D.T.-A.. and a gift from the MatTei family. Milan. Italy. This work has been presented in part at the 21st annual interactions with the neurotrophins or their receptors. meetmg of the Society for Neuroscience, New Orleans, LA, I99 I Two NGF receptorshave beendescribed. The 75 kDa protein Correspondence should be addressed to C. Dominique Toran-Allerand. Colum- (Chao et al., 1986; Johnsonet al., 1986; Radekeet al., bla University College of Physicians and Surgeons, 630 West 168th Street, New P75N”“K York. New York 10032. 1987) has been shown to bind with low affinity at least three Copyright (c? 1994 Society for Neuroscience 0270-6474/94/140459-13$05.00/O membersof the neurotrophin family; NGF, brain-derived neu- 480 Sohrabji et al. l Estrogen and NGF Receptor in RNA Regulation in DRG

Ere GGTCANNNTGACC IIII ::: III Humngfr AGAGGkACAGGGAACCGCAGTGGACGCCCCTGGTCCCCGGGACGACGCGCAGTAGCGCCGG 50 60 70 80 90 100

Humngfr GAACTGGGTACCAGGGCGGGATGGGTGAGAGGCTCTAAGGGACAAGGCAGGGAGAAGCGCGC 110 120 130 140 150

Ere GGTCANNNTGACC IIIIl:::I II HuntrkA CGATGTGGGCCGGGCAGAGGTCTCTGTTCAGGTCAACGTCTCCTTCCCGGCCAGTGTGCA 930 940 950 960 970 980

HumtrkA GCTGCACACGGCGGTGGAGATGCACCACTGGTCGATCCCCTTCTCTTTGGATGGGCAGCC 990 1000 1010 1020 1030 1040 Figure 1. ERE-like sequences in ~75~~~~ and trkA genes. Computer-assisted comparison of the 13 base palindromic ERE sequence (described for vitellogenin and C-$X genes) and the NGF receptor genes revealed close homology to sequences in the promoter region of the ~75~~~~ gene and the 5’ region flanking the trk oncogene breakpoint in the t&A gene. Homologous nucleotides are indicated by vertical bars. Humngfr, human p7YGFR sequence; HumtrkA, human t&A sequence.

Cortex ctx ctx A Uterus DRG 6 P6 P8 P14 c P4 DRG P6

28S- 28S- * 28S- 28S-

18S-

18S- 18S- r 18S-

Figure 2. Specificity of oligonucleotide probes. Specificity of the oligonucleotide probes for the estrogen receptor, p75NCFR,and trkA was determined by Northern blot analysis of total RNA. A, Hybridization of the estrogen receptor probe to a single (arrowhead) transcript (6.5 kb) in rat uterus and DRG RNA. B, Hybridization of the t&A oligonucleotide probe to a single (arrowhead) transcript (3.2 kb) in mouse cortex RNA, aged postnatal day (P) 6, P8, and P14. C, Hybridization of the ~75~~~~ oligonucleotide probe to a single (arrowhead) transcript (3.8 kb) in P4, P6 mouse cortex and rat DRG RNA. The Journal of Neuroscience, February 1994, f4(2) 461

Figure 3. A-E, Estrogen receptor mRNA expression in DRG neurons. Twenty micrometer cryostat sections through DRG of intact and OVX females were hybridized to a 48 base digoxigenin-labeled oligonucleotide probe for the estrogen receptor. Note that the extent of hybridization to this probe was much greater in sensory neurons of the OVX female (A) than the intact female during proestrus (B) or the OVX female replaced with estrogen for 52 hr (C). D, Control for hybridization specificity; DRG sections (from OVX animals) incubated with the estrogen receptor probe q/ier prior incubation with a 90 base oligonucleotide, overlapping the 48 base experimental probe, were always unstained. E. Immunohistochemical control; DRG sections from OVX animals hybridized with the noncomplementary (sense) oligonucleotide were also unstained. A, C, D, and E were photographed with Nomarski optics; B, bright-field photograph. Scale bar, 100 pm.

