Formation of viable chimaeras by aggregation between teratocarcinomas and preimplantation mouse embryos

Chimaeras, which are individuals composed of two or more populations of cells derived from different embryos, have been used to investigate a wide variety of problems in the reproduction, genetics and development of mice (for reviews, see McLaren, 1976; Russell, 1978). Recently, one particular class of chimaeras has attracted much attention, namely those formed between mouse embryos and embryonal carcinoma (EC) cells, the malignant stem cells of teratocarcinomas (Brinster, 1974; Mintz, Illmensee & Gearhart, 1975; Mintz & Illmensee, 1975; Papaioannou, McBurney, Gardner & Evans, 1975). Such chimaeras have been of interest since it is hoped that EC cells may be used as vectors for the introduction of new or mutant genes into the germ line of mice. Furthermore, the colonization of the embryo by the EC cells has been of interest in the study of malignancy (Pierce et al. 1979; Papaioannou, Evans, Gardner & Graham, 1979).

Chimaeras have usually been produced using one of two techniques. Either embryos, prior to blastocyst formation, have been aggregated together to form a single composite embryo (Tarkowski, 1961 ; Mintz, 1964), or a single cell or groups of cells have been injected into the blastocoele cavity of a blastocyst using a micromanipulator (Gardner, 1972). It is this latter method that has been used to produce EC ⟷ embryo chimaeras. Here, a method is described for producing viable EC ⟷ embryo chimaeras by the alternative aggregation technique and this method provides a simpler and as efficient means for introducing EC cells into embryos. In addition to this, the colonization of tissues in the post implantation embryo is described in detail for the first time.

In this study, two cell lines were tested for their ability to form viable EC cell chimaeras. These lines were derived from cloned cells and were PSA-l/NG-2 (obtained from Dr D. Martin) and PSA-4/TG12 (obtained from Dr M. Hooper). The parent lines PSA-1 and PSA-4 were isolated from the same tumour by Martin & Evans (1975 a). The tumour OTT5568 was induced by the transfer of a 3-day 129/SvSlCP embryo to the testis of an A₂G x 129/J F₁ male (Stevens, 1970). Both lines were resistant to thioguanine and thus were deficient in the activity of the enzyme hypoxanthine phosphoribosyltransferase (HPRT, IMP: pyrophosphase phosphoribosyltransferase, EC 2.4.2.8). PSA-l/NG-2 was mutagenized with nitrosoguanidine prior to selection (Dewey, Martin, Martin & Mintz, 1977) and PSA-4/TG12 was a spontaneous mutant (Slack, Morgan & Hooper, 1977). Both PSA-l/NG-2 and PSA-4/TG12 (hereafter referred to as NG-2 and TG12) were chromosomally aneuploid, having a trisomic chromosome no. 6, and also were XO (Cronmiller & Mintz, 1978; Hooper, personal communication).

NG-2 and TG12 were routinely maintained on mitomycin treated STO feeder cell layers (Martin & Evans, 1975b) in alpha medium (Stanners, Eliceiri & Green, 1971) without added nucleosides or deoxynucleosides, containing 10 % (v/v) heat-inactivated foetal calf serum (FCS) (Flow Laboratories, Irvine, Scotland) under a 10 % (v/v) CO₂ in air gas mixture. The dishes of feeder cells were prepared as follows: 50 or 35 mm diameter plastic tissue culture dishes (Sterilin Ltd, Richmond, Surrey, U.K.) were covered with a 1 % (w/v) gelatin (Swine skin, type 1, Sigma Chemical Co., U.K.) solution and left at 4 °C for 30 min. Subsequently the gelatin was decanted and the dishes were covered to confluency with STO feeder cells (Ware & Axelrad, 1972) which had previously been treated with mitomycin C to prevent their growth (Sigma Chemical Co., U.K.); (technique described by Martin & Evans, 1975b, and McBurney, 1976). The EC cells grew as clumps on the surface of the feeder cells and the culture medium was changed every day.

A number of different strains of mice were used as the source for the embryos. The strain used in each experiment depended on the particular intention of the experiment.

