Electron Microsc. Rel:., Vol. 5. pp, I 24, 1992. Printed in Great Britain. All rights reserved.
0892-0354~92 $15.00 ~ 1991 Pergamon Press plc,
FORMATION AND ULTRASTRUCTURE S O M A T I C CELL H Y B R I D S A. Y U . K E R K I S
OF
a n d N . S. Z H D A N O V A
Institute o f Cytology and Genetics q[' the Siberian Branch of the Academy of Sciences of the U.S.S.R., 630090, Novosibirsk, U.S.S.R.
Abstract--Taken altogether, the EM evidence we have obtained indicates that the induced (both viral and PEG) and spontaneous (entrance of a splenocyte into a cell) fusion of mammalian somatic cells are associated with alterations in the structure of fusing cells. For example there are alterations in the structure of not only the surfaces of fusing cells but also in the nucleus envelopes and cytoplasmic organelles after PEG treatment. Also, there is long retention of cellular plasma membrane remnants in virally-induced heterokaryons. In short, for each case the alterations were unquestionably specific, in response to the imposed challenge. These specific features not only determine the efficiency and rate of fusion, but also the mode by which the hybrid nucleus is formed. This mode directly determines the fate of the synkaryon and the stability of the so formed hybrid genome. It might be thought that an increase in the inner nuclear envelope observed in some hybrids would counteract the consequences of the disproportion arising between the increase in cell volume and nuclear surface. The finger-like invaginations of the hybrid nuclei nuclear envelope, surrounded by replicatively and transcriptionally active chromatin, appear to be EM demonstrations of such counteracting mechanisms. These invaginations, by augmenting the available inner layer, most likely increase the anchorage sites for chromatin. It is noteworthy that the invaginations occur mainly in multichromosomal hybrids with little chromosome loss. It appears possible that some of the hybrids may contain particular chromosomes from the more differentiated parent cell.
CONTENTS I. II. III. IV. V. VI. VII.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electron microscopic analysis of the process of cell fusion under the effect of UV-inactivated Sendai virus . . . . . . . . . . Electron microscopic analysis of cell fusion induced by the action of polyethylene glycol (PEG) . . . . . . . . . . . . . . . . . . . Formation of hybrid cells during spontaneous fusion between splenocytes and fibroblasts . . . . . . . . . . . . . . . . . . . . . . . . . Ultrastructural features of the nuclei of hybrid cell clones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electron microscopic analysis of hybrid clones containing different chromosome number . . . . . . . . . . . . . . . . . . . . . . . . . . Summary and conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
I. I N T R O D U C T I O N Since the a d v e n t o f v e r s a t i l e m e t h o d s a l l o w i n g us to a c h i e v e the f u s i o n o f cells, the n e w discipline o f s o m a t i c cell genetics h a s b e e n in the m a k i n g . W i t h the aid o f f u s o g e n i c agents, there a p p e a r s to be u n p r e c e d e n t e d possibilities f o r d e r i v i n g h y b r i d cel! lines w h o s e g e n o m e s c o n t a i n g e n e t i c m a t e r i a l c o n t r i b u t e d by t w o o r m o r e p a r e n t a l species, frequently quite remote from each other. These EMR 5/I--A
1
2 7 11
14 18 23 24
a p p r o a c h e s m a k e the c h r o m o s o m a l l o c a t i o n o f genetic d e t e r m i n a n t s p o s s i b l e for m a n y m a m m a l i a n species. T h e h e t e r o k a r y o n s r e s u l t i n g f r o m f u s i o n o f d i f f e r e n t t y p e s o f cells (or species) a n d the h y b r i d c o n s t r u c t s offer real systems for s t u d y i n g the e x p r e s s i o n o f d i s t i n c t genes, p a r t i c u l a r l y t h o s e w h i c h are tissue-specific. N u m e r o u s successful exp e r i m e n t s b e a r w i t n e s s to the v a l i d i t y o f these a p p r o a c h e s . I n d e e d , m u c h has b e e n a t t a i n e d in this g r o w i n g discipline. H o w e v e r , little a t t e n t i o n has
2
A. Yu. Kerkis and N. S. Zhdanova
been given to the ways and means by which the heterokaryons and hybrids are formed. Scant is our knowledge about the morphology of the cytoplasmic and nuclear structures at the early steps of the formation of the hybrid cell. To our knowledge, cell fusion under the effect of fusogenic agents has not been described in accurate morphological terms. Furthermore, the data for virusinduced fusion have been mainly derived from studies of the mechanism of virus-cell interaction and they date back to the 1960s and 1970s. As a result, much remains unclear as to how the hybrid cell may be reconstructed with the aid of fusogenic agents such as Sendal virus or polyethylene glycol (PEG). This information is needed because fusion efficiency, yield of viable hybrids and, most importantly, genome stability of the derived hybrid clones depend upon the degree to which the mechanism by which the cell plasma membranes are altered during cell fusion and hybrid cell formation. Quite plausibly, the segregation of the chromosomes contributed by one of the parent cells, as well as the frequently appearing chromosome rearrangements (according to our summarized 30 40% of the primary hybrid clones contain genetic material of a partner inserted or translocated to another chromosome), are dependent upon the rate and degree to which the structures normally ensuring the spatially ordered organization of genetic material in the nucleus are reestablished. It is clear that this determines the fate of the hybrid cell. Another aspect demanding clarification is how the chromosomes contributed by different genomes may interact. This interaction, termed "chromosome compatibility", underlies the type and degree of expression of definite species and tissue-specific gene determining cell phenotype. We present here a comparative EM analysis of the consecutive steps during the formation of heterokaryons achieved with the aid of the currently used fusogenic agents, inactivated Sendal virus and PEG. In addition, we attempt to establish patterns of hybrid cell formation during spontaneous (without the aid of fusogenic agents) fusion of splenocytes with cultured somatic cell lines when the two cell types were grown together. We also present the results of EM analysis of
several sets of hybrid clones generated by the fusion of fibroblasts of mammals: human and Chinese hamster, mink and human, mink and mouse. We have applied a method we developed for the identification and choice of single cells under the light microscope for their subsequent visualization and analysis under the electron microscope.
II. E L E C T R O N M I C R O S C O P I C ANALYSIS OF T H E P R O C E S S OF CELl. FUSION U N D E R THE EFFECT OF UV-INACTIVATED S E N D A l VIRUS It is well known that orthomixoviruses can elicit the fusion of cells (Harris and Watkins, 1965: Okada and Murayama, 1965). It has been established that these viruses are capable of producing fusion as a result of adsorption to the cell surface and dissolution of the cell membrane with involvement of the viral neuroaminidase (Bratt and Gallaher, 1972). The early course of the process, namely the fusion of cells as a result of the adherence of the viruses to the cell surface, the breakdown of the membrane, followed by lysis in sites where the cells are brought into close viral contact, have been amply described (see reviews, Harris, 1970: Ringertz and Savage, 1976). We are concerned here with the further fate of fusing cells, with thc question of, how a heterokaryon may be formed from two fusing cells'? In fusion experiments aimed at producing heterokaryons and somatic cell hybrids, UV-inactivated Sendal virus is a routinely used fusogenic agent (Harris and Watkins, 1965). This virus i~ advantageous because after UV-exposure it loses infectivity, yet retains fusogenic ability. The cells we utilized in the present experiments were as follows. Chinese hamster fibroblasts, human embryonic and skin fibroblasts (lines kindly provided by the Institute of Medical Genetics, U.S.S.R.L Our choice of cells for fusion experiments was based on the conspicuous morphological differences between long cultured Chinese hamster fibroblasts and human primary fibroblast. The latter are spindle-shaped, they contain irregular nuclei and are several fold larger than the polygonal Chinese
Somatic Cell Hybrids hamster fibroblasts (Fig. l a). The differences between the cell partners were more conspicuous at the ultrastructural level. The human primary fibroblasts contained, as a rule, numerous polysaccharide granules not observed in Chinese hamster fibroblasts. Furthermore, the "total" electron density of the cytoplasm was much higher in human than Chinese hamster fibroblasts (Fig. lb). As
3
shown, there are conspicuous structural differences in the Golgi apparatus, granular endoplasmic reticulum, polysome number and the number and size of the mitochondria. Cells were fused in suspension as described (Harris et al., 1966). Before fusion, Sendai virus was inactivated by UV-light. Heterokaryons were made by addition of 1000 hemagglutinating units
Fig. 1. The morphological appearance of parental cells. (a) Chinese hamster fibroblasts (clone MI5). (b) Human skin primary fibroblasts. Scale markers indicate 1 #m.
