Sertoli Cell Behaviors in Developing Testis Cords and Postnatal Seminiferous Tubules of the Mouse

Abstract

Sertoli cells are the primary structural component of the fetal testis cords and postnatal seminiferous tubules. Live imaging technologies facilitate the visualization of cell morphologies and behaviors through developmental processes. A transgenic mouse line was generated using a fragment of the rat Gata4 gene to direct the expression of a dual-color fluorescent protein reporter in fetal and adult Sertoli cells. The reporter encoded a red fluorescent protein, monomeric Cherry (mCherry), fused to histone 2B and enhanced green fluorescent protein (EGFP) fused to a glycosylphosphatidylinositol sequence, with a self-cleaving 2A polypeptide separating the two fusion proteins. After translation, the red and green fluorescent proteins translocated to the nucleus and plasma membrane, respectively, of Sertoli cells. Transgene expression in testes was first detected by fluorescent microscopy around Embryonic Day 12.0. Sertoli cell division and migration were visualized during testis cord formation in organ culture. Initially, the Sertoli cells had mesenchyme-like morphologies and behaviors, but later, the cells migrated to the periphery of the testis cords to become epithelialized. In postnatal seminiferous tubules, Sertoli nuclei were evenly spaced when viewed from the external surface of tubules, and Sertoli cytoplasm and membranes were associated with germ cells basally in a rosette pattern. This mouse line was bred to previously described transgenic mouse lines expressing EGFP in Sertoli cytoplasm or a nuclear cyan fluorescent protein (Cerulean) and mCherry in plasma membranes of germ cells. This revealed the physical relationship between Sertoli and germ cells in developing testis cords and provided a novel perspective on Sertoli cell development.

Sertoli cells are visualized using novel fluorescent protein reporters in transgenic mice to study their morphology and behavior in embryonic and postnatal testes.

INTRODUCTION

The mouse testis forms soon after midgestation from a gonad primordium that is bipotential (i.e., having the ability to develop into either a testis or an ovary). The seemingly indifferent gonad primordium contains primordial germ cells that later give rise to spermatozoa or oocytes. In mice, expression of the sex-determining region of the Y chromosome (Sry) gene on Embryonic Day (E) 10.5 initiates a cascade of differentiation events [1–3]. This includes up-regulation of several sex-determining factors, differentiation of Sertoli cells, testis cord formation, and male-specific vascularization of the gonad (for review, see [4]). By E13.5, testis cords containing germ cells have become defined, and the final number that will develop into seminiferous tubules has been established [5, 6].

Sertoli cells are the primary structural component of the testis cords and seminiferous tubules. They differentiate from the somatic cell population within the urogenital ridge shortly after the Sry gene is activated. The earliest sign of this differentiation is up-regulation of sex-determining genes, including Sox9 [7–11] and Gata4 [12–14]. The resulting morphological changes involve growth by an increase in proliferation [15, 16] and rearrangement of Sertoli cells to form testis cords that contain germ cells and exclude endothelial and other interstitial cell types.

Although gross morphological events that occur during testis cord formation have been described, most reports have not characterized the behavior of specific cell types within the cords during this developmental process [6, 17]. Various imaging and reconstruction techniques have been used to portray tissue morphogenesis in the fetal testis, but these techniques have had shortcomings in their inability to identify individual cells and their physical relationships with neighboring cells. Without this ability, it is very difficult to characterize individual cell behaviors and cell-cell interactions. A four-dimensional analysis of endothelial cell behaviors in the mouse fetal gonad using time-lapse live imaging and transgenic reporter mice demonstrated the power of fluorescent live imaging approaches for studying organogenesis [18]. In the current study, we therefore developed a transgenic mouse line that labels Sertoli cells using a previously described dual-color fluorescent protein gene construct [19] driven by a 4.7-kb region of the rat Gata4 promoter. This sequence can direct transgene expression exclusively in Sertoli cells in the testis, although it is not testis specific [20]. Live imaging with organ culture permits the observation of fluorescently labeled cells over time and provides insights regarding their origin, migration patterns, differentiation, and interaction with other cell types.

