Germ cell analysis

The methodology used to process germ cells has been previously described by our research group (Solé et al. 2021) and it consists of several steps:

Testicular tissue was obtained from four 12-week-old C57BL/6 J mice. The tissue was mechanically and enzymatically disaggregated following the procedure described by Garcia-Quevedo et al. (2012). Testicular cells were adhered to customized polylysine-coated slides (1 mg/ml), fixed with 4% paraformaldehyde, and subjected to a permeation treatment with 0.1 N hydrochloric acid, 0.5% triton, liquid nitrogen, and 0.005% pepsin (protocol adapted from Cremer et al. 2008). Prior to application of the FISH procedure, slides were incubated in 50% formamide for a minimal period of 2 months, with the purpose of achieving a prehybridization slow DNA denaturation.

Three successive rounds of FISH were performed using a custom designed Chromoprobe Multiprobe® OctoChrome Murine System ™ kit (Cytocell Ltd, Cambridge, UK). This kit contains a multiprobe device that consists of three different coverslips with seven delimited independent regions. Each of these regions presents a specific combination of three painting labeled with a different fluorochrome (Aqua DEAC, FITC, and Texas Red). Therefore, after sequential application of three coverslips, it is possible to identify up to nine chromosomes per region in the same nuclei. Analysis of all the coverslips and regions provides identification of all chromosomes of the mouse karyotype as well as all possible combinations of chromosome pairs.

In each round of FISH, the slide and the corresponding coverslip were mounted together in formamide solution and then subjected to denaturation for 5 min at 75 °C in a Hybridization Vysis HYBrite System. Hybridization was performed for 60 h at 37 °C (procedure adapted from Cytocell manufacturer instructions). After hybridization, the coverslip was removed and slides were transferred to 1 × saline-sodium citrate (SSC) buffer, incubated for 2 min, and washed in 2 × SSC/0.05% Tween-20 at RT for 30 s. Finally, hybridized areas were mounted in the antifade provided by the kit. After each FISH round, a washing step was performed with 0.0625 × SSC at 73 °C for 5 min in order to remove previous hybridization signals.

A Leica TCS-SP5 confocal microscope coupled to an image analysis system (LAS AF v.1.8.1) was used to capture serial optical sections of nuclei after each hybridization round (Fig. 1). A hybrid detector (HyD) and HCX PL APO lambda blue 63.0 × 1.40 OIL UV objectives were used. Specifically, lasers and excitation frequencies applied were a 405 nm UV diode laser, a 488 nm argon laser, and a 561 nm DPSS laser for the Aqua DEAC, FITC, and Texas Red fluorochromes, respectively. The HyD detector was configured at 415–470 nm for Aqua DEAC, at 500–550 nm for FITC, and at 571–750 nm for Texas Red. A high-speed resonant scanner module was used to obtain serial optical sections on the X, Y, and Z axes, with a distance between sections of 0.17 μm, a 512 × 256 pixel format, and an optical zoom of 5X. The number of sections was different for each nucleus (i.e., according to the corresponding nuclear volume). All captures were associated with their coordinates in order to relocate and capture the same nuclei after each hybridization round.

Fig. 1

3D-FISH confocal image captures of each round of hybridization (a, b, and c) for four different mice germ cell nuclei (spermatogonia-early preleptotene, mid-preleptotene-zygotene, pachytene, and round spermatids). Different combinations of three chromosomes (Chr.) displayed in FITC, Texas Red, or Aqua DEAC are observed in each nucleus. For each hybridization round are shown maximum intensity projections of confocal serial sections and 3D composite reconstructions using Imaris 9.3 software in both RGB merge and split channels (to view the planes’ sequence of the maximum intensity confocal images and the 3D composite reconstruction of premeiotic cells, see Online resource 1)

Image analysis was performed using customized developed scripts designed within the Fiji software environment (Schindelin et al. 2012) and Matlab R2013b. The Fiji scripts permitted segmenting nuclei and chromosome territories in different serial binary images as shown in the animation Online resource 1. Following this, Matlab scripts permitted 3D reconstruction of the nuclei and chromosome territories, enabling the extraction of numerical data concerning chromosome position and volume, as well as nuclei volume.

We considered that two homologous chromosomes were paired when all pixels of a specific signal formed a unique and continuous group of voxels (Online resource 2). Conversely, homologous chromosomes were classified as unpaired when two groups of voxels were observed as two separate entities, so sharing no voxel (Online resource 2). Similarly, we considered that two nonhomologous chromosomes (also referred to as heterologous chromosomes) were associated when their territories overlapped. The percentage of overlapping was determined as the number of shared voxels respect the voxels each chromosome territory occupy.

An immunofluorescence procedure was designed to unequivocally identify premeiotic cells in relation to somatic cells and meiotic cells. With this purpose, three proteins were identified: synaptonemal complex protein 3 (SYCP3), which allowed identification of germ cells from mid-preleptotene to pachytene stages; testicle-specific histone H1 (H1T), which enabled identification of germ cells ranging from the late pachytene stage to round spermatids; and a germ cell-specific nuclear antigen recognized by the monoclonal antibody TRA98 (Tanaka et al. 1997), which allowed discrimination between somatic cells and germ cells. Various cell fractions were then identified based on the presence or absence of these proteins as previously shown by our research group (Solé et al. 2021):

I. Premeiotic cells. This category includes cells from A-type spermatogonia to early preleptotene spermatocytes (the onset of the premeiotic S-phase in which DNA replication is performed). These nuclei stained positively for TRA98, with negative staining for SYCP3 and histone H1T (Fig. 2).