rotrophic factor, and neurotrophin-3. Neurotrophin signal pelvic ganglia (Melvin and Hamill, 1986, 1989) have been re- transduction, however, appears to require the neurotrophins to ported to be sensitive to gonadal steroids. associate with members of the tyrosine kinase receptor family We also addressed the general question of whether colocaliza- of proto-oncogene products, trkA (gp140”k) (Kaplan et al., tion of estrogen and neurotrophin receptor mRNAs implies the 199 la,b) and the structurally related genes trkB (Soppet et al., potential for regulatory interactions, by examining whether es- I99 1; Squint0 et al., 199 1) and trkC (Lamballe et al., 199 1). trogen responsiveness in DRG neurons would be manifested We have recently shown that estrogen receptor systems colo- functionally as an effect of this gonadal hormone on neurotroph- calize with ~75~“~~ mRNA and protein and trkA mRNA in in receptor gene expression. In the present study, we first com- neurons of the developing rodent medial septum, vertical and pared NGF receptor mRNA expression in DRG of ovariecto- horizontal limbs of the nuclei of the diagonal band, substantia mized (OVX; estrogen deficient) and intact animals at proestrus innominata, and nucleus basalis of Meynert (Toran-Allerand et (physiologically high estrogen levels) and found that not only al., 1992a; Miranda et al., 1993), as well as the cerebral cortex, are estrogen receptor mRNA and estrogen-binding sites present striatum (caudate/putamen), and hippocampus (Miranda et al., in adult female rat DRG neurons but that the expression of the 1993) regions previously described as targets of both NGF mRNAs for the estrogen receptor and the two NGF receptors (Taniuchi et al., 1986; Yan and Johnson, 1987, 1988; Ecken- p75NGFR and trkA is differentially modulated by the ovarian stein, 1988; Koh and Loy, 1989; Mobley et al., 1989) and es- status of the animal. To assess directly the action of estrogen trogen (Loy et al., 1988; Toran-Allerand et al., 1992b). The itself on NGF receptors mRNAs, we further compared OVX widespread colocalization ofp75N”FR with estrogen-binding sites animals with those replaced with estrogen levels comparable to suggested that the central and peripheral targets of the neuro- the morning of proestrus, and found a time- and receptor-spe- trophins may respond to estrogen as a general phenotypic char- cific regulation of NGF receptor mRNAs by estrogen. Our data acteristic. To address this issue we investigated the presence of support the hypothesis that the actions of estrogen on some estrogen receptor mRNA and estrogen-binding sites in a pro- target genes may involve intermediate steps, by modulating totypical peripheral target of NGF, the dorsal root (sensory) other transcription-regulating factors such as the neurotrophin ganglion (DRG) neuron of the adult female rat. DRG neurons receptors. are dependent on NGF for their survival during development (Barde, 1980; Johnson et al., 1980; Thoenen and Barde, 1980; Materials and Methods Eichler and Rich, 1989) but only following injury in adulthood Tissue samples (Csillick, 1987; Lindsay, 1988; Rich et al., 1989). Neither es- Dorsal root ganglia were obtained from intact, OVX, and estrogen- trogen receptor protein nor its mRNA has previously been dem- treated OVX adult female rats (Sprague-Dawley, Zivic-Miller, Allison Park, PA). All animals were maintained in a 14 hr: 10 hr light/dark cycle onstrated in DRG neurons, or indeed in other peripheral ganglia, and given food and water ad libitum. although neurons of the superior cervical ganglia (SCG; Wright ProestruslO VX study. The estrus cycle of intact females (n = 8) was and Smollen, 1983a,b), hypogastric ganglion, and the major monitored for at least 2 weeks (3-4 cycles) before use and animals were 462 Sohrabji et al. l Estrogen and NGF Receptor in RNA Regulation in DRG killed between I:00 and 2:00 P.M. on the day of proestrus (determined antibody-conjugated alkaline phosphatase, using nitroblue tetrazolium by vaginal smears). Bilaterally ovariectomized females were housed for salt (0.225 &ml) and 5-bromo-4-chloro-3-indolyl-phosphate (0. I75 IO d before death and handled daily. &ml) as substrates in buffer A, with 0.24 m&ml levamisole to inhibit Estrogen replacement study. Bilaterally OVX animals purchased from endogenous alkaline phosphatase activity. The color of the alkaline Zivic-Miller (Allison Park, PA) were housed for 10 d before the exper- phosphatase reaction product depends critically on the pH of the sub- iment. They were then injected, subcutaneously, with IO r+! of estradiol strate buffer (A), and can vary from a blue-purple (pH 9.5) to a brown- benzoate in sesame oil (IO:00 A.M., day l), and killed 4 hr (2:00 P.M., black (pH 7.5-8.0) coloration. For any given probe, one animal each day I) or 52 hr later (2:00 P.M., day 3). Controls were injected with from the following comparison groups was run concurrently, that is, vehicle (sesame oil) only. proestrus/OVX or OVX + vehicle/OVX + 4hrE/OVX + 52hrE. Color For in situ hybridization histochemistry, animals (n = S/group) were development was closely monitored, and the reaction was terminated anesthetized with pentobarbital(O.5 ml/l00 gm) and perfused transcar- for all groups (5 min in 10 mM Tris + I mM EDTA), when color appeared dially with isotonic saline, followed by 4% paraformaldehyde, contain- fully develoued in one condition. Slides were dehvdrated brieflv. throuahI ing 2.5% dimethyl sulfoxide (Sigma) and 0.1% diethyl-pyrocarbonate ethanol grades, cleared in Histoclear (National Diagnostics), and cov- (Sigma), an RNase inhibitor. Cervical and lumbar ganglia were removed erslipped in Permount. Hybridization and immunohistochemical con- and postfixed for 2 hr in the same fixative. Tissue was allowed to sucrose- trols for the estrogen receptor oligonucleotide are described in Results. equilibrate overnight [ 15% sucrose in phosphate-buffered saline (PBS)] Extensive controls, ensuring the specificity of the nonisotopic in situ and embedded in Ml embedding matrix (Lipshaw). Cryostat sections hybridization method and the probes used here, are described in Miran- (20 Fm) were thaw mounted onto gelatin/polylysine-coated slides and da and Toran-Allerand (1992) and Toran-Allerand and Miranda (I 993). processed immediately for in situ hybridization histochemistry. For RNA preparation, animals (OVX/proestrus, n = 3/group; estrogen re- Hybridization specificity of oligonucleotideprobes placement, n = S/group) were anesthetized (as before), the spinal column Hybridization specificity of each oligonucleotide probe was character- exposed, and cervical and lumbar DRGs gathered immediately into ized by Northern blot analysis. TotalRNA, prepared by a modification liquid nitrogen. DRGs were stored at -70°C until use. (Lu et al., 1989) of the Chomczynski and Sacchi (1986) method, was size fractionated on a I .2% agarose denaturing gel and capillary trans- Preparation qf oligonucleotideprobes ferred to nylon membrane (Hybond-N, Amersham). Filters were prehybridized overnight at 42°C in 4 x SSPE, 40% formamide, 5% dextran Complementary (“antisense”) and noncomplementary (“sense”) oli- sulfate, 25 &ml salmon sperm DNA, 100 &ml tRNA, 0.04x Den- gonucleotide sequences were synthesized on an Applied Biosystems hardt’s solution, and 0.2% SDS. The probes were 5’ end-labeled with 380B DNA Synthesizer (Protein Core, Columbia University/Howard Y-~~P-ATP, using T4-polynucleotide kinase, and hybridization was per- Hughes Medical Institute). formed at 42°C for 24 hr. Washes were performed under conditions of Estrogenreceptor. A previously described (Miranda and Toran-Al- gradually increasing stringency (4 x SSC to 0.2 x SSC, and from room lerand, 1992) 48 base oligonucleotide sequence was used, corresponding temperature to SOOC).Nylon membranes were apposed to Kodak X-Omat to nucleotides 1923-l 970 in the ligand-binding domain of the uterine film between intensifying screens at -80°C for 14 d. estrogen receptor (Koike et al., 1987) with less than 25% homology with Northern analysis of total rat uterine RNA probed with the estrogen the other members of the steroid receptor superfamily of DNA-binding receptor oligonucleotide probe revealed hybridization to a single tran- proteins. script (Fig. 2A) of the previously reported size [6.5 kilobases (kb) Koike NGF receptorA’.(1) For p75N”FH, a previously reported (Emfors et al., et al., 19871. A similar-sized transcript was seen in RNA obtained from 1988; Toran-Allerand et al., 1992a), 46 base sequence from the putative OVX rat DRG. The 46 base ~75~“~~ oligonucleotide sequence was membrane-spanning domain of the sequenced receptor in the chick, identical to that reported by Emfors et al. (1988), and Northern blot corresponding to nucleotides 807-852 of the rat homolog (Radeke et analysis of DRG-derived mRNA probed with this oligonucleotide re- al., 1987) was used. (2) For IrkA, a unique 60 base oligonucleotide vealed hybridization to the expected single transcript, approximately sequence directed against nucleotides 208-267 in the putative extra- 3.8 kb in size (Fig. 2C). Total RNA probed with the olipnucleotide cellular domain of the human cDNA sequence (Martin-Zanca et al., probe for t&A showed a single band of hybridization (Fig. 2B) to a 3.2 1989) corresponding to a region of the mouse t&A first exon (Martin kb transcript (Martin-Zanca et al., 1989). Zanca et al.. 1990). was used to detect trkA mRNA. Complementary and noncomplementary oligonucleotide probes were Northern blot analysis 3’ end-labeled with digoxigenin-deoxyuridine-triphosphate (dig-dUTP; an average of 4 molecules/tail), using terminal deoxynucleotidyl trans- Total RNA was size fractionated (IO-25 &lane) on a 1.2% agarose, ferase (Bethesda Research Labs) as previously described (Miranda and 18% formaldehyde gel and capillary transferred to nylon membrane Toran-Allerand, 1992; Toran-Allerand and Miranda, 1992). Labeled (Hybond-N, Amersham). Filters were prehybridized (4-l 8 hr) and hy- probes were purified by sodium acetate/ethanol precipitation and di- bridized (36-40 hr) at 37°C in 50% formamide, 3 x SSC, 0. I M Tris, luted with TES buffer (IO mM Tris-HCI, 1 mM EDTA, 0. I% SDS, pH 5 x Denhardt’sand 10% dextran sulfate. Probes were labeled with a-“P- 8.0). dCTP, using a random priming kit (Boehringer Mannheim). Trk4 mRNA was identified by hybridization with pDM97, a 450 base pair (bp) t&A- specific probe corresponding to the extracellular and transmembrane In situ hybridization histochemistrv domains ofthe trkA gene (Martin-Zancaet al., 1989; gift of L. F. Parada). In situ hybridization was performed according to the procedure de- P75N”FR was probed by hybridization with p5b, a 2 kb probe corre- scribed in Miranda and Toran-Allerand (1992). Briefly, sections were soondine. to the extracellular domain of the ~75~“~~ (Buck et al.. 1988). dehydrated in increasing ethanol grades, rehydrated without drying and Blots w&e washed in 0.2 x SSC and 0.5% ‘SDS, briefly at room tem- rinsed for 5 min in I x SSC (0.15 M NaCI, 0.0 I5 M sodium citrate, pH perature first and then for 2 hr at 6o”C, with frequent changes. Filters 7.2). Sections were prehybridized in a buffercontaining 50% formamide, were then exposed to film (Kodak X-Omat) between intensifying screens I M NaCI, I x Denhardt’s solution, 10 mM dithiothreitol (in 0.01 M and stored at -70°C for l-4 d before developing. sodium acetate, pH 5.2) and 0.5 mg/ml tRNA for 2 hr at room tem- Using a standard morphometrics package (Jandel Scientific), optical perature. Sections were then incubated 16-l 8 hr with the oligonucleotide density measurements were obtained from x-ray film. t&A and ~75~“~~ orobe (diluted 20 t&ml in the above buffer) at 42°C. rinsed twice in transcript densities were normalized to an internal control to account i x SSC, washed for2 hr in either 0.2 x SSC at 42°C (estrogen receptor for variation in the amount of RNA loaded. Since estrogen has been and ~75~“~~ probes) or in 0.1 x SSC at 60°C (t&A probe), and washed reported to enhance the expression of commonly used controls such as overnight (42°C) in I x SSC. Tissue sections were then blocked for 30 p-actin (Guo and Gorski, 1988) and 28s (Jones et al., 1986). but not min with 5% nonfat dry milk (Carnation) in Tris-buffered saline (TBS; I8S, ribosomal RNA (Yu et al., 199 I), NGF receptor transcripts in this IO mM Tris, 140 mM NaCl, pH 7.4) containing 0.3% Triton X-100 and study were normalized to 18s ribosomal RNA bands, visualized by then incubated in a I:500 dilution of the polyclonal sheep anti-digox- ethidium bromide or by a riboprobe specific for 18s (Ambion, Austin, igenin antibody, Fab fragment, conjugated to alkaline phosphatase TX). Group differences for the proestrus/OVX experiments were eval- (Boehringer Mannheim) prepared in the blocking solution, for 48 hr at uated using a t test while group differences for the estrogen replacement 4°C. Sections were washed three times for 15 min in TBS and soaked study were evaluated using a one-way ANOVA, with Student-Newman- for 5 min in buffer A (100 nM Tris, 200 mM NaCI, and 50 mM MgCI,, Keuls test of planned post hoc comparisons. Differences were considered pH 9.5). Hybrids were visualized with a color reaction catalyzed by significant at the 0.05 level. The Journal of Neuroscience, February 1994, 14(2) 463

