Transcriptome analysis of cPGCs

To correlate in vivo observations of epigenetic marks with epigenetic modifier expression in early chicken embryo germ cells, we elucidated germ cell-specific expression profiles. Comparisons were made between cPGCs and ESCs, the in vitro derivatives of early embryonic germ and somatic cells, respectively. RNA-seq, followed by differential gene expression analysis and transcript per million (TPM) analysis, was performed for a comprehensive estimation of gene expression levels (Additional file 1: Table S1; excerpt in Additional file 2: Table S2). As expected, pluripotency-associated genes, such as NANOG, OCT4/POUV, SOX2, SOX3, KLF2, KLF4, and KLF5, exhibited robust expression in both cell types. Differential analysis confirmed significantly higher expression of germinal genes, including DAZL, DDX4, PIWIL1, and MAEL, in cPGCs. Interestingly, HOX genes were markedly repressed in cPGCs, with a mean TPM of 0.05 compared with 6.61 in ESCs. This repression aligns with observations in mammals, where HOX genes are downregulated in migrating PGCs compared with somatic neighboring cells, likely reflecting the necessity to inhibit somatic transcriptional programs for germ specification [13, 14].

DNA methylation in chicken gonadal PGCs

In mammals, DNA methylation in PGCs reaches its nadir upon settling at the genital ridges and proliferating in the developing gonads. Therefore, we examined postmigratory chicken PGCs in the embryonic gonads during the period of germ and somatic gonadal cell proliferation, before female germ cells initiate meiotic arrest after approximately 14 days of development [51]. Fluorescence immunodetection in tissue sections of the gonads from 6‑, 8‑, 10‑, and 14‑day-old chick embryos of both sexes was used to visualize epigenetic marks. Germ cells were identified by the presence of CVH/DDX4 and DAZL proteins as germ-specific markers in early chicken embryos [52, 53]. Fluorescent signals were observed in the nuclei of these germ cells and surrounding somatic cells. At all stages and for both sexes, 5mC was clearly detected in germ cells and neighboring somatic cells (Fig. 1A). Signal quantification revealed consistently higher global 5mC levels in germ cells compared with somatic cells, contrasting with the global DNA demethylation observed in mammalian gonadal PGCs. Regarding DNA hydroxymethylation on CpG, a pronounced reduction in 5hmC has previously been observed in mouse and human gonadal germ cells, with this decrease following a transient increase upon the arrival of mouse PGCs at the genital ridges [9, 20]. In chicken gonads, 5hmC was abundant in somatic cells but barely detectable in germ cells, regardless of stage and sex (Fig. 1B). Importantly, the global 5mC and 5hmC levels in gonadal PGCs were similar to those quantified previously in cPGCs relative to somatic cells [30].

Fig. 1

5mC and 5hmC in chicken embryo gonads. A Immunodetection of 5mC (gray) and germ cell marker (red) in tissue sections. Scale bar: 10 µm. Quantification of fluorescence intensity in germ and somatic cell nuclei is shown below. Number of analyzed nuclei in 6‑, 8‑, 10‑, and 14‑day-old embryos: for females, 44, 50, 60, and 54 germ cells and 74, 101, 68, and 57 somatic cells, respectively; for males, 28, 22, 56, and 34 germ cells and 71, 76, 76, and 47 somatic cells, respectively. B Immunodetection of 5hmC (gray) and germ cell marker (red) in tissue sections. Scale bar: 10 µm. Quantification of fluorescence intensity in germ and somatic cell nuclei is shown below. Number of analyzed nuclei in 6‑, 8‑, 10‑, and 14‑day-old embryos: for females, 48, 27, 49, and 34 germ cells and 76, 81, 54, and 41 somatic cells, respectively; for males, 37, 13, 39, and 33 germ cells and 61, 46, 42, and 48 somatic cells, respectively

