High-pressure freezing and freeze substitution of rat myocardium for immunogold labeling of connexin 43.код для вставкиСкачать
THE ANATOMICAL RECORD PART A 288A:1059–1067 (2006) High-Pressure Freezing and Freeze Substitution of Rat Myocardium for Immunogold Labeling of Connexin 43 CHRISTIAN MÜHLFELD* AND JOACHIM RICHTER Division of Electron Microscopy, Department of Anatomy, University of Göttingen, Göttingen, Germany ABSTRACT The value of high-pressure freezing (HPF) and freeze substitution (FS) for immunoelectron microscopy (immuno-EM) of the heart was investigated in bioptic specimens taken from isolated hearts of 0-, 5-, and 14day-old rats at baseline and at 15, 30, 45, and 60 min after induction of ischemia. The target antigen chosen here was the gap junction protein connexin 43 (Cx43). After HPF and FS, immunogold labeling was applied for detection of Cx43. Gold particles associated with gap junction areas, free plasma membrane, and annular gap junctions (AGJs) were counted and distributions compared by contingency table analysis. HPF and FS resulted in excellent preservation of antigenicity for Cx43. The mostly good preservation of the ultrastructure was limited by mechanical damage at the border and by ice crystal formation in the center of the tissue blocks. In normal myocardium of newborns, gold particles associated with free plasma membrane were frequently observed, with AGJs only seldom. In older rats, the opposite relation was found. During ischemia, no distribution changes occurred in newborn or 14-day-old rats. In 5-day-old rats, however, ischemia induced a shift of Cx43 from gap junction plaques to AGJs. In conclusion, HPF and FS are an ideal alternative to chemical ﬁxation for immuno-EM as the excellent preservation of antigenicity is combined with a well-preserved ultrastructure. The results indicate that the process of degradation of gap junctions via AGJs gradually increases during postnatal rat heart development, a process that may be accelerated by ischemia in an early developmental state. Anat Rec Part A, 288A: 1059–1067, 2006. Ó 2006 Wiley-Liss, Inc. Key words: high-pressure freezing; stereology; connexin 43 High-pressure freezing (HPF) has been introduced into transmission electron microscopy (EM) as a method of rapidly ﬁxing small specimens without prior chemical ﬁxation (McDonald, 1999). In comparison with aldehyde ﬁxation (routinely used in EM), HPF leaves the antigenicity of the frozen specimens unaltered by the ﬁxation process and therefore seems to be an ideal tool for immuno-EM (Monaghan et al., 1998). Further, it was shown that HPF is superior to freezing at ambient pressure, which results in formation of large ice crystals and consequently in destruction of tissue ultrastructure (Monaghan et al., 1998). If high-pressure frozen tissue is further processed by freeze substitution (FS) and embedded in methacrylates, it should yield materials highly suitable for immuno-EM. However, the speciﬁc value of HPF and FS for immunoÓ 2006 WILEY-LISS, INC. myocardium; ischemia; EM may vary with the biological material of interest (Nicolas and Bassot, 1993). The mammalian heart, for instance, is mainly composed of cardiomyocytes that are connected with several surrounding cardiomyocytes by gap junctions. Investigations comparing immersion and *Correspondence to: Christian Mühlfeld, Department of Anatomy, Division of Electron Microscopy, Kreuzbergring 36, D-37075 Göttingen, Germany. Fax: 49-551-397004. E-mail: firstname.lastname@example.org Received 12 May 2006; Accepted 17 July 2006 DOI 10.1002/ar.a.20380 Published online 1 September 2006 in Wiley InterScience (www. interscience.wiley.com). 1060 MÜHLFELD AND RICHTER perfusion ﬁxation have shown that the mechanical damage by cutting out a piece of myocardium may lead to alterations within the whole specimen (Gavin et al., 1991). Since specimens suitable for HPF have to be of a very small size (approximately 0.2 mm in diameter), this problem may arise here, too, although the time window from biopsy to ﬁxation can be reduced to less than 40 sec (Hohenberg et al., 1996). To our knowledge, there is no study that uses HPF and FS to investigate the distribution of a particular antigen in mammalian myocardium systematically. Therefore, we performed an immuno-EM investigation on normal and ischemic myocardium taken from postnatal rats of different age groups using HPF and FS. Immunogold labeling and ischemic damage of myocardium were evaluated by design-based stereological methods. The target antigen chosen for the present study was the major myocardial gap junction protein connexin 43 (Cx43) because immunogold labeling of Cx43 is known to be difﬁcult in myocardium treated with crosslinking ﬁxatives. Speciﬁcally, glutaraldehyde, even in low concentrations, markedly decreases Cx43 immunoreactivity (Luke et al., 1989). Cx43 is synthesized at the rough endoplasmic reticulum and transported to the free lateral membrane of cardiomyocytes via the Golgi apparatus, from where it is translocated to the specialized areas of the gap junction plaques. Due to their high turnover, gap junctions are frequently removed from the specialized membrane areas and degraded by lysosomes or proteasomes. One classical feature of this degradation pathway is the formation of annular gap junctions (AGJs), which consist of a ring of double