Quantitative and Qualitative Microscopy

By P. Michael Conn

Methods in Neurosciences, Volume 3: Quantitative and Qualitative Microscopy is a collection of papers that deals with microscopic techniques in statistical measures. This volume describes microscopy using sophisticated stains and dyes to advance observation of tests and experiments. Section I describes autoradiography including micro chemical methods, high-resolution autoradiography, and single- or double-label quantitative autoradiography for use in imaging of brain activity patterns or determining cerebral physiology. Section II discusses the quantification of structures through statistical and computational methods including dynamic video imaging technology. Section III explains the use of tracers, toxins, or dyes in tracing neuronal connections. One paper addresses the use of small injections of axonally transported fluorescent tracers. Section IV explains staining technology such as using the silver impregnation method for frozen sections of human nervous tissue that are gathered from tissues preserved in formalin. Section V addresses freezing techniques and those using freeze-fracture methods in neurobiology. The text also discusses cryoprotection and other freezing methods to control ice crystals found in fixed or unfixed brain tissues. Section VI presents the combined and high-resolution methods in polarization microscopy and microscopic investigations. Cellular biologists, micro-chemists, and scientific researchers in the field of micro- and cellular biology will appreciate this book.

• Language: English
• Publisher: Academic Press
• Release date: Oct 22, 2013
• ISBN: 9781483268170

Book preview

volume.

Methods in Neurosciences

Edited by P. Michael Conn

Volume 1 Gene Probes

Volume 2 Cell Culture

Volume 3 Quantitative and Qualitative Microscopy

Volume 4 Electrophysiology and Microinjection (in preparation)

Section I

Autoradiography

[1]

Autoradiographic and Microchemical Methods for Quantitation of Steroid Receptors

N.J. MacLusky, T.J. Brown, E. Jones, C. Leranth and R.B. Hochberg

Publisher Summary

During the past few years, considerable progress has been made toward the development of specific new methods for the analysis of steroid receptor concentrations in discrete, anatomically defined regions of the brain. These methods are based on two experimental approaches. The first method includes a variety of modifications of earlier subcellular fractionation-based steroid receptor assay procedures combined with the Palkovits microdissection method to achieve the necessary anatomical resolution. The second approach is based on quantitation of the uptake of radiolabeled steroids by the tissue through densitometric analysis of autoradiograms prepared by the exposure of tissue sections against film. This chapter outlines the principal features of these two methodological strategies: microchemical measurement of steroid receptors in the central nervous system (CNS) and quantitative autoradiography for the measurement of steroid receptor levels in the central nervous system. Measurement of steroid receptor levels in individual microdissected cell groups from the CNS, however, presents special technical problems that restrict the range of methodology that can be successfully applied. For more than two decades, the primary method for high-resolution localization of steroid binding sites in the brain has been autoradiography. Perhaps the most exciting potential application of these methods reviewed in the chapter is the possibility for integrating high-resolution measurements of steroid receptor levels with other biochemical and anatomical techniques. Using combinations of experimental approaches, it may be possible to achieve a far greater understanding of the mechanisms of steroid hormone action on the brain.

Introduction

Over the past two decades, biochemical methods have been established for measurement of all five major classes of steroid hormone receptor in the central nervous system (CNS). Many of these procedures also allow selective estimation of the extent to which the receptors are occupied by endogenous circulating hormones. Despite the wide range of methodology available, however, progress toward defining the relationship between steroid–receptor interactions in different regions of the brain and responsiveness to changes in circulating steroid levels has been relatively slow. A major problem is that the majority of the available steroid receptor assay methods are relatively insensitive, which necessitates use of large tissue samples and/or pooling of tissue from several animals for each determination. This limitation seriously constrains attempts to correlate receptor binding with physiological responses. Studies on the effects of intracerebral hormone implants strongly suggest that neuroendocrine and behavioral responses to gonadal steroids may require exposure of only one, or a few, highly localized target structures to estrogen (1–3). Moreover, several reports have suggested that steroid receptor concentrations in different regions of the brain may be selectively influenced by such factors as reproductive status and afferent neural input (4–7). Thus, overall measurements of receptor concentrations in crudely dissected brain samples may give a misleading picture of the relationship between receptor binding and responses to circulating steroids.

The CNS presents a unique challenge with respect to the methodology for steroid receptor assays. Not only are overall receptor concentrations relatively low, in comparison to steroid target organs such as the liver and reproductive tract, but the target cells in the brain are distributed heterogeneously, dispersed in groups among much larger numbers of nontarget neurons and glia (8,9). This makes measurements of steroid receptors in the brain particularly difficult, since the assay methods must combine levels of sensitivity and specificity that are considerably greater than those required for most other steroid–sensitivity structures.

