Membrane systems of guinea pig myocardiumUltrastructure and morphometric studies.код для вставкиСкачать
THE ANATOMICAL RECORD 222:362-379 (1988) Membrane Systems of Guinea Pig Myocardium: UItrastructure and Morphometric Studies M.S. FORBES AND E.E. VAN NIEL Electron Microscopy Laboratory, Department of Physiology, University of Virginia School of Medicine, Charlottesuille, Virginia ABSTRACT The structure and quantitative contribution of membrane systems (transverse-axial tubular system [TATS] and sarcoplasmicreticulum [SR])have been investigated in the heart of the adult guinea pig. Although previous quantitative studies have been made of guinea pig myocardium, this is the first such study that has utilized tissue in which membrane system elements were clearly identified by selective staining (in this case by the osmium-ferrocyanide [OsFeCN] postfixation method). Both membrane systems are highly developed in ventricular cells, but a TATS is essentially absent from atrial myocytes. The ventricular TATS consists principally of large-bore elements which may be oriented transversely, axially, or obliquely, making numerous anastomoses with one another to form a highly interconnected system of extracellular spaces that penetrate to all myoplasmic depths of the ventricular cell. The cell coat that lines the lumina of these tubules is structured, containing fibrillar structures that run along the length of the tubule. The volume fraction (V,) of the ventricular TATS is low (2.5-3.2%), in consideration of the qualitative prominence of the TATS in these cells. The relative total population of sarcoplasmic reticulum is higher in the atria (V, of 10-11%) than in the ventricles (Vv of ca. 8%).In all guinea pig myocytes, several major structural divisions of SR can be discerned, which include network SR, junctional SR, corbular SR, and cisternal SR. Junctional SR (J-SR) in the atrial cells is limited almost exclusively to peripheral saccules of junctional SR (PJSR), whereas both interior J-SR and P J S R are present in the ventricle. Two distinct morphological types of PJSR appear in atrial cells, including both flattened and distended saccules, the latter resembling PJSR of lower vertebrate heart. Spheroidal bodies of SR with opaque contents (corbular SR) are prominent at or near Z-line levels of the sarcomeres of atrial and ventricular cells. Cisternal SR is likely a subset of network SR, but some examples appear related to rough endoplasmic reticulum. An overall impression obtained from this study is that guinea pig atria are composed of structurally primitive cells, whereas the ventricular cardiac muscle cells are more highly developed entities. In the parlance of our times, the term guinea pig has come to denote an experimental subject. Nowhere today is this more true than in the field of heart research, in which (for example) isolated atrial muscle cells from guinea pig have been found to be ideal subjects for electrophysiological studies with modern techniques such as patch-clamping (Bechem et al., 1983). In addition, the study of myocardial ultrastructure has long depended on the guinea pig as a source of material. Early electron microscopy of the intercalated disc demonstrated extensive gap junctions in guinea pig heart (Sjostrand et al., 1958; Dewey and Barr, 1964).The welldeveloped system of transverse tubules in guinea pig ventricle has been studied both in thin sections (Simpson, 1965) and in freeze-fracture replicas (Rayns et al., 1967, 1968). Further investigation revealed that a Tsystem is virtually nonexistent in the atrial cells of guinea pig heart, however (Sperelakis and Rubio, 19711, and that transverse-axial tubular system (TATS) devel0 1988 ALAN R. LISS, INC. opment in guinea pig ventricle is qualitatively complete by the time of birth (Forbes and Sperelakis, 1976). ReAbbreuiatiom IJSR JSR Interior junctional sarcoplasmic reticulum junctional sarcoplasmic reticulum network SR network sarcoplasmic reticuluin ferrocyanide-reduced osmium; "osmium ferrocyanide" OsFeCN peripheral junctional sarcoplasmic reticulum PJSR sarcoplasmic reticulum SR surface density; surface area of cell component per unit sv volume of cell, expressed in units of pm"/pm3 or pm-l TATS Vv transverse-axial tubular system volume fraction; volume of cell component per unit volume of cell, expressed as a percentage Received March 9, 1988;accepted May 24, 1988. Address reprint requests to Michael S. Forbes, Ph.D., Electron Microscopy Laboratory, Box 449,Department of Physiology, University of Virginia School of Medicine, Charlottesville, VA 22908. ULTRASTRUCTURE OF GUINEA PIG HEART cently the guinea pig atrioventricular node and bundle-of particular interest in functional studies-have been described in great detail in terms of the qualitatively differing structure of sarcoplasmic reticulum (SR) among the five anatomically discernible regions in that area (Tomita and Ferrans, 1987). Despite this intense interest in its function and structure, only a limited amount of quantitative work has been devoted to the guinea pig heart (Denoit and Coraboeuf, 1965; Hirakow and Gotoh, 1980).Neither of these earlier studies utilized tissue in which selective contrasting of membrane system elements (SR and T tubules) had been carried out. Our previous studies of mouse heart (Forbes e t al., 1984, 1985) have demonstrated that stereological measurement of routinely prepared heart leads to considerable underestimation of membrane system volume fractions (V), and surface densities (SV).We undertook the present study of guinea pig heart because of the important role this tissue has played to date in cardiac research. The stereological data generated by this investigation are analyzed together with equivalent parameters obtained from our previous study of mouse myocardial sarcoplasmic reticulum (Forbes et al., 1985);this is part of a n ongoing set of comparative stereological studies of mammalian hearts which differ from one another in physiological properties such as ambient rate of heartbeat. The use of selectively stained tissue has in addition emphasized some interesting qualitative aspects of the membrane systems of guinea pig myocardial cells, and these are noted as well in this communication. 363 and dehydrated in a graded series of ethanol solutions (70%, 95%, loo%), passed through propylene oxide, and gradually infiltrated with PolyBed 812 resin (Polysciences, Inc., Warrington, PA). Since our original study of selective staining of muscle membrane systems by means of postfixation in ferrocyanide-reduced osmium solution (OsFeCN Forbes et al., 19771, we have found three factors to be of crucial importance during this type of tissue processing. First, postfixation in OsFeCN should be carried out without agitation; use of a rotator at this stage has the undesirable effect of producing a low degree of SR delineation, to the point that no selective staining a t all may be found in many blocks. Second, it should be kept in mind that the osmium-ferrocyanide-cacodylate solution is not truly buffered, as pointed out by Neiss (1984); as made up, its intrinsic pH is generally in the range of 10.010.2, In our hands, postfixation at this alkaline pH tends to produce tracing of the TATS, with or without a degree of SR staining, whereas deliberate establishment of a postfixative solution pH of ca. 7.6 gives a more general delineation of the SR. The third crucial step is infiltration. Difficulties in securing adequate embedment of OsFeCN-treated tissues have been reported previously (Hoshino et al., 1976); we have found satisfactory infiltration to be promoted by a gradual introduction of the epoxy resin into the tissue through intermediary solutions of propylene oxide and resin used in the following approximate schedule: MATERIALS AND METHODS Tissue Preparation 2: 1propylene oxide:PolyiBed, overnight; 1 : l propylene oxide:Poly/Bed, 8 h r the next day; 1:2 propylene oxide:Poly/Bed, overnight; pure PolyBed, 6 h r under vacuum. Adult specimens of the Hartley strain of guinea pig (either sex) were used for these studies. Each animal was anesthetized with intraperitoneal injections of pentobarbital and its chest cavity opened to expose the heart. Whole-body perfusion was accomplished by insertion of a cannula into the tip of the left ventricle and injection of 30-40 ml of fixative solution, which left the circulatory system through a small incision in the right atrium. The primary fixative employed in all cases consisted of a n aqueous solution containing 3% glutaraldehyde, 3% dextrose, 3% dextran (81,500 avg. MW), and 50 mM CaC12 (pH 7.4-7.6). Following the period of perfusion (ca. 3 rnin), each heart was