- Email: [email protected]
A microscopic survey on the efficiency of well-known routine chemical fixatives on cryosections
ARTICLE IN PRESS Acta histochemica 108 (2006) 487—496
www.elsevier.de/acthis
A microscopic survey on the efficiency of wellknown routine chemical fixatives on cryosections Ozgur Cinara, Olcay Semizb, Alp Cana, a
Department of Histology and Embryology, Ankara University School of Medicine, Sihhiye, 06100 Ankara, Turkey Sakarya University School of Health, Sakarya, Turkey
b
Received 2 March 2006; received in revised form 27 May 2006; accepted 29 May 2006
KEYWORDS Fixation; Cryosections; Acetone; Formaldehyde; Paraformaldehyde
Summary This study was designed to analyze and compare tissue preservation efficiency of acetone (AC), formaldehyde (FA) and paraformaldehyde (PFA) on cryosections. Brain, kidney, heart and liver tissue of adult Balb/c mice were fixed with either FA or PFA prior to cryosectioning, or fixed with AC alone immediately after cryosectioning. Hematoxylin and eosin staining showed that AC is a poor fixative in preserving the general tissue and cellular organization. PFA, and to a lesser extent FA, produced significantly better results. Another set of cryosections were further analyzed to test the properties of those fixatives to preserve proteins from specific cell structures. Cytokeratin filaments, F-actin filaments and nuclei were immunolabeled and examined using confocal microscopy. Results demonstrated that, overall, PFA is the best fixative tested. However, FA fixation gave poor results in preserving neuronal tissues. Immunofluorescence confirmed the inefficiency of AC fixation, after which no specific labelling of cytokeratin filaments was detectable. Nevertheless, actin filaments were detectable on AC-fixed samples, a finding that was supported by the quantification of fluorescein-phalloidin binding to F-actin. Overall, the data suggest that AC fixation is unacceptable for preservation of most samples, whereas FA and PFA fixation should be chosen according to the tissues and proteins to be studied. & 2006 Elsevier GmbH. All rights reserved.
Introduction Corresponding author. Tel.: +90 312 3103010x369;
fax: +90 312 3106370. E-mail addresses: [email protected] (O. Cinar), [email protected] (O. Semiz), [email protected]. edu.tr (A. Can).
The structure of an animal tissue is determined largely by the configuration of its proteins. The main contributors to structure are the lipoproteins, which are major components of the plasma
0065-1281/$ - see front matter & 2006 Elsevier GmbH. All rights reserved. doi:10.1016/j.acthis.2006.05.008
ARTICLE IN PRESS 488 membranes and membranous organelles of cells; the fibrous glycoproteins of extracellular elements, such as collagen and basal laminae; and the globular proteins, which are dissolved in the cytoplasm and extracellular fluid. For histology and immunohistochemistry, all these substances must be stabilized by fixation. Fortunately, most fixatives preserve the form and location of insoluble proteins, polynucleotides and mucosubstances (Kiernan, 1990), although some substances may be more completely preserved than others by specific fixatives. In the last three decades, cryosectioning has gained prominence in comparison to classical fixation, processing and paraffin wax embedding of tissue blocks, since many immunolabelling procedures, as well as routine hematoxylin and eosin (H&E) stains, can be performed on cryosections. ‘‘Cryo’’ procedures are not only timesaving, but also cause less damage to antigenic sites; therefore antigen-antibody reaction is maximized during immunolabelling of proteins. This phenomenon is evidenced by the minimal appearance of background labelling in comparison to that seen on paraffin wax embedded material, and the lesser degree of cross reactivity of immunoglobulin molecules (Onodera et al., 1992). However, cryosectioning alone does not usually preserve tissue structures sufficiently for use in diagnostic and research activities, and therefore some alternative approaches have been sought to promote preservation and subsequent labelling quality (Onodera et al., 1992; Hall et al., 1987; Richter et al., 1999). Using a chemical fixative prior to, or after, cryosectioning is the first choice to increase the tissue quality for immunolabelling. However, relatively few reports have appeared in the literature describing a systematic analysis of the advantages and disadvantages of different chemicals on tissue preservation of cryosections (Hall et al., 1987; Onodera et al., 1992; Richter et al., 1999). In this study, we first tested the simple organic fixative, acetone (AC), and then two other commonly used fixatives, formaldehyde (FA) and paraformaldehyde (PFA), in order to evaluate whether rapid AC fixation is sufficient to preserve the main cellular and extracellular matrix protein molecules, or whether relatively slow fixation agents – such as PFA and FA – should be used prior to cryosectioning. Four different organs with different tissue architecture were chosen to compare the fixation and staining efficiency of cellular proteins. Heart, kidney, brain and liver were fixed with PFA or FA for 24 or 48 h prior to cryosectioning or directly cryosectioned and postfixed with AC. All sections were visualized with multiple staining/
O. Cinar et al. immunolabelling procedures, including routine H&E stains, immunohistochemistry using anti-pancytokeratin antibody for cytokeratin filaments, fluorescently-labeled phalloidin for actin microfilaments and red emission dye 7-aminoactinomycin-D (7AAD) for nuclear labelling. Samples were examined both in a qualitative and a semi-quantitative manner in which line densitometric profiles were used to analyze the F-actin preservation efficiency of these three fixatives.
Materials and methods Tissue selection and fixation protocols To assess the overall efficiency of selected fixatives, four organs with different embryonic origins and with different proportions of epithelial, connective tissue, neuronal cells and extracellular matrices were evaluated in this study. The experimental protocol on animals was approved by the Institutional Review Board. Heart, kidney, brain and liver were sampled as follows: twelve adult Balb/c mice were divided into three groups. Mice in the first group were killed by cervical dislocation and the organs were immediately removed. 5 5 mm tissue blocks from the organs were immersed into Cryomatrixs (Shandon Scientific Ltd., Pittsburgh, PA, USA) and frozen for 10 min at 60 1C. The specimen temperature was raised to 15 1C, and 8 mm-thick cryosections were cut, air-dried for 10 min, and then fixed for 10 min with AC at 20 1C, pH 6.9 (ultra grade purity 99.9%, Applichem Co., Darmstadt, Germany) then stored at 4 1C until required. Animals in the second and third groups were anesthetized by intraperitoneal injection of 50 mg sodium pentobarbital (Sigma, St. Louis, MO, USA) per 100 g of body weight. After animals became insensitive to painful stimuli (within 5 min), animals in the second group were perfused with 10% FA solution containing 0.08 M dibasic and 0.04 M monobasic sodium phosphate (Merck Co., Darmstadt, Germany), final pH 7.4, for 10 min at room temperature, followed by immersion in the same fixative for 48 h at the same temperature. Animals in the third group were perfused for 10 min with freshly prepared 3.5% PFA, pH 7.4, (Merck Co., Darmstadt, Germany), and then immersed in the PFA solution for 24 h at 4 1C. Prior to cryosectioning, by the same procedure as stated for AC samples, both FA- and PFA-fixed blocks were immersed in 20% (0.58 M) and 30% (0.88 M) (w/v in distilled water) sucrose solutions (Merck Co., Darmstadt, Germany), for 12 h each, at 4 1C for cryoprotection (Romijn et al., 1999).
ARTICLE IN PRESS Fixative efficiency on cryosections
489 mental design, assessed the quality of tissue protection. Fluorescence intensity and staining/ immunolabelling patterns were evaluated using a line densitometric profile analysis, a component of the LSM 510 Meta software. The resolution of the line was set to 1000 points along 20 mm in all samples, corresponding to approximately nine sarcomeres. The grey value (8 bit ¼ 256) of each point along the designated line was determined with regard to distance (Ogiwara et al., 1999). This approach was used (i) to assess the fine staining property of actin filaments and (ii) to quantify the phalloidin binding efficiency to F-actin after fixation, both of which, as a whole, reflect the degree of tissue preservation. Representative micrographs (n ¼ 12) from each FITC-phalloidin-bound heart muscle section were chosen in order to compare the three fixatives. Staining intensities from randomly selected cardiac myocytes were compared by using the computer-based SigmaStat software package (version 3.0; Jandel Scientific Corporation, San Rafael, CA, USA). The differences among each group were separately tested utilizing one-way ANOVA and post hoc Dunn tests, with a level of significance set at po0:05.
