Characterization of a cell-surface differentiation antigen of mouse spermatogenesis: timing and localization of expression by immunohistochemistry using a monoclonal antibody

Spermatogenesis is an ideal system for the study of development because of several useful and well-documented characteristics. First, the system is relatively simple involving only two types of cells, the developing germ cells and the epithelial Sertoli cells. Second, there are clear morphological markers of the progress of cells in spermatogenic development. Third, because of the regular timing of entry of new generations of germ cells into spermatogenesis and the consistent progression of the germ cells through spermatogenesis, recurrent associations of germ cells are formed within each tubule section. These associations allow identification of the precise state of development of each germ cell to within plus or minus 22 h within the 34 ·5-day (mouse) differentiation sequence (Oakberg, 1956a,b). Lastly, relatively large amounts of material are available. Spermatogenesis is a developmental system which proceeds from stem cell to product in the adult, and moreover, does so repeatedly.

The cell surface (integral membrane components, junctions, peripherally associated components) is the means through which the cell interacts with its environment. It is thought to play important roles in regulating development and in producing differentiated function (Subtelny & Wessels, 1980). One approach to study the role of the cell surface in differentiation is analysis of differentiation antigens. Several populations of cell-surface differentiation antigens of testicular germ cells have been described using polyclonal immune sera, and recently monoclonal antibodies have been used to identify individual antigens or sets of antigens crossreactive at a single antigenic determinant. A set of antigens (defined by a rabbit anti-mouse spermatogonia serum exhaustively absorbed with somatic cells) is present on all of the germ cells while they reside within the epithelium of the testicular tubule (Millette, 1979; Millette & Bellvé, 1980). Presumably during elongation of the spermatids these antigens (determinants) become differentially distributed to the cytoplasmic lobe of the spermatid with little or none of these antigens remaining on the head or tail of the spermatid. At release of the spermatozoa from the epithelium, the physical budding off of the residual cytoplasmic body from the sperm separates most or all of this set of antigens from the sperm proper (Millette, 1979; Millette & Bellvé, 1980). Antigens which are not detected on spermatogonia, but are found on spermatocytes have been detected by several different immune sera (O’Rand & Romrell, 1977 ; Millette & Bellvé, 1977 ; Tung & Fritz, 1978; O’Rand & Romrell, 1981) and monoclonal antibodies (Bechtol, Brown & Kennett, 1979; Gaunt, 1982). Some of these antigens continue to be expressed on the sperm surface, while others become undetectable by the elongated-spermatid stage. Still other surface antigens are not detectable on either spermatogonia or spermatocytes and first appear on late spermatids (Tung, Bebe-han & Evan, 1979; O’Rand & Romrell, 1980). Quantitative and qualitative shifts in cell surface composition of germ cells as they progress through spermatogenesis have also been observed by comparing electrophoretic gel patterns of proteins and glycoproteins solubilized from germ cells separated according to size (developmental stage) (Boitani, Geremia, Rossi & Monesi, 1980; Millette & Moulding, 1981; Romrell, O’Rand, Sandow & Porter, 1982; Kramer & Erickson, 1982).

The XT-1 antigen, bound by monoclonal antibody XT-I (IgG_2a), is a testisspecific differentiation antigen of mouse germ cells. By quantitative-absorption radioimmunoassay using cell suspensions from whole testis of several juvenile ages and adult, the average concentration of XT-1 in the testis is seen to increase from undetectable in the 8-day juvenile mouse to a peak around 22 days after birth, followed by a decline to about 50 % of that concentration in the adult testis. Antigen is detectable on epididymal sperm in concentrations per mg of protein about 4 % of that in the adult testis (Bechtol et al. 1979; Bechtol, Jonak & Kennett, 1980). The antigen is found in the 13-day juvenile mouse, and this is consistent with its first appearance on either spermatogonia or early spermatocytes (Nebel, Amaroso & Hackett, 1961). The present study was undertaken to determine more precisely the developmental expression of XT-1 and to describe its distribution on spermatozoa. As seen in immunoperoxidase-stained tissue sections, the antigen becomes detectable on early (leptotene/zygotene) spermatocytes and increases in staining during the long pachytene spermatocyte phase. During haploid cell development the distribution of the determinant shifts from its relatively uniform surface distribution on spermatocytes to a more restricted localization on the base of the head, tail and cytoplasmic lobe of the elongating spermatid. The XT-1 determinant is thus a marker of spermatocytes and later germ cells and of a surface molecule which is related to their differentiation. It also provides a paradigm for alteration of the cell surface during spermatid elongation.