membranes where connexins can be localized by immuno-EM (Dermietzel et al., 1991; Safﬁtz et al., 2000; Segretain and Falk, 2004). We quantiﬁed gold particles associated with gap junctional and nonjunctional plasma membrane and AGJ to study whether the distributions undergo changes during postnatal development and in the course of ischemia. This approach was used for two reasons. First, it is well known that profound remodeling processes of Cx43 distribution take place during postnatal development (Fromaget et al., 1992; Gourdie et al., 1992; Peters et al., 1994). Starting with discontinuous gap junction plaques around the entire cell surface in the newborn rat, after the ﬁrst postnatal weeks the typical localization of Cx43 at the longitudinal parts of the intercalated disks is reached. However, most studies on this topic rely on light microsopy and no data exist concerning possible changes in the preferential localization of Cx43 in free sarcolemma, AGJs, and gap junctions, respectively. Second, the redistribution of Cx43 during ischemia has important clinical impact because of ischemia-associated arrhythmias. Importantly, Cx43 has been shown to be an essential part in the protective response of the myocardium to ischemic preconditioning (Schwanke et al., 2002). In a previous study, Vetterlein et al. (2006) have shown that ischemic preconditioning results in an increased Cx43 immunoreactivity in the free plasma membrane of cardiomyocytes. As neonatal hearts are more tolerant to ischemia than those of adult rats (Singer, 1999) and neonatal ischemia tolerance shares similarities with ischemic preconditioning, we hypothesized that the myocardium of newborn rats might be equipped with a similar distribution like that shown by Vetterlein et al. (2006) for ischemic preconditioning and that this distribution may get lost during postnatal development. TABLE 1. Number of animals, body and heart weight 0d 5d 14d Number of 6 5 6 animals Body weight (g) 6.16 6 0.18 16.59 6 1.07 33.08 6 2.37 Heart weight (mg) 42 6 4 124 6 9 209 6 18 Our study clearly demonstrates that HPF and FS yield a good ultrastructural preservation combined with an excellent preservation of antigenicity of Cx43. HPF and FS may therefore represent an ideal alternative to aldehyde perfusion ﬁxation of the heart, especially if the antigenicity of the structure of interest is vulnerable to aldehyde ﬁxation. In accordance with our hypothesis, Cx43 immunoreactivity within the free plasma membrane was markedly higher in newborn than in older rat hearts, whereas the association of immunogold particles with AGJs was increased in the older rat hearts. Ischemia did only alter Cx43 distribution in the 5-day-old rats, a result we cannot explain so far. MATERIALS AND METHODS Animals The experiments described in this study comply with the current German laws and conform to the Guide for the Care and Use of Laboratory Animals published by the U.S. National Institutes of Health (NIH publications 85-23, revised 1996). The painless sacriﬁce of the animals was reported to the Bioethical Committee of the District of Braunschweig, Germany. Experimental Procedure Postnatal rats of different ages (day 0, n ¼ 6; day 5, n ¼ 5; day 14, n ¼ 6) were used in the present study. For further animal data, see Table 1. Animals were weighed, then sacriﬁced either by decapitation (newborn) or by cervical dislocation (5- and 14-day-old rats). The thorax was opened by a median sternotomy, the rib cage was spread by means of two needles and the beating heart was carefully excised, weighed, and washed in a Petri dish ﬁlled with isotonic NaCl solution. The heart was then transferred to a new Petri dish ﬁlled with fresh NaCl solution at room temperature and subjected to extracorporeal global ischemia (Mühlfeld et al., 2006). Immediately after this procedure, a biopsy was performed, taken from the left ventricular free wall using the microbiopsy transfer system described by Vanhecke et al. (2003) (Leica, Vienna, Austria). After transfer of the specimen to a ﬂat specimen carrier, the latter was placed into a pod and inserted into the cryoﬁxation machine using a loading device. The specimen was then high-pressure-frozen at a mean pressure of 2,023 6 5 bar as described by Studer et al. (2001) (Leica EM Pact). Further biopsies were performed after 15, 30, 45, and 60 min of ischemia, which underwent the same procedure. In all cases, the time from biopsy to freezing was approximately 30 sec. For the comparison of HPF with freezing at manual pressure and with chemical ﬁxation, respectively, two 14-day-old rat hearts were used. For each ﬁxation method, three tissue blocks per animal were ﬁxed. Freezing under HIGH-PRESSURE FREEZING OF MYOCARDIUM manual pressure was performed with the help of a forceps precooled in liquid nitrogen. For chemical ﬁxation, a ﬁxative with 4% paraformaldehyde in 0.2 M HEPES buffer was used. Tissue Processing and Immunolabeling The high-pressure frozen biopsy material was transferred to an AFS cryosubstitution unit (Leica, Bensheim, Germany) precooled to 908C. They were then immersed in 0.5% uranyl acetate in methanol for 72 hr, warmed to 458C at a rate of 58C/hr, and washed with methanol for several times. Thereafter, specimens were removed from the specimen carrier and immersed in Lowicryl HM20/ methanol 1:1 and 2:1 for 2 hr