Expression of DNA methylation modifiers in germ cells (Additional file 2: Table S2) was explored by comparing cPGCs to ESCs, which exhibit 5mC and 5hmC levels typical of somatic cell types [30]. In mouse PGCs, reduced expression or protein presence is observed for several factors involved in 5mC maintenance, including DNA methyltransferases (DNMTs) and UHRF1/NP95, whereas TET1 and TET2 enzymes, converting to 5mC to 5hmC, showed increases [3, 5, 7, 13, 54]. In chicken cPGCs, we found that DNMT genes were not underexpressed, aligning with previous reports indicating higher expression of DNMT1, DNMT3A, and DNMT3B in PGCs compared with somatic gonadal cells [55]. UHRF1 exhibited slightly higher expression in cPGCs relative to ESCs, whereas TET1, TET2, and TET3 showed lower expression, suggesting a balance that does not favor DNA demethylation. Notably, LSH/HELLS, necessary for repetitive element methylation, especially at PCH [56], showed higher expression in cPGCs, consistent with the presence of large nuclear foci of 5mC (Fig. 1A).

Histone PTMs in chicken gonadal PGCs

To determine whether certain histone-related events of mammalian epigenome reprogramming occurred in chicken germ cells, we analyzed H3K9me2 and H3K27me3, the heterochromatic histone PTMs that exhibit the most pronounced changes in mammalian PGCs. In females, the H3K9me2 level was slightly higher in germ cells than in somatic gonadal cells initially, then gradually decreased (by approximately twofold), whereas in males, it remained consistently higher at all stages (Fig. 2A). Thus, the enduring loss of H3K9me2 observed in mammalian gonads [3, 4, 20, 21] did not manifest in chickens. Initially, the H3K27me3 level was marginally higher in germ cells than in somatic gonadal cells in females but was consistently higher (up to 2.4-fold) in males (Fig. 2B). Chromatin enrichment for H3K27me3 occurs transiently in human PGCs [20, 21] and transiently or for an extended period, depending on the study, in mouse PGCs [3, 4]. Our results suggest the presence of sex-specific variations in H3K27me3 levels in chicken PGCs. However, the observed changes in H3K9me2 and H3K27me3 were not consistent with the hypothesis of the latter replacing the former, as proposed in the mammalian model [3].

Fig. 2

H3K9me2 and H3K27me3 in chicken embryo gonads. A Immunodetection of H3K9me2 (gray) and germ cell marker (red) in tissue sections. DNA staining (cyan). Scale bar: 10 µm. Quantification of fluorescence intensity in germ and somatic cell nuclei is shown below. Number of analyzed nuclei in 6‑, 8‑, 10‑, and 14‑day-old embryos: for females, 24, 29, 52, and 45 germ cells and 133, 179, 65, and 56 somatic cells, respectively; for males 33, 34, 41, and 38 germ cells and 125, 126, 104, and 112 somatic cells, respectively. B Immunodetection of H3K27me3 (gray) and germ cell marker in tissue sections. DNA staining (cyan). Scale bar: 10 µm. Quantification of fluorescence intensity in germ and somatic cell nuclei is shown below. Number of analyzed nuclei in 6‑, 8‑, 10‑, and 14‑day‑old embryos: for females, 42, 48, 34, and 84 germ cells and 411, 329, 202, and 314 somatic cells, respectively; for males, 31, 21, 34, and 44 germ cells and 434, 224, 225, and 244 somatic cells, respectively

Next, we examined H3K9me3, the histone PTM specific for constitutive heterochromatin. Remarkably, the global level of H3K9me3 was substantially higher in chicken gonadal germ cells compared with surrounding somatic cells (Fig. 3A). This enrichment surpassed that observed for H3K9me2 and H3K27me3, being markedly higher (2.0 to 4.5-fold) and observed in both sexes at all stages. This result aligns with our previous observation that chicken cPGCs exhibit a high level of H3K9me3 compared with several somatic cell types [30]. However, it contrasts with findings in mammals, where H3K9me3 levels are known to be low in germ cells or similar to those in surrounding gonadal cells [4, 20, 21, 57]. According to the present RNA-seq analysis and our previous findings [30], KMT1B/SUV39H2, encoding the main enzyme responsible for H3K9 trimethylation in heterochromatin, was more expressed in PGCs than in ESCs and several other somatic cell types, potentially contributing to the observed high H3K9me3 level. Moreover, several genes encoding enzymes involved in H3K9 trimethylation deposition or removal were more highly expressed in chicken PGCs than in ESCs, with KDM4C being significantly overexpressed, and KMT1E, KDM3A, KMT1D, KDM4B and KDM7A exhibiting slightly higher expression levels (Additional file 2: Table S2).

