Primordial germ cells of the mouse cultured on feeder layers with leukemia inhibitory factor, Steel factor and basic fibroblast growth factor give rise to cells that resemble undifferentiated blastocyst-derived embryonic stem cells. These primordial germ cell-derived embryonic germ cells can be induced to differentiate extensively in culture, form teratocarcinomas when injected into nude mice and contribute to chimeras when injected into host blastocysts. Here, we report the derivation of multiple embryonic germ cell lines from 8.5 days post coitum embryos of C57BL/6 inbred mice. Four independent embryonic germ cell lines with normal male karyotypes have formed chimeras when injected into BALB/c host blastocysts and two of these lines have transmitted coat color markers through the germline. We also show that pluripotent cell lines capable of forming teratocarcinomas and coat color chimeras can be established from primordial germ cells of 8.0 days p.c. embryos and 12.5 days p.c. genital ridges. We have examined the methylation status of the putative imprinting box of the insulin-like growth factor type 2 receptor gene (Igf2r) in these embryonic germ cell lines. No correlation was found between methylation pattern and germline competence. A significant difference was observed between embryonic stem cell and embryonic germ cell lines in their ability to maintain the methylation imprint of the Igf2r gene in culture. This may illustrate a fundamental difference between these two cell types.

Cells that give rise to the primordial germ cells (PGCs) of the mouse are first identified by lineage tracing as a population of cells within the proximal margin of the 6.5 days post coitum (p.c.) epiblast, close to the junction with the extraembryonic ectoderm (Lawson and Hage, 1994). By 7.5 days p.c. PGCs have migrated to the extraembryonic mesoderm at the posterior of the primitive streak near the base of the allantois where they can be identified by staining for endogenous alkaline phosphatase (AP) (Ginsburg et al., 1990). The PGCs then become associated with the hindgut endoderm and migrate through the gut mesentery and begin to arrive at the genital ridges by 10.5 days p.c. During the process of moving from the posterior primitive streak to the genital ridges PGCs increase in number from about 150 cells at 8.5 days p.c. to approximately 25,000 at 13.5 days p.c. (Tam and Snow, 1981). By 13.5 days p.c. PGCs within the genital ridge cease dividing; those in the female enter meiosis and those in the male undergo mitotic arrest (Ginsburg et al., 1990).

Given the importance of PGCs to the continuation of the species, it is of interest to define the in vivo factors that control their proliferation, survival and migration. Progress has come from the study of mouse mutants, such as Dominant White-spotting (W) and Steel (Sl), in which PGCs do not follow their normal tightly regulated pattern of proliferation and differentiation, resulting in severe defects in fertility in homozygotes (Mintz and Russell, 1957; McCoshen and McCallion, 1975). The finding that W and Sl encode, respectively, the c-kit tyrosine kinase transmembrane receptor (Chabot et al., 1988; Geissler et al., 1988) and its ligand, Steel factor (SF, also known as stem cell factor, mast cell growth factor and c-kit-ligand), (Williams et al., 1990; Flanagan and Leder, 1990; Zsebo et al., 1990; Huang et al., 1990,) suggests that polypeptide growth factors, their receptors and downstream intracellular signalling pathways play key roles in regulating the proliferation, migration and differentiation of PGCs.

Previous work from our laboratory (Matsui et al., 1991, 1992) and others (Dolci et al., 1991; Godin et al., 1991; Resnick et al., 1992) has shown that specific soluble and membrane-associated polypeptide growth factors are necessary for PGCs to survive and proliferate in culture. When PGCs from 8.5 days p.c. embryos are cultured with a mixture of membrane associated and soluble SF, leukemia inhibitory factor (LIF) and basic fibroblast growth factor (bFGF), PGCs continue to divide in culture and give rise to lines of undifferentiated cells. These cells have been termed embryonic germ cells or EG cells to distinguish them from blastocyst-derived embryonic stem (ES) cells (Resnick et al., 1992). Once established, these EG cells can proliferate without added FGF and SF, but they still require feeder cells and serum factors. EG cells can differentiate into embryoid bodies in vitro and form teratocarcinomas in vivo. Recent studies have shown that EG cell lines derived from 129/Sv 8.5 days p.c. PGCs can con-tribute to the germline of chimeras (Stewart et al., 1994) and we show here that the same is true for EG cell lines derived from C57BL/6 PGCs.

One of the events associated with germ cell differentiation is a change in the methylation status of parentally imprinted genes (Barlow, 1993). Because it is not feasible to isolate large numbers of primordial germ cells from very early embryos, we have studied EG cell lines established from PGCs at various times during development in the hope of gaining some information about the imprinting process in vivo. We have analyzed the methylation state of the putative imprinting box of the maternally expressed insulin-like growth factor 2 receptor (Igf2r) gene (Stöger et al., 1993) in different EG cell lines and show that the methylation of region 2 of the Igf2r gene in most EG cell lines differs from that characteristic of normal somatic cells and ES cells. Moreover, in one cell line this methylation is lost in culture. However, there appears to be no correlation between the methylation pattern and the ability of the cells to contribute to the germline of chimeric mice. This illustrates a difference between ES cells and EG cells, a distinction that may reflect the difference in the developmental program of an inner cell mass cell versus a PGC.