The overall objective of the current study was to use this newly developed transgenic mouse line as a live imaging tool to characterize Sertoli cell behaviors during testis cord development. Subsequently, this mouse line was bred to previously described Sox9-EGFP and Oct4-HS-CR mouse lines to study the relationship between the Sertoli and germ cells in embryonic and adult testes. The Sox9-EGFP (official symbol Sox9^tm1Haak) line expresses enhanced green fluorescent protein (EGFP) in Sertoli cell cytoplasm [6], whereas germ cell nuclei and membranes of the Oct4-HS-CR line express Cerulean, a nuclear cyan fluorescent protein, and monomeric Cherry (mCherry) fluorescent protein, respectively [19]. The individual and combined fluorescent protein expression patterns in these mouse lines provide a new perspective on the anatomy and development of testis cords and seminiferous tubules.

MATERIALS AND METHODS

Mice

B6D2F1 mice were purchased from The Jackson Laboratory. All animal manipulations were conducted in accordance with the National Research Council Guide for Care and Use of Laboratory Animals and approved by the University of Texas M.D. Anderson Cancer Center Institutional Animal Care and Use Committee.

Dissection and Culture Medium

Unless otherwise stated, all reagents were purchased from GIBCO (Invitrogen). Culture medium consisted of Dulbecco modified Eagle's medium with 4.5 g/L of d-glucose but without l-glutamine, sodium pyruvate, and phenol red (DMEM) and supplemented with 25% fetal bovine serum (HyClone; Thermo Fisher Scientific), 2 mM GlutaMax, 1 mM sodium pyruvate, 0.1 mM β-mercaptoethanol (Sigma-Aldrich), 10 μl/ml of MEM nonessential amino acids, 100 IU of penicillin, and 0.1 mg/ml of streptomycin. During dissection, urogenital ridges were maintained in culture medium containing 20 mM Hepes buffer (Sigma-Aldrich).

Gene Construction and Generation of Transgenic Mice

The pCAG-HS-XY represents a set of fluorescent reporter constructs employing different combinations of mCherry, enhanced green, Cerulean, and T-sapphire fluorescent protein reporters to visualize nuclei and plasma membranes [19]. In the current study, X is human histone 2B (H2B) fused with mCherry fluorescent protein, and Y is EGFP fused with the glycosylphosphatidylinositol signal sequence. This plasmid was transiently transfected into HeLa cells to confirm the predicted fluorescent protein subcellular localization expression pattern in live cells. HeLa cells were cultured in DMEM supplemented with 10% fetal bovine serum in a 37°C incubator with 5% CO₂. HeLa cells were transfected using Lipofectamine (Invitrogen) following the manufacturer's instructions.

A 4.7-kb fragment of the rat Gata4 promoter was inserted into pBluescript (pBS) as follows: A modified pXP1 plasmid (RV-164), which contains the rat Gata4 sequence previously shown to be sufficient for reporter gene expression in Sertoli cells [20], was digested with BamH1 and Sal1 to obtain 1.6- and 3.1-kb fragments. These fragments were sequentially ligated into pBS (previously linearized with BamH1 and Sal1). The complete, 4.7-kb promoter fragment was then isolated from pBS using Sal1 and Spe1.

To insert the isolated promoter fragment upstream of the dual-labeling color construct, the Red-Green (RG) version of pCAG-HS-XY [19] was digested with Spe1 and Xho1 followed by ligation with the isolated Gata4 promoter fragment (Fig. 1A). The 6.9-kb Gata4-RedGreen (or, more simply, Gata4-RG) gene construct was isolated from plasmid vector sequences by digestion with Spe1 and Ase1, gel purified, and microinjected into B6D2F2 zygotes using standard procedures [21]. Gata4-RG transgenic mice were identified by PCR using GFP primers, 5′-ACCCTGAAGTTCATCTGCACCACCG-3′ and 5′-CGTCGTCCTTGAAGAAGATGGTGCG-3′, that yield a 173-bp amplified DNA product. Founder animals were bred to B6D2F1/J mice, and subsequent hemizygous embryos and offspring were evaluated for fluorescent protein expression.

FIG. 1.

