Partial characterization of a unique growth factor secreted by human Sertoli cells*†

FERTILITY AND STERILITY Copyright~ 1988 The American Fertility Society

Vol. 49, No.4, April1988 Printed in U.S.A.

Partial characterization of a unique growth factor secreted by human Sertoli cells*t

Jeffrey P. Buch, M.D.:j: Dolores J. Lamb, Ph.D.§ Larry I. Lipshultz, M.D. Roy G. Smith, Ph.D. Scott Department of Urology, Baylor College of Medicine, Houston, Texas

Human Sertoli cells were grown in a serum-free environment, and the Sertoli cell conditioned medium (hSCCM) was tested for mitogenic activity. The presence of a potent growth factor(s), termed Sertoli cell secreted growth factor (SCSGF), in hSCCM was confirmed and supports previous observations based on experiments using rat SCCM. Mitogenicity of hSCSGF was demonstrated in cell proliferation assays with the A431 (human epidermoid carcinoma) cell line and in [methyl- 3 H]-thymidine incorporation (DNA synthesis) assays with the Swiss 3T3 (mouse embryo fibroblast) cell line. In a dose-dependent manner, hSCSGF stimulated A431 cell growth up to 4-fold over control values (P < 0.0001) and stimulated thymidine incorporation up to 4.5-fold over control values (P < 0.0002). Importantly, SCSGF stimulated A431 proliferation 2-fold over control values (P < 0.0002) in the presence of 5% serum. With the exception of rat SCSGF, human SCSGF is the only growth factor known to stimulate A431 cells. SCSGF also demonstrated epidermal growth factor (EGF)-like activity based upon displacement of EGF from its receptor in a radioreceptor assay. However, SCSGF is not EGF since it is a potent stimulator of A431 cells, whereas EGF is inhibitory. The growth factor was stable to heat, freeze-thaw, acid (pH 3), and trypsin treatment. Furthermore, it did not bind heparin agarose and is thus distinct from the endothelial cell growth factor family. Highpressure liquid chromatography on size exclusion (TSK G2000 SW) columns revealed an approximate size of 8000 daltons. Human SCSGF is a unique growth factor and may play a key role in the regulation of normal spermatogenesis. Fertil Steril 49:658, 1988

Received June 1, 1987; revised and accepted December 30, 1987. • Supported in part by the National Institutes of Health grants DA-03431 and CA-36264. t Presented at the Forty-Third Annual Meeting of The American Fertility Society, September 28 to 30, 1987, Reno, Nevada. Award-winning paper in Urology Research. t Recipient of the American Urological Association Research Scholarship Award. § Reprint requests: Dolores J. Lamb, Ph.D., Department of Urology, Baylor College of Medicine, One Baylor Plaza, Room 440E, Houston, Texas 77030.

barrier/· 2 phagocytosis of degenerating germ cells and spermatozoan cytoplasmic debris, 2 •3 and the secretion of several proteins in response to folliclestimulating hormone (FSH). 4 - 7 The Sertoli cell possesses several unique physical characteristics which relate to its function as a "nursing cell" for developing germ cells. Numerous processes extend from the Sertoli cell's surface to envelop the maturing germ cells. 2 Adjacent Sertoli cells thus provide an intertwining support structure for the germ cells with extensive opportunities for direct intercellular communication. Furthermore, adjacent Sertoli cells form junctional complexes at a level separating the spermatogonia and pre-leptotene spermatocytes from the more mature spermatocytes, thus establishing a functional blood-testis

Buch et al. Human Sertoli cell secreted growth factor

Fertility and Sterility

Evidence suggests an important role for the Sertoli cell in spermatogenesis. Among the functions of the Sertoli cell which appear important to spermatogenesis are establishment of the blood-testis

