Pachytene spermatocyte and round binding to Sertoli cells in vitro

GEORGE C. ENDERS* and CLARKE F. MILLETTE

Laboraturv of and Reproductive Biology, Department of Anatomy anil Cellular Hiohigy, llan-ard Medical School, Hoston, MA 021 IS, USA

•Author for correspondence at: Harvard Medical School, 45 Shattuck Street, Boston, MA 02115, USA

Summary

Spermatogenic cells differentiate in vivo while in spermatocyte and round spermatid binding sites continuous contact with the . During on the Sertoli cell monolayer were similar, but the differentiation, the spermatogenic cells and Ser- kinetics of binding were different. Pachytene toli cells form a number of morphologically spermatocytes were able to inhibit significantly distinct stage-specific adhesions. We describe an round spermatid binding, while round sperma- in vitro assay system for studying the adhesion of tids did not significantly inhibit pachytene sper- spermatogenic cells to Sertoli cell monolayers. matocyte binding. After coculture for 24-48 h, Mixed populations of spermatogenic cells or en- spermatocytes form junctional structures with riched fractions of pachytene spermatocytes and Sertoli cells that are similar to desmosome-like round were labelled with the vital junctions. These results suggest that pachytene dye, fluorescein diacetate, prior to their addition spermatocytes and round spermatids bind to Ser- to Sertoli cell monolayers so that the adhesion of toli cells by different mechanisms. viable spermatogenic cells could be quantified. Using this assay system, the number of pachytene Key words: Sertoli cell, spermatogenic cell, adhesion.

Introduction junctional specializations, which are unique to the testis, are formed between the Sertoli cell and round is the process of male and elongating spermatids (Flickingcr & Fawcett, proliferation and differentiation that occurs in continu- 1967; Ross 6c Dobler, 1975; Ross, 1976; Russell, ous contact with a , the Sertoli cell. 19776, 1979; Romrell & Ross, 1979). Originally Throughout differentiation the spermatogenic cells referred to as junctional specializations by Flickingcr & and the Sertoli cells share morphologically distinct Fawcett (1967), and later as ectoplasmic specializations plasma membrane junctions (for review, see Russell, by Russell (19776), they appear similar to half of a 1980). There are at least five morphologically distinct Sertoli-Sertoli cell junction within the Sertoli cell and cell-cell interactions, each present throughout certain form on the plasma membrane overlying the acrosomal stages of spermatogenic cell differentiation. Desmo- membrane of the developing spermatid. These special- some-like junctions are seen between the Sertoli cells izations hold elongating spermatids to Sertoli cells and the following spermatogenic cells: spermatogonia, prior to release into the lumen (Russell, 19776; Rom- primary and secondary spermatocytes, and early round rell & Ross, 1979) and thus will be referred to as spermatids (Altorfer

