Nuclear transfer from teratocarcinoma cells into mouse oocytes and eggs

Teratocarcinomas are tumors that develop spontaneously in the testes and ovaries of mammals and are composed of a variety of aberrantly differentiated tissues. These gonadal tumors are either benign, nontransplantable and are then referred to as teratomas, or malign, transplantable and therefore called teratocarcinomas. As their stem cell population, teratocarcinomas contain embryonal carcinoma (EC) cells, which are scattered throughout the solid tumors in small groups, continue to divide and proliferate, and sometimes even invade and metastasize, thereby retaining the malignant properties of transplantable tumors (Stevens, 1967). Some of those teratocarcinomas have been converted to a modified ascites form dubbed embryoid bodies (EBs) because of their close resemblance to normal embryos. The EBs can be propagated in other syngeneic mice via intraperitoneal inoculation and, in this way, established as ascitic tumors for many transplant generations. The similarities between EC cells and normal embryonic cells have led to the widespread use of teratocarcinomas as a model for studies of cell differentiation (see Shermann and Solter, 1975; Muramatsu et al. 1982; Silver et al. 1983).

Under conditions in vivo, when the EC cells were transplanted into mouse blastocysts, they were able to participate in normal tissue differentiation of adult chimeric mice. During this developmental process, the EC cells revealed their true pluripotentiality and even lost their malignant properties (Brinster, 1974; Mintz and Illmensee, 1975; Papaioannou et al. 1975). Some of the chimeric mice were free of tumors, although the EC cells had contributed to many organs (reviewed by Illmensee and Stevens, 1979). The developmental potentiality of nuclei from the EC cells can be tested via nuclear transplantation (reviewed by DiBerardino, 1987). Is the nucleus from a mammalian tumor cell able to bring about development when transferred into an enucleated oocyte or egg?

In 1982, one of us (K.I.) had carried out such nuclear transfer experiments taking cells of an ascitic teratocarcinoma derived from a spontaneous ovarian tumor. The nuclei of EC cells from this ascitic tumor had been microinjected into fertilized but enucleated mouse eggs. Development of some preimplantation embryos originating from the injected EC nuclei was obtained, and a few of them showed overall morphological features resembling embryos at the morula and blastocyst stage.

The aim of our experiments reported here was to repeat and extend the previous work. We have transferred nuclei from EC cells of another similarly derived ovarian teratocarcinoma into mouse oocytes and eggs. Such nuclear transfers were carried out using two different methods, i.e. fusion with polyethylene glycol or electric field and microinjection. Some of these results have been published recently (Illmensee et al. 1989). In cytological studies, we show that EC nuclei seem to be morphologically remodelled in the foreign cytoplasm and show structural features very similar to pronuclei from activated oocytes or eggs. Furthermore, from cytological and biochemical analyses, we show that, at least, some nuclei from EC cells are able to initiate preimplantation development which resembles normal embryogenesis.

Spontaneous tumors were dissected from the ovaries of female mice of the inbred strain LT/Sv originating from The Jackson Laboratory and subsequently bred in our animal colony. Small amounts of the solid tumors were injected subcutaneously or intraperitoneally into syngeneic females. From a total of 11 teratocarcinomas that could be established, three of them gave rise to an ascites tumor containing EBs.

For the experiments reported here, the ascites line LT/123 was used between its 25^th and 55^th intraperitoneal passages. The crude ascitic tumor was washed in DMEM (GIBCO). A sample was taken for microscopic observation. The other part was resuspended in DMEM without Ca²⁺ and Mg²⁺ or supplemented with glucose/EDTA and preincubated about 5 min at 37°C. Following centrifugation, the pellet was resuspended in 0.04–0.05 % Trypsin solution (GIBCO) and incubated about 5 min. During this period, the suspension was pipetted, and large aggregates were removed. Following repeated centrifugation and resuspension in DMEM with 30% FCS (GIBCO) or M2+15%FCS (Fulton and Whittingham, 1978), a volume of supernatant was taken and transferred into culture medium for storage under 5 %CO₂ at 37 °C.

Occasionally, we isolated a small ascites volume (about 0.1ml) 4 days after inoculation and transferred it into M2+15%FCS. Under these conditions, it was not necessary to trypsinize the EBs since at this early stage of tumor progression a sufficient number of single EC cells was present.