During the past few years, considerable progress has been made toward the development of specific new methods for the analysis of steroid receptor concentrations in discrete, anatomically defined regions of the brain. These methods are based on two experimental approaches. The first includes a variety of modifications of earlier subcellular fractionation-based steroid receptor assay procedures combined with the Palkovits microdissection method (10) to achieve the necessary anatomical resolution. The second approach is based on quantitation of the uptake of radiolabeled steroids by the tissue through densitometric analysis of autoradiograms prepared by exposure of tissue sections against film. This article briefly outlines the principal features of these two methodological strategies.

Microchemical Measurement of Steroid Receptors in the Central Nervous System

General Principles

In common with nonneural target tissues, steroids appear to act on the brain through receptor proteins that are recovered in the soluble cytoplasmic (cytosol) fraction of the cell after subcellular fractionation. Binding of the steroid to its receptor site results in transformation of the receptor to a form that binds tightly within the cell nucleus, from which it can be extracted only by disruption of the chromatin using DNase or high-ionic-strength buffers (11–14). The different properties of the native and hormone-bound forms of steroid receptors necessitate use of somewhat different methodologies for assay of occupied and unoccupied receptor sites. Since the majority of unoccupied receptors are extracted into the cytosol after tissue homogenization, these receptors can be assayed relatively easily by equilibrating tissue cytosol fractions with appropriate stereospecific labeled ligands, followed by separation of the receptor-bound and free components. Nuclear bound steroid–receptor complexes are separated by preparation of a cell nuclear fraction from which the majority of unbound receptors are eliminated by washing with low-ionic-strength buffer. The remaining receptors are then labeled by exchanging a radioactive ligand for the unlabeled steroid in the nuclear-bound receptor complex by incubating either the intact nuclei or a soluble nuclear extract with the labeled ligand (for a general review of steroid receptor assay procedures, see Ref. 11).

Cytosol Steroid Receptor Assays

The major differences between cytosol steroid receptor assay methodologies involve the technique used for separation of bound and unbound ligand. For studies of the steroid receptor content of relatively large regions of the brain (e.g., the entire hypothalamus and preoptic area) a wide variety of such techniques have been employed. Measurement of steroid receptor levels in individual microdissected cell groups from the CNS, however, presents special technical problems which greatly restrict the range of methodology that can be successfully applied. There are three main problems. The first concerns the process for dissection of the tissue, which must be accomplished without significant degradation of the relatively sensitive steroid receptor molecules. Second, the extremely small tissue samples must be homogenized under conditions which will ensure reproducible extraction of the nonnuclear-bound receptor sites. Finally, separation of the receptor-bound hormone must be accomplished with maximum efficiency to allow detection of the small number of receptors present in individual microdissected tissue samples.

The procedure currently in use in our laboratories for measurement of cytosol steroid receptors is as follows.

General Cytosol Steroid Receptor Assay Procedure

For cytosol steroid receptor assays in microdissected regions of the brain, the tissue must initially be frozen and sectioned to allow dissection using the Palkovits punch technique (10). In the rat brain, no special precautions are necessary for preservation of the receptors during the microdissection procedure. However, perfusion of the animals via the left ventricle with ice-cold aqueous 10% (v/v) dimethyl sulfoxide prior to removal of the brain may offer some advantages for progestin receptor assays (15); this point will be discussed further below. The brain is rapidly removed, blocked with a razor blade, mounted in the required orientation on a cryostat chuck, and frozen on dry ice. Serial 300-μm-thick coronal sections are cut and thaw-mounted on glass slides. Regions of interest are then removed from the frozen sections under a dissecting microscope using stainless steel needles ranging in diameter from 0.3 to 1 mm. Figure 1 illustrates the scheme that we currently use for studies of the regulation of estrogen and progestin receptor levels in the adult rat brain (PVPOA, periventricular preoptic area; mPOA, medial preoptic area; BNST, bed nucleus of the stria terminalis; ARC, arcuate nucleus; ME, median eminence; VMN, ventromedial nucleus of the hypothalamus; AMYG, pooled cortical and medial nuclei of the amygdala).