Journal of Cull Science

106 G. C. Enders and C. /•'. Millette before being added to the Sertoli cell monolayers. During the contrast microscopy. Acridine Orange was occasionally washing procedures the spertnatogenic cell numbers were used to aid the process of cell counting and to examine adjusted so that lml of medium was added to 9 ml of for myoid cell contamination. Sertoli cells are easily DMEM/F-12 plus 10% FCS already present in the distinguishable from peritubular myoid cells because of 60 mm X IS mm culture dishes containing the coverslips. their extensive perinuclear acidic compartments. Local The culture dishes were then placed in a tray on a orbital cell density varied, but generally was between 1 -2x 104 shaker table (~30 cycles min~', ~linch of rotation) for 4 2 60 min at room temperature. Alternative procedures are and 5X 10 cells cm~ , meaning individual Sertoli cells 2 described in Results. At 15-min intervals the tray containing occupied between 1600 and 8300 jum , and thus are the culture dishes was gently agitated linearly by hand to much more squamous than Sertoli cells /// vivo. redistribute cells from the centre of the dishes. The tray was used to ensure that each dish received the same agitation. Effect, of temperature After 60 min the coverslips were removed from the culture Mixed populations of cultured spermatogenic cells (not dishes and placed into DMEM/F-12 in coverslip Coplin jars. separated by unit gravity sedimentation, but cultured To remove loosely adherent spermatogenic cells, each cover- overnight) were used to examine the effect that tem- slip was dropped from just above the surface of the perature had on adhesion of spermatogenic cells to DMEM/F-12 twice into three separate Coplin jars. The enriched Sertoli cell monolayers. It was found that coverslips were then carefully placed cell side down on slides similar numbers of mixed germ cells bound to the in a humidified box kept at 4°C or on ice until examined Sertoli cell monolayer when incubated at either 24°C or (generally within 4 h) for epifluorescence with fluorescein 33 °C, while incubation at 4°C significantly reduced filter combinations. The total number of fluorescently labelled cells in 10 randomly chosen fields (6-3x or 10X binding (Fig. 1). If the culture dishes were gently objective) on each coverslip was counted, resulting in four shaken at room temperature during the incubation values for each data point unless otherwise indicated. In some period of 60 min, a significant increase in the number of instances, especially when large numbers of adherent cells adherent spermatogenic cells was observed. Thus, all were present, each field was photographed or the video image future incubations were performed with gentle agi- was stored on laser disk for later analysis. In these cases the tation at room temperature. number of cells in 8-10 fields on four coverslips was independently recorded. The statistical significance of the Effect of time on adhesion results was determined by using Student's /-test. The effect of incubation time on spermatogenic cell— Sertoli cell adhesion was examined. When mixed Ultrastmctural techniques populations of spermatogenic cells were gently agitated In order to examine the nature of the adhesion between on top of enriched Sertoli cell monolayers at room spermatogenic cells and Sertoli cells at the ultrastructural temperature, cell binding peaked after about one hour level, mixed populations of spermatogenic cells were added to of incubation (Fig. 2). Two additional experiments Sertoli cell monolayers, which were grown directly on tissue (not illustrated), which continued the incubation culture plastic as opposed to coverslips. Mixed populations of spermatogenic cells were added to the Sertoli cell monolayers periods for up to 5h, did not show an increase in and gently shaken as described above. Non-adherent sperma- spermatogenic cell binding above that observed at 60 or togenic cells were removed by rinsing the culture dishes three Temperature effect to four times with fresh DMEM/F-12, and then placed in DMEM/F-12 plus 10% FCS and cultured for 24-48 h. The 400- medium was then removed and the cells fixed with 1 %