Diagram of Gata4-RG gene construct (A) and fluorescent protein expression in HeLa cells transfected with the CAG-RG construct (B) and within the transgenic line carrying the Gata4-RG transgene (C–G). A) Gata4 is a 4.7-kb rat promoter fragment; RG is an H2B-mCherry fragment fused to an EGFP glycosylphosphatidylinositol (GPI) fragment. Arrow indicates start of transcription; black box indicates the Gata4 5′ untranslated region (5′UTR). PolyA shows the polyA signal sequence. 2A, self-cleaving 2A polypeptide; SS, endoplasmic reticulum signal peptide; pA = bovine growth hormone polyadenylation signal sequence. B) HeLa cells transfected with the RG construct expressing EGFP in the membranes and mCherry in the nuclei. The arrow indicates a mitotic nucleus. C) Seminiferous tubules from 10-wk-old mice (arrowhead) expressing RG in Sertoli cells under control of the Sertoli-specific Gata4 promoter fragment. The outer cellular layer was damaged (asterisk) to demonstrate that only Sertoli cells express nuclear mCherry. D) Three-week-old seminiferous tubule expressing RG in Sertoli cells under control of the Sertoli-specific Gata4 promoter fragment. Green membranes formed rosette-like patterns around single red Sertoli nuclei (dashed circle), surrounding spermatogonia. E–G) Intact wild-type (wt) and transgenic (tg) E12.5 embryos under bright-field (E) microscopy showing EGFP (F) and mCherry (G) expression under fluorescent microscopy. Bars = 20 μm (B and D); 100 μm (C); and 2 mm (E–G).

Immunofluorescence

Immunofluorescence analysis of adult testes was performed as described previously [6] with modification. Briefly, 6-wk-old testes were dissected and rinsed in Ca²⁺/Mg²⁺-free Dulbecco PBS, and the tunica albuginea was punctured using forceps. The organs were fixed in 4% paraformaldehyde for 1 h at room temperature, serially transferred into 15% and 30% sucrose, 50% O.C.T. compound (Tissue-Tek, Sakura Finetek USA, Inc.) in 30% sucrose-PBS, and then frozen in 100% O.C.T. on dry ice. A primary rabbit anti-SOX9 antibody (1:200 dilution; Chemicon International), paired with a goat anti-rabbit Alexa-Fluor (AF) 633-conjugated (far red) secondary antibody (1:400 dilution; Molecular Probes), was used to label Sertoli cell nuclei in frozen sections (thickness, 10–12 μm). Slides were subsequently stained with 0.1 mg/ml of Hoechst 33342 in PBS for 10 min, rinsed in PBS for 5 min, and mounted with Gel/Mount mounting media (Biomeda Corp.). Slides were imaged using a PerkinElmer spinning disc laser confocal (SDLC) microscope (PerkinElmer Life and Analytical Sciences, Inc.) using 405-nm (blue for Hoechst), 640-nm (far red for AF 633), and 488-nm (green for EGFP) lasers.

Static Imaging

Postnatal seminiferous tubule imaging was performed using live tissue from 3-wk-old males. Testes were isolated, and the tunica albuginea was torn using forceps to expose the seminiferous tubules. The tissue was imaged in PBS in a minimal volume to ensure that the tubules were stationary on a glass-bottom Petri dish (MatTek corporation). Gata4-RG hemizygous males were crossed to Sox9-EGFP homozygous females [6] for more detailed visualization of postnatal Sertoli cells. The tissue was stained with Draq5 (Biostatus Ltd.), a vital nuclear fluorescent stain, before imaging to label nuclei.

To further investigate the relationship between embryonic Sertoli and germ cells, static imaging of tissues obtained by mating Gata4-RG hemizygous females with Oct4-HS-CR hemizygous males [19] was performed. Embryos were dissected at 0.5-day intervals to determine the timing and pattern of transgene expression in live tissues. Tail somite stage (TS) was determined for precise staging and comparison between different embryos, with E11.5, E12.5, and E13.5 corresponding to TS18, TS25, and TS30, respectively. To determine the genetic sex of embryos, amnions were fixed and stained with 1% toluidine blue to observe the presence (XX) or absence (XY) of a condensed chromatin body [22]. Gonads were imaged in dissection medium within 2 h of dissection on the SDLC microscope using the above-specified lasers as well as a 564-nm laser for mCherry fluorescent protein.

