fetuin: a serum component associated with rat sertoli and spermatogenic cells in coculture

JOURNAL OF CELLULAR PHYSIOLOGY 127:463-472 (1986)

Fetuin: A Serum Component Associated With Rat Sertoli and Spermatogenic Cells

in Coculture MUNIR ABDULLAH, JAMES A. CROWELL, LAURA L. TRES, AND

ABRAHAM L. KIERSZENBAUM' Departments of Anatomy and Pediatrics and The Laboratories for Reproductive Biology, School of Medicine, The University of North Carolina at Chapel Hill, Chapel Hill, North

Carolina 27514

Cocultures of rat Sertoli-spermatogenic cells plated in a culture medium supplemented with 10% fetal bovine serum for 6-12 h and then maintained in serum free, hormone/growth factor-supplemented medium accumulated an acidic glycoprotein of molecular weight of 68,000 dalton (68 kD) and isoelectric point range of about 4.2-3.5. Anion exchange chromatography has allowed the partial purification of this protein, which consists of a major protein band of 68 kD and two minor, low molecular weight components. A rabbit antiserum raised against the 68 kD component also crossreacts with the two low molecular weight components, thus suggesting that these two minor components are antigenically related to t h e 68 kD protein. The 68 kD protein has been identified as fetuin, the major component of fetal bovine serum, based on similar molecular weight, isoelectric point, immunoreactivity and trypsin inhibitory activity. Labeling experiments with [14C]amino acid mixture show that 68 kD protein i s not synthesized by cocultured rat Sertoli and spermatogenic cells. Immunocytochemistry and Western blot approaches carried out under various experimental conditions support the view that the fetuin-68 kD protein is taken u p from serum by both Sertoli cells and pachytene spermatocytes. Because fetuin 1) behaves as a carrier protein for growth factors, 2) has protease inhibitory activity, 3) is preferentially internalized by Sertoli cells and pachytene spermatocytes and 4) fetal bovine serum-supplemented medium impairs spermatogenic cell viability, there is a need to further define appropriate conditions for optimizing long-term viability and differentiation of spermatogenic cells in vitro.

It has been shown that primary cell cultures can be established in a medium in which the usual serum sup- plement is replaced with hormones, growth factors, nutrients, and attachment factors (Barnes and Sato, 1980). Most protocols use a cell plating step in serum-supplemented medium because serum facilitates proper attachment and spreading of cells to plastic or glass surfaces (Higuchi, 1973; Temin et al., 1972; Barnes and Sato, 1980).

We have reported that a discrete population of rat spermatogenic cells in culture are able to synthesize DNA and differentiate into meiosis I1 when they are cocultured with Sertoli cells in a serum-free medium supplemented with hormones and growth factors (Tres and Kierszenbaum, 1982). However, when spermatogenic cells are cocultured with Sertoli cells in fetal bovine serum-supplemented medium, the progression of early meiotic prophase cells (preleptotene) into more advanced stages of meiosis was impaired by rapid deple- tion of spermatogenic cells from the cocultures. Studies on the effect of serum on rat testicular organogenesis in vitro have shown that fetal bovine serum prevented cell aggregation and organization of seminiferous cords

0 1986 ALAN R. LISS, INC.

(Chartrain et al., 1984). The differentiation of seminiferous cords was unaffected when primordia of fetal rat testes were maintained in serum-free medium.

We have searched for newly-synthesized secretory proteins in rat Sertoli-spermatogenic cell cocultures that could provide insights into the culture conditions re- quired for long-term viability of spermatogenic cells in vitro. During these studies we detected an abundant acidic glycoprotein with molecular weight (Mr) 68,000 daltons (68 kD) that accumulated in serum-free culture medium and then disappeared gradually 5-6 days after plating. Protein purification, radiolabeling, and immu- nological methods have allowed us to conclude that the 68 kD protein is not synthesized by cultured Sertoli or spermatogenic cells. We have identified this protein as fetuin (a major component of fetal bovine serum) and established its trypsin inhibitory property. The presence of this 68 kD protein has been attributed to the initial

Received December 11, 1985; accepted January 27, 1986. *To whom reprint requestsicorrespondence should be addressed.

464 ABDULLAH, CROWELL, TRES, AND KIERSZENBAUM

use of fetal bovine serum-supplemented medium for achieving rapid and effective attachment of Sertoli and spermatogenic cells to inert substrates.

MATERIALS AND METHODS Rat Sertoli-spermatogenic cell cocultures

Primary cultures were prepared from 20-22-day-old rats (Charles River CD) as described (Kierszenbaum and Tres, 1981). Briefly, trypsidcollagenase-dissociated seminiferous epithelial cells were suspended in Eagle’s min- imum essential medium (EMEM) supplemented with 10% fetal bovine serum (Sterile Systems, Logan, Utah), nonessential amino acids (0.1 mM), glutamine (4 mM), sodium pyruvate (1 mM), penicillin (100 U/ml), and streptomycin (100 pgiml). Samples were plated in culture flasks (for collection of conditioned medium) and on glass coverslips (for immunocytochemistry) at very high cell density (approximately 3 x lo6 cellsiml). Serum- supplemented medium was changed as soon as cell aggregates attached to the substrate (within 6-12 h).