Spermatogenic cell-Sertoli cell binding 107 Effect of time of incubation Specificity of cell binding 600 800 • 0No. of PSC bound QNo. of RST bound 600-

400-

200- * **r?7\ V 1 V?M(/A 30 60 90 120 150 Sertoli 3T3 Coverslip Trypsin Time of incubation (min) Different substrata

Fig. 2. The time of incubations of spermatogenic cells Fig. 3. Spermatogenic cells adhere preferentially to Sertoli with the enriched Sertoli cell monolayers affects the cells and few bind to either 3T3 cells or glass coverslips. In number of spermatogenic cells that bind to the Sertoli cell addition, when pachytene spermatocytes and round monolayer. 3'6X 10"1 spermatogenic cells per cm2 were spermatids were trypsinized (1 mgml"1 in DMEM/F-12 added to Sertoli cell monolayer. In this experiment n = 6. for 30 min at room temperature, then washed three times in The error bars = standard error of the mean. DMEM/F-12 plus 10% FCS) subsequent binding to the Sertoli cell monolayer was reduced substantially. 4-6X 104 pachytene spermatocytes (PSC) per cm were used; 90 min. Thus, for all the following experiments the ** =P<0-0001 when compared with binding to Sertoli cell adhesion assay was performed at room temperature monolayer (control). 4-OxlO4 round spermatids (RST) per with gentle agitation, for 60 min. cm were used; * =P<0'01 when compared with binding to Sertoli cell monolayer (control). In this experiment Specificity of the adhesion ;; = 6. The error bars = standard error of the mean. The specificity of spermatogenic cell—Sertoli cell ad- hesion was examined by placing either pachytene spermatocytes or round spermatids on other substrata. Both spermatogenic cell types showed little affinity for bound was equal to or greater than the number of either 313 cells, bare coverslips, or the enriched Sertoli pachytene spermatocytes binding (data not illus- cell monolayer if the spermatogenic cells were first trated). trypsinized (Fig. 3). In control preparations the pachy- tene spermatocytes and round spermatids cells that adhered to the Sertoli cell monolayer generally bound Competitive binding studies with a seemingly random distribution to the Sertoli In order to determine if binding of pachytene sper- cells (Figs 4, 5). While some clumping of cells was matocytes and round spermatids occurred indepen- seen, the cell clumps were counted as single cells. dently of the presence of each other, the following experiments were performed. A given number of Saturation binding studies pachytene spermatocytes that had been labelled with The numbers of pachytene spermatocytes and round fluorescein diacetate were added to the enriched Sertoli spermatids added to enriched Sertoli cell monolayers cell monolayers along with twice the number of un- were varied over a wide range to examine the binding labelled round spermatids. The reciprocal experiment kinetics of the two cell types (Fig. 6). When the total was also performed; that is, fluorescently labelled number of spermatogenic cells added was low (below round spermatids were added to the Sertoli cell mono- =5xl05cm~2), a greater percentage of pachytene layer along with twice the number of unlabelled spermatocytes bound than round spermatids in eight pachytene spermatocytes. Unlabelled pachytene sper- out of 10 trials in eight different experiments (different matocytes were able to inhibit in a statistically signifi- Sertoli cells and spermatogenic cells on different cant manner labelled round spermatid binding dates). The data are not illustrated. However, when (Fig. 7). However, the converse was not true. Un- the number of spermatogenic cells added was large labelled round spermatids did not significantly inhibit (above ~5Xl05cm~2), the total number of round the binding of labelled pachytene spermatocytes. In spermatids that bound often exceeded the number of four separate experiments pachytene spermatocyte pachytene spermatocytes that bound (Fig. 7; Table 1). binding was reduced an average of 11 % by unlabelled In four out of six separate experiments when more than round spermatids, while round spermatid binding was 5x 103 cells cm~2 were added to enriched Sertoli cell reduced an average of 72% by the addition of un- monolayers, the number of round spermatids that labelled pachytene spermatocytes (Table 1).