each. Finally, the specimens were placed into Lowicryl HM20, which was polymerized under UV light for 2 days at 458C. From these blocks, ultrathin sections were cut and mounted either on onehole copper grids or on 100 mm nickel mesh grids. The copper grid sections were stained with lead citrate and uranyl acetate and used for stereological estimation of ischemic injury. Ultrathin sections mounted on mesh grids were used for immunogold labeling for Cx43. First, the grids were ﬂoated on 0.02% glycine in Tris-buffered saline (pH 7.6) for 15 min and then on blocking buffer containing 5% fetal calf serum, 0.1% Tween 20, 0.25% bovine serum albumin in Tris-buffered saline (pH 7.5) for 30 min. Afterward, the grids were transferred to the anticonnexin43 antiserum (diluted 1:50 in blocking buffer to a ﬁnal concentration of 20 mg/mL; catalog number MAB3068; Chemicon, Canada) for 30 min. Grids were rinsed six times each for 5 min with blocking buffer, and visualization of immunoreactivity was performed by incubation with a secondary 10 nm gold-coupled antimouse antibody diluted 1:20 in blocking buffer (catalog number EM.GAM10; British Biocell, U.K.). Grids were washed four times for 4 min in distilled water, stained on a drop of uranyl acetate, rinsed three times with distilled water, and dried for 4 hr at 408C. Control sections of normal and ischemic myocardium were prepared by the same procedure as described above except for the omission of the primary antibody. Apart from very few gold particles without structural preference, all control sections were negative, i.e., no immunoreactivity of the secondary antibody with cardiomyocytes was observed. The chemically ﬁxed probes were ﬁrst immersed for 24 hr in 2.3 M sucrose and plunge-frozen before transfer to the cryosubstitution unit. The protocol then applied to the myocardium frozen under manual pressure or ﬁxed by aldehyde immersion was the same as the one applied to the high-pressure frozen tissues. Stereology and Electron Microscopy All electron microscopic investigations were performed with an EM 900 (Zeiss, Oberkochen, Germany) equipped with a digital camera (Megaview III, Soft Imaging Systems, Münster, Germany) and computer imaging software (AnalySIS, Soft Imaging Systems, Münster, Germany). For each animal and time point, one section was investigated to evaluate parameters of ischemic damage. At a primary magniﬁcation level of 7,0003, 60 images were 1061 obtained by systematic random sampling from each section. As ischemia is accompanied by swelling of myocyte mitochondria and cardiomyocytes as a whole, the degree of these alterations was established by an estimation of the volume-weighted mean volume of mitochondria and the cellular edema index as previously described in detail (Mühlfeld et al., 2006). In short, a test system with 18 lines and 36 points was projected onto each test ﬁeld. Whenever a test point hit a mitochondrial proﬁle, the edge-to-edge chord length was measured along the corresponding test line. From these measurements, the volume-weighted mean volume of mitochondria was calculated according to Gundersen and Jensen (1985). Additionally, points hitting myoﬁbrils, nuclei, mitochondria, and sarcoplasm were counted to estimate the volume densities of these compartments according to Weibel (1979). From these volume densities, the cellular edema index was estimated according to DiBona and Powell (1980). This parameter increases when cardiomyocyte swelling occurs. Immunoelectron microscopy for Cx43 was employed to estimate the mean number of gold particles associated with one of the following compartments. The myocyte cell surface was divided into specialized membrane area where plasma membranes of two adjacent myocytes were close enough to be bridged by two connexons, and nonspecialized membrane area where no direct contact with other myocytes was present or where adjacent myocyte membranes were divided by a gap too large to be bridged by connexons. The third compartment under investigation consisted of AGJs, which were deﬁned as intracytoplasmic organelles surrounded by two membranes. Only once was an extracellular AGJ found and was not included in the analysis. For quantiﬁcation of gold particles, one section per animal and time point was investigated. At a primary magniﬁcation of 7,0003, each section was scanned exhaustively by systematic random sampling. Thus, a total of roughly 600 test ﬁelds (with an individual area of ca. 9 mm2) per section was used to count gold particles associated with one of the three compartments described above. On each section, a total of approximately 150 gold particles was counted. The observed gold counts were used to calculate mean gold counts for each time point and age group. These mean gold counts were entered into the statistical analysis described below. Statistics To minimize the inﬂuence of the interindividual and developmental composition of cardiomyocytes on the evaluation of ischemic swelling of mitochondria and cardiomyocytes, cellular edema index (CEI) and volumeweighted mitochondrial volume were normalized by dividing these parameters of each time point by the normal value of the same animal, e.g., CEI (45 min)normalized ¼ CEI (45 min)/CEI (0 min). In order to analyze whether ischemia had induced a signiﬁcant cell and mitochondrial swelling over 60 min, Friedmans ANOVA