Mouse embryos and cell lines

C57BL/6, BALB/c, Swiss Webster, ICR, (C57BL/6 × DBA)F1 and nude mice were obtained from Jackson Labs, Harlan Labs or Taconic Farms. Noon on the day of plug is 0.5 days post coitum (p.c.).

The ES cell lines tested were all derived from 129/Sv mice. Cell lines D3, J1 and R1 were kindly provided by Drs Thomas Doetschman, Rudolph Jaenisch and Janet Rossant, respectively. SC6 was derived from AB-1 cells obtained from Dr Allan Bradley. TL1 was derived in this laboratory.

Tumors were initiated in nude mice by the subcutaneous injection of 1×106 cells at each of four locations on the back. Tumors were dissected and processed for histological analysis as described (Matsui et al., 1992).

PGC culture

Cultures were initiated as described (Matsui et al., 1992) by dissecting C57BL/6 8.5 days p.c. embryos free of extraembryonic tissues. Fragments comprising the posterior third of the embryo (from the base of the allantois to the first somite) were then pooled, rinsed with Dulbecco’s Ca2+-, Mg2+-free phosphate-buffered saline (PBS) and dissociated with 0.25% trypsin, 1 mM EDTA (GIBCO) and gentle pipetting. This single cell suspension is then plated in 0.1% gelatincoated 24-well dishes (Corning) with irradiated Sl/Sl4 m220 cells as feeder layers at a concentration of approximately 0.5 embryo equivalents per well. The cultures were grown in Dulbecco’s modified Eagle’s medium (DMEM) (Specialty Media, Lavallette, NJ) supplemented with 4.5 g/l glucose, 0.01 mM non-essential amino acids (GIBCO), 2 mM glutamine (GIBCO), 50 μg/ml gentamycin (Sigma), 15% fetal bovine serum (selected batches, Hyclone) and 0.1 mM 2-mercaptoethanol (Sigma). For these primary cultures, the medium was additionally supplemented with soluble recombinant rat SF at 60 ng/ml, bFGF at 20 ng/ml (GIBCO) and LIF at 20 ng/ml. After 6 days some of the cultures are stained for alkaline phosphatase (AP) as described (Matsui et al., 1992) in order to assess the survival and proliferation of PGCs. After 10 days, parallel cultures are dissociated into single cells and plated onto mouse embryo fibroblast (mef) feeder layers with LIF (ESGRO, GIBCO 1000 U/ml). These cultures were monitored for the appearance of colonies of EG cells. Individual EG colonies were isolated with a micropipette and lines established. EG cultures were then maintained in the same manner as ES cell lines with irradiated mefs as feeder cells and LIF (Smith et al., 1988; Williams et al., 1988). Cultures from 12.5 days p.c. genital ridges or 8.0 days p.c. embryos were initiated and maintained in a similar manner.

Blastocyst injection

All EG cell lines were karyotyped (Robertson, 1987) and only lines with 80% or more of the cells containing 40 chromosomes were used for blastocyst injection. Ten to twenty EG cells at passage numbers 6 to 10 were injected into 3.5 days p.c. blastocysts from BALB/c mice. Foster mothers were (C57BL/6 × DBA)F1 females mated to vasec-tomized Swiss Webster males. Injected blastocysts were transferred to the uterus of 2.5 days p.c. foster mothers (Hogan et al., 1986) and chimeric pups were identified by their coat color. Chimeras were bred to either BALB/c or ICR mice and germline transmission was judged on the day of birth by the presence of eye pigment.

Southern analysis

EG cell lines from passage number 6 to 10 were cultured without feeder layers on gelatin-coated 6 cm tissue culture dishes (Corning) for at least two passages in order to remove contaminating feeder cells while differentiation was inhibited by the continued presence of LIF in the medium. Cultures were dissociated with 0.25% trypsin, 1 mM EDTA, washed 3× with Dulbecco’s Ca2+-, Mg2+-free phosphate-buffered saline (PBS) and lysed in TENS (100 mM EDTA, 50 mM Tris-Hcl pH8.0, 100 mM NaCl) with 1% SDS. The lysate was incubated overnight at 55°C with 0.5 mg/ml Proteinase K. The lysates were extracted first with phenol, then with phenol:chloroform:isoamyl alcohol (25:24:1) and then with chlorform:isoamyl alcohol (24:1). Genomic DNA was ethanol precipitated and spooled onto a capillary tube. Southern blots were performed as described (Stöger et al., 1993). Briefly, DNA was dissolved in 10 mM Tris-HCl pH 7.6, 1 mM EDTA (TE), digested with the restriction enzymes HpaII and PvuII, and separated on a 0.8% agarose gel. Southern blots were generated by standard methods (Sambrook et al., 1989) and filters were hybridized with an intron-specific probe covering region 2 of the Igf2r receptor (pPP4, see Stöger et al., 1993). A probe from region 1 of the Igf2r gene was used to check for complete digestion of genomic DNA (pXIgf2r, see Stöger et al., 1993). All Southern blots were repeated with at least two independent restriction digests of the genomic DNA.