108 G. ('. Enders and C. F. Millette Fig. 4. Fluorescent photomicrograph of fluorescein diacetate-labelled pachytene spermatocytes adhering to an enriched Sertoli cell monolayer. Only cells that fluoresced strongly were counted. The percentage of fluorescent pachytene spermatocytes was generally 88-94%. Clumps of cells were counted as one cell, not two or several. X49. Fig. 5. Fluorescent photomicrograph of fluorescein diacetate-labelled round spermatids adhering to an enriched Sertoli cell monolayer. The intensity of the round spermatids was more variable than that of the pachytene spermatocytes. Typically 84-91 % of the round spermatids fluoresced. Faintly fluorescent cells were not counted and clumps of round spermatids (there are at least 16 shown here) were counted as one adherent cell. X49.

Competitive binding study Effect of cell number 100 1400

60-

40-

20-

PSC PSC RST RST 0 10 20 30 40 + 2XURST +2XUPSC No. of cells x 10s added cm"2 Cell type added

Fig. 6. The total number of pachytene spermatocytes (•) Fig. 7. Unlabelled round spermatids show little ability to and round spermatids (•) which bound to Sertoli cell compete for pachytene spermatocyte binding sites on monolayers saturated at similar levels. At low cell numbers Sertoli cells, while unlabelled pachytene spermatocytes (below =5xlO5 cells added per cm2) a greater percentage significantly inhibited the binding of round spermatids of pachytene spermatocytes bound than round spermatids. (* = p< 0-00001). 8-4X105 labelled pachytene In other experiments not illustrated the saturating level of spermatocytes (PSC) or round spermatids (RST) per cm" round spermatids to bind exceeded that of pachytene were added to the Sertoli cell monolayers. 2XURST = spermatocytes. In this experiment n = 4. The error 16-8X105 unlabelled round spermatids; 2xUPSC = bars = standard error of the mean. 16-8X105 unlabelled pachytene spermatocytes. In this experiment // = 34. The error bars = standard error of the Ultrastrnctural observations mean. Despite the fact that mixed populations of spermato- genic cells were added to the Sertoli cell monolayers, increases in the electron density of Sertoli cell cyto- by the time the cells were cocultured for 24-48 h, fixed, plasm (Fig. 9). Endoplasmic reticulum was not associ- and processed for thin-section electron microscopy, ated with the regions of cell-cell contact. Spermatids spermatocytes were the exclusive cell type observed were not observed adhering to the Sertoli cell mono- adhering to the Sertoli cell monolayer (Fig. 8). Sper- layers. matocytes were typically held to underlying Sertoli cells by regions of cell-cell contact that displayed small Discussion quantities of electron-dense extracellular material be- tween adjacent cells. In some instances the areas of The cell—cell adhesion assay we introduce in this paper Sertoli cell-spermatocyte contact also displayed focal has a number of unique features not present in other