A serum-free, hormone/growth factor-supplemented medium consisting of EMEM supplemented as above but lacking serum was used with the following addi- tions: transferrin (5 pgiml) insulin (5 pgiml) and epider- mal growth factor (10 ng/ml) (all these components from Collaborative Research, Lexington, Mass), human growth hormone (133 pmirnl, AB Kabi, Stockholm, Swe- den), retinol(5 pM, Sigma) and 0.1 pM each of testoster- one and dihydrotestosterone (Calbiochem). After an adaptation period of 24 h, the cocultures (consisting of Sertoli cells, spermatogonia, and meiotic prophase spermatocytes) were maintained in serumfree medium as above with the addition of follicle-stimulating hormone (FSH, 5 pg/ml, NIH-oFSH-S16). Cells were cocultured in this medium for 6 h, the medium removed and replaced with serum-free medium as above except that FSH was omitted. This “minus-FSH” medium was replaced after 18 h with “plus-FSH” medium containing 5 pg/ml of FSH in addition to other hormones, growth factors and nutrients. Further changes of medium followed the alternating minus-FSWplus-FSH, 6/18 h schedule for up to 10 days. Serum-free, hormone/growth factor-supplemented medium designated TKM (Tres et al., 1986) was freshly prepared every 3 days. The first two samples of conditioned medium were discarded. Future medium samples were collected and pooled for protein purification. Phenylmethylsulfonyl fluoride was added to a final concentration of 1 mM.

In additional experiments, samples were plated and maintained in 1) serum-free EMEM (supplemented with glutamine, sodium pyruvate, nonessential amino acids and antibiotics) or 2) EMEM supplemented as above and containing 10% of either new born bovine, adult bovine, horse or chicken sera. Cocultures maintained in either serum-free or in any of the four experimental serum- supplemented media were placed in the same 6/18 h alternating medium change schedule and conditioned media collected during 5 consecutive days.

Protein purification procedure All steps were carried out at 4°C. Conditioned serum-

free medium (500 ml) from rat Sertoli-spermatogenic cell cocultures was concentrated to about 50 ml by ultrafil- tration using a PM-30 membrane (nominal Mr exclusion greater than 30 kD, Amicon, Danvers, Mass), The me-

dium was dialyzed overnight against 4 liters of 0.01 M sodium acetate buffer, pH 4.4, and then applied to a QAE-Sephadex column (Pharmacia, Piscataway, NJ? (5 x 22 cm), which had been previously equilibrated with the same buffer. After loading the sample, the column was washed with the sodium acetate buffer containing 50 mM NaCl until protein absorption (monitored at 280 nm) was less than 0.02 units. Proteins bound to the anion exchange column were eluted with a linear gradient of 50-500 mM NaCl in the same buffer (total volume: 400ml). Fractions comprising the peak were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) for the presence of a 68 kD protein (see below).

Production of rabbit antisera against a 68 kn protein Two female rabbits (2-3 kg) were immunized for the

production of antibodies. Protein samples obtained from the QAE-Sephadex column were resolved by one-dimensional SDS-PAGE, gels stained with Coomassie Blue, destained, and rinsed in deionized water. Portions of the gel displaying a major protein band of Mr 68 kD were cut with a razor blade and homogenized in 2 ml of phosphate-buffered saline (PBS) using a Dounce manual tissue grinder. For immunization, the protein-polyacrylamide mixture was emulsified with Freund’s complete adjuvant (Sigma). A total amount of 200 pg of protein was injected subcutaneously into each rabbit on either side of the neck. After 6 weeks, the rabbits were injected again with about 110 pg of antigen emulsified with Freund’s adjuvant for secondary immunization. Ani- mals were bled by central ear vein puncture 10-15 days following the booster. To increase antibody titer, the rabbits were further immunized twice using 110 pg of antigen per rabbit 10 and 15 weeks following the primary immunization. In both cases, animals were bled one week after each booster. Serum was separated and stored at -20°C for further characterization.

Gel electrophoresis Protein samples were resolved on 10% polyacrylamide

gels containing 0.1% SDS after reduction and denatura- tion of the sample. Both one-dimensional (Laemmli, 1970) and two-dimensional (2D) PAGE (O’Farrell, 1975) methods were used. For 2D-PAGE, lyophilized samples were resuspended in lysis buffer (O’Farrell, 1975), resolved in cylindrical isoelectric focusing gels for the first dimension and then separated in second dimension using a 10% polyacrylamide gel overlaid with a 4% stack- ing gel. Estimates of Mr and isoelectric points (PI) were obtained by running Mr/pI standards in the first andor second dimension. Gels were stained with Coomassie Blue or periodic acid Schiff (PAS) (Fairbanks et al., 1971) and dried according to standard procedure.

Western blotting Purified protein samples resolved by one- and 2D-

PAGE were electrophoretically transfered to nitrocellulose paper (Towbin et al., 1979). Paper blots were treated with 1% gelatin in Tris buffered-saline (TBS), pH 7.4, for 1 h at 40°C to block nonspecific reactive sites. Paper blots were incubated for 2 h at room temperature in a TBS solution of primary antibody generated against the Mr 68 kD protein being probed (working dilution 1:20) and containing 3% bovine serum albumin and 1% nor-

FETUIN IN SERTOLI-SPERMATOGENIC CELL MEDIUM 465

ma1 goat serum. After washing in TBS, blots were incubated 1 h in a 1:lOOO dilution of goat anti-rabbit IgG- horseradish peroxidase in TBS as the second antibody. Finally, immunoreacted protein blots were incubated in a solution containing substrate for horseradish peroxidase (4-chloronaphthol) and rinsed with distilled water to stop development of the reaction.

