Ultrastructure of mitotic arrest induced by isoprenaline in rat parotid acinar cells

The observation that isoprenaline can induce proliferation of acinar cells in rodent salivary glands (Selye, Veilleux & Cantin, 1961) has led to the use of this drug in studies of the control of cell multiplication (reviewed by Baserga, 1970). Recent investigations have shown that isoprenaline has other effects on these cells, depending on their stage in the division cycle (Radley & Hodgson, 1971, 1973a). In particular, the drug can cause mitotic arrest. The processes leading to the arrest are unknown, and as a preparatory step towards their elucidation the effects of the drug on mitotic cells of the rat parotid gland have been examined at the ultrastructural level.

Female rats of the Canberra Black strain, aged about 8 weeks and weighing about 140 g, were used for this study. The frequency of mitosis among acinar cells of the parotid gland is normally very low in these rats, and to facilitate the study all animals were given an initial injection of isoprenaline to stimulate a wave of cell division (Radley, 1968). The mitotic frequency is appreciably increased at about 28 h after drug administration, although still averaging below 1 % of the acinar population. At this time the rats were either given a second injection of isoprenaline to determine the effect of the drug on mitotic cells, or examined as controls without further treatment.

The marked changes in ultrastructural appearance of mitotic cells which occur following an injection of isoprenaline are of a transitory nature, disappearing after several hours, and the use as controls of rats which had received a stimulatory dose of isoprenaline 28 h previously would appear to be justified. This assertion is supported by the similarities in appearance of mitotic cells in the control animals with those in the parotid glands of young rats (Redman & Sreebny, 1970).

A solution containing 20 mg isoprenaline sulphate per ml of normal saline was freshly prepared, and each rat given an intraperitoneal injection of 0·5 ml to stimulate a wave of cell division. Mitotic arrest was achieved by a further injection of the same amount of the drug 28 h after the initial dose.

Parotid glands were removed for examination at various time intervals up to 10 h after the second injection of isoprenaline. The rats were lightly anaesthetized with ether, and pieces of the parotid gland removed and fixed in a solution containing 2 % paraformaldehyde and 2 % glutaraldehyde in 0·06 M phosphate buffer (pH 7·2) at room temperature for several hours. After washing, the tissue was postfixed in 1 % osmium tetroxide in 0·06 M s-collidine buffer (pH 7·2) for 1 h, and then dehydrated in graded ethanol solutions and acetone, and embedded in Araldite. Semi-thin sections were cut at 0·5 μm thickness and stained with a 1 % borax solution containing 1 % toluidine blue. Ultrathin sections were stained in a 2 % aqueous solution of uranyl acetate, followed by a 1 % solution of lead citrate, or in the latter solution alone.

Cells blocked in mitosis are best considered as 2 groups – those in mitosis, and those in the G₂ phase – at the time of isoprenaline administration. A feature of cells in mitosis at the time of drug administration is that they do not become depleted of secretory granules, whilst cells in all other phases of the division cycle are emptied within 2 h of drug administration (Radley & Hodgson, 1971). For convenience, cells which are in mitosis at the time of administration of isoprenaline, and become blocked, have been designated Mf cells (f, filled with secretory granules). Mf cells include all cells in mitosis up to the metaphase stage; telophase cells appear to complete division.

Cells blocked in G. at the time of isoprenaline administration are blocked in this phase for about 2 h, during which time they are depleted of secretory granules, and then enter mitosis where they again undergo a temporary block (Radley & Hodgson, 1971, 1973b). These cells have been designated Md cells (d, depleted of secretory granules).