Spennatogenic cell—Sertoli cell binding 109 Table 1. Competitive binding assay

Number of germ cells bound/field to Sertoli cell monolavers Experiment % PSC binding % RST binding (no. of labelled germ PSC RST reduced by reduced by cells added/cm2) PSC + 2xUnRSr r RST + 2xUnPSC n +2xUnRST + 2xUnPSC 1 (6-25 X105) 67 ±5 53 ±9 111=+= 10 27 ±3 34 21% 76%** 2 (8-375xlO5) 75 ±4 67 ±4 86 ±5 40 ±4 34 11% 54 %** 3 (9-875x10s) 51+3 53 ±3 85 ±9 26 ±2 34 (4%) 70%** 4 (9-625X105) 244 ± 56 233 ± 33 250 ± 56 35 + 12 4 9% 86 %* Average reduction

11 % ± 4 72% ±8

PSC, pachytene spermatocyte; RST, round spennatids; Un, unlabelled spermatogenic cells. Values arc mean ± standard error of the mean; ( ), incrca.- binding. ;/, the number of microscopic fields counted (experiments 1-3) or the sum of 10 microscopic fields (experiment 4). * = /J< 0-006; *• = /'< 0-00001; comparing RST binding with RST+UnPSC binding.

assays used to examine Sertoli cell—spermatogenic cell vary from one experiment to the next, despite the fact interactions. Our assay is a fast, quantitative method that the relative trends in binding (e.g. pachytene (time of incubation is only 1 h) that does not rely on the spermatocytes versus round spermatids) was highly use of radioisotopes. Most other Sertoli cell—spermato- reproducible. Within the literature descriptions of genic cell adhesion assays used require incubation spermatogenic cell—Sertoli cell adhesion are very vari- periods of 3-24 h or more (Ziparo et al. 1980; Tres & able. Ziparo et al. (1980) reported that about 90% of Kierszenbaum, 1983; Galdieri et al. 1984; Le Maguer- pachytene spermatocytes but only 10-30% of round esse et al. 1986), with the exception of the assay of spermatids bound to Sertoli cell monolavers when cell Grootegoed et al. (1982), who used concanavalin A to numbers added were below 4x 105 cells cm"2. Rivarola induce adhesion, and that of D'Agostino et al. (1984), et al. (1985) reported that only 20-40% of the germ who reported rat pachytene spermatoeyte binding to cells added (5x 105 cell ml"1) adhered to Sertoli cell rat Sertoli cells saturated after 90min at 33 °C. The monolayers after coculture periods of 4 and 24 h. They spermatogenic cell-Sertoli cell adhesion in our assay state that in their assay system "very few germ cells may require less time than others have reported for a were observed on the surface of Sertoli cells after number of reasons. First, we use gentle agitation collection" (of the spermatogenic cells). Grootegoed et during our assay to increase the number of cell-cell al. (1982), using mouse pachytene spermatocytes and collisions during the assay. Second, we preincubate our round spermatids, reported that only 10-15% of cells spermatogenic cells overnight in order to allow them to initially added (2-5 X 105 cells cm"2) to the Sertoli cells resynthesize surface components that may have been bound to the monolayers after coculture for up to 20 h cleaved enzymically during the isolation procedure. in medium 199. This is important, since the enzymic treatment used to Previous investigations using both pachytene sper- isolate the various cell types destroys the cell-cell matoeyte and round spermatid cells in binding studies adhesion present in vivo. Third, our assay uses mouse reported that a greater percentage of added pachytene cells, while most previous work has been done using rat spermatocytes than round spermatids were bound cells, which may display species differences. Moreover, (Ziparo et al. 1980; Galdieri et al. 1984; Le Maguer- our assay procedure is designed to examine the early esse, 1986; D'Agostino & Stefanini, 1987). In these recognition events between Sertoli cells and spermato- experiments spermatogenic cell densities of genic cells, while much of the previously described 3 2 4xl0 cm~ or less were generally used. We found work was designed to study other aspects of Sertoli similar results at these cell densities. However, at cell-spermatogenic cell interactions, such as changes in 2 higher germ cell density (5X lO'cm" or more), equal Sertoli cell secretory activity in response to spermato- or greater numbers of round spermatids than pachy- genic cells (Le Magueresse et al. 1986). tene spermatocytes were bound. The slope of round During the development of this assay we had trouble spermatids binding versus the number of cells added with variability, both within a given experiment, and was less than that for pachytene spermatocytes, sugges- from one experiment to the next. Much of the intraex- ting that Sertoli cells have a lower affinity for round periment variability was reduced by paying close spermatids than pachytene spermatocytes, but similar attention to the details of the assay. However, the or slightly greater number of binding sites for round absolute number of spermatogenic cells adhering to a spermatids. In vivo each pachytene spermatoeyte gives given Sertoli cell monolayer preparation still tends to rise to four round spermatids, suggesting that on