Trypsin activity assay Tryptic activity of the purified protein sample was

assayed using the chromogenic substrate N-benzoyl-DL- arginine p-nitro-anilide (BAPA) according to the method of Erlanger et al. (1961). Tryptic hydrolysis of the substrate produced p-nitroaniline and this yellow product was measured by recording the linear increase in ab- sorbance at 410 nm over time using a Beckman DB-GT spectrophotometer. To follow the inhibitory effect of the purified protein fraction on trypsin activity, the enzyme was first mixed with increasing concentration of the purified protein sample and then the mixture applied to the assay system (Galembeck and Cann, 1974).

[14C] amino acid mixture labeling of rat Sertoli- spermatogenic cell cocultures

Cells were labeled in Earle’s balanced salt solution supplemented with the normal concentration of amino acids absent from the [14C]amino acid mixture (New England Nuclear, Boston, Mass, Cat.# NEC-445E), [ 14C] amino acid mixture (62.5 pCi/ml), glutamine (4 mM), sodium pyruvate (1 mM), vitamins and antibiotics. Co- cultures were labeled for 24 h, the medium collected, microfuged and then lyophilized. Lysis buffer (1OOpl) was added to the lyophilized sample and 2D-PAGE carried out using a 250,000 cpm sample. The sample also contained about 20 pg of non-radioactive purified 68 kD protein. Gels were stained with Coomassie Blue to iden- tify Mr standards and the added 68 kD protein. Gels were soaked in Amplify (Amersham, Arlington Heights, Illinois), dried and exposed to X-ray film for autoradiography. In additional experiments, radioactively-labeled secretory proteins were electrotransferred to nitrocellulose paper, immunoreacted with anti-68 kD protein serum as described above, soaked in Amplify, dried and exposed to X-ray film for autoradiography (not shown).

Neuraminidase treatment Both purified 68 kD protein from conditioned TKM

and standard fetuin (Sigma, St. Louis, Mo, type lV, lyophilized from fetal bovine serum, Cat. # F 3004) were treated with neuraminidase (Lampreave and Pineiro, 1984). Reaction mixtures contained 400 pg of either protein in 400 pl 0.1 M sodium acetate buffer (pH 5.01 with or without 40 pg of neuraminidase (Sigma, type VIII from Clostridium perfringens, Cat. # N 5631) dissolved in 200 pl of distilled water. Reaction mixtures were incubated at 37°C for 2 h after which a 100 pl aliquot of each mixture was neutralized with 10 pl of 0.5 M Tris- HCl buffer (pH 8.0) and analyzed by SDS-PAGE.

Indirect immunofluorescence Rat Sertoli-spermatogenic cell cocultures were grown

on 18 x 18 mm glass coverslips in 60-mm plastic culture dishes. Coverslips were rinsed in PBS and fixed in 3.7% formaldehyde (with or without permeabilization with 0.2% Triton X-100 in PBS for 2 min). Cells were incu-

bated with the antisera generated in the rabbit against the purified 68 kD protein (working dilutions: 1:25-150) for 1 h in a moist chamber a t room temperature. To localize bound 68 kD protein antibody, fluorescein iso- thiocyanate-conjugated goat anti-rabbit IgG was applied to the cells for 30-60 min, rinsed in PBS and coverslips mounted with Elvanol (Rodriguez and Deinhardt, 1960). Preparations were observed with oil immersion phase- contrast/fluorescence objectives in a Leitz photomicro- scope equipped with epi-illumination for fluorescence microscopy and transmitted light and phase-contrast microscopy. Images were recorded on Kodak Tri-X film with typical exposures of 15 sec. Immunocytochemical controls included omission of the primary antiserum and absorption of antiserum with concentrated conditioned medium, purified 68 kD protein, and standard fetuin purified from fetal bovine serum (Sigma type IV) carried out as described (Kierszenbaum et al., 1981). Control experiments resulted in the lack of fluorescent staining patterns in Sertoli and spermatogenic cells (not shown). Antibody specificity was evaluated in Western blot experiments as described above.

RESULTS Anion exchange chromatography of rat

Sertoli-spermatogenic cell conditioned serum-free culture medium

Concentrated medium was fractionated on a QAE- Sephadex column using 0.01 M sodium acetate buffer, pH 4.4. A number of proteins present in the medium have a net positive charge at this pH and, therefore, do not bind to QAE-Sephadex. Unbound proteins were removed from the column by washing with 50 mM NaCl (Fig. 1,A). One-dimensional SDS-PAGE analysis of the unbound fraction showed that albumin, transferrin (both proteins identified in Western blots using rabbit antibodies (not shown)), and other minor unidentified proteins were present (Fig. 1,B-lane 2). The presence of albumin and transferrin was expected because albumin is a carrier present in the FSH preparation and transferrin, as well as FSH, are components of TKM used for coculturing Sertoli and spermatogenic cells. A major protein that remained bound to QAE-Sephadex under the conditions used was eluted as a single peak (Fig. 1,A). SDS-PAGE analysis of the fractions comprising the peak showed the presence of a major protein with apparent Mr of 68 kD and two slighty lower Mr proteins of about 62 and 56 kD (Fig. 1,B-lane 1). In a 2D-PAGE system, the major protein and the two minor components migrated towards the acidic end with a PI range of 4.2-3.5 (Fig. 4,B). PAS staining of electrophoresed proteins revealed that the 68 kD protein and the two minor low Mr components were PAS positive (not shown). Based on the binding properties of the 68 kD protein to QAE-Sephadex at pH 4.4, its PI range and PAS reactivity, this protein was identified as an acidic glycoprotein and the two low Mr components as either breakdown products or protein aggregates.