was used. In order to recognize those points in time that signiﬁcantly contribute to the swelling, the normalized CEI and mitochondrial volume were compared by the Wilcoxon test for paired samples. In order to establish whether the degree of ischemic damage differs at a given time point of ischemia among the age groups, the corresponding values of a 1062 MÜHLFELD AND RICHTER TABLE 2. Contingency table analysis of gold counts in normal myocardium among the age groups SpM NSpM AGJ Column total Nobs (Nexp) 0d Nobs (Nexp) 5d Nobs (Nexp) 14d Row total partial v2 0d partial v2 5d partial v2 14d 87.2 (83.6) 19.2 (13.0) 3.3 (13.1) 109.7 (109.7) 67.8 (71.8) 13.4 (11.1) 13.0 (11.3) 94.2 (94.2) 137.0 (136.6) 12.7 (21.2) 29.5 (21.4) 179.2 (179.2) 292.0 45.3 45.8 383.1 0.153 2.984 7.307 10.444 0.223 0.465 0.265 0.953 0.001 3.407 3.032 6.441 In the ﬁrst column, the specialised plasma membrane (SpM), non-specialised membrane (NSpM) and annular gap junctions (AGJs) compartments are denoted. For each of the compartments, a mean number of associated gold particles was observed (Nobs). According to the equation given in the methods section the number of expected gold particles was calculated (Nexp). To see whether the distributions of observed gold particles within the three compartments were altered during postnatal development, partial v2 values and total v2 were calculated. For 4 degrees of freedom (3-1 compartments 3 3-1 groups) and with a total chi-squared value of 17.838 the null hypothesis that distributions are similar has to be rejected. Those compartments that make a substantial contribution to total chi-squared (more than 10% of total), i.e. the intergroup difference are highlighted in bold letters within the table. Thus, gold particles with non-specialised membrane are higher than expected in newborn rats and lower than expected in 14-day-old rats. The opposite holds true for AGJ. given point in time were compared by the Kruskal-Wallis ANOVA. The statistical evaluation of the immunoelectron microscopic data was performed according to Mayhew et al. (2004). The observed numbers of gold particles were compared by use of a contingency table analysis to depict distributional changes within one age group during the progression of ischemia or among the age groups in normal myocardium. From the number of observed particles, the number of expected particles was calculated according to the equation: column total 3 row total/grand total ¼ expected number of gold particles. Partial chi-square values for each group and compartment are given by: (observed golds expected golds)2/expected golds. The total chi-square value makes it possible to determine whether the observed distributions differ signiﬁcantly. If the observed gold distributions are different, the partial chisquare values will indicate those compartments that make a substantial contribution to the total chi-square value, i.e., the difference between groups. A substantial contribution is given by 10% or more of the total chi-square value (Mayhew et al., 2004). An example for this evaluation is reported for the age group differences of Cx43 distribution (Table 2). RESULTS Preservation of Ultrastructure and Antigenicity The preservation of myocardial ultrastructure by the given protocol was heterogeneous, as can be seen in Figures 1, 4, and 5. The quality of preservation depended on the age and the depth of the tissue block from which the section was cut. Mechanical damage at the border of the tissue block affected larger tissue volumes in the older rats due to the larger cell size. Ice crystal formation was frequently observed in the center of the tissue block commencing at a depth of approximately 75 mm. Ice crystal formation began nearer to the surface in tissue blocks from newborn than from older rats. Therefore, this artifact was more frequently observed in newborns than in older rats. In several sections obtained between 20 and 75 mm, a good ultrastructural preservation was found where neither damage due to mechanical distortion nor ice crystal formation was seen (Fig. 1a). However, in several sections, Fig. 1. Immunogold labeling of gap junctions. a: 60 min of ischemia. In 14-day-old rats, gap junction plaques (thin arrows) were usually in close connection with adherens junctions (white asterisks). Note the division of membranes after the gap junction area (thick arrow). b: 15 min of ischemia. In newborn rats, no conncetion with adherens junctions was observed. Gap junctions between adjacent myocytes (thin arrows) were interrupted by areas of nonjunctional membrane (thick arrows). HIGH-PRESSURE FREEZING OF MYOCARDIUM 1063 Fig. 3. Antigenicity in chemically ﬁxed myocardium. The myocardium shown here was ﬁxed with 4% paraformadehyde in 0.2 M HEPES buffer and shows Cx43 immunogold labeling of the longitudinal part (arrows) of an intercalated disk. The white asterisks mark the transversal parts of the intercalated disk. Although speciﬁc, the labeling was rather weak compared with HPF (see Fig. 1a for comparison). mi, mitochondria; mf, myoﬁbrils. Immunogold Labeling Fig. 2. Ice crystals in high-pressure frozen and manually frozen myocardium. a: Normal myocardium, 14-day-old rat. High-pressure frozen myocardium was usually well-preserved. However, within some areas, ice crystals (arrows) were observed that mainly affected the