The appearance of acinar cells in the adult rat parotid gland has received the attention of several authors (Parks, 1961; Amsterdam, Ohad & Schramm, 1969; Simson, 1969). The only additional information pertinent to the present study relates to the centrioles. These organelles were situated in the apical region of the cell, often close to a canalicular or luminal space. In sections in which both centrioles of a cell were intercepted they appeared close together (Fig. 1), and in most instances the pair had lost the orthogonal arrangement which is present during telophase (Robbins, Jentzsch & Micali, 1968). The very low rate of proliferation in these cells implies that they are almost entirely in the G₁ phase, and consequently, movement and replication of centrioles associated with division will not generally be seen (Robbins et al. 1968; Tokuyasu, 1972).

Prophase. The centrioles existed as duplicate pairs in prophase, which had moved away from each other and towards the nucleus (Fig. 2). The shape of the cell changed during prophase from the pyramid form of interphase to a more rounded appearance.

The general appearance of metaphase cells resembled those observed in the parotid gland of young rats (Redman & Sreebny, 1970). A pair of centrioles, arranged orthogonally, was situated at each pole and microtubules extended out from the centriolar complex towards the chromosomes aligned on the equatorial plate (Fig. 3). Rough endoplasmic reticulum (RER) was arranged around most of the cell periphery, with mitochondria interspersed among the layers. Secretory granules, of average diameter 1·0 μm, occupied an intermediate position between the RER and the mitotic spindle.

Short lengths of RER, ribosomes, and clusters of small vesicles were present within the spindle volume. Metaphase cells were rounded in shape but displayed the same surface irregularities as interphase cells (Fig. 4). Microvilli were present on the plasma membrane exposed to the canalicular and luminal spaces, whilst interdigitations occurred frequently over the remainder of the cell surface.

Fig. 5 shows a typical example of an early telophase cell in which the daughter chromosomes have merged to form dense masses at opposite ends of the cell. Secretory granules were present in the zone between the forming nuclei, in the volume previously occupied by the spindle.

Later stages of telophase were seen in which the nuclear membranes had reformed and cytokinesis was well advanced. In Fig. 6 the intercellular bridge is clearly visible. With favourable sectioning, jhe cleavage furrow gave the appearance of having developed unilaterally from an invagination of the plasma membrane at the base of the cell (see Fig. 16). It would appear unlikely that this effect is due to the isoprenaline since it was observed 28 h after the injection. Furthermore, eccentricity of the inter-cellular bridge has been observed in the division of other types of cells in the rat (Buck & Tisdale, 1962; Dougherty & Lee, 1967).

The first observations on mitotic cells following isoprenaline administration were made after 15 min. Virtually all mitotic cells appeared to be blocked in metaphase at this time – an observation confirmed by the rapid decline of prophases and telophases (Radley & Hodgson, 1971, 1973b). Whenever a centriole was visible in an Mf cell it occupied a central location (Figs. 7, 8). In some cells serial sections were cut to verify that both pairs of centrioles were located near the centre, and it was noted that the orthogonal relationship between the centrioles of a pair did not appear to be affected by the block. In interphase cells the centrioles remained in the apical zone following isoprenaline treatment.

In sections of Mf cells in which centrioles were visible, they were surrounded by a circular array of chromosomes, suggesting a spherical arrangement in the intact cell (Fig. 7). Examination of 0·5-μm-thick serial sections under the light microscope confirmed that the chromosomes were arranged in the form of a thick spherical shell, but with some irregularities in coverage.

Numerous microtubules were seen in the vicinity of the centriolar complexes. In addition, microtubules were observed attached to kinetochores. Occasionally, with fortuitous sectioning, both kinetochores of a chromosome were visible, and had attached microtubules (Fig. 8). Secretory granules were present in much of the volume formerly occupied by the spindle, and were intermingled with chromosomes. The juxtaposition of some of the secretory granules and chromosomes visible in Figs. 7 and 8 was not confined to cells blocked in metaphase, but was also seen on the periphery of the metaphase plate during normal division (Fig. 3 A). The distribution of RER around the cell periphery was not affected by the drug.