Staining and immunolabelling procedures In this study, no detectable differences were noted between previously fixed sections stored at room temperature, 4 1C and –20 1C (unpublished observations). Sections stored at 4oC were initially stained with Harris’s hematoxylin (Sigma, St. Louis, MO, USA) and eosin using routine protocol for general orientation of the sections. Bright-field colored images were recorded from 36 different samples taken from the four organs. Sections adjacent to H&E stained samples were incubated with: (a) 50 mg/ml fluorescein isothiocyanate (FITC)-labelled phalloidin (a bicyclic heptapeptide toxin of Amanita phalloides) in phosphate buffered saline (PBS) containing 1% methanol, for 1 h, at room temperature (Sigma) (Can and Semiz, 2000); (b) affinity-purified mouse monoclonal anti-pancytokeratin antibody (Sigma) at a dilution of 1:100 in PBS, for 90 min at 37 1C, followed by Cy3-labelled goat anti-mouse IgG (Jackson ImmunoResearch, West Grove, PA, USA) at a dilution of 1:100 in PBS, for 90 min at 37 1C; or (c) 10 mM 7-AAD (Sigma, St. Louis, MO, (Sigma)) for 1 h at room temperature.
Confocal microscopy
Results
Slides were analyzed using a Zeiss LSM 510 Meta confocal microscope (Zeiss, Jena, Germany) equipped with two lasers (488 nm Argon ion and 543 nm He–Ne). A color-coded palette was used to optimize the grey value for proper acquisition of fluorescent images from each label. Detection parameters such as laser intensity, pinhole diameter, detector gain, amplifier offset and amplifier gain were set to identical values. 2048 2048 pixel single optical sections were recorded using Zeiss LSM Meta 3.2 version software.
Qualitative staining and immunolabelling results are summarized in Table 1.
Hematoxylin and eosin staining Results are illustrated in Fig. 1. Acetone: In brain, a loss of tissue organization and cytoarchitecture, artificial widening and swelling throughout the brain cortex and in the medulla (not shown) predominated. Similar findings were also seen in kidney and heart, and to a lesser extent in liver preparations. AC fixation resulted in an extreme loss of interstitial tissues around tubules in kidney, and in interfibrillar connective tissue in the
Evaluation of staining pattern and intensity Two different histologists, who were not informed about the study purpose or the experiTable 1.
Qualitative staining results of sections from four different organs (n ¼ 42)
Brain
AC FA PFA
Kidney
Liver
Heart
GS
CYK
F-A
Nuc
GS
CYK
F-A
Nuc
GS
CYK
F-A
Nuc
GS
CYK
F-A
Nuc
+ ++ +++
+ + +++
+ +++
+++ +++ +
+ ++ +++
+ ++ +++
+ ++ +++
+++ ++ +
++ + +++
+ ++ +++
+ +++ +
+ +++
+ + +++
++ ++ +++
+++ +++ +++
+++ +++ +
GS: General structure; CYK: Pancytokeratin immunolabelling; F-A: F-Actin staining: Nuc: Nuclear staining : no staining; +: poor; ++: moderate; +++: well.
ARTICLE IN PRESS 490
O. Cinar et al.
Figure 1. Representative images obtained from H&E-stained cryosections of AC-, FA- and PFA-fixed brain, kidney, liver and heart. AC-fixed samples display relatively poor fixation results as seen by the appearance of interstitial swelling and widening. FA-fixed samples provide better results, especially in brain sections. Loss of tissue organization and disrupted capillary walls are seen as prominent features in kidney, liver and heart. PFA-fixed samples gave optimum results regardless of the tissue studied. Scale bar ¼ 50 mm.
heart. Hepatic plates and capillary sinusoids were relatively poorly preserved by AC fixation. Formaldehyde: Brain sections were relatively well preserved compared to kidney, liver and heart sections. Giant pyramidal cells of brain cortex and intercellular matrices were clearly identified. In contrast, interstitial and interfibrillar tissues were severely disrupted in kidney and heart sections, respectively. More striking features were noted in liver samples. Paraformaldehyde: This fixative provided wellpreserved cell and tissue organization, particularly in brain, liver and heart sections. In kidney samples, glomeruli and tubules were moderately maintained with a minor disorganization of interlobular arteries and veins.
Fluorescence labelling Three different fluorescent markers were used to analyze the effects of fixatives on specific cellular
proteins. Anti-pancytokeratin antibody that was conjugated with Cy3 (red signal in Fig. 2) detects cytokeratin types 4, 5, 6, 8, 10, 13, and 18, and thus differentiates the epithelial cells from the other cell types (Bartek et al., 1991). The green marker in Fig. 3 is FITC-phalloidin, and this readily binds to F-actin in almost all cell types (Prochniewicz-Nakayama et al., 1983). Red signals in Fig. 3 correspond to nuclear dye 7-AAD. The following results were obtained by detecting those three cellular markers that were constitutively found in all sections. The effectiveness of the three fixatives was qualitatively and quantitatively evaluated using exactly the same imaging parameters in all observations and recordings. Acetone: In brain sections, a few foci were immunolabelled with the anti-pancytokeratin antibody; however they did not resemble any of the expected cellular phenotypes (Fig. 2). No F-actin labelling was detectable (Fig. 3). In contrast, nuclear 7-AAD staining was well preserved (Fig. 3). In kidney and liver sections, the pancytokeratin
ARTICLE IN PRESS Fixative efficiency on cryosections
491
Figure 2. Pancytokeratin antibody used to detect the antigen preservation efficacy of AC, FA and PFA fixation on cryosections. Pancytokeratin immunolabelling of epithelial cells was used to visualize the different amounts of epithelial components in the different tissues. AC-fixed samples yield blurry and non-specific results in all sections examined, with the exception of heart samples. FA provided more specific immunolabelling, as the capillary endothelial cells are identified particularly in kidney, liver and heart sections. More striking results are obtained on PFA-fixed samples that enable detection of both endothelial and other epithelium-originated cell types. Intertubular and intercellular capillaries are obvious against a black background, indicating an absence of non-specific labelling. Scale bar ¼ 40 mm.
immunolabelling pattern appeared fuzzy and nonspecific (Fig. 2). Inconsistent and uneven F-actin labelling was noted in kidney and liver sections (Fig. 3). However, 7-AAD labelling was confined to well-preserved nuclei (Fig. 3). Interestingly, sections from heart showed clearer labelling with all markers tested (Figs. 2 and 3). Formaldehyde: This fixative did not preserve any cellular cytokeratin filaments in brain. A blurry appearance of pancytokeratin-immunopositive areas in brain was not specifically localized to any recognisable cell type (Fig. 2). Similarly, phalloidin labelling did not seem to be specific to F-actin in brain sections (Fig. 3). Nuclei however were clearly identified by 7-AAD labelling, in a similar manner to AC-fixed groups (Fig. 3). FA successfully maintained the antigenic sites for binding of the anti-pancytokeratin antibody in kidney, liver and heart tissues (Fig. 2). A specific filamentous immunolabelling pattern was confined to epithelial cells in kidney
and liver sections (Fig. 2). Few pancytokeratinimmunopositive areas were detected in heart sections, due to the relatively low number of capillaries among cardiac myocytes (Fig. 2). FA also preserved actin in kidney, liver and heart. Slender actin filaments were clearly identified in glomeruli, tubules and in the vasculature of kidney (Fig. 3). Similarly fine labelling of intracellular actin filaments was seen in liver sections. Central veins of hepatic lobules, sinusoidal capillaries and bile canaliculi were clearly identified at high magnification in liver preparations. Paraformaldehyde: In general, all sections from all tissue samples displayed well-preserved intraand intercellular features, illustrated in Figs. 2 and 3. Nevertheless, pancytokeratin immunolabelling was restricted to capillary endothelial cells in all tissues tested (Fig. 2), and a consistently weak labelling of hepatocytes was observed (Fig. 3). FITC-phalloidin bound to all F-actin containing
ARTICLE IN PRESS 492
O. Cinar et al.
Figure 3. Dual labelling of AC-, FA- and PFA-fixed cryosections with FITC-phalloidin (green signal) and 7-AAD (red signal) for screening the actin microfilaments and cell nuclei, respectively. AC-fixed brain sections feature an absence of microfilament labelling, whereas a substantial increase in labelling intensity is noted after FA fixation, albeit in a non-filamentous pattern. Strikingly, bundles of F-actin positive-axons are identified in PFA-fixed brain sections. Kidney samples fixed with AC give poor and non-specific labelling results since the labelled areas do not seem to correspond to actin filaments, which are particularly found in the cell periphery. In contrast, FA and PFA preserve those filaments, facilitating differentiation of the general tissue components (see FA- and PFA- fixed kidney and liver samples). PFA shows the best F-actin preservation among the other two fixatives, as the delicate structure of cardiac myocytes with large amounts of actin microfilaments are clearly illustrated after fixation with PFA. Nuclear staining is generally acceptable in all specimens, except that little or cell-type specific staining is noted in FA-fixed liver and PFA-fixed kidney sections. Scale bar in brain ¼ 40 mm, other ¼ 20 mm.