Organ Culture and Time-Lapse Imaging

Organ culture was performed as previously described [6]. Briefly, gonads were cultured at the air-medium interface on Millicell tissue culture plate inserts (Millipore Corp.) in 35-mm, poly-d-lysine-coated, glass-bottom culture dishes. Hemizygous Gata4-RG gonads were placed on the membrane with a minimal amount of medium. Time-lapse imaging was performed on the SDLC microscope at 37°C and 5% CO₂ in a humidified environmental chamber; control gonads were cultured in a separate humidified incubator at 37°C in 5% CO₂. Images were typically acquired at 30-min intervals using the lasers described above at 50% power and 500 msec of exposure. These specifications varied slightly depending on stage and brightness of the tissue as well as on the magnification used (typically 20× dry objective with 0.75 numerical aperture for time-lapse imaging).

RESULTS

Gata4-RG Expression Labels Sertoli Cells in Adult Testis and Embryos

When HeLa cells were transfected with the RG version of the pCAG-HS-XY plasmid, mCherry fluorescent protein was localized in nuclei and mitotic chromosomes, whereas EGFP marked the plasma membranes (Fig. 1B). This reporter construct facilitated the visualization of individual cells and clearly labeled the nuclei, chromosomes, and plasma membranes in two different fluorescent colors.

Six transgenic founder mice were generated by pronuclear injection of Gata4-RG into zygotes. These founders were bred with wild-type animals to generate progeny and embryos for analysis. For line 1 (Gata4-RG-1), 3- and 10-wk-old males had Sertoli cells that fluoresced red in the nuclei and green in the plasma membranes, indicating that the Gata4-RG transgene was being expressed and that mCherry and EGFP were present in the nuclei and membranes, respectively (Fig. 1, C and D). Interestingly, Sertoli nuclei were surrounded by Sertoli plasma membranes in a rosette pattern, suggesting an orderly association with germ cells (Fig. 1D). Transgenic embryos were easily identifiable in whole mount by mCherry and EGFP fluorescence in the head and tail (Fig. 1, E–G). Somatic cells in the testis expressed the transgene, but only after E11.5. These results corresponded to the known expression pattern of Gata4 in the mouse testis [14]. However, a subpopulation of interstitial cells also expressed Gata4-RG. It was not clear if these interstitial cells were Sertoli cells that had not been incorporated into the seminiferous tubules or if this resulted from spurious transgene expression.

In the second transgenic line (Gata4-RG-2), Sertoli cells of 3-wk-old mice did not express the transgene. Instead, Leydig cells had brightly fluorescing red nuclei and green plasma membranes, whereas peritubular myoid cells had lower levels of expression (data not shown). Expression by Leydig cells was only detected at E18.5, however, and fetal Sertoli cells did not express the transgene (data not shown). Leydig and peritubular myoid cell identity were confirmed by histology. Furthermore, expression of the transgene was observed in other tissues, including heart muscle, skeletal muscle, fat cells, kidney glomerular epithelium, hepatocytes, and oviduct epithelium. Transgenic embryos showed strong fluorescence in the apical epidermal ridge of the limb, which served as a simple identifier of transgenic embryos in this line.

The third transgenic line examined (Gata4-RG-3) expressed the transgene in specific parts of the vascular system, but mostly in the brain. Interestingly, only a subset of vessels fluoresced (data not shown). No expression was detected in the fetal testis (data not shown).

The remaining transgenic lines all had unique expression patterns, with one expressing only EGFP and no mCherry. The variation in expression between the Gata4-RG transgenic mouse lines suggests that the Gata4-RG construct is influenced by the chromosomal location of integration. Thus, we focused our analysis on the Gata4-RG-1 line. The other lines were not evaluated further.

Gata4-RG-1 Transgene Expression Colocalizes with SOX9 in Seminiferous Tubules

To confirm Sertoli cell-specific expression of the transgene, immunostaining for SOX9 was performed on frozen sections of Gata4-RG-1 adult testes. Although the membrane-bound EGFP fluorescence was not preserved during the immunostaining procedure, H2B-mCherry fluorescence was detected in immunostained sections. Colocalization with the SOX9 antibody confirmed that transgene-driven H2B-mCherry expression occurred specifically in Sertoli cells of the testis (Fig. 2).