Identification of the 68 kD acidic glycoprotein Rat Sertoli-spermatogenic cell cocultures were care-

fully washed three times with Hanks Balanced Salt Solution before placing them in serum-free TKM. In addition, the first two conditioned TKM samples collected every 24 h were discarded. However, we consid-


FRACTION NUMBER A Fig. 1. A. Anion exchange chromatography of concentrated ( X 10) serum-free medium conditioned by cocultured rat Sertoli-spermatogenic cells. The arrow indicates the application of a linear gradient of NaCl (50-500 mM) in 0.01 M sodium acetate buffer (pH 4.4). Column size: 5 x 22 cm; flow rate: 40 m l h and fraction size: 5 ml. B. SDS PAGE (10% polyacrylamide) of the bound (lane 1) and unbound (lane 2) pro-

ered the possibility that residual components of fetal bovine serum supplemented-EMEM used during the initial cell attachment phase, could explain the abundance of the 68 kD glycoprotein accumulated in serum-free TKM.

To address this possibility, antibodies were raised in rabbits against the major 68 kD protein component ex- cised from preparative SDS-PAGE. Western blot experiments demonstrated that antiserum generated against this protein crossreacted with both the major 68 kD and the two low Mr proteins (Fig. 2,A). In addition, the antiserum crossreacted slightly with purified preparations of bovine a-fetoprotein (Calbiochem) and bovine serum albumin (Sigma) (data not shown) thus indicating crossreactivity with two albumin-like proteins of related Mr but less acidic PI (albumin: Mr 67.6 kD/ PI 4.7; a- fetoprotein: Mr 68 kD/ PI 4.8).

It was reported that cultured cells initially grown in the presence of fetal bovine serum-supplemented medium adsorb a protein of Mr 68 kD that was then released into serum-free medium (human embryonic lung cells, Rohrlich and Rifkin, 1981; and breast epithelial cells, Gendler and Tokes, 1984). This protein, identified as fetuin, consisted of a major protein band of 68 kD and two protein bands of lower Mr (apparent Mr 62 and 56 kD, Rohrlich and Rifkin, 1981).

Based on the electrophoretic similarities between fetuin and our 68 kD protein purified from conditioned serum-free TKM, studies were carried out to determine whether the 68 kD protein was indeed fetuin. When the 68 kD protein and a standard preparation of fetuin (Sigma) were run on SDS-PAGE, both preparations displayed a major 68 kD protein band and two minor protein bands of lower Mr (Fig. 2,B). In addition, antiserum generated against the 68 kD protein crossreacted in Western blot experiments with both the major and two minor components of fetuin and the protein purified from conditioned TKM (Fig. 2,A). Furthermore, rabbit

1 2 Mr.kD

94 - 68 -

43-

30 -

20 -

tein fractions. Lane 1 shows a major protein band Mr 68 kD and two minor, low Mr components (arrowheads). Lane 2 shows two major protein bands identified as transferrin (76 kD) and albumin (68 kD) found in the major peak preceding application of salt gradient. Coom- assie Blue-stained gels.

antiserum raised against fetuin purified from serum- free medium conditioned by human embryonic lung cells (a generous gift from Drs. S.T. Rohrlich and D.B. Rifiin) also recognized the 68 kD protein and the two minor, low Mr components of standard fetuin and of our purified 68 kD acidic glycoprotein (data not shown).

Because fetuin is known to possess terminal sialic acid residues that account for its Mr heterogeneity (Graham, 1972), we decided to determine the effects of neuraminidase on both fetuin and the 68 kD protein. Figure 2B shows that neuraminidase treatment: 1) reduces the density of the major 68 kD protein band, 2) increases the density of the two low Mr protein bands, and 3) generates additional low Mr protein bands. This finding demonstrated that the removal of sialic acid residues from fetuin and the 68 kD protein resulted in the breakdown of the major 68 kD protein band into low Mr components. This conclusion was also supported by the antigenic relationship between the 68 kD protein and the two low Mr fragments detected in control samples (Fig. 2,A).

Trypsin inhibitory activity of the 68 kn acidic glycoprotein

It was reported that fetuin has serine protease inhibitory activity (Galembeck and Cann, 1974; Rohrlich and Rifkin, 1981). Therefore, it was important to investigate whether the purified 68 kD acidic glycoprotein shared this additional property with fetuin. A trypsin inhibitory assay was carried out and trypsin activity monitored by spectrophotometry. Figure 3 shows that the addition of increasing amounts of the 68 kD protein resulted in a progressive loss of trypsin activity leading to the complete inhibition of the enzyme.

Because of the similarities in Mr, PI and protease inhibitory activity of the 68 kD protein and al-antichymotrypsin (Travis et al., 19781, we wanted to eliminate the possibility of the two being the same protein. Based

FETUIN IN SERTOLI-SPERMATOGENIC CELL MEDIUM

94 - 68 -

I

Mr (kD) Mr (kB)

- 94 - 68

43 - -43

30 - - 30

20 - 14 -

A B - 20

Fig. 2. A. Western blots of standard fetuin and purified 68 kD protein from conditioned serum-free TKM. Arrowheads indicate two low Mr protein bands below the major band component of fetuin that cross- react with antibody raised against 68 kD purified protein. Similar protein migration pattern and immunoreactivity are observed in the 68 kD sample. B. Effect of neuraminidase on standard fetuin and 68 kD protein. Coomassie Blue-stained gel. Note that the major 68 kD component of both fetuin and 68 kD protein are reduced in density after neuraminidase treatment and that low Mr components increase in density and number.

on the property of q-antichymotrypsin to bind to DNA- cellulose (Siddiqui et al., 1980), we have tested the be- havior of the 68 kD protein and found that it does not bind to DNA-cellulose (data not shown).