ultrastructure of myoﬁbrils (mf) and nuclei (nuc). Mitochondria (mi) were usually not altered. b: Normal myocardium, 14-day-old rat. Freezing under manual pressure led to a disturbed ultrastructure with large ice crystal formation. Nuclei (nuc) and myoﬁbrils (mf) were completely distorted, whereas mitochondria (mi) and the plasma membrane (arrows) were damaged less severely. ice crystal formation was present in varying degrees (Fig. 2a). In comparison to myocardium ﬁxed by manual freezing followed by freeze substitution (Fig. 2b), however, high-pressure freezing always yielded better results. The preservation of antigenicity of Cx43 was excellent in high-pressure frozen myocardium (Figs. 1, 4, 5, and 6) and superior to all previously tested chemical ﬁxation protocols (Fig. 3). Ischemic Injury In all age groups, signiﬁcant increases both in cellular edema index and in volume-weighted mean mitochondrial volume were observed, reﬂecting the ischemic injury (Table 3). Qualitatively, Cx43 immunoreactivity was observed in each of the three compartments deﬁned in the Materials and Methods section. Gold particles associated with the specialized membrane compartments were localized in small gap junction plaques in newborn myocardium (Fig. 1b). These were observed over the entire cell surface and interspersed with nonjunctional sarcolemma. In 5and 14-day-old animals, immunoreactive gap junction plaques occupied larger areas and were usually localized in close proximity to the intercalated disks (Fig. 1a). Only few immunogold particles were observed at the plicate parts of the intercalated disks. Gold particles associated with the nonspecialized plasma membrane were usually single particles. Sometimes a higher number of gold particles linked with the free plasma membrane was detected near gap junction plaques (Fig. 4). AGJs immunoreactive for Cx43 were observed in all age groups and were frequently located close to intercalated disks (Fig. 5). The mean number of gold particles associated with AGJs increased during postnatal development. Often, mitochondria were localized in close proximity to gap junction plaques or AGJs. Gold particles were sometimes observed with mitochondria; however, no systematic Cx43-labeling of mitochondria was observed (Fig. 6). The quantitative analysis revealed that the distribution of gold particles in normal myocardium differs between the age groups investigated (total chi-square value: 17.84; P < 0.01). In neonatal myocardium, the number of gold particles was lower than expected with AGJs and higher than expected with the nonspecialized membrane compartment. In 14-day-old rats, the opposite was observed (Fig. 7a). Ischemia did not signiﬁcantly inﬂuence the distribution of gold particles in neonatal and 14-day-old rats (total chi-square values: 7.33 and 10.79, respectively). In 5-dayold rats, ischemia led to a signiﬁcant difference in gold 1064 mV(Mi) [mm ] 3 CEI CEI, cellular edema index; mV(Mi), volume-weighted mean volume of mitochondria. *p < 0.05 towards control (0 minutes of ischemia); y p < 0.05 towards the previous sample. At no point in time did the values signiﬁcantly differ between the age groups. 0.10 6 0.07*y 0.65 6 0.31*y 0.87 6 0.06* 0.42 6 0.10* 0.83 6 0.08*þ 0.42 6 0.16*þ 0.76 6 0.08* 0.32 6 0.14 1.01 6 0.09* 0.45 6 0.18* 0.92 6 0.15* 0.36 6 0.14* 0.85 6 0.13* 0.41 6 0.24 0.70 6 0.10 0.25 6 0.08 0.94 6 0.16* 0.46 6 0.10* 0.98 6 0.11* 0.49 6 0.11* 1.17 6 0.14* 0.50 6 0.15* 0.58 6 0.04 0.22 6 0.10 0.79 6 0.1* 0.34 6 0.13 0.86 6 0.05* 0.33 6 0.09* 0.59 6 0.05 0.22 6 0.08 30 15 0 60 45 30 15 0 30 45 60 0 15 5d 0d Ischemia (min) TABLE 3. Stereological data on ischemic injury 14d 45 60 MÜHLFELD AND RICHTER Fig. 4. Immunogold labeling of free plasma membrane. Next to a capillary, a gap junction area between two myocytes is visible (thin arrows). After separation of the membranes (thick arrows), immunogold particles are found associated with the free myocyte membrane (14-dayold rat, 15 min of ischemia). endo, endothelial cell; cap, capillary lumen. distribution (total chi-square value: 29.35; P < 0.001) with increasing gold counts on AGJs and decreasing gold counts on specialized membrane compartments in the course of ischemia (Fig. 7b–d). DISCUSSION The present study clearly demonstrates that high-pressure freezing and freeze substitution are highly suitable tools to generate myocardial specimens for immunogold labeling and electron microscopy. HPF is a relatively new tool for ﬁxation of biological specimens for electron microscopy. In contrast to plunge freezing, HPF uses a pressure of approximately 2,000 bar in order to make the water vitrify and prevent formation of large ice crystals (Dubochet, 1995). Since no chemical ﬁxation is used, the ultrastructure is thought to be closer to the living state and antigenicity is remarkably well preserved (Monaghan et al., 1998). Similar results using HPF and FS were recently obtained by Rostaing et al. (2004) and Wang et al. (2005). Our experiences with this method demonstrate an excellent preservation of Cx43 antigenicity in myocardial preparations of postnatal rats combined with a good ultrastructure. The formation of ice crystals that was observed in several sections can be explained by the fact that before freeze substitution the frozen specimens were rewarmed to 908C. It is known that above 1008C ice