Journal of Histochemistry & Cytochemistry

In situ hybridization provides invaluable information regarding the localization of gene expression in heterogeneous tissues. The technique is extremely sensitive and can detect the amount of mRNA contained in a single cell. This review provides a starting point for those who wish to begin using in situ hybridization in their own laboratories. The procedure outlined here is based on 35S-labeled riboprobes and has been used with many probes and tissues with a greater than 90% success rate on the first hybridization. The importance of appropriate controls is stressed. Clusters of silver grains after hybridization do not necessarily indicate specific mRNA localization. Regions of the tissue rich in nuclei often appear to cause spurious binding of probes and have high backgrounds often mistaken as positive signals. The most difficult aspect of in situ hybridization is not to get clusters of silver grains on the slide but rather to do the appropriate controlled experiments to ensure that the signal is real and is not due to some artifactual binding of the probe to the tissue.

In this article we discuss strategies for selecting oligonucleotides to target isoform-specific mRNAs, drawing on our experience with isotopically labeled oligonucleotides for ISH of Na,K-ATPase mRNA alpha- and beta-subunit isoforms. Oligonucleotide probes based on one of these isoforms have a high probability of forming nonspecific hybrids with related isoform mRNAs. The design and selection of isoform-specific ISH and how their nucleotide structure influences hybridization are reviewed, as well as basic principles in identifying and evaluating candidate probes. Controls such as Tm analysis and GC content are evaluated. For distinguishing among multiple isoforms of gene families, choose lowest possible homology between isoforms consistent with other factors that influence probe performance.

We present a simple, reliable method for simultaneous detection of two distinct mRNAs within the same tissue sample by double in situ hybridization histochemistry. Sections are hybridized with a cocktail of radiolabeled and digoxigenin-labeled cRNA probes. The digoxigenin-labeled probe is detected with an alkaline phosphatase (AP)-dependent chromogen reaction and then the radiolabeled probe is detected by conventional autoradiography. The sensitivities of the two detection methods are comparable and demonstrate the feasibility of using double in situ hybridization histochemistry to investigate the co-localization and co-regulation of mRNAs that are expressed at relatively low levels.

Biotinylated probes for in situ hybridization (ISH) are now widely used to detect RNAs and viral genomes at the light and electron microscopic levels. Many protocols for ISH with biotinylated probes are now available. The most critical parameters that influence sensitivity and specificity of the procedure are the nature of the probe, the mode of fixation or incorporation of biotin, and the mode of detection of biotin. Biotin can be detected with antibodies or with avidin (or streptavidin) to which is coupled a marker that can be identified under the microscope. Here we review the protocols, pitfalls, advantages, and disadvantages of biotinylated probes for ISH.

We describe a procedure for detection of low-abundance cellular RNAs by in situ hybridization histochemistry, using single-stranded DNA probes produced by amplified primer extension labeling with Taq polymerase. We have used this approach to detect a number of high- and low-abundance RNA species and have found it to be a simple and reproducible method of obtaining sensitive probes for in situ hybridization studies. For example, DNA probes generated by amplified primer extension labeling can detect low-abundance heteronuclear RNAs in individual neurons. Since this procedure does not involve recombinant DNA technology or microbiological facilities, it should prove useful to a wide variety of investigators studying the regulation of gene expression at the cellular level.

The goal of quantitative autoradiography (QAR) in in situ hybridization (ISH) is to determine the amount of radioactive oligonucleotide or riboprobe present in the corresponding area of the tissue slice that produced an autoradiographic image. This article discusses (a) some of the considerations related to selection and use of computer image analysis systems for accomplishing this objective, (b) development of 14C plastic autoradiographic standards for ISH QAR with 33P, (c) using QAR to develop Tm curves for ISH probes, and (d) measurement of resolution with video imaging systems for QAR.