Lack of [14C]amino acid mixture labeling of the 68 kD acidic glycoprotein

Labeling of rat Sertoli-spermatogenic cell cocultures with [14C]amino acid mixture was carried out to determine whether the 68 kD acidic glycoprotein was a newly- synthesized product of the cultured cells. We decided to use a mixture of radioactively-labeled amino acids in- stead of [35S]methionine since fetuin lacks methionine residues (Graham, 1972; Salomon et al., 1982). A purified 68 kD protein preparation was added as a tracer to [14C]amino acid-labeled secretory proteins accumulated in the medium of cocultured Sertoli and spermatogenic cells and the samples were fractionated by 2D-PAGE and resolved by autoradiography (Fig. 4,A). Figure 4,B illustrates the Coomassie Blue staining of the 68 kD acidic glycoprotein and its expected position in the au-

0.12 A

E 7 - ," 0.10

E 0.08 3 a >-̂ 0.06 k 1 & 0.04 a z 3 0.02 a >- E

t .-

0 20 40 60 8 0 7

467

I

Fig. 3. Inhibition of trypsin activity by 68 kD protein purified from Sertoli-spermatogenic cell coculture medium. AU/min indicates ab- sorbance unitsimin at 410 nm.

toradiogram (Fig. 4,A). As shown, the 68 kD protein tracer does not correlate with any of the [14C]amino acid mixture-labeled secretory proteins. Some of the adjacent proteins are Sertoli cell-specific secretory proteins (Kier- szenbaum et al., 1986). In additional experiments, radioactively-labeled secretory protein samples also containing non-radioactive 68 kD tracer were resolved by 2D-PAGE, transferred to nitrocellulose paper, immunoreacted with anti-68 kD serum and then exposed to X- ray film. The resulting 68 kD immunoreactive product did not match with any radioactively-labeled protein spot in the autoradiogram (not shown).

Localization of the 68 kD acidic glycoprotein in Sertoli-spermatogenic cell cocultures by

indirect immunofluorescence We decided to use immunocytochemistry to determine

the possible location of the 68 kD protein in cultured Sertoli or spermatogenic cells. Rat Sertoli-spermatogenic cell cocultures plated on glass coverslips were maintained under identical conditions used for collect- ing TKM conditioned medium for protein purification. As control, cells were directly plated in TKM, thus omit- ting the use of fetal bovine serum-supplemented medium. Cells were processed for immunofluorescence using as primary antibody the antiserum raised against the 68 kD acidic glycoprotein purified from conditioned TKM. Figure 5,A shows a large number of spermatogenic cells cultured on top of subjacent Sertoli cells. Each spermatogenic cell depicts a spot-like, excentric immunoreactive product. Spermatogenic cells were identified by phase-contrast microscopy as pachytene spermatocytes by the characteristic display of autosomal and XY bivalents (Fig. 5,B). A very different immunoreactive pattern was detected in other spermatogenic cells (Fig. 5,C): fluorescent coarse granules were re- stricted to the cell periphery, possibly representing cell

468

94 - 68 -

43-

30-

20 -

ABDULLAH, CROWELL, TRES, AND KIERSZENBAUM

Fig. 4. A. Autoradiogram of [14C]amino acid mixture-labeled secretory proteins accumulated in the medium of rat Sertoli-spermatogenic cell cocultures. The expected position of 68 kD protein is indicated. The position of transferrin (W, and three major Sertoli cell-specific acidic secretory proteins 670: Mr 72-70 kD; S45: Mr 45 kD; S35: Mr

surface sites. While the circumferential immunoreactive pattern was observed in most meiotic prophase spermatocytes, and less frequently in spermatogonia, the spot- like 68 kD immunoreactive product was typically associated with pachytene spermatocytes. Cocultured Sertoli cells showed a punctuate immunoreactive pattern that appeared scattered throughout the cytoplasm (Fig. 5,D). We have presumed that the punctuate immunofluorescent product corresponds to the internalized 68 kD protein since it decreased gradually with time in cells maintained in serum-free TKM and it was not observed in cells plated and maintained in either serum-free EMEM or TKM (not shown).

To determine whether a relationship existed between the decline of cell immunoreactivity and the presenct the 68 kD protein in the culture medium, a time-cour, study was carried out using a Western blot approach. Figure 6 demonstrates that the relative amount of the 68 kD protein decreases in the culture medium within a 96-h time period. In addition, media from cocultures plated and maintained in serum-free TKM did not show the immunoreactive 68 kD protein (Figure 6).

We then wanted to determine whether other sera were responsible for the accumulation of the 68 kD acidic glycoprotein. Rat Sertoli-spermatogenic cell cocultures were plated in EMEM supplemented with fetal bovine, new born bovine, adult bovine, horse, or chicken sera. After a 6-12 h cell attachment period, cells were placed in serum-free TKM and medium collected during a 6- day period. Pooled conditioned media were analyzed by Western blotting. Figure 7 demonstrates that the immunoreactive protein was detected only in samples from cocultures initially plated in fetal bovine and new born bovine serum but was absent in cultures plated in adult bovine, chicken, or horse sera.

35 kD) are also indicated. Labeling time: 24 h. Sample: 250,000 cpm. Exposure time of Amplify-enhanced gel: 1 week. B. Corresponding Coomassie Blue-stained gel showing the location of non-radioactive 68 kD protein used as a tracer.