intracellular structures in all samples, including axonal fibres, epithelial cells of the pia mater, glomerular podocytes, and sinusoidal endothelial cells (Fig. 3). Interestingly, varying degrees of nuclear labelling with 7-AAD were consistently observed in PFA-fixed samples. Only in liver sections was the nuclear staining of a quality sufficient to differentiate the cells clearly (Fig. 3). In the remainder of the samples, extremely weak nuclear staining was seen.
Evaluation of labelling pattern and intensity Twelve heart sections from each group of fixative group were selected randomly to compare the F-actin decoration (a non-immunological direct
binding of phalloidin to F-actin) using the line profile digitization technique, illustrated in Fig. 4. Signal intensities along a selected 20 mm line were statistically evaluated. In AC-fixed groups, dye staining specificity appeared low, as seen by the inconsistent and uneven peaks of signal intensities; no sarcomere pattern could be recognized in AC-fixed groups. In FA-fixed groups, dye distribution gave more specific results, albeit in a lower intensity (39% lower values compared to AC-fixed samples). Relatively better results were obtained in PFA-fixed groups, as successively sharp peaks of signals were noted that clearly identify nine successive sarcomeres. The cumulative results show that AC and PFA have higher labelling intensities than FA. There was no statistically significant difference between AC and PFA regarding the signal
ARTICLE IN PRESS Fixative efficiency on cryosections
493
Figure 4. Cryosections of heart labelled with FITC-phalloidin and subjected to densitometric analysis to reveal the staining quality and intensity of successive sarcomeres along a 20 mm line (white line) on randomly selected cardiac myocytes. Corresponding profiles below each micrograph demonstrates the degree of regularity of fluorescent peaks. AC fixation results in irregular peaks with noisy background, whereas more consistent peaks corresponding to successive nine sarcomeres are noted in FA- and PFA-fixed samples. Although labelling intensity in FA is significantly lower than PFA and AC (39%73.4), AC fixation is still not favoured in comparison to the other two fixatives. Y-axis refers to grey-scale value (0–255); x-axis refers to distance in mm.
intensities. However, the differences between PFA and FA and between FA and AC were statistically significant (po0:05).
Discussion During the last 30 years, a limited number of studies have focused on the comparison of fixatives and fixation methods aimed at preserving vital tissue organization as much as possible (Hall et al., 1987; Onodera et al., 1992; Richter et al., 1999; Gokhale et al., 2002). Frozen sectioning has achieved popularity for two main reasons. One is that false positive background staining can be an undesirable problem in paraffin wax embedded sections, particularly during observations using fluorescence labelling, and this makes it impossible in many cases to differentiate minute structures in sections. The second reason is that frozen sections provide an invaluable tool in diagnostic pathology; as the tissue processing procedure is extremely fast, it can be undertaken while a surgical operation is ongoing. In recent years, frozen sections have been considered as the preferred method for preparation of tissues for immunoenzymatic and immunofluorescent labelling techniques, particularly for preserving delicate protein molecules. Freezing is
generally thought of as a convenient method to protect antigenic determinants, although it is not often adequate for preserving general tissue organization (Prento and Lyon, 1997; Gokhale et al., 2002; Hoetelmans et al., 2002). However, protection of antigenic epitopes means very little if results are wrongly interpreted because tissue structure is distorted. In this sense, the effectiveness of tissue preservation of several fixation approaches still remains questionable. PFA and FA are aldehyde fixatives and form both intra- and intermolecular cross-links with protein molecules, which result in the formation of more rigid heteropolymers. The monoaldehyde FA is the simplest member of this family of fixatives, and has been commonly used for both light and electron microscopy. The reactions of FA with proteins are numerous and well understood. The first step involves nucleophilic attack by amino groups with the formation of amino methylol products, which then condense with other groups such as phenol, imidazole and indole moieties to form methylene bridges. The methylene bridges are considered responsible for the fixation of protein by formalin under conditions of fixation for electron microscopy (Griffiths, 1993). PFA is preferred over FA because it contains fewer impurities. Cross-linking does not always inhibit enzymatic activity, and even the preservation of a small percentage of enzymatic activity is usually
ARTICLE IN PRESS 494 sufficient for precise localization of the enzyme (Hayat, 1970). Like methanol and ethanol, AC is a simple organic coagulant that displaces water from proteinaceous materials, thereby breaking hydrogen bonds and disturbing tertiary structure to produce denaturation. Soluble proteins in the cytoplasm are coagulated and organelles are destroyed. Nucleic acids are not precipitated. Precipitated protein that has not been denatured retains enzymatic or other biological properties, and remains soluble in aqueous solution. AC extracts lipid from tissues but carbohydrate-containing components are largely unaffected. AC is used alone for fixing films and smears of cells, and previously for otherwise unfixed cryostat sections. It is not suitable for blocks of tissue. Unless they are very small, because it causes considerable shrinkage and hardening of tissue (Kiernan, 1990). This was seen in our AC-fixed samples, particularly of brain, kidney and heart sections. In contrast, swelling and enlargement of hepatocytes was observed. AC fixation has been reported to have severe detrimental effects in cell surface protein molecules and major cytoskeletal elements (Haas et al., 1988; Vielkind and Swierenga, 1989). Vielkind and Swierenga (1989) reported that microtubules were poorly preserved in AC-fixed cells, while, immunohistochemistry using a monoclonal anti-cytokeratin antibody produced an incomplete immunolabelling pattern against a diffuse background after FA fixation. Haas et al. (1988) demonstrated that PFA fixation maintained the morphologic integrity of the sperm acrosome and plasma membranes, while AC fixation was unacceptable. In this study, we tested the labelling pattern of two stable cytoskeletal proteins and the nucleus. In all of the organs studied, AC fixation resulted in poor results in preserving the cytoskeletal elements. In contrast, nuclear binding of 7-AAD to DNA revealed excellent results. The latter was previously tested on different tissues and cells (Can et al., 2003) and the results were consistent with the findings of the current study that AC fixation is fastest and best for subsequent 7-AAD labelling of cell nuclei. Similarly, Gillespie et al. (2002) found that AC was superior to aldehyde fixatives when the aim was to visualize the details of the nucleus. Taken together, AC fixation, particularly for frozen sections, is not a suitable method either for general microscopic evaluations or for immunohistochemistry (Gillespie et al., 2002). Hall et al. (1987) made efforts to improve the cellular morphology in cryostat sections using AC and/or periodate-lysine-PFA fixatives (PLP) on tonsil sections. They separated specimens into four groups, which were AC alone (25 min), PLP