Mf cells were blocked for at least 2 h following isoprenaline treatment. At later times the number of Mf cells in arrest had declined, and examples were seen of Mf cells with the normal metaphase configuration, and also undergoing cytokinesis, indicating that after release from the block some cells at least were able to complete mitosis. A conspicuous increase in the frequency of small secretory granules was noted in Mf cells several hours after the drug was given (not shown). This suggests that the cells are able to form secretory granules during the period of block.

The observations on Md cells were made over the period 2–10 h following iso-prenaline treatment, and included cells in various stages from prophase to metaphase arrest, and also cells released from mitotic block. During prophase, the centriole pairs moved apart from one another and towards the nucleus (Fig. 9 A). Numerous micro-tubules were visible in the vicinity of the centrioles (Fig. 9B). Small dense bodies, approximately 0·5 μm diameter, were present in Md cells. These dense bodies were identified as small secretory granules since they were membrane bound, and had a similar amorphous structure to large secretory granules. The granules did not appear to be surrounding the centriole complexes in prophase cells, but by early prometaphase there was the suggestion that this was occurring (Fig. 10). At the stage of metaphase arrest clustering of secretory granules around the centriolar complexes was pronounced (Figs, 11, 12, 14A).

Small secretory granules were also present in the interphase cells, and their appearance coincided with the time at which the amylase content of the gland begins to rise (Byrt, 1966). In these cells, granules were not clustered around the centrioles.

In the metaphase-arrested cell the centrioles had moved to the cell centre, and were surrounded by chromosomes. In several cells serial sections were cut to verify that all 4 centrioles were at the centre of the cell. Microtubules were observed to radiate out from the vicinity of both centriolar complexes (Fig. 12). Attachment of microtubules to kinetochores was seen (Fig. 13). In addition, evidence of the presence of non-kinetochore microtubules, which radiated out from the cell centre and terminated in the cytoplasm, was obtained in serial sections (Fig. 14).

Occasional sections of both Mf and Md cells revealed the presence of a microtubule which had penetrated a chromosome (Fig. 8). When serial sections containing micro-tubules were examined it was found that they passed completely through the chromosome. Such microtubules fall into the non-kinetochore category, and have been previously observed during metaphase in mammalian cells (Robbins & Gonatas, 1964; Jokelainen, 1967; Brinkley & Stubblefield, 1970). The question of whether these microtubules are of the interpolar type, or play a direct role in chromosome movement, has been discussed by Luykx (1970).

Evidence of recovery of Md cells from metaphase block was found. Cells were seen in which the chromosomes were aligned on the metaphase plate, with the centrioles located at the poles (Fig. 15). Microtubules extended from the vicinity of the centrioles towards the metaphase plate. Small secretory granules had been distributed to both ends of the cell, remaining near the centrioles. Clustering of secretory granules was also observed in telophase cells (Fig. 16).

Blocked metaphase cells were no longer seen 10 h after the injection of isoprenaline. There was an increase in the frequency of secretory granules in all acinar cells compared with earlier times after the drug, and the average size of the granules had increased to 0·75 μm diameter. The secretory granules were distributed peripherally in the mitotic cells and there was no evidence of clustering of the granules near the poles (Fig. 17). Apart from the size and number of secretory granules, these mitotic cells resembled those in glands from control animals.

Interest in the effect of another injection of isoprenaline on these cells was stimulated by the observation that the clustering of secretory granules around the centrioles appeared to be less pronounced in Mf cells than in Md cells. It was thought that this difference might be related to an effect of isoprenaline on the Md cells whilst they were in G₂. To test this, experiments were carried out in which cells were depleted with isoprenaline, allowed to accumulate secretory granules for 10 h, and then subjected to another injection of isoprenaline. The interval between the injections was sufficient to have allowed all cells blocked in G₂ by the initial injection to have completed mitosis. Fifteen minutes after the final injection of isoprenaline cells were again blocked in metaphase (Fig. 18). There was marked clustering of secretory granules around the centrioles, from which it was concluded that the distribution of secretory granules in Md cells was not dependent on an effect of the drug during G₂. Rather, it would appear that the intensity of clustering is related to the size of the granules in the blocked cells.