DISCUSSION

Results presented in this paper show that, when seminiferous epithelial samples are plated for a brief period of time (6-12 h) in fetal bovine serum-supplemented medium, a major 68 kD acidic glycoprotein of the serum, fetuin, is taken up by Sertoli and spermatogenic cells. This protein is then concentrated and gradually released into a serum-free medium until its intracellular storage is depleted. While most cultured spermatogenic cells show immunofluorescent evidence of fetuin adsorption to cell surfaces, only Sertoli cells and pachytene spermatocytes displayed patterns compatible with a mechanism of endocytosis.

In vivo, meiotic and postmeiotic spermatogenic cells are excluded from direct interaction with blood plasma and lymphatic components by a physiological barrier (Setchell and Waites, 1975; Sharpe, 1984; Waites and Glaswell, 1982) whose major structural components are tight junctions linking adjacent Sertoli cells (Dym and Fawcett, 1970). Therefore, a serum-supplemented me- c’iurn is likely to provide spermatogenic cells in vitro u-”h a variety of components that have been previously screened but not admitted by the physiological barrier. Using a serum-free, hormoneigrowth factor-supplemented medium, we have shown that the the highly differentiated spermatogenic cells cocultured with Ser- toli cells have an extended life-span (Tres and Kierszen- baum, 1983; Tres et al., 1986). In addition to providing a combination of hormones and growth factors to cocultured Sertoli and spermatogenic cells, the serum-free medium was replaced twice a day following a precise alternating daily schedule as described in Materials and Methods. It is likely that the 6/18 h medium replace- ment schedule has a favorable effect on spermatogenic


Fig. 5. Indirect immunofluorescent localization of 68 kD protein using rabbit antibody. A. Pachytene spermatocytes showing a spot-like, immunoreactive excentric site. White lines indicate immunoreactivity in a few cells. x 1,760. B. Phase-contrast microscopy corresponding to Fig. 5,A. The location of the XY bivalent is indicated in the same cells labeled in Fig. 5,A. Criteria for identification of spermatogenic cell types in cocultures were described (Tres and Kierszenbaum, 1983).

x 1,760. C. High magnification photograph illustrating the peripheral, cell surface location of 68 kD immunoreactivity in spermatocytes. ~5,550. D. Cytoplasmic distribution of 68 kD-immunoreactive punctuate granules in Sertoli cells. The arrow indicates a characteristic juxtanuclear (N) lipid droplet found in most cultured Sertoli cells. x 1,050.


z w 2 w + 2

Y I- m n n_

94 - 68 -

43 - 30 - 20 -

Fig. 6. Western blot of media collected from rat Sertoli-spermatogenic cell cocultures plated in serum-free TKM or serum-supplemented EMEM for 12 h and then maintained in serum-free TKM for 4 days. Immunoreactivity for 68 kD is only observed in the medium samples plated in fetal bovine serum-supplemented medium and decreases with time.

cell long-term viability by consistently removing fetuin and other unwanted cell products from the immediate cell environment in the cocultures. In vivo, metabolic cell products find their way into seminiferous tubular luminal and intertubular space compartments (Waites and Glaswell, 1982).

Findings described in the paper present two interest- ing aspects for consideration: 1) the possible effects of fetuin on Sertoli and spermatogenic cell function and 2) the localization of fetuin in pachytene spermatocytes.

Two properties attributed to fetuin are relevant to the first aspect: 1) fetuin is a carrier protein for growth factors (Fisher et al., 1958; Graham, 1972; Temin et al., 1972; Lai et al., 1981; Lieberman and Ove, 1981). and 2) fetuin has a protease inhibitory activity (Galembeck and Cann, 1974; Rohrlich and Rifkin, 1981). It is widely accepted that, when serum is supplemented to various culture media, it provides low- and high-Mr components for cell growth and attachment. Mitogenic activity present in serum is ascribed to various hormones and growth factors that occur in either a free form or bound to carrier proteins (Hadden and Rout , 1964; Higuchi, 1973; Hoffman et al., 1973; Nevo and Laron, 1979; Salomon, 1980). An appropriate example is somatomedin-C, an

94 - 68 -

43 -

30

20 - 14 -

w

> z w o w Z m Z

Fig. 7. Western blot illustrating the effect of five different serum- supplemented media on the accumulation of 68 kD protein. Only cocultures plated in EMEM supplemented with fetal bovine or, to a lesser extent, newborn bovine sera exhibit immunoreactivity for the 68 kD protein in pooled medium collected for 5 consecutive days.

insulin-like growth factor present in blood plasma that is also produced by cultured rat Sertoli and peritubular cells and has binding affinity to Sertoli cells and, inter- estingly enough, to pachytene spermatocytes (Tres et al., 1986).

Fetuin, present in fetal bovine serum, contains growth- promoting activity (Fisher et al., 1958; Graham, 1972; Temin et al., 1972; Lieberman and Ove, 1981). Growth promoting activity of crude fetuin preparations is associated with platelet-derived growth factor (Libby et al., 1985) and a high Mr protein (270 kD) designated embryonin that resembles a fragment of q-macroglobulin protomers (Salomon et al., 1982; Feldman et al., 1984). Embryonin differs from fetuin in Mr, PI, amino acid composition and immunoreactivity (Salomon et al., 1982). An accepted view is that the biological effects of fetuin on cultured cells are related to its carrier function for molecules with growth-promoting properties. We have not carried out studies with crude fetuin-supplemented medium to determine possible growth-promoting properties of the 68 kD acidic glycoprotein on cocultured Sertoli and spermatogenic cells. However, we have reported that the viability of spermatogenic cells cocultured with Sertoli cells in fetal bovine serum-supplemented medium is considerably reduced in spite of an alternating 6/18 h medium change schedule (Tres and Kierszenbaum, 1983). Therefore, it is possible that other properties independent of the growth-promoting


effect on fetuin-containing serum may contribute to spermatogenic cell limited viability in culture.