4
A. Yu. Kerkis and N. S. Zhdanova
of UV-inactivated Sendai virus. Monolayers of fused hybrid cells growing logarithmically in Petri dishes were cultured in Eagle's minimal essential medium, supplemented with 10% fetal calf serum, in 5% CO 2 humidified air for 6, 12, 24 and 48 hr.
The cells were stained with azure-eosine and then examined with the light microscope. Multinuclear cell counts demonstrated a constant proportion during the observation period, amounting to about 18% of the total number of cells.
Fig. 2a to c.
Fig. 2d to g. Fig. 2(a-g). The consecutive steps of the fusion Chinese hamster fibroblasts (C) and human fibroblasts (H) induced by Sendai virus. (For explanations see text.) Arrows point to the sites where the plasma membranes have fused. At these sites the cytoplasm of the parental type cells mix. Scale markers indicate 1 ,am.
6
A. Yu. Kerkis and N. S. Zhdanova
EM studies of cells during the fusion process, as well as of the hybrid cells, were carried out using the method we developed for this purpose. It has been designated as the "flat parallel embedding method" (Christolubova and Kerkis, 1972). After identification of definite cells and marking under the light microscope, chosen cells were examined under the electron microscope. Briefly outlined, the procedures were as follows. The cells were cultured in Petri dishes with carbon-coated flat bottoms. The cells were fixed and embedded in Araldite directly in the dishes by changes of solution. Glutaraldehyde fixed, osmium postfixed monolayers of hybrid cells were dehydrated through a graded series of ethanols. After dehydration and presoaking with Araldite, the final polymerization mixture was poured into each dish so that an even 2 3 mm covered its bottom. Thereafter the bottom of each dish was covered with Araldite for polymerization. The dishes were left horizontally in a thermostat at 60" for 24 hr. The preparations were removed from the dishes using dry ice or liquid nitrogen. The preparations thus obtained were Araldite discs comprising embedded cells amenable to observation by light microscopy at any magnification. For autoradiography the preparations were coated with emulsion (Ilford type L-4) and exposed for 5 7 days. In the case of light microscopic analysis of the preparations, cells containing several nuclei, of which at least one was 3H-thymidine labelled, were considered to be heterokaryons, whereas cells, which were morphologically fibroblasts and whose nuclei (one or several) were ~H-thymidine labelled, were considered to be hybrid. By means of a device fixed on the objective of the light microscope, the cells to be electronmicroscopically analyzed further were marked. Based on the results of the present EM analysis of cells undergoing fusion, the process may be conceived as consisting of the following steps. Step 1: Agglutination; the adherence of cell plasma membranes by the interaction of attached viral particles (Fig. 2a and b). Step 2: Destruction: the plasma membrane starts to lose its structure at sites where cells come into contact (Fig. 2c and d). Step 3: Disintegration; as a result of a breakdown of the membrane, cytoplasmic bridges connect the parental cells (Fig. 2e), the number of bridges varies
largely, being obviously dependent on the number of adsorbed viral particles (Okada, 1962: Okada et al., 1975). Step 4: Formation of larger cytoplasmic junctions; the membranes at the contacl sites now assume the appearance of structureless electron dense bands (Fig. 2f). Step 5: Disappearance of the remnants of the cytoplasmic membranes at some contact sites and their retention at others: this seems to be the initiation of the formation of hybrid cytoplasm and the true heterokaryon (Fig. 2g). Heterokaryons at each of the above steps were observed at all time points, up to 48 hr after treatment of the cells with inactivated virus. At the earlier fusion steps, the proportion of heterokaryons was low and, accordingly, the number of heterokaryons increased with incubation time. Examination of serial ultra thin sections demonstrated that 2 days after viral treatment 80% of the heterokaryons chosen under the light microscope were clearly heterokaryons at the ultrastructural level; the rest were represented by cells proceeding through the earlier steps of fusion, including the first. According to Rechsteiner and Parson (1976), synkaryons start to form by the end of the first day after treatment of cells with Sendal virus. We observed hybrid cells (cells morphologically similar to human fibroblasts and containing ~H-thymidine labelled nucleus) 2 4 h r after hybridization and, consequently, it may be assumed that the process of synkaryon formation in the system of Rcchsteiner and Parson and in ours is similar. It thus appears that, 24 hr after viral treatment o[" the cells, the cultures contain synkaryons, as well as heterokaryons passing through the suggested steps of the fusion process, i.e. the transformation rate of the aggregated cells into a heterokaryon apparentlx depends upon how close the cells come into contact and/or the number of adsorbed virsues. There is a feature of the virus-induced heterokaryons to be noted. Although outwardly similar under the light and electron microscopes, these heterokaryons in fact represent a wide variety of transitional forms from closely adhering cells bridged cytoplasmically to true heterokaryons with hybrid cytoplasm. These remnants of the cell membranes, as it may seem, hinder the fusion of nuclei
Somatic Cell Hybrids and mixing of parental cytoplasm, particularly of structural components and membrane connected factors. This may be one of the reasons why multinuclear cells are occasionally retained, up to 9 days, in cultures (Velasquez, 1971). Other evidence indicating that mixing of parental cytoplasm is very slow comes from the studies of Siniscalco et al. (1969) on the location of the activity of glucose6-phosphate dehydrogenase around one of the nuclei in 2-day-old heterokaryons resulting from the fusion of human fibroblasts with normal and zero enzyme activity.
III. E L E C T R O N M I C R O S C O P I C ANALYSIS OF CELL FUSION INDUCED BY T H E ACTION OF P O L Y E T H Y L E N E G L Y C O L (PEG) The use of inactivated Sendai virus for the fusion of cells and further selection of hybrids has a number of limitations. A major condition must be met, the cell surface must have receptors for the adsorption of the virus. For this reason, many kinds of cells, particularly of plant origin, cannot be fused; the technical difficulties imposed by an, albeit inactivated yet infectious virus, are too great. Starting in the 1970s there was an increasing search for chemical and physical agents with fusogenic action (Poole and Nowell, 1970. Ahkong et al., 1972). PEG proved to be a promising chemical on account of its high fusogenic efficiency and relatively low toxicity. PEG has gained wide acceptance because it has made feasible the production of hybrid forms in plants (Pontecorvo, 1975). We studied the morphology of cells and their fusion patterns in cultured Chinese hamster and human fibroblasts (embryonic and skin) for 2 days after treatment with 50% PEG. Fusion was performed in a suspension of cells collected after centrifugation of a mixture of 5 x 106 of both cell types. Chinese hamster cells were first grown in an 3H-thymidine medium, as in the experiments with virus-induced fusion. The genetic material derived from Chinese hamster was, therefore, labelled in the heterokaryons and hybrids.