Fig. 1 Microdissection scheme for the hypothalamus and preoptic area. Twelve consecutive 300-μm thick sections are cut from each brain and mounted onto microscope slides. Brain nuclei are removed from the frozen sections with a 500-μm or 1000-μm diameter stainless steel punch. The PVPOA (A) is dissected by taking two midline 500-μm punches from sections 1–3; the mPOA (B) and BNST (C) are dissected by taking bilateral 1000-μm punches from sections 1–3; the ARC-ME (D) is dissected by taking a single partial 1000-μm punch from sections 8–12; the VMN (E) is dissected by taking bilateral 1000-μm punches from sections 9–12; and the AMYG (F) is dissected by taking bilateral double 1000-μm punches from sections 9–12. [Reproduced with permission from Brown et al. (62).]

The frozen plugs of tissue at the tip of the dissecting needle are pushed out into 200–250 μl of ice-cold TEGD buffer [10 mM Tris–HCl, 1.5 mM EDTA, 10% (v/v) glycerol, 1 mM dithiothreitol] using a fine wire stylet inserted through the lumen of the needle. The tissues are then homogenized, the homogenates are centrifuged (105,000 g for 20 min, using either a Beckman Airfuge or the Beckman TL-100 bench top ultracentrifuge), and the supernatant cytosol fractions decanted and used immediately.

Incubations of aliquots (75–100 μl) of the cytosol fractions with labeled ligands are performed under conditions appropriate for the specific class of receptors under study. No specific modifications are required in adapting incubation conditions developed for larger tissue samples to microdissected specimens. The labeled steroids are added to each aliquot of cytosol in a small volume (50 μl) of TEGD buffer and incubated at 0–4°C for a period of between 4 and 18 hr. Control incubates, required in order to correct the results for nonspecific binding, contain equal volumes of cytosol and the labeled ligand solution, in the presence of a 50- to 100-fold molar excess of an unlabeled competitor chosen to inhibit specifically binding of the radioactive ligand to the receptors. Bound radioactivity in the incubates is separated on small columns (7 × 32 mm) of Sephadex LH-20 (16) prepared in plastic (2 ml capacity) pipette tips maintained in either a cold room or a refrigerator at 0–4°C. Aliquots (100 μl) of each incubate are loaded onto the columns and washed into the column bed with 100 μl of ice-cold TEGD buffer. It is not necessary to keep the column surfaces covered, since surface tension prevents the column bed from drying out. Thirty minutes after sample loading, the macromolecular fraction is eluted directly into scintillation vials with 400 μl of buffer and counted.

Factors Affecting Performance of Cytosol Receptor Assays

These basic procedures have now been used successfully in a number of laboratories for studies of steroid receptor levels in the brains of several mammalian and avian species (6,17–23). Although the methods appear to be fairly easy to adapt to different steroids and new species, considerable variations have emerged in specific aspects of the methodology as well as the results obtained.

Dissection Procedure

Microdissection of the CNS can be accomplished using either fresh, chilled tissue slices (24) or frozen tissue using the Palkovits punch dissection methodology described above. We have found the latter procedure to be preferable in terms of the greater anatomical resolution that can be achieved using frozen tissue slices. Our initial attempts to adapt the Palkovits methodology for use with steroid receptors, however, encountered serious problems, particularly in the case of the more labile receptor populations, such as progestin receptors (20). We reasoned that the most likely cause for these losses was degradation of the receptors during the microdissection procedure. During the course of this procedure, the tissue is frozen and thawed twice; once during preparation and mounting of the tissue sections on glass slides, and a second time during dissection and subsequent homogenization of the tissue samples. To minimize possible damage to the receptors during freezing, we examined the effects of equilibrating the tissues with a cryopreservative solution containing 10% dimethyl sulfoxide (DMSO) prior to freezing. Although this modification appeared to be successful, in that for the first time it became possible to assay estrogen and progestin receptors in microdissected regions of the rat brain (20), confusion subsequently emerged in the literature with respect to the effects of DMSO with which different laboratories obtained substantially different results. The original reports by Parsons et al. (20) and by Rainbow et al. (21) suggested that perfusion of the tissues with aqueous DMSO prior to freezing was essential for accurate determination of levels of progestin receptors, but that this procedure increased losses of estrogen receptors by about 30% compared to results obtained in fresh tissues. Other studies failed to confirm either the extent of the protection conferred by DMSO on the progestin receptor assays or the DMSO-induced loss of estrogen binding (15,18,25). We observed no losses of androgen receptors in DMSO-preserved frozen rat brain tissue (15); in contrast, Roselli et al. (25) have reported that DMSO results in enhanced losses of nuclear androgen receptors compared to normally frozen brain specimens.