129/SV and BALBcJ mice were maintained as inbred colonies at The Wistar Institute. Results obtained using the two strains were indistinguishable. W^x/+(C3H/HeJ) (Russell, Lowson & Schabtach, 1957) and W^V/+(C57B1/6J) were obtained from Jackson Laboratory, Bar Harbor, Maine, and were bred to produce W^x/W^v black-eyed-white and W/-I-white-spotted heterozygotes.

The XT-I (IgG_2a) monoclonal anti-testis antibody was produced by a subclone of the original XT-I hybridoma (Bechtol et al. 1979) which has lost the γ₁ heavy chain of the P3-X63Ag8 myeloma parent. Antibody was concentrated from culture supernatant (HY medium, Kennett, 1980) by precipitation with 50 % saturated (NH₄)₂SO₄ or was purified from serum-free culture supernatant (Murakami et al. 1982) by pH 4 ·5 elution from a protein A affinity column (Ey, Prowse & Jenkin, 1978; Protein A Sepharose, Pharmacia). Negative controls were similar preparations of an IgG_2a monoclonal antibody, UK8, which does not bind testis, and of supernatants of the SP2/0 myeloma and 0 ·5 % bovine serum albumin. All three gave similar results.

Immunoperoxidase staining was performed on 5 μm sections of tissue fixed for 1 to 2 h in Bouin’s fixative or 20 % formalin, followed by 75 % ethanol overnight and standard embedding in wax. Paraffin was removed from the sections by the method of Sainte-Marie (1962), and immunoperoxidase staining was performed using affinity-purified, horseradish peroxidase-coupled rabbit anti-mouse Ig (Kirkegaard and Perry Laboratories, Inc.), diaminobenzidine (Aldrich Chemical Company, Inc.), and OsO₄ (Electron Microscopy Sciences), as described by Jonak (1980).

The XT-1 antigenic activity was not detectable after embedding in several water-permeable plastics tested. Therefore, to obtain 1 μm immunoperoxidase-stained sections, sections were processed as described above except that they were placed on plastic coverslips, and, following immunoperoxidase staining, they were embedded in epon and 1 μm sections were cut parallel to the plane of the coverslip.

The ‘stage’ of spermatogenesis present in each tubule cross-section was estimated from haematoxylin (H)- and eosin (E)-stained and periodic acid-Schiffs base (PAS)- and fast-green-stained sections serially adjacent to the immunoperoxidase-stained section using the criteria (e.g. morphology of spermatids), described in Oakberg (1956a). Photomicrographs of immunoperoxidase staining were taken with Tri-X Pan film, ASA 400, using an Olympus BH microscope with green filter.

Immunofluorescent staining was performed on cells derived from adult testis by mincing the tubules. Cells were incubated with XT-I or UK8 for 1 h at 0°C, washed three times, and resuspended in a 1/120 dilution of biotinylated-rabbit anti-mouse Ig (Vector Labs) for Ih at 0°C. The cells were then washed and incubated with FITC-avidin DCS (Vector Labs) for 1 h at O °C. The unfixed cells were observed with a Leitz Orthoplan microscope equipped with phase-contrast and a Phoem epi-illumination system. Photomicrographs were taken with Tri-X Pan film, ASA 800.

Immunoperoxidase-stained tubule sections from juvenile and adult testes demonstrate the developmental expression of XT-1 antigen in a segment of testis tubule (Fig. 1). The cell associations (12 stages, Oakberg, 1956a) of mouse spermatogenesis are numbered around the circle, and stained tubules in most of these stages are shown in the outer circle. This outer circle reconstructs the events in continuously cycling epithelium of the adult testis. As is well documented, new generations of germ cells enter spermatogenesis near the rim of the tubule and progress from mitotic spermatogonia through meiotic spermatocytes to haploid spermatids as time proceeds. In effect, the whole tubule section with time passes repeatedly around this cycle with the individual germ cell progressing farther in development and nearer the centre of the tubule until its release as a sperm in stage VII. Each turn of the circle requires 8 ·6 days of the 34 ·5-day development of the germ cell (Oakberg, 1956b).