O. Cinar et al. (10 min), PLP (8 min) followed by AC (2 min), and AC (2 min) followed by PLP (8 min). They found that AC alone gives excellent morphology, in contrast to our AC results. Nevertheless, they indicated that AC alone also gave poor immunolabelling and that this could be improved by using AC (2 min) followed by PLP (8 min). Prento and Lyon (1997) compared the effects of buffered formalin, Clarke’s ethanol-acetic acid, and ethanol alone on rat intestine, kidney and liver tissues, and for bright-field and fluorescence microscopy. Buffered formalin and Clarke’s ethanol-acetic acids gave better results compared to fixatives that contain some commercial substitutes or ethanol. Histological distortion, cell shrinkage and vacuolization were prominent when substitutes or ethanol were used. In contrast, they found that artefacts were occasionally found when buffered formalin or Clarke’s fixative were used. Immunohistochemical studies demonstrated a total loss of low molecular weight antigens after treatment with all fixatives, except buffered formalin. They concluded that the best immunolabelling was obtained by combining formalin fixation with subsequent antigen retrieval. Ethanol was tested as a fixative in the present study. Ethanol fixation resulted in poor morphology, histological distortion, cellular shrinkage and vacuolisation, consistent with the results reported by Prento and Lyon (1997). Rustad et al. (1989) compared the cell-protective and staining effects of FA, ethanol, and a fixative cocktail (a combination of ethanol, formalin and glacial acetic acid) on dermal tissues after cryosectioning. They concluded that FA fixation was characterized by a significant increase in the staining intensity of basal cells of epidermis. Delicate and precise cellular staining was remarkably good after the cocktail fixative in comparison to FA and ethanol. All studies cited above aimed to analyze the relationship between fixation and frozen sectioning. However, the effects of fixatives regarding general morphology and antigen protective efficiency on tissues derived from diverse embryonic layers, and having different cellular and extracellular matrix organization, have not been previously evaluated in any systematic way. The aim of the present study was to focus on that issue, and to address the overall efficiency of selected fixatives on four different organs of different embryonic origin and including different amounts of epithelial, connective tissue, neuronal cells and extracellular matrices. PFA fixation was superior to either FA or AC for subsequent H&E staining at the light microscope level. Further observations using higher resolution imaging systems may reveal
ARTICLE IN PRESS Fixative efficiency on cryosections the mechanisms by which tissue protective efficiency is achieved by these fixatives. In fact, confocal microscopy results have enabled a substantially higher degree of detection and quantification. Using several optical sections, we were able to demonstrate that PFA preserves actin filaments (the most abundant filamentous protein in muscle cells) compared to FA or AC. Similarly, pancytokeratin-immunolabelled filaments in epithelial cells were better retained by PFA compared to FA or AC. However, AC displays better fixation quality for nuclear labelling by 7-AAD. We recently proposed a simple and reliable fluorescent staining method that is specific to DNA for the visualization of cell nuclei using 488 and 543-nm visible lasers or conventional mercury lamps (Can et al., 2003). The staining properties of two different reagents, chromomycin-A3 (C-A3) and 7-AAD were analyzed in several types of mammalian cells and tissues with different fixation and preparation methods. We also evaluated the antigen preservation efficiency of PFA, FA and AC on cultured Chinese hamster ovary (CHO) cells, mouse oocytes, rat aorta and rat dermal connective tissue, both in paraffin wax embedded sections and frozen sections. C-A3 and 7-AAD were highly specific fluorescent markers in both PFA- and FA-fixed cultured cells, and in paraffin wax embedded and cryostat sections. The chromatin and chromosome fine structure was visualised depending on the stage of the cell cycle. As supporting data for our previous results, here we re-illustrate that AC is better than aldehyde fixatives in labelling the nuclear fine structure and morphology. Quantification of fixation and dye-binding efficiency presented in this work using a series of densitometric line profiles revealed the fine pattern of actin labelling. A digital signal processor embedded in the confocal system provided excellent image comparison to relatively thin optical slices obtained using the same control parameters of the imaging elements. Gray-scale values corresponding directly to staining intensities of randomly selected cardiac myocytes were analyzed. Our results showed that the mean intensities from FA-group were statistically lower than others (po0:05). No statistically significant distinction was noted between PFA and AC-fixed groups. However, based on the labelling quality and detailed analysis of the densitometric profiles obtained, we assume that PFA or FA yields finer results than AC. In conclusion, interactions between the fixative and cellular protein structures may result in enormous variations if controlled experiments have not been performed using alternative methods.
495 Regarding the type of fixative, PFA is likely to be the first choice among tested fixatives for solid tissue components. On the other hand, amplified staining or over-staining does not simply imply that good antigenic protection was obtained, due to the problem of false or/and increased positivity. Additionally, the loss of or variation in tissue composition raises the question about whether changes arise from either the fixation method or the experiment performed. Continued improvement of tissue processing methods is still considered as a critical step toward ultimately determining the complete structure of normal and altered cells.
Acknowledgements The authors would like to Dr. Mitch Halloran, Ph.D. for his valuable criticism of the manuscript; to Remzi Ata who handled the animal laboratory ¨ zkavukcu and Sercin Karahuseyiand Drs. Sinan O noglu for the blind observations and interpretations of the H&E-stained slides. This project was financially supported by Ankara University Biotechnology Institute.
References Bartek J, Vojtesek B, Staskova Z, Bartkova J, Kerekes Z, Rejthar A, et al. A series of 14 new monoclonal antibodies to keratins: characterization and value in diagnostic histopathology. J Pathol 1991;164:215–24. Can A, Semiz O. Diethylstilbestrol (DES)-induced cell cycle delay and meiotic spindle disruption in mouse oocytes during in-vitro maturation. Mol Hum Reprod 2000;6:154–62. Can A, Semiz O, Cinar O. Two convenient methods for nuclear labeling in confocal microscope with visible lasers. J Histotechnol 2003;26:147–56. Gillespie JW, Best CJ, Bichsel VE, Cole KA, Greenhut SF, Hewitt SM, et al. Evaluation of non-formalin tissue fixation for molecular profiling studies. Am J Pathol 2002;160:444–57. Gokhale S, Ololade O, Adegboyega PA. Effect of tissue fixatives on the immunohistochemical expression of ABH blood group isoantigens. Appl Immunohistochem Mol Morphol 2002;10:282–8. Griffiths G. Fixation for fine structure preservation and immunocytochemistry. Fine structure immunohistochemistry. Berlin: Springer; 1993. p. 26–80. Haas Jr. GG, DeBault LE, D’Cruz O, Shuey R. The effect of fixatives and/or air-drying on the plasma and acrosomal membranes of human sperm. Fertil Steril 1988;50:487–92.
ARTICLE IN PRESS 496 Hall PA, Stearn PM, Butler MG, D’Ardenne AJ. Acetone/ periodate-lysine-paraformaldehyde (PLP) fixation and improved morphology of cryostat sections for immunohistochemistry. Histopathology 1987;11:93–101. Hayat MA. Biological applications. Principles and techniques of electron microscopy, Vol. 1. New York: Van Nostrand Reinhold Company; 1970. p. 65–95. Hoetelmans RW, van Slooten HJ, Keijzer R, van de Velde CJ, van Dierendonck JH. Comparison of the effects of microwave heating and high pressure cooking for antigen retrieval of human and rat Bc1-2 protein in formaldehyde-fixed, paraffin-embedded sections. Biotech Histochem 2002;77:137–44. Kiernan JA. Fixation. Histological histochemical methods, 2nd ed. Oxford: Pergamon Press; 1990. p. 10–32. Ogiwara N, Usuda N, Yamada M, Johkura K, Kametani K, Nakazawa A. Quantification of protein A-gold staining for peroxisomal enzymes by confocal laser scanning microscopy. J Histochem Cytochem 1999;47: 1343–9. Onodera Y, Shimizu H, Yamashita S, Nishikawa T. Cryofixed, freeze-dried and paraffin-embedded skin enables successful immunohistochemical staining of
O. Cinar et al. skin basement membrane antigens. Histochemistry 1992;98:87–91. Prento P, Lyon H. Commercial formalin substitutes for histopathology. Biotech Histochem 1997;72:273–82. Prochniewicz-Nakayama E, Yanagida T, Oosawa F. Studies on conformation of F-actin in muscle fibers in the relaxed state, rigor, and during contraction using fluorescent phalloidin. J Cell Biol 1983;97:1663–70. Richter T, Nahrig J, Komminoth P, Kowolik J, Werner M. Protocol for ultrarapid immunostaining of frozen sections. J Clin Pathol 1999;52:461–4. Romijn HJ, van Uum JFM, Breedijk I, Emmering J, Radu I, Pool CW. Double immunolabeling of neuropeptides in the human hypothalamus as analyzed by confocal laser scanning fluorescence microscopy. J Histochem Cytochem 1999:229–35. Rustad OJ, Kaye V, Cerio R, Zachary CB. Postfixation of cryostat sections improves tumor definition in Mohs surgery. J Dermatol Surg Oncol 1989;15:1262–7. Vielkind U, Swierenga SH. A simple fixation procedure for immunofluorescent detection of different cytoskeletal components within the same cell. Histochemistry 1989;91:81–9.
