Download structures of physiological interest in the frog heart ventricle

J. Cell Sri. II, 179-203 (1972) Printed in Great Britain

STRUCTURES OF PHYSIOLOGICAL INTEREST IN THE FROG HEART VENTRICLE SALLY G. PAGE AND R. NIEDERGERKE Biophysics Department, University College London, Govier Street, London WCiE 6BT, England

SUMMARY Two structures of physiological interest in frog heart ventricles have been examined in detail: (a) the layer of endothelial cells which encloses each bundle of heart fibres, and (b) the sarcoplasmic reticulum (SR) inside the heart fibres. Some additional observations on fibre sizes and types have been made. Movement across the endothelial cell layer of molecules (molecular or ionic size < 12-5 nm) occurs through narrow clefts separating each endothelial cell from its neighbour. This conclusion results from experiments made with the extracellular markers ferritin and horseradish peroxidase. A diffusion equation describing the movement of solutes into and out of the fibre bundle has been derived using several geometrical parameters, such as the length and width of the clefts and the size of the extracellular aqueous space inside the bundle, all of which were determined from electron micrographs of the tissue. The theoretical solution for a stepwise change of external calcium concentration gives a halftime of 23 s (± c 8 s, s.D. of 13 bundles) for diffusion equilibrium at the surface of the heart fibres; this value, however, is likely to be an overestimate, by some 20-30%, on account of several systematic errors which are described. The sarcoplasmic reticulum in heart fibres consists of a network of thin tubules which partially encircle the myofibrils at Z-line level and also form occasional longitudinal connexions. Branches extend to peripheral regions of the cell and terminate in close apposition to the inner surface of the cell membrane. The volume of the SR is estimated to be approximately 0-5 % of the myofibrillar volume of the cells. Cross-sectional areas of heart fibres (and also their shapes) vary considerably, from less than 2 to more than 100 fim1 (average 174 /4m1). Fibres of large size and small surface/volume ratio contain many fewer myofibrils and more glycogen granules than fibres of the same size but larger surface/volume ratio. Physiological implications of these results are discussed.

INTRODUCTION

Much work has been devoted to the fine structure of heart muscle fibres in various lower vertebrates (e.g. fish: Kilarski, 19640,6; amphibia: Lindner, 1957; Scheyer, i960; Naylor & Merrillees, 1964; Staley & Benson, 1968; Sommer & Johnson, 1969; Gros & Schrevel, 1970; Baldwin, 1970; and certain reptilia: Fawcett & Selby, 1958; Slautterback, 1963; Leak, 1967; Forbes & Sperelakis, 1971), and several features common to these tissues have become apparent. For example, the heart cells of these animals are small and devoid of a T- (transverse tubular) system, and the sarcoplasmic reticulum (SR) is sparse by comparison with that in most skeletal muscle cells. During recent work on the function of the frog heart (Niedergerke, Page & Talbot, 1969;

180

S. G. Page and R. Niedergerke

Chapman & Niedergerke, 1970) some further points of structural detail have come to light requiring closer investigation, which has been undertaken in the present work. One of the structures examined is the sheath of endothelial cells which surrounds individual heart fibre bundles or trabeculae (cf. Ecker & Wiedersheim, 1896), and our attention has been focused on the nature and size of the pathways, present in this sheath, for the movement of ions and other solutes between the ventricular cavity and the surface of the heart cells. On the basis of histological data which we obtained a diffusion equation has been set up describing the theoretical time course of such ion movements, in particular that for calcium. From the comparison of this writh the experimentally determined time course of the tension responses to changes of external calcium concentration (Chapman & Niedergerke, 1970) essential information for the interpretation of these responses is obtained. Another subject of study has been the structure of the sarcoplasmic reticulum of frog ventricle fibres. Our intention here was to assess the precise distribution and volume of this organelle within the cells and so obtain an estimate of the intracellular storage capacity of this tissue for calcium, the activator of contraction. Finally, some observations on fibre sizes and types have been made. These are relevant to the understanding of impulse conduction in this tissue and, in addition, provide data on certain parameters used for physiological measurements. METHODS Ventricles from the frog heart (Ranapipiens) were attached to glass cannulae inserted through the atrial-ventricular orifice, and the aortic trunk together with the aortic valves was removed to allow rapid perfusion of the ventricle. For routine fixation, the ventricles were first perfused with a Ringer's fluid containing io~° g/ml tetrodotoxin, which prevented contraction during the subsequent perfusion with the fixative. In the course of this work, various methods of fixation were used (usually at 20-22 °C, but occasionally at 4 °C): (i) fixation in a solution of 1 % OsO 4 and 005 M phosphate buffer (pH 70-7-4) for 0-5 h; (ii) fixation for 2-3 h in a solution of 5 % glutaraldehyde (also buffered with 0 0 5 M phosphate buffer at p H 7-0-7-4), followed by a period (several hours) of washing of the tissue in a 005 M phosphate buffer solution and a final 0 5 - h period of fixation in a solution of 1 % OsO 4 and 005 M phosphate buffer (pH 7-0—7-4); (lii) a 2-4 h period of fixation in a fluid containing 2-5 % glutaraldehyde and 2 % paraformaldehyde (Karnovsky, 1967) and 0-08 M cacodylate buffer (pH 7-2-7-4), followed by washing for several hours in a 0 1 M cacodylate buffer fluid and a final 0'5-h fixation in a solution of 1 % OsO 4 and either phosphate (005 M, pH 7-0-7-4) or veronal-acetate (0075 M> pH 7-2-7-4) buffer. Of these methods, (iii) was found to give the most consistent preservation of the sarcoplasmic reticulum, and was therefore used for the quantitative examination of this structure; nevertheless, most features of the SR to be described were also observed in tissue fixed by methods (i) and (ii). After fixation in the case of (i), or after the first fixation period in (ii) and (iii), the ventricles were detached from the cannulae and cut in half by a longitudinal incision from base to tip. Short lengths (1-2 mm) of bundles were then dissected from regions in which adjacent bundles were roughly parallel and in which little branching occurred. Many of these bundles, taken from the wall of the ventricular subcavities, had been oriented, approximately, in a longitudinal direction within the ventricle (i.e. from its base to tip), and they are therefore representative of the bundles studied by Chapman & Niedergerke (1970) in experiments with which the results of the present work are to be compared. Other bundles, however, were from different regions of the ventricle, but a systematic study of any of their properties relating to orientation or location has not been made. The isolated bundles were dehydrated in an ethanol series (in the case of methods (ii) and (iii) after the final fixation in OsO 4 ), then soaked in propylene oxide and