crystal formation already commences (Dubochet et al., 1991). In general, ice crystal formation was more often found in the newborn than in older rat heart preparations. A possible reason for this may be the higher volume fraction of the watery content of cardiomyocytes, the sarcoplasm, in newborns (Mühlfeld et al., 2005). Moreover, HPF is only possible for very small specimens. Mechanical damage at the edges of the specimen will occur in every organ under examination; however, cardiomyocytes are extensively interconnected via gap junctions. Therefore, one must emphasize that between biopsy and ﬁxation of a myocardial tissue block, there may be a ‘‘conduction’’ of mechanical damage from the outer to the inner myocytes of the block. HIGH-PRESSURE FREEZING OF MYOCARDIUM 1065 Fig. 6. Localization of Cx43 with mitochondria. Not only were mitochondria (mi) observed in connection with AGJs and gap junction plaques (thin arrows), they sometimes even showed immunogold labeling for Cx43 (thick arrow). Fig. 5. Immunogold labeling of annular gap junctions. AGJs (thick arrows) were frequently found in 5-day-old (a) and 14-day-old rat hearts (b). Note the usual colocalization of AGJs with mitochondria (mi). a: 45 min of ischemia. The AGJ is located near a poorly preserved and labeled gap junction area (thin arrows). b: 60 min of ischemia. Two annular gap junctions are shown in the interior of a cardiomyocyte. Mitochondria with normal ultrastructure (mi) and such showing severe ischemic injury (smi) were often coexistent within one cell. In the present study, the myocardium was usually wellpreserved without signs of mechanical distortion within a depth of 20–75 mm of the tissue block. However, one should be aware that this range may easily be shifted toward greater distances from the border zone if the size of cardiomyocytes increases with age or in other species. The immunogold labeling technique used in the present study provided an excellent method for detection of Cx43 in postnatal rat myocardium. It should be mentioned, however, that connexin immunogold labeling is possible with other techniques and may also yield excellent labeling results (Luke et al., 1989; Fromaget et al., 1992; Gros et al., 1994; Laird, 1996). In a pilot study, we tested several paraformaldehyde/glutaraldeyhde ﬁxation protocols followed by cryosubstitution (Fig. 3) and compared them to HPF. As shown, the latter was superior to aldehyde ﬁxation in preserving myocardial Cx43 immu- noreactivity. However, if a protocol of chemical ﬁxation works for immuno-EM, it offers several advantages compared to HPF. For example, perfusion ﬁxation, as performed by Gros et al. (1994) for detection of Cx40, provides larger blocks of material in which different parts of an organ can be studied in a comparative manner within the same preparation. Stereological investigations on such material allow determination of the reference space and the experimental and ﬁnancial effort is less. The choice of method surely depends on which advantages and disadvantages are more relevant in a speciﬁc investigation. With respect to the application of stereological methods in the present study, several issues have to be discussed. First, after removal of the specimens from their carrier, they were dropped into the embedding capsules without preferential orientation. Thus, isotropy was assumed (Stringer et al., 1982). The application of more accurate methods to obtain isotropic uniform random materials such as the isector (Nyengaard and Gundersen, 1992) may only be possible after embedding and requires a second step of embedding the isectors. Second, no information about the reference space was available. Therefore, we only used stereological methods that do not require the estimation of the reference space (Howard and Reed, 1998). Third, the comparison of the gold labelings by the method described by Mayhew et al. (2004) served as a suitable tool for detecting differences in the labeling pattern. However, for a fuller understanding of the mechanisms behind the differences in labeling distribution, further data would be required such as surface density of the membrane compartments (Mayhew and Desoye, 2004). To verify our model, it was tested whether alterations typical of ischemia occur. Indeed, ischemia led to a significant swelling of cardiomyocytes and their mitochondria. Interestingly, the injury was similarly severe in all age groups. Given the concept of increased neonatal hypoxia and ischemia tolerance (Ostadal et al., 1999; Singer, 1999), this result is surprising. However, although there is a loss of ischemia tolerance in the rat already within the ﬁrst 14 postnatal days (Mühlfeld et al., 2005), this 1066 MÜHLFELD AND RICHTER Fig. 7. Relative distributions of gold particles. Relative distributions of gold particles associated with specialized membrane (white), nonspecialized membrane (black), and AGJs (gray). The relative distributions for a speciﬁc time point and age group were calculated by dividing the mean gold count of a particular compartment by the mean of the total particle number. The distributions for normal myocardium (0 min) were signiﬁcantly different between the age groups (a; see also Table 2). Neither in newborn nor in 14-day-old rats did a signiﬁcant distribution change occur in the course of ischemia (b and d). In 5-day-old