www.elsevier.de/acthis
A microscopic survey on the efficiency of wellknown routine chemical fixatives on cryosections Ozgur Cinara, Olcay Semizb, Alp Cana, a
Department of Histology and Embryology, Ankara University School of Medicine, Sihhiye, 06100 Ankara, Turkey Sakarya University School of Health, Sakarya, Turkey
b
Received 2 March 2006; received in revised form 27 May 2006; accepted 29 May 2006
KEYWORDS Fixation; Cryosections; Acetone; Formaldehyde; Paraformaldehyde
Summary This study was designed to analyze and compare tissue preservation efficiency of acetone (AC), formaldehyde (FA) and paraformaldehyde (PFA) on cryosections. Brain, kidney, heart and liver tissue of adult Balb/c mice were fixed with either FA or PFA prior to cryosectioning, or fixed with AC alone immediately after cryosectioning. Hematoxylin and eosin staining showed that AC is a poor fixative in preserving the general tissue and cellular organization. PFA, and to a lesser extent FA, produced significantly better results. Another set of cryosections were further analyzed to test the properties of those fixatives to preserve proteins from specific cell structures. Cytokeratin filaments, F-actin filaments and nuclei were immunolabeled and examined using confocal microscopy. Results demonstrated that, overall, PFA is the best fixative tested. However, FA fixation gave poor results in preserving neuronal tissues. Immunofluorescence confirmed the inefficiency of AC fixation, after which no specific labelling of cytokeratin filaments was detectable. Nevertheless, actin filaments were detectable on AC-fixed samples, a finding that was supported by the quantification of fluorescein-phalloidin binding to F-actin. Overall, the data suggest that AC fixation is unacceptable for preservation of most samples, whereas FA and PFA fixation should be chosen according to the tissues and proteins to be studied. & 2006 Elsevier GmbH. All rights reserved.
Introduction Corresponding author. Tel.: +90 312 3103010x369;
fax: +90 312 3106370. E-mail addresses: [email protected] (O. Cinar), [email protected] (O. Semiz), [email protected]. edu.tr (A. Can).
The structure of an animal tissue is determined largely by the configuration of its proteins. The main contributors to structure are the lipoproteins, which are major components of the plasma
0065-1281/$ - see front matter & 2006 Elsevier GmbH. All rights reserved. doi:10.1016/j.acthis.2006.05.008
ARTICLE IN PRESS 488 membranes and membranous organelles of cells; the fibrous glycoproteins of extracellular elements, such as collagen and basal laminae; and the globular proteins, which are dissolved in the cytoplasm and extracellular fluid. For histology and immunohistochemistry, all these substances must be stabilized by fixation. Fortunately, most fixatives preserve the form and location of insoluble proteins, polynucleotides and mucosubstances (Kiernan, 1990), although some substances may be more completely preserved than others by specific fixatives. In the last three decades, cryosectioning has gained prominence in comparison to classical fixation, processing and paraffin wax embedding of tissue blocks, since many immunolabelling procedures, as well as routine hematoxylin and eosin (H&E) stains, can be performed on cryosections. ‘‘Cryo’’ procedures are not only timesaving, but also cause less damage to antigenic sites; therefore antigen-antibody reaction is maximized during immunolabelling of proteins. This phenomenon is evidenced by the minimal appearance of background labelling in comparison to that seen on paraffin wax embedded material, and the lesser degree of cross reactivity of immunoglobulin molecules (Onodera et al., 1992). However, cryosectioning alone does not usually preserve tissue structures sufficiently for use in diagnostic and research activities, and therefore some alternative approaches have been sought to promote preservation and subsequent labelling quality (Onodera et al., 1992; Hall et al., 1987; Richter et al., 1999). Using a chemical fixative prior to, or after, cryosectioning is the first choice to increase the tissue quality for immunolabelling. However, relatively few reports have appeared in the literature describing a systematic analysis of the advantages and disadvantages of different chemicals on tissue preservation of cryosections (Hall et al., 1987; Onodera et al., 1992; Richter et al., 1999). In this study, we first tested the simple organic fixative, acetone (AC), and then two other commonly used fixatives, formaldehyde (FA) and paraformaldehyde (PFA), in order to evaluate whether rapid AC fixation is sufficient to preserve the main cellular and extracellular matrix protein molecules, or whether relatively slow fixation agents – such as PFA and FA – should be used prior to cryosectioning. Four different organs with different tissue architecture were chosen to compare the fixation and staining efficiency of cellular proteins. Heart, kidney, brain and liver were fixed with PFA or FA for 24 or 48 h prior to cryosectioning or directly cryosectioned and postfixed with AC. All sections were visualized with multiple staining/
O. Cinar et al. immunolabelling procedures, including routine H&E stains, immunohistochemistry using anti-pancytokeratin antibody for cytokeratin filaments, fluorescently-labeled phalloidin for actin microfilaments and red emission dye 7-aminoactinomycin-D (7AAD) for nuclear labelling. Samples were examined both in a qualitative and a semi-quantitative manner in which line densitometric profiles were used to analyze the F-actin preservation efficiency of these three fixatives.
Materials and methods Tissue selection and fixation protocols To assess the overall efficiency of selected fixatives, four organs with different embryonic origins and with different proportions of epithelial, connective tissue, neuronal cells and extracellular matrices were evaluated in this study. The experimental protocol on animals was approved by the Institutional Review Board. Heart, kidney, brain and liver were sampled as follows: twelve adult Balb/c mice were divided into three groups. Mice in the first group were killed by cervical dislocation and the organs were immediately removed. 5 5 mm tissue blocks from the organs were immersed into Cryomatrixs (Shandon Scientific Ltd., Pittsburgh, PA, USA) and frozen for 10 min at 60 1C. The specimen temperature was raised to 15 1C, and 8 mm-thick cryosections were cut, air-dried for 10 min, and then fixed for 10 min with AC at 20 1C, pH 6.9 (ultra grade purity 99.9%, Applichem Co., Darmstadt, Germany) then stored at 4 1C until required. Animals in the second and third groups were anesthetized by intraperitoneal injection of 50 mg sodium pentobarbital (Sigma, St. Louis, MO, USA) per 100 g of body weight. After animals became insensitive to painful stimuli (within 5 min), animals in the second group were perfused with 10% FA solution containing 0.08 M dibasic and 0.04 M monobasic sodium phosphate (Merck Co., Darmstadt, Germany), final pH 7.4, for 10 min at room temperature, followed by immersion in the same fixative for 48 h at the same temperature. Animals in the third group were perfused for 10 min with freshly prepared 3.5% PFA, pH 7.4, (Merck Co., Darmstadt, Germany), and then immersed in the PFA solution for 24 h at 4 1C. Prior to cryosectioning, by the same procedure as stated for AC samples, both FA- and PFA-fixed blocks were immersed in 20% (0.58 M) and 30% (0.88 M) (w/v in distilled water) sucrose solutions (Merck Co., Darmstadt, Germany), for 12 h each, at 4 1C for cryoprotection (Romijn et al., 1999).
ARTICLE IN PRESS Fixative efficiency on cryosections
489 mental design, assessed the quality of tissue protection. Fluorescence intensity and staining/ immunolabelling patterns were evaluated using a line densitometric profile analysis, a component of the LSM 510 Meta software. The resolution of the line was set to 1000 points along 20 mm in all samples, corresponding to approximately nine sarcomeres. The grey value (8 bit ¼ 256) of each point along the designated line was determined with regard to distance (Ogiwara et al., 1999). This approach was used (i) to assess the fine staining property of actin filaments and (ii) to quantify the phalloidin binding efficiency to F-actin after fixation, both of which, as a whole, reflect the degree of tissue preservation. Representative micrographs (n ¼ 12) from each FITC-phalloidin-bound heart muscle section were chosen in order to compare the three fixatives. Staining intensities from randomly selected cardiac myocytes were compared by using the computer-based SigmaStat software package (version 3.0; Jandel Scientific Corporation, San Rafael, CA, USA). The differences among each group were separately tested utilizing one-way ANOVA and post hoc Dunn tests, with a level of significance set at po0:05.