658

barrier .1·2 Early studies of the properties of the blood-testis barrier by Waites and Setchell8 revealed a restrictive permeability function based, in part, upon molecular size. Larger molecules found in testicular lymph were noted to be absent in rete testis fluid. This blood-testis barrier, therefore, results in the separation of the seminiferous epithelium into a "basal" and an "adluminal" compartment2 with differing extracellular environments. Although the precise function and significance of the blood-testis barrier is still undefined, its role in the compartmentalization of maturing germ cells appears critical for normal spermatogenesis. The Sertoli cell secretes several proteins in response to FSH. Included among these are androgen-binding protein, 4 inhibin, 5 gamma-glutamyl transpeptidase, 6 and transferrin. 7 Androgen-binding protein secretion, as regulated by FSH, can control local levels of active (unbound) androgen. Inhibin secretion results in a negative feedback upon the production of FSH by the pituitary. Sertoli cell-derived transferrin may play a key role in the iron metabolism of pachytene spermatocytes. 7 Appropriate stimulation of the Sertoli cell, therefore, provides a favorable physiologic environment for normal spermatogenesis. Recently, we reported that rat Sertoli cells secrete a growth factor(s) that blocks the binding of epidermal growth factor (EGF) to its receptor. 9 However, this factor is not EGF since it stimulates proliferation of A431 cells (derived from human epidermoid carcinoma) in culture, whereas EGF has been shown to be inhibitory to A431 growth. 9 This growth factor(s) has been termed Sertoli cell secreted growth factor (SCSGF). Although other growth factors have been reported as products of the seminiferous epithelium,9- 12 SCSGF has unique characteristics which differentiate it from the other reported factors. The seminiferous growth factor (SGF), partially purified by Feig and co-workers, 11 was obtained from seminiferous epithelium cytosols of several mammalian species, has a molecular weight of 15,700 daltons, and is heat-sensitive. Brown et al. 10 have detected a growth factor in the rete testis fluid of rams with a molecular weight of 45,000 daltons. Furthermore, a somatomedin-like peptide with a molecular weight of 25,000 daltons has been partially characterized by Smith et al. 12 from the spent medium of cultured rat Sertoli cells. None of these factors is known to displace EGF from binding to its receptor. Furthermore, SCSGF has a molecular weight of 8000 and is heat-stable. 9 In our previous report, possible roles for SCSGF Vol. 49, No.4, April1988

as an autocrine regulator of prepubertal Sertoli cell proliferation and as a paracrine factor for other testicular cells (germ cells and Leydig cells) were postulated. Since a role for EGF in male reproductive function has been suggested recently/ 3 the EGF-like activity (EGF-LA) noted in SCSGF has gained further significance. Although growth factors have been derived from the seminiferous epithelium of several mammalian species, data regarding human seminiferous epithelium is lacking. Our laboratory has previously reported an excellent technique for the culture of human Sertoli cells in vitro, as well as documenting secretory activity in response to FSH. 6•7 This culture technique was modified slightly to obtain human Sertoli cell cultures, which were tested for the presence of growth factor activity. In the present study, we report on the identification and partial characterization of SCSGF from human Sertoli cells in culture. MATERIALS AND METHODS Tissue

Testes were obtained from patients undergoing male-to-female sexual reassignment surgery (SRS, n = 3), patients undergoing therapeutic orchiectomy for metastatic carcinoma of the prostate (CAP, n = 5), or postpubertal cryptorchid patients undergoing therapeutic orchiectomy (CRP, n = 2). The SRS patients had been treated with 0.5 to 1.0 mg/day estinyl estradiol for a minimum of 1 year preceding orchiectomy, and, as a result of estrogen therapy, had significant enrichment of the Sertoli cell population. 6 Isolation and Culture of Sertoli Cells

The isolation and culture methods of Lipshultz et al. 6 were employed with the following modifications. Enzymatic disruption of the minced seminiferous tubules was accomplished with a solution containing 0.1% collagenase (class IV, Worthington Biochemical Corporation, Freehold, NJ) and 0.1% dispase (Boehringer Mannheim, Indianapolis, IN). Cells were plated onto 100-mm plastic culture dishes (Corning Corporation, Corning, NY) containing Dulbecco's Modified Eagle's medium (DME), 5% fetal bovine serum (FBS), and 0.1% penicillin-streptomycin (Grand Island Biological Company, Grand Island, NY) at 34 °C in an atmosphere of 5% C02 and 95% air. After 48 hours, the Buch et al. Human Sertoli cell secreted growth factor