PCR for Zfy

DNA from EG cell lines was analyzed by PCR for the presence of a Y chromosome. Primers used were 5′-AAGATAAGCTTA-CATAATCACATGGA and 3′-CCTATGAAATCCTTTGCTGCA-CATGT (Nagamine et al., 1989). PCR reactions were cycled at 94°C for 45 seconds, 62°C for 25 seconds and 72°C for 1 minute for 30 cycles. The presence of a Y chromosome was indicated by a 600 base pair reaction product.

Skeletal preparations

Skeletal preparations were performed essentially as described (McLeod, 1980). Briefly, newborn mice were first killed using Metofane, then incubated in distilled water at room temperature overnight, scalded with hot water for thirty seconds, skinned and eviscerated. Cartilage was stained with Alcian Blue and bone counter-stained with Alizerin Red. Skeletons were cleared with KOH and glycerin then stored in glycerin.

EG cell lines derived from C57BL/6 8.5 days p.c. PGCs can contribute to the germline of chimeric mice

Initial experiments from our laboratory to derive EG cells in culture used embryos from matings of ICR random bred females with (C57BL/6 × DBA)F1 males (Matsui et al., 1991, 1992). Here, we have used the inbred mouse strain C57BL/6 to control for any genetic differences that may affect the derivation of EG cell lines. C57BL/6 blastocysts have been used previously to generate totipotent ES lines (Ledermann and Burki, 1991).

EG cell lines were first obtained from 8.5 days p.c. C57BL/6 embryos and early passage EG cells (passage number 6-10) were used for both blastocyst injection and genomic DNA isolation. Cell lines were also tested for the presence of a Y chromosome by performing PCR with primers specific for Zfy (see Methods). All cell lines were karyotyped and male cell lines with a normal karyotype were injected into BALB/c blastocysts. Table 1 shows the number of chimeras obtained from four of these EG lines. Male and female chimeras were bred to either BALB/c or ICR mice to test for the ability of the cells to differentiate into functional germ cells. The results of these matings are shown in Table 2. Male chimeras from two independent cell lines (TGC8.510 and TGC8.519) were able to transmit the C57BL/6 genome through the germline (Table 2; Fig. 1). No female chimeras gave birth to pigmented pups.

Table 1.

Chimeras from blastocyst injection of EG cell lines

Chimeras from blastocyst injection of EG cell lines
Chimeras from blastocyst injection of EG cell lines
Table 2.

Germline transmission of the EG cell lines TGC8.510 and TGC8.519

Germline transmission of the EG cell lines TGC8.510 and TGC8.519
Germline transmission of the EG cell lines TGC8.510 and TGC8.519
Fig. 1.

Germline transmission of TGC8.510. Male chimera from EG cell line TGC8.510, ICR random bred female, and pups from two separate litters. Pigmented pups (agouti and non-agouti) illustrate germline transmission of the C57BL/6 derived EG cell line.

Fig. 1.

Germline transmission of TGC8.510. Male chimera from EG cell line TGC8.510, ICR random bred female, and pups from two separate litters. Pigmented pups (agouti and non-agouti) illustrate germline transmission of the C57BL/6 derived EG cell line.

Generation of EG cell lines from gonadal PGCs

To ask whether PGCs become restricted in their ability to form EG cell lines after entering the gonad, we repeated these experiments with germ cells from later stage C57BL/6 embryos. Germ cells from the gonads of 15.5 days p.c. embryos and newborn mice did not give rise to EG cell lines under the conditions used previously. However, we were able to derive EG cell lines from the genital ridges of 12.5 days p.c embryos, at a time when most of the PGCs have migrated into them. At 12.5 days p.c. it is possible to identify the sex of the genital ridge by its morphology, so male and female genital ridges were pooled separately. The frequency of EG cell line formation is much lower with genital ridges from 12.5 days p.c. embryos compared to posterior tissue of 8.5 days p.c. embryos. From several separate experiments, ten cell lines were obtained from the equivalent of 6/100 of a genital ridge (approximately 432 PGCs, as calculated from Tam and Snow, 1981) as compared to more than 20 from the equivalent of the posterior portion of 1/6 of an 8.5 days p.c. embryo (approximately 24 PGCs, as calculated from Tam and Snow, 1981). All ten cell lines were derived from male gonads; no cell lines were obtained from female gonads. Two karyotypically normal cell lines were tested and shown to form differentiated tumors in nude mice and contribute to coat color chimeras. Table 1 shows the frequency of chimerism for these two EG cell lines (TGC12.51 and TGC12.52). The chimeras did not show a bias toward maleness; of the 9 surviving adult chimeras from both cell lines, there are 6 females and 3 males. These were bred to test for germline transmission but no pigmented pups resulted from these matings. Two of eleven identified chimeras from these two cell lines exhibited skeletal abnormalities and died soon after birth (within one to two days). The skeleton of one of these two newborns exhibited malformations of the rib cage and sternum (data not shown).