Fine structure of frog heart cells

181

finally embedded in Araldite (CIBA). In some instances, the tissue was stained before dehydration in a 0 5 or 2 % uranyl acetate solution which was buffered either with maleate (Karnovsky, 1967) or with veronal-acetate (Farquhar & Palade, 1965). However, since this 'en bloc' staining extracted the glycogen granules from the myocardial cells, it was used mainly when examining the endothelial sheath. Some ventricles were perfused before fixation (by methods (i) or (ii) above) with a Ringer's fluid containing the extracellular marker ferritin ( 0 1 - 0 2 mg/ml). For others, the perfusion fluid contained 1-2 mg/ml horseradish peroxidase, and subsequent fixation was with 1% paraformaldehyde, 1-5 % glutaraldehyde (o-i M cacodylate buffer, pH 7-2-7-4). In the latter case the isolated bundles were subsequently rinsed, first in o-i M cacodylate buffer fluid and then in distilled water, before a 10—20 min incubation period in a medium containing 005 % 3,3'diaminobenzidine tetrahydrochloride, o-oi % H a O 2 and 0-05 M Tris maleate buffer (pH 76) (Karnovsky, 1967). After incubation the bundles were again rinsed in water and finally fixed in 1% OsO< (method (i)). In experiments designed to identify intracellular structures which might be capable of storing calcium ions, the ventricles (cannulated as above) were first treated in one of two ways to make cell membranes permeable to the ions of the incubation medium: (a) by perfusion for 15-20 min with a 1 mM calcium Ringer's fluid containing 2 mM diaminoethanetetra-acetic acid (Thomas, i960); (b) by exposure, at 4 °C, to a glycerol-water (1:1) mixture for 13-18 h and then for 10 min to a solution containing 30 mM KC1, 5 mM MgCl 2 and 30 mM potassium phosphate buffer (pH 6-5) (Pease, Jenden & Howell, 1965). After either treatment, the ventricle was incubated for 1-1-5 h in a medium from which calcium uptake has been obtained with other muscles (Pease et al. 1965; S. Page, 1969): a solution of 5 mM adenosine triphosphate, 6 mM creatine phosphate, 5 mM potassium oxalate, 2 mM ethyleneglycol bis(aminoethylether)-7v^iVvtetra-acetic acid, o-8 mM CaCl s and 30 mM Tris-maleate buffer (pH 64). The subsequent fixation was according to method (i), except that the fixative was saturated with calcium oxalate. Finally, the tissue was dehydrated in 100% ethanol and propylene oxide, and embedded in Araldite. Sections were cut with glass or diamond knives on a Porter-Blum microtome and usually stained, first at 40-50 °C with a saturated solution in 5 0 % ethanol of uranyl acetate and afterwards with lead hydroxide (Karnovsky, 1961) or lead citrate (Venable & Coggeshall, 1965). The sections were viewed in a Siemens Elmiskop I microscope. For many of the measurements of bundle cross-sectional areas io-/*m thick sections were cut with a glass knife, and afterwards viewed with a phase-contrast microscope. For the purpose of calibrating the electron microscope when the precise magnification factor was required, a diffraction grating replica was photographed immediately after the tissue section had been examined, i.e. without an intervening change in lens current. At low magnification, < 3000 times, image distortion was reduced by working with the intermediate lens current switched off. Cross-sectional areas (of the subendothelial space, and of fibres and bundles) were determined either by planimetry or by cutting out and weighing appropriate areas of the print (see Results section for more detail). Length measurements, such as the circumference of fibres or fibre bundles, were made on the prints with a curvimetre map-reader. It should be mentioned that the values of the various parameters so determined were from bundles whose fibres had sarcomeres of about 2-2-2-5 Z4111) a n d had been fixed, therefore, in a slightly stretched condition.

RESULTS

As is well known (e.g. Ecker & Wiedersheim, 1896), muscle cells of frog heart ventricles are arranged in bundles, or trabeculae, surrounded by a layer of endothelial cells. These bundles branch and interconnect in a complex network to form the wall of the ventricle and the divisions between the various cavities inside it (Gompertz, 1884). The bundle size varies considerably, as is shown by the histogram of the distribution of the cross-sectional area of 91 such bundles in Fig. 1. Appreciable variation also exists in the outline of bundle cross-sections, of which only the smallest are approxi-

S. G. Page ad R. Niedergerke

182 20 16

ii 12 JO

i—11 :—in . n n;—! 4

8 12 16 20 24 Cross-sectional area, 103 //m 2

iu •

28

B

8 -i

4>

is

6 •