Females of the inbred strain C57BL/6 used for egg collection and C57BL/6xCBA-H females for oocyte collection were superovulated at the age of about 4 weeks, using 5 i.u. PMS and 5 i.u. hCG (ORGANON). To obtain fertilized eggs, the hormonally treated females were mated to C57BL/6 males. Oocytes and eggs were removed from the oviducts, treated with 1 mg ml^-1 hyaluronidase (SIGMA) in modified Whit ten’s medium (Hoppe and Pitts, 1973) to remove the cumulus cells.

Fertilized eggs at the pronuclear stage were transferred in modified Whitten’s medium containing 2.5μgml^-1 cytochalasin D (SIGMA) and preincubated for at least 30min. About 20 eggs were transferred into a microdrop for subsequent enucleation, following two different methods (Illmensee and Hoppe, 1981; McGrath and Solter, 1984).

A group of dissociated EC cells was transferred into the microdrop containing the preincubated eggs. EC cells of medium size with smooth surfaces, homogeneous cytoplasm and clearly visible nuclei (under Nomarski optics) were selected. Nuclear transfers were carried out as already described (Illmensee and Hoppe, 1981).

Unfertilized oocytes from which the zona pellucida had been removed in 0.5 % pronase (SIGMA) were transferred to M2 without albumin. The selected EC cells were added to the same well coated with 1% agar. About 5–10 min later, the oocytes were tranferred into 150–300 μg ml^-1 phytohemagglutinin (SIGMA) and incubated for about 4min. The oocytes were then returned to the well containing the EC cells and placed on top of them in such a way that each oocyte was surrounded by several cells. Such aggregates were treated for PEG fusion as previously described (Czolowska et al. 1984). Between 30–60 min after PEG treatment, one group of oocytes was fixed, as described below. About 30–45 min after PEG treatment, oocytes were activated with 8% ethanol (Cuthbertson, 1983) and then cultured in vitro for various time intervals from 2 to 24 h.

• Unfertilized and zona-free oocytes from which the spindle of the second metaphase had been removed microsurgically as previously established (Waksmundzka, personal communication), were treated for fusion with EC cells following the above mentioned scheme. The oocytes were then activated 45 min after PEG treatment, washed and transferred to culture medium. Fixation was done in the majority of cases between 18–22 h later.

• Unfertilized and zona-free oocytes were divided mechanically into five or six pieces (Tarkowski and Rossant, 1976). After removal of the piece with the spindle of the second metaphase, the oocytoplasts were brought into contact, each of them, with one to three EC cells. Only those that fused with a single cell were selected and injected under the zona pellucida of unfertilized oocytes. Following this microinjection, the oocytes were exposed to electric field (Kubiak and Tarkowski, 1985; Ozil and Modlinski, 1986). Between 30 and 45min after the beginning of fusion, the oocytes were activated as described above and transferred to ligated oviducts.

Two to five EC cells were microinjected through the zona pellucida into the perivitellin space. These eggs were then transferred into culture medium without albumin. Fusion was achieved under the following conditions: 50% PEG 2000 (FLUKA) for 1–1.5 min. The eggs were transferred to culture medium with albumin (MILES) and washed twice. They were controlled under the microscope to follow the process of fusion. The eggs were then cultured in vitro or in vivo.

Part of those eggs that have survived micromanipulation following fusion or injection was cultured in drops of modified Whitten’s medium and kept in a gassed and sealed desiccator at 37°C. Daily up to the fourth day, development of the experimental eggs was registered twice and compared with normal C57BL/6 embryos. Part of the manipulated oocytes was cultured in M2 up to 24 h. Another part of the oocytes and eggs was injected into the oviducts, usually five in each one, of anesthesized immature females of the Swiss/Albino strain. As a precaution, the oviducts had been ligated prior to oocyte or egg transfer. After one to three days, the oviducts were flushed with culture medium to recover the experimental embryos.

Samples of EC cell-fused oocytes were fixed in Heidenhain’s fixative. Whole mount preparations were carried out according to the method described (Tarkowski and Wroblewska, 1967) and stained with Ehrlich’s hematoxylin.