Imprinting of the Igf2 receptor gene

In the somatic tissues of the adult mouse, the Igf2r gene is parentally imprinted with expression only from the maternal allele (Barlow et al., 1991). Recent work has identified a specific region in the Igf2r gene, region 2, within the second intron 27 kilobases downstream of the transcriptional start site of the gene, that is hypermethylated only on the maternally inherited allele (Stöger et al., 1993). This methylation pattern may be one factor regulating the transcription of Igf2r, as mice that possess two mutated copies of the methyltransferase gene display an altered methylation pattern in this region as well as downregulated expression of Igf2r (Li et al., 1993). Fig. 2 shows the methylation pattern of Igf2r region 2 for some of the EG cell lines tested. Half of the 8.5 days p.c. derived EG cell lines (9 out of 18), including TGC 10, one of the two cell lines to transmit through the germline, show both a 3 kilobase and a 500 base pair hybridizing fragment implying that they carry one methylated and one non-methylated allele (Fig. 2C,D). This pattern of methylation is characteristic of somatic cells (Fig. 2A, mefs) and 5 different pluripotent ES cell lines (Fig. 2A for D3 and R1, and data not shown for J1, SC6 and TL1. The remaining cell lines, including TGC8.519, the other totipotent cell line, have a different pattern of methylation in which both alleles are unmethylated (Fig. 2C, D; Table 3) and show only a 500 base pair hybridizing fragment.

Table 3.

The methylation status and sex of all EG cell lines

The methylation status and sex of all EG cell lines
The methylation status and sex of all EG cell lines
Fig. 2.

Methylation status of region 2 of the Igf2r gene. Southern blots of genomic DNA from different cell lines digested with restriction enzymes HpaII and PvuII and hybridized with a probe for region 2 of the Igf2r gene. (A) D3 and R1, two different ES cell lines, and mouse embryo fibroblasts (mef). (B) EG cell lines derived from 8.0 days p.c. primordial germ cells. (C) EG cell lines derived from 8.5 days. p.c., primordial germ cells. (D) The two EG cell lines derived from 8.5 days p.c. primordial germ cells which transmitted through the germline. (E) EG cell lines derived from 12.5 days p.c. genital ridges.

Fig. 2.

Methylation status of region 2 of the Igf2r gene. Southern blots of genomic DNA from different cell lines digested with restriction enzymes HpaII and PvuII and hybridized with a probe for region 2 of the Igf2r gene. (A) D3 and R1, two different ES cell lines, and mouse embryo fibroblasts (mef). (B) EG cell lines derived from 8.0 days p.c. primordial germ cells. (C) EG cell lines derived from 8.5 days. p.c., primordial germ cells. (D) The two EG cell lines derived from 8.5 days p.c. primordial germ cells which transmitted through the germline. (E) EG cell lines derived from 12.5 days p.c. genital ridges.

In order to evaluate the methylation pattern of Igf2r region 2 during PGC maturation, EG cell lines were established from PGCs at different times. EG cell lines from earlier PGCs (8.0 days p.c.) were generated and these cell lines were assayed for their methylation pattern. The results were similar to those with the EG cell lines derived from 8.5 days p.c. PGCs. Some cell lines (10 out of 18) showed both alleles to be unmethylated (for example TGC8.09, TGC8.010, TGC8.014, TGC8.015 and TGC8.016 in Fig. 2B; Table 3) whereas others (8 out of 18) showed a methylated and an unmethylated allele (for example TGC8.012 in Fig. 2B; Table 3). All ten EG cell lines derived from 12.5 days p.c. genital ridges showed an unmethylated pattern for region 2 (Fig. 2E).

Occasionally, when longer exposures of Southern blots were examined, some cell lines showed a faint 3 kb hybridizing fragment that was much less intense than the 3 kb fragment of either ES cells or fibroblasts. We interpret this as partial methylation of region 2, due to heterogeneity in the cultures. A summary of the methylation pattern of region 2 and the sex of the cell lines is presented in Table 3. There appears to be no correlation between the methylation pattern and the sex of the cells.

In order to test the possibility that the methylation pattern of region 2 of the Igf2r could change as ES and EG cell lines are grown in culture, several lines were cultured to late passage numbers (between 17 and 34). We have cultured ES cells to later passages (p=17) and have never seen alteration of the normal somatic pattern of methylation. Two of the three EG cell lines tested (TGC8.519 and TGC12.58) had both alleles of Igf2r unmethylated at early passage. As shown in Fig. 3, both of these cell lines maintained this methylation pattern in culture, even as late as passage 34. In contrast, TGC8.55, which had partial methylation at early passages (p=7, 11, 15), lost this methylation and the 3 kb hybridizing fragment by late passage numbers (p=20, 26, 29).

Fig. 3.

Temporal changes in methylation status of region 2 of the Igf2r gene in EG cell lines. Southern blots of genomic DNA from different cell lines digested with restriction enzymes HpaII and PvuII and hybridized with a probe for region 2 of the Igf2r gene. (A) TGC8.55 at passage numbers 7, 11, 15, 20, 29; (B) TGC8.519 at passage numbers 8, 17, 20, 26, 30; (C) TGC12.58 at passage numbers 8, 16, 21, 25, 31, 34.