Fig. 3

Constitutive heterochromatin marks and their nuclear distribution in gonadal germ cells. A H3K9me3 in chicken embryo gonads. Immunodetection of H3K9me3 (gray) and germ cell marker (red) in tissue sections. DNA staining (cyan). Scale bar: 10 µm. Quantification of fluorescence intensity in germ and somatic cell nuclei is shown below. Number of analyzed nuclei in 6‑, 8‑, 10‑, and 14‑day-old embryos: for females, 61, 53, 34, and 91 germ cells and 685, 423, 311, and 210 somatic cells, respectively; for males, 88, 46, 64, and 40 germ cells and 377, 266, 183, and 229 somatic cells, respectively. B Immunodetection of H3K9me3 (magenta) and HP1beta and gamma (green) in a 10‑day-old male embryo gonadal tissue section. Stars indicate germ cell nuclei, identified using their specific H3K9me3 enrichment and nuclear morphology. DNA staining (cyan). Scale bar: 5 µm. C Analysis of the radial distributions of H3K9me3, HP1, and DNA signal intensities in germ and somatic cell nuclei. Number of analyzed nuclei for germ and somatic cells: 44 and 55 for H3K9me3, 44 and 55 for DNA, 25 and 30 for HP1alpha, and 19 and 25 for HP1gamma, respectively. D Quantification of fluorescence intensity for HP1beta and gamma immunodetection in gonadal germ and somatic cell nuclei from 10-day-old male embryos. Number of analyzed nuclei for germ and somatic cells: 27 and 60 for HP1beta, and 19 and 45 for HP1gamma, respectively

We also examined several histone PTMs specific to active chromatin: H3K4me3, H3K4me1, H3K27ac, and H3K9ac. Regarding H3K4 methylation, germ cell levels did not markedly differ from somatic gonadal cell levels, with a decreasing tendency observed in females. In males, germ cell H3K4 methylation levels remained higher than those in somatic cells at all stages (Additional file 2: Fig. S1A and B). Regarding acetylation, H3K27 in germ cells followed a similar trajectory in both sexes, starting at similar levels relative to somatic cells and ending at lower levels (Additional file 2: Fig. S1C). H3K9 acetylation constantly maintained significantly lower levels in germ cells compared with somatic cells (Additional file 2: Fig. S1D). These observations resembled those documented for gonadal mouse PGCs, characterized by low levels of active histone PTMs [4], as opposed to findings reported for human PGCs [21].

Nuclear organization and chromatin density in chicken PGCs

As mammalian PGCs differ from somatic cells in the abundance and nuclear distribution of epigenetic marks and architectural proteins, particularly at chromocenters, we investigated whether this distinction extended to chicken PGCs. Initially, we examined the presence of epigenetic marks at PCH, which forms chromocenters in chicken cPGC nuclei [30], on gonad tissue sections. In chicken gonadal PGCs, 5mC displayed a typical pattern of enrichment at chromocenters, similar to somatic cells (Fig. 1A). Although 5hmC was barely visible in PGC nuclei, overexposure after immunodetection revealed a faint modification without the relatively homogenous pattern found in somatic cells. Rather, it was concentrated in large foci (not shown), akin to chromocenters, reminiscent of the transient enrichment observed in mouse gonadal PGCs [9]. Regarding repressive histone PTMs, H3K9me3 was strongly enriched at chromocenters in PGCs, whereas H3K9me2 was unenriched, not differing from the H3K9me2 levels observed in somatic cells (Fig. 2A). Some of the largest chromocenters in gonadal PGC nuclei were enriched in H3K27me3, as previously observed for cPGCs and chicken PSCs [30]. Surprisingly, they also exhibited enrichment of H3K4me3, indicative of a unique type of PCH, perhaps related to bivalent chromatin (not shown). Therefore, no loss of heterochromatin marks at PCH, as initially reported for mouse PGCs [4], was observed. On the contrary, heterochromatin marks were present at PCH in chicken gonadal PGCs, consistent with later reports for mouse and human PGCs [20, 57].