Although it is not possible post hoc to determine the precise reasons for these disparities, the experience with these assays in our laboratories suggests a number of possible explanations. While DMSO can afford a degree of protection for steroid receptors in frozen brain tissue, the extent of this protection is dependent on several factors, including (a) the nature of the tissue and the receptor population, (b) the speed with which the dissection procedure is carried out, and (c) less well-defined factors related to the purity and storage characteristics of commercially available DMSO. (a) In the rat brain, while estrogen, androgen, and progestin receptors all appear to be protected to some extent by freezing with DMSO, corticosteroid receptor levels are lower after freezing with DMSO than in unprotected frozen tissue (15). For gonadal steroid receptors the extent of the protection is greater in the pituitary than in the brain, and, at least in the case of estrogen and androgen receptors, appears to be more marked for nuclear than for cytosolic receptor sites (15). (b) Although initial studies suggested that recovery of progestin receptors from rat brain is low in the absence of DMSO (20), subsequent work failed to demonstrate any significant protective effect of DMSO on progestin receptor levels in the rat brain. Studies in our laboratory indicate that DMSO is not essential even for measurement of progestin receptors in frozen microdissected samples of brain tissue (15). However, for the Palkovits punch dissection procedure this is only true once a reasonable amount of technical experience has been accumulated, so that processing of the tissue sections proceeds rapidly without any extended periods of exposure of the tissue to temperatures above 0°C. Thus, variable or low estimates of progestin receptor concentrations may at first be obtained in non-DMSO-preserved tissue, unless care is taken to minimize the time taken between sectioning the tissue and processing of the dissected tissue samples. (c) Commercially available DMSO preparations may contain impurities that interfere significantly with steroid receptor assays. These impurities accumulate slowly on standing, probably as a result of oxidative reactions—we have repeatedly observed that stock DMSO bottles opened more than 1–2 months previously do not protect steroid receptor sites in frozen rat brain as reliably as freshly opened material. Therefore, if DMSO is to be used, it is vitally important that it be fresh and of the highest possible quality.

Homogenization and Incubation Conditions

We have found the above general procedure to be suitable for measurement of androgen, estrogen, progestin, and corticosteroid binding sites in the rat brain. Modifications to the homogenization buffer compositions and incubation conditions may, however, be required under some circumstances. Although we have not noted any significant positive effects of molybdate ion (26,27) on the recovery of cytosol steroid-receptor complexes from the rat brain, the addition of 10–20 mM sodium molybdate to the homogenization medium does not interfere with the microdissection-based assays and may be advantageous under some circumstances, particularly for androgen and glucocorticoid receptor assays (15,28,29) or for assays in tissues from species other than the rat.

A technical problem that has considerable impact on the efficiency of both cytosol and nuclear receptor assays concerns the homogenization procedure. Efficient, reproducible homogenization of the extremely small microdissected samples is required if reliable results are to be obtained. Excessively vigorous homogenization may lead to damage to the cell nuclei, as well as the possible release of lysosomal enzymes that may degrade the receptors during incubation. On the other hand, if homogenization is not sufficiently vigorous, small tissue fragments may escape homogenization completely. Whatever homogenization method is used, the entire process must be accomplished in a very small volume, with minimal losses in transferring the homogenate for subsequent subcellular fractionation. Our homogenization procedure is based on the general method used by Luine and co-workers (30,31) for studies of neurotransmitter levels in punch microdissected samples of the rat brain. The homogenization buffer and tissue fragments are placed into small (500 μl capacity) disposable glass tubes at the time of dissection. After the dissection has been completed, homogenization is accomplished by making 2 × 10 passes up and down each tube with a slow-speed motor-driven Teflon pestle, machined to give an all-round clearance in the tubes of 0.125 mm. This procedure minimizes losses of tissue and ensures even and reproducible extraction of cytosol steroid receptors.