Staining and immunolabelling procedures In this study, no detectable differences were noted between previously fixed sections stored at room temperature, 4 1C and –20 1C (unpublished observations). Sections stored at 4oC were initially stained with Harris’s hematoxylin (Sigma, St. Louis, MO, USA) and eosin using routine protocol for general orientation of the sections. Bright-field colored images were recorded from 36 different samples taken from the four organs. Sections adjacent to H&E stained samples were incubated with: (a) 50 mg/ml fluorescein isothiocyanate (FITC)-labelled phalloidin (a bicyclic heptapeptide toxin of Amanita phalloides) in phosphate buffered saline (PBS) containing 1% methanol, for 1 h, at room temperature (Sigma) (Can and Semiz, 2000); (b) affinity-purified mouse monoclonal anti-pancytokeratin antibody (Sigma) at a dilution of 1:100 in PBS, for 90 min at 37 1C, followed by Cy3-labelled goat anti-mouse IgG (Jackson ImmunoResearch, West Grove, PA, USA) at a dilution of 1:100 in PBS, for 90 min at 37 1C; or (c) 10 mM 7-AAD (Sigma, St. Louis, MO, (Sigma)) for 1 h at room temperature.
Confocal microscopy
Results
Slides were analyzed using a Zeiss LSM 510 Meta confocal microscope (Zeiss, Jena, Germany) equipped with two lasers (488 nm Argon ion and 543 nm He–Ne). A color-coded palette was used to optimize the grey value for proper acquisition of fluorescent images from each label. Detection parameters such as laser intensity, pinhole diameter, detector gain, amplifier offset and amplifier gain were set to identical values. 2048 2048 pixel single optical sections were recorded using Zeiss LSM Meta 3.2 version software.
Qualitative staining and immunolabelling results are summarized in Table 1.
Hematoxylin and eosin staining Results are illustrated in Fig. 1. Acetone: In brain, a loss of tissue organization and cytoarchitecture, artificial widening and swelling throughout the brain cortex and in the medulla (not shown) predominated. Similar findings were also seen in kidney and heart, and to a lesser extent in liver preparations. AC fixation resulted in an extreme loss of interstitial tissues around tubules in kidney, and in interfibrillar connective tissue in the
Evaluation of staining pattern and intensity Two different histologists, who were not informed about the study purpose or the experiTable 1.
Qualitative staining results of sections from four different organs (n ¼ 42)
Brain
AC FA PFA
Kidney
Liver
Heart
GS
CYK
F-A
Nuc
GS
CYK
F-A
Nuc
GS
CYK
F-A
Nuc
GS
CYK
F-A
Nuc
+ ++ +++
+ + +++
+ +++
+++ +++ +
+ ++ +++
+ ++ +++
+ ++ +++
+++ ++ +
++ + +++
+ ++ +++
+ +++ +
+ +++
+ + +++
++ ++ +++
+++ +++ +++
+++ +++ +
GS: General structure; CYK: Pancytokeratin immunolabelling; F-A: F-Actin staining: Nuc: Nuclear staining : no staining; +: poor; ++: moderate; +++: well.
ARTICLE IN PRESS 490
O. Cinar et al.
Figure 1. Representative images obtained from H&E-stained cryosections of AC-, FA- and PFA-fixed brain, kidney, liver and heart. AC-fixed samples display relatively poor fixation results as seen by the appearance of interstitial swelling and widening. FA-fixed samples provide better results, especially in brain sections. Loss of tissue organization and disrupted capillary walls are seen as prominent features in kidney, liver and heart. PFA-fixed samples gave optimum results regardless of the tissue studied. Scale bar ¼ 50 mm.
heart. Hepatic plates and capillary sinusoids were relatively poorly preserved by AC fixation. Formaldehyde: Brain sections were relatively well preserved compared to kidney, liver and heart sections. Giant pyramidal cells of brain cortex and intercellular matrices were clearly identified. In contrast, interstitial and interfibrillar tissues were severely disrupted in kidney and heart sections, respectively. More striking features were noted in liver samples. Paraformaldehyde: This fixative provided wellpreserved cell and tissue organization, particularly in brain, liver and heart sections. In kidney samples, glomeruli and tubules were moderately maintained with a minor disorganization of interlobular arteries and veins.
Fluorescence labelling Three different fluorescent markers were used to analyze the effects of fixatives on specific cellular
proteins. Anti-pancytokeratin antibody that was conjugated with Cy3 (red signal in Fig. 2) detects cytokeratin types 4, 5, 6, 8, 10, 13, and 18, and thus differentiates the epithelial cells from the other cell types (Bartek et al., 1991). The green marker in Fig. 3 is FITC-phalloidin, and this readily binds to F-actin in almost all cell types (Prochniewicz-Nakayama et al., 1983). Red signals in Fig. 3 correspond to nuclear dye 7-AAD. The following results were obtained by detecting those three cellular markers that were constitutively found in all sections. The effectiveness of the three fixatives was qualitatively and quantitatively evaluated using exactly the same imaging parameters in all observations and recordings. Acetone: In brain sections, a few foci were immunolabelled with the anti-pancytokeratin antibody; however they did not resemble any of the expected cellular phenotypes (Fig. 2). No F-actin labelling was detectable (Fig. 3). In contrast, nuclear 7-AAD staining was well preserved (Fig. 3). In kidney and liver sections, the pancytokeratin
ARTICLE IN PRESS Fixative efficiency on cryosections
491
Figure 2. Pancytokeratin antibody used to detect the antigen preservation efficacy of AC, FA and PFA fixation on cryosections. Pancytokeratin immunolabelling of epithelial cells was used to visualize the different amounts of epithelial components in the different tissues. AC-fixed samples yield blurry and non-specific results in all sections examined, with the exception of heart samples. FA provided more specific immunolabelling, as the capillary endothelial cells are identified particularly in kidney, liver and heart sections. More striking results are obtained on PFA-fixed samples that enable detection of both endothelial and other epithelium-originated cell types. Intertubular and intercellular capillaries are obvious against a black background, indicating an absence of non-specific labelling. Scale bar ¼ 40 mm.
immunolabelling pattern appeared fuzzy and nonspecific (Fig. 2). Inconsistent and uneven F-actin labelling was noted in kidney and liver sections (Fig. 3). However, 7-AAD labelling was confined to well-preserved nuclei (Fig. 3). Interestingly, sections from heart showed clearer labelling with all markers tested (Figs. 2 and 3). Formaldehyde: This fixative did not preserve any cellular cytokeratin filaments in brain. A blurry appearance of pancytokeratin-immunopositive areas in brain was not specifically localized to any recognisable cell type (Fig. 2). Similarly, phalloidin labelling did not seem to be specific to F-actin in brain sections (Fig. 3). Nuclei however were clearly identified by 7-AAD labelling, in a similar manner to AC-fixed groups (Fig. 3). FA successfully maintained the antigenic sites for binding of the anti-pancytokeratin antibody in kidney, liver and heart tissues (Fig. 2). A specific filamentous immunolabelling pattern was confined to epithelial cells in kidney
and liver sections (Fig. 2). Few pancytokeratinimmunopositive areas were detected in heart sections, due to the relatively low number of capillaries among cardiac myocytes (Fig. 2). FA also preserved actin in kidney, liver and heart. Slender actin filaments were clearly identified in glomeruli, tubules and in the vasculature of kidney (Fig. 3). Similarly fine labelling of intracellular actin filaments was seen in liver sections. Central veins of hepatic lobules, sinusoidal capillaries and bile canaliculi were clearly identified at high magnification in liver preparations. Paraformaldehyde: In general, all sections from all tissue samples displayed well-preserved intraand intercellular features, illustrated in Figs. 2 and 3. Nevertheless, pancytokeratin immunolabelling was restricted to capillary endothelial cells in all tissues tested (Fig. 2), and a consistently weak labelling of hepatocytes was observed (Fig. 3). FITC-phalloidin bound to all F-actin containing
ARTICLE IN PRESS 492
O. Cinar et al.
Figure 3. Dual labelling of AC-, FA- and PFA-fixed cryosections with FITC-phalloidin (green signal) and 7-AAD (red signal) for screening the actin microfilaments and cell nuclei, respectively. AC-fixed brain sections feature an absence of microfilament labelling, whereas a substantial increase in labelling intensity is noted after FA fixation, albeit in a non-filamentous pattern. Strikingly, bundles of F-actin positive-axons are identified in PFA-fixed brain sections. Kidney samples fixed with AC give poor and non-specific labelling results since the labelled areas do not seem to correspond to actin filaments, which are particularly found in the cell periphery. In contrast, FA and PFA preserve those filaments, facilitating differentiation of the general tissue components (see FA- and PFA- fixed kidney and liver samples). PFA shows the best F-actin preservation among the other two fixatives, as the delicate structure of cardiac myocytes with large amounts of actin microfilaments are clearly illustrated after fixation with PFA. Nuclear staining is generally acceptable in all specimens, except that little or cell-type specific staining is noted in FA-fixed liver and PFA-fixed kidney sections. Scale bar in brain ¼ 40 mm, other ¼ 20 mm.
intracellular structures in all samples, including axonal fibres, epithelial cells of the pia mater, glomerular podocytes, and sinusoidal endothelial cells (Fig. 3). Interestingly, varying degrees of nuclear labelling with 7-AAD were consistently observed in PFA-fixed samples. Only in liver sections was the nuclear staining of a quality sufficient to differentiate the cells clearly (Fig. 3). In the remainder of the samples, extremely weak nuclear staining was seen.