removed in toto and immersed in fresh fixative solution for 30-45 min, following which time small pieces were removed to vials of fixative and left a n additional 60 min. The tissue samples were then washed overnight in aqueous 3% dextrose-3% dextran-50 mM CaC12 (pH 7.4-7.6). Samples were subsequently processed by one of three regimens: 1) “conventional” preparation; 2) selective staining of sarcoplasmic reticulum or TATS; or 3) freezefracture. Solutions used for processing procedures 1)and 2) differed only in the postfixative. Tissue for procedure 1)was postfixed 2 h r in a n aqueous solution of 2% OsO4 in 0.1 M sodium cacodylate, pH 7.2-7.4. In procedure 21, SR staining was accomplished by immersion at room temperature for 2 hr in 2% OsO4 and 0.8% potassium ferrocyanide in 0.1 M sodium cacodylate, pH 7.6 (TATS tracing was effected at a higher pH [see below]). Tissues for both procedures 1)and 2) were then stained en bloc for 30 rnin in saturated aqueous uranyl acetate solution, In contrast to the postfixation step, all dehydration and infiltration steps up to the placement of tissues under vacuum should be carried out on a rotating device. For final curing, tissues are transferred to a 60°C oven (a vacuum oven is not necessary) and left there for 2 days. All resin-embedded tissues were sectioned with diamond knives mounted in Sorvall MT-2 or LKB Ultrotome-I11ultramicrotomes. Thin (50-70 nm) sections were collected on bare or Formvar-coated copper-mesh grids and stained sequentially with uranyl acetate (saturated solution i n 50% acetone: 2 min) and 0.4% alkaline lead citrate (Venable and Coggeshall, 1965: 1 min). Thicker sections (up to 2 pm) of selectively stained heart were collected and allowed to dry down on Formvar-coated grids (Forbes and Sperelakis, 1980a). Such sections were either examined unstained or stained for 2 min with lead citrate solution. For freeze-fracture (procedure 31, glutaraldehyde-fixed tissue samples were gradually infiltrated with glycerol in dextrose-dextran solution up to a final glycerol concentration of 30%. The samples were frozen in liquid Freon 22, fractured at -100°C in a Balzers BAI? 300 freeze-fracture apparatus, and allowed to etch for 2 min prior to coating with a 2-2.5 nm layer of platinum a t a shadowing angle of 39”, followed by application of a stabilizing layer of carbon. The platinum-carbon replicas were digested away from the underlying tissue in bleach solution, and collected on hexagonal-mesh copper grids. 364 M.S. FORBES AND E.E. VAN NIEL Most specimens were examined and photographed in a Zeiss EM 10CA transmission electron microscope operated at accelerating voltages of 60, 80 or 100 keV. Selected thick sections of TATS-traced ventricle were investigated with a Zeiss EM902 equipped with a Castaing-Henry type energy-filtering spectrometer, operated at 80 keV; “energy filtering electron microscopy” (EFEM) has been demonstrated to be quite effective in resolving fine structure in thick sections, particularly those of specimens treated with selective staining (Peachey et al., 1987). Stereological Measurements Volume fraction (Vv) and surface density (SV) Measurements of VV and Sv of the TATS and SR were carried out in a manner essentially identical with that previously described (Forbes et al., 1984,1985).Samples used for stereology came from tissues prepared by the variation on method 2) which produced staining of the SR. The regions sampled were right and left atria (exclusive of any regions containing conducting tissue), right and left ventricular wall, and right and left ventricular papillary muscles. Preliminary evaluation of ventricular wall by means of a nested analysis of variance (Shay, 1975)indicated that a total of six animals was sufficient for adequate quantitation (at a confidence level of 90% with a sensitivity of 0.1%)of SR categories of relatively low incidence (e.g., peripheral junctional SR). Th‘in sections were prepared from two blocks of each region of interest (e.g., right ventricular wall), collected on Formvar-coated 150-mesh copper grids, and micrographs taken randomly of these sections (according to methods described in Forbes et al., 1985). VV values were obtained by point counts of micrographs printed at 14,500x magnification with a cross-hatch test lattice pattern photographically superimposed (scale interval of 0.63 pm). SVvalues were determined by printing, for each block, five of the micrographs used for Vv determinations at a magnification of 20,200 x together with a semicircular test overlay (in order to minimize the effects of any preferential orientation [anisotropy] of the SR: Weibel, 1973). VV values were also obtained in ventricular wall and papillary muscle for other cell components, including myofibrils, mitochondria, the TATS (in the guinea pig these large tubular elements are identifiable without selective staining), and nuclei. For atrial cells, all the components listed above were measured, as well as specific atrial granules, elements of the Golgi apparatus (including saccules and small vesicles), and lysosomes (the latter comprising multivesicular bodies, peroxisomes, and lipofuscin).SVvalues for the TATS were also obtained for all regions. Pairwise comparison of atrial or ventricular parameters was made by means of the Mann-Whitney test, and significant differences were judged to exist when P < 0.05. Myofibril-associatedsarcoplasmic reticulum In a previous study (Forbes et al., 1985), we carried out measurements of surface density of total SR and network SR, and compared these figures with surface density of myofibrils. Data used for this relationship were obtained from micrographs (printed at 25,000 x with semicircle test overlays) of transverse or near- transverse thin sections of SR-stained myocardium, in which myofibrillar surfaces could be distinguished not only by dint of their own substance, but also because of the closely apposed mitochondria and opacified tubules of sarcoplasmic reticulum. The relationship between SR and myofibril surface densities reduces t o the simple ratio of intersections, (I (sR)/I(myofibril)or 1(N-SR)&m ofibril). The quotient approximates the amount of myofiirillar surface covered by SR tubules. For a more stringent appraisal of this parameter, we recounted both the SV data from mouse right ventricular wall and guinea pig right and left papillary muscle, counting only the actual surface intersections of total SR and network SR with myofibrillar surfaces, and then divided each of these figures by the number of myofibrillar intersections obtained in the previous measurements. Another parameter of interest is the relationship between interior junctional SR and the TATS. One way of expressing the contribution of interior couplings formed by JSR-TATS apposition is to quantify the area of the TATS surface which is covered by interior JSR. This was obtained for each of the ventricular regions by dividing &~JSR) by Sv ATS), and then halving this value (since in general onyy one face of the IJSR abuts a TATS membrane: Forbes et al., 1985). RESULTS The most notable structural difference between the “working” myocardial cells of the atrial and ventricular regions of guinea pig heart is found in their overall degree of differentiation. In general, the atrial muscle cells are visibly thinner. The myofibrillar architecture of atrial cells is less strongly defined, furthermore, and their mitochondria1 complement is less prominent. Closer qualitative inspection and quantitative measurement of the ultrastructure of the two cell types has underscored this general impression, and uncovered additional differences, notably between their membrane systems. Atrial Cardiomyocytes Filling of the system of extracellular fluid spaces with osmium precipitate, in combination with the preparaFigs. 1-8. Left atrium of guinea pig heart. Fig. 1. Thick (ca. 130 nm) longitudinal section of several muscle cells. The sarcolemmal borders (SL) in this field are delineated by osmium precipitate, and for the most part display only small vesicular inpouchings (caveolae); a single distinct invagination is seen at the arrow. Nu, muscle cell nucleus. ~ 7 4 4 0 . Fig. 2. Detail of the transverse tubule (TT) in Figure 1. Aside from single caveolar vesicles (C), such short tubules are the sole sort of sarcolemmal invagination found in guinea pig atrial cells. x 25,000. Fig. 3. Longitudinal, 150-nm-thicksection of SR-stained atrial muscle cell. A variety of sarcoplasmic reticulum configurations is evident. The predominant category, network SR (N-SR), in association with myofibrillar surfaces displays a sarcomere-specific pattern of tubules that are heavily concentrated over the A-bands. A spheroid of corbular SR (C-SR) is anastomosed with network SR, this association distinguishing it from the specific atrial granules (SAG) that are found near the nucleus and scattered throughout the cell. Distended segments of SR (“cisternal” SR: Cs-SR) are intercalated with network SR, and a large cisternal mass of ribosome-decorated, rough endoplasmic reticulum (RER) is located in the nuclear pole myoplasmic region. The Golgi apparatus (GA) is usually not stained by osmium-ferrocyanide postfixation. x 16,000. 