We describe a new stereological method for analyzing data derived from the in situ hybridization procedure. This method should prove important, since data summarization in terms of grains per anatomic area by sampling of tissue sections may lead to faulty interpretations. Using computer simulation of measurements taken from a two- and a three-dimensional perspective, we show how the detection of molecular changes can be influenced by multiple structural events. Cell volume, the volume of the structure, and the number of labeled cells in an anatomic structure are all important parameters that must be assessed to obtain accurate results. We then outline in detail a simple and efficient method for estimation of the number of mRNA molecules in cells and in the total structure. By estimation of the abundance of mRNA molecules in three dimensions, we can make better inferences concerning gene expression in a structural context.

Simultaneous study of intracellular quantification and distribution of fluorescent probes is difficult when cell staining is not homogeneous. This occurs after mitochondrial staining with rhodamine 123 (R123). Classical techniques for evaluation of intracellular R123 fluorescence, such as flow cytometry, are based on measurement of the global fluorescence intensity but do not take into account parameters that reflecting cellular distribution of the probe. For simultaneously studying intracellular quantification and distribution of R123 with fluorescence image analysis, we delineated a mask of the cell, generated from a fluorescent image of the plasma membrane stained by nile red (NR). After a preliminary study of the fluorescence characteristics of R123 and NR to avoid artifacts and optimize conditions of staining, quantification and distribution of intracellular R123 studies were performed by superimposition of the mask on the R123 fluorescence image. This protocol was applied to leukemic cells and allowed estimation of individual cell parameters such as mean fluorescence intensity and standard deviation, the latter providing information of the cellular distribution of R123. Moreover, it permitted demonstration of the redistribution of R123 in the whole cell when coincubated in the presence of nigericin.

In this study we determined acetylcholinesterase (AChE) activity by electron probe microanalysis (EPMA). AChE was localized in rat brain sections by a histochemical reaction that produced a Cu-ferricyanide precipitate. This reaction is based on the hydrolysis of acetylcholine substrate with AChE. The copper (iron) concentration in the deposit is proportional to acetylcholine involved in the hydrolysis reaction. Absolute determination of the size of the reaction was calculated from EPMA measurements compared with those made from a dextran standard containing a known amount of Cu(Fe). Electron probe microanalysis allows AChE activity to be assayed quantitatively on thick brain sections with a spatial resolution of 10 microns.

Granins are acidic proteins co-localized with peptides in secretory granules of many endocrine cells. They are thought to participate in certain steps of the regulated secretory pathway. This is of particular interest in rat pituitary gonadotropes, which in most cases contain both gonadotropins (follicle-stimulating hormone, FSH and luteinizing hormone, LH) and two granins (chromogranin A, CgA and secretogranin II, Sg II). Therefore, we investigated male rat gonadotropes ultrastructurally and for the cellular and subcellular localization of gonadotropins/granins under normal conditions and after stimulation by luteinizing hormone-releasing hormone (LHRH) or castration. Typical gonadotropes of controls contained differently composed secretory granules: small granules showed immunoreactivity for LH and SgII and larger granules were immunoreactive for FSH and CgA and to a lesser extent, for LH. Stimulated gonadotropes showed hypertrophy or hyperplasia and RIA for plasma LH levels showed a 30-70-fold increase. In stimulated cells a third type of secretory granule became prominent. These "intermediate" granules had an electron-dense core immunoreactive for LH and SgII, whereas CgA labeling was confined to a less electron-dense outer region. (In stimulated gonadotropes, FSH immunoreactivity could be shown effectively only at the light microscopic level.) Intermediate granules developed from structures resembling condensing vacuoles. They began to exhibit their typical double structure as they budded off from the trans-Golgi network. It therefore appears that CgA and SgII participate in establishing two different routes of the regulated pathway in gonadotropes. Therefore, immunocytochemistry of the granins seems to be a suitable approach to investigating secretory pathways in these endocrine cells.