Three ascites bearing females were injected with 25 μg Colce-mid and killed 1 h later. The ascitic fluid was transferred into tubes containing 0.5 %KC1. The supernatant was transferred to other tubes and centrifuged. The resulting pellet was fixed with methanol: acetic acid (3:1). Small amounts of the pellet were dropped onto clean and dry slides. For differential staining of the chromosomes, the slides were processed for G-banding (Seabright, 1971), for C-banding (Sumner, 1972) and for NOR-staining (Albini et al. 1984).

For electrophoresis, the embryos were sucked individually or, occasionally, pooled into micropipettes containing 0.1 μl of 50mM-Tris-HCl, 1mM-EDTA (MERCK) and 1mM-β-mer-captoethanol (SERVA) at pH 7.5. The embryos were lyzed within the pipettes by freeze-thawing. Each sample was then applied onto a marked dot of the cellulose acetate plate. Electrophoresis and staining of the gels followed established procedures (Eppig et al. 1977).

We first had to establish new teratocarcinoma lines, since the original tumor LT/263 used for nuclear transplantation in 1982, could no longer be maintained. Among 11 newly isolated tumors that were growing after subcutaneous passage in LT/Sv females, three of them could be converted into an ascites form. The three lines were analyzed morphologically and tested in preliminary nuclear transfer experiments by injecting nuclei from their EC cells into fertilized but enucleated C57BL/6 eggs. However, only one line, dubbed LT/ 123, gave satisfactory results, i.e. some nuclei from EC cells of this ascitic line were able to initiate preimplantation development, and some of the resulting embryos looked morphologically similar to 8- to 12-cell embryos. In contrast, nuclei from the other two lines resulted in fragmentation and abnormal cleavage (Illmensee, unpublished data). We therefore concentrated our work on the LT/123 line and the results reported here were obtained during a period from the 25^th to the 55^th tumor passages.

The solid tumor LT/123 is composed of several tissue derivatives including embryonal carcinoma (EC) cells, that is typical for a multidifferentiated teratocarcinoma (Fig. 1). The ascitic tumor is composed of clusters of cells morphologically similar to EC cells and EBs (Fig. 2).

During the 46^th and 48^th passages, the karyotype of the ascitic line was analyzed in more detail. Almost half of the cells (49 %) showed 41 acrocentric chromosomes. In 26% of the cells, 40 acrocentric chromosomes (the standard karyotype) were present (Table 1). After C-band staining, all chromosomes were endowed with C-band-positive segments around the centromere. No distal or intercalary C-bands were found. One chromosome possessed a large C-band-positive block. In G-banded prometaphase, one chromosome 12 exhibited an exceptional large Al segment that corresponds to the C-positive segment after C-banding. After NOR staining, five to nine chromosomes showed proximal to the centromere a positive signal with varying intensities. The NORs were localized at the same position as on chromosomes of normal mouse cells.

After G-banding, twelve metaphases were analyzed in more detail by constructing the whole karyotype. (As a comparative standard, see Nesbitt and Franke, 1973). In all cells, only one X chromosome was present. The cell line had gained one chromosome 6 and one chromosome 15 (Fig. 3). In addition to the consistent MsX, Ts6 and Tsl5, some other chromosomes could show rearrangements. In some cells, one chromosome 2 was smaller than normal. According to comparative G-band analysis, the small size is due to an intercalary deletion. The deleted segment comprises the region C2 to D. On the other hand, sometimes a large chromosome 5 has been identified. The enlargement can be characterized as a duplication of the terminal segments F and G of the same chromosome. Yet, cells with an altered chromosome 5 do have a partial Ts5. The alteration sometimes observed in chromosome 10 is explainable by a paracentric inversion of the segment B3 to C2.

Oocytes without zona pellucida were PEG-fused with EC cells, activated with ethanol and then cultured in vitro up to 22 h. From a total of 63 PEG-treated oocytes, 45 were briefly subjected to ethanol for activation and subsequently cultured in vitro. Between 30 min and 22 h after PEG treatment samples of experimental oocytes were fixed and further processed for morphological analysis. From 35 oocytes analyzed, 21 were heterokaryons containing the female pronucleus and the EC nucleus, 8 were non-fused oocytes and 6 were fused but non-activated oocytes. The latter ones happened to be fixed 30–40 min after ethanol treatment. In three of them, the introduced EC nucleus already showed typical premature chromatin condensation (PCC reaction; Fig. 4).