366 M.S. FORBES AND E.E. VAN NIEL tion of thick sections, provides a “negative” view of the sarcolemmal surface of guinea pig atrial cells and indicates a near absence of transverse tubular invaginations as well as a paucity of surface-connected caveolar vesicles (Fig. 1).The few tubular incursions that can be found are quite short, extending only to the level of the outermost myofibrils (Figs. 1,2). A wide variety of configurations of the sarcoplasmic reticulum is evident in atrial cells (Figs. 3-8). In OsFeCN-treated material, opacified profiles of SR tubules and cisternae can be distinguished from elements of the Golgi apparatus, which rarely fill with osmium precipitate. Sections through the nuclear pole regions of atrial cells typically contain profiles of the Golgi, accumulations of mitochondria, and collections of specific atrial granules (Fig. 3). Prominent as well in these regions are examples of rough endoplasmic reticulum, often massed into a localized collection of parallel cisternae; such cisternae are frequently resolved in grazing sections as extensive sheetlike structures (Fig. 3). The network (“free,” “longitudinal”) SR displays a variety of patterns, including meshworks applied beneath the surface sarcolemma (Fig. 4)and on the myofibrillar surfaces (Figs. 3-5). Connected with these coarse tubular retes there also appear simple longitudinal tubules (Figs. 3-5) and perforated cisternae (“fenestrated collars”), these last SR elements located predominantly over the middle levels (pseudo-H zones) of the sarcomeres (Figs. 3,5). Two divisions of the SR, corbular SR and Z tubules, are located a t or near the Z-line levels of atrial sarcomeres. Corbular SR elements are rounded or ovoid bodies, ca. 120 nm in average diameter (ovoid examples have a major diameter of up to 200 nm); they are anastomosed with the network SR, and often occur in clusters (Figs. 5, 6). Z tubules are elements of the network SR which form partial or complete collars around the Zline material; this encircling configuration is most obvious in transverse sections (Fig. 6). The absence of a system of transverse tubules from guinea pig atrial cells precludes, by definition, the existence of a significant population of interior junctional SR (IJSR), that specialized SR division which forms complex appositions (couplings) with transverse and axial tubules. Virtually all atrial couplings, therefore, are formed between the surface sarcolemma and the peripheral J-SR. In grazing thick sections that pass longitudinally along the sarcolemma (Fig. 4), the peripheral J-SR is evident a s distended, platelike bodies that are apposed to the inner sarcolemmal surface and interlinked with the subsarcolemmal tubular meshwork of network SR. Two distinct types of peripheral J-SR-as distinguished by their shape-exist in working atrial cells. Both forms of peripheral J-SR can coexist in the same cell, adjacent to one another along the sarcolemmal border, and examples of either type can be traced to connections with network SR. One type is the “conventional” sort of J-SR, each example a flattened saccule containing a linear array of intrasaccular ‘Ijunctional granules” (Fig. 7); the second comprises cisternae that are markedly distended in their vertical plane (Figs. 6, 8) and filled with lightly opaque granular contents similar in appearance to junctional granules in the attenuate peripheral J-SR (Fig. 8). In both types of couplings, bridging junctional processes (“pillars”) can be discerned in the space between the J-SR and sarcolemmal membrane leaflets (Figs. 7, 8). Estimates of relative frequencies (made from counts on the prints used for stereological data) indicate an approximate ratio between the flattened and distended varieties of peripheral junctional SR of 1:2 in both the right and left atria; that is, in guinea pig atrial working muscle cells, the engorged variety of’ PJSR is the predominant one. Ventricular Cardiornyocytes Transverse tubules Early studies of guinea pig ventricular myocardial cells established that the large-diameter invaginations from the sarcolemma are not merely a collection of transversely oriented tubules, but rather a n intricate and interconnected array of transversely, longitudinally, and obliquely oriented elements, which has come to be known as a “transverse-axial tubular system” (now more conveniently abbreviated to TATS). The anastomosis of TATS elements into latticeworks is particularly evident in thick sections (0.25 pm or greater) of OsFeCN-traced tissues (Figs. 9-11). It becomes especially apparent through examination of stereoscopic micrograph pairs that the TATS is not a simple architectural entity consisting of strictly transversely running tubules joined at precise right angles to absolutely vertical (“axial”) tubules. In particular, it can be noted that the axial tubules frequently display arclike profiles (Figs. 9, 10). Substantial TATS luminal interconnection exists for great distances within the cell, both in the longitudinal (Fig. 10) and transverse axes (Fig. 11). Adding to the complexity of the guinea pig ventricular TATS is the frequent presence of “doublets,” usually composed of pairs of transverse tubules (Figs. 12, 15). Such doublets are usually not derived from branching of transverse Fig. 4. Thick (200-nm) longitudinal section. Toward the right side of the field, close to the surface of a n atrial cardiac muscle cell, an extensive array of network SR tubules blends in numerous places with expanded saceules of peripheral junctional SR (PJ-SR) that form couplings with the inner surface of the sarcolemma. Proximity of this SR to the cell surface is indicated by caveolar profiles (C) interspersed with the meshes of the network SR. At deeper levels of the cell, network SR forms simple longitudinal tubules (LT) that lie in myofibrillar clefts. x 20,000. Fig. 5. Another field in the same section. This SR array contains spherules of corbular SR (C-SR)that are restricted to Z-line levels of the underlying sarcomeres, longitudinal N-SR segments (LT), distended cisternal SR (Cs-SR), and a perforated “fenestrated collar” region (FC). ~ 3 5 , 5 0 0 . Fig. 6. Transverse, 100-nm-thick section through SR-stained atrial cell. The SR is most prominent at the levels of the Z discs (Z), where it forms closely adherent uninterrupted stretches of so-called Z tubules (ZT). Corbular SR (C-SR) is also present at these levels. The simple cylindrical form of the longitudinal N-SR tubules (LT) is confirmed by their appearance in cross section. Expanded bodies of peripheral junctional SR (PJ-SR) abut the inner sarcolemma. ~36,000. Figs. 7, 8. Stereoscopic micrograph pairs of the two different forms of peripheral junctional SR in guinea pig atrial cells. Though the flattened saccule in Figure 7 is the more conventional category in appearance, the distended cisterna in Figure 8 represents the predominant type of PJ-SR. Strands of opaque material (junctional processes) bridge the myoplasmic gap between the SR and sarcolemmal surfaces; in some places these can be resolved into linear “pillars” (between arrows). Stereo angles: Figure 7, 10”; Figure 8, 5”. Both figures x 110,000. 