II. Mid preleptotene-zygotene spermatocytes. These nuclei showed positive staining for TRA98 and for SYCP3 protein and negative staining for histone H1T (Fig. 2). Despite using the same fluorophores to detect SYCP3 and TRA98, both proteins were easily distinguishable because TRA98 showed a uniform labeling pattern that was distinct from SYCP3’s dotted or thread-like appearance (Fig. 2).

III. Pachytene spermatocytes. Pachytene spermatocyte nuclei stained positive for TRA98, demonstrated a thread-like staining pattern for SYCP3, and were either negative (early pachytene) or positive (late pachytene) for histone H1T (Fig. 2).

IV. Round spermatids. Spermatid nuclei stained positive for TRA98 and histone H1T and were negative for SYCP3 (Fig. 2).

Fig. 2

Graphical representation of the labeling pattern of synaptonemal complex protein 3 (SYCP3), testicle-specific histone 1 (H1T), and a testis-specific nuclear protein known as TRA98 during the spermatogenesis process, along with some relevant meiotic events that occur in primary spermatocytes. The colored bars indicate the presence of each protein: SYCP3 (dark green bar), H1T (red bar), and TRA98 (light green bar), throughout spermatogenesis (the top grey bar)

The protein labeling process was carried out after chromosomal FISH analysis in two sequential rounds. In the first round, detection of SYCP3 and histone H1T was performed while, in the second round, samples were immunostained with the monoclonal antibody against TRA98. Briefly, cells were fixed in 3% paraformaldehyde for 10 min and permeabilized in a 1 × PBS-0.5% Triton-X100 solution for 5 min. Next, samples were incubated for 15 min with blocking solution (1% w/v bovine serum albumin). After blocking, the cells were incubated at 4 °C with the primary antibodies, specifically rabbit anti-SCP3 (Abcam, Cambridge, UK) plus either guinea pig anti-H1T (The Jackson Laboratory, USA) (first round) or rat anti-TRA98 (Abcam, Cambridge, UK) (second round). Next, samples were incubated with secondary antibodies for 40 min at 37 °C. Specifically, the secondary antibodies were goat anti-rabbit FITC (Jackson ImmunoResearch Inc., Cambridge, UK) and goat anti-guinea pig CY3 (Jackson ImmunoResearch Inc., Cambridge, UK) (first round) or goat anti-rat FITC (Abcam, Cambridge, UK) (second round).

Results were analyzed and processed in the following manner: data regarding chromosome position and volume, which were extracted from the application of Matlab scripts, permitted to calculate (1) the average percentage of paired and unpaired homologous chromosomes; (2) the average percentage of nonhomologous chromosomes sharing the same chromosomal territory; (3) the average of nonhomologous overlapping percentage; and (4) the nuclear volume proportion occupied for each chromosomal territory analyzed.

Pearson’s correlations were performed to evaluate the degree of linear relationship between chromosome size, GC content, and chromosome gene density (these parameters are detailed on Table 1) and the average rate of homologous chromosome pairing. Besides, the percentage of paired homologous chromosomes was compared between NOR-bearing-chromosomes and no NOR-bearing-chromosomes by a T-test. These analyses were only performed at the first two stages studied (i.e., spermatogonia-early preleptotene spermatocytes and mid preleptotene-zygotene spermatocytes), as there were no unpaired chromosomes in the remaining stages.

Table 1 Chromosome features used to test for possible conditioning factors of homologous pairing. Data extracted from The Genome Reference Consortium, Mus musculus GRCm38.p6. aData extracted from The Genome Reference Consortium, Mus musculus GRCm38.p6. bData extracted from (Evans et al. 1974; Henderson et al. 1974; Atwood et al. 1976; Dev et al. 1977; Kurihara et al. 1994; Britton-Davidian et al. 2011). Crosses (x) indicate the presence of nucleolus organizer regions (NORs)

Concerning heterologous associations, the Wald's asymptotic method was used to calculate 95% confidence intervals for the overlap of each pair of chromosomes. For those with an overlap of 0%, Wilsons’ score was used. Chromosomes with a confidence interval above or below the overall weighted mean of associations were considered statistically significant.

Analyses were performed by SAS v9.4 software, SAS Institute Inc., Cary, NC, EEU and the significance level was set to 0.05.

Lymphocyte analysis

As somatic pairing does not occur in mice cells, we additionally analyzed the distribution of homologous chromosomes in lymphocytes, in order to obtain a basal level at which two homologous chromosomes are observed nearby due to non-pairing related causes. Accordingly, spleens were removed from two C57BL/6 J mice. After injecting 5 ml of RPMI into the spleens using a syringe and recovering the solution, lymphocytes were isolated by Ficoll-Paque gradient separation, fixed with 4% paraformaldehyde, and subjected to permeation treatment with 0.1 N hydrochloric acid, 0.5% triton, liquid nitrogen, and 0.005% pepsin (protocol adapted from Cremer et al. 2008). Next, FISH was again performed using the custom designed Chromoprobe Multiprobe® OctoChrome Murine System ™ kit (Cytocell Ltd, Cambridge, UK). In this case, only one coverslip was used; specifically, this coverslip permitted labeling the following chromosomes: 1, 3, 4, 6, 7, 8, 9, 10, 11, 12, 14, 16, 17, 18, and 19. The analysis was performed using an Olympus BX60 epifluorescence microscope equipped with filter sets for FITC, Texas Red, Aqua DEAC, and DAPI/Texas Red/FITC. The same criteria applied in the germ cell analysis were then used in order to classify the homologous chromosomes as paired or unpaired.