Estrogen receptor binding assay The estrogen receptor probe used here is highly specific for Estrogen-binding sites were identified by a modification of the nuclear the estrogen receptor (Miranda and Toran-Allerand, 1992) and, exchange assay (MacLusky et al., 1990). Ten days after surgery, bilat- by Northern blot analysis, recognizes a band of approximately erally ovariectomized and adrenalectomized rats (Sprague-Dawley, 6.5 kb (Koike et al., 1987) in rat uterine RNA and DRG RNA Charles River, Kingston, NY) were deeply anesthetized (pentobarbital, from OVX rats (Fig. 2A). To ensure that the reaction product 0.5 ml/l00 pm) and their (cervical/lumbar) DRGs removed and collected in ice-cold medium [Dulbecco’s minimum essential medium was the result of specific hybridization to the oligonucleotide, (MEM). 0.65% glucose, 25% gelded horse serum). Ganglia were crudely some sections were incubated with the estrogen receptor probe freed of meningeal tissue and basement membrane and incubated with qfter prior incubation in a lOO-fold excess of an unlabeled, 90 a 2 nM concentration of the synthetic estrogen I I/J-methoxy-‘H-mox- base oligonucleotide (deduced amino acids 564-593) that com- estrol (R2858; Du Pont-New England Nuclear; specific activity, 80 Ci/ mmol: n = 3). for 2 hr at 37°C (n = 3). For some animals (n = 2). DRG pletely overlapped the 48 base experimental probe. These sec- were incubated with 2 nM l6u-“‘I-iodo-3,17@-estradiol (specific activ- tions always remained unstained (Fig. 30). As an immunohis- ity, 2200 Ci/mmol; Du Pont-New England Nuclear). In all cases, half tochemical control, some sections from each DRG were theganglia harvested were incubated with the radioligand in the presence hybridized with a noncomplementary (“sense”) oligonucleotide of 2 pc~ unlabeled estrogen (17/3-estradiol or diethylstilbestrol, DES). After incubation, DRG were washed three times in ice-cold MEM + and, as seen in the representative example in Figure 3E. these 0.1% BSA, for 5 min each, transferred to a homogenizer (clearance, sections always remained unstained, indicating a lack of hy- 0. I25 mm), and homogenized in 300 ~1 ofa I mM potassium phosphate, bridization. 3 mM magnesium chloride, and 0.25% Triton X-100 (pH 6.5) solution. A modification of the nuclear exchange assay (MacLusky et The homogenate was centrifuged briefly to recover a pellet and washed al., 1990; see Materials and Methods) was used to determine twice in the same buffer without Triton X-100. Pellets were then eluted with I ml of 95% ethanol (estrogen is fully soluble in >50% ethanol), the presence ofestrogen-binding sites in sensory ganglia. In these and the radioactivitv counted in a liauid scintillation counter. Pellets experiments, DRG were obtained from ovariectomized and ad- were dried and resuspended in I x TE (I 0 mM Tris, I mM EDTA) and renalectomized rats to minimize receptor occupancy by any assayed fluorometrically for DNA content, using the Hoechst dye 33258 endogenous estrogen. Similar levels ofcompetable binding were (Cesarone et al., 1979). Specific binding was determined as the difference seen using either 1 6oI-1’s1-iodo-3, 17/3-estradiol (3 I .36 fmol/mg between total binding and nonspecific binding (in the presence of an unlabeled competing ligand), and the data are expressed as femtomoles DNA) or 1 I@-methoxy-‘H-moxestrol (39.43 fmol/mg DNA), of specifically bound ligand/mg DNA. with a mean specific nuclear binding of 34.5 (k8.58) fmol/mg To demonstrate specificity ofestrogen binding sites in DRG neurons, DNA. The nonestrogenic ligands DHT and progesterone were additional experiments were performed using four groups of DRG. DRG not able to compete moxestrol binding. suggesting that binding were harvested as before and incubated in phenol red-free RPM1 I640 media supplemented with 25% gelded horse serum with ‘H-moxestrol was specific to the estrogen receptor. Since estrogen receptor alone or ‘H-moxestrol and 2 PM DES, or the nonestrogenic hormones mRNA localized primarily to sensory neurons, specific binding dihydrotestosterone (DHT) or progesterone. All subsequent analyses in these ganglia suggests that estrogen receptor mRNA is trans- wcrc as described before. Iatcd into its encoded protein in these neurons.

Estrogen assay Ovarian hormonesdownregulate estrogen receptor tnRNA in Circulating estrogen levels were assayed from the plasma of proestrus DRG neurons and untreated/estrogen-treated OVX animals. Blood was obtained from Not only was estrogen receptor mRNA present in all DRG the internal carotids of deeply anesthetized (pentobarbital, 0.5 ml/l00 gm) animals and collected in ice-cold, heparin-coated vials. Samples neurons, but the extent of hybridization to the estrogen receptor were then spun on a clinical centrifuge for IO min. and the supematant probe appeared to be regulated by the estrogen status of the (plasma) stored at -20°C until assayed. Estrogen content was deter- animal. In the OVX rat (Fig. 3/i) the pattern of hybridization mined by a solid-phase “‘I radioimmunoassay (Diagnostic Products, (staining) was virtually homogeneous within a ganglion. Thus, Los Angeles. CA). in the estrogen-deficient state, every DRG neuron appeared to cxprcss estrogen receptor mRNA strongly. In plasma obtained Results from OVX animals I I d aftergonadectomy, estrogen levels were I:‘strogen receptor nl RN,4 and estrogenbinding sitesare below sensitivity of the assay (~8 pgml; see Fig. 4). In contrast, pwscnt in adult IIRG on the afternoon of proestrus, soon after estrogen levels peak Estrogen receptor mRNA expression was found in virtually all physiologically and shortly before the estrus-related decline in DRG neurons of the intact and OVX animal, using nonisotopic estrogen, the pattern of hybridization to the estrogen receptor (digoxigcnin) in situ hybridization histochemistry with an oli- probe was more heterogeneous, as compared to the OVX con- gonucleotide probe (Miranda and Toran-Allerand, 1992) to the dition (Fig. 3R). While there were a few darkly stained neurons ligand-binding domain of the rat uterine estrogen receptor. The prcscnt. the hybridization signal was markedly reduced in the staining intensity of the hybridization product appeared strong- majority of neurons in this group. Plasma estrogen levels for est in ganglion neurons obtained from OVX rats. As seen in this group were on average 30 pg/ml, consistent with expected Figure 3:1, all large and small neurons in the ganglion hybridized estrogen levels on the afternoon of proestrus (Butcher et al.. the probe, as evidenced by the dark blue/purple cytoplasmic 1974). Hybridization to estrogen receptor mRNA also appeared reaction product. Nuclei generally remained unstained. How- markedly reduced in OVX animals. 52 hr after a single dose of ever. although hybridization was characteristically restricted to estrogen (Fig. 3C). While some neurons strongly hybridized the the cytoplasm, alkaline phosphatase results in a flocculent re- probe, in most instances the hybridization signal in DRG of action product, and where the probe hybridized abundantly, as estrogen-replaced OVX animals at 52 hr was markedly reduced in this instance, overstaining makes the nucleus less readily or virtually absent, although this effect was variable at 4 hr after apparent. Non-neuronal cells, such as satellite cells and fibro- estrogen treatment (data not shown). Plasma estrogen levels at blasts. did not hybridize the estrogen receptor probe, although 4 hr after a single injection of estrogen (Fig. 4) were high, re- occasional hybridization was seen in Schwann cells, and there sembling levels typically seen on the morning ofproestrus, while was no hybridization signal in neuropil and axon bundles. at 52 hr, estrogen levels had fallen almost threefold. It should 464 Sohrabji et al. - Estrogen and NGF Receptor in RNA Regulation in DRG