The second aspect relates with the intracellular localization of the 68 kD protein in pachytene spermatocytes. It has been reported that rat pachytene spermatocytes have binding affinity for transferrin (Holmes et al., 1983) and somatomedin-C (Tres et al., 1986). Immunofluores- cent patterns of somatomedin-C binding and internalization in pachytene spermatocytes (Tres et al., 1986) correlate with those described for the 68 kD protein. In addition, electron micrographs of pachytene spermatocytes cocultured with Sertoli cells display abundant coated pits (Tres et al., 1986), thus suggesting that an operational mechanism exists in these meiotic prophase cells for internalization of ligands. However, an intrigu- ing question is the possible effect of the 68 kD, serum- derived protein on pachytene spermatocytes.

In this context, the trypsin inhibitory activity associated with fetuin (Galembeck and Cann, 1974; Rohrlich and Rifkin, 1981) is of particular significance. We have shown in this paper that the fetuin-like 68 kD acidic glycoprotein, detected in both Sertoli and spermatogenic cells and then released into serum-free TKM, retains its known trypsin inhibitory activity. We do not know whether fetuin’s internalization by Sertoli or spermatogenic cells is preceded by cell endocytosis or receptor binding by a fetuin-transported growth factor. Another possibility is that, because of its protease inhibitory activity, fetuin could bind to cell membrane associated proteases (Nakamura et al., 1984). In fact, it has been reported that various inhibitors of trypsin-like proteins have binding affinity for purified membrane preparations of rat hepatocytes (Nakamura et al., 1984). Inacti- vation of poteases occurs by forming complexes with protease inhibitors thus protecting cells from proteolysis (Neurath, 1984). At present, little is known about the presence of proteases on the surfaces of testicular cells.

In addition to fetuin, fetal bovine serum and other sera contain a number of protease inhibitors that in- clude q-antitrypsin, c-u-2-macroglobulin and cq-antichymotrypsin (Travis and Salvesen, 1983). Furthermore, it has been reported that human breast epithelial cells can actively synthesize and secrete q-antichymotrypsin into serum-free medium (Gendler and Tokes, 1984). Further studies should determine the effect of various protease inhibitors on the function of both Sertoli and spermatogenic cells. From the foregoing account, it can be con- cluded that a number of cell culture variables still need to be considered in our efforts for achieving long-term viability and differentiation of spermatogenic cells in vitro.

ACKNOWLEDGMENTS We are grateful to Dr. Lubna Abdullah for her help in

raising antibodies and to Dr. Eric P. Smith for part of the imnunocytochemical studies.

This research was supported by USPHS grants HD18315 (to L.L.T.) and HD11884 (to A.L.K.).

LITERATURE CITED Barnes, D. and Sato, G. (1980) Methods for growth of cultured cells in

serum-free medium. Anal. Biochem., 102255-270. Chartrain, I., Magre, S., Maingourd, M. and Jost, A. (1984) Effect of

serum on organogenesis of the rat testis in vitro. In Vitro, 20t912- 922.

and the physiological compartmentalization of the seminiferous epi- thelium. Biol. Reprod., 3,308-326.

Erlanger, B.F., Kokowsky, N. and Cohen, W. (1961) The preparation and properties of two new chromogenic substrates of trypsin. Arch. Biochem. Biophys., 95t271-278.

Fairbanks, G., Steck, T.L., and Wallach, D.F.H. (1971) Electrophoretic analysis’of the major polypeptides of the human erythrocyte membrane. Biochemistry, ZUt2606-2617.

Feldman, S.R., Gonias, S.L., Ney, K.A., Pratt, C.W. and Pizzo, S.V. (1984) Identification of “embryonin” as bovine w2-macroglobulin. J. Biol. Chem., 259.4458-4462,

Fisher, H.W., Puck, T.T. and Sato, G. (1958) Molecular growth requlre- ments of single mammalian cells: the action of fetuin in promoting cell attachment to glass. Proc. Natl. Acad. Sci. USA 44.4-10.

Galembeck, F. and Cann, J.R. (1974) Fetuin as a trypsin inhibitor. Arch. Biochem. Biophys., 264t326-331.

Gendler, S.J. and Tokes, Z. A. (1984) Active proteinase inhibitors associated with human breast epithelial cells. J. Cell. Biochem., 26157- 167.

Graham, E.R.B. (1972) Fetuin In: Glycoproteins: Their Composition, Structure and Function. A. Gottschalk. ed. Elsevier. Amsterdam, Part A. Vol. 5, pp. 717-731.

Drotein in normal human serum. Nature 2U2t1342-1343. Hadden, D. R. and Prout, T. E. (1964) A growth hormone binding

Higuchi, K. (1973) Cultivation of animal cells in chemically defined media, a review. Adv. Appl. Microbiol., 16.111-136.

Hoffman, R., Ristow, H-J., Veser, J . and Frank, W. (1973) Properties of two growth-stimulating proteins isolated from fetal calf serum. Exp. Cell Res., 85275-280.

Holmes, S. D., Bucci, L.R., Lipshultz, L. I. and Smith, R. G. (1983) Transferrin binds specifically to pachytene spermatocytes. Endocri- nology, 213t1916-1918.

Kierszenbaum, A.L., Crowell, J.A., Shabanowitz, R.B., DePhilip, R.M. and Tres, L.L. (1986) Protein secretory patterns of rat Sertoli and peritubular cells are influenced by culture conditions. Bioi. Reprod. (in press).