We examined the nuclear distribution of heterochromatin components in greater detail to understand which genome compartments were affected by the global enrichment of H3K9me3 in PGCs. In somatic cells, H3K9me3 was concentrated in large foci corresponding to centrally located bright chromocenters, typical of constitutive PCH. In germ cells, H3K9me3 was more homogeneously distributed in the nucleoplasm, suggesting its presence in other, more diffuse, regions as well as PCH (Fig. 3B). Heterochromatin proteins HP1beta and gamma were mainly enriched at chromocenters in both cell types (Fig. 3B). Some large peripheral H3K9me3-rich domains in germ cells were not enriched for HP1, suggesting a chromatin type wherein H3K9me3 plays a role other than constitutive heterochromatin maintenance. The radial distributions of these heterochromatin components in nuclei confirmed differences between the two cell types (Fig. 3C). Maximal H3K9me3 enrichment was at the nuclear center in somatic cells but near the periphery in PGCs. HP1 isoforms’ radial distributions were similar between the two cell types, with a slight shift to the periphery observed for PGCs. Notably, heterochromatin proteins HP1beta and gamma were more abundant in chicken germ cell nuclei than in somatic cell nuclei (Fig. 3D), thereby differing from mammalian PGCs, in which HP1 isoforms in the nuclei tend to be reduced or absent during PGC epigenetic reprogramming [4, 20, 57]. Overall, the high levels and distributions of heterochromatin components in chicken PGCs did not correspond to a mere intensification of the somatic cell pattern, suggesting the germ-specific functions of these epigenetic actors.

We also investigated the presence of macroH2A, a histone variant that undergoes marked redistribution during mammalian PGC reprogramming, including concentration at PCH or depletion from the nucleus [4, 20]. Notably, macroH2A1, a variant stabilizing nucleosome in chicken cells [58], predominantly localized to large nuclear foci in gonadal germ cells at early stages compared with somatic cells, becoming barely detectable later (Fig. 4A). This distinctive nuclear distribution was also observed in cPGCs but not in cultured somatic cells, such as ESCs and chicken embryonic fibroblasts (CEFs) (Fig. 4B). Consistently, chicken cPGCs exhibited low levels of macroH2A1 mRNA (Additional file 2: Table S2) and protein (not shown) compared with ESCs. To confirm that macroH2A1 foci in chicken PGCs corresponded to PCH, we examined the presence of other PCH characteristics in cPGC nuclei (Fig. 4C). At macroH2A1 foci, the centromere protein CENP‑T was, on average, threefold more concentrated than in the whole nucleus, indicating proximity to centromeres. H3K9me3 and DNA were 1.9-fold and 1.4-fold more concentrated, respectively, suggesting that macroH2A1 foci tended to contain heterochromatin.