Regulation of P75”C8fiRrnRNA. As seenin Figure 5. .,I and H. DRG neurons, hybridized with the oligonucleotide probe for P75N”FRmRNA, exhibited some level of expression for this receptor mRNA irrespective of the hormonal status of the animal. While both large and small neurons were equally likely to express ~75~“~~ mRNA, there was a marked difference in the pattern of hybridization to the ~75~“~~probe, dependingupon the estrogen status of the animal. While some DRG neurons from OVX animals did hybridize the ~75~“~~probe, Figure 5il shows that there were many neurons present that did not de- tectably expressthis mRNA. The proportion of DRG neurons, from OVX animals,expressing p7YjFR mRNA varied from one experimental run to another; hence, their numbers were not quantified, althoughthese ranged approximately from one-fourth to one-half of the neurons present in a section. Among those neuronsthat did expressthe mRNA, a few were darkly stained but the majority had a pale hybridization product, indicating that the expressionof hybrids in theseneurons was significantly low. This pattern of hybridization was in striking contrast to the one seenin DRG neurons of the proestrus animal probed for ~75~“~~ mRNA. In these ganglia, virtually every neuron appeared to expressthe mRNA, and the homogeneousdark blue-black hybridization product (Fig. 54 indicated that the expression of this mRNA within a given neuron was much higher on average in proestrus than OVX animals. Northern blot analysis(Fig. 6A) of total RNA from OVX and proestrus DRGs confirmed the differencesobserved in the in situ hybridization analysis.Optical density measurementsofthe P75N”FRtranscript (normalized to 18s ribosomal RNA) indicated a 62% (p < 0.05) increase in ~75~“~~ mRNA during proestrus,as compared to OVX controls. The Northern analysis Figure 4. Plasma estrogen levels. Plasma estrogen levels were deter- confirmed the direction of proestrus induced changesseen by mined in OVX, proestrus, and estrogen-replaced OVX animals using in situ hybridization analysis,although the calorimetric reaction an iodinated estradiol antibody kit. Bar graph represents the mean appearsto exaggeratethe difference in ~75~‘;‘~ mRNA expres- (*SEM) estrogen levels detected, consisting of three animals from each sion. However, the alkaline phosphatasereaction is not linear hormonal condition. The level of estrogen in untreated (OVA’) and and therefore unsuitable for quantitation, unlike the Northern vehicle-treated OVX (OVX+ Vehicle) animals was undetectable by this assay (below 8 pg/ml) while the mean estrogen levels ofintact (Proestrus) blot analysis.Our resultsthus indicate that the hormonal con- animals agrees closely with those reported the afternoon of proestrus ditions existing at proestrus thus appeared to upregulateNGF (Butcher et al., 1974). OVX animals~given a single injection of EB (10 receptor mRNA, in contrast to the downregulation of estrogen ~a) had high levels of estrogen 4 hr later (OVA’+4hE). comparable to receptor mRNA. those seen&r the morning of proestrus, which declined threefold when measured 52 hr (0 KY+ 52hE) after the injection. Asteriskindicates 8 Regulation of trkA mRNA. Hybridization with the oligonu- pgml, the lower limit of the sensitivity of the assay. cleotide probe for trkA mRNA in DRG neurons revealed an expressionpattern similar to that of ~75~“~~mRNA, that is, a reduced level of expression in OVX animals, as compared to proestrus. In the OVX animal (Fig. 5C), a few darkly labeled be noted that in the intact animal, estrogentiters that are very neurons were generally found scattered in the periphery and low at metestrus,rise gradually to peak on the morning of pro- center of the ganglion. The majority of neurons,however, were estrus(2-2.5 d later) and fall shortly thereafter. Hence, the de- pale colored, indicative of significantly loweredlevels of mRNA creasein estrogen receptor mRNA hybridization during pro- expression.However, the pattern of hybridization was consid- estrusand at 52 hr after estrogentreatment suggests,as has been erably altered during proestrus, where almost every neuron in shown in the brain (Lauber et al., 1990; Simerly and Young, the ganglion appearedto hybridize strongly with the trkA probe, I99 I ; Shughrueet al., 1992) that prolonged exposure to estro- as shown by the dark blue-black staining of the DRG neurons gen downregulatesthe expressionof its own receptor mRNA in of Figure 5D. Both the number of hybridizing neuronsand the sensoryneurons. intensity of the hybridization signal appeared greater in the proestrus condition. Ovarian hormonesupregulate NGF receptor mRNAs Northern blot analysisof trkA RNA obtained from DRGs of proestrusand OVX animals confirmed the observed differences We alsoanalyzed the expressionofthe two NGFreceptor mRNAs in the pattern ofhybridization seenby in situ hybridization (Fig. P7SNCiFRand trkA in OVX, proestrus,and estrogen-replacedOVX 6B). The levels of trkA mRNA, measureddensitometrically and animals, by in situ hybridization histochemistry, using oligo- normalized to an internal control (18s ribosomal RNA), were, deoxynucleotide probes, and by Northern blot analysis, using on average, threefold (285%) greater during proestrusas com- cDNA probes. pared to the OVX condition. Thus, both NGF receptor mRNAs The Journal of Neuroscience, February 1994, 14(2) 465

Figure 5 hybridized to digoxigenin-labeled oligonucleotide probes for the ~75 N°FR(46 basesjand t&A (60 bases) mRNA. While NGF receptor mRNAs were expressed in both hormonal conditions, the extent of hybridization to ~75~~~~ (upper panels) and trk4 (lower panels) mRNA was much less in OVX (A and C) animals as compared to intact animals during proestrus (B and D). (A-C were photographed with Nomarski optics; D, bright-field photograph.) Scale bar, 100 km. appeared to be increasedin the proestruscondition, when compared to vehicle treated controls (p < 0.0 1). The levels of trkA pared to the OVX animals, trkA more so than p7YFK. mRNA observedin the OVX + 4hrE group wasnot significantly different from the OVX + 52hrE group, indicating an early and Estrogen regulation of ~75~‘~‘~and trkA mRNA prolonged upregulation of trkA mRNA by estrogen. Estrogen To assessdirectly the effect ofestrogen on NGF receptor mRNA, regulation of trkA mRNA was also detected by in situ hybrid- we then compared OVX animals with those replaced with a ization analyses, using a digoxigenin-labeled oligonucleotide single doseof estrogen(estradiol benzoate, EB) and killed 4 or probe for trkA. Hybridization patterns of the trkA probe, in situ 52 hr later. Plasmaestrogen assays indicated that, while OVX (Fig. 8A,B), closely paralleled Northern analysesof the t&A animals had undetectable levels of estrogen, 4 hr after a single transcript (Fig. 7B), with respectto estrogentreatment. In gen- dose of estrogen,estrogen titers rose to 58 pg/ml, comparable eral, the extent of hybridization of the trkA probe in DRG to levels seenin the morning of proestrus. Fifty-two hours after obtainedfrom vehicle-treatedcontrols appearedlow. While there a single injection of estrogen, estrogenlevels had fallen to 17 were a few darkly staining neurons, the majority of neurons pgml (Fig. 4). were either unstained or had a pale blue color, indicating an Northern blot analysis of ~75~~~~mRNA from DRGs ob- overall low level of hybridization product in DRG neurons tained from animals replaced with estrogenindicated that es- within this group (Fig. 8A). In contrast, hybridization product trogen transiently downregulatedthis receptormRNA @J< 0.05; was seenin virtually every neuron of DRG obtained from OVX Fig. 7A). Levels of ~75~‘~~~measured densitometrically (nor- animals exposed to estrogen for 52 hr (Fig. 8B), with many malized to 18s ribosomal RNA) were reduced by half, 4 hr after darkly staining neurons present throughout the ganglion. The a singledose of EB and were no different in OVX animals killed extent of hybridization to the trk4 probe in DRG from OVX 52 hr after a singleinjection ofestrogen, ascompared to vehicle- animals, 4 hr after estrogentreatment (data not shown), ap- treated controls. peared intermediate to those seenin the OVX and OVX + However, in the caseof trkA (Fig. 7B), in contrast, densito- 52hrE groups. ~75~~~~mRNA regulation by estrogenwas not metric measurements ofthis transcript in Northern blot analysis amenableto in situ hybridization analysesgiven that the already indicated that estrogentreatment resultedin a twofold increase low hybridization to the ~75~“~~ probe in DRG neurons of in mRNA expressionof this NGF receptor 4 hr after a single untreated OVX animals (example shown in Fig. 5A) would be injection of estrogen (p < 0.0 1) and a threefold increasein trkA further decreasedin DRG neurons of estrogen-treatedOVX mRNA in OVX animals exposed to 52 hr of estrogen, as com- animals (as indicated by Northern blot analyses),requiring de- 466 Sohrabji et al. l Estrogen and NGF Receptor in RNA Regulation in DRG

”