368 M.S. FORBES AND E.E. VAN NIEL tubules, but rather (as best seen in freeze-fracture replicas) from the close but separate origin of T tubules from the surface sarcolemma (Fig. 12). TATS elements display substantial variation in their profiles, from round to flattened ovoid shapes; their diameters may be as small as 40 nm in extreme cases, but generally fall within the range of 200-360 nm. This compares well with previous estimates (310 nm: Rubio and Sperelakis, 1971; 340 nm: Rayns et al., 1975). Though cursory inspection of the TATS gives the impression of a highly ordered system, the geometric regularity of the TATS is in fact limited. This can be shown both by inspection of thick longitudinal sections (Fig. 9) and freeze-fracture replicas of the surface sarcolemma (Figs. 12-14). Although large areas appear in which T-tubule ostia are present in regularly spaced arrays-at or near the Z-line levels of the underlying myofibrils (Fig. 12)-many areas of the ventricular myocardial surface display irregular patterns of ostial location. There are regions in which discernible T-tubule openings are missing from many points in the surface lattice (Fig. 13), or in which ostia are altogether absent (Fig. 14).Given the large diameters of both the ostia (up to 600 nm) and the corresponding tubules, the origination points of the T tubules can readily be distinguished from the smaller surface-connected sarcolemmal invaginations, the caveolae (Figs. 12-14). Caveolae are not so prominent in guinea pig ventricle as they are in some other mammalian hearts (e.g., mouse), but appear in limited numbers attached to surface sarcolemma and to the TATS elements themselves (Figs. 15, 23). Caveolae of guinea pig heart are almost always found as single entities, in contrast to the multiunit “caveolar chains” found frequently in mouse (Forbes et al., 1984). Thin sections of conventionally stained tissue reveal the complexity of the TATS luminal contents, which frequently can be resolved into fibrillar structures of 12-14-nm diameter (Figs. 17, 18).These fibrils in longitudinal sections seem to correspond for the most part to structures located toward the centers of the lumina (Fig. 17). In addition, smaller (ca. 9 nm), roughly circular profiles can be made out, which possibly represent condensed regions of the cell coat that lines the margins of the tubules (Figs. 17,18). Sarcoplasmic reticulum The sarcoplasmic reticulum in ventricular cells is more highly organized than that of the atrial cells. The network SR tends to appear in more closely arranged, sarcomere-specificpatterns when apposed to the myofibrils than when it is located in a subsarcolemmal position (Fig. 19). Guinea pig ventricular peripheral junctional SR is best visualized in en face sections of the cell surface (Fig. 19). Interior J-SR is far more abundant, forming numerous appositions with the TATS (Figs. 15-20, 23). Junctional SR in ventricle, unlike atrium, consists solely of flattened saccules that contain linearly arranged junctional granules. Pillarlike structures are evident among the junctional processes that connect the JSR and TATS membranes (Fig. 16). Possibly because of the great size of the TATS elements, interior J-SR in guinea pig ventricle form couplings that are only partial enwrapments around transverse or axial tubules. The form of these couplings is best appreciated in thick SR-stained sections (Fig. 20) or-with thin sections-in fortuitous grazing planes through the TATS surface (Fig. 23). Another notable SR component in ventricular cells is cisternal SR. Each example of cisternal SR is a distended region intercalated with the regular small tubules of network SR, and not necessarily restricted to a specific sarcomere level (Figs. 20, 21); the surfaces of cisternal SR are frequently studded with ribosomes (Fig. 21). Fenestrated segments of SR (Figs. 20,26) also appear t o be a type of cisternal SR; these are rare, but when found tend to lie either at Z-line levels or over the pseudo-H zones of the sarcomeres. There is no ultrastructural evidence, in cisternal SR, of junctional granules or surface projections which might identify these segments as examples of the “extended junctional SR” as described in other mammalian hearts (Dolber and Sommer, 1980; Scales, 1981). Corbular SR bulbules are a notable feature of guinea pig ventricle; they are found attached primarily to the perimyofibrillar network SR, but not to the subsarcolemma1SR arrays (Figs. 19,ZO).Examples of ventricular corbular SR are similar in size and structure to their atrial counterparts, but are not strictly limited in location to the Z-line levels (Figs. 19,201.Structural similarities between corbular SR and the definitive junctional SR include intraluminal granular densities and opaque projections from the limiting membranes, the latter of which resemble the junctional processes of couplings (Fig. 22). Z tubules are prominent components of guinea pig ventricular SR (Figs. 23-26). As is the case in atrial cells, the close relationship between the SR and the Zline material is best demonstrated in transverse sections (Figs. 23, 24). Examination of longitudinal sections of SR-stained material, on the other hand, reveals a number of different SR configurations at the Z-line levels which correspond to “Z tubules” (Figs. 25,26). Stereological Observations In general, our stereological measurements of guinea pig heart indicate that the contributions of the various cell constituents are quite similar within the major heart regions (right atrium compared to left atrium, right ventricle compared to left ventricle) (Tables 1-3). Between the atria, in fact, the only statistically significant difference which can be discerned is in the myofibrillar volume fraction, which is greater in the right atrium than in the left (Table 1). These measurements also confirm the qualitative observations which previously had indicated that neither a T system nor a significant Figs. 9-1 1. Thick sections through TATS-traced myocardial cells in guinea pig right ventricular wall. Fig. 9. Half-micrometer-thick longitudinal section. The transverse tubules (TT) and longitudinally oriented (axial I tubules (AxT) are continuous at numerous points, thus forming an extensive interconnected tubular latticework. x 11,000. Fig. 10. In this field (0.5-pm-thicksection) the continuity of the TATS can be traced for a distance that extends for seven sarcomere lengths. Stereo angle, 20”. ~ 6 , 5 0 0 . Fig. 11. Quarter-micrometer-thick section which demonstrates the interconnected TATS in the transverse plane of the cell. Distended regions appear at various places in the latticework, particularly at the branching points (arrow). x 12,500. 370 M.S. FORBES AND E.E. VAN NIEL population of interior junctional SR is present in guinea pig atrial myocytes. Cells from the different ventricular regions sampled are likewise quite similar in their overall constitution (Tables 2, 3). Here small yet statistically significant differences are seen only in the VV(TATS) between right ventricular wall and right papillary (Table a), and in SV(IJSR), SV~JSR), and &(TATS) in some ventricular regions (as detailed in Table 3). The relationship between surface densities of interior junctional SR and the TATS of guinea pig ventricular myocardial cells is detailed in Table 4; this demonstrates that between 10.5%and 14%of the TATS surface is occupied by couplings. This is somewhat lower on average than equivalent values calculated for mouse ventricular cells (13-16%) in our previous studies (Forbes et al., 1984; Forbes and Sperelakis, 1987). When the question of coverage of the myofibrils by tubules of sarcoplasmic reticulum was investigated, it was found (Table 5) that the relative amount of myofibrillar surface directly apposed to network SR is similar in right and left papillary of guinea pig. This amount (ca. 43.5% for network SR) is, however, significantly lower than the value measured for mouse ventricular wall (ca. 49%). DISCUSSION Comparison of Guinea Pig Atrial and Ventricular Cells It is established that in vertebrates the primitive heart tube differentiates in a craniocaudal direction; the result of this is that the ventricle forms first and initiates the heart beat, with the atrial musculature and cells of the atrioventricular conducting system becoming defined and electrophysiologically functional at a later time during development. This delay in development may account for the less highly structured architecture of the guinea pig atrial muscle cells. Both the volume and cross-sectional area of guinea pig atrial cells are smaller than those of the ventricular cells, furthermore (Campbell et al., 1987). In functional terms, furthermore, mammalian atria are now proposed to constitute endocrine organs, in view of their production and secretion of vasoactive peptides such as atrial natriuretic factor and cardiodilatin (Forssmann et al., 1984). The appearance in adult guinea pig heart of "dilated" examples of peripheral junctional SR in the atrial working cells and nodal conducting cells (but not in the ventricular working cells) suggests a difference in the relative degree of differentiation of the major myocardial regions. The distended peripheral junctional SR of guinea pig is particularly reminiscent of myocardial junctional SR of lower vertebrate heart (lizard heart: Forbes and Sperelakis, 1974; salamander myocardium: Anderson et al., 1976). In mammalian heart, distended peripheral PJSR seems limited in general to the more primitive regions of the heart, including the conducting system (Rybicka, 1977; Tranum-Jensen, 1978; Tomita and Ferrans, 19871, and