FIG. 2.

Frozen sections of Gata4-RG adult testis immunostained for SOX9. Seminiferous tubules are composed of Sertoli cells (arrowheads) that express mCherry (red) in the nuclei (A and E) and labels positive for SOX9 (AF 488, green) antibody (B and F). Germ and somatic cell nuclei are stained blue (C and G). Colocalization of mCherry and SOX9 immunostaining (D and H) confirms that Sertoli cells are expressing the Gata4-RG transgene. Bar = 50 μm.

Gata4-RG-1 Allows Visualization of Sertoli Cells in Live Seminiferous Tubules

We crossed mice from the Gata4-RG-1 line with mice from the previously described Sox9-EGFP knock-in line and stained live, 3-wk-old seminiferous tubules with Draq5, a vital fluorescent nuclear stain. Sertoli cells had red nuclei with green plasma membranes (from Gata4-RG-1) and cytoplasms (from Sox9-EGFP), further confirming the Gata4-RG-1-expressing cell type. This allowed us to observe in detail the relationship between Sertoli cells and different stages of germ cells, labeled by the vital nuclear stain (Fig. 3 and Supplemental Movie 1; all Supplemental Data are available online at www.biolreprod.org). At the outer cell layers of the tubules, Sertoli cell nuclei were evenly distributed. EGFP-positive cytoplasm was visible between the basement membranes and the nuclei, apparent as a green layer covering the red nuclei (Fig. 3, A and D). Sertoli cell cytoplasm and spermatogonia occupied the areas between Sertoli cell nuclei (Fig. 3, A and D). More luminal to the Sertoli and spermatogonia layer, spermatocytes were surrounded by green cytoplasm of the Sertoli cells that extended toward the tubule interior (Fig. 3, B and E). Deeper into the tubule, round spermatids did not appear to have Sertoli cell cytoplasm between neighboring spermatids (Fig, 3, E and F). Further toward the lumen, elongating spermatids were embedded in cytoplasmic projections of Sertoli cells that extended into the center of the tubules (Fig. 3, C and F).

FIG. 3.

Live seminiferous tubule from a 3-wk-old Sox9-EGFP::Gata4-RG mouse stained with Draq5. Sertoli cell nuclei express mCherry (red) in their nuclei (arrowheads in A, B, and D) and EGFP (green) in their cytoplasm and membranes (double arrows in D and E). Sertoli cell nuclei (arrowheads in A and B) are located only in the outer layer of the tubule surrounded by spermatogonia (arrows in A) in a rosette arrangement. Their cytoplasms extend deep into the tubule (arrowheads in C) to surround the spermatocytes (arrows in B) and round spermatids (arrows in C) toward the tubule interior. Toward the center of the tubule, elongated spermatids are embedded in the Sertoli cell crypts (circled areas in B and C) before they are released into the lumen. D and E are zoomed images as denoted by dashed squares in A–C directly above for a more detailed view. s, spermatogonium; s*, spermatocyte; s**, round spermatid; s', elongated spermatid. Bar = 25 μm. See Supplemental Movie 1.

Gata4-RG-1 Transgene Expression in Embryonic Testes

At E11.5 (TS18, at the onset of testis differentiation), Gata4-RG transgene expression was not detected in XY gonads. Initial fluorescent protein expression could be visualized between E11.5 and E12.0 (∼TS22), and the initial, faint expression became brighter as differentiation progressed (Fig. 4, A–C, and Supplemental Movie 2). The nuclear mCherry became visible shortly before EGFP in plasma membranes could be detected. Transgene expression was limited to the Sertoli cell population within the testis cords, although a subpopulation of unidentified somatic cells in the interstitium also expressed the transgene. No expression was detected in XX fetal gonads (Fig. 5).

FIG. 4.