0 0 L

tection of differencesbeyond the sensitivity of most, including cells,another prototypical target of NGF, have estrogen-binding this, calorimetric assay. sites, and in this cell type as well, estrogenappears to regulate differentially both p7YGFRand t&A mRNA (F. Sohrabji, L. A. Discussion Greene, R. C. Miranda, and C. D. Toran-Allerand, unpublished The present data indicate that adult femalerat sensoryneurons observations). are estrogentargets, in viva, and that estrogentreatment differ- Although estrogen receptors have not been previously re- entially modulatesthe expressionof estrogenand NGF receptor ported in adult sensoryneurons, the widespreadexpression of mRNAs. These findings suggest that estrogen may regulate pe- estrogenreceptor mRNA observed in DRG neurons and the ripheral neuronal sensitivity to both estrogenand the neuro- presenceof nuclear estrogen-bindingsites argue that theseneu- trophins. Our data indicate that estrogenreceptor mRNA and rons are estrogentargets. While the levels of estrogen-binding specific, competable, nuclear estrogen-binding sites(receptors) sites in sensory ganglia appear low, they are within the range are present in adult sensory ganglia. Moreover, the estrogen reported for estrogentargets in the adult brain (Friedman et al., status of the animal appears to regulate reciprocally the ex- 1983). It should be noted, however, that the estimatesof estro- pressionof estrogenreceptor mRNA and the mRNAs for the gen-binding sitesreported here may underestimatethe propor- NGF receptorsp7YCFK and irkA in adult sensoryneurons. While tion of estrogenreceptors in DRG neurons.First, binding sites estrogenreceptor mRNA is downregulated and NGF receptors were normalized to total DNA content of the ganglia, while the upregulatedduring proestrus, estrogenreplacement in OVX an- in situ hybridization analysis indicates that estrogenreceptor imals, in contrast, appearsto downregulate mRNA for the es- mRNA was predominantly expressedonly in neuronsand not trogen receptorand the pan-neurotrophin receptorp7YiFR, while support cells of the DRG. Additionally, binding assaysthat used upregulating trkA mRNA. The present findings are consistent Dulbecco’s MEM (seeMaterials and Methods) were not phenol with our hypothesis that estrogensensitivity may be a general red free, and a contaminant found intermittently in phenol red feature of the neural targetsof neurotrophins, in this caseNGF, dye lots is reportedly estrogenic (Berthois et al., 1986; Bindal and that some of estrogen’saction on neural tissue may result and Katzenellenbogen, 1988), which may contribute to under- via estrogen-induced transcriptional regulation of the neuro- estimates of actual binding sites. In fact, preliminary assays trophin receptors. The potential universality of these interac- using phenol red-free media suggestthat estrogenbinding es- tions is suggested,moreover, by preliminary findingsthat PC12 timates in DRG are substantially higher. In addition to the The Journal of Neuroscience, February 1994. 14(2) 467

* -L

52hE

Figure 7. Northern blot analyses of estrogen regulation of NGF receptor mRNAs. Total RNA, loaded 25 pg per lane, was obtained from DRG of vehicle-treated OVX animals (lanes I and 2) and OVX animals killed 4 hr (lanes 3 and 4) and 52 hr (lanes 5 and 6) after a single injection of EB. Nylon filters were hybridized with random-primed, 32P-labeled cDNA probes for either a 2 kb cDNA probe corresponding to the extracellular domain of the ~75~‘~~~ gene (A) (p5b; Buck et al., 1988),or a 450bp trlul-specificcDNA probecorresponding to the extracellularand transmembrane domainsof the t&A gene(B) (pDM97; Martin-Zancaet al., 1989;gift of L. F. Parada).To quantify groupdifferences, optical density of the NGF receptortranscripts was normalized to 18sribosomal RNA, to control for variationsin the amountof total RNA loadedonto the gel. Bar graph represents mean (+SEM) of five animals from each group. Asterisks indicatestatistically significant (p < 0.05) differencesas comparedto the vehicle-treated OVX group. binding assays,the effect of estrogenon NGF receptor mRNA estrogenreceptor mRNA in sensoryneurons are similar to that expressionfurther supportsthe concept of the existence of func- observed in the adult brain (Simerly and Young, 1991; Lauber tional estrogenreceptors in DRG neurons.Our presentdata are et al., 1992)and extraneural targetsofestrogen suchas the uterus also consistent with reports that developing DRG and other (Shupnik et al., 1989). In the case of proestrus animals, fur- peripheral neurons are responsive to estrogen. For example, thermore, downregulation of estrogenreceptor mRNA in sen- neonatalexposure to estrogenreportedly increasesboth neuron sory neuronsis similar to that observedin the intact adult brain and synapsenumber in the SCG (Wright and Smollen, 1983a). compared to the OVX condition. For example, in established More recently, it has been shown that those lumbosacral spine targets of estrogen such as the medial preoptic area and the motoneuronsthat innervate the bulbocavernosusmuscle [spinal ventromedial and arcuate nuclei of the hypothalamus, ovari- nucleusof the bulbocavemosus(SNB)], if exposed to early es- ectomy reportedly increasesthe level ofestrogenreceptor mRNA, trogen treatment, initiate elaborate dendrite arborizations (D. when compared to any stagein the estrus cycle of the intact Sengelaub,personal communication). Sinceneither estrogenre- female (Shughrue et al., 1992). Similarly, OVX animals with ceptors (binding sites)(Pfaff and Keiner, 1973; Sar and Stumpf, hormone replacementalso show marked downregulation of es- 1977; Breedlove and Arnold, 1980) nor mRNA (Simerly et al., trogen receptor mRNA in the ventrolateral aspect of the ven- 1990)have been found in adult SNB motoneurons,this dendrite tromedial nucleus and in the arcuate nucleus of the hypothal- growth may well result indirectly from the action of estrogen amus 18 hr and later, although this effect is not apparent at 4 on those neurons and processesafferent to the SNB, namely, hr following estrogentreatment (Lauber et al., 1992). Exposure the L.5, L6, and Sl DRG neurons. to estrogen in the form of an implanted silastic capsule also In the presentstudy, the expressionofestrogen receptor mRNA depressedestrogen mRNA levels in theseregions, when mea- was markedly reducedin DRG neuronsof the proestrusanimal, sured 24 and 72 hr later (Simerly and Young, 1991). In the as well as OVX animals 52 hr after exposure to estrogen, as present study, while estrogenconsistently downregulatedits re- compared to ovariectomized controls. Both the direction and ceptor mRNA at 52 hr, the effectsof estrogenat 4 hr were more relative time course of estrogen-dependentdownregulation of variable, being downregulated in some cases,and unchanged 466 Sohrabjl et al. * Estrogen and NGF Receptor in RNA Regulation in DRG

Figure 8. In situ hybridization analyses of t&A mRNA. DRG sections(20 r.rrn thick), obtained from vehicle- and estrogen-treated OVX animals, were hybridized with a digoxigenin-labeled oligonucleotide probe for trkA mRNA. The extent of hybridization to the trkA probe was low in sensory neurons of vehicle-treated OVX animals (A), but increased dramatically in OVX animals exposed to estrogen for 52 hr (B). Pho- tographed with Nomarski optics. Scale bar, 100 pm.