Separation of Bound and Free Steroid

A wide variety of methods have been used in the past to separate bound and free labeled steroids for steroid receptor assays. The Sephadex LH-20 gel filtration method, however, is currently the only one used with microdissected samples on CNS tissue. This method, originally introduced by Ginsburg et al. (16), is capable of essentially complete separation of receptor-bound and unbound steroids. As compared to other methods for separation of macromolecular-bound steroids, Sephadex LH-20 has a number of advantages, chief among which is the fact that its efficiency is relatively unaffected by wide variations in either protein concentration or ionic strength. Compared to Sephadex G-25, from which it is derived, it usually exhibits negligible nonspecific adsorption of steroid-receptor complexes (16). Occasionally, however, there can be significant variations between different manufacturers’ lot numbers of Sephadex LH-20 in this aspect of their performance. Several laboratories using Sephadex LH-20-based receptor assays have observed that some lot numbers of Sephadex LH-20 may absorb as much as 35–40% of the receptor complexes applied (E. J. Roy, personal communication; N. J. MacLusky, unpublished observations). Fortunately this is rare; the problem can be obviated by pretesting samples of different lot numbers of Sephadex LH-20, to eliminate lots prior to use.

The second major variable that can affect the outcome of steroid receptor assays performed using Sephadex LH-20 is the level of nonspecific bound radioactivity in the macromolecular bound fraction. This nonspecific binding has two components: radiolabeled steroid bound to nonreceptor proteins, and unbound radioactivity not retained by the columns. Since the separation of bound and free steroid on Sephadex LH-20 is based on both gel filtration and hydrophobic adsorption, steroids are essentially quantitatively separated from the macromolecular-bound fraction by even very small columns. However, this is not the case for polar radiochemical decay products which can build up to a significant extent during storage of radiolabeled steroids. Such impurities are not removed by Sephadex LH-20 columns as efficiently as the intact steroid, seriously degrading the performance of the assay. We have in the past used a number of complementary approaches to minimize these problems. First, where nonspecific binding remains a problem after passage of the macromolecular-bound steroid through Sephadex LH-20, a further reduction in nonspecific binding may be possible by selective precipitation of the receptor complexes from the column eluates, e.g., using protamine sulfate (32) or hydroxylapatite (33). Second, monthly repurification of stock radiolabeled ligands by column partition chromatography or high-performance liquid chromatography (HPLC) tends to minimize blank levels of radioactivity (16). Finally, overnight extraction of the column eluates into a nonaqueous miscible scintillator prior to counting can also significantly reduce assay blanks, since any polar radiolabeled impurities carried through the columns will remain in the aqueous phase and not be counted.

Nuclear Receptor Exchange Assays

Nuclear receptor exchange assays are carried out on microdissected brain tissue specimens using essentially the same procedures as those described above, except that a high-salt extract of cell nuclei is used in place of a cytosol fraction. The key to successful assay of nuclear-bound steroid receptors in microdissected tissue samples is the preparation of the cell nuclear fraction. This preparation must be achieved rapidly and in high yield.

The majority of work carried out so far on methods for assay of nuclear steroid receptors in microdissected regions of the brain has centered on studies of estrogen binding. We will therefore initially restrict discussion of nuclear steroid receptor assays to the estrogen receptor, and consider applications of the methodology to other receptor systems at the end of the section.

Nuclear Estrogen Receptor Assays

Buffers

BI: 1 mM KH2PO4, 3 mM MgCl2, 0.32 M sucrose, pH 6.5

BII: 1 mM KH2PO4, 1 mM MgCl2, 2.4 M sucrose, 0.25% (v/v) Triton X-100, pH 6.5

BIII: 1 mM KH2PO4, 1 mM MgCl2, 1.8 M sucrose, pH 7.0

TEGD: 10 mM Tris-HCl, 1.5 mM ethylenediaminetetraacetic acid (EDTA), 10% (v/v) glycerol, 1 mM dithiothreitol, pH 7.4

TEGDB: TEGD buffer, containing 0.5 mM bacitracin.

TEGDBK: TEGDB buffer containing 0.8 M KCl

Tissue Dissection

For nuclear estrogen receptor (ERn) assays in microdissected regions of the rat brain, we routinely perfuse the animals via the left ventricle with 40 ml of ice-cold aqueous 10% dimethyl sulfoxide under ether anesthesia (15). This perfusion procedure is probably not essential for protection of ERn in rat brain tissue. However, as discussed above, our experience is that ERn concentrations in DMSO-frozen tissue are identical to those in fresh brain, and the inclusion of DMSO provides an added measure of protection for the tissue during the freezing and thawing process. Subsequent dissection of the brain proceeds exactly as described above for the cytosol receptor assays.