Two genetic markers were used to distinguish the EC cells’ progeny from those of the host embryos. These were two allozymal variants of the glucose phosphate isomerase Gpi-1 locus and albino or pigmentation in the coat colour (Staats, 1980). The NG-2 and TG12 EC lines were homozygous for the Gpi-1^a allele. Thus, the host embryos used in experiments with these lines were homozygous for the Gpi-1^b allele and were either (CBA × C57BL/6) F₂ embryos derived from (CBA × C57BL/6) F₁ matings or they were PO (Pathology, Oxford) or MFI (Olac) mice. The PO and MFI strains were outbred and thus individuals homozygous for the Gpi-1^b allele were selected from an initially random-bred stock. The F₂ embryos were used in experiments where the experiment was to be concluded before birth and the PO and MFI embryos were used in experiments where the embryos were allowed to develop to term.

The basic procedure was identical to that already described by Stewart (1980), although there were changes in preparing the EC cells prior to aggregation. The first series of experiments was performed with NG-2 EC cells. This line was chosen, since it has the highest reported rate of chimaerism by microinjection (Papaioannou, 1979), and thus it provided a good test of the efficiency of aggregation.

The clumps of EC cells were picked from the feeder layers, using a mouth-controlled micropipette. Care was taken to minimize fibroblast feeder cell contamination. These clumps were briefly washed in phosphate buffered saline (PBS solution A of Dulbecco & Vogt, 1954) and then incubated for 5 min at 37 °C in 0·125 % (w/v) trypsin (Difco Ltd, U.K.) and 0·2 mM ethylene diamine tetraacetic acid (EDTA) in PBS. Dissociation of the clumps of EC cells into single cells or small groups was completed by gentle pipetting. Groups consisting of 3–5 EC cells were then picked and placed into microdrops of prewarmed tissue culture medium, maintained under paraffin oil (Boots Pure Drug Co., U.K.) where they were allowed to recover for 2 h at 37 °C. In later experiments, an important modification to the method of dissociating the clumps of EC cells was introduced. This involved incubating the clumps of EC cells in 0·5 mM ethylene glycol bis tetraacetic acid (EGTA) in PBS for 30 min at 37 °C rather than using EDTA-trypsin. All subsequent parts of the procedure were identical.

The host embryos were dissected from the oviducts into prewarmed Whitten’s medium (Whitten, 1971), the zona pellucidae were removed by brief exposure of the embryos to acidified Tyrodes solution (Nicolson, Yanagimachi & Yanagimaçhi, 1975) and. then they were cultured in prewarmed, preequilibrated microdrops of Whitten’s medium under paraffin oil.

For each aggregation, two embryos were placed into a single microdrop of Whitten’s medium supplemented with 10 % (v/v) heat-inactivated FCS. A single group of 3–5 EC cells was then sandwiched between the embryos, making sure that the EC cells remained sticking to the embryos (Fig. 1). In a modification of this experiment, a single group containing the same number of EC cells was aggregated with a single embryo. This experiment was only performed with EGTA-dissociated NG-2 cells.

Once all the embryos and EC cells remained sticking together, the aggregates were cultured for 24 h, i.e. to the following afternoon ( = to the 4th day of gestation; day 1 = day of plug). The aggregates were then transferred to the uteri of (CBA × C57BL/6) F₁) foster mothers that had been mated with vasectomized males of proven sterility. Whenever possible, controls consisting of two embryos aggregated together were transferred to the contralateral horn of the recipient. The aggregates were either allowed to develop to term or they were dissected from the uteri on the 12th or 13th days of gestation. The conceptuses recovered were dissected into the following tissues: the trophoblast, parietal endoderm, embryo, amnion, yolk-sac mesoderm and yolk-sac endoderm. The yolk sacs were separated into their mesoderm and endoderm components by first rinsing in PBS and then incubating in a mixture of 0·5 % trypsin and 2·5 % pancreatin (Difco) in calcium- and magnesium-free Tyrodes saline at pH 7·7 for a minimum of 4 h at 4 °C (Levak-Svajger, Svajger & Skreb, 1969). Thereafter, they were transferred to PBS at room temperature and the tissues separated using forceps.