The spiral part of Fig. 1 shows XT-I-stained sections of tubules from juvenile mice. Each cross section in a pie-shaped segment of the figure, for example in the pie segment including roman numeral IV, is in the same stage of spermatogenesis, though the more mature generations of germ cells are missing from some tubules, depending on the number of generations that have been initiated in that tubule. Some of the earlier cells are more easily seen in the juvenile tubules where there are fewer cells, the diameter of the tubule is less and the younger cells form layers several cells thick.

The complete development of expression of XT-1 on a single generation of germ cells can be seen by following that generation as it develops in the adult tubules of the outer circle of the diagram or by following the lumenal layer of germ cells in each tubule of the spiral proceeding from the centre tubule to the outer circle, as described below.

The central tubule of Fig. 1 (tubule A) and the three subsequent sections (B,C,D) contain only spermatogonia and immature Sertoli (somatic epithelial) cells, neither of which stain with XT-I.

The central cell layer of the next two tubules (E,F) contain antigen-positive leptotene/zygotene spermatocytes. Early-pachytene spermatocytes are developing in the following two tubules (G,H; stages I to V), and these show increased XT-1 expression. These spermatocytes are surrounded by a more peripheral layer of unstained spermatogonia. Intense XT-I staining is seen on the late-pachytene spermatocytes of the next two tubules (I,J; stages VIII through X). In those tubules the peripheral layers contain both lightly stained early spermatocytes and unstained spermatogonia.

Diplotene primary spermatocytes and spermatocytes in meiosis are found in the centre of the next two tubules (K,L; stages XI and XII). These cells are past the peak of XT-I staining (compare with the previous two tubules), though they still stain more intensely than the more peripheral zygotene spermatocytes in the same tubules. Two cells in this stage XII tubule are in the process of division. Such cells are intensely stained by the reagents in the absence of XT-I, and their antigen expression, therefore, cannot be determined in these experiments.

Antigen-positive, round spermatids are found in tubules of stages I through VIII following meiosis (i.e. M –P, the next four sections in Fig. 1). At stage IV (N), the round spermatids are approximately similar in staining to the more peripheral pachytene spermatocytes. As the tubule section progresses through stages VI and VII the round spermatids remain apparently unchanged in staining intensity, while the more peripheral pachytene spermatocytes increase in staining intensity as seen in the previous turn around the spiral.

At stage VIII, elongation of the spermatid’s nucleus begins. Along with this, the bulk of the spermatid cytoplasm begins to move caudally, away from the nucleus to form the cytoplasmic lobe which protrudes into the lumen of the tubule (Oakberg, 1956a). All tubules in the outer circle of Fig. 1 (Q –Y) contain elongating spermatids lining the lumen. Because the elongating cells are less than 5 μm in diameter, they are best observed in the thinner, 1 μm sections of Fig. 2. The antigen-positive spermatids in Fig. 2A are very near the beginning of elongation (as determined by adjacent serial sections stained with H and E, and with PAS and fast green). By Fig. 2B the spermatids have clearly begun to elongate and the cytoplasmic lobe of the spermatid is clearly visible by antigen staining (arrow). In Fig. 2D the elongating spermatids are near release and the antigen-positive midpiece of the tail, housing the mitochondria, is clearly visible (arrow).

A representative spermatid (C & D) and a released testicular sperm (A & B) stained in suspension using indirect immunofluorescence are shown in Fig. 3. The lobe of cytoplasm (large arrowhead, 3D) and the midpiece of the tail (from sperm head to small arrowhead, 3D and total length shown below head, 3B) show surface staining. The principal piece of the tail (from small arrowhead downward, 3D) also shows detectable uniform staining, though it appears much less intense, at least in part for geometric reasons since the distance looked through in that part of the tail is shorter than in the midpiece. There appears to be a small area of staining at the base of the head at the point of attachment of the tail. The remainder of the head is negative.

Immature Sertoli cells are not stained by XT-I (Fig. 1, centre tubule). To test for antigen expression on mature Sertoli cells, the genetically germ-cell-depleted testes of adult W7W^V mice were immunoperoxidase stained with XT-I. These mice are deficient in germ cells because in the embryo the primordial germ cells do not increase normally in number and few migrate to the genital ridge (Mintz & Russell, 1957). Nevertheless, the Sertoli cells mature, showing the characteristic changes in mitochondrial and nuclear morphology, appearance of specialized junctions and pumping of fluid (Handel & Eppig, 1979). XT-1 is not detectable in the testis tubules of adult W^x/W^v mice (Fig. 1, lower left insert), though it is expressed normally in their W/+ sibs (not shown).