Nuclear Isolation Procedures

We have developed two basic strategies for the preparation of cell nuclear fractions from frozen microdissected regions of the brain. The first (which we have termed the one-tube method) is designed to achieve a high degree of purity in the cell nuclear fraction. As such, it completely eliminates contributions from cytoplasmic factors—including proteolytic enzymes that may cause considerable problems during subsequent incubation of the receptors with labeled estradiol. However, this method is not without drawbacks in terms of the difficulties associated with preparing purified nuclei from brain tissue in sufficiently high yield. For this reason, a second, simpler procedure, the TEGD-wash method, was subsequently developed which compromises on the purity of the nuclear preparation in order to maximize the recovery of nuclear receptor complexes.

One-Tube Method

This procedure essentially represents a modification of the method first described by Roy and McEwen (34), which has been widely used for studies of ERn levels in crudely dissected regions of the rat brain. Initial attempts to study ERn levels in punch-microdissected tissue samples using the Roy and McEwen procedure were unsuccessful. Although some binding was observed, the recovery of nuclei was frequently poor (10–30%), which necessitated pooling tissues from many animals in order to obtain sufficiently high ERn levels. The problem appeared to be that small numbers of nuclei were lost during each of the multiple washing and centrifugation steps involved in this technique. A procedure introduced by Keiner et al. (35,36) offered a potential means of reducing the extent of these losses. With this method, tissue homogenates are overlayed directly onto dense sucrose solution and centrifuged, the nuclei sedimenting as a pellet while cytosol-extracted estrogen receptors remain near the top of the tube. Ultrastructural examination of nuclear pellets prepared using the Keiner et al. (35,36) methodology, however, revealed considerable cytoplasmic contamination (C. Leranth and N. J. MacLusky, unpublished observations). The one-tube procedure was developed by fusing the principal features of the Roy–McEwen and Keiner et al. methods to achieve a rapid, high-yield nuclear isolation without compromising the purity of the final nuclear preparation.

Our current protocol for the one-tube nuclear isolation procedure is as follows: Punch-microdissected tissue samples are homogenized in 300 μl of ice-cold BI buffer and mixed with 500 μl of BII buffer. The homogenizer is rinsed with 2 × 100 μl of BI buffer, and the rinses combined with the homogenate. BIII buffer (1.5 ml) is added down the side of the tube wall, displacing the homogenate–BII mixture upward. The tubes are left on ice for 2–3 min to allow the two layers to settle, then centrifuged for 35 min at 20,000 g. The supernatants are discarded and the sides of the centrifugation tubes are carefully wiped clean with tissues. The nuclear pellets are resuspended in 100–200 μl of TEGDB buffer and salt-extracted by addition of an equal volume of TEGDBK buffer. After 30 min on ice, the nuclear residue is sedimented by centrifugation at 105,000 g for 10 min. The supernatants, containing the extracted ERn, are decanted and assayed immediately.

TEGD-Wash Method

The second procedure for preparation of nuclei prior to ERn assay is based on the method of Anderson et al. (37). Tissues are homogenized in 200 μl of ice-cold TEGD buffer and centrifuged in a Microfuge (Beckman Instruments) at 8700 g for 1 min at 0–4°C. The pellets are washed twice by resuspension in 200 μl of TEGD buffer, followed by centrifugation at 8700 g for 1 min. The final crude nuclear pellet is resuspended in TEGDB buffer and salt-extracted by addition of an equal volume of TEGDBK buffer as described above for the pellets from the one-tube assay. The extracted residue is sedimented by centrifugation for 1 min at 8700 g in the microfuge.

ERn Assay Incubation Conditions

Labeling of the extracted receptors is performed in exactly the same way for both the one-tube and TEGD-wash procedures. The incubation conditions used in the assay depend on whether it is necessary to obtain separate measurements of unoccupied and total ERn concentrations. If measurements of unoccupied nuclear receptor sites are required, a two-step incubation is performed (38). Selective labeling of the unoccupied receptor fraction is accomplished by initially incubating the nuclear extracts with 2 nM [³H]estradiol for 2.5 hr at 0–4°C. Aliquots of the incubates are withdrawn and chromatographed on Sephadex LH-20, as described above for the cytosol receptor assays. The remainder of each incubate is then warmed to 25°C for 4.5 hr to allow labeling of all of the receptor complexes present through exchange of any bound unlabeled hormone (38). If measurements of unoccupied nuclear receptors are not required, the initial low-temperature incubation can be omitted. Results are standardized on the basis of the quantity of DNA recovered in the insoluble nuclear residue after salt extraction.