110 G. C. Enders and C. F. Millette average each Sertoli cell should bind four times as that there are 2-4 pachytene or diplotene spermatocytes many round spermatids as pachytene spermatocytes. and 7-9 step 1-10 round spermatids per rat Sertoli cell Indeed, morphometric study of the rat testis indicates (Wing & Christensen, 1982). Yet a given Sertoli cell

A

Fig. 8. A pachytene spermatocyte (p) adhering to a Sertoli cell (5) grown on tissue culture plastic. The pachytene spermatocyte was cocultured with the Sertoli cell for 24 h prior to fixation. Note, while there are several regions of contact, the only specialization appears to be a slight increase in the electron density of material between the two adjacent phospholipid bilayers, and this only occurs in focal spots. X 12 000. Fig. 9. A spermatocyte adhering to a Sertoli cell (5) grown on tissue culture plastic cocultured for 24 h. Note the increase in electron density of the extracellular material between the two plasma membranes of adjacent cells. There is also an increase in the electron density of the cytoplasm of the Sertoli cell adjacent to the regions of cell-cell contact (arrowheads). X 25 000.

Spennatogenic cell—Sertoli cell binding 111 makes many more cell-cell contacts, since diploid Sertoli cells (Russell et al. 1983). About half of the step spermatogenic cells are frequently in contact with more 5 round spermatids examined had desmosome/gap than one Sertoli cell. The stage V Sertoli cell that junctions, and about one third had ectoplasmic junc- Russell and covvorkers (Russell et al. 1983) semiserial tional specializations. The desmosome/gap junctions sectioned made contact with eight pachytene spermato- that were present in round spermatids were much cytes and 27 step 5 round spermatids. Therefore, it is smaller (about 0-5 fim) than those present in pachytene somewhat surprising that previous workers have spermatocytes (up to 1-2,14m). This may explain the reported larger numbers of pachytene spermatocyte lower slope of the round spermatid binding curve binding sites than those for round spermatids. This compared to that for pachytene spermatocytes. Russell result may be due to a limited ability of round (1980) believes that the ectoplasmic junctional special- spermatids to resynthesize surface components respon- ization does not serve an adhesive function until the sible for cell binding compared to that of pachytene acrosomal membrane contacts the plasma membrane spermatocytes. Gerton & Millette (1986) have, how- within the step 7 round spermatid. Our round sperma- ever, demonstrated that both pachytene spermatocytes tid fraction contains steps 1-8 round spermatids. As and round spermatids are capable of in vitro synthesis differentiation in vivo progresses, each spermatid be- of plasma membrane proteins. In our assay at saturat- comes held to the Sertoli cell by one ectoplasmic ing levels there were only slightly larger numbers of junctional specialization per spermatid, and loses all bound round spermatids than pachytene spermatocytes desmosome-like junctions. Perhaps the initial adhesion on the Sertoli cell monolayers. We still are not able to events that ultimately develop into an ectoplasmic mimic the /;/ vivo ratio of pachytene spermatocytes to junctional specialization are much weaker than the round spermatids bound per Sertoli cell. This may be initial adhesion events that ultimately develop into due to the fact that the Sertoli cells used in our assay desmosome-like junctions. It is also possible that only a were isolated from immature animals; however, pro- limited subset of our round spermatid fraction is cedures for rcproducibly isolating Sertoli cells from capable of adhering to Sertoli cells; say, round sperma- adult animals are not available, to our knowledge. We tids of steps 7 and 8. Alternatively, the slightly higher estimate that at most four to five pachytene spermato- number of round spermatid binding sites versus pachy- cytes and round spermatids bind per Sertoli cell in our tene spermatocyte binding sites may thus be due to in vitro assay. Perhaps a closer approximation to the round spermatid binding to both desmosome-like rec- situation in vivo may be achieved by using Sertoli cell ognition sites and ectoplasmic-junctional-specializ- monolayers cultured in bicameral chambers on extra- ation-like recognition sites, each with a different subset cellular matrix so that the epithelial cells assume a more of round spermatids. The binding of pachytene columnar morphology. spermatocytes to desmosome recognition sites might sterically inhibit round spermatid binding to both In our competitive binding experiments pachytene desmosome-like and ectoplasmic-junctional-specializ- spermatocytes were able to inhibit round spermatid ation-like recognition sites on the Sertoli cells /// vitro. binding, but round spermatids did not inhibit pachy- tene spermatocyte binding. This suggests that either With this cell binding assay we cannot accurately the two cell types share some common binding site, determine the extent of non-specific binding as can be with pachytene spermatocyte-Sertoli cell binding be- done for other receptor-ligand binding systems by ing of higher affinity than round spermatid-Sertoli cell adding 100-fold excess of unlabelled ligand, due to binding, or that round spermatids bind a separate site, limited cell numbers available. Thus, an alternative which is both of lower affinity and sterically hindered explanation of the data could be that round spermatids by the presence of bound pachytene spermatocytes. bind with lower affinity to Sertoli cells and also D'Agostino & Stefanini (1987) reported that an anti- undergo more non-specific binding than do pachytene serum raised against rat pachytene spermatocytes, spermatocytes. which inhibits their in vitro binding to Sertoli cell Further differences between pachytene spermato- monolayers, retains its ability to inhibit pachytene cyte and round spermatid binding are evident from the spermatocyte binding even when preabsorbed against ultrastructural observations. When mixed spermato- spermatids. In vivo, rat pachytene spermatocytes are genic cell populations were added to the Sertoli cells bound to stage V Sertoli cell by one to four desmosome- monolayer and gently shaken for 60min, then fixed /gap junctions, with an average of 2-4 desmosome/gap immediately with paraformaldehyde and glutaralde- junctions per pachytene spermatocyte (Russell et al. hyde and processed for electron microscopy, very few 1983; the magnification used for this ultrastructural spermatogenic cells adhere to the monolayer (unillus- study was not sufficient to distinguish between desmo- trated observations). When adhering spermatogenic somes and gap junctions). Step 5 round spermatids, on cells are cocultured for 24-48 h and then fixed and the other hand, form both desmosome/gap junctions processed for electron microscopy, spermatocytes are and ectoplasmic junctional specializations with rat observed adhering to Sertoli cells by structures similar