In normal adult testes there are occasionally found sections of tubule in which one or more generations of germ cells is missing from the normal stage association or where no apparent spermatogenesis is occurring. When such sections are stained for XT-1, germ cells stain as expected based on their morphology, irrespective of absence of younger or older cells and the Sertoli cells remain unstained.

The XT-1 differentiation antigen of mouse spermatogenesis is expressed on the surface of spermatocytes and later germ cells. There is an increase in the intensity of surface immunostaining as germ cells proceed through spermatocyte differentiation. This presumably is due to an increase in concentration of the XT-1 determinant on the cell surface, and not to an optical artifact based on the increasing size of the cell. All of the spermatocytes have diameters greater than the 5 μm thickness of the sections. Moreover, at high magnification (×1000), examination of individual cells over their surface in many different focal planes confirms the staining difference between cells in different stages observed at lower magnification. The XT-1 determinant continues to be expressed after meiosis, on spermatids, and at elongation of the spermatid, detection of the determinant becomes restricted to the base of the head, cytoplasmic lobe and tail. The determinant(s) detected by XT-I binding on spermatocytes and sperm are probably identical or nearly identical since by Scatchard plot the antibody binds to both cells with the same K₀ (unpublished observation).

Two other individual antigens of spermatogenesis which are also expressed on sperm have been identified, RSA-1 (rabbit) and 1B3 (mouse). Both of these stain the whole sperm surface with some variation in intensity and patchiness of the distribution (O’Rand & Romrell, 1981; Gaunt, 1982). Surface antigens that arise during spermatogenesis and are then found on the spermatozoon can either be distributed over the entire sperm surface (O’Rand & Romrell, 1981; Gaunt, 1982) or restricted to one or more areas of the sperm head and tail (Millette & Bellvé, 1977; O’Rand & Romrell, 1980; Tung, Yanagimachi & Yanagimachi, 1982). Several domains of sperm surface antigens (intrinsic membrane components and molecules adhering from fluids of the reproductive tract) have been described, including the entire sperm head, the acrosomal and post-acrosomal regions, the whole tail, the midpiece and the principal piece of the tail (e.g. Myles, Primakoff & Bellvé, 1981; Feuchter, Vernon & Eddy, 1981; Schmell, Gulyas, Yuan & August, 1982a,b). The restricted distribution of XT-1 staining on sperm may reflect 1) a redistribution of the existing surface molecule bearing the XT-1 determinant or restricted localization of incorporation of new antigen molecules during spermatid elongation, 2) chemical modification of the antigenic molecule on the developing head which alters the XT-1 determinant so that it no longer binds the XT-I antibody, and/or 3) masking of the XT-1 determinant by other molecules newly associated with it.

The XT-1 antigen is detected on the surface of germ cells in situ (Figs 1 –3) and in cell suspension (Bechtol, Jonak & Kennett, 1980), provided the cells have not previously been exposed to proteolytic enzymes (Bechtol et al. 1979). The antigen is not detected on juvenile Sertoli cells in normal testis, on Sertoli cells in aspermatogenic segments of normal adult testis, or on Sertoli cells in genetically germ-cell-depleted adult mice. The XT-1 determinant thus appears to be a marker for spermatocytes and later germ cells, though it cannot be ruled out from the above data that Sertoli cells express the antigen where they contact or lie near the antigen-positive germ-cell stages. Detection of the antigen is in any case dependent on the presence of germ cells of the appropriate developmental stage.

The monoclonal antibody (XT-I) and the antigenic moiety it binds can serve 1) as useful markers for spermatocytes and later germ cells, 2) as markers for at least one form of a surface molecule which is quantitatively related to male germcell differentiation, and 3) as markers of a molecular redistribution or modification which occurs at the surface of the germ cell during spermatid elongation. The description presented herein serves as a foundation on which to build molecular and functional studies of the antigen.

This research was supported by grants GM-23892 and CA-10815 from the National Institutes of Health. The author wishes to thank Ms Marian Butts and Ms Elsa Agio for their excellent technical assistance, Mr Joe Weibel for embedding and cutting the 1 /im plastic sections, and Drs Barry Zirkin and Gerd Maul for their interested discussion.