from controls (OVX) in other instances(data not shown). In proestrus, it is not the only gonadal hormone present at this general, though, it appearsthat the direction and time course time. In addition to estrogen,there are high levels of testoster- of estrogen receptor mRNA regulation in DRG neurons are one, prolactin, and follicle-stimulating hormone presentduring similar to that seenin the central neural targets of estrogen. the afternoon of proestrus (Brown-Grant et al., 1970; Neil], While the hormonal conditions during proestrusand estrogen 1972; Butcher et al., 1974) which may influence transcription replacement of OVX animals both downregulate estrogen re- of NGF receptors. It may prove to be the case that estrogen ceptor mRNA, they appearto have different effects on the two interactions with other hormonespresent at proestrusmay result NGF receptor mRNAs present in DRG neurons. Both NGF in markedly different outcomes from that of estrogenacting receptorsare present in DRG neurons:the p7YGFRprotein, the alone, as in the estrogen-replacedOVX animals. Thus, the first cloned NGF receptor (Chao et al., 1986; Johnson et al., changesin NGF receptor mRNA during proestrus must rep- 1986; Radekeet al., 1987) and trkA (Martin-Zanca et al., 1990). resent a sum total of several hormonesexerting diverse control Compared to OVX animals, expression of trkA mRNA was of the samegene, although the similarity on the direction and enhanced comparably in both the proestrus animal as well as amount of trkA mRNA regulation in both paradigms (OVX/ the estrogen-replacedOVX animals. trkA mRNA expression proestrusand estrogen-replacedOVX) suggeststhat t&A mRNA was significantly higher than controls 4 hr after estrogentreat- upregulation during proestrus may be due specifically to the ment and remained upregulated 52 hr after estrogentreatment. high titers of estrogenpresent at this time. There appeared to be no statistically significant increment in Although estrogenappears to have differential effects on the trkA mRNA expressionbetween 4 and 52 hr of estrogentreat- two components of the NGF receptor, its actions appear to be ment, although at the presenttime wecannot determine whether consistentwith thosereported for other membersofthe steroid- estrogenacts by continuous stimulation of the trkA gene or by thyroid hormone receptor superfamily of transcription factors. increasingthe half-life of its transcript. While trkA mRNA levels For example, replacingtestosterone, a relatedgonadal hormone, remained equally high at both time points, estrogentiters at 52 downregulates~75~~~~ mRNA in Sertoli cells of the adult testis, hr postinjection were threefold lower than at 4 hr postinjection. an androgen target (Persson et al., 1990). In the brain, long- Hence, unlike ~75~‘~~~mRNA, which fell at 4 hr of estrogen term (16-30 d) exposure to a silastic implant of estrogenis treatment and returned to basal levels at 52 hr, trkA mRNA reported to decrease~75~~~~ mRNA in the hippocampus and levels appearnot to have fallen in responseto decreasingavail- septum (Gibbs and Pfaff, 1992). Retinoic acid, another member ability of estrogen. However, since estrogenlevels did not fall of this superfamily of DNA-binding proteins, has been shown to basal(OVX) levels by 52 hr we can only speculatethat even to increaselZ51-NGF binding in Lan- 1 neuroblastomacells (Has- the low levels of estrogenpresent were capable of stimulating kell et al., 1991) and in developing chick sympathetic neurons this gene. To test accurately whether trkA mRNA levels fall if (Rodriguez-Tebar et al., 1991) and to increaset&A mRNA in estrogenis subsequentlyremoved would be more feasiblein a neuroblastomacells (D. R. Kaplan, personalcommunication). tissueculture systemwhere exposureand deprivation ofa ligand In fact, our own preliminary data from PC12 cells indicate that are more feasible. estrogenupregulates trkA mRNA and downregulates~75~~~~ While trkA mRNA was upregulated in both the proestrusand mRNA (Sohrabji, Greene, Miranda, and Toran-Allerand, un- estrogenreplacement condition, ~75~~~~mRNA, on the other published observations), to a similar extent as in the DRG. hand, which appeared enhanced during proestrus, was down- At the present time, we can only speculateon the biological regulated 50% following estrogen treatment, as compared to and functional significance of differential regulation of NGF OVX controls. The most obvious explanation for the discrep- receptor mRNAs by estrogen, since the precise role of each ancy in p75NGFRmRNA regulation between the two paradigms receptor is not clearly known. While ~75~~~~and trkA individ- lies in the fact that, although estrogen titers are very high at ually bind NGF with low affinity, together they demonstrate a The Journal of Neuroscience, February 1994, 14(2) 469

high-affinity binding characteristic (Hempstead et al., 1991). trophin family. Steroid/neurotrophin interactions suggest a Some evidence suggests, however, that cellular requirements for mechanism by which transcription factors such as the estrogen the two NGF receptors may differ. In a mutant form (nnr) of receptor, via actions on other transcription-regulating factors, PC 12 cells, which does not respond to NGF with neurite elon- may exert wide influences on the nervous system. gation (Greene et al., 1986), p75 N(;FKis present at the usual levels but trkA mRNA is severely depleted (Loeb et al., 199 1). Trans- fecting these mutants with an expression vector encoding trkA References cDNA reinstated the capacity for NGF-induced neurite out- Bajusz E (1959) Effect of hormones upon regression of muscle atrophy growth (Loeb et al., 199 l), suggesting that trkA is necessary for of nervous origin. Endocrinology 64:262-269. neurotrophin-mediated signal transduction. Hence, if trkA is Barde YA, Edgar D, Thoenen H (1980) Sensory neurons in culture. Changing requirements for survival factors during embryonic devel- the signal transducing arm of the NGF receptor, estrogen, by opment. Proc Nat1 Acad Sci USA 77: 1199-l 203. increasing trkA mRNA in sensory ganglia, may be in a position Battleman DS, Geller AI, Chao MV (1993) HSV-I vector-mediated to alter neuronal responsivity to NGF, by altering the ratio of gene transfer of the human nerve growth factor receptor p75hN”‘-H and trkA, and consequently, the formation of the hy- defines high-affinity NGF binding. J Neurosci 13:94 l-95 I. P75N”‘-K Beato M (1989) Gene regulation by steroid hormones. Cell 56:335- pothetical NGF receptor complex as recently suggested by Chao 344. and colleagues (Battleman et al., 1993). In addition, it is inter- Berthois Y, Katzenellenbogen JA, Katzenellenbogen BS (I 986) Phenol esting to note that preliminary evidence from transfected NIH- red in tissue culture is a weak estrogen: implications concerning the 3T3 cells indicates that overexpression of p7SNGFR inhibits au- study of estrogen-responsive cells in culture. Proc Nat1 Acad Sci USA 83:2496-2500. tophosphorylation of trkA (D. R. Kaplan, personal communi- Bindal RD, Katzenellenbogen JA (1988) Bis(4-hydroxyphenyl)‘2- cation). It may be that estrogen, by increasing trkA mRNA while (phenoxysulfonyl)phenyl-methane: isolation and structure elucida- decreasing p7Z+jFK mRNA, may modulate ligand-independent tion of a novel estrogen from commercial preparations of phenol red signal transduction by receptor tyrosine kinases. Although the (phenolsulfonphthalein). J Med Chem 3 1: 1978-l 983. present study focused on estrogen regulation of neurotrophin Breedlove SM, Arnold AP (1980) Hormone accumulation in a sexually dimorphic motor nucleus of the rat spinal cord. Science 210:564- receptors that bind NGF, DRG neurons also express trkB and 566. trkC, which bind other members of the neurotrophin family of Brown-Grant K, Exley D, Naftolin F (1970) Peripheral plasma oes- growth factors. Localization ofestrogen receptors in neurotroph- tradiol and luteinizing hormone concentration during the oestrus cycle in target neurons both in the DRG and CNS suggests that es- of the rat. J Endocrinol 48:261-295. trogen may be in a position to regulate mRNA expression of Buck CR, Martinez HJ, Chao MV, Black IB (1988) Differential expression of nerve growth factor receptor gene in multiple brain areas. other tyrosine kinase receptors as well. Dev Brain Res 44:259-268. Estrogen replacement may downregulate ~75~“~~ mRNA in Butcher RL, Collins WE, Fugo NW (1974) Plasma concentrations of neurotrophin targets such as the DRGs and the basal forebrain LH, FSH. prolactin. progesterone and estradiol- 170 throughout the (Gibbs and Pfaff, 1992) of the uninjured adult, but its actions 4 dav estrous cycle of the rat. Endocrinolonv 94: 1704-l 708. Cesaroie C, Bolognesi C, Santi L (I 979) Im$oved microfluorometric on this gene may be different during periods of critical neuro- DNA determination in biological material using 33258 Hoechst. Anal trophin requirement such as development or following injury. Biochem 102:188-197. For example, early neonatal exposure to estrogen appears to Chao MV, Bothwell MA, Ross AH, Koprowski H, Lanahan AA, Buck increase high-affinity NGF receptor binding, as determined by CR, Sehgal A (I 986) Gene transfer and molecular cloning of the human NGF receptor. Science 232:5 18-52 I. lZSI-NGF binding in DRG neurons (Wright et al., 1988), sug- Chomczynski P, Sacchi N (I 987) Single step method of RNA isolation gesting that estrogen may increase both trkA and ~75~‘~~~ tran- by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal scriptionally or posttranscriptionally. Moreover in the adult Biochem 162:156-159. brain, where estrogen downregulates ~75~“~~ mRNA in the in- Csillick B (1987) Transganglionic regulation of the primary sensory tact basal forebrain, it appears not to augment the loss of p75NC”K neuron. Acta Physiol Acad Sci Hung 69:355-36 I DeKloet ER, Voorhuis DAM, Boschma Y, Elands J (I 986) Estradiol mRNA, following fimbria-fornix lesions (Gibbs and Pfaff, 1992). modulates density ofputative ‘oxytocin receptors’ in discrete rat brain Furthermore, estrogen treatment has been reported to have a regions. Neuroendocrinology 44:4 15-42 1. beneficial effect on peripheral nerve regeneration (Bajusz, 1959), Dohanich GP, Witcher JA, Weaver DR, Clemens LG (1982) Alter- consistent with the hypothesis that estrogen, via interactions ations of muscarinic binding in specific brain areas following estrogen with growth factors known to ameliorate injury in the adult treatment. Brain Res 241:347-350. Eckenstein F (1988) Transient expression of NGF-receptor-like im- PNS, such as the neurotrophins, may have a protective role munoreactivity in postnatal rat brain and spinal cord. Brain Res 446: following neural injury. 149-154. The modulation ofneurotrophin receptor mRNAs bygonadal Eichler ME, Rich KM (I 989) Death of sensory ganglion neurons after hormones, particularly estrogen, in adult peripheral neural tar- acute withdrawal of nerve growth factor in dissociated cell cultures. Brain Res 482:340-346. gets ofNGF suggests a critical role for these steroids in mediating Emfors P, Hallbrook F, Ebendal T, Shooter EM, Radeke MJ, Misko NGF-dependent regulation of neuronal survival, plasticity, and TP, Persson H (1988) Developmental and regional expression of repair. Evidence from this laboratory indicates that estrogen P-nerve growth factor receptor mRNA in the chick and rat. Neuron receptor mRNA or protein colocalizes with t&A and trkB mRNA 1:983-996. as well as ~75~“~~ mRNA or protein in neurotrophin targets of Evans RM (1988) The steroid and thyroid hormone receptor superfamily. Science 240:889-895. the developing forebrain (Toran-Allerand et al., 1992a; Miranda Frankfurt M, Gould E, Woolley C, McEwen BS (1990) Gonadal ste- et al., 1993), suggesting the potential for estrogen regulation of roids modify spine density in the ventromedial hypothalamus neu- NGF receptors in the CNS, as well as the potential for estrogen rons: a Golgi siudy. Neuroendocrinology 51:530-jj5. regulation of other members of the tyrosine kinase family of Friedman WJ. McEwen BS. Toran-Allerand CD. Gerlach JL (1983) Perinatal development of hypothalamic and cortical estrogen ;ecep: neurotrophin receptors. Moreover, since ~75~~~~ binds BDNF tars in mouse brain: methodological aspects. Dev Brain Res I I: l9- and NT-3 as well as NGF, gonadal hormones may also alter 27. responsivity of target neurons to other members of the neuro- Gibbs R, Pfaff DW (I 992) Effects of estrogen and IimbriaIfornix tran- 470 Sohrabji et al. . Estrogen and NGF Receptor in RNA Regulation in DRG