rats, ischemia increased a translocation of Cx43 from specialized membranes to AGJs (c). Asterisks mark those compartments that contributed substantially to the intergroup difference. decrease is rather small and was not detected with the current methodology. Cx43 distribution was signiﬁcantly different among specialized, nonspecialized plasma membranes of myocytes and AGJs throughout postnatal development. Cx43 that is transported to the nonspecialized membrane via the Golgi apparatus and then translocated to gap junction plaques is frequently removed via annular gap junctions (Jordan et al., 2001; Segretain and Falk, 2004). Our results may allow the interpretation that in newborn rat hearts the new formation and transport of Cx43 to gap junction plaques exceed the removal. This interpretation corresponds well with the observed abundance of Cx43 mRNA in newborn rats and the subsequent postnatal decrease in Cx43 mRNA (Fromaget et al., 1990). Ischemia did not induce Cx43 distribution changes in newborn and 14-day-old rats. Only in 5-day-old rats was an increase in Cx43 immunogold labeling found in AGJs with a decrease of labeling in the specialized membrane compartment. This result is difﬁcult to interpret and we have no current explanation. In light of the intermediate distribution values of the 5-day-old rats under normal conditions, it could be hypothesized that ischemia induces a response that favors the Cx43 distribution of the older rat hearts. It remains to be determined whether age-dependent distribution differences are related to a different susceptibility of the postnatal hearts to ischemia. In fact, in a recent study (Vetterlein et al., 2006), we found that ische- mic preconditioning (Murry et al., 1986) induced a cardioprotection associated with a redistribution of Cx43 from gap junction plaques to the free plasma membrane, a distribution pattern similar to the one we found in newborn rat hearts in the present investigation. It is tempting to speculate that this distribution pattern is related to a higher tolerance to ischemia since both ischemically preconditioned and newborn rat hearts are more tolerant to ischemia (Murry et al., 1986; Singer, 1999; Mühlfeld et al., 2005). Since heterozygous Cx43-deﬁcient mice do not beneﬁt from ischemic preconditioning (Schwanke et al., 2002), it would be of particular interest whether they show the phenomenon of a higher myocardial ischemia tolerance in the newborn. In connection with the phenomenon of ischemic preconditioning, a study by Boengler et al. (2005) needs to be mentioned. These authors found immunoreactivity for Cx43 in mitochondria of rat myocardium, which was increased by ischemic preconditioning. In the present study, mitochondria were often observed in close proximity to gap junction plaques or AGJs and gold particles were sometimes associated with mitochondria. This may hint at a role of mitochondria in the function and metabolism of Cx43 molecules or vice versa. High-pressure freezing and freeze substitution of rat myocardium can successfully be employed to study the distribution of the gap junction protein connexin 43 by electron microscopy. The specimens generated by this ﬁxation and embedding technique are suitable to perform design-based stereology with respect to both the HIGH-PRESSURE FREEZING OF MYOCARDIUM ultrastructure and the immunogold labeling. However, one should be aware of the problems that may arise from tissue orientation and lack of knowledge about the reference space. Cx43 distribution in gap junction plaques, free plasma membrane, and annular gap junctions undergoes profound changes during postnatal development, which may have functional signiﬁcance in the course of myocardial ischemia. ACKNOWLEDGMENTS The authors thank Ms. S. Freese, Ms. H. Hühn, Ms. S. Kosin, and Ms. S. Wienstroth for excellent technical support, Ms. C. Maelicke for proofreading the English language, and Professor F. Vetterlein for proofreading the manuscript and discussing the results with us. The present study was supported by the research program of the Faculty of Medicine, Georg-August-University Göttingen (to C.M.), and by a grant from the Youth Travel Fund at the Federation of European Biochemical Societies (to C.M.). LITERATURE CITED Boengler K, Dodoni G, Rodriguez-Sinovas A, Cabestrero A, Ruiz-Meana M, Gres P, Konietzka I, Lopez-Iglesias C, Garcia-Dorado D, Di Lisa F, Heusch G, Schulz R. 2005. Connexin 43 in cardiomyocyte mitochondria and its increase by ischemic preconditioning. Cardiovasc Res 67:234–244. Dermietzel R, Hertzberg EL, Kessler JA, Spray D. 1991. Gap junctions between cultured astrocytes: immunocytochemical, molecular, and electrophysiological analysis. J Neurosci 11:1421– 1432. DiBona DR, Powell WJ. 1980. Quantitative correlation between cell swelling and necrosis in myocardial ischemia in dogs. Circ Res 47: 653–665. Dubochet J, Richter K, Roy HV, McDowall AW. 1991. Freezing: facts and hypothesis. Scanning Microsc 5(Suppl):S11–S15. Dubochet J. 1995. High-pressure freezing for cryoelectron microscopy. Trends Cell Biol 5:366–368. Fromaget C, Aoumari AE, Dupont E, Briand JP, Gros D. 1990. Changes in the expression of connexin 43, a cardiac gap junctional protein, during mouse heart development. J Mol Cell Cardiol 22: 1245–1258. Fromaget C, Aoumari AE, Gros D. 1992. Distribution pattern of