now is seen to exist in atrial working cells as well. Though appearing in the same cells with the more conventional flattened JSR saccules, the distended variety is actually twice as populous in guinea pig atrium. The primitive nature of the distended JSR is further underscored by its transient appearance in embryonic guinea pig ventricle, there also coexisting with the more conventional flattened JSR saccules (unpublished observations). TABLE 1. Summary of volume fractions of myocardial cell components in guinea pig atrial muscle cells' Parameter VV(SR) VV(N-SR) VWIJSR) VV(PJSR) VV(CSR) VV(CsSR) VV(rnyofib) VV(rnito) VWTATS) VV(Il"C) VV(SAGj VV(GAj VV(1vs) Right atrium Left atrium 9.97 f 0.80 8.98 f 0.61 10.92 k 0.67 9.96 + 0.56 -0- 0.51 f 0.14 0.32 + 0.09 0.15 + 0.02 45.27 2.04* 17.25 f 0.61 0.10 f 0.02 4.42 f 1.09 0.13 i 0.06 0.34 & 0.11 0.12 f 0.04 * -0- 0.40 f 0.06 0.30 k 0.06 0.26 0.06 41.14 1.01* 18.54 2 0.88 0.06 0.03 3.15 0.29 0.21 f 0.05 0.32 0.04 0.12 0.03 ** ** ** 'SR, sarcoplasmic reticulum; N-SR, network SR; IJSR, interior junctional SR; PJSR, peripheral junctional SR; CSR, corbular SR; CsSR, cisternal SR; myofib, myofibrils; mito, mitochondria; TATS, transverse-axial tubular system; nuc, nucleus; SAG, specific atrial granules; GA, Golgi apparatus; lys, lysosomes. * P < 0.05 in pairwise comparison with Mann-Whitney test. Thus in terms of morphological features of its atrial cells (namely, sarcomere structure, the virtual lack of the TATS, and appearance of the peripheral JSR), one can view the guinea pig heart as being composed of two distinct yet joined entities: a relatively primitive portion, comprising the conducting system and the atrial working tissue, and a highly differentiated portion, represented by the components of the ventricular chambers of the heart. It should be recognized that this argument is a generalization, in consideration of the transitional cell forms which have been described in other hearts (Martinez-Palomoet al., 1970), the presence of conducting tissue (the Purkinje system) within the ventricles, the admixture of atrial and ventricular myosin isozymecontaining cells in the ventricular tissue (Sartore et al., 1981), and the physiological superiority of atrial tissue in some respects (Urthaler et al., 1975). A feature common to both atrial and ventricular cells of guinea pig is the presence of distended portions of nonjunctional SR; similar structures have been termed "cisternal" SR by Scales (1981). In many cases a substantial subset of cisternal SR is the myocardial cell's complement of "rough" endoplasmic reticulum, shown by Slade and Severs (1985) to be prominent in rabbit heart. Even though rough endoplasmic reticulum is Figs. 12-16. Longitudinal freeze-fracture replicas of guinea pig right ventricular wall. Fig. 12. Stereoscopic micrograph pair (10" stereo angle). In this Pface replica there appears, at the 2-line levels, a fairly regular array of transverse tubule apertures (examples shown a t TT). Double T-tubule apertures (D) are commonly found. ~ 8 , 0 0 0 . Fig. 13. Stereo pair of sarcolemmal E-face (stereo angle of 10"). In contrast t o the field shown in Figure 12, the array of T-tubule apertures is neither regularly spaced into a lattice nor strictly located at Zline levels. In fact, many points at which a T-tubule opening would be expected to appear are devoid of any such profile. ~ 7 , 0 0 0 . Fig. 14. Expanse of E-face sarcolemmal leaflet (longitudinal axis of myocardial cell runs horizontally in this micrograph). Although the Zline levels of the myofibrils can readily be made out in this replica, the sarcolemma is virtually devoid of T-tubule apertures, and is instead decorated with numerous caveolar inpouchings (C). x 13,000. 372 M.S. FORBES AND E.E. VAN NIEL TABLE 2. Summary of volume fractions of cell components in guinea pig ventricular muscle cells Right ventr. wall Left ventr. wall Right papillary Left papillary 8.27 f 0.49 7.40 0.41 0.45 f 0.07 0.15 f 0.03 0.17 f 0.01 0.16 & 0.04 44.91 f 1.13 24.95 & 0.68 2.71 f 0.18" 2.94 i 0.52 8.17 f 0.43 7.35 i 0.33 0.41 f 0.05 0.12 k 0.03 0.15 k 0.03 0.16 i 0.03 45.54 & 0.98 25.71 i 0.75 2.53 0.18 2.58 k 0.25 8.25 i 0.31 7.46 f 0.28 0.35 f 0.05 0.12 0.02 0.17 f 0.01 0.15 0.02 45.08 k 1.87 25.45 f 0.77 3.23 i 0.09" 3.20 i 0.63 7.95 i 0.47 7.11 f 0.44 0.37 f 0.05 0.11 0.01 0.18 0.02 0.18 f 0.03 43.88 i 1.42 25.64 f 0.67 3.06 k 0.35 2.11 k 0.15 Parameter VV(SH1 VWN~SRI VV(IJSK, VV(P.JSR) VV(CXR) VV(CsSR) VVC myafi h 1 Vv(rnito) VV~TATS) VV(1,"d * * ** :kP< 0.05 between right VW and right papillary i n TATS parameter; no other statistically significant differences exist among ventricular wall and papillary parameters a s measured by Mann-Whitney test. TABLE 3. Summary of surface density values in atria and ventricles of guinea pig Parameter L. atr. Rt. atr. 2.187 f 0.212 SV(SR) SV(N-SR) 1.987 i 0.201 SV(IJSR) SV~JSR) 2.633 2.431 0.003 0.082 0.062 0.055 L. vw Rt. VW f 0.221 f 0.189 f 0.003 2.000 f 0.102 2.064 i 0.158 1.797 i 0.085 1.873 i 0.148 0.109 f 0.009""" 0.110 i 0.015 0.028 f 0.003** 0.015 f 0.003"" 0.037 f 0.005 0.031 0.004 0.028 f 0.007 0.036 i 0.013 0.439 & 0.019 0.396 f 0.040 Rt. papillary Left papillary 1.947 f 0.083 1.747 f 0.080 0.110 0.007" 0.035 f 0.004"" 0.025 5 0.004 0.031 0.006 0,477 k 0.020"" 1.941 & 0.143 1.773 f 0.133 0.080 i 0.011*,**" 0.020 f 0.003"" 0.032 + 0.004 0.037 0,011 0.376 & 0.018"" + 0.003 f 0.002 0.021 0.079 i 0.008 SV(CSR) 0.084 f 0.017 f 0.013 i 0.017 SV(C~SR) 0.035 k 0.013 SV(TATS) 0.018 i 0.008 0.009 i 0.004 * P < 0.05 i n pairwise comparison. "*P < 0.05 for t h e following pairs: right ventricular wall vs. left ventricular wall; right papillary vs. left papillary; right ventricular wall vs. left papillary; left ventricular wall vs. right papillary. ***P < 0.05 (left VW vs. left papillary). * * TABLE 4. Density of interior junctional sarcoplasmic reticulum saccules on transverse-axial tubules in guinea pig ventricular myocardial cells' Right ventr. wall 0.124 i 0.015 Left ventr. wall Right papillary .. Left papillary 0.141 f 0.011" 0.116 i 0.009 0.105 i 0.010" 'Derived from the equation SV,I.TSRJSV(TATS~~. < 0.05 (left ventricular wall versus left papillary). qualitatively more evident in the perinuclear cytoplasm son et al., 1973), each example so called because of its of guinea pig atrial cells, the total volume fraction of all strict and close adherence to the perimeter of a myoficisternal SR is quite similar throughout the heart (ranging from 0.15% to 0.26%). For the purposes of this comFig. 15. In this replica, a portion of the TATS latticework is exposed, munication, the distinction made between network SR encompassing both axial tubules (AxT) and a doublet (D) composed of and cisternal SR should be considered valid in structural two parallel transverse tubules in close apposition. Caveolalike bodies terms alone, since no physiological differences between (C) are fused with the TATS membranes. The plane of fracture also through a flattened saccule of interior junctional SR (IJ-SR) the two have as yet been determined, nor has the signif- tpasses h a t forms a coupling with a n axial tubule. ~31,500. icance of myocardial rough endoplasmic reticulumFig. 16. Detail of the coupling in Figure 15. The interior of the often joined with smooth-surfaced SR tubules-been established. The question raised by Slade and Severs (1985) junctional SR is exposed (IJ-SR), and particles can be discerned in the space between the SR and axial tubule (AxT). Some of these particles as to the tubular or cisternal conformation of rough form a thin pillarlike body (arrow). X 111,000. endoplasmic reticulum in mammalian heart is readily Fig. 17. Thin section of a transverse tubule i n right ventricular wall. answered-in the guinea pig at least-by the inspection Interior junctional SR (IJ-SR) is apposed to t h e tubule (cf. Figs. 15, 16). of thick, OsFeCN-stained sections, in which numerous Within t h e tubule is a distinct surface coat (arrows) which interconexamples of extensive cisternae of rough endoplasmic nects with fibrillar bodies (F)oriented along t h e long axis of the tubule. reticulum can be found. Distinct circular profiles ("1 appear i n a n adjacent T tubule caught in cross section. ~ 6 1 , 0 0 0 . Z tubules A particular structural specialization long associated with the SR of guinea pig heart is the "Z tubule" (Simp- Fig. 18. Stereoscopic micrograph pair (20" stereo angle) of a transverse tubule in thin section. The electron-opaque portion of the surface coat is separated from t h e tubule membrane by a space of ca. 20 nm, and some of t h e opaque substance is collected into circular profiles (arrows), implying a fibrillar substructure. ~86,000. 