Fig. 3.

Temporal changes in methylation status of region 2 of the Igf2r gene in EG cell lines. Southern blots of genomic DNA from different cell lines digested with restriction enzymes HpaII and PvuII and hybridized with a probe for region 2 of the Igf2r gene. (A) TGC8.55 at passage numbers 7, 11, 15, 20, 29; (B) TGC8.519 at passage numbers 8, 17, 20, 26, 30; (C) TGC12.58 at passage numbers 8, 16, 21, 25, 31, 34.

Derivation of EG cell lines by treating PGCs with growth factors

Previous work from our laboratory (Matsui et al., 1991, 1992; Labosky et al., 1994) and others (Dolci et al., 1991; Godin et al., 1991; Resnick et al., 1992) and the data reported here show that the combination of SF, LIF and bFGF allows the establishment of EG cell lines. The mechanism by which this is achieved is not clear. One possibility is that this combination of growth factors somehow alters the developmental program of the PGCs so that they now behave like the epiblast or inner mass cells from which they were derived. Alternatively, the combination of growth factors may block the terminal differentiation of the PGCs, arresting them as an immature or precursor

PGCs. At present the only difference that we have observed between EG cells and blastocyst-derived ES cells is the pattern of DNA methylation within region 2 of the Igf2r gene. Both cell types differentiate in vitro and in vivo, and both EG cells and ES cells can differentiate into functional sperm and thereby transmit their genome through the germline (Table 2; Matsui et al., 1992; Labosky et al., 1994; Stewart et al., 1994). Our results regarding the methylation pattern of region 2 of the Igf2r gene support the second alternative, that the EG cells resemble an immature germ cell rather than an inner cell mass cell, since most of the EG cell lines have a pattern of methylation more similar to that of late germ cells (Brandeis et al., 1993) than somatic cells and ES cells. Additionally, we have used our culture conditions with epiblasts from earlier embryos (6.5 days p.c.) and so far have been unable to establish stem cell lines, further supporting the theory that the EG cell lines derived from 8.0 and 8.5 days p.c. PGCs do not resemble epiblast cells.

Our data here suggest that PGCs from 8.5 days p.c. embryos and 12.5 days p.c. genital ridges are not equivalent in their response to the combination of SF, bFGF and LIF. Although pluripotent EG cell lines can be derived from both 8.5 and 12.5 days p.c. germ cells, the process is much more efficient with younger PGCs. This is illustrated by the observation that only ten cell lines were obtained from approximately four hundred 12.5 days p.c. PGCs while over twenty cell lines were obtained from an estimated twenty four 8.5 days p.c. PGCs. It appears that germ cells in 15.5 days p.c., newborn and 8 days postnatal gonads do not respond to this combination of growth factors since we have been unable to establish EG cell lines from these stages with C57BL/6 inbred mice. Germ cells become postmi-totic after they reach the genital ridge and it is possible that as they become quiescent they no longer respond to SF, LIF and bFGF perhaps because they downregulate one or more receptors for these factors. In support of this hypothesis, it has been shown (Manova and Bachvarova, 1991) that c-kit, the receptor for SF, is down regulated in PGCs shortly after they reach the genital ridge.

Our data raise the possibility that the EG cell lines derived from 12.5 days p.c. PGCs may be qualitatively different from the EG cell lines derived from 8.5 days p.c. PGCs. The EG cell lines derived from 12.5 days p.c. PGCs, while pluripotent, do not appear to be totipotent. A sex bias is not seen in chimeras after blastocyst injection and these resulting chimeras are less extensive than chimeras from EG cell lines derived from 8.5 days p.c. PGCs (as judged by coat color contribution) even when as many as 20 EG cells are used for each blastocyst injection (Table 2). Possible reasons for a qualitative difference are discussed below.

Imprinting of the Igf2r gene in EG cells

We have shown here that in approximately half of the EG cell lines derived from 8.0-8.5 days p.c. PGCs and all of the EG cell lines derived from 12.5 days p.c. genital ridges the methylation of region 2 of theIgf2r gene differs from normal somatic cells. Instead of the normal region 2 methylation pattern, with the maternal allele methylated and the paternal allelle unmethylated, these EG cell lines have both alleles unmethylated in the majority of the cells.

There are several possible explanations for these methylation differences. First, it may reflect the in vivo methylation status of the Igf2r gene in the founding PGCs. The mouse embryo undergoes massive general demethylation of genomic DNA at the 8-to 16-cell stage and a subsequent remethylation, which is completed by implantation (Monk et al., 1987). In germ cells, the general methylation changes are not identical to those seen in the rest of the embryo (Monk et al., 1987, Kafri et al., 1992). While the DNA in the somatic cells of the embryo has begun to be remethylated by 4.5 days p.c., germ cells remain relatively unmethylated at 12.5-13.5 days p.c.. By 15.5 days p.c. they are partially remethylated, with the process completed by 18.5 days p.c.. These observations on general demethylation of genomic DNA do not apply to region 2 of the Igf2r gene since the maternal-specific methylation of this region, directly inherited from the oocyte, is preserved through preimplantation (Stöger et al., 1993, Brandeis et al., 1993). However, region 2 does become unmethylated in both male and female germ cells by 13.5 days p.c. (Brandeis et al., 1993), implying that the genomic imprint for Igf2r is erased sometime between the blastocyst and 13.5 days p.c.. If the methylation pattern for Igf2r region 2 of our EG cell lines represents the methylation pattern of founding PGCs, the absence of methylation reported here in most EG cell lines reflects erasure of the Igf2r genomic imprint. Our results suggest that some PGCs have erased their imprint by 8.0-8.5 days p.c., with all having done so by 12.5 days p.c..