7
In analysis of the cell suspension promptly after washing away the PEG (15rain after its application), we observed 16% of the cell population to be multinuclear cells which, judging by our criteria, were referred to heterokaryons. Surveying the electron micrograms of these cells, it is clear that the fused plasma membranes have become structure~ less bands at many sites so that the cytoplasm of the clumped cells is hardly, if ever, disconnected by membranes; yet the cytoplasmic structures of the parental cells have not, so far, mixed together (Fig. 3a). There were about 10% multinuclear cells in the culture 20-24 hr after treatment with PEG; rare stray ones were heterokaryons. The electron micrograph shown in Fig. 3b demonstrates a heterokaryon with four centrioles. Uni- and multinuclear synkaryons were observed 48 hr after treatment of the cells with PEG, but no heterokaryons appeared at that point in time. The majority of heterokaryons had transformed to synkaryons by 20 hr after treatment with PEG. This indicates the significant difference in the time course of formation of hetero- and synkaryons, after cellular treatment with PEG or Sendai virus. With the aid of PEG, fusion is achieved faster and more synchronously. In the case of PEG, cell fusion is effected through the direct action of the fasogen on the membranes. PEG is hydrophobic and surface-acting, and it also binds calcium ions. This property provides the structural modification of the plasma membrane, the redistribution of the electric charge over the membrane surface, the resulting fusion and ultimate lysis of the cell membrane (Kao and Michaljluk, 1974). If the structure of the non-fused plasma membrane cannot be restored, the cells die. Clearly, at minimum toxicity of PEG, cells fuse rapidly, and soon thereafter the entire agglutinated plasma membrane surfaces breakdown. Observations of the ultrastructural changes produced by PEG revealed that they concerned not only the cell plasma membrane; changes occur also in the nuclear-nuclear (Fig. 3c), nuclear-mitochondrial (Fig. 3d, e, f), mitochondrial-mitochondrial (Fig. 3g) membranes and lipid inclusions. Fragments of nuclear membranes are very frequently seen in the nuclei. Examination of serial
,~
A. Yu. Kerkis and N. S. Zhdanova
(a)
Fig. 3a to c. ultrathin sections assured us that they are, indeed, fragments and not invaginations of the nucleus. We did not observe fragments of this kind in parental cells, heterokaryons and synkaryons produced by virus-induced fusion. In contrast, these
fragments were present in almost all the hybrid nuclei of cells fused by PEG. A plausible mechanism for the origin of these fragments in the hybrid nuclei is offered. After penetrating into the cells. PEG may exert its action on the nuclear
Somatic Cell Hybrids
f)
9
I
~:•
Fig, 3d to h.
membrane, as it does on the plasma membrane, and, as a result, heterokaryon nuclei fuse not only during mitosis, when the nuclear membrane disappears [this is the case when hybrid cell formation is virus-induced (Harris et al., 1966; Reichsteiner
and Parson, 1976)], but also during interphase. Suggestive evidence for the possible direct fusion of nuclei formed after PEG treatment may be the numerous multinuclear cells we observed with hybrid nuclei. No such heterokaryons were
A. Yu. Kerkis and N. S. Zhdanova
Fig. 3i and j. Fig. 3. The stepwise formation of a hybrid cell under the influence of polyethylene glycol (PEG). tt H u m a n fibroblast: C hamster fibroblast. Scale markers indicate 1/~m. (a) The cells start to fuse 15 min after PEG has been washed off. Arrows point to the region of disintegrating plasma membrane. (b) A formed heterokaryon 4 hr after treatment with PEG. Arrows point to the double set of centrioles. (c) Arrows pinpoint tile sites where the nuclei of a heterokaryon fuse during the formation of a hybrid nucleus. 2 6 hr after exposure to PEG. (d f) Possible fusion of mitochondria with the nucleus in a hybrid cell, resulting from treatment with PEG. N--nucleus: M mitochondrion, Arrows point to the site of the disintegrating nuclear envelope. (g) Fusion of mitochondrial membranes (M). (h) Fusion of lipid inclusions (L) in a hybrid cell under the effect of PEG. (i j) Fragmentation of nuclei and nuclear envelope during the formation of a hybrid cell. Arrows point to the site where the nuclear membrane is destroyed and fragmented.
Somatic Cell Hybrids found in the case of Sendai virus-induced fusion of cells. Figures 3i and j demonstrates another effect of PEG, the fragmentation of nuclei in the hybrid cells. In some instances, the fragmentation of the hybrid nucleus was complete, and in others small fragments break away from the nucleus. The virusinduced synkaryons showed no such patterns. Thus, rapidity and synchronicity are the features of PEG-induced cell fusion. Furthermore, according to our present observations, cell treatment with PEG produces definite changes not only in the membranes of the cell itself, but also in the nucleus and cytoplasm. As a consequence, direct fusion of cells becomes feasible at any stage of the cell cycle: it ceases to be restricted to mitosis. Reasoning still further, it appears that the effects of preliminary condensation, pulverization of chromosomes (the known concomitants of the fusion of cells at different stages of their cycle), among others, on chromosome segregation, as well as the appearance frequency of chromosome rearrangements, are exaggerated in the PEG-induced compared to the virus-induced hybrids. The fragmentation of nuclei definitely contributes to chromosome elimination. When comparing the two fusogenic agents, it becomes more evident that PEG has the merit of simplicity, whereas Sendai virus spares the integrity of genetic material in the hybrid cells.
IV. F O R M A T I O N OF H Y B R I D CELLS DURING S P O N T A N E O U S FUSION B E T W E E N S P L E N O C Y T E S AND FIBROBLASTS
When the formation of hybrid cells is included by Sendai virus or PEG, the plasma membrane and in the case of PEG the majority of intracellular membranes also lose their structural integrity. The cytotoxic effects of these fusogenic agents may be attributed to their destructive action on the membranes, which was stronger in the case of PEG. This action may be one of the causes of the numerous chromosome rearrangements. These rearrangements make it difficult to locate chromosomal genes in somatic cell hybrids and occasionally
11
renders impossible any analysis of the early steps of hybrid cell formation. With this in mind, we attempted to generate hybrid cells in a way that is not traditional, by taking advantage of the ability of lymphocytes to rapidly penetrate into target cells. In the relevant experiments Chinese hamster fibroblasts were cocultivated with fox or mouse splenocytes. Fox and mouse spleens were isolated under sterile conditions, were sliced, washed and dissolved in a solution ( D M E M with supplements, 10% fetal calf serum and 250 ng/ml PHA) so that there were 10 splenocytes for each fibroblast. Thereafter a splenocyte suspension was transferred to a Petri dish containing growing Chinese hamster fibroblasts. The results were best when fibroblasts were grown to semiconfluent monolayers. This was followed by fixation of cells after 1, 4, 8, 24, 48 and 72 hr. The procedures for the fixation, embedding, and choice of cells for EM analyses were standard, as described above (Section II). Splenocytes are able to freely enter the cytoplasm of the fibroblast, while the nucleus, being protected from the attack of lysosomal enzymes, is spared and hence it shows no morphological changes up to 72 hr. Some of the splenocyte nuclei are able to enter the fibroblast nuclei at 1 - 4 h r (Fig. 4a, b,c and d). After 24-48 hr blocks of condensed chromatin from the splenocyte nuclei are seen in the nuclei of the fibroblasts (Fig. 4e, f and g). The amount of condensed chromatin in the fused nuclei decreases in the process of cocultivation presumably as a result of chromatin activation. Based on the data obtained, the sequence of events providing the formation of somatic cell hybrids may be envisaged as follows (1) The synkaryon would be formed as a result of the rapid passage of the splenocyte into the nucleus of the fibroblast; as a consequence, its plasma and nuclear membranes would be destroyed. (2) The chromatin of the splenocyte nucleus, residing in the fibroblast cytoplasm for 15-20 hr of cocultivation, would fall apart into chromosome-like blocks. Each block is seen outlined by a double membrane presumably formed from the destroyed membrane of the splenocyte. If this is the case, the synkaryon
12
A. Yu. Kerkis and N. S. Zhdanova
lib
t
(d)~ ~k 3
~t ¸~ \ .
I
I
Fig. 4a to d.
would be formed at that time point of mitosis, when separate membrane-outlined fragments may be integrated into the assembling fibroblast nucleus. As in the condensed chromatin of the splenocyte fragments would be rather well identified up
to 72 hr (Fig. 4h). At this time point, the membrane fragments attached to the remains of condensed chromatin are still observed. The data obtained demonstrate the lklcility ol high frequency production of somatic cell hybrids
Somatic Cell Hybrids
Fig. 4e to h. Fig. 4. Formation o f a hybrid cell during the cocultivation o f splenocytes (Sp) with fibroblasts (F). Scale markers indicate I/~ m. (a) Early fusion 1 hr after cocultivation. (b, c) Steps of the penetration of a splenocyte into fibroblast cytoplasm 4-5 hr after cocultivation. Arrows point to the sites of fusing membranes. (d) Fragmentation of a splenocyte nucleus in the cytoplasm of a fibroblast 18-24 hr after cocultivation. (e-h) Distribution of chromatin blocks of splenocytes (arrows) in the nuclei of hybrid cells 48 hr (e, f), 6 hr (g) and 72 hr after cocultivation (h).