374 M.S. FORBES AND E.E. VAN NIEL TABLE 5. Ratio of apposed surfaces of sarcoplasmic reticulum and myofibrils: comparison of mouse and guinea pig ventricular muscle’ SV(tota1SR) SV(networkSR) Mouse right ventricular wall Guinea pig right papillary Guinea pig left papillary 49.60 k 2.10* 48.90 & 2.10* 44.52 f 1.76 43.35 i 1.75 44.55 f 1.83 43.46 i 1.84 ‘Measured in transversely sectioned cells; n = 10 for mouse, n = 6 for guinea pig. Values given as means k S.E.M. * P < 0.05 as determined in pairwise comparison with Mann-Whitney test. brillar Z disc. Some morphological evidence has suggested that Z tubules are physically connected to the Zline substance (Forbes and Sperelakis, 1980b, 1987). Z tubules are far more clearly discernible in transversely cut sections that pass through the Z-line levels of the sarcomeres. In such views, they appear as a partial or complete girdling ring around the myofibril. As reported by Simpson et al. (19731, guinea pig Z tubules are very difficult to detect in longitudinal thin sections. The use of thick sections of SR-stained material is therefore more efficacious in the visualization of longitudinal SR organization at the Z bands. Although simple slim tubules can on rare occasions be found over Z bands, it appears that for the most part in the guinea pig-as in the mouse (Forbes et al., 1985)-discrete Z tubules as seen in cross sections are largely an illusory optical phenomenon that is the result of the coalescence of a melange of tubular, cisternal, and fenestrated elements of network SR components at or near the Z-lines. Comparison of guinea pig and mouse heart When comparing the results of our quantitative measurements of guinea pig (the present study) and mouse heart (Forbes et al., 1984, 1985; Forbes and Sperelakis, 1987),one must keep in mind the difference between the total proportionality (volume fraction: see Table 2) versus the proportional location of the SR (Table 5). The first parameter merely indicates the total average fraction of the cell volume that is occupied by SR, and thus certain parameters (as measured in ventricular tissue) such as VV(SRflV(myofibril) are similar between the two species (ca. 0.16 in mouse, 0.18 in guinea pig). A small but significant difference is found, however, between the actual amounts of SR apposed directly to the myofibrillar surfaces in the two hearts, with the greater coverage found in mouse ventricle (ca. 50% in mouse, 45% in guinea pig [total SR]; 49% in mouse, 43-44% in guinea pig [network SR]). This indicates that in mouse ventricle there is a greater amount of network SR that is strategically located in the immediate proximity of the contractile mechanism of the myocardial cell. It remains for further comparative and quantitative studies of additional mammalian species to determine whether structural parameters such as this one can be generally and positively correlated with such physiological parameters as speed of shortening of the muscle cell (as roughly indicated by average heart rate of a particular species’ heart). One correlation of this sort has been postulated to exist between the relative incidence of specialized SR segments and the velocity of sarcomere shortening (Nassar et al., 1987). These workers noted that in isolated neonatal (3-week) rabbit ventricular cells, corbular SR is more prominent, and sarcomere shortening velocity lower, in comparison to the situation in isolated adult rabbit ventricular cells. Corbular SR is thought-on the basis of its calsequestrin content (Jorgensen et al., 1985) and its ultrastructural features (junctional granules and junctional processes: Dolber and Sommer, 1980, 1984)to be a form of junctional SR, specifically the mammalian equivalent of “extended” junctional SR (in bird heart these saccules presumably take the place of interior junctional SR: Jewett et al., 1971). Nassar and coworkers conclude that the rate andor amount of activator calcium release to the myofibrils from corbular SR stands to be lower than that from the definitive junctional SR which is in contact with the surface sarcolemma or TATS, through which is relayed the electrical or electrochemical signal that elicits calcium release. Many workers believe that calcium release can occur only across the “junctional face membrane” (that surface of the junctional SR which directly abuts the sarcolemma or TATS membrane) of the coupling (e.g.,Costello et al., 1986),and that the flow of calcium across network SR membranes is unidirectional: from the myoplasm into the SR lumen. In intuitive terms, this indicates that the efficiency of the contraction of myofibrils is lower than that of their relaxation; alternative hypotheses have been advanced which consider the possibility of the network SR’s parFig. 19. Survey micrograph of a 200-nm-thick longitudinal section through SR-stained right ventricular wall. The plane of section captures myoplasmic regions both at the cell surface (bottom of Geld) and deeper within the cell (toward the top of the field). The network SR (NSR) displays different patterns according to its location: subsarcolemma1 N-SR is a simple meshwork that intcrdigitates with caveolar profiles (C) and anastomoses with peripheral junctional SR saccules (PJ-SR).The N-SR surrounding the myofibrils observes a tighter, sarcomrre-oriented confieuration. and its comaonent tubules meree with bulblike bodies of corbular SR <C-SRjand sa’ccules of interior jugctional SR (IJ-SR). X20,OOO. ~~~~~~~~ ~~~~ Fig. 20. Stained SR array in a longitudinal 200-nm-thick section of right ventricular wall. Fused with the N-SR are numerous examples of corbular SR (C-SRj,not all of which are strictly aligned with 2 bands of the underlying myofibrils. Perforated cistctrnae C‘fenestrated collars”: FC) are located over both Z-line and A-band levels of the under^ lying sarcomeres, and simple distensions of the SR (cisternal SR: CsSR) also are present in connection with the M-SR, as are saccules of interior junctional SR (IJ-SR) that form coupiings with a transverse tubule (TT). ~ 3 2 , 5 0 0 . Fig. 21. Cisternal SR seen in a thin longihdinal section of right ventricular wall. This example of distended SR is decorated with granular bodies, which may be either ribosomes or 0-particles of glycogen. x 39,000. Fig. 22. Stereoscopic pair of a body of corbular SR, whose connection with N-SR can be seen (top of field). The corbule membrane surface is studded with projections of opaque material which in side view (arrows) resemble the junctional processes seen in couplings (cf. Fig. 171, and in en face view (double-headed arrow) appear circular. 10” stereo angle. x 141,500. 376 M.S. FORBES AND E.E. VAN NIEL ticipation in both phases of the excitation-contraction cycle (e.g., Forbes et al., 1985). However, if calcium indeed can only exit the SR across the faces of the various categories of junctional SR, then the contribution of corbular SR, along with its location, must be given consideration. Even though the excitation signal, which evokes calcium release from the corbular SR, may be delayed by its being transmitted through a circuitous route (from the interior junctional SR to the network SR, thence to the corbular SR), each of the numerous bulbules of corbular SR represents a nearly spherical face across which calcium can be liberated. In guinea pig ventricle, furthermore, corbular SR accounts for 2227% of the total junctional SR, and can exist at sarcomere levels other than the 2 lines (to which many of the definitive interior JSR saccules, coupled with transverse tubules, are relegated). The existence of an extensive population of corbular SR could therefore be viewed as a compensatory mechanism through which a more widespread release of activator calcium could be accomplished. This distribution of release sites would be augmented to some degree by the presence of couplings along axial segments of the TATS, as well as by the participation of the units of peripheral junctional SR, which are not positioned according to specific sarcomere levels. It is interesting in terms of the comparison of different species and heart regions to note that the incidence of corbular SR in mouse ventricle is very low. Though we have previously described corbular SR in mouse ventricle, in the context of its gravitation to the Z-line levels of the ventricular myofibrils (Forbes and Sperelakis, 1980b), its quantitative occurrence in comparison to other SR divisions is negligible (Forbes et al., 1985; Forbes and Sperelakis, 1987).In contrast, corbular SR is both a qualitatively and quantitatively prominent division of the sarcoplasmic reticulum in both atria and ventricles of