Evaluation of labelling pattern and intensity Twelve heart sections from each group of fixative group were selected randomly to compare the F-actin decoration (a non-immunological direct
binding of phalloidin to F-actin) using the line profile digitization technique, illustrated in Fig. 4. Signal intensities along a selected 20 mm line were statistically evaluated. In AC-fixed groups, dye staining specificity appeared low, as seen by the inconsistent and uneven peaks of signal intensities; no sarcomere pattern could be recognized in AC-fixed groups. In FA-fixed groups, dye distribution gave more specific results, albeit in a lower intensity (39% lower values compared to AC-fixed samples). Relatively better results were obtained in PFA-fixed groups, as successively sharp peaks of signals were noted that clearly identify nine successive sarcomeres. The cumulative results show that AC and PFA have higher labelling intensities than FA. There was no statistically significant difference between AC and PFA regarding the signal
ARTICLE IN PRESS Fixative efficiency on cryosections
493
Figure 4. Cryosections of heart labelled with FITC-phalloidin and subjected to densitometric analysis to reveal the staining quality and intensity of successive sarcomeres along a 20 mm line (white line) on randomly selected cardiac myocytes. Corresponding profiles below each micrograph demonstrates the degree of regularity of fluorescent peaks. AC fixation results in irregular peaks with noisy background, whereas more consistent peaks corresponding to successive nine sarcomeres are noted in FA- and PFA-fixed samples. Although labelling intensity in FA is significantly lower than PFA and AC (39%73.4), AC fixation is still not favoured in comparison to the other two fixatives. Y-axis refers to grey-scale value (0–255); x-axis refers to distance in mm.
intensities. However, the differences between PFA and FA and between FA and AC were statistically significant (po0:05).
Discussion During the last 30 years, a limited number of studies have focused on the comparison of fixatives and fixation methods aimed at preserving vital tissue organization as much as possible (Hall et al., 1987; Onodera et al., 1992; Richter et al., 1999; Gokhale et al., 2002). Frozen sectioning has achieved popularity for two main reasons. One is that false positive background staining can be an undesirable problem in paraffin wax embedded sections, particularly during observations using fluorescence labelling, and this makes it impossible in many cases to differentiate minute structures in sections. The second reason is that frozen sections provide an invaluable tool in diagnostic pathology; as the tissue processing procedure is extremely fast, it can be undertaken while a surgical operation is ongoing. In recent years, frozen sections have been considered as the preferred method for preparation of tissues for immunoenzymatic and immunofluorescent labelling techniques, particularly for preserving delicate protein molecules. Freezing is
generally thought of as a convenient method to protect antigenic determinants, although it is not often adequate for preserving general tissue organization (Prento and Lyon, 1997; Gokhale et al., 2002; Hoetelmans et al., 2002). However, protection of antigenic epitopes means very little if results are wrongly interpreted because tissue structure is distorted. In this sense, the effectiveness of tissue preservation of several fixation approaches still remains questionable. PFA and FA are aldehyde fixatives and form both intra- and intermolecular cross-links with protein molecules, which result in the formation of more rigid heteropolymers. The monoaldehyde FA is the simplest member of this family of fixatives, and has been commonly used for both light and electron microscopy. The reactions of FA with proteins are numerous and well understood. The first step involves nucleophilic attack by amino groups with the formation of amino methylol products, which then condense with other groups such as phenol, imidazole and indole moieties to form methylene bridges. The methylene bridges are considered responsible for the fixation of protein by formalin under conditions of fixation for electron microscopy (Griffiths, 1993). PFA is preferred over FA because it contains fewer impurities. Cross-linking does not always inhibit enzymatic activity, and even the preservation of a small percentage of enzymatic activity is usually
ARTICLE IN PRESS 494 sufficient for precise localization of the enzyme (Hayat, 1970). Like methanol and ethanol, AC is a simple organic coagulant that displaces water from proteinaceous materials, thereby breaking hydrogen bonds and disturbing tertiary structure to produce denaturation. Soluble proteins in the cytoplasm are coagulated and organelles are destroyed. Nucleic acids are not precipitated. Precipitated protein that has not been denatured retains enzymatic or other biological properties, and remains soluble in aqueous solution. AC extracts lipid from tissues but carbohydrate-containing components are largely unaffected. AC is used alone for fixing films and smears of cells, and previously for otherwise unfixed cryostat sections. It is not suitable for blocks of tissue. Unless they are very small, because it causes considerable shrinkage and hardening of tissue (Kiernan, 1990). This was seen in our AC-fixed samples, particularly of brain, kidney and heart sections. In contrast, swelling and enlargement of hepatocytes was observed. AC fixation has been reported to have severe detrimental effects in cell surface protein molecules and major cytoskeletal elements (Haas et al., 1988; Vielkind and Swierenga, 1989). Vielkind and Swierenga (1989) reported that microtubules were poorly preserved in AC-fixed cells, while, immunohistochemistry using a monoclonal anti-cytokeratin antibody produced an incomplete immunolabelling pattern against a diffuse background after FA fixation. Haas et al. (1988) demonstrated that PFA fixation maintained the morphologic integrity of the sperm acrosome and plasma membranes, while AC fixation was unacceptable. In this study, we tested the labelling pattern of two stable cytoskeletal proteins and the nucleus. In all of the organs studied, AC fixation resulted in poor results in preserving the cytoskeletal elements. In contrast, nuclear binding of 7-AAD to DNA revealed excellent results. The latter was previously tested on different tissues and cells (Can et al., 2003) and the results were consistent with the findings of the current study that AC fixation is fastest and best for subsequent 7-AAD labelling of cell nuclei. Similarly, Gillespie et al. (2002) found that AC was superior to aldehyde fixatives when the aim was to visualize the details of the nucleus. Taken together, AC fixation, particularly for frozen sections, is not a suitable method either for general microscopic evaluations or for immunohistochemistry (Gillespie et al., 2002). Hall et al. (1987) made efforts to improve the cellular morphology in cryostat sections using AC and/or periodate-lysine-PFA fixatives (PLP) on tonsil sections. They separated specimens into four groups, which were AC alone (25 min), PLP
O. Cinar et al. (10 min), PLP (8 min) followed by AC (2 min), and AC (2 min) followed by PLP (8 min). They found that AC alone gives excellent morphology, in contrast to our AC results. Nevertheless, they indicated that AC alone also gave poor immunolabelling and that this could be improved by using AC (2 min) followed by PLP (8 min). Prento and Lyon (1997) compared the effects of buffered formalin, Clarke’s ethanol-acetic acid, and ethanol alone on rat intestine, kidney and liver tissues, and for bright-field and fluorescence microscopy. Buffered formalin and Clarke’s ethanol-acetic acids gave better results compared to fixatives that contain some commercial substitutes or ethanol. Histological distortion, cell shrinkage and vacuolization were prominent when substitutes or ethanol were used. In contrast, they found that artefacts were occasionally found when buffered formalin or Clarke’s fixative were used. Immunohistochemical studies demonstrated a total loss of low molecular weight antigens after treatment with all fixatives, except buffered formalin. They concluded that the best immunolabelling was obtained by combining formalin fixation with subsequent antigen retrieval. Ethanol was tested as a fixative in the present study. Ethanol fixation resulted in poor morphology, histological distortion, cellular