At present we cannot distinguish between an erased or an androgenetic imprint of the Igf2r gene (both alleles unmethy-lated), but our observations for the Igf2r gene may reflect a general absence or erasure of imprinting signals in the EG cells and not simply an androgenetic pattern. In support of this speculation, our results showing that EG cells from 8.5 days p.c. PGCs can form healthy chimeras and in some cases contribute to the germline illustrates that these EG cells differ from andro-genetic ES cells (Mann et al., 1990; Mann and Stewart, 1991). However, we cannot eliminate the possibility that some of the EG cells from gonadal 12.5 days p.c. PGCs have acquired an androgenetic imprint since some of the skeletal defects of the chimeras obtained with these EG cells resemble the defects of androgenetic ES cell chimeras, although they are less severe and not as prevalent.

An alternative explanation for the unmethylated pattern in the EG cells is that methylation of Igf2r region 2 is not stable during in vitro culture. We have shown here that the somatic pattern of methylation seen in blastocyst-derived ES cells remains stable for up to 17 passages in culture. In addition, the unmethylated pattern in EG cells remains stable in several different lines for up to 34 passages. However, one EG cell line (TGC8.55) that displayed methylation at early passage (p=7, 11, 15) lost this methylation by later passages (p=20, 26, 29).

This observation implies that the methylation differences between different EG cell lines may not pre-exist before the lines were established but instead are the result of a difference in their response to in vitro culture. It is also possible that the methylation is unstable in EG cells because of their origin from PGCs. This instability may reflect an in vivo difference between PGC-derived EG cell lines and inner cell massderived ES cell lines, reflecting the difference in the developmental program of the founding cells. It is also noteworthy that, once unmethylated, the methylation of region 2 is not regained, implying that this erased pattern may be the normal end result for a PGC as it reaches the gonad.

The difference in the methylation pattern of region 2 in TGC8.510 and TGC8.519 (Fig. 2D) implies that the imprinted status of the Igf2r gene in an EG cell line does not affect its ability to pass through the germline. This result is perhaps not surprising in view of the fact that, although parthenogenetic embryos cannot form normal mice (Stevens, 1978; Clarke et al., 1988; and Fundele et al., 1990), it has been observed that parthenogenetic ES cells can contribute to chimeras and form functional germ cells giving rise to normal offspring (J. Mann and C. Stewart, unpublished data).

In conclusion, the two hypotheses that we have proposed for the differences in methylation of region 2 of the Igf2r gene between ES cells and EG cells may not be mutually exclusive. It is possible that the methylation differences between the various early passage EG cell lines reflect the changing methylation status of the migrating PGCs. It is also possible that the heterogeneity and plasticity of this methylation pattern in vitro is another property of PGCs. PGCs are programmed to erase their imprint in vivo and may be carrying out that program in vitro. In the future, the identification of specific regions such as region 2 in the Igf2r gene that regulate other imprinted genes will allow the generalities of these ideas to be explored.

We should like to thank Babru Samal, AMGEN, Inc. for recombinant LIF and SF, Drs. Jeff Mann and Chris Wright for helpful comments on the manuscript, and Linda Hargett for her excellent technical assistance. P.A.L. is an Associate and B.L.M.H. an Investigator of the Howard Hughes Medical Institute.