guinea pig heart. Furthermore, corbular SR is quite prominent in the atrial cells of mouse heart (Forbes et al., 1985),and its volume contribution appears to be of the same general magnitude (40-50% of the entire junctional SR) as that in guinea pig atrium (unpublished observations). Symmetry of the guinea pig TATS A particularly useful technique for evaluating the geometric regularity of the guinea pig ventricular TATS is the inspection of E-face sarcolemmal replicas. In these replicas, the T-tubule apertures are clearly displayed as broken-off stumps which protrude from their origin at the sarcolemma toward the cell interior, whereas caveolae appear as single smooth-surfaced hemispheres. Our overall impression from our inspection of such replicas is that the arrays of ventricular T-tubule openings in guinea pigs are far from regular and geometric. Large areas of the sarcolemma may lack these openings (Fig. 14);such defects in the TATS surface arrays are reflected deep in the cell, where substantial volumes of myoplasm may lack a population of TATS elements. These observations can be reconciled with our stereological findings that the volume fraction of the guinea pig ventricular TATS is rather small (ranging from ca. 2.5-3.2%, depending on the ventricular region). Our measurements for VV(TATS) are also in general agreement with those of Denoit and Coraboeuf (1965)(3.24% in adult guinea pig right papillary) and Hirakow et al. (1980) (1.7-2.1% in 2-week postnatal ventricle). Previous qualitative examinations of thin sections (Sperelakis and Rubio, 1971)or freeze-fracture replicas alone (Rayns et al., 1967, 1968) have indeed suggested a far more substantial contribution of this membrane system element; in all likelihood this has been a result of the natural tendency to concentrate upon the most orderly TATS formations that come to view and to generalize from them. Comparison with previous stereological studies As pointed out in the preceding section, our measurements of volume fractions of the TATS are comparable to those found in previous studies (Denoit and Coraboeuf, 1965; Hirakow and Gotoh, 1980). An additional study of guinea pig left atrium was published by Frank et al. (1975). The major disparity between the present study and that of Frank et al. is the volume fraction of network SR (9.96%versus 1.7%),whereas the measurements for peripheral JSR are quite similar (0.4% and 0.5%, respectively). We have pointed out previously that a study based on selectively stained material tends to weight the measurement toward the side of the contrasted constituents (the SR, in the present case), particularly in the case of a cell constituent which would otherwise be difficult to resolve against the background of more structured components such as the myofibrils and mitochondria. In conventionally prepared and stained thin sections of heart, network SR is just this sort of constituent; accordingly, in stereological studies of such preparations its volume contribution has been substantially underestimated. On the other hand, an entity such as a saccule of junctional SR, which presents a characteristic shape and complement of junctional processes, is readily recognized in either conventionally or selectively stained material; therefore, it is not surprising that our study is in agreement with that of Frank et al. (1975) in terms of VV(~JSR) but differs substantially for VV(NSR).The same sort of argument holds true for measurements of the TATS, and again our values are similar to those previously published (see above). It was initially puzzling, therefore, t o compare our figures for volume fractions of the two major myocardial cell comFig. 23. Thin transverse section of SR-stained right ventricular wall. The plane of section passes at or near the 2-line level of many of the myofibrils in this field (Z); profiles of elements of the TATS are therefore prominent, some of which have associated caveolar projections (C). Interior junctional SR (IJ-SR), seen en face, forms a partial collar around a portion of a T tubule. Adjacent to the Z-band material are closely adherent tubules of SR (Z tubules: ZT). X39,OOO. Fig. 24. Transverse thin section of SR-stained right ventricular wall. A Z tubule forms a complete encirclement of a small myofibril at the level of its Z band. This Z tubule branches at several points (arrows)to lead into other tubules of network SR. AxT, axial tubule of the TATS. ~62,500. Flg. 25. Longitudinal thick (ca. 160 nm) section through the myofibrillar surface of a cell in right ventricular wall. Superimposed on the Z line is a simple profile of network SR, apparcntly corresponding to a Z tubule of the sort seen in transverse sections (cf. Figs. 23, 24). ~62,500. Fig. 26. Same section and orientation as in Figure 25. Network SR is coalesced over four successive Z lines (arrows). Though from this vantage point the SR varies in conformation from incidental N-SR tubules to fenestrated cisternal elements, transverse sections taken at any of these levels would yield profiles of “ 2 t,ubules,” since superposition of the stained SR components would mask their complex vertical organization. X 35,500. 378 M.S. FORBES AND E.E. VAN NIEL ponents-myofibrils and mitochondria-with those of Hirakow and Gotoh (1980), which substantially exceeded ours in the cases of VV(myofibril)of both atria and ventricles, and VV(mitochon&ia) for ventricular tissue. An extreme example is found in the case of the myofibrillar complement of left atrium (our value, 41.14%, as compared to 65.5% in the study by Hirakow and Gotoh). One possible source of difference is the age of the animals; Hirakow and Gotoh have presented a table of age-related changes in developing guinea pig heart, ending with the 2-week postnatal animal (our results come from the fully adult heart). Although much of the qualitative development of guinea pig heart seems to have been completed by the time of birth (Toth and Schiebler, 1967; Forbes and Sperelakis, 1976), it may be that further quantitative changes occur as the animal matures. One indication of this is seen in the fact that the results of Hirakow and Gotoh on the younger animals fail to reveal the difference in VV(myofibril) between the right and left atria that our study has shown in fully adult guinea pigs. As mentioned before, a stereological study of SRstained heart may yield overestimated values for SR components, particularly the network SR. This is a n example of the “Holmes effect,” as a result of which structures occupying only a fraction of the section thickness appear to exist at all levels of the section (because of the artifact of optical projection). Thus profiles of network SR stand out in OsFeCN-treated cells and-in measurement of Vv and SVvalues-are counted at the expense of underlying structures. Therefore, at the same time the SR is overestimated, myofibrils and mitochondria are subjected to some degree of underestimation. Even to consider every SR count of our study to represent a lost myofibril count (which obviously is not the case) would not account totally for the differences we have noted above, however. One critical advantage in making a quantitative study of specifically stained material is that of recognition. In comparing studies of SR-stained myocardial embedments versus conventionally prepared heart, it becomes apparent that this recognition factor applies not only to the SR, but also-by default-to adjacent structures (most often myofibrils and mitochondria). Simply put, it is far easier to determine whether a lattice point or surface intersection lies on a myofibril or on the juxtaposed SR when the S R can be clearly identified. Therefore, in lowmagnification electron micrographs of conventionally prepared tissue, the SR encircling a myofibril is virtually invisible-even if it is not superimposed on the myofibril, as is the case in transverse sections of the cell-and thus can be mistakenly scored as part of that myofibril. We believe, then, that the problem of structure discrimination that is intrinsic to the examination of conventionally prepared tissue likely accounts in substantial part for what we consider to be the overestimated volume fractions of SR-associated structures (such as myofibrils) in studies such as that of Hirakow and Gotoh (1980) on guinea pig and of Bossen et al. (1978) on mouse heart (cf. Forbes et al., 1985). ACKNOWLEDGMENTS This research was supported by a grant from the Public Health Service (HL 28329). Most of the tissues used in this study were generously provided by Dr. Luiz Belardinelli (Department of Internal Medicine, University of Virginia School of Medicine) and Mrs. Susan I. PurdyRamos (Department of Physiology, IJniversity of Virginia School of Medicine). Freeze-fracture replicas were prepared by Ms. Bonnie Sheppard of the Central Electron Microscope Facility of the University of Virginia School of Medicine. We are grateful as well to Ms. Lisa Kremer for her skillful preparation of the electron micrographs used for the stereology studies reported in this communication. Special thanks go to Mr. Guenter Lamprecht of Carl Zeiss, Inc., Thornwood, New York, for providing access to the Zeiss EM 902 instrument, and to Mr. Jean-Marc Theler (Department of Physiology, University of Virginia School of Medicine) for valuable discussions concerning aspects of this research. LITERATURE CITED Anderson, PAW., A. Manring, J.R. Sommer, and E.A. Johnson 1976 Cardiac muscle: An attempt to relate structure to function. J. Mol. Cell. Cardiol., 8t123-143. Bechem, M., L. Pott, and H. Rennebaum 1983 Atrial muscle cells from hearts of adult guinea-pigs in culture: A new preparation for cardiac cellular electrophysiology. Eur. J. Cell Biol., 31:366-369. Bossen, E.H., J.R. Sommer, and R.A. Waugh 1978 Comparative stereology of the mouse and finch left ventricle. Tissue Cell 10:773784. Campbell, S.E., A.M. Gerdes, and T.D. Smith 1987 Comparison of regional differences in cardiac myocyte dimensions in rats, hamsters, and guinea pigs. Anat. Rec., 219t53-57. Costello, B., C. Chadwick, A. Saito, A. Chu, A. Maurer, and S. Fleischer 1986 Characterization of the junctional face membrane from terminal cisternae of sarcoplasmic reticulum. d. Cell Biol., 103:741753. Denoit, F., and E. Coraboeuf 1965 Etude comparative de l’ultrastructure du myocarde chez le Rat et Ie Cobaye. C.R. SOC. Biol. (Paris), 159:2118-2121. Dewey, M.M., and L. Ban- 1964 A study of the structure and distribution of the nexus. J. Cell Biol., 28:553-585. Dolber, P.C., and J.R. Sommer 1980 Freeze-fracture appearance of rabbit cardiac sarcoplasmic reticulum. In: 38th Annual Proceedings, Electron Microscopy SOC.Amer. G.W. Bailey, ed. Claitor’s, Baton Rouge, pp. 630-631. Dolber, P.C., and J.R. Sommer 1984 Corbular sarcoplasmic reticulum of rabbit cardiac muscle. J. Ultrastruct. Res., 87:-190-196. Forbes, M.S., L.A. Hawkey, S.K. Jirge, and N. Sperelakis 1985 The sarcoplasmic reticulum of mouse heart: Its divisions, configurations, and distribution. J. Ultrastruct. Res., Y3:l-16. Forbes, M.S., L.A. Hawkey, and N. Sperelakis 1984 The transverseaxial tubular system (TATS) of mouse myocardium: Its morphology in the developing and adult animal. Am. J. Anat., 170:143-162. Forbes, M.S., B.A. Plantholt, and N. Sperelakis 1977 Cytochemical staining procedures selective for sarcotubular systems of muscle: Applications and modifications. J. Ultrastruct. Res., 60:306-327. Forbes, M.S., and N. Sperelakis 1974 Spheroidal bodies in the junctional sarcoplasmic reticulum of lizard myocardial cells. J. Cell Biol., 60r602-615. Forbes, M.S., and N. Sperelakis 1976 The presence of transverse and axial tubules in the ventricular myocardium of embryonic and neonatal guinea pigs. Cell Tissue Res., 166:83-90. Forbes, M.S., and N. Sperelakis 1980a Membrane systems in skeletal muscle of the lizard Anolis carolinensis. J. Ultrastruct. Res., 73.245261. Forbes, M.S., and N. Sperelakis 1980b Structures located at the level of the Z bands in mouse ventricular myocardial cells. Tissue Cell, 12t467-489. Forbes, M.S., and N. Sperelakis 1988 Ultrastructure of mammalian cardiac muscle. In: Physiology and Pathophysiology of the Heart, second ed. N. Sperelakis, ed. Kluwer Acad. Pub., Norwell, pp. 3-14 (in press). Forssmann, W.G., C. Birr, M. Carlquist, M. Christmann, R. Finke, A. Henschen, D. Hock, H. Kirchheim, V. Kreye, F. Lottspeich, J. Metz, V. Mutt, and M. Reinecke 1984 The auricular myocardiocytes of the heart constitute an endocrine organ. Characterization of a porcine cardiac peptide hormone, cardiodilatin-126. Cell Tissue Res., 238:425-430. Frank, M., I. Albrecht, W.W. Sleator, and R.B. Robinson 1975 Stereological measurements of atrial ultrastructures in the guinea-pig. Experientia, 31:578-579. ULTRASTRUCTURE OF GUINEA PIG HEART Hirakow, R., and T. Gotoh 1980 Quantitative studies on the ultrastructural differentiation and growth of mammalian cardiac muscle cells. 11. The atria and ventricles of the guinea pig. Acta Anat. (Basel), 108:230-237. Hirakow, R., T. Gotoh, and T. Watanabe 1980 Quantitative studies on the ultrastructural differentiation and growth of mammalian cardiac muscle cells. I. The atria and ventricles of the rat. Acta Anat., 108:144-152. Hoshino, Y., W.A. Shannon, Jr., and A.M. Seligman 1976 A study of ferrocyanide-reducedosmium tetroxide as a stain and cytochemical agent. Acta Histochem. Cytochem., 9:125-136. Jewett, P.H., J.R. Sommer, and E.A. Johnson 1971 Cardiac muscle: Its ultrastructure in the finch and hummingbird with special reference to the sarcoplasmic reticulum. J. Cell Biol., 49:50-65. Jorgensen, A.O., A.C.-Y. Shen, and K.P. Campbell 1985 Ultrastructural localization of calsequestrin in adult rat atrial and ventricular muscle cells. J. Cell Biol. 101:257-268. Martinez-Palomo, A,, J. Alanis, and D. Benitez 1970 Transitional cardiac cells of the conductive system of the dog heart: Distinguishing morphological and electrophysiological features. J. Cell Biol., 47: 1-17. Nassar, R., M.C. Reedy, and P.A.W. Anderson 1987 Developmental changes in the ultrastructure and sarcomere shortening of the isolated rabbit ventricular myocyte. Circ. Res., 61:465-483. Neiss, W.F. 1984 Electron staining of the cell surface coat by osmiumlow ferrocyanide. Histochemistry, 80231-242. Peachey, L.D., J.P. Heath, G.G. Lamprecht, and R. Bauer 1987 Energy filtering electron microscopy (EFEM) of thick sections of embedded biological tissues at 80 kV. J. Electron Microsc. Technique, 6:219230. Rayns, D.G., C.E. Devine, and C.L. Sutherland 1975 Freeze fracture studies of membrane systems in vertebrate muscle. I. Striated muscle. J. Ultrastruct. Res., 50t306-321. Rayns, D.G., F.O. Simpson, and W.S. Bertaud 1967 Transverse tubule apertures in mammalian myocardial cells: Surface array. Science, 156:656-657. Rayns, D.G., F.O. Simpson, and W.S. Bertaud 1968 Surface features of striated muscle. I. Guinea-pig cardiac muscle. J. Cell Sci., 3:467474. Rybicka, K. 1977 Sarcoplasmic reticulum in the conducting fibers of the dog heart. Anat. Rec., 189.237-262. Sartore, S., L. Gorza, S. Pierobon Bormioli, L. Dalla Libera, and S. Schiafino 1981 Myosin types and fiber types in cardiac muscle. I. Ventricular myocardium. J. Cell Biol., 88:226-233. 379 Scales, D.J. 1981 Aspects of the mammalian cardiac sarcotubular system revealed by freeze fracture electron microscopy. J. Mol. Cell. Cardiol., 13:373-380. Shay, J. 1975 Economy of effort in electron microscope morphometry. Am. J. Pathol., 81:503-512. Simpson, F.O. 1965 The transverse tubular system in mammalian myocardial cells. Am. J. Anat., 117:1-18. Simpson, F.O., D.G. Rayns, and J.M. Ledingham 1973 The ultrastructure of ventricular and atrial myocardium. In: Ultrastructure of the Mammalian Heart. C.E. Challice and Sz. Viragh, eds. Academic Press, New York, pp. 1-41. Sjostrand, F.S., E. Andersson-Cedergren, and M.M. Dewey 1958 The ultrastructure of the intercalated discs of frog, mouse and guinea pig cardiac muscle. J. Ultrastruct. Res., 1:271-287. Slade, A.M., and N.J. Severs 1985 Rough endoplasmic reticulum in the adult mammalian cardiac muscle cell. J. Submicrosc. Cytol., 17:531536. Sperelakis, N., and R. Rubio 1971 An orderly lattice of axial tubules which interconnect adjacent transverse tubules in guinea-pig ventricular myocardium. J. Mol. Cellular Cardiol., 2211-220. Tomita, Y., and V.J. Ferrans 1987 Morphological study of sarcoplasmic reticulum in the atrioventricular node and bundle cells in guinea pig hearts. Am. J. Anat., 180:100-122. Toth. A,, and T.H. Schiebler 1967 Uber die Entwicklung der Arbeitsund Erregungsleitungsmuskulatur des Herzeus v6n Ratte und Meerschweinchen. Histologische, histochemische und elektrophysiologische Untersuchungen. Z. Zellforsch., 69:22-40. Tranum-Jensen, J. 1978 The fine structure of the sinus node: A survey. In: The Sinus Node. F.I.M. Binke, ed. Martinus Nijhoff, Boston, pp. 149-165. Urthaler, F., A.A. Walker, L.L. Hefner, and T.N. James 1975 Comparison of contractile performance of canine atrial and ventricular muscles. Circ. Res., 37~762-771. Venable, J.H., and R. Coggeshall 1965 A simplified lead citrate stain for use in electron microscopy. J. Cell Biol., 25:407-408. Weibel, E.R. 1973 Stereological techniques for electron microscopic morphometry. In: Principles and Techniques of Electron Microscopy. M.A. Hayat, ed. Van Nostrand Reinhold, New York, pp. 237296.