Fig. 4

MacroH2A1 and chromatin density in chicken embryo germ cells. A Immunodetection of macroH2A1 (gray) and germ cell marker (red) in gonad tissue sections. DNA staining (cyan). Scale bar: 10 µm. Quantification of fluorescence intensity in germ and somatic cell nuclei is shown below. Number of analyzed nuclei in 6‑, 10‑, and 14‑day-old-embryos: for females, 43, 44, and 38 germ cells and 76, 45, and 30 somatic cells, respectively; for males, 25, 39, and 34 germ cells and 46, 60, and 60 somatic cells, respectively. B macroH2A1 nuclear distribution in chicken germ and somatic cells. Immunodetection of macroH2A1 (green) in cPGCs, ESCs, and CEFs. DNA staining (cyan). Scale bar: 5 µm. C macroH2A1 foci compared to pericentric heterochromatin. Immunodetection of macroH2A1 (green) and CENP‑T or H3K9me3 (magenta) in the nucleus of cPGCs. DNA staining (cyan). Scale bar: 5 µm. Box plots show the enrichment in CENP‑T and H3K9me3 (mean intensity in each focus compared to mean intensity in whole nucleus) for more than 100 foci. D Representative germ and somatic cell nuclei in gonadal tissue sections with DNA staining (gray). Stars indicate germ cell nuclei. Scale bar: 5 µm. E Quantification of chromatin density (measured as the intensity of the DNA staining by unit area) in germ and somatic cell nuclei. Number of analyzed nuclei in 6‑, 8‑, 10‑, and 14‑day-old embryos: for females, 311, 284, 308, and 489 germ cells and 1873, 1754, 1101, and 919 somatic cells, respectively; for males, 323, 241, 323, and 316 germ cells and 1899, 1220, 924, and 1069 somatic cells, respectively. F Ultrastructure of germ and somatic cell nuclei observed using transmission electron microscopy in 14‑day-old embryo gonads. Nuclei of germ (G) and somatic (S) cells are indicated. Scale bar: 1 µm. Magnified views of the nuclear envelope and the associated chromatin are shown

Subsequently, we examined indicators of chromatin density in chicken PGCs, searching for possible “loosening,” similar to the epigenome reprogramming of mouse PGCs. Chromocenters have been reported to disappear around E10.5 [4] or become less visible [8, 10, 57] in mouse PGCs. Although DNA labelling intensity tended to be lower in chicken gonadal PGCs than in somatic cells, chromocenters remained visible at all stages, either as clearly defined foci near the nuclear center or embedded in the peripheral rim of heterochromatin (Fig. 4D). The nuclear distribution of the centromere protein CENP‑T did not differ between germ cells and somatic cells (data not shown); however, CENP‑T foci tended to be smaller and fainter in the former, suggesting chromatin decondensation, possibly resulting from the declustering of pericentromeres, akin to the occurrence in mouse PGCs [57]. We estimated chromatin density in gonadal PGC nuclei by measuring the fluorescence intensity of DNA counterstaining per unit area of nucleus section. For both sexes and at all stages, germ cell chromatin appeared about twofold less dense than somatic cell chromatin (Fig. 4E). Therefore, we analyzed the ultrastructure of gonadal cell nuclei using transmission electron microscopy in 14-day-old chick embryos (Fig. 4F). Germ cell nucleoplasm appeared less electron-dense compared with somatic cell nucleoplasm and lacked a discernible dense chromatin layer beneath the nuclear envelope. This pattern was similar to our previous observation in cPGCs [30], as well as observations of PGCs settling in the genital ridges of chicken embryos [59]. Thus, low-density fluorescent DNA staining in germ cell nuclei corresponded to low chromatin compactness.

Establishing the germ-specific epigenetic signature during embryogenesis

We attempted to determine when the distinctive epigenetic profile observed in gonadal PGCs emerged during early embryo development. Initially, we examined PGCs at the blastoderm center in unincubated eggs (stage EG&K X–XII embryos). At this stage, PGC nuclei were indistinguishable from blastodermal somatic cell nuclei (Fig. 5A). H3K9me3, H3K27me3, and 5mC primarily localized at chromocenters, whereas H3K9me2, H3K9ac, 5hmC, and macroH2A1 exhibited a more homogeneous distribution. By the HH4 stage (Fig. 5B), when PGCs had relocated to the germinal crescent, their nuclei displayed a similar pattern to somatic cell nuclei for H3K9me3, H3K27me3, and 5mC, with macroH2A1 more frequently exhibiting foci in PGCs. Remarkably, about 61% of PGC nuclei (46 cells from 6 embryos) showed lower 5hmC richness compared with the surrounding somatic cell nuclei. By the HH13 stage (Fig. 5C), as PGCs began migration through the vascular system, some displayed features of the distinctive epigenetic profile, i.e., high H3K9me3 levels at diffuse regions outside chromocenters and less intense DNA staining of their nuclei compared with surrounding cell nuclei, suggesting chromatin decondensation. The macroH2A1 pattern with large foci was occasionally observed. Examining PGCs newly arrived at the genital ridges from the HH15 to HH23 stage, we found that the PGC-specific macroH2A1 nuclear pattern was present in most cells (Fig. 5D). Quantification of H3K9me3 labelling (as previously performed on the gonads) revealed significantly higher modification levels in PGCs compared to the surrounding somatic cells, with an increase over time during this period (Fig. 5E) and beyond until the maximum level was reached in the gonads. Concurrently, chromatin density in PGC nuclei decreased (Fig. 5F). Therefore, the first elements of the PGC-specific epigenetic signature appear early during embryonic development, preceding PGC migration in the bloodstream, with subsequent steps occurring progressively during migration and in the differentiating gonads.