13
14
A. Yu. Kerkis and N. S. Zhdanowl
by cocultivation of fibroblasts with splenocytes, w i t h o u t the aid of fusogenic agents. It is possible that this approach should allow to generate more somatic cell hybrids without chromosome rearrangements, thereby facilitating investigation of the patterns of hybrid cell formation.
V. U L T R A S T R U C T U R A L F E A T U R E S OF THE NUCLEI OF H Y B R I D CELL C L O N E S Genetic material of species frequently differing in number, size and other chromosomal features is internalized within the hybrid nucleus. Ordered spatial distribution of all the chromosomes within such a tetraploid nucleus is a requisite for its function. Consequently, the structure of the nuclear matrix is of importance for the formation of the hybrid nucleus. There is now convincing evidence indicating that discrete elements of the nuclear matrix are involved in the determination of the functionally ordered internal structure of the nucleus. Thus, it has been shown that the chromatin fibrils are bound to the nuclear matrix, thereby ensuring the higher order of chromatin organization (Cooch and Brazells, 1975), that transcription is initiated in the binding sites (Comings and Okada, 1975; Berezney, 1979) and that it is here that the actin genes are, as a rule, located (Small et al., 1985), thereby providing evidence for the important role of this structure in the organization of active chromatin (Lafond and Woodcock, 1983). Elements of the nuclear matrix accomplish important regulatory functions, being involved in D N A replication and intranucleus transport of R N A (Zbarsky, 1985). These elements seem to promote the regular passage of mitotic chromosomes to the opposite cell poles, and they can turn to scaffold-like structures during mitosis (Bekers et al., 1981). As a whole, the nuclear matrix may be regarded as a flexible structure capable of reducing in a reversible way so that the volume and shape of the nucleus can change during interphase (Berezney and Coffey, 1977). The nuclear lamina, one of the structurally fundamental units, was found to consist of three protein components which are different from the
other proteins of the nuclear matrix. The nuclear lamina structurally interacts with the inner nuclear membrane and nuclear pore complexes, on the one hand, and chromatin elements on the other (Gerace and Blobel, 1980). During mitosis, the lamina can reversibly pass to the monomeric state with one, at least, of the constituent proteins remaining bound to the membrane fragments o1" the disrupted nuclear membrane, It is known that the fragments of the nuclear membrane remain associated with single chromsomes during mitosis (Kerkis and Christolubova, I191; Zatsepina et al.. 1976) and that the new nuclear membrane starts to assemble from these bound structures during telophase. It appears that the specificity and number of attachment sites of the chromosomes to the lamina and, therefore, the spatial distribution of the chromosomes in the interphase nucleus are all determined through the mediation of these bound structures. When taking this body of evidence together, it may be assumed, with good reason, that genomic stability, i.e. the segregation rate of chromosomes in a hybrid cell, is dependent upon the structure o1" the nuclear lamina. We have carried out EM studies of the nuclei from different hybrid clone cells with reference to the structures associatcd with the nuclear matrix. Chinese h a m s t e ~ h u m a n somatic cell hybrids were generated by fusion induced between Chinese hamster fibroblasts (clone M I5, H P R T ) and human embryonic fibroblasts at 7 10 passages. Hybridization was performed as described above with the aid of UV-inactivated Sendal virus. 10 ' ~1 ouabain was added to the standard selective medium H A T and killed human fibroblasts. Five clones of this type containing a tetraploid set of Chinese hamster chromosomes and from 11 20 human chromosomes were taken for EM analysis. We also studied five mouse hepatoma cells (clone BW8)-American mink fibroblast ( M V T K ) hy~ brids. These hybrid clones contained a tetraploid set of mouse chromosomes and from 2 17 chromosomes contributed by American mink. These clones were obtained using PEG, as described above (Zhdanova et al., 1985). For EM analysis, the cells were frst chosen under the light microscope, and then serially
Somatic Cell H y b r i d s
sectioned with the use of our methods. Replication of the hybrid cell nuclei was studied by the modified method of electron autoradiography (Kerkis and Khristolubova, 1974). For this pur-
~, U I
15
pose, 18.5~tc/ml 3H-thymidine (18C/mM; Amersham) was added to the culture medium 30 min before the beginning of fixation. Ultrathin sections were covered with emulsion of Ilford L4 type.
i
Fig. 5a to d.
1()
A. Yu. Kerkis and N. S. Zhdanova
Fig. 5e to g. Fig. 5. Ultrastructural fragmentation of the nuclei of hybrid clones. Scale markers indicate l #m. (a c) Tubular finger-like invaginations (arrows) of internal nuclear envelope in longitudinal (a, c) and transversal sections {b). N--Nucleus, C---cytoplasm, Ch-q:hromatin. (d) Electron microscopic autoradiograph of 3H-thymidine-labelled chromatin surrounding the tubular structures (arrows). (e) Specific histochemical staining for transcriptionally active chromatin surrounding the tubular structure (arrows) according to Frenster's reaction (for explanation see text). N--Nucleus, C-~:ytoplasm. (f) Tubular structures in the nuclei of hybrid cells formed as a result of combined treatment with Triton X-100 and DNase. Arrows point to the longitudinal sections, and triangles to the transversal sections of pore complex. (g) Blocks of condensed chromatin in the nuclei of some hybrid clones.
Somatic Cell H y b r i d s
Transcriptional activity of the nuclei was analyzed by means of Frenster's method (Frenster, 1971; Losecke et al., 1980). In studies of the structure of the nuclear matrix composition cells were treated before fixation with 1/~g/ml DNase I (Serva) for 1.5 hr, then 0 . 2 ~ Triton X-100 for 20 min, or Triton X-100 together with D N a s e I. The electron micrograph in Fig. 5a shows deep invaginations of the inner layer of nuclear membrane of hybrids, thereby greatly increasing its nuclear surface area. This makes the appearance of the hybrid cell strikingly different from the parental cells. O f interest are the structures within the hybrid nuclei, which have not been previously described. In cross sections, they appear as numerous round cavities with a monolayer membrane and an internal diameter of about 40 nm (Fig. 5b). As many as 70% of all the cells had this structure in some clones. In longitudinal sections, they appeared as tubular cavities surrounded by chromatin. The structures traversed the nucleus in all the directions, and they were well defined far from the nuclear envelope (Fig. 5c). Examination of serial sections demonstrated that these structures are finger-like invaginations of inner nuclear membrane surrounded by closely adhering loose chromatin. When 3H-thymidine was added 30 min before fixation, the silver grains were mainly distributed over chromatin associated with these tubular structures. This was firm evidence for the replicative activity in these chromatin regions. Figure 5e shows that the electron-dense products of the D N a s e - A O reaction occur, as a rule, in the nucleolus and chromatin surrounding the tubular
17
cavities. Thus, we obtained experimental evidence indicating that these chromatin regions are also active in transcriptional terms. Treatment of the cultured cells with DNase I solution promptly before fixation removes the bulk of chromatin from the nucleus and, as a result, branched tubular cavities are clearly seen in the nucleus. In the case when the cells are, in addition to DNase I, treated with Triton X-100 (i.e. subjected to a procedure routinely used for isolation of the nuclear matrix), numerous elements of the intracellular matrix associated with the tubular cavities containing the characteristic components of the pore complexes and lamina are seen in the section (Fig. 5F). Thus, our observations indicate that the numerous nuclear structures formed by the finger-like invaginations of the inner nuclear membrane are closely associated with the elements of the intracellular matrix bound to the transcriptionally active chromatin. The chromatin associated with these structures is probably that part of the laminabound chromatin which determines the location of the chromsomes in the hybrid cell, both on the nuclear membrane and, in a broader sense, the spatial arrangement of chromatin in the nucleus. Interestingly, among clones studied there occurred some containing the above described structures and others with cells having nuclei containing large globules of condensed electron-dense chromatin, occasionally occupying about 35% of a section area (Fig. 5g). According to the measurement data, the amount of condensed chromatin is not related to the number of human chromosomes present in the hybrid nucleus of a clone, nor is it associated with any appreciable increase in the total area of a
Table 1. Relative n u m b e r of c o n d e n s e d c h r o m a t i n in the cell nuclei of Chinese h a m s t e r x h u m a n fibroblast hybrids
Clone 3S 2 3S
0892-0354~92 $15.00 ~ 1991 Pergamon Press plc,
FORMATION AND ULTRASTRUCTURE S O M A T I C CELL H Y B R I D S A. Y U . K E R K I S
OF
a n d N . S. Z H D A N O V A
Institute o f Cytology and Genetics q[' the Siberian Branch of the Academy of Sciences of the U.S.S.R., 630090, Novosibirsk, U.S.S.R.