All of these tissues, as were the adult tissues, were screened for chimaerism by running aqueous extracts of tissues on starch gel electrophoresis, to separate out the allozymal GPI variants (Chapman, Whitten & Ruddle, 1971). The technique used was sensitive enough to detect a minimum contribution of 5 % of the minor band.

The ability of two EC cell lines to aggregate with 8-to 16-cell embryos and form viable chimaeras after transfer to foster mothers has been investigated. Only one of the lines, NG-2, has been reported to colonize embryos and produce chimaeras (Dewey et al. 1977), whereas the other line, TGI 2, has not until now been investigated.

The results of all the aggregation experiments are presented in Table 1. In the preliminary series of experiments performed with trypsin-EDTA dissociated NG-2 cells, 103 aggregates were transferred and 61 conceptuses were recovered (only in those aggregates that became pregnant were the number of transferred aggregates counted). Of these 61, 10 were found to be chimaeric solely in the embryonic fraction (Table 2). None of the 61 conceptuses exhibited any colonization in the yolk sac or amnion and in none of the 37 conceptuses from which the trophoblast and parietal endoderm had also been dissected, were these tissues colonized.

Of the 46 MFl^bb ⟷ NG-2 aggregates that were transferred and allowed to develop to term, 20 were born. Two were overtly chimaeric and showed extensive EC derived pigmentation in the coat. One of these also contained pigment in the eyes and both exhibited chimaerism in some internal organs (Table 3). No evidence for an EC derived cell population was found in the remaining 18 offspring.

These results were encouraging. However, the rate of chimaerism was less than that reported for the microinjection experiments performed with the same cell line (Dewey et al. 1977). Thus, a variety of methods were attempted to try to improve the rate of chimaerism. Eventually, substitution of trypsin-EDTA by 0·5 mM EGTA in PBS was found to be the simplest and most effective (Table 1).

Disaggregation of the NG-2 cells with EGTA resulted in an increase in the rate of chimaerism in the conceptuses (from 16 to 37 %). However, this was due to colonization of the extraembryonic membranes, in particular the amnion and yolk sac rather than the embryo proper. An increase in the percentage of NG-2 chimaeras born did also occur (12·5 to 25 %) (Table 1). Similar results were also obtained with the TG12 EC cell line where the yolk-sac mesoderm was colonized at the highest frequency (17 out of 24 chimaeras). In this series of experiments, using TG12 EC cells, the rate of chimaerism was relatively high (24 out of the 33 conceptuses recovered), although only 7 of these were chimaeric in the embryonic fraction; a figure that was almost identical to the percentage of live MFI ⟷TG12 chimaeras that were born (Table 1).

A series of experiments were also conducted, in which a single embryo was aggregated with a group of 3–5 NG-2 EC cells, and although the rate of chimaerism was high, 7 out of the 8 conceptuses that were chimaeric were severely malformed (Table 1).

A total of 10 viable NG-2 and TGI 2 chimaeric offspring were born (Figs. 2 and 3). The majority exhibited coat colour pigmentation, with mouse no. 4 (Fig. 2) being the most extensive. Analysis of the internal tissues of the individuals revealed some to be chimaeric in many tissues and in others the chimaerism was confined to one or two organs (Table 3). No tumours were found in any of these chimaeras, although a histological examination of the chimaeric tissues was not undertaken. However, the presence of pigmentation in the coat and in the eyes of some individuals, and with two individuals chimaeric in the skeletal muscle, where the intermediate Gpi- 1^ab band was present indicating fusion between myoblasts from EC cells and embryonic cells (Mintz & Baker, 1967), was evidence for the EC cells having undergone the appropriate differentiation in the tissues they colonized. Two of NG-2 chimaeras (numbers 1 and 4, Table 3) were test mated, but neither produced any offspring. Mouse no. 1 was autopsied at the age of 17 weeks, after he had been caged with a total of 10 females, none of which became pregnant. His testes and epididymis were examined histologically and although morphologically normal sperm were present, many of the tubules were devoid of sperm and were packed with many round nucleated and often necrotic cells, indicating abnormal spermatogenesis. Mouse no. 4 was caged with seven successive males, all of known fertility. She did not become pregnant and was autopsied at 27 weeks. Her uteri appeared to be normal, although her ovaries were small and yellow.