Myocardial capillary endothelial cells, arteriolar endothelial cells, and the arterial adventitia show positive alkaline phosphatase (AP) enzyme reaction and immunoreactivity in both rat and human hearts. In guinea pigs, however, capillary endothelial staining is discontinuous and arterial adventitia is negative. The ultrastructural correlate of discontinuous capillary staining is a pronounced labeling of pericytes in guinea pig heart and relatively weak endothelial staining. In rat and human heart, enzyme reaction products are localized mainly on plasma membranes and cytotic vesicles of endothelial cells. Comparison of two strains of rat reveals a more dense deposition of enzyme reaction product along the luminal and particularly along the abluminal plasma membrane of Sprague-Dawley rats than of Wistar rats. Quantitative analysis of immunogold labeled anti-AP antibody density confirms the pronounced polarity of capillary endothelial cell labeling in Sprague-Dawley rats. More than 80% of total endothelial AP protein in Sprague-Dawley rats is localized over the abluminal plasma membrane and basal lamina, as compared with less than 30% in Wistar rats. Moreover, the total endothelial cell labeling is almost sixfold higher in Sprague-Dawley than in Wistar rats. Total endothelial labeling and proportion of labeling on the abluminal endothelial plasma membrane in human hearts is intermediate between the two strains of rat. The strain and species differences in enzyme distribution could provide important information concerning enzyme function.

The multifunctional adhesive glycoprotein vitronectin (VN) undergoes a unique conformational transition from the plasma form into a multimeric form that represents the reactive heparin-binding form. In this study we investigated the interaction of multimeric vitronectin (VNmult) or VN-gold conjugates (which are equivalent in biochemical properties) with confluent and subconfluent monolayers of porcine endothelial cells. Time-dependent direct binding of radiolabeled VNmult to the luminal face of endothelial cells at 37 degrees C was observed which was competed by heparin, whereas plasma VN showed hardly any binding. At 4 degrees C binding of VNmult remained cell-associated, whereas after 6 hr at 37 degrees C a major portion of the ligand was translocated through cells and was associated with the subcellular matrix. Cytochemical studies with VN-gold conjugates were performed to demonstrate uptake of VNmult. At 4 degrees C only surface decoration of cells with gold label was seen, which was totally reversible in the presence of heparin. Subsequent incubation for various time intervals at 37 degrees C revealed disappearance of gold label from the surface and accumulation of conjugates in a perinuclear distribution inside the cells as judged both by electron microscopy and after silver enhancement by light microscopy. Cross-sections of endothelial cells demonstrated the inclusion of VN-gold conjugates in coated pits, endosomes, and in lysosomal compartments close to the nucleus. Within 2-6 hr a portion of VN-gold conjugates had accumulated with proteoglycans at the matrix face. These data provide strong evidence for specific routing of a portion of VNmult from the circulation into extravascular spaces, where the protein is believed to fulfill major adhesive and regulatory functions particularly as co-factor in plasminogen activation and immune defense.

To ascertain the ability of commercial and home-made anti-fading media to reduce the decrease of fluorescein isothiocyanate (FITC) fluorescence, we studied the bleaching characteristics of FITC-stained Reh 6 cells mounted in buffered glycerol and in anti-fading media. We measured the intensity of fluorescence over time with a confocal laser scanning microscope and a standard epifluorescence microscope coupled to an image analysis system. Most of the anti-fading media effectively retard fading but each has drawbacks. Better results were obtained with media containing p-phenylenediamine (solutions in buffered glycerol, Vectashield, Fluorstop). However, Mowiol, Slowfade, n-propyl gallate (20 g/liter) were also effective in retarding fading. Most of them, except Mowiol, reduced fluorescence intensity. We concluded that the choice of anti-fading medium would depend on the desired results: a slower decay of fluorescence despite an initial quenching of fluorescence or a lower retardant effect with no decrease in initial fluorescence intensity. Moreover, the combination of Mowiol with another anti-fading medium may be a useful compromise when a strong retardant effect is required without marked quenching of the initial fluorescence.