In the 21 activated oocytes the morphological changes in the EC nucleus were studied at the different time intervals during which this nucleus was exposed to the cytoplasm of the activated oocyte. Morphological remodelling of the EC nucleus was composed of initial swelling, decondensation of chromatin, and considerable increase in size (Fig. 5). Of four eggs fixed about 3h after PEG treatment, two contained a small female pronucleus and an EC nucleus that showed some swelling when compared with its original size. About another 3 h later the process of nuclear swelling and decondensation of chromatin had continued (Fig. 6). Of six activated oocytes that were fixed 6 h after PEG treatment, three contained an EC nucleus equal in size to the female pronucleus whereas in the other three oocytes, the EC nucleus was slightly smaller. About 13 to 15 h after PEG treatment, the EC nucleus did not further increase in volume and had reached about the same size as the female pronucleus (Fig. 7). They had approached each other in the centre of the oocyte. Usually the EC nucleus could be recognized by its darker staining. Otherwise the two nuclei showed similar chromatin and nucleoli structure, and sometimes looked very much alike. About 17 h after fusion and activation, the EC nucleus and the female pronucleus were tightly touching each other (Fig. 8). Approximately another 5h later, following dispersion of both nuclear membranes, the chromosomes from the EC nucleus and female pronucleus had formed a common metaphase (Fig. 9), as seen in four oocytes out of seven fixed and stained about 22 h after development in vitro.

In the first series, microsurgically enucleated oocytes from which the zona pellucida was removed were PEG-fused with EC cells, activated and cultured in vitro for up to 22 h. In the second series, small oocytoplasts were PEG-fused with EC cells and then injected into the perivitelline space underneath the zona pellucida of enucleated oocytes. After electric field (EF) treatment and ethanol activation, the nucleocytoplasmic hybrids were cultured in vivo up to three days.

In the first series, from 69 PEG-treated oocytes, 43 were subjected to ethanol of which 36 could be cultured in vitro. 19 EC cell-fused oocytes showed activation as well as 12 non-fused oocytes, whereas 5 non-fused oocytes remained non-activated. The majority of oocytes of this series were fixed for cytological analysis between 18 to 22 h after PEG treatment. From the 19 nuclear transplant oocytes, 13 could be analyzed after staining. Three of them degenerated, one contained a relatively large EC nucleus, five had entered the first mitosis or were showing aberrant karyokinesis and four embryos had passed the first cleavage division normally to the 2-cell stage.

In the second series, from a total of 85 EF- and ethanol-treated oocytes, 79 were transferred to ligated oviducts of young recipient females and 62 were recovered after one to three days later (Table 2). In seven of the twelve embryos at the 2-cell stage, the two blastomeres, each containing an EC nucleus with several nucleoli, looked morphologically normal with respect to their size and structure. In the remaining four, differences in size and morphology were observed for nuclei and the two blastomeres when compared with normal development. In one of the 3-cell embryos, one cell appeared to lack a nucleus, whereas in the other 3-cell embryo one cell contained two small nuclei. The 5-cell embryo was composed of equally sized blastomeres but with some structural differences between their nuclei. The 6- to 8-cell embryo also showed slightly abnormal features. The two morulae were in a semicompact and premature state, and variation in size of nuclei was seen in some blastomeres (Fig. 10).

In 11 experimental series we have investigated the behaviour and fate of EC nuclei in fertilized eggs from which the two pronuclei have been microsurgically removed. Two different methods have been applied to transfer nuclei from EC cells into enucleated eggs, that is fusion of the EC cell injected under the zona pellucida with the egg membrane via PEG (a) and microinjection of the EC nucleus directly into the egg (b). From 174 eggs that survived the various micromanipulations, 130 were cultured in vitro and 44 were transferred into ligated oviducts for further development in vivo (Table 3).