112 G. C. Enders and C. F. Millette to desmosome-like junctions, which are present in model for the study of somatic-germ cell interactions. .7- vivo. However, round spermatids apparently do not Andml. 5, 249-254. form junctions stable enough to be retained through GERTON, G. L. & MILLETTE, C. F. (1984). Generation of culture and processing for electron microscopy. The flagella by cultured mouse spermatids. jf. Cell Biol. 98, round spermatids' lack of ability to form stable junc- 619-628. tions may be due to a more limited ability of ectoplas- GERTON, G. L. & MILLETTE, C. F. (1986). Stage-specific synthesis and fucosylation of plasma membrane proteins mic junctional specialization sites to form stable adher- by mouse pachytene spermatocytes and round spermatids ent junctions in vitro, or that desmosome-like junction in culture. Biol. Reprod. 35, 1025-1035. adhesion mechanisms normally disappear rather than GROOTEGOED, J. A., JUTTE, N. H. P. M., ROMMERTS, F. form at this stage in development. Ultimately the use of F. G., VAN DER MOLEN, H. J. & OHNO, S. (1982). our assay system in conjunction with inhibitory anti- Concanavalin A-induced attachment of spermatogenic sera that could specifically inhibit the binding of a cells to Sertoli cells in vitro. Expl Cell Res. 139, given stage of spermatogenic cells should be able to 472-475. answer some of the questions raised by our results. KAYA, M. & HARRISON, R. G. (1976). The infrastructure relationships between Sertoli cells and spermatogenic This work was supported by National Institutes of Health cells in the rat. J. Anal. 121, 279-290. grants HD06468 to G.C.E. and HD 15269 to C.F.M. LE MAGUERESSE, B., LE GAC, F., LOIR, M. & JEGOU, B. (1986). Stimulation of rat Sertoli cell secretory activity in vitro by germ cells and residual bodies..7. Reprod. Vert. 77, 489-498. References MATHER, J. P. & PHILLIPS, D. M. (1984). Methods for Serum-Free Culture of Cells of the Endocrine System ALTORFER, J., FUKUDA, T. & HEDINGER, C. (1974). (ed. D. W. Barnes, D. A. Sirbasky & G. H. Sato), Desmosomes in human seminiferous epithelium. p. 29. New York: Alan R. Liss. Virchows Arch. 13 Cell Path. 16, 181-194. McGlNLEY, D. M., POSALASKY, Z., PORVAZN1K & BELLVI5, A. R., CAVICCHIA, J. C, MILLETTE, C. F., RUSSELL, L. (1979). Gap junctions between Sertoli and O'BRIEN, D. A., BHATNAGAR, Y. M. & DYM, M. (1977). germ cells of rat seminiferous tubules. Tissue & Cell 11, Spermatogenic cells of prepuberal mouse. Isolation and 741-754. morphological characterization. J. Cell Biol. 74, 68-85. MILLETTE, C. F. & BELLVE, A. R. (1977). Temporal D'AGOSTINO, A., MONACO, L., STEFANINI, M. & GEREMIA, expression of membrane antigens during mouse R. (1984). Study of interaction between germ cells and spermatogenesis.J- Cell Biol. 74, 86-97. Sertoli cells in vitro. Expl Cell Res. 150, 340-435. MILLETTE, C. F. & MOULDING, C. T. (1981). Cell surface D'AGOSTINO, A. & STEFANINI, M. (1987). A rat marker protein during mouse spermatogenesis: Two- spermatocyte surface protein is involved in adhesion of dimensional electrophoretic analysis..7- Cell Sci. 48, pachytene spermatocytes to Sertoli cells in vitro. Molec. Cell. Biol. 7, 1250-1255. 367-382. DEPHILIP, R. M. (1985). Germ cell binding to rat Sertoli MILLETTE, C. F. & SCOTP, B. K. (1984). Identification of cells. J. Cell Biol. 101, 373a. spermatogenic cell plasma membrane glycoproteins by two-dimensional electrophoresis and lectin blotting. 7- EDDY, E. M. & KAHRI, A. I. (1976). Cell associations and surface features in cultures of juvenile rat seminiferous Cell Sci. 65, 233-248. tubules. Anal. Rec. 185, 333-358. MlSHELL, B. B. & SHIIGI, S. M. (1980). Select Methods in ENDERS, G. C, HENSON, J. II. & MILLETTE, C. F. (1986). Cellular Immunology (ed. B. B. Mishell & S. M. Shiigi), Sertoli cell binding to isolated testicular basement p. 19. New York: W. II. Freeman & Co. membrane. J. Cell Biol. 103, 1109-1119. PALOMBI, F., ZIPARO, E., GALDIERI, M., RUSSO, M. A. & ENDERS, G. C. & MILLETTE, C. F. (1984). An in vitro STEFANINI, M. (1980). Ultrastructural evidence for germ assay for spermatogenic cell—Sertoli cell adhesion. J. Cell cell-Sertoli cell interaction in culture of rat seminiferous Biol. 99, 392a. epithelium. Biol. Cell. 39, 249-252. FAWCETT, D. W. & PHILLIPS, D. M. (1969). Observations PALOMBI, F., ZIPARO, E., ROMMERTS, F. F. G., on the release of spermatozoa and on changes in the head GROOTEGOED, J. A., ANTONINI, M. & STEFANINI, M. during passage through the . .7. Reprod. Vert. (1979). Morphological characteristics of male germ cells (Suppl.) 6, 405-418. of rats in contact with Sertoli cells in vitro..7- Repmd. FLICKINGER, C. & FAWCETT, D. W. (1967). The junctional Vert. 57, 325-330. specializations of Sertoli cells in the seminiferous RIVAROLA, A. M., SANCHEZ, P. & SAEZ, J. M. (1985). epithelium. Anal. Rec. 158, 207-222. Stimulation of ribonucleic acid and deoxyribonucleic acid GALDIERI, M., MONACO, L. & STEFANINI, M. (1984). synthesis in spermatogenic cells by their coculture with Secretion of androgen binding protein by Sertoli cells is Sertoli cells. Endocrinology 117, 1796-1802. influenced by contact with germ cells. J. And ml. 5, ROMRELL, L. J., BELLVE, A. R. & FAWCETT, D. W. (1976). 409-415. Separation of mouse spermatogenic cells by GALDIERI, M., ZIPARO, E., PALOMBI, F., RUSSO, M. A. & sedimentation velocity. A morphological characterization. STEFANINI, M. (1981). Pure Sertoli cell cultures: A new Devi Biol. 49, 119-131.