Fig. 5

Setting of the germ-specific epigenetic signature during chicken embryo early development. Immunodetection of histone PTMs, macroH2A1, 5mC, and 5hmC (gray) in chicken embryos. Representative nuclei of PGCs identified with a germ cell marker (red) and somatic surrounding cells are shown. For 5hmC, additional labelling of RNA pol II was performed to locate nuclei when necessary. DNA staining (cyan). Scale bar: 10 µm. A In the blastoderm (stage EG&K X). B In the germinal crescent (stage HH4). C In blood vessels near the head (stage HH13). D At the genital ridges (stage HH15-17). E Quantification of fluorescence intensity for H3K9me3 in germ and somatic cell nuclei in stage HH15 to HH23 embryos. Number of analyzed nuclei in HH15-17, HH20, and HH23 embryos (two embryos per stage): 39, 46, and 54 germ cells and 157, 144, and 259 somatic cells, respectively. F Quantification of chromatin density (intensity of the DNA stain by unit area) in germ and somatic cell nuclei, for the same stage HH15 to HH23 embryos

Genome-wide profiling of H3K9me3 in chicken PGCs

H3K9me3 global enrichment in chicken germ versus somatic cells is a striking distinctive feature of chicken PGCs compared to mammalian PGCs. It is also an uncommon epigenome feature, since observed H3K9me3 levels are generally similar in all cell types of a species, the modification being mostly present in constitutive heterochromatin such as PCH. However, fluorescence immunodetection did not suggest that the enrichment of H3K9me3 in chicken PGCs compared to somatic cells was located at PCH. To investigate precisely where this enrichment took place in the genome of chicken PGCs, we performed chromatin immunoprecipitation followed by high-throughput sequencing (ChIP-seq). We aimed at studying PGCs before they underwent sexual differentiation in the gonads. The low number of germ cells at migrating or early gonadal stages made the use of freshly isolated cells technically difficult. Consequently, as we had shown that cPGC maintained the epigenomic features of embryonic PGCs, we chose to analyze cPGCs versus ESCs, the in vitro derivatives of early embryonic germ and somatic cells. Studying such homogeneous cell populations also sharpened the identification of germ-specific features using combined epigenome and transcriptome analysis.

Initially, we determined whether the modification was enriched at repetitive sequences, given that H3K9me3 is present in constitutive heterochromatin in chickens [30, 60]. Evaluating the proportion of repeated element sequences in the immunoprecipitated DNA, we found that approximately 26% of sequenced bases were located in repeated elements for cPGCs compared with around 39% for ESCs (Fig. 6A). This indicates that the higher global H3K9me3 level in cPGCs did not result from preferential increase of enrichment at repetitive sequences. To estimate H3K9me3 enrichment at PCH, we examined sequences of the chicken nuclear membrane (CNM) repeat, enriched at PCH in chickens [60]. The proportion of CNM-containing reads in the immunoprecipitated DNA was approximately 12% and 15% for cPGCs and ESCs, respectively (Fig. 6B), indicating that H3K9me3 was not preferentially concentrated at PCH in germ cells relative to somatic cells. This finding aligns with our observations regarding the nuclear localization of H3K9me3, suggesting that the enrichment in PGCs, in comparison to somatic cells, is not primarily concentrated at PCH but rather distributed across other genomic regions. Analyzing the presence of H3K9me