In order to determine whether the introduced EC nuclei were involved in development, electrophoretic enzyme tests for glucose phosphate isomerase (GPI) were carried out on the nuclear transplant embryos including the abnormal and fragmented ones. Donor nuclei (GPI-A) and recipient eggs (GPI-B) were genetically marked with different allelic GPI variants to distinguish them. The experimental embryos were blotted individually for enzyme analysis or, rarely, pooled as primarily done with the fragmented embryos. We have tried to carry out the enzyme tests on all embryos and have presented our results in Table 4. In several embryos, in addition to the EC-specific GPI-A the eggspecific GPI-B was also detected. In the other embryos only EC-specific GPI was found. (As examples, see Figs 11–14).

About 15 years ago it was shown independently by several investigators that EC cells from teratocarcinomas, when transplanted into mouse blastocysts, participate in orderly differentiation of many tissues in adult chimeric mice and can even give rise to functional germ cells (cited in Bradley et al. 1984). Reversion to normalcy of the originally malignant EC cells has led to the conclusion that the cellular environment of the embryo may be important for this process to occur. Besides the striking histogenetic potential of EC cells, the developmental capacity of their nuclei remained unsettled. It should be of interest to determine whether the nucleus of such an embryo-related tumor cell can be returned into an oocyte or egg. In amphibians, the transfer of nuclei from Lucke’s renal adenocarcinoma cells into eggs gave rise to tadpoles and thus revealed the remarkable developmental plasticity of these tumor nuclei (McKinnell et al. 1969).

Karyotype analysis on the ascitic tumor LT/123 used in our studies has revealed that in the majority of cells the chromosome numbers are in the diploid range. All chromosomes are of the same acrocentric shape as the standard karyotype. Furthermore, the localization of NORs and C-bands is comparable to that in normal mouse chromosomes. However, the consistent abnormality in these cells was MsX, Ts6 and Tsl5. It is well known that numerical aberrations, especially trisomy of certain chromosomes, are associated with tumori-genesis (Wiener et al. 1981; Herbst and Gropp, 1982). On the other hand, trisomy within the developing mouse embryo exerts detrimental effects. All auto-somal whole arm trisomies are lethal during gestation or soon thereafter (Epstein, 1986). Tsl5 embryos show a retarded and abnormal development and are viable up to 11 days of gestation. The effect of combined chromosomal aberrations on embryonic development like in our LT/123 line cannot be envisaged exactly, but we assume that its altered genome reduces the developmental capacity. In future experiments, it will be desirable to utilize other EC cells corresponding more closely to a normal diploid karyotype that is certainly more advantageous for nuclear transfers.

After fusion of EC cells with non-activated oocytes, we have observed a similar reaction of the introduced EC nucleus, called premature chromatin condensation (PCC), as reported for thymocyte nuclei (Czolowska et al. 1984). This PPC reaction, which was initiated already about 30 min after fusion, is induced by a factor(s) released after germinal vesicle breakdown into the cytoplasm of metaphase II oocytes and influences the chromatin structure of the EC nucleus towards the formation of meiotic metaphase chromosomes as it is postulated for other experimental conditions (Matsui and Clarke, 1979; Tarkowski, 1982; Sörensen et al. 1985).

Following fusion of EC cells or oocytoplasts carrying EC nuclei with activated oocytes, the introduced tumor nuclei changed morphologically and passed through a sequential series of remodelling events that have also been seen in other embryonal carcinoma cell nuclei (McGrath and Solter, 1985) and thymocyte nuclei (Szöllösi et al. 1988). They comprise nuclear swelling, decondensation of chromatin, formation of chromatin structure resembling the female pronucleus, appearance of several nucleoli (but not seen in PCC3 nuclei, McGrath and Solter, 1985) and considerable increase in nuclear size. About 22 h after fusion and activation, the chromosomes of EC nuclei formed a common metaphase with the chromosomes of the female pronucleus. Such heterokaryons have already been documented, taking place in non-enucleated eggs injected with morula cell nuclei (Modlinski, 1978).

Although remodelling of EC nuclei seems to require several hours, it is not possible to give an exact timing due to several parameters such as variations in the state of the cell cycle for the different EC nuclei before fusion, developmental variations between the activated oocytes, or different periods of time between fusion and activation during which the EC nucleus was already exposed to the oocytoplasm. In the thymocyte system it was found that when nuclei entered the oocyte one hour after activation, they were no longer able to show remodelling (Czolowska et al. 1984).