Comparative Properties of One-Tube and TEGD-Wash Procedures

The one-tube procedure represents a minor modification of a similar method we first reported 5 years ago (39). The changes made since this initial publication, although small, are significant in that they dramatically improve the performance of the assay with microdissected tissue samples. Compared to the earlier method (39), the sucrose concentrations in both the top sample layer and the lower dense sucrose layer through which the nuclei are centrifuged have been slightly reduced, and the duration of the final centrifugation step has been increased (from 30 to 35 min). These modifications increase the yield of nuclei in the final nuclear pellet, without compromising the purity of the nuclear preparation. In addition, the g force used to sediment the nuclear residue remaining after salt extraction has been increased more than 5-fold, from 20,000 to 105,000 g. This modification is vitally important for studies in relatively small regions of the brain (e.g., the periventricular preoptic area). With very small tissue samples (<5 μg DNA), centrifugation at 20,000 g for 10 min may not be sufficient to achieve quantitative sedimentation of the extracted chromatin. By using a higher speed centrifugation after salt extraction, this problem is eliminated.

Over the past few years, we have extensively compared the performance of the one-tube and Roy–McEwen procedures in order to validate the former thoroughly. At relatively high tissue sample loads, ERn data obtained with the two methods are indistinguishable (Fig. 2). The quality of the nuclear preparations obtained with the one-tube method is at least the equal of that obtained with the conventional Roy–McEwen technique. The Roy–McEwen and one-tube methods both yield nuclei that are essentially free of cytoplasmic contamination, the pellets from the Roy–McEwen method exhibiting somewhat greater evidence of damage to the nuclear envelope (Fig. 3) presumably because of the greater length of this procedure and the effects of repeated centrifugation and resuspension of the nuclei. Size distribution studies of the nuclei in the pellets also provide evidence of slightly greater damage in the case of the Roy–McEwen method (Fig. 4). Some disparity between the methods is apparent with respect to the relative proportions of small (<7 μm) diameter nuclei, possibly resulting from more pronounced nuclear swelling in the case of the Roy–McEwen procedure during exposure to Triton X-100 (40). No significant differences are observed between the two methods, however, with respect to the intranuclear morphology of the purified nuclear fractions: both yield nuclear preparations containing approximately 30% oligodendrocyte and other small glial nuclei, the remainder representing neuronal or astrocytic nuclei (N.J. MacLusky and C. Leranth, unpublished observations). Thus, overall, the Roy–McEwen and one-tube methodologies can be considered to be essentially comparable in terms of the data that they provide.

Fig. 2 Dose–response curves for nuclear estrogen receptor (ERn), concentrations in (A) the pituitary and (B) pooled mediobasal hypothalamus + preoptic area after i.v. injection of 17β-estradiol. Ovariectomized rats were injected with estradiol (0.02–200 μg/kg body weight, via the jugular vein) 1 hr prior to sacrifice and ERn concentrations determined by exchange assay using either the Roy–McEwen (solid symbols) or one-tube (open symbols) procedures.

Fig. 3 Morphology of the nuclei recovered after subcellular fractionation using the Roy–McEwen and one-tube procedures. Nuclear pellets were fixed immediately after preparation in 0.1 M phosphate buffer, pH 7.4, containing 0.5% (v/v) glutaraldehyde and 3% (w/v) paraformaldehyde. The pellets were transferred to BEEM capsules (Polyscience Inc., Philadelphia, PA), osmicated in 1% (w/v) osmium tetroxide for 15 min, dehydrated in ethanol, then embedded in Em-Bed 812 (EMS, Philadelphia, PA). Ultrathin sections were contrasted with lead citrate and examined in a Hitachi HU-12 electron microscope. (B) and (C) Nuclei from the one-tube and Roy–McEwen procedures, respectively, at identical magnifications. (A) Higher magnification of the edge of a neuronal cell nucleus isolated using the one-tube procedure: the nuclear envelope has been almost completely removed by the Triton X-100 added to the sample (BII) buffer. Bars: 1 μm.

Fig. 4 Nuclear pellets prepared from the pooled hypothalamus, preoptic area, and amygdala of ovariectomized female rats were examined by light microscopy (magnification: × 100). Photographs of the nuclei were then projected at a final magnification of ×550 onto a white sheet of paper placed over an Apple graphics tablet attached to an Apple 11+ microcomputer. Nuclear diameters were calculated by tracing the images of the nuclei using the graphics tablet pen and software. Equivalent spherical nuclear diameters (D) were calculated using the formula D = 2(A/ϕ)⁰.⁵ where A is the area of the projected image of the nucleus, converted to μm². Four measurements were averaged for each nucleus. Frequency distributions are plotted for the percentage of the total number of nuclei studied from each procedure (Roy–McEwen, N = 553; one-tube, N = 423) falling within each size classification. Areas of overlap between the two distributions are indicated by the stippled bars. Mean nuclear diameters were as follows: Roy–McEwen, 8.7 ± 2.2 mm (SD); one-tube, 7.6 ± 2.1 mm (SD).