Kierszenbaum, A. L., Feldman, M., Lea, O., Spruill, W.A., Tres, L. L., Petrusz, P. and French, F. S. (1981) Localization of androgen-binding protein in proliferating Sertoli cells in culture. Proc. Natl. Acad. Sci. USA, 77.5322-5326.

Kierszenbaum, A. L. and Tres, L. L. (1981) The structural and func- tional cycle of Sertoli cells in culture. In: Bioregulators of Reproduc- tion. (J.G. Jagiello and H.J. Vogel eds). Academic Press, New York, pp. 207-228.

Laemmli, U. K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227t680-685.

Lai, P.C.W., Huang, L.L., Panrucker, D.E., Church, R.B. and Lorschei- der, F.L. (1981) Distribution of bovine fetuin and albumin in plasma, allantoic and amniotic fluids during development. J. Reprod. Fert., 63t53-60.

Lampreave, F. and Pineiro, A. (1984) The major serum protein of fetal and newborn pigs: Biochemical properties and identification as a fetal form of wl-acid glycoprotein. Int. J. Biochem., 16t47-53.

Libby, P., Raines, E.W., Cullinane, P.M. and Ross, R. (1985) Analysis of the mitogenic effect of fetuin preparations on arterial smooth muscle cells: The role of contaminant platelet-derived growth factor. J . Cell. Physiol., 225t357-366.

Lieberman, I. and Ove, P. (1958) A protein growth factor for mammalian cells in culture. J. Biol. Chem., 233t637-642.

Nakamura, T., Asami, O., Tanaka, K. and Ichihara, A. (1984) Increased survival of rat hepatocytes in serum-free medium by inhibition of a trypsin-like protease associated with their plasma membranes. Exp. Cell Res., 15423-91.

Neurath, H. (1984) Evolution of proteolytic enzymes. Science, 224:350- 357.

Nevo, 2. and Laron, 2. (1979) Growth factors, Am. J. Dis. Child., 133.419-428.

O’Farrell, P. H. (1975) High resolution two-dimensional gel electrophoresis of proteins. J. Biol. Chem., 250.4007-4021.

Rodriguez, J. and Deinhardt, F. (1960) Preparation of a semipermanent mounting medium for fluorescent antibody studies. Virology, 12t316- 317.

Rohrlich, S.T. and Rifkin, D.B. (1981) Isolation of the major serine protease inhibitor from 5-day serum-free conditioned medium of human embryonic lung cells and demonstration that it is fetuin. J. Cell Physiol., 209tl-15.

Ronne, H., Anundi, H., Rask, L. and Peterson, P. A. (1979) Nerve growth factor binds to serum alpha-2-macroglobulin. Biochem. Bio- phys. Res. Commun., 87.330-336.

Salomon, D. S. (1980) Growth of mouse embryonal carcinoma cells in Dym, M. and Fawcett, D.W. (1970) The blood-testis barrier in the rat serum-free, hormone-supplemented medium correlated with recep-


tors for growth factors. Exp. Cell Res., 128t311-327. Salomon, D. S., Bano, M., Smith, K.B. and Kidwell, W. R. (1982)

Isolation and characterization of a growth factor (embryonin) from bovine fetuin which resembles az-macroglobulin. J. Biol. Chem., 257t14093-14101.

Setchell, B.P. and Waites, G. M. H. (1975) The blood testis barrier. In: Handbook of Physiology. R.O. Greep and R.B. Astwood, eds. Ameri- can Physiological Society, Washington, D.C. Section 7, Vol. 5, pp. 143-172.

Sharpe, R. M. (1984) Intratesticular factors controlling testicular function. Biol. Reprod., 30129-49.

Siddiqui, A. A., Hughes, A. E., Davies, R. J. H. and Hill, J. A. (1980) The isolation and identification of a1 -antichymotrypsin as a DNA- binding protein from human serum.-Biochem. Biophys. Res. Com- mun., 95:1737-1742.

Temin, H. M., Pierson, R. W. and Dulak, N. C. (1972) The role of sei-um in the control of multiplication of avian and mammalian cells in culture. In: Growth, Nutrition and Metabolism of Cells in Culture. G.H. Rothblat and V. J. Cristotolo, eds. Academic Press, New York, Vol. 1, pp. 50-81.

Towbin, H., Staehelin, T. and Gordon, J. (1979) Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci. USA, 76t4350- 4354.

Travis, J., Garner, D. and Bowen, J. (1978) Human al-antichymotrypsin: Purification and properties. Biochemistry, 175647-5651.

Travis, J., and Salvesen, G. S. (1983) Human plasma proteinase inhibitors. Ann. Rev. Biochem., 52t655-709.

Tres, L. L. and Kierszenbaum, A. L. (1983) Viability of rat spermatogenic cells in vitro is facilitated by their coculture with Sertoli cells in serum-free hormone supplemented medium. Roc. Natl. Acad. Sci. USA, 8Ot3377-3381.

Tres, L. L., Smith, E. P., Van Wyk, J. J. and Kierszenbaum, A. L. (1986) Immunoreactive sites and accumulation of somatomedin-C in rat Sertoli-spermatogenic cell co-cultures. Exp. Cell Res., 162:33-50.

Waites, G. M. H. and Glaswell, R. T. (1982) Physiological significance of fluid secretion in the testis and blood-testis barrier. Physiol. Rev., 62624-671.

fetuin: a serum component associated with rat sertoli and spermatogenic cells in coculture

Documents