To study the origin of glucose in the oviduct fluid, we cytochemically examined glucose-6-phosphatase (G6Pase) activity in rat oviduct. The activity in the whole oviduct was also assayed biochemically. During proestrous, estrous, and metestrous phases, staining reaction for the activity was moderate in the epithelium of the caudal isthmus (CaI) and uterotubal junction (UJ), whereas it was weak in that of the ampulla (A) and cephalic isthmus (CeI). In the diestrous phase, staining reaction in the epithelium of CaI and UJ became strong although it remained weak in that of A and CeI. Reaction product for the activity was localized in the endoplasmic reticulum and nuclear envelope of all cell types in the epithelium. The amount of reaction product in secretory cells was small to moderate in CaI and UJ, and small in A and CeI during proestrus, estrus, and metestrus. In diestrous the amount became abundant in CaI and UJ and moderate in A and CeI. However, the amount in ciliated cells remained small in the four segments during the four phases. The biochemical activity in diestrous was greater than that in proestrus, estrus, or metestrus. This shows that the activity is high in secretory cells in the epithelium of CaI and UJ in the diestrous phase and suggests that the role of the high activity is to release glucose into the oviduct fluid for use by the embryo passing down the CaI and UJ to the uterus.

The goal of this study was to localize phosphoenolpyruvate carboxykinase (PEPCK), glycogen synthase (GS), and glycogen phosphorylase (GP) in the liver lobule by immunocytochemical techniques and to describe the effects of feeding and fasting on the distribution and quantity of these enzymes. Livers from ad lib fed and overnight fasted normal adult male rats were frozen in liquid nitrogen after transcardial perfusion with 30% sucrose. Serial cryostat sections of tissue were collected on slides, fixed by immersion in 4% paraformaldehyde, and incubated with antibodies against PEPCK, GS, and GP. Antibodies to these enzymes were visualized with a gold-conjugated secondary antibody and a silver enhancement technique. Fed animals demonstrated a periportal to pericentral gradient of PEPCK. Fasting increased the periportal content of PEPCK, induced the midlobular and centrilobular cells to express the enzyme, and steepened the periportal to pericentral gradient. The increase of PEPCK was confirmed by Western blot analysis. GS and GP were distributed throughout the lobule in the fed animal but often showed a centrilobular pattern, and fasting did not alter the lobular distribution of either enzyme. Western blot analysis revealed no changes in the amount of these enzymes in the fed or fasted state. The cellular distribution of the three enzymes is similar to that of hepatic glycogen, in that the immunoreactive material has a clumped appearance in the periportal hepatocytes and is more dispersed in the pericentral cells. On fasting the periportal hepatocytes lose the dense compact localization of the enzymes and the protein becomes more homogeneously distributed throughout the cytosol. Further studies are needed to elucidate the functional significance of the regional heterogeneity of the glycogen-metabolizing enzymes and the molecular mechanisms regulating their gene expression.

Cystatins represent a widely distributed superfamily of cysteine proteinase inhibitory proteins. We investigated the expression of the cystatin C gene, belonging to the family 2 of cystatins, in the hearts of female rats. Using a highly sensitive reverse transcriptase-polymerase chain reaction (RT-PCR) we have detected cystatin C mRNA in the ventricule and atrium, as well as in liver and submandibular gland. A digoxigenin-labeled cystatin C probe, generated by PCR, hybridized to a single mRNA species of about 700 nucleotides on Northern blots. Northern blot hybridizations established that neither an acute inflammation produced by injection of turpentine nor administration of the beta-adrenergic agonist isoproterenol had an effect on the level of cystatin C mRNA in the heart. In situ hybridizations with digoxigenin-labeled probe localized the expression of the cystatin C gene to cardiac muscle fibers but not to other cardiac cellular elements. Cystatin C may be released by cardiac muscle fibers under physiological and pathological conditions and may modify inflammatory and necrobiotic processes.