Cells arrested in metaphase in response to isoprenaline were characterized by displacement of the poles to the cell centre, around which the chromosomes formed an approximately spherical shell. Microtubules radiated out from both pairs of centrioles, demonstrating that the ‘spindle’ system in the arrested cell is bipolar. Kinetochores had attached microtubules, some of which, following favourable sectioning, were seen to extend a considerable way towards the cell centre. Kinetochore microtubules up to several microns long, and terminating near the pole, have been observed during normal metaphase (Jokelainen, 1967; Brinkley & Cartwright, 1971). Examination of serial sections established that non-kinetochore microtubules also were present in metaphase arrested cells. The action of isoprenaline on acinar cells differs from that reported for low doses of colcemid (Brinkley, Stubblefield & Hsu, 1967), or cold (Brinkley & Cartwright, 1970), on Chinese hamster cells, where metaphase arrest is characterized by the retention of kinetochore microtubules alone. Although both kinetochore and non-kinetochore micro-tubules were visible in the isoprenaJine-arrested cells, the possibility remains that the drug might have an influence on the number of micro tubules of either type present. A quantitative study would be necessary to resolve this point.

Metaphase arrest following isoprenaline treatment would thus appear to be related to the failure of the poles to separate, with the concomitant formation of a normal spindle. The mechanism underlying the separation of the centriolar complexes, which occurs during prophase in normal mitosis, has not been established. It has been suggested that the growth and interaction between microtubules which radiate out from the centriolar complexes are responsible for the movement (Mclntosh, Hepler & Van Wie, 1969), such microtubules forming the interpolar spindle. If this is the mechanism involved then it suggests that isoprenaline interferes with the ability of the microtubules from opposite poles to interact with each other, since the generation of microtubules occurred during prophase in isoprenaline-treated glands. The collapse of the established spindle in cells already in metaphase at the time of drug administration would indicate that the bond linking the microtubules is labile. It was not possible to reach a conclusion on whether or not microtubules formed direct links between the 2 poles in the arrest state because of the comparative rarity with which both centriolar complexes were intercepted in the same section in the present study.

The results pose other questions concerning displacement of organelles in the arrested cells. It is evident that force is necessary to move chromosomes and centriolar complexes from their positions at the end of prophase to those seen in the arrested state. Microtubules would appear to be involved in the positioning of the chromosomes since most drugs which elicit metaphase block cause complete dissolution of the spindle (Eigsti & Dustin, 1957; George, Journey & Goldstein, 1965; Krishnan, 1968), and a random arrangement of chromosomes in the cell. When drug treatment results in retention of kinetochore microtubules, the chromosomes are assembled in a spherical array (Brinkley et al. 1967). This suggests that kinetochore microtubules function to tether the chromosomes to the centriolar complexes, either by terminating within a centriolar complex, or by forming a link with non-kinetochore microtubules radiating out from the poles. The movement of the chromosomes could be simply explained if tension existed between kinetochore and pole. There is evidence of a poleward force acting at the kinetochore during normal mitosis, but no satisfactory explanation for its origin has been advanced (Izutzu, 1959). According to the sliding-microtubule hypothesis of Mclntosh et al. (1969) such a force is provided by active mechano-chemical bridges between kinetochore and interpolar microtubules. In the metaphase arrested cell interaction between kinetochore and non-kinetochore microtubules might provide the necessary force. This would seem to be at variance with the suggestion above that isoprenaline interferes with the ability of microtubules to interact with each other. However, in a modified version of the sliding microtubule hypothesis, which better fits the known facts of mitosis, Nicklas (1971) has proposed that the interaction between interpolar microtubules from opposite poles may not involve a bond of the force-generating type.