Abstract--Taken altogether, the EM evidence we have obtained indicates that the induced (both viral and PEG) and spontaneous (entrance of a splenocyte into a cell) fusion of mammalian somatic cells are associated with alterations in the structure of fusing cells. For example there are alterations in the structure of not only the surfaces of fusing cells but also in the nucleus envelopes and cytoplasmic organelles after PEG treatment. Also, there is long retention of cellular plasma membrane remnants in virally-induced heterokaryons. In short, for each case the alterations were unquestionably specific, in response to the imposed challenge. These specific features not only determine the efficiency and rate of fusion, but also the mode by which the hybrid nucleus is formed. This mode directly determines the fate of the synkaryon and the stability of the so formed hybrid genome. It might be thought that an increase in the inner nuclear envelope observed in some hybrids would counteract the consequences of the disproportion arising between the increase in cell volume and nuclear surface. The finger-like invaginations of the hybrid nuclei nuclear envelope, surrounded by replicatively and transcriptionally active chromatin, appear to be EM demonstrations of such counteracting mechanisms. These invaginations, by augmenting the available inner layer, most likely increase the anchorage sites for chromatin. It is noteworthy that the invaginations occur mainly in multichromosomal hybrids with little chromosome loss. It appears possible that some of the hybrids may contain particular chromosomes from the more differentiated parent cell.
CONTENTS I. II. III. IV. V. VI. VII.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electron microscopic analysis of the process of cell fusion under the effect of UV-inactivated Sendai virus . . . . . . . . . . Electron microscopic analysis of cell fusion induced by the action of polyethylene glycol (PEG) . . . . . . . . . . . . . . . . . . . Formation of hybrid cells during spontaneous fusion between splenocytes and fibroblasts . . . . . . . . . . . . . . . . . . . . . . . . . Ultrastructural features of the nuclei of hybrid cell clones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Electron microscopic analysis of hybrid clones containing different chromosome number . . . . . . . . . . . . . . . . . . . . . . . . . . Summary and conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
I. I N T R O D U C T I O N Since the a d v e n t o f v e r s a t i l e m e t h o d s a l l o w i n g us to a c h i e v e the f u s i o n o f cells, the n e w discipline o f s o m a t i c cell genetics h a s b e e n in the m a k i n g . W i t h the aid o f f u s o g e n i c agents, there a p p e a r s to be u n p r e c e d e n t e d possibilities f o r d e r i v i n g h y b r i d cel! lines w h o s e g e n o m e s c o n t a i n g e n e t i c m a t e r i a l c o n t r i b u t e d by t w o o r m o r e p a r e n t a l species, frequently quite remote from each other. These EMR 5/I--A
1
2 7 11
14 18 23 24
a p p r o a c h e s m a k e the c h r o m o s o m a l l o c a t i o n o f genetic d e t e r m i n a n t s p o s s i b l e for m a n y m a m m a l i a n species. T h e h e t e r o k a r y o n s r e s u l t i n g f r o m f u s i o n o f d i f f e r e n t t y p e s o f cells (or species) a n d the h y b r i d c o n s t r u c t s offer real systems for s t u d y i n g the e x p r e s s i o n o f d i s t i n c t genes, p a r t i c u l a r l y t h o s e w h i c h are tissue-specific. N u m e r o u s successful exp e r i m e n t s b e a r w i t n e s s to the v a l i d i t y o f these a p p r o a c h e s . I n d e e d , m u c h has b e e n a t t a i n e d in this g r o w i n g discipline. H o w e v e r , little a t t e n t i o n has
2
A. Yu. Kerkis and N. S. Zhdanova
been given to the ways and means by which the heterokaryons and hybrids are formed. Scant is our knowledge about the morphology of the cytoplasmic and nuclear structures at the early steps of the formation of the hybrid cell. To our knowledge, cell fusion under the effect of fusogenic agents has not been described in accurate morphological terms. Furthermore, the data for virusinduced fusion have been mainly derived from studies of the mechanism of virus-cell interaction and they date back to the 1960s and 1970s. As a result, much remains unclear as to how the hybrid cell may be reconstructed with the aid of fusogenic agents such as Sendal virus or polyethylene glycol (PEG). This information is needed because fusion efficiency, yield of viable hybrids and, most importantly, genome stability of the derived hybrid clones depend upon the degree to which the mechanism by which the cell plasma membranes are altered during cell fusion and hybrid cell formation. Quite plausibly, the segregation of the chromosomes contributed by one of the parent cells, as well as the frequently appearing chromosome rearrangements (according to our summarized 30 40% of the primary hybrid clones contain genetic material of a partner inserted or translocated to another chromosome), are dependent upon the rate and degree to which the structures normally ensuring the spatially ordered organization of genetic material in the nucleus are reestablished. It is clear that this determines the fate of the hybrid cell. Another aspect demanding clarification is how the chromosomes contributed by different genomes may interact. This interaction, termed "chromosome compatibility", underlies the type and degree of expression of definite species and tissue-specific gene determining cell phenotype. We present here a comparative EM analysis of the consecutive steps during the formation of heterokaryons achieved with the aid of the currently used fusogenic agents, inactivated Sendal virus and PEG. In addition, we attempt to establish patterns of hybrid cell formation during spontaneous (without the aid of fusogenic agents) fusion of splenocytes with cultured somatic cell lines when the two cell types were grown together. We also present the results of EM analysis of
several sets of hybrid clones generated by the fusion of fibroblasts of mammals: human and Chinese hamster, mink and human, mink and mouse. We have applied a method we developed for the identification and choice of single cells under the light microscope for their subsequent visualization and analysis under the electron microscope.
II. E L E C T R O N M I C R O S C O P I C ANALYSIS OF T H E P R O C E S S OF CELl. FUSION U N D E R THE EFFECT OF UV-INACTIVATED S E N D A l VIRUS It is well known that orthomixoviruses can elicit the fusion of cells (Harris and Watkins, 1965: Okada and Murayama, 1965). It has been established that these viruses are capable of producing fusion as a result of adsorption to the cell surface and dissolution of the cell membrane with involvement of the viral neuroaminidase (Bratt and Gallaher, 1972). The early course of the process, namely the fusion of cells as a result of the adherence of the viruses to the cell surface, the breakdown of the membrane, followed by lysis in sites where the cells are brought into close viral contact, have been amply described (see reviews, Harris, 1970: Ringertz and Savage, 1976). We are concerned here with the further fate of fusing cells, with thc question of, how a heterokaryon may be formed from two fusing cells'? In fusion experiments aimed at producing heterokaryons and somatic cell hybrids, UV-inactivated Sendal virus is a routinely used fusogenic agent (Harris and Watkins, 1965). This virus i~ advantageous because after UV-exposure it loses infectivity, yet retains fusogenic ability. The cells we utilized in the present experiments were as follows. Chinese hamster fibroblasts, human embryonic and skin fibroblasts (lines kindly provided by the Institute of Medical Genetics, U.S.S.R.L Our choice of cells for fusion experiments was based on the conspicuous morphological differences between long cultured Chinese hamster fibroblasts and human primary fibroblast. The latter are spindle-shaped, they contain irregular nuclei and are several fold larger than the polygonal Chinese
Somatic Cell Hybrids hamster fibroblasts (Fig. l a). The differences between the cell partners were more conspicuous at the ultrastructural level. The human primary fibroblasts contained, as a rule, numerous polysaccharide granules not observed in Chinese hamster fibroblasts. Furthermore, the "total" electron density of the cytoplasm was much higher in human than Chinese hamster fibroblasts (Fig. lb). As
3
shown, there are conspicuous structural differences in the Golgi apparatus, granular endoplasmic reticulum, polysome number and the number and size of the mitochondria. Cells were fused in suspension as described (Harris et al., 1966). Before fusion, Sendai virus was inactivated by UV-light. Heterokaryons were made by addition of 1000 hemagglutinating units
Fig. 1. The morphological appearance of parental cells. (a) Chinese hamster fibroblasts (clone MI5). (b) Human skin primary fibroblasts. Scale markers indicate 1 #m.