section on P75NGFR and ChAT expression in the medial septum Martin-Zanca D, Barbacid M, Parada LF (1990) Expression of the rrk and diagonal band of Broca. Exp Neural 116:23-39. proto-oncogene is restricted to the sensory cranial and spinal ganglia Greene Sfi, Rydel RE, Conolly JL, Greene LA (1986) PC12 cell mu- of neural crest origin in mouse development. Genes Dev 4:683-694. tants that possess low- but not high-affinity nerve growth factor re- Matsumoto A, Arai Y (198 1) Neuronal plasticity in the deafferented ceptors neither respond to nor internalize nerve growth factor. J Cell hypothalamic arcuate nucleus of adult female rats and its enhance- Biol 102:830-843. ment by treatment with estrogen. J Comp Neural 197: 197-206. Guo JZ, Gorski J (1988) Estogen effects on histone messenger ribo- Matsumoto A, Murakami S, Arai Y (1988) Neurotrophic effects of nucleic acid levels in the rat uterus. Mol Endocrinol 2:693-700. estrogen on the neonatal preoptic area grafted into the adult brain. Hammer RP. Jacobson CD (1984) Sex difference in dendritic devel- Cell Tissue Res 252:33-37. opment of the sexually dimbrphic nucleus of the preoptic area in the Melvin JE, Hamill RW (1986) Gonadal hormone regulation of neu- rat. Int J Dev Neurosci 2:77-85. rotransmitter synthesizing enzymes in the developing hypogastric Haskell BE, Stach RW, Werbach-Perez K, Perez-Polo JR (1987) Effect ganglion. Brain Res 383:38-46. of retinoic acid on nerve growth factor receptors. Cell Tissue Res 247: Melvin JE, Hamill RW (1989) Androgen-specific critical periods for 67-73. the organization of the major pelvic ganglion. J Neurosci 9:736-742. Hempstead BL, Martin-Zanca D, Kaplan DR, Parada LF, Chas MV Miranda RC, Toran-Allerand CD (1992) Developmental expression (1991) High-affinity NGF binding requires the co-expression of the of estrogen receptor mRNA in the rat cerebral cortex: a non-isotopic trk proto-oncogeneand the low-affinity receptor. Nature 350:678:683. in situ hybridization histochemistry study. Cereb Cortex 2: l-1 5. Johnson D, Lanahan A, Buck CR, Sehgal A, Morgan C, Mercer E, Miranda RC, Sohrabji F, Toran-Allerand CD (1993) Estrogen target Bothwell M, Chao M (1986) Expression and structure of the human neurons express mRNAs for the neurotrophins and their receptors NGF receptor. Cell 47:545-554. during development: a basis for potential interactions ofestrogen with Johnson EM, Gorin PD. Brandies LD, Pearson J (1980) Dorsal root the neurotrophins. Mol Cell Neurosci, in press. ganglion neurons are destroyed by exposure in utero to maternal Mobley CM, Woo JE, Edwards RH, Riopelle RJ, Longo FM. Wescamp antibody to nerve growth factor. Science 2 IO:9 16-9 18. G, Otten U, Valetta JS, Johnston MV (1989) Developmental reg- Jones KJ, Chikaraishi DM, Harrington CA, McEwen BS, Pfaff DW ulation of nerve growth factor and its receptor in the rat caudate- (1986) In situ hybridization detection of estradiol-induced changes putamen. Neuron 3:655-664. in ribosomal RNA levels in rat brain. Mel Brain Res 1:145-l 52. Neil1 JD (1972) Comparison of plasma prolactin levels in cannulated Kaplan DR, Hempstead BL, Martin-Zanca D, Chao MV, Parada LF and decapitated rats. Endocrinology 90:568-572. (I 99 1a) The trk protooncogene product: a signal transducing receptor Nishizuka N, Arai Y (1981) Sexual dimorphism in synaptic organi- for nerve growth factor. Science 252:554-558. zation in the amygdala and its dependence on neonatal hormone Kaplan DR, Martin-Zanca D, Parada LF (199 I b) Tyrosine phos- environment. Brain Res 2 12:3 I-38. phorylation and tyrosine kinase activity of the trk protooncogene Nishizuka N, Arai Y (1982) Synapse formation in response to estrogen product induced by NGF. Nature 350:158-160. in the medial amygdala developing in the eye. Proc Nat1 Acad Sci Klein-Hitpass L. Schorpp M. Wagner U. Rvffel GU (19861 An estro- USA 79~7024-7026. gen-responsive element derive2 from thk 5’ flanking region of the O’Malley B (1990) The steroid receptor superfamily: more excitement Xenopus vitellogenin A2 gene functions in transfected human cells. predicted for the future. Mel Endocrinol 4:363-369. Cell 46:1053-1061. Persson H, Ayer-Le Lievre C, Sodor 0, Villar MJ, Metsis M. Olson L, Koh S, Loy R (I 989) Localization and development of nerve growth Ritzen M, Hokfelt T (1990) Expression of &nerve growth factor factor-sensitive rat basal forebrain neurons and their efferent projec- mRNA in Sertoli cells down:reguiated by testbsterone-Science 247: tions to hippocampus and neocortex. J Neurosci 9:2999-30 18. 704-707. Koike S, Sakai M, Maramatsu M (I 987) Molecular cloning and char- Pfaff D, Keiner M (1973) Atlas of estradiol concentrating cells in the acterization of the rat estrogen receptor cDNA. Nucleic Acids Res central nervous system of the female rat. J Comp Neural 15 1:121- 15:2499-2513. 158. Lamballe F, Klein R, Barbacid M (I 99 1) trkC, a new member of the Radeke MJ, Misko TP, Hsu C, Herzenberg LA, Shooter EM (1987) family of the trk family of tyrosine protein kinases, is a receptor for Gene transfer and molecular cloning of the rat nerve growth factor neurotrophin-3. Cell 66:967-979. receptor. Nature 325:593-597. Lauber AH, Roman0 GJ, Mobbs CV, Pfaff DW (1990) Estradiol reg- Rich KM, Disch SP, Eichler ME (1989) The influence of regeneration ulation of estrogen receptor messenger ribonucleic acid in rat me- and nerve growth factor on the neuronal cell body reaction to injury. diobasal hypothalamus: an in situ hybridization study. J Neuroen- J Neurocytol 18:569-576. docrinol 2:605-6 12. Rodriguez-Tebar A, Rohrer H (1991) Retinoic acid induces NGF- Lindsay RM (1988) Nerve growth factors (NGF, BDNF) enhance dependent survival response and high-affinity NGF receotors in im- axonal regeneration but are not required for survival of adult sensory mature chick sympathetic neurons.Development 1 12:8 i3-820. neurons. J Neurosci 8:2394-2405. Roman0 GJ, Harlan RE, Shivers BD, Howells RD. Pfaff DW (1988) Loeb DM, Maragos J, Martin-Zanca D, Chao MV, Parada LF, Greene Estrogen increases proenkephalin messenger ribonucleic acid levels LA (199 1) The trk proto-oncogene rescues NGF responsiveness in in the ventromedial hypothalmus of the rat. Mol Endocrinol 2: 1320- mutant NGF-nonresponsive PC1 2 cell lines. Cell 66:96 1-966. 1328. Loy R, Gerlach JL, McEwen BS (1988) Autoradiographic localization Sar M, Stumpf WE (1977) Androgen concentration in motor neurons of estradiol-binding neurons in the rat hippocampal formation and of cranial nerves and spinal cord. Science 197:77-79. entorhinal cortex. Brain Res 467:245-25 1. Sehgal A, Patil N, Chao MV (1988) A constitutive promoter directs Lu B, Buck CR, Dreyfus CF, Black IB (1989) Expression of NGF and expression of the nerve growth factor receptor gene. Mol Cell Biol NGF receptor mRNAs in the developing brain: evidence for local 8:3160-3167. delivery and action of NGF. Exp Neural 104: 19 1-199. Shughrue PJ, Refsdal CD, Dorsa DM (1992) Estrogen receptor mes- Lustig RH, Sudol M, Pfaff DW, Federoff HJ (1991) Estrogenic regu- senger ribonucleic acid in female rat brain during the estrus cycle: a lation and sex dimorphism ofgrowth-associated protein 43 kDa (GAP- comparison with ovariectomized females and intact males. Endocri- 43) messenger RNA in the rat. Mol Brain Res 11:125-132. nology 131:381-388. MacLuskv NJ, McEwen BS (1978) Oestrolren modulates oroaestin Shupnik MA, Rosenzweig BA (1991) Identification of an estrogen receptor concentrations in some rat brain regions but not cn others. response element in the rat luteinizing hormone beta gene. DNA- Nature 2741276-278. estrogen receptor interactions and functional analysis. J Biol Chem MacLusky NJ, Brown TJ, Jones E, Leranth C, Hochberg RB (1990) 266:17084-17091. Autoradiographic and microchemical methods for quantitation of Shupnik MA, Gordon MS, Chin WW (1989) Tissue specific regulation steroid receptors. In: Methods in neurosciences, Vol 3, Quantitative of rat estrogen receptor mRNAs. Mol Endocrinol 3:660-665. and qualitative microscopy (Conn PM, ed), pp l-35. New York: Simerly RB, Young BJ (1991) Regulation of estrogen receptor mes- Academic. senger ribonucleic acid in rat hypothalamus by sex steroid hormones. Martin-Zanca D, Oskam R, Mitra G, Copeland T, Barbacid M (1989) Mol Endocrinol 5:424432. Molecular and biochemical characterization of the human trk proto- Simerly. . RB,^. Chang C, Muramatsu M, Swanson LW (1990) Distri- oncogene. Mol Cell Biol 9:24-33. butlon ot androgen and estrogen receptor mRNA containing cells in The Journal of Neuroscience-, February 1994. f4(2) 47*