Barlow
,
D. P.
,
Stöger
,
R.
,
Herrmann
,
B. G.
,
Saito
,
K.
and
Schweifer
,
N.
(
1991
).
The mouse insulin-like growth factor type-2 receptor is imprinted and closely linked to the Tme locus
.
Nature
349
,
84
87
.
Barlow
,
D. P.
(
1993
).
Methylation and imprinting: From host defense to gene regulation?
Science
260
,
309
310
.
Brandeis
,
M.
,
Kafri
,
T.
,
Ariel
,
M.
,
Chaillet
,
J. R.
,
McCarrey
,
J.
,
Razin
,
A.
and
Cedar
,
H.
(
1993
).
The ontogeny of allele-specific methylation associated with imprinted genes in the mouse
.
EMBO J
.
12
,
3669
3677
.
Chabot
,
B.
,
Stephenson
,
D. A.
,
Chapman
,
V. M.
,
Besmer
,
P.
and
Bernstein
,
A.
(
1988
).
The proto-oncogene c-kit encoding a transmembrane tyrosine kinase maps to the mouse W locus
.
Nature
335
,
88
89
Clarke
,
H. J.
,
Varmuza
,
S.
,
Prideaux
,
V. R.
and
Rossant
J.
(
1988
)
The developmental potential of partenogenetically derived cells in chimeric mouse embryos: implications for actions of imprinted genes
.
Development
104
,
175
182
.
Dolci
,
S.
,
Williams
,
D. E.
,
Ernst
,
M. K.
,
Resnick
,
J. L.
,
Brannan
,
C. I.
,
Lock
,
L. F.
,
Lyman
,
S. D.
,
Boswell
,
H. S.
and
Donovan
,
P. J.
(
1991
)
Requirement for mast cell growth factor for primordial germ cell survival in culture
.
Nature
352
,
809
811
.
Flanagan
,
J. G.
and
Leder
,
P.
(
1990
).
The kit ligand: A cell surface molecule altered in steel mutant fibroblasts
.
Cell
63
,
185
194
.
Fundele
,
R. H.
,
Norris
,
M. L.
,
Barton
,
S. C.
,
Fehlau
,
M.
,
Howlett
,
S. K.
,
Mills
,
W. E.
and
Surani
,
M. A.
(
1990
).
Temporal and spatial selection against parthenogenetic cells during development of fetal chimeras
.
Development
108
,
203
211
.
Geissler
,
E. N.
,
Ryan
,
M. A.
and
Housman
,
D. A.
(
1988
).
The dominantwhite spotting (W)locus of the mouse encodes the c-kit proto-oncogene
.
Cell
55
,
185
192
Ginsburg
,
M.
,
Snow
,
M. H. L.
and
McLaren
,
A.
(
1990
).
Primordial germ cells in the mouse embryo during gastrulation
.
Development
110
,
521
528
.
Godin
,
I.
,
Deed
,
R.
,
Cooke
,
J.
,
Zsebo
,
K.
,
Dexter
,
M.
and
Wylie
,
C.
(
1991
).
Effects of the steel product on mouse primordial germ cells in culture
.
Nature
352
,
807
809
.
Hogan
,
B.
,
Costantini
,
F.
and
Lacy
,
E.
(
1986
).
Manipulating the Mouse Embryo: A Laboratory Manual
.
Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press
.
Huang
,
E.
,
Nocka
,
K.
,
Beier
,
D. R.
,
Chu
,
T.-Y.
,
Buck
,
J.
,
Lahm
,
H. W.
,
Wellner
,
D.
,
Leder
,
P.
and
Besmer
,
P.
(
1990
)
The hematopoietic growth factor KL is encoded by the
Sl locus and is the ligand of the c-kit receptor, the gene product of the W locus. Cell
63
,
225
233
.
Kafri
,
T.
,
Ariel
,
M.
,
Brandeis
,
M.
,
Shemer
,
R.
,
Urven
,
L.
,
McCarrey
,
J.
,
Cedar
,
H.
and
Razin
,
A.
(
1992
).
Developmental pattern of gene-specific DNA methylation in the mouse embryo and germ line
.
Genes Dev
.
6
,
705711
.
Labosky
,
P. A.
,
Barlow
,
D. P.
and
Hogan
B. L. M.
(
1994
) Embryonic germ cell lines and their derivation from mouse primordial germ cells. In
Germline Development Ciba Foundation Symposium 182
, pp.
157
-
178
. Chichester:
Wiley
.
Lawson
,
K. A.
and
Hage
,
W. J.
(
1994
). Clonal analysis of the origin of primordial germ cells in the mouse. In
Germline Development Ciba Foundation Symposium 182, pp 68-91. Chichester: Wiley
.
Ledermann
,
B.
and
Burki
,
K.
(
1991
).
Establishment of a germ-line competent C57BL/6 embryonic stem cell line
.
Exp. Cell Res
.
197
,
254
258
.
Li
,
E.
,
Beard
,
C.
and
Jaenisch
,
R.
(
1993
).
Role for DNA methylation in genomic imprinting
.
Nature
366
,
362
364
.
Mann
,
J. R.
,
Gadi
,
I.
,
Harbison
,
M. L.
,
Abbondanzo
,
S. J.
and
Stewart
,
C. L.
(
1990
).
Androgenetic mouse embryonic stem cells are pluripotent and cause skeletal defects in chimeras: implications for genetic imprinting
.
Cell
62
,
251
260
Mann
,
J. R.
and
Stewart
,
C. L.
(
1991
).
Development to term of mouse androgenetic aggregation chimeras
.
Development
113