The only meaningful differences between the one-tube and Roy–McEwen methods concern the time taken to obtain the final nuclear pellets, and the efficiency of the nuclear isolation procedure. Because the one-tube method entails only a single centrifugation step, as opposed to the four spins used in the Roy–McEwen procedure, it is considerably shorter. Moreover, the yields of nuclei from the one-tube method tend to be slightly higher and more consistent, since repeated washing of the nuclei is avoided. Even with very small tissue samples, yields of nuclei from the one-tube method consistently approach 40% of the DNA present in the original tissue homogenate.

Despite these advantages, however, the one-tube method still suffers from serious drawbacks with respect to studies of regional ERn occupation in the brain. The recovery of nuclei, although reasonable for a procedure designed to achieve a high degree of nuclear purity, is still far from ideal. In practice, using [³H]estradiol as the labeled ligand, microdissected tissues must be pooled from at least 2–3 rats in order to obtain sufficiently high count rates in the macromolecular-bound fraction. Incorporation of Triton X-100 into the BII buffer with which the homogenate is mixed is necessary to achieve complete removal of cytoplasmic material from the nuclei; however, this step precludes using the supernatant from the initial centrifugation to obtain simultaneous measurements of cytosol estrogen receptor capacity. Moreover, the necessity to use high-speed centrifugation to sediment the nuclear pellet through the dense sucrose layer, as well as after KCl extraction, limits the number of samples that can be processed at one time.

Early studies by Anderson et al. (37) demonstrated that crude washed nuclear pellets could be used for ERn assays in the rat brain. In these studies, nuclei were incubated intact with the labeled estrogen and then separated from the unbound steroid by repeated washing (37). The sensitivity of this procedure is inadequate for studies of microdissected tissue samples due to difficulties in completely washing out the unbound steroid. Attempts to improve the sensitivity of the method by salt extracting the crude nuclear pellet and labeling the resultant extract in vitro were unsuccessful due to proteolytic degradation of the receptor complexes during incubation with the labeled steroid. Incorporation of protease inhibitors, such as leupeptin (41), into the incubates only partially inhibited this receptor degradation. During the course of exploring the possibility of using crude, buffer-washed nuclear preparations, however, we noted that the extent of the receptor losses appeared to be dependent on the quantity of tissue processed in each tube. As the amount of tissue contributing to each assay incubate was reduced, the recovery of ERn increased until, with very small tissue samples, receptor losses became negligible (Fig. 5A). These findings suggested that, with punch-microdissected regions of the rat brain, it might be possible to use a crude nuclear fraction and still obtain results comparable to those provided by the one-tube and Roy–McEwen procedures.

Fig. 5 ERn and DNA recoveries using the TEGD-wash assay procedure for assay of ERN concentrations in dilutions of homogenates from the pooled hypothalamus and preoptic area from estradiol-primed ovariectomized rats. (A) ERn concentrations (fmol/mg DNA) using either the TEGD-wash procedure (filled circles), the one-tube method (open triangle), or the TEGD-wash procedure incorporating 0.5 mM leupeptin and 20 mM sodium molybdate in the exchange assay buffers (filled triangle). Results represent means ± SEM of 4–7 observations in each case. (B) Recoveries of DNA through the TEGD-wash procedure at different homogenate dilutions. Essentially all the DNA was recovered across the entire range of sample DNA concentrations (means ± SEM of triplicate observations; where no error bars are visible, the errors are smaller than the symbol diameters).

This prediction proved to be correct. Using crude TEGD-washed nuclear pellets from microdissected brain regions of individual rats, ERn assay results are virtually identical to those obtained using a larger number of animals with the considerably more arduous one-tube procedure (Table I). Most important, losses of nuclei are negligible: using the TEGD wash procedure, essentially all of the DNA present in the original tissue sample is recovered in the final extracted nuclear pellet (Fig. 5B). Moreover, use of TEGD buffer for homogenization considerably enhances the potential utility of the method, since it is possible to use the supernatant from the initial centrifugation step for other assays. Recentrifugation of this supernatant at higher g provides a cytosolic fraction suitable for measurement of either nonnuclear bound estrogen receptors, or receptors for other hormonal