Spermatogenic cell-Sertoli cell binding 113 ROMRELL, L. J. & Ross, M. H. (1979). Characterization of RUSSELL, L. D. & MALONE, J. P. (1980). A study of Sertoli cell-germ cell junctional specializations in Sertoli-spermatid tubulobulbar complexes in selected dissociated testicular cells. Anat. Rec. 193, 23-42. mammals. Tissue & Cell 12, 263-285. Ross, M. H. (1976). The Sertoli cell junctional RUSSELL, L. D., TALLON-DORAN, M., WEBER, J. E., specialization during sperniiogenesis and at spermiation. WONG, V. & PETERSON, R. N. (1983). Three- dimensional reconstruction of a rat stage V Sertoli cell: Anat. Rec. 186, 79-104. III. A study of specific cellular relationships. Am.J. Ross, M. H. & DOBLER, J. (1975). The Sertoli cell Anat. 167, 181-192. junctional specializations and their relationship to the SZOLLOSI, A. & MARCAILLOU, C. (1980). Gap junctions germinal epithelium as observed after efferent ductule between germ and somatic cells in the testis of the moth, ligation. Anal. Rec. 183, 267-292. Anagasta kiiehniella (Insecta: Lepidoptera). Cell Tiss. RUSSELL, L. (1977«). Desmosome-like junctions between Res. 213, 137-147. Sertoli and germ cells in the rat testis. Am. J. Anat. 148, TRES, L. L. & KIERSZENBAUM, A. L. (1983). Viability of 301-312. rat spermatogenic cells in vitro is facilitated by their RUSSELL, L. D. (19776). Observations on rat Sertoli coculture with Sertoli cells in serum-free - supplemented medium. Pivc. natn. Acad. Sci. U.S.A. 80, ectoplasmic ('Junctional') specializations in their 3377-3381. association with germ cells of the rat testis. Tissue & TUNG, P. S., LACROIX, M. & FRITZ, I. B. (1980). Effects Cell 9, 475-498. of cytosine arabinoside on properties of testicular RUSSELL, L. D. (1979). Further observations on preparations in culture. Biol. Reprod. 22, 1255-1261. tubulobulbar complexes formed by late spermatids and VAN DER DONK, J. A., RUITER-BOOTSMA, A. L., ULTEE- Sertoli cells in the rat testis. Anat. Rec. 194, 213-232. VAN GESSEL, A. M. & WAUBEN-PENRIS, P. J. (1986). RUSSELL, L. D. (1980). Sertoli-germ cell interactions: A Cell-cell interactions between rat Sertoli cells and mouse review. Res. 3, 179-202. germ cells in vitro. J. exp. Cell Res. 164, 191-198. RUSSELL, L. D. (1984). Spermiation-the release WING, T.-Y. & CHRISTENSEN, A. K. (1982). Morphometric process: Ultrastructural observations and unresolved studies on rat seminiferous tubules. Ant. J. Anat. 165, problems. In infrastructure of Reproduction (ed. J. van 13-25. ZIPARO, E., GEREMIA, R., RUSSO, M. A. & STEFANINI, M. Blerkom & P. M. Motta), Electron Microscopv in Biology (1980). Surface interactions in vitro between Sertoli cells and Medicine, p. 46. Boston: Martinus Nijhoff. and germ cells at different stages of spermatogenesis. RUSSELL, L. & CLERMONT, Y. (1976). Anchoring device Am.J. Anat. 159, 385-388. between Sertoli cells and late spermatids in rat seminiferous tubules. Anat. Rec. 185, 259-278. {Received 15 October 1987 - Accepted 11 January 1988)

114 G.C. Enders and C. F. Millette