the rat brain: an i~r silrr hybridization study. J Comp Neurol 294:76- amus and preoplic area in vitro. III. Efl&ts of estrogen on dcndritic 95. differentiation. Dev Brain Res 7:97-101. Soppet D, Escandon E, Marangos J, Middlemas DS, Reid SW, Blair J, Toran-Allerand CD, Miranda RC, Bentham W, Sohrabji F, Brown EJ, Burton LE, Stanton BR, Kaplan DR, Hunter T, Nikolics K, Parada Hochberg RB, MacLusky NJ (1992a) Estrogen receptors co-localize LF (1991) The neurotrophic factors brain-derived neurotrophic fac- with low-affinity NGF receptors in cholinergic neurons of the basal tor and ncurotrophin-3 are ligands for the rrkB tyrosine kinase re- forebrain. Proc Nat1 Acad Sci USA 89:46684672. ceplor. Cell 65:895-903. Toran-Allcrand CD, Miranda RC, Hochberg RB, MacLusky NJ Squinlo SP, Stilt TN, Aldrich TH. Davis S, Bianco SM, Radziejewski (1992b) Cellular variations in estrogen receptor mRNA translation C, Glass DJ, Masiakowski P, Furth ME, Valenzuela DM, DcStefano in the developing brain: evidence from combined [“‘Iestrogen au- F’S, Yancopoulos GD (199 I) IrkB encodes a functional receptor for toradiography and non-isotopic in si!ii hybridization histochcmistry. brain derived ncurotrophic factor and ncurotrophin-3 but not nerve Brain Res 576:254 I. growth factor. Cell 65:885-893. Waterman ML, Adler S. Nelson C. Greene GL. Evans RM, Rosenfeld Stanley HF, Fink G (1986) Synthesis of specific brain proteins is MG (1988) A single domain of the estrogen receptor confers de- influenced by testosterone at mRNA level in the neonatal rat. Brain oxyribonucleic acid binding and transcriptional activation of the rat Kes 370~223-231, prdactin gcnc. Ma1 Endocrinol 2: 14-2 L. Stanley HF, Borthwick NM, Fink G (1986) Brain Drotein changes Weisz A, Rosalcs R (1990) Identification of an estrogen response during development and sexual differentiaiion in th; rat. Brain ties element upstream of the human C#US gene that binds-the estrogen 370:2 15-222. receptor and the AP- 1 transcription factor. Nucleic Acids Res IS: Taniuchi M, Schweitzer JB, Johnson EM (I 986) Nerve growth factor 5097-5 106. receptors in the tar brain. Proc Nat1 Acad Sci-USA 83~i95&1954. Wilcox JN, Roberts JL (1985) Estrogen decreases rat hypothalamic Thoenen H. Barde YA (1980) Phvsioloav of nerve arowth factor. proopiomelanocortin messenger ribonucleic acid levels. Endocrinol- Physiol Rev 60:1284-1335. . -- ogy I 17~2392-2396. Toran-Allerand CD (1976) Sex steroids and the development of new- Wright LL, Smolen AJ (1983a) Effcctsof Ii’-fl-eslradiol on developing born mouse hypothalamus and preoptic areas in rirru: implications superior cervical ganglion neurons and synapses. Dev Brain Res 6:299- for sexual differentiation. Brain Res 106:407-4 12. 303. Toran-Allerand CD (I 980) Sex steroids and the development of new- Wright LL, Smolen AJ (1983b) Neonatal lcstosteronc treatment in- born mouse hypothalamus and prcoptic area in vifra. II. Morpholog- creases neuron and synapse numbers in male rat superior cervical ical correlates and hormonal specificity. Brain Res I89:4 13-427. ganglion. Dev Brain Res 8: 145-I 53. Toran-Allerand CD (I 984) On the genesis of sexual differentiation of Wright LL, Marchetti D, Perez-Polo JR (1988) ElTects of gonadal the central nervous system: morphogenetic consequences of stcraidal steroids on nerve growth factor receptors in sympathetic and sensory exposure and possible role of a-feloprotcin. Prog Brain Res 6 I :63- ganglia of neonata rats. Int J Dev Neurosci k:2i7-222. 98. Yan 0.-.. Johnson EM (1987) A quantitative study ofthedcvclonmental Toran-Allerand CD, Miranda RC (I 993) Combining non-isotopic in expression of nervd growth fa’Ctor (NGF) rcciptor in rats. dev Biol si/rr hybridization with steroid auloradiography. In: Ijt si~r~ hybrid- 121:139-148. ization in neurobiology, 2d ed (Eberwine J, Barchus J, Vatentino K, Yan Q, Johnson EM (1988) An immunohistochemical study of the eds). New York: Oxford UP. nerve growth factor receptor in developing rats. J Neurosci 8:348 l- Toran-Allerand CD, Gerlach JL, McEwcn BS (I 980) Autoradiograph- 3496. - ic localization of [3H]estradiol related to steroid responsiveness in Yu WHA, Gibbs RB, Pfaff DW (199 1) Effects of testosterone on levels cultures of the newborn mouse hypothalamic and preoptic area. Brain of ribosomal RNA (rRNA) in hypoglossal neurons following nerve Res 184:s 17-522. injury. Sot Neurosci Abstr 17:226. Toran-Allerand CD, Hashimoto K, Grcenough WT, Saltarelli M (1983) Sex steroids and the development of the newborn mouse hypothal-