shrinkage and vacuolisation, consistent with the results reported by Prento and Lyon (1997). Rustad et al. (1989) compared the cell-protective and staining effects of FA, ethanol, and a fixative cocktail (a combination of ethanol, formalin and glacial acetic acid) on dermal tissues after cryosectioning. They concluded that FA fixation was characterized by a significant increase in the staining intensity of basal cells of epidermis. Delicate and precise cellular staining was remarkably good after the cocktail fixative in comparison to FA and ethanol. All studies cited above aimed to analyze the relationship between fixation and frozen sectioning. However, the effects of fixatives regarding general morphology and antigen protective efficiency on tissues derived from diverse embryonic layers, and having different cellular and extracellular matrix organization, have not been previously evaluated in any systematic way. The aim of the present study was to focus on that issue, and to address the overall efficiency of selected fixatives on four different organs of different embryonic origin and including different amounts of epithelial, connective tissue, neuronal cells and extracellular matrices. PFA fixation was superior to either FA or AC for subsequent H&E staining at the light microscope level. Further observations using higher resolution imaging systems may reveal
ARTICLE IN PRESS Fixative efficiency on cryosections the mechanisms by which tissue protective efficiency is achieved by these fixatives. In fact, confocal microscopy results have enabled a substantially higher degree of detection and quantification. Using several optical sections, we were able to demonstrate that PFA preserves actin filaments (the most abundant filamentous protein in muscle cells) compared to FA or AC. Similarly, pancytokeratin-immunolabelled filaments in epithelial cells were better retained by PFA compared to FA or AC. However, AC displays better fixation quality for nuclear labelling by 7-AAD. We recently proposed a simple and reliable fluorescent staining method that is specific to DNA for the visualization of cell nuclei using 488 and 543-nm visible lasers or conventional mercury lamps (Can et al., 2003). The staining properties of two different reagents, chromomycin-A3 (C-A3) and 7-AAD were analyzed in several types of mammalian cells and tissues with different fixation and preparation methods. We also evaluated the antigen preservation efficiency of PFA, FA and AC on cultured Chinese hamster ovary (CHO) cells, mouse oocytes, rat aorta and rat dermal connective tissue, both in paraffin wax embedded sections and frozen sections. C-A3 and 7-AAD were highly specific fluorescent markers in both PFA- and FA-fixed cultured cells, and in paraffin wax embedded and cryostat sections. The chromatin and chromosome fine structure was visualised depending on the stage of the cell cycle. As supporting data for our previous results, here we re-illustrate that AC is better than aldehyde fixatives in labelling the nuclear fine structure and morphology. Quantification of fixation and dye-binding efficiency presented in this work using a series of densitometric line profiles revealed the fine pattern of actin labelling. A digital signal processor embedded in the confocal system provided excellent image comparison to relatively thin optical slices obtained using the same control parameters of the imaging elements. Gray-scale values corresponding directly to staining intensities of randomly selected cardiac myocytes were analyzed. Our results showed that the mean intensities from FA-group were statistically lower than others (po0:05). No statistically significant distinction was noted between PFA and AC-fixed groups. However, based on the labelling quality and detailed analysis of the densitometric profiles obtained, we assume that PFA or FA yields finer results than AC. In conclusion, interactions between the fixative and cellular protein structures may result in enormous variations if controlled experiments have not been performed using alternative methods.
495 Regarding the type of fixative, PFA is likely to be the first choice among tested fixatives for solid tissue components. On the other hand, amplified staining or over-staining does not simply imply that good antigenic protection was obtained, due to the problem of false or/and increased positivity. Additionally, the loss of or variation in tissue composition raises the question about whether changes arise from either the fixation method or the experiment performed. Continued improvement of tissue processing methods is still considered as a critical step toward ultimately determining the complete structure of normal and altered cells.
Acknowledgements The authors would like to Dr. Mitch Halloran, Ph.D. for his valuable criticism of the manuscript; to Remzi Ata who handled the animal laboratory ¨ zkavukcu and Sercin Karahuseyiand Drs. Sinan O noglu for the blind observations and interpretations of the H&E-stained slides. This project was financially supported by Ankara University Biotechnology Institute.
References Bartek J, Vojtesek B, Staskova Z, Bartkova J, Kerekes Z, Rejthar A, et al. A series of 14 new monoclonal antibodies to keratins: characterization and value in diagnostic histopathology. J Pathol 1991;164:215–24. Can A, Semiz O. Diethylstilbestrol (DES)-induced cell cycle delay and meiotic spindle disruption in mouse oocytes during in-vitro maturation. Mol Hum Reprod 2000;6:154–62. Can A, Semiz O, Cinar O. Two convenient methods for nuclear labeling in confocal microscope with visible lasers. J Histotechnol 2003;26:147–56. Gillespie JW, Best CJ, Bichsel VE, Cole KA, Greenhut SF, Hewitt SM, et al. Evaluation of non-formalin tissue fixation for molecular profiling studies. Am J Pathol 2002;160:444–57. Gokhale S, Ololade O, Adegboyega PA. Effect of tissue fixatives on the immunohistochemical expression of ABH blood group isoantigens. Appl Immunohistochem Mol Morphol 2002;10:282–8. Griffiths G. Fixation for fine structure preservation and immunocytochemistry. Fine structure immunohistochemistry. Berlin: Springer; 1993. p. 26–80. Haas Jr. GG, DeBault LE, D’Cruz O, Shuey R. The effect of fixatives and/or air-drying on the plasma and acrosomal membranes of human sperm. Fertil Steril 1988;50:487–92.
ARTICLE IN PRESS 496 Hall PA, Stearn PM, Butler MG, D’Ardenne AJ. Acetone/ periodate-lysine-paraformaldehyde (PLP) fixation and improved morphology of cryostat sections for immunohistochemistry. Histopathology 1987;11:93–101. Hayat MA. Biological applications. Principles and techniques of electron microscopy, Vol. 1. New York: Van Nostrand Reinhold Company; 1970. p. 65–95. Hoetelmans RW, van Slooten HJ, Keijzer R, van de Velde CJ, van Dierendonck JH. Comparison of the effects of microwave heating and high pressure cooking for antigen retrieval of human and rat Bc1-2 protein in formaldehyde-fixed, paraffin-embedded sections. Biotech Histochem 2002;77:137–44. Kiernan JA. Fixation. Histological histochemical methods, 2nd ed. Oxford: Pergamon Press; 1990. p. 10–32. Ogiwara N, Usuda N, Yamada M, Johkura K, Kametani K, Nakazawa A. Quantification of protein A-gold staining for peroxisomal enzymes by confocal laser scanning microscopy. J Histochem Cytochem 1999;47: 1343–9. Onodera Y, Shimizu H, Yamashita S, Nishikawa T. Cryofixed, freeze-dried and paraffin-embedded skin enables successful immunohistochemical staining of
O. Cinar et al. skin basement membrane antigens. Histochemistry 1992;98:87–91. Prento P, Lyon H. Commercial formalin substitutes for histopathology. Biotech Histochem 1997;72:273–82. Prochniewicz-Nakayama E, Yanagida T, Oosawa F. Studies on conformation of F-actin in muscle fibers in the relaxed state, rigor, and during contraction using fluorescent phalloidin. J Cell Biol 1983;97:1663–70. Richter T, Nahrig J, Komminoth P, Kowolik J, Werner M. Protocol for ultrarapid immunostaining of frozen sections. J Clin Pathol 1999;52:461–4. Romijn HJ, van Uum JFM, Breedijk I, Emmering J, Radu I, Pool CW. Double immunolabeling of neuropeptides in the human hypothalamus as analyzed by confocal laser scanning fluorescence microscopy. J Histochem Cytochem 1999:229–35. Rustad OJ, Kaye V, Cerio R, Zachary CB. Postfixation of cryostat sections improves tumor definition in Mohs surgery. J Dermatol Surg Oncol 1989;15:1262–7. Vielkind U, Swierenga SH. A simple fixation procedure for immunofluorescent detection of different cytoskeletal components within the same cell. Histochemistry 1989;91:81–9.