Another factor which would be expected to play a role in the arrangement of chromosomes in the arrested cell is whether one or both kinetochores are linked to the poles via microtubular links. In the present study chromosomes were observed in which both kinetochores appeared to have microtubular attachments. Extensive serial sectioning would be required to demonstrate whether each kinetochore was linked to its respective pole, or if only one kinetochore was so linked (Brinkley & Stubblefield, 1966, 1970). Possibly both arrangements occur, depending on the orientation of the kinetochores with respect to the poles at the time of nuclear membrane breakdown. It should be mentioned, however, that a situation in which only one kinetochore and pole are linked has been demonstrated to be unstable in normal metaphase; Nicklas & Koch (1969) found from micromanipulation studies that without tension from bipolar pulling, constant reorientation of the unattached kinetochore took place. Another possible misarrangement, the connexion of both kinetochores to the same pole, is an error which has been observed in early prometaphase but is normally soon corrected (Bajer & Mol è-Bajer, 1972).

In addition to movement of chromosomes, force is required to effect the distribution of secretory granules in the arrested cell. Clustering of granules has not been observed in metaphase-arrested cells before, but evidence of poleward displacement of particles, leading eventually to their expulsion from the spindle volume, has been observed in a number of studies (Bajer, 1967; Luykx, 1970; Nicklas & Koch, 1972). The nature of the force responsible for particle movement during mitosis is an open question, and it still remains to be established whether or not the force is the same as that responsible for chromosome movement. Models for mitosis have been postulated in which a common force has been assumed. The results of the present study do not support the molecular pump hypothesis (Ostergren, Molè-Bajer & Bajer, 1960) since it would be expected that countercurrents would be produced which would tend to prevent the clustering of secretory granules in the vicinity of the centriolar complexes. In the sliding-microtubule model of Mclntosh et al. (1969) it is proposed that work performed by microtubule arms or bridges produces particle movement. This model would be compatible provided the poleward-directed forces of the non-kinetochore microtubules on the secretory granules exceed those in the opposite direction exerted by the kinetochore microtubules. The difficulty of the dispersive force exerted by the kinetochore microtubules is avoided in the modified version of the sliding-microtubule model proposed by Nicklas (1971), in which the only force is that exerted by the non-kinetochore microtubules in a poleward direction.

An alternative explanation for the poleward movement of particles which merits attention is that they adhere to microtubules which are continuously growing at the kinetochore and being disassembled at the poles (Nicklas, 1971; Bajer & Mole-Bajer, 1972). The concept that microtubular material was undergoing continuous movement was advanced by Forer (1965) to account for the birefringence patterns observed when the spindle was irradiated with an ultraviolet microbeam. The displacement of secretory granules in response to isoprenaline is compatible with this explanation; kineto-chore microtubule formation occurs at the time of nuclear membrane breakdown (Brinkley & Stubblefield, 1970) and some clustering of secretory granules was observed in early prometaphase, but further investigations of the relationship between granule movement and microtubule growth in isoprenaline-treated cells is necessary. In particular, it would be important to establish from u.v. microbeam irradiation studies whether microtubule formation at the kinetochore, and disassembly at the poles, takes place in cells arrested in metaphase. If microtubule movement does provide the motive force for granule displacement then possibly the appendages seen on the walls of the microtubules may act as the link between the two (Burkholder, Okado & Comings, 1972).

Recently it has been postulated that actin–myosin interactions are responsible for motility of all kinds in higher cells, following the detection of actin-like filaments in spindles (Behnke, Forer & Emmersen, 1971; Forer & Behnke, 1972; Gawadi, 1971). Whatever the nature of the poleward force or forces acting on the chromosomes and granules during normal metaphase, it would appear that they remain operative during metaphase arrest. This factor has to be considered in any general model for mitosis.

The author is grateful to Mr I. Kohlman for his excellent technical assistance, and to Dr G. S. Hodgson and Dr R. M. Moore for their helpful discussions.