The experiments described here have demonstrated that the aggregation technique is a simple but effective alternative to microinjection, for producing EC cell chimaeras. It is possible that the method used for introducing the EC cells into the embryos may have had an effect on the EC cell distribution in the chimaeras. However, a comparison between the adult NG-2 chimaeras produced here and those produced using the microinjection technique (Dewey et al. 1977) revealed few differences, suggesting that the technique does not influence the distribution of the EC cells in adult chimaeras, except that in those adults produced by aggregation, no colonization of the lungs, thymus, spleen, gut or gonads was detected, although these tissues may be found to be chimaeric if a larger number of chimaeras was studied. Both methods did, however, produce individuals in which there was colonization of many tissues, and others in which an EC cell presence was confined to a few tissues.

Little information, apart from the study of Papaioannou et al. (1979), has been published on the pattern of EC cell colonization in conceptuses recovered from in utero. Thus, the results presented here were the first detailed description of the pattern of in utero chimaerism of EC cells in conceptuses. The two most striking observations were the relatively high rate of colonization by the EC cells, especially TG12, of extraembryonic tissues either completely derived from the mesoderm (the yolk-sac mesoderm) or containing a high proportion of mesodermal tissue (the amnion) (Rossant & Papaioannou, 1977). The second striking observation was the relatively low rate of colonization of the endodermal tissues of the conceptuses. No contribution by the EC cells was found in the parietal endoderm and only six individuals (two of them including conceptuses with extensive malformations) were chimaeric in the visceral yolk-sac endoderm.

The precise reasons for such a pattern are unknown, since these cells in vivo have not mimicked the pattern of differentiation that occurs in vitro. In vitro these cell lines, when permitted to differentiate, usually produce, at first, a mixture of cell types with the morphological and biochemical characteristics of visceral and parietal endoderm of the conceptus (Martin & Evans, 1975 b; Adamson, Gaunt & Graham, 1979; reviewed Graham, 1977; and Martin, 1978). However, in vivo these cells, on the basis of the results presented here, only colonized these tissues at a very low rate, and the reasons for this are difficult to explain. Possibly, due to the EC cells’ abnormal karyotype, selective pressures exerted by the embryonic tissues restricted the EC cell differentiation (Graham, 1977). Alternatively, the relatively low rate of colonization of the extraembryonic endoderm could be explained on the basis of the results from Gardner & Rossant’s (1979) investigation of the inner cell mass (ICM) from 5th day blastocysts. They showed that primitive endoderm and primitive ectoderm isolated from 5th day ICM’s were already committed tissues, in that they produced two distinct distributions when reinjected into 4th day blastocysts. The primitive endoderm exclusively colonized the extraembryonic endoderm of the yolk sac, whereas the primitive ectoderm colonized the yolk-sac mesoderm and embryo. Thus, it may have been possible that EC cells maintained for prolonged periods in vitro required time to adjust to an embryonic environment, i.e. inside the ICM, and this prevented them being exposed to the appropriate conditions permitting endoderm formation.

Apart from again invoking the argument of selective pressures, there is no explanation for the very high tendency, especially of the TG12 cells, to colonize the mesoderm in chimaeras, since there is no report of these cells predominantly producing mesodermal cells when undergoing differentiation in vitro. The high frequency of the TG12 cells colonizing the yolk-sac mesoderm may have resulted in a high proportion of the adults being chimaeric in the blood, since it is the yolk-sac mesoderm that the first blood cells are formed (Metcalf & Moore, 1971). However, within the haematopoietic system, cells lacking the HGPRT enzyme are known to be selected against (Kelley & Wyngaarden, 1978) and this may have accounted for the absence of haematopoietic cells, derived from the TG12 cells being present in the adults. Therefore, experiments are being performed to discover if the wild-type parent line, PSA-4, does have a similar tendency to colonize the yolk-sac mesoderm at high frequency and hence the haematopoietic cells of the adult.

I would like to thank Drs C. F. Graham and V. E. Papaioannou for their advice during the time this work was being performed. The author was supported by an M.R.C. research studentship.