Still frame images from Supplemental Movies 2–4 from live imaging of Gata4-RG gonads. A–C) E12.0 gonad cultured and imaged for 21 h, showing initiation of Gata4-RG transgene expression. Initially, the gonad (inside rectangle) in not visible under fluorescent microscopy (A and B). As mCherry and EGFP expression increases, the red and green gonad (arrow in C) becomes visible. D–F) E12.5 gonad cultured for 42 h. Testis cords become more distinct as individual cords (arrow) becomes separated from adjacent cords. Some cords are completely eliminated over the culture period by retracting toward the mesonephros (asterisk in D–F). G–I) Late E12.5 gonad showing individual Sertoli cell behavior. Initially, red nuclei (arrows) are visible inside the cord interior (G). As time progresses, these nuclei migrate toward the cord surface and become more epithelial-like, leaving the cord interior (stars) appearing hollow (H–I). Bar = 100 μm.

FIG. 5.

Live gonads (A and E) with their attached mesonephroi (m) of E14.5 Gata4-RG XX and XY embryos. Gonads are outlined by yellow dashed ovals (A–H). Testis cords are visible in XY gonads (A–D), with Sertoli cells expressing mCherry in their nuclei (B) and EGFP in their cytoplasms (C), making the testis cords red and green (D). XX gonads do not have cords and do not express the fluorescent proteins (E–H). I–N represent a magnification at the cord exterior (I–K) and interior (L–M) of the yellow rectangle depicted in D. Sertoli cells (arrowheads) located in the exterior cell layer are in close contact with neighboring Sertoli cells (I–K) and appear to form clusters or sheets at the cord surface. Toward the cord interior (L–N), Sertoli cells are located between nonfluorescing areas (arrows), which represent germ cells. This causes the Sertoli cells to appear squeezed in between these cells. Bar = 500 μm (A–H) and 10 μm (I–N).

With high-magnification static imaging on the SDLC microscope, Sertoli cell morphology could be seen in great detail (Figs. 3, 5, and 6). Bright green fluorescent plasma membranes surrounded individual red fluorescent nuclei and allowed visualization of individual Sertoli cell shape (Fig. 5, I–N). The fluorescent Sertoli cell membranes also allowed indirect description of the shape of germ cells (no fluorescence) among the Sertoli cell population (Figs. 5, I–N, and 6). Sertoli cell membranes highlighted variable shapes, whereas germ cells were round and smooth in appearance (Fig. 5, I–N). Sertoli cell nuclei were brightly fluorescing red, which allowed visualization of different mitotic stages of the nuclei (not shown).

Sertoli Cells Exhibit Mesenchymal to Epithelial Behaviors During Cord Formation

Initially, all the Gata4-RG-1-expressing cells in the fetal testis had a mesenchymal morphology and appeared unpolarized. Their shape and behavior were reminiscent of fibroblasts, moving seemingly nonspecifically within the Sertoli germ cell mass (SGCM) as testis cords developed (Fig. 4, A–C, and Supplemental Movie 2). While cord formation proceeded between E12.0 and approximately E13.5, the fluorescing cells maintained mesenchymal morphology and movements and did not appear to actively contribute in shaping the SGCM (Fig. 4, D–F, and Supplemental Movie 3).

Throughout this process, individual Sertoli cells could be distinguished by their location and fluorescent protein expression. We witnessed migration of fluorescent Sertoli cells between adjacent, connected cords between E12.5 and E14.5. Migration between two cords occurred mostly toward one cord and appeared to be a mass migration of Sertoli cell groups rather than individual cells migrating from one cord to the other (data not shown).

Once testis cords were established at approximately E13.0, the majority of Sertoli cells became polarized such that nuclei were adjacent to the basement membrane, with their cytoplasms extending into the interior of the cord (Fig. 6). The migration to the cord surface was visible during live imaging by movement of cell nuclei from the interior of the cords toward the periphery (Fig. 4, G–I, and Supplemental Movie 4). Sertoli cells therefore became the main structural component of the individual cords with germ cells located centrally.

FIG. 6.

Live, isolated testis cord from an E13.5 Gata4-RG embryo in cross section. Sertoli cells inside the cords (dashed outline) are expressing mCherry in their nuclei (A and C) and EGFP in their membranes (B and C). Toward the exterior of the cords, Sertoli cells contribute to the cord border and appear polarized, with the nuclei in close proximity to the basement membrane surrounding the cord and cytoplasm extending inward (arrows). Sertoli cells toward the cord interior (arrowheads) exhibit no apparent polarity and appear squeezed in between germ cells (unlabeled areas). Bar = 50 μm.