659

medium was aspirated and replaced with DME containing 0.1% penicillin-streptomycin, 5 J.Lg/ml insulin (Sigma, St. Louis, MO), and 200 ng/ml FSH (ovine-FSH-16; a gift from the NIH Hormone Distribution Program, NIADDK). Final serum-free conditions for collection of Sertoli cell-conditioned medium (SCCM) were provided with repeated washing and medium replacement 24 hours later. Thereafter, SCCM was collected under serum-free conditions at 3-day intervals, and cultures were terminated after media collection on day 9. Quality of the Sertoli cell cultures was verified by phase-contrast microscopic morphology6 and by the production of transferrin, as measured in a previously reported radioimmunoassay (RIA). 7 Cultures were obtained from three SRS patients (humans 1, 2, and 8), five CAP patients (humans 4 to 7, and 10), and two CRP patients (humans 3 and 9). Histologic examination and transferrin assays of cultures from CAP patients 4 and 5 (both of whom had estrogen therapy and pelvic irradiation prior to orchiectomy) revealed only peritubular fibroblasts in the cultures, and the results were excluded from the study. SRS patient 8 did not successfully culture secondary to an unusual degree of peritubular fibrosis, presumably associated with his estrogen therapy. Our experience confirms the previously noted difficulty in obtaining human Sertoli cell cultures of high purity. 7 Postpubertal CRP patient 3 was cultured at a lower plating density and was therefore excluded from the study. Rat Sertoli cells were cultured for the collection of conditioned medium, as previously described. 9 Epidermal Growth Factor Binding Assays

The binding displacement of 1251-EGF (New England Nuclear, Boston, MA) from its receptor was evaluated with a radioreceptor assay (RRA) employing subconfluent monolayers of formalin fixed A431 cells (human epidermoid carcinoma line) as modified by Holmes et al. 9 The standard curve generated a range of sensitivity from 0.2 to 12.8 ng, with samples tested in 450-J.Ll aliquots in triplicate. Intra-assay variation was no more than 8%, and interassay variation was no more than 10%. Growth Studies

Subconfluent A431 cells were plated onto 35-mm plastic dishes (Corning Corporation, Corning, NY) in DME-5% FBS at an initial density of 10 X 103 / plate to a volume of 2 ml/plate. After 24 hours, the dishes were made serum-free and standards or 660

Buch et al.

Human Sertoli cell secreted growth factor

samples were added in triplicate, typically at a concentration of 33% (i.e., 0.5 ml sample/standard plus 1.0 ml DME). At the end of 72 hours, these dishes were treated with 0.25% trypsin-EDTA (Grand Island Biological Company, Grand Island, NY) and cell counts quantitated with a Coulter counter. Control dishes typically contained 33% DME with the same concentrations of insulin and FSH (DME plus IIFSH) as that in SCCM. Mitogenic Assay

Mitogenic activity was measured by the incorporation of [methyPH]-thymidine into trichloroacetic acid (TCA) precipitable material by confluent quiescent cultures of the mouse embryo fibroblast line, Swiss 3T3. The Swiss 3T3 cells were plated into 24-well dishes (Corning Corporation, Corning, NY) in DME-10% calf serum (CS) and allowed to reach confluence at 37°C in an atmosphere of 5% C02 /95% air. Twenty-four hours later, media was aspirated and samples were added at various concentrations, as indicated in the presence of DME-2% CS. From 18 to 24 hours after the samples were added, a 1-hour pulse-chase with 1 J.LCi/ well of [methyPH]-thymidine was performed at 37°C. The cells were trypsinized, and then harvested and precipitated onto filters with 10% TCA using a cell harvester (Brandel, Gaithersburg, MD). The filters then were dried, equilibrated in 5 ml of scintiverse (Fisher Scientific, Springfield, NJ), and counted in the scintillation counter. Chromatography