,
1325
1333
.
Manova
,
K.
and
Bachvarova
,
R. F.
(
1991
).
Expression of c-kit encoded at the W locus of mice in developing embryonic germ cells and presumptive melanoblasts
.
Dev. Biol
.
146
,
312
324
.
Matsui
,
Y.
,
Toksoz
,
D.
,
Nishikawa
,
S.
,
Nishikawa
,
S.-I.
,
Williams
,
D.
,
Zsebo
,
K.
and
Hogan
,
B. L. M.
(
1991
).
Effect of Steel factor and leukaemia inhibitory factor on murine primordial germ cells in culture
.
Nature
353
,
750
752
.
Matsui
,
Y.
,
Zsebo
,
K.
,
Hogan
,
B. L. M.
(
1992
).
Derivation of pluripotent embryonic stem cells from murine primordial germ cells in culture
.
Cell
70
,
841
847
.
McCoshen
,
J. A.
and
McCallion
,
D. J.
(
1975
).
A study of the primordial germ cells during their migratory phase in Steel mutant mice
.
Experientia
31
,
589590
.
McLeod
,
M. J.
(
1980
).
Differential staining of cartilage and bone in whole tissue fetuses by alcian blue and alizarin red. S
.
Teratology
22
,
299
301
.
Mintz
,
B.
and
Russell
,
E. S.
(
1957
)
Gene-induced embryological modifications of primordial germ cells in the mouse
.
J. Exp. Zool
.
134
,
207
237
.
Monk
,
M.
,
Boubelik
,
M.
and
Lehnert
,
S.
(
1987
).
Temporal and regional changes in DNA methylation in the embryonic, extraembryonic and germ cell lineages during mouse embryo development
.
Development
99
,
371
382
.
Nagamine
,
C. M.
,
Chan
,
K.
,
Kozak
,
C. A.
and
Lau
,
Y.
(
1989
).
Chromosome mapping and expression of a putative testis-determining gene in mouse
.
Science
243
,
80
83
.
Resnick
,
J. L.
,
Bixler
,
L. S.
,
Cheng
,
L.
and
Donovan
,
P. J.
(
1992
).
Long-term proliferation of mouse primordial germ cells in culture
.
Nature
359
,
550
551
.
Robertson
,
E. J.
(
1987
). Embryo-derived stem cell lines. In
Teratocarcinomas and Embryonic Stem Cells
. (ed.
E. J.
Robertson
), pp.
71
112
.
Oxford
:
IRL Press
.
Sambrook
,
J.
,
Fritsch
,
E. F.
and
Maniatis
,
T.
(
1989
).
Molecular Cloning: A Laboratory Manual
.
Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press
.
Smith
,
A. G.
,
Heath
,
J. K.
,
Donaldson
,
D. D.
,
Wong
,
G. G.
,
Moreau
,
J.
,
Stahl
,
M.
and
Rodgers
,
D.
(
1988
).
Inhibition of pluripotential embryonic stem cell differentiation by purified peptides
.
Nature
336
,
688
690
.
Stevens
,
L. C.
(
1978
).
Totipotent cells of parthenogenetic origin in a chimeric mouse
.
Nature
276
,
266
267
.
Stewart
,
C. L.
,
Gadi
,
I.
and
Blatt
,
H.
(
1994
).
Stem cells from primordial germ cells can reenter the germ line
.
Dev. Biol
.
161
,
626
628
.
Stöger
,
R.
,
Kubicka
,
P.
,
Liu
,
C.-G.
,
Kafri
,
T.
,
Razin
,
A.
,
Cedar
,
H.
and
Barlow
,
D. P.
(
1993
).
Maternal-specific methylation of the imprinted mouse Igf2r locus identifies the expressed locus as carrying the imprinting signal
.
Cell
73
,
61
71
.
Tam
,
P. P. L.
and
Snow
,
M. H. L.
(
1981
).
Proliferation and migration of primordial germ cells during compensatory growth in mouse embryos
.
J. Embryol. Exp. Morph
.
64
,
133
147
.
Williams
,
D. E.
,
Eisenman
,
J.
,
Baird
,
A.
,
Rauch
,
C.
,
Van Ness
,
K.
,
March
,
C. J.
,
Park
,
L. S.
,
Martin
,
U.
,
Mochizuki
,
D. Y.
,
Boswell
,
H. S.
,
Burgess
,
G. S.
,
Cosman
,
D.
and
Lyman
,
S. D.
(
1990
).
Identification of a ligand for the c-kit proto-oncogene
.
Cell
63
,
167
174
Williams
,
R. L.
,
Hilton
,
D. J.
,
Pease
,
S.
,
Willson
,
T. A.
,
Stewart
,
C. L.
,
Gearing
,
D. P.
,
Wagnew
,
E. F.
,
Metcalf
,
D.
,
Nicola
,
N. A.
and
Gough
,
N. M.
(
1988
).
Myeloid leukaemia inhibitory factor maintains the developmental potential of embryonic stem cells
.
Nature
336
,
684
687
.
Zsebo
,
K. M.
,
Wypych
,
J.
,
McNiece
,
I. K.
,
Lu
,
H. S.
,
Smith
,
K. A.
,
Karkare
,
S. B.
,
Sachdev
,
R. K.
,
Yuschenkoff
,
V. N.
,
Birkett
,
N. C.
,
Williams
,
L. R.
,
Satyagal
,
V. N.
,
Tung
,
W.
,
Bosselman
,
R. A.
,
Mendiaz
,
E. A.
and
Langley
,
K. E.
(
1990
).
Identification, purification, and biological characterization of hematopoietic stem cell factor from buffalo rat liverconditioned medium
.
Cell
63
,
195
201
.