Sertoli Cell Membranes Envelop Germ Cells During Cord Formation

When mice from the Gata4-RG-1 line were bred with mice from the Oct4-HS-CR line, Sertoli cells had red fluorescing nuclei and green plasma membranes, whereas germ cells had blue nuclei and red plasma membranes (Fig. 7). Therefore, the two cell types could be distinguished unequivocally by their fluorescent protein color code and their physical relationship described. Before the Sertoli cells moved to the cord periphery (E12.5), the two cell types were apparently uniformly mixed within the SGCM, with intimate contact between the different colored membranes. Some germ cells were completely enveloped by green plasma membranes (Fig. 7, D and H, and Supplemental Movie 5), indicating a very intimate relationship with Sertoli cells. Other germ cells were aligned in linear groups within the testis cords with no green plasma membranes between them but with green plasma membranes enveloping the entire cluster (Fig. 7 and Supplemental Movie 5). This suggests that different types of Sertoli cell-germ cell relationships exist in the testis cords.

FIG. 7.

Live testis cord from an E12.5 Gata4-RG::Oct4-CR embryo showing Sertoli cells (arrowheads) expressing mCherry in their nuclei and EGFP in their membranes, whereas the germ cells (arrows) are expressing mCherry in their membranes and Cerulean in their nuclei. Toward the cord interior, germ cells are clustered and often appear to have two nuclei surrounded by a single membrane (arrows in A–D). Throughout the cords, Sertoli cell membranes (double arrows in E–H) are in close contact with those of the germ cells (arrows) and appear to engulf the germ cells by extending their membranes toward the germ cells. Bar = 25 μm. See Supplemental Movie 5.

Between E11.5 and E13.5, the SGCM changes its shape to give rise to the testis cords. While doing so, Sertoli cells migrate outward to align and form the outer layer of the cords, and germ cells move toward the interior (Fig. 6). Sertoli cell nuclei remain close to the basement membrane, and their cytoplasms and cell membranes extend inward to fill the cord interior and engulf the germ cell clusters. At E13.5, however, a large population of Sertoli cells is still found among the internally localized germ cells (Fig. 6).

DISCUSSION

The combination of fluorescent proteins with live imaging in morphogenetic studies has become increasingly useful for studying cell behaviors in forming tissues. Much information is available regarding the development and use of animal models that express various FPs under the control of cell-specific promoters [23]. These models can be used to study cell- and tissue-type differentiation and organ formation. Several existing transgenic mouse lines express fluorescent proteins that label testicular cells, including germ cells [24], interstitial and peritubular myoid cells [25], endothelial cells [26, 27], and Sertoli cells [6]. We recently described a Sox9-EGFP knock-in mouse line to study the behavior of the SGCM during testis cord formation [6]. Although valuable information was obtained from studying how the SGCM behaved as a unit, this model did not allow individual cell distinction; therefore, cell behavior and interactions between neighboring cells could not be observed. The purpose of the current study was to generate a mouse model that can be used to follow testis cord formation and, specifically, Sertoli cell behavior by monitoring individual cells. We used a previously described red-green (nucleus-membrane) dual fluorescent protein reporter construct [19] to generate mice with Sertoli cells expressing red fluorescent nuclei and GFP plasma membranes under control of rat Gata4 regulatory sequences. Previously, a similar strategy was used to produce transgenic mice with ubiquitous dual-labeled cells [28] and, specifically, for labeling germ cell nuclei and membranes with different fluorescent proteins [19]. Our results demonstrate that this dual fluorescent protein reporter construct robustly labeled the nuclei and membranes of Sertoli cells, enabling us to identify the appropriate cell type and characterize their behavior.