4
A. Yu. Kerkis and N. S. Zhdanova
of UV-inactivated Sendai virus. Monolayers of fused hybrid cells growing logarithmically in Petri dishes were cultured in Eagle's minimal essential medium, supplemented with 10% fetal calf serum, in 5% CO 2 humidified air for 6, 12, 24 and 48 hr.
The cells were stained with azure-eosine and then examined with the light microscope. Multinuclear cell counts demonstrated a constant proportion during the observation period, amounting to about 18% of the total number of cells.
Fig. 2a to c.
Fig. 2d to g. Fig. 2(a-g). The consecutive steps of the fusion Chinese hamster fibroblasts (C) and human fibroblasts (H) induced by Sendai virus. (For explanations see text.) Arrows point to the sites where the plasma membranes have fused. At these sites the cytoplasm of the parental type cells mix. Scale markers indicate 1 ,am.
6
A. Yu. Kerkis and N. S. Zhdanova
EM studies of cells during the fusion process, as well as of the hybrid cells, were carried out using the method we developed for this purpose. It has been designated as the "flat parallel embedding method" (Christolubova and Kerkis, 1972). After identification of definite cells and marking under the light microscope, chosen cells were examined under the electron microscope. Briefly outlined, the procedures were as follows. The cells were cultured in Petri dishes with carbon-coated flat bottoms. The cells were fixed and embedded in Araldite directly in the dishes by changes of solution. Glutaraldehyde fixed, osmium postfixed monolayers of hybrid cells were dehydrated through a graded series of ethanols. After dehydration and presoaking with Araldite, the final polymerization mixture was poured into each dish so that an even 2 3 mm covered its bottom. Thereafter the bottom of each dish was covered with Araldite for polymerization. The dishes were left horizontally in a thermostat at 60" for 24 hr. The preparations were removed from the dishes using dry ice or liquid nitrogen. The preparations thus obtained were Araldite discs comprising embedded cells amenable to observation by light microscopy at any magnification. For autoradiography the preparations were coated with emulsion (Ilford type L-4) and exposed for 5 7 days. In the case of light microscopic analysis of the preparations, cells containing several nuclei, of which at least one was 3H-thymidine labelled, were considered to be heterokaryons, whereas cells, which were morphologically fibroblasts and whose nuclei (one or several) were ~H-thymidine labelled, were considered to be hybrid. By means of a device fixed on the objective of the light microscope, the cells to be electronmicroscopically analyzed further were marked. Based on the results of the present EM analysis of cells undergoing fusion, the process may be conceived as consisting of the following steps. Step 1: Agglutination; the adherence of cell plasma membranes by the interaction of attached viral particles (Fig. 2a and b). Step 2: Destruction: the plasma membrane starts to lose its structure at sites where cells come into contact (Fig. 2c and d). Step 3: Disintegration; as a result of a breakdown of the membrane, cytoplasmic bridges connect the parental cells (Fig. 2e), the number of bridges varies
largely, being obviously dependent on the number of adsorbed viral particles (Okada, 1962: Okada et al., 1975). Step 4: Formation of larger cytoplasmic junctions; the membranes at the contacl sites now assume the appearance of structureless electron dense bands (Fig. 2f). Step 5: Disappearance of the remnants of the cytoplasmic membranes at some contact sites and their retention at others: this seems to be the initiation of the formation of hybrid cytoplasm and the true heterokaryon (Fig. 2g). Heterokaryons at each of the above steps were observed at all time points, up to 48 hr after treatment of the cells with inactivated virus. At the earlier fusion steps, the proportion of heterokaryons was low and, accordingly, the number of heterokaryons increased with incubation time. Examination of serial ultra thin sections demonstrated that 2 days after viral treatment 80% of the heterokaryons chosen under the light microscope were clearly heterokaryons at the ultrastructural level; the rest were represented by cells proceeding through the earlier steps of fusion, including the first. According to Rechsteiner and Parson (1976), synkaryons start to form by the end of the first day after treatment of cells with Sendal virus. We observed hybrid cells (cells morphologically similar to human fibroblasts and containing ~H-thymidine labelled nucleus) 2 4 h r after hybridization and, consequently, it may be assumed that the process of synkaryon formation in the system of Rcchsteiner and Parson and in ours is similar. It thus appears that, 24 hr after viral treatment o[" the cells, the cultures contain synkaryons, as well as heterokaryons passing through the suggested steps of the fusion process, i.e. the transformation rate of the aggregated cells into a heterokaryon apparentlx depends upon how close the cells come into contact and/or the number of adsorbed virsues. There is a feature of the virus-induced heterokaryons to be noted. Although outwardly similar under the light and electron microscopes, these heterokaryons in fact represent a wide variety of transitional forms from closely adhering cells bridged cytoplasmically to true heterokaryons with hybrid cytoplasm. These remnants of the cell membranes, as it may seem, hinder the fusion of nuclei
Somatic Cell Hybrids and mixing of parental cytoplasm, particularly of structural components and membrane connected factors. This may be one of the reasons why multinuclear cells are occasionally retained, up to 9 days, in cultures (Velasquez, 1971). Other evidence indicating that mixing of parental cytoplasm is very slow comes from the studies of Siniscalco et al. (1969) on the location of the activity of glucose6-phosphate dehydrogenase around one of the nuclei in 2-day-old heterokaryons resulting from the fusion of human fibroblasts with normal and zero enzyme activity.
III. E L E C T R O N M I C R O S C O P I C ANALYSIS OF CELL FUSION INDUCED BY T H E ACTION OF P O L Y E T H Y L E N E G L Y C O L (PEG) The use of inactivated Sendai virus for the fusion of cells and further selection of hybrids has a number of limitations. A major condition must be met, the cell surface must have receptors for the adsorption of the virus. For this reason, many kinds of cells, particularly of plant origin, cannot be fused; the technical difficulties imposed by an, albeit inactivated yet infectious virus, are too great. Starting in the 1970s there was an increasing search for chemical and physical agents with fusogenic action (Poole and Nowell, 1970. Ahkong et al., 1972). PEG proved to be a promising chemical on account of its high fusogenic efficiency and relatively low toxicity. PEG has gained wide acceptance because it has made feasible the production of hybrid forms in plants (Pontecorvo, 1975). We studied the morphology of cells and their fusion patterns in cultured Chinese hamster and human fibroblasts (embryonic and skin) for 2 days after treatment with 50% PEG. Fusion was performed in a suspension of cells collected after centrifugation of a mixture of 5 x 106 of both cell types. Chinese hamster cells were first grown in an 3H-thymidine medium, as in the experiments with virus-induced fusion. The genetic material derived from Chinese hamster was, therefore, labelled in the heterokaryons and hybrids.