Relevant to this phenomenon are observations on mouse blastomeres transplanted into activated oocytes (Bolesta and Modlinski, unpublished data). When the time lapse between activation and fusion was enlarged, the tendency for non-proper remodelling of the blastomere nuclei increased. Furthermore, in other relevant experiments, it was shown that between 1 to 6h after activation of the oocyte, the sperm could still penetrate it but showed less and less normal pronucleus formation (Tarkowski and Tittenbrun, unpublished data) similar to what was seen in polyspermy (Witkowska, 1981).

At the biochemical level, there is only rudimentary information about a possible relationship between in-corporation of oocytoplasmic proteins into transplanted nuclei and nuclear remodelling (reviewed by DiBerardino, 1987). In the mouse, it was shown that proteins involved in nuclear decondensation are located in the oocytoplasm (Clarke and Matsui, 1987). Besides the structural changes during nuclear remodelling, important changes in gene expression should occur during this process (reviewed by Brachet, 1987).

In relation to other nuclear transfer experiments carried out with late preimplantation nuclei (McGrath and Solter, 1984; Surani et al. 1984), we wanted to find out more about the potential of EC nuclei. After having kept the nuclear transplant embryos under culture conditions in vitro and in vivo for several days, the majority of them fragmented or cleaved abnormally. This could have been due to an unfavorable stage in the cell cycle of donor EC cells or recipient oocytes and eggs, variations in the karyotype among EC cells, non-optimal time lapse between EC cell fusion and oocyte activation, experimental and chemical treatment or a combination of these possibilities.

The nuclear transfer embryos originating from fertilized and enucleated eggs were analyzed, mostly individually, for two allelic GPI variants to distinguish nuclei from the EC cell (GPI-A) and the egg (GPI-B). The embryos flushed from oviducts looked more healthy than those cultured in vitro, indicating superior conditions in vivo. During this period in the oviduct, the developing embryos also remained in a closed physiological milieu not amenable to possible fluctuations during culture in vitro. Morphological analysis, as shown for the oocyte series, would not be sufficient since incomplete enucleation of eggs could give artifactual results. This would be revealed unequivocally in the enzyme tests. It should be emphasized, that these tests were carried out on the embryos which had been kept in culture over four days. It was found that several developmentally arrested or fragmented embryos contained the egg- and EC-specific GPI, as it was reported for other nuclear transfer embryos (Gilbert and Solter, 1985). Direct comparison, however, is hampered by the fact that in these studies, pronuclei not showing GPI-specific gene activity before transplantation were used, whereas in our studies the EC nuclei were already genetically active for GPI. We did not observe an intermediate hybrid band of GPI-A/B that had been found occasionally in embryos derived from transplanted nuclei of ascitic LT/263 cells (Illmensee, unpublished data) or from nuclei of ICM cells (Illmensee and Hoppe, 1981). The more advanced embryos contained only the tumor-specific GPI, while the eggspecific GPI must have been metabolized during development. Otherwise, we would have detected it on the gels. In some instances, GPI-B could have been still present but was difficult or uncertain to deduce from the gels. Since some embryo samples showed only weak GPI-A staining, we do not know whether GPI-B was present but undetectable under these conditions.

Most of the embryos were arrested during early preimplantation, predominantly at the 2- to 4-cell stage, when normally the embryonic genome becomes transcriptionally active (e.g. Clegg and Piko, 1983). It was also shown that the largest qualitative changes in protein synthesis during preimplantation occur between the 2- and 8-cell stage (reviewed by Magnuson and Epstein, 1981). The increased developmental arrest of the nuclear transplant embryos around this crucial period might be explained by the failure of the teratocarcinoma genome to interact properly with the egg cytoplasm, although this genome originally derived from parthenogenetically activated diploid eggs (Stevens and Vamum, 1974; reviewed by Kaufman, 1983).

We would like to thank F. Abbet, M. Ackermann, M.F. Blanc, D. and F. Chatelain for careful technical assistance, B. Maichel for typing the manuscript, L. Slomkowski for photography and the University of Geneva for supporting part of this work. We thank Professor A. K. Tarkowski for carefully reading the manuscript. B.L. was awarded with a fellowship from the Italian Government. J.M. was supported by funds from the University of Geneva. K.I. would like to express his gratitude to Professor G. Czihak for his generous support and encouragement during the final period of this work.