We generated a total of six different transgenic mouse lines using the Gata4-RG construct. The Gata4-RG-1 line, which labels Sertoli cells, also has a non-Sertoli cell population within the interstitium that expresses the transgene. This differs from the results of Mazaud Guittot et al. [20], who reported the same rat Gata4 fragment only labeled Sertoli cells in the testis. We believe that this difference in expression pattern likely results from the influence of chromosomal location where the transgene integrated. Indeed, we found a wide variety of expression patterns using the Gata4-RG construct, suggesting that it is sensitive to position effects. Although we do not know what cell types are expressing the transgene in the interstitium, the Sertoli identity of the transgene-expressing cells within the testis cords is clear. One drawback of the Gata4-RG-1 line compared to our Sox9-EGFP knock-in mouse line is that the Gata4-RG transgene apparently becomes activated or expresses sufficient levels of fluorescent proteins only between E11.5 and E12.0, compared to E11.5 in Sox9-EGFP testes [6]. This suggests that the Gata4-RG-1 line is less useful for following Sertoli cells before E12.0, during the first testis differentiation events. Because GATA4 protein is detectable at E10.5 in the urogenital ridge [29], there may be a lag period of fluorescent protein accumulation in the appropriate cell compartment before fluorescent detection becomes possible [20]. This corresponds with the Sox9-EGFP knock-in line that only exhibits sufficient EGFP for visualization at E11.5, although endogenous Sox9 transcripts can be detected in the E10.5 urogenital ridge [6]. After expression starts, however, the Sertoli cell nuclei and membranes are clearly distinguishable in the Gata4-RG-1 line, which allows detailed description of cell behavior and makes this line especially valuable for studying cellular behavior after cord formation has been initiated.

In our live imaging movies, we could monitor Sertoli cell proliferation as the red fluorescing nuclei condensed into mitotic chromosomes. Sertoli cell proliferation, one of the earliest effects of Sry, is essential for cord formation [15, 16]. Although a brief cessation of proliferation occurs at E11.5, proliferation continues after E12.5 and throughout cord formation [15]. In the present study, Sertoli cell divisions were found in the transgenic fetal testes throughout the culture period; however, this was not quantified. Future experiments could focus on the rate of proliferation in the different regions of the forming cords.

Previously, we described the behavior of the SGCM as a single tissue mass composed of two cell types—namely, Sertoli and primordial germ cells. During cord formation, this tissue behaved as a unit, apparently being molded by surrounding tissues. Furthermore, we noticed various interesting cord morphologies during cord formation and witnessed their formation using live imaging [6]. In the current study, we witnessed Sertoli cell migration between connected cords. This corresponds to data from our previous studies in which neighboring cords were fused, branched, or connected by bridges [6]. Although tissue morphologies were described, the aforementioned study did not allow us to see how individual cells behaved in these cord structures. Our new, dual-labeled, Sertoli-specific transgenic mouse line allowed description of this Sertoli cell behavior and will be a useful tool in future studies of embryonic and adult Sertoli cells. These future studies will include visualizing Sertoli cell morphology in real time and the complex interactions of these somatic cells with spermatogenic cells.

Fetal testis cords are composed of an apparently uniform mixture of Sertoli and germ cells. In the current study, Sertoli cells initially had a fibroblast-like morphology, with apparently random migration events taking place throughout the SGCM. This corresponds with our previous hypothesis that the SGCM is shaped by external forces from surrounding tissues rather than from intrinsic Sertoli cell behavior [6]. Our movies indicate that after the cords have been defined and are increasing in length (after E13.5), the Sertoli cells migrate toward the periphery of the cords and become polarized. Individual cells migrated from the cord interior to the cord surface to form an outer cord lining with neighboring Sertoli cells. They also acquired an epithelial-like morphology, with the nuclei in closer proximity to the basal lamina and the membranes extending toward the future cord lumen. These observations suggest that genes regulating mesenchymal-to-epithelial transitions may be important downstream effectors of testis cord morphogenesis.

Our crossbreeding experiments showed that Sertoli membranes were folded intimately around individual germ cells and germ cell clusters so that the cords remained solid, without an apparent lumen, even after epithelialization of Sertoli cells. In the mouse, the lumen only appears around the third week after birth [30], so until then, individual germ cells may, theoretically, be in close contact with different Sertoli cells from different sides of the cord. It would be interesting to see how these germ cells finally segregate toward specific Sertoli cells later in development.

In summary, our dual fluorescent protein reporter system provides an innovative tool for following the behaviors of live cell types during testis differentiation. Labeling other testicular cell types (e.g., Leydig, peritubular myoid, and vascular cells) with other color codes should provide more clarity regarding the dynamics of cell differentiation and cell-cell interactions during gonadogenesis.