7
In analysis of the cell suspension promptly after washing away the PEG (15rain after its application), we observed 16% of the cell population to be multinuclear cells which, judging by our criteria, were referred to heterokaryons. Surveying the electron micrograms of these cells, it is clear that the fused plasma membranes have become structure~ less bands at many sites so that the cytoplasm of the clumped cells is hardly, if ever, disconnected by membranes; yet the cytoplasmic structures of the parental cells have not, so far, mixed together (Fig. 3a). There were about 10% multinuclear cells in the culture 20-24 hr after treatment with PEG; rare stray ones were heterokaryons. The electron micrograph shown in Fig. 3b demonstrates a heterokaryon with four centrioles. Uni- and multinuclear synkaryons were observed 48 hr after treatment of the cells with PEG, but no heterokaryons appeared at that point in time. The majority of heterokaryons had transformed to synkaryons by 20 hr after treatment with PEG. This indicates the significant difference in the time course of formation of hetero- and synkaryons, after cellular treatment with PEG or Sendai virus. With the aid of PEG, fusion is achieved faster and more synchronously. In the case of PEG, cell fusion is effected through the direct action of the fasogen on the membranes. PEG is hydrophobic and surface-acting, and it also binds calcium ions. This property provides the structural modification of the plasma membrane, the redistribution of the electric charge over the membrane surface, the resulting fusion and ultimate lysis of the cell membrane (Kao and Michaljluk, 1974). If the structure of the non-fused plasma membrane cannot be restored, the cells die. Clearly, at minimum toxicity of PEG, cells fuse rapidly, and soon thereafter the entire agglutinated plasma membrane surfaces breakdown. Observations of the ultrastructural changes produced by PEG revealed that they concerned not only the cell plasma membrane; changes occur also in the nuclear-nuclear (Fig. 3c), nuclear-mitochondrial (Fig. 3d, e, f), mitochondrial-mitochondrial (Fig. 3g) membranes and lipid inclusions. Fragments of nuclear membranes are very frequently seen in the nuclei. Examination of serial
,~
A. Yu. Kerkis and N. S. Zhdanova
(a)
Fig. 3a to c. ultrathin sections assured us that they are, indeed, fragments and not invaginations of the nucleus. We did not observe fragments of this kind in parental cells, heterokaryons and synkaryons produced by virus-induced fusion. In contrast, these
fragments were present in almost all the hybrid nuclei of cells fused by PEG. A plausible mechanism for the origin of these fragments in the hybrid nuclei is offered. After penetrating into the cells. PEG may exert its action on the nuclear
Somatic Cell Hybrids
f)
9
I
~:•
Fig, 3d to h.
membrane, as it does on the plasma membrane, and, as a result, heterokaryon nuclei fuse not only during mitosis, when the nuclear membrane disappears [this is the case when hybrid cell formation is virus-induced (Harris et al., 1966; Reichsteiner
and Parson, 1976)], but also during interphase. Suggestive evidence for the possible direct fusion of nuclei formed after PEG treatment may be the numerous multinuclear cells we observed with hybrid nuclei. No such heterokaryons were
A. Yu. Kerkis and N. S. Zhdanova
Fig. 3i and j. Fig. 3. The stepwise formation of a hybrid cell under the influence of polyethylene glycol (PEG). tt H u m a n fibroblast: C hamster fibroblast. Scale markers indicate 1/~m. (a) The cells start to fuse 15 min after PEG has been washed off. Arrows point to the region of disintegrating plasma membrane. (b) A formed heterokaryon 4 hr after treatment with PEG. Arrows point to the double set of centrioles. (c) Arrows pinpoint tile sites where the nuclei of a heterokaryon fuse during the formation of a hybrid nucleus. 2 6 hr after exposure to PEG. (d f) Possible fusion of mitochondria with the nucleus in a hybrid cell, resulting from treatment with PEG. N--nucleus: M mitochondrion, Arrows point to the site of the disintegrating nuclear envelope. (g) Fusion of mitochondrial membranes (M). (h) Fusion of lipid inclusions (L) in a hybrid cell under the effect of PEG. (i j) Fragmentation of nuclei and nuclear envelope during the formation of a hybrid cell. Arrows point to the site where the nuclear membrane is destroyed and fragmented.
Somatic Cell Hybrids found in the case of Sendai virus-induced fusion of cells. Figures 3i and j demonstrates another effect of PEG, the fragmentation of nuclei in the hybrid cells. In some instances, the fragmentation of the hybrid nucleus was complete, and in others small fragments break away from the nucleus. The virusinduced synkaryons showed no such patterns. Thus, rapidity and synchronicity are the features of PEG-induced cell fusion. Furthermore, according to our present observations, cell treatment with PEG produces definite changes not only in the membranes of the cell itself, but also in the nucleus and cytoplasm. As a consequence, direct fusion of cells becomes feasible at any stage of the cell cycle: it ceases to be restricted to mitosis. Reasoning still further, it appears that the effects of preliminary condensation, pulverization of chromosomes (the known concomitants of the fusion of cells at different stages of their cycle), among others, on chromosome segregation, as well as the appearance frequency of chromosome rearrangements, are exaggerated in the PEG-induced compared to the virus-induced hybrids. The fragmentation of nuclei definitely contributes to chromosome elimination. When comparing the two fusogenic agents, it becomes more evident that PEG has the merit of simplicity, whereas Sendai virus spares the integrity of genetic material in the hybrid cells.
IV. F O R M A T I O N OF H Y B R I D CELLS DURING S P O N T A N E O U S FUSION B E T W E E N S P L E N O C Y T E S AND FIBROBLASTS
When the formation of hybrid cells is included by Sendai virus or PEG, the plasma membrane and in the case of PEG the majority of intracellular membranes also lose their structural integrity. The cytotoxic effects of these fusogenic agents may be attributed to their destructive action on the membranes, which was stronger in the case of PEG. This action may be one of the causes of the numerous chromosome rearrangements. These rearrangements make it difficult to locate chromosomal genes in somatic cell hybrids and occasionally
11
renders impossible any analysis of the early steps of hybrid cell formation. With this in mind, we attempted to generate hybrid cells in a way that is not traditional, by taking advantage of the ability of lymphocytes to rapidly penetrate into target cells. In the relevant experiments Chinese hamster fibroblasts were cocultivated with fox or mouse splenocytes. Fox and mouse spleens were isolated under sterile conditions, were sliced, washed and dissolved in a solution ( D M E M with supplements, 10% fetal calf serum and 250 ng/ml PHA) so that there were 10 splenocytes for each fibroblast. Thereafter a splenocyte suspension was transferred to a Petri dish containing growing Chinese hamster fibroblasts. The results were best when fibroblasts were grown to semiconfluent monolayers. This was followed by fixation of cells after 1, 4, 8, 24, 48 and 72 hr. The procedures for the fixation, embedding, and choice of cells for EM analyses were standard, as described above (Section II). Splenocytes are able to freely enter the cytoplasm of the fibroblast, while the nucleus, being protected from the attack of lysosomal enzymes, is spared and hence it shows no morphological changes up to 72 hr. Some of the splenocyte nuclei are able to enter the fibroblast nuclei at 1 - 4 h r (Fig. 4a, b,c and d). After 24-48 hr blocks of condensed chromatin from the splenocyte nuclei are seen in the nuclei of the fibroblasts (Fig. 4e, f and g). The amount of condensed chromatin in the fused nuclei decreases in the process of cocultivation presumably as a result of chromatin activation. Based on the data obtained, the sequence of events providing the formation of somatic cell hybrids may be envisaged as follows (1) The synkaryon would be formed as a result of the rapid passage of the splenocyte into the nucleus of the fibroblast; as a consequence, its plasma and nuclear membranes would be destroyed. (2) The chromatin of the splenocyte nucleus, residing in the fibroblast cytoplasm for 15-20 hr of cocultivation, would fall apart into chromosome-like blocks. Each block is seen outlined by a double membrane presumably formed from the destroyed membrane of the splenocyte. If this is the case, the synkaryon
12
A. Yu. Kerkis and N. S. Zhdanova
lib
t
(d)~ ~k 3
~t ¸~ \ .
I
I
Fig. 4a to d.
would be formed at that time point of mitosis, when separate membrane-outlined fragments may be integrated into the assembling fibroblast nucleus. As in the condensed chromatin of the splenocyte fragments would be rather well identified up
to 72 hr (Fig. 4h). At this time point, the membrane fragments attached to the remains of condensed chromatin are still observed. The data obtained demonstrate the lklcility ol high frequency production of somatic cell hybrids
Somatic Cell Hybrids
Fig. 4e to h. Fig. 4. Formation o f a hybrid cell during the cocultivation o f splenocytes (Sp) with fibroblasts (F). Scale markers indicate I/~ m. (a) Early fusion 1 hr after cocultivation. (b, c) Steps of the penetration of a splenocyte into fibroblast cytoplasm 4-5 hr after cocultivation. Arrows point to the sites of fusing membranes. (d) Fragmentation of a splenocyte nucleus in the cytoplasm of a fibroblast 18-24 hr after cocultivation. (e-h) Distribution of chromatin blocks of splenocytes (arrows) in the nuclei of hybrid cells 48 hr (e, f), 6 hr (g) and 72 hr after cocultivation (h).