connexin 43, a gap junctional protein, during the differentiation of mouse heart myocytes. Differentiation 51:9–20. Gavin JB, Maxwell L, Sage MD. 1991. Interrelationships of ultrastructure and function in the microvasculature of normal and ischaemic myocardium. J Electron Microsc Tech 19:429–438. Gourdie RG, Green CR, Severs NJ, Thompson RP. 1992. Immunolabeling patterns of gap junction connexins in the developing and mature rat heart. Anat Embryol 185:363–378. Gros D, Jarry-Guichard T, Velde IT, de Maziere A, van Kempen MJA, Davoust J, Briand JP, Moorman AFM, Jongsma HJ. 1994. Restricted distribution of connexin40, a gap junctional protein, in mammalian heart. Circ Res 74:839–851. Gundersen HJG, Jensen EB. 1985. Stereological estimation of the volume-weighted mean volume of arbitrary particles observed on random sections. J Microsc 138:127–142. Hohenberg H, Tobler M, Muller M. 1996. High-pressure freezing of tissue obtained by ﬁne-needle biopsy. J Microsc 183:133–139. Howard CV, Reed MG. 1998. Unbiased stereology: three-dimensional measurement in microscopy. Oxford: BIOS. Jordan K, Chodock R, Hand AR, Laird DW. 2001. The origin of annular junctions: a mechanism of gap junction internalization. J Cell Sci 114:763–773. Laird DW. 1996. The life cycle of a connexin: gap junction formation, removal, and degradation. J Bioenerg Biomembr 28:311–318. 1067 Luke RA, Beyer EC, Hoyt RH, Safﬁtz JE. 1989. Quantitative analysis of intercellular connections by immunohistochemistry of the cardiac gap junction protein connexin-43. Circ Res 65:1450–1457. Mayhew TM, Desoye G. 2004. A simple method for comparing immunogold distributions in two or more experimental groups illustrated using GLUT1 labeling of isolated trophoblast cells. Placenta 25:580–584. Mayhew TM, Grifﬁths G, Lucocq JM. 2004. Application of an efﬁcient method for comparing immunogold labeling patterns in the same sets of compartments in different groups of cells. Histochem Cell Biol 122:171–177. McDonald K. 1999. High-pressure freezing for preservation of high resolution ﬁne structure and antigenicity for immunolabeling. Methods Mol Biol 117:77–97. Monaghan P, Perusinghe N, Muller M. 1998. High-pressure freezing for immunocytochemistry. J Microsc 192:248–258. Mühlfeld C, Singer D, Engelhardt N, Richter J, Schmiedl A. 2005. Electron microscopy and microcalorimetry of the postnatal rat heart (Rattus norvegicus). Comp Biochem Physiol A Mol Integr Physiol 141:310–318. Mühlfeld C, Urru M, Rümelin R, Mirzaie M, Richter J, Schöndube FA, Dörge H. 2006. Myocardial ischemia tolerance in the newborn rat involving opioid receptors and mitochondrial Kþ-channels. Anat Rec 288A:297–303. Murry CE, Jennings RB, Reimer KA. 1986. Preconditioning with ischemia: a delay of lethal cell injury in ischemic myocardium. Circulation 74:1124–1136. Nicolas M-T, Bassot J-M. 1993. Freeze substitution after fast-freeze ﬁxation preparation for immunocytochemistry. Microsc Res Tech 24:286–290. Nyengaard JR, Gundersen HJG. 1992. The isector: a simple method for generating isotropic, uniform random sections from small specimens. J Microsc 165:427–431. Ostadal B, Ostadalova I, Dhalla NS. 1999. Development of cardiac sensitivity to oxygen deﬁciency: comparative and ontogenetic aspects. Physiol Rev 79:635–659. Peters NS, Severs NJ, Rothery SM, Lincoln C, Yacoub MH, Green CR. 1994. Spatiotemporal relation between gap junctions and fascia adherens junctions during postnatal development of human ventricular myocardium. Circulation 90:713–725. Rostaing P, Weimer RM, Jorgensen EM, Triller A, Bessereau J-L. 2004. Preservation of immunoreactivity and ﬁne structure of adult C. elegans tissues using high-pressure freezing. J Histochem Cytochem 52:1–12. Safﬁtz JE, Laing JG, Yamada KA. 2000. Connexin expression and turnover: implications for cardiac excitability. Circ Res 86:723–728. Schwanke U, Konietzka I, Duschin A, Li X, Schulz R, Heusch G. 2002. No ischemic preconditioning in heterozygous connexin43deﬁcient mice. Am J Physiol Heart Circ Physiol 283:H1740–H1742. Segretain D, Falk MM. 2004. Regulation of connexin biosynthesis, assembly, gap junction formation, and removal. Biochim Biophys Acta 1662:3–21. Singer D. 1999. Neonatal tolerance to hypoxia: a comparative-physiological approach. Comp Biochem Physiol A 123:221–234. Stringer BMJ, Wynford-Thomas D, Williams ED. 1982. Physical randomization of tissue architecture: an alternative of systematic sampling. J Microsc 126:179–182. Studer D, Graber W, Al-Amoudi A, Eggli P. 2001. A new approach for cryoﬁxation by high-pressure freezing. J Microsc 203:285–294. Vanhecke D, Graber W, Herrmann G, Al-Amoudi A, Eggli P, Studer D. 2003. A rapid microbiopsy system to improve the preservation of biological samples prior to high-pressure freezing. J Microsc 212:3–12. Vetterlein F, Mühlfeld C, Cetegen C, Volkmann R, Schrader C, Hellige G. 2006. Redistribution of connexin43 in regional acute ischemic myocardium: inﬂuence of ischemic preconditioning. Am J Physiol Heart Circ Physiol 291:H813–H819. Wang L, Humbel BM, Roubos EW. 2005. High-pressure freezing followed by cryosubstitution as a tool for preserving high-quality ultrastructure and immunoreactivity in the Xenopus laevis pituitary gland. Brain Res Protoc 15:155–163. Weibel ER. 1979. Stereological methods, vol. 1, practical methods for biological morphometry. London: Academic Press.