13
14
A. Yu. Kerkis and N. S. Zhdanowl
by cocultivation of fibroblasts with splenocytes, w i t h o u t the aid of fusogenic agents. It is possible that this approach should allow to generate more somatic cell hybrids without chromosome rearrangements, thereby facilitating investigation of the patterns of hybrid cell formation.
V. U L T R A S T R U C T U R A L F E A T U R E S OF THE NUCLEI OF H Y B R I D CELL C L O N E S Genetic material of species frequently differing in number, size and other chromosomal features is internalized within the hybrid nucleus. Ordered spatial distribution of all the chromosomes within such a tetraploid nucleus is a requisite for its function. Consequently, the structure of the nuclear matrix is of importance for the formation of the hybrid nucleus. There is now convincing evidence indicating that discrete elements of the nuclear matrix are involved in the determination of the functionally ordered internal structure of the nucleus. Thus, it has been shown that the chromatin fibrils are bound to the nuclear matrix, thereby ensuring the higher order of chromatin organization (Cooch and Brazells, 1975), that transcription is initiated in the binding sites (Comings and Okada, 1975; Berezney, 1979) and that it is here that the actin genes are, as a rule, located (Small et al., 1985), thereby providing evidence for the important role of this structure in the organization of active chromatin (Lafond and Woodcock, 1983). Elements of the nuclear matrix accomplish important regulatory functions, being involved in D N A replication and intranucleus transport of R N A (Zbarsky, 1985). These elements seem to promote the regular passage of mitotic chromosomes to the opposite cell poles, and they can turn to scaffold-like structures during mitosis (Bekers et al., 1981). As a whole, the nuclear matrix may be regarded as a flexible structure capable of reducing in a reversible way so that the volume and shape of the nucleus can change during interphase (Berezney and Coffey, 1977). The nuclear lamina, one of the structurally fundamental units, was found to consist of three protein components which are different from the
other proteins of the nuclear matrix. The nuclear lamina structurally interacts with the inner nuclear membrane and nuclear pore complexes, on the one hand, and chromatin elements on the other (Gerace and Blobel, 1980). During mitosis, the lamina can reversibly pass to the monomeric state with one, at least, of the constituent proteins remaining bound to the membrane fragments o1" the disrupted nuclear membrane, It is known that the fragments of the nuclear membrane remain associated with single chromsomes during mitosis (Kerkis and Christolubova, I191; Zatsepina et al.. 1976) and that the new nuclear membrane starts to assemble from these bound structures during telophase. It appears that the specificity and number of attachment sites of the chromosomes to the lamina and, therefore, the spatial distribution of the chromosomes in the interphase nucleus are all determined through the mediation of these bound structures. When taking this body of evidence together, it may be assumed, with good reason, that genomic stability, i.e. the segregation rate of chromosomes in a hybrid cell, is dependent upon the structure o1" the nuclear lamina. We have carried out EM studies of the nuclei from different hybrid clone cells with reference to the structures associatcd with the nuclear matrix. Chinese h a m s t e ~ h u m a n somatic cell hybrids were generated by fusion induced between Chinese hamster fibroblasts (clone M I5, H P R T ) and human embryonic fibroblasts at 7 10 passages. Hybridization was performed as described above with the aid of UV-inactivated Sendal virus. 10 ' ~1 ouabain was added to the standard selective medium H A T and killed human fibroblasts. Five clones of this type containing a tetraploid set of Chinese hamster chromosomes and from 11 20 human chromosomes were taken for EM analysis. We also studied five mouse hepatoma cells (clone BW8)-American mink fibroblast ( M V T K ) hy~ brids. These hybrid clones contained a tetraploid set of mouse chromosomes and from 2 17 chromosomes contributed by American mink. These clones were obtained using PEG, as described above (Zhdanova et al., 1985). For EM analysis, the cells were frst chosen under the light microscope, and then serially
Somatic Cell H y b r i d s
sectioned with the use of our methods. Replication of the hybrid cell nuclei was studied by the modified method of electron autoradiography (Kerkis and Khristolubova, 1974). For this pur-
~, U I
15
pose, 18.5~tc/ml 3H-thymidine (18C/mM; Amersham) was added to the culture medium 30 min before the beginning of fixation. Ultrathin sections were covered with emulsion of Ilford L4 type.
i
Fig. 5a to d.
1()
A. Yu. Kerkis and N. S. Zhdanova
Fig. 5e to g. Fig. 5. Ultrastructural fragmentation of the nuclei of hybrid clones. Scale markers indicate l #m. (a c) Tubular finger-like invaginations (arrows) of internal nuclear envelope in longitudinal (a, c) and transversal sections {b). N--Nucleus, C---cytoplasm, Ch-q:hromatin. (d) Electron microscopic autoradiograph of 3H-thymidine-labelled chromatin surrounding the tubular structures (arrows). (e) Specific histochemical staining for transcriptionally active chromatin surrounding the tubular structure (arrows) according to Frenster's reaction (for explanation see text). N--Nucleus, C-~:ytoplasm. (f) Tubular structures in the nuclei of hybrid cells formed as a result of combined treatment with Triton X-100 and DNase. Arrows point to the longitudinal sections, and triangles to the transversal sections of pore complex. (g) Blocks of condensed chromatin in the nuclei of some hybrid clones.
Somatic Cell H y b r i d s
Transcriptional activity of the nuclei was analyzed by means of Frenster's method (Frenster, 1971; Losecke et al., 1980). In studies of the structure of the nuclear matrix composition cells were treated before fixation with 1/~g/ml DNase I (Serva) for 1.5 hr, then 0 . 2 ~ Triton X-100 for 20 min, or Triton X-100 together with D N a s e I. The electron micrograph in Fig. 5a shows deep invaginations of the inner layer of nuclear membrane of hybrids, thereby greatly increasing its nuclear surface area. This makes the appearance of the hybrid cell strikingly different from the parental cells. O f interest are the structures within the hybrid nuclei, which have not been previously described. In cross sections, they appear as numerous round cavities with a monolayer membrane and an internal diameter of about 40 nm (Fig. 5b). As many as 70% of all the cells had this structure in some clones. In longitudinal sections, they appeared as tubular cavities surrounded by chromatin. The structures traversed the nucleus in all the directions, and they were well defined far from the nuclear envelope (Fig. 5c). Examination of serial sections demonstrated that these structures are finger-like invaginations of inner nuclear membrane surrounded by closely adhering loose chromatin. When 3H-thymidine was added 30 min before fixation, the silver grains were mainly distributed over chromatin associated with these tubular structures. This was firm evidence for the replicative activity in these chromatin regions. Figure 5e shows that the electron-dense products of the D N a s e - A O reaction occur, as a rule, in the nucleolus and chromatin surrounding the tubular
17
cavities. Thus, we obtained experimental evidence indicating that these chromatin regions are also active in transcriptional terms. Treatment of the cultured cells with DNase I solution promptly before fixation removes the bulk of chromatin from the nucleus and, as a result, branched tubular cavities are clearly seen in the nucleus. In the case when the cells are, in addition to DNase I, treated with Triton X-100 (i.e. subjected to a procedure routinely used for isolation of the nuclear matrix), numerous elements of the intracellular matrix associated with the tubular cavities containing the characteristic components of the pore complexes and lamina are seen in the section (Fig. 5F). Thus, our observations indicate that the numerous nuclear structures formed by the finger-like invaginations of the inner nuclear membrane are closely associated with the elements of the intracellular matrix bound to the transcriptionally active chromatin. The chromatin associated with these structures is probably that part of the laminabound chromatin which determines the location of the chromsomes in the hybrid cell, both on the nuclear membrane and, in a broader sense, the spatial arrangement of chromatin in the nucleus. Interestingly, among clones studied there occurred some containing the above described structures and others with cells having nuclei containing large globules of condensed electron-dense chromatin, occasionally occupying about 35% of a section area (Fig. 5g). According to the measurement data, the amount of condensed chromatin is not related to the number of human chromosomes present in the hybrid nucleus of a clone, nor is it associated with any appreciable increase in the total area of a
Table 1. Relative n u m b e r of c o n d e n s e d c h r o m a t i n in the cell nuclei of Chinese h a m s t e r x h u m a n fibroblast hybrids
Clone 3S 2 3S