SCCM obtained from cultures in the presence of DME without phenol red was concentrated 5- to 10-fold with a YM-2 (>1000 daltons molecular weight) retention filter (Amicon Corporation, Danvers, MA). The SCCM concentrates were subjected to size exclusion high-performance liquid chromatography (HPLC) in 1-ml aliquots, after filtration through 0.2 J.LM acrodiscs (Gelman Sciences, Ann Arbor, MI). A Beckman Instruments (Palo Alto, CA) model 342 HPLC was used. The sizing columns employed were Spherogel TSK G2000 SW (7.5 mm X 600 mm, Beckman Instruments), including a Spherogel TSK guard column (Beckman Instruments). !socratic elutions were performed at room temperature, and DME without phenol red containing 25 mM 4-(2-hydroxyethyl)1-piperazinethane-sulfonic acid (HEPES, pH 7.2) served as the mobile phase. One-milliliter fractions from successive runs were collected, pooled, and Fertility and Sterility

filtered with an acrodisc, and tested directly in A431 growth studies, as described previously, at concentrations of33% (0.5 ml sample to 1.5 ml total volume). Characterization of the Mitogen in Sertoli Cell-Conditioned Medium

The sensitivity to protease was tested by incubation with trypsin (500 ~g/ml) for 1 hour at 37°C. Proteolysis was terminated by the addition of soybean trypsin inhibitor (2 mg/ml). Some samples were frozen and thawed twice, while others were tested for heat lability (100°C for 2 or 10 minutes). Stability to acid treatment was assessed with 1 N HCl to pH 3 (100 ~1/ml SCCM) for 1 hour at room temperature, and neutralized with 5 N NaOH (20 ~l/ml SCCM). Ability to bind heparin agarose (Sigma, St. Louis, MO) also was tested by batch binding (0.5 g/ml SCCM) at 37°C for 1 hour. After these treatments, SCCM was tested in A431 growth studies, as described previously. Protein Determination

Protein assays were performed according to the method of Bradford14 using the Bio-Rad protein reagent (Richmond, CA) with bovine serum albumin (BSA) as a standard. RESULTS

We examined the effect of increasing doses of hSCCM on A431 cell proliferation (Fig. 1). In the absence of serum after 3 days in culture, a 4-fold

70 60

0

"'a; 0

'ii c

u::

50 40 30 20 10 0 0

10

20

30

40

50

60

70

80

90

100

% hSCCM

Figure 1 Dose-response effect of hSCCM on A431 proliferation. The cells were incubated for 3 days with varying concentrations of hSCCM in serum-free DME (total volume, 1.5 ml/35-mm plate) up to 100% hSCCM. Data points represent the mean of triplicate samples ± SD. Vol. 49, No.4, April1988

stimulation above control levels of A431 growth was achieved with a concentration of 67% hSCCM in DME. Increasing the concentration of hSCCM above this (i.e., 100%) did not stimulate further proliferation. A similar dose-response effect was noted in testing the ability of hSCCM to stimulate [methyl3H]-thymidine incorporation in confluent quiescent cultures of Swiss 3T3 cells. An increase in thymidine incorporation more than 4-fold over control levels was noted in the presence of 50% hSCCM (not shown). There was some concern that the mitogenic activity of SCCM might be a generalized protein effect. In order to rule out this possibility, A431 proliferation in response to SCCM was evaluated in the presence and absence of serum. In the first experiment, we compared the ability of hSCCM 1 (SRS patient) to stimulate A431 cell growth in the presence or absence of 1% FBS. In the absence of serum, hSCCM stimulated cell proliferation (34.1 ± 3.5 X 103 cells) 2.8-fold as compared with control (12.2 ± 0.8 X 103 cells, P < 0.002). At a serum concentration of 1% FBS, the fold increase over control of A431 cell proliferation is preserved (2.6fold; 14.9 ± 1.1 X 103 cells in the control; 39.1 ± 3.7 X 103 cells for hSCCM-treated cells, P < 0.002). In a separate experiment, even in the presence of 5% FBS, hSCCM stimulated A431 proliferation greater than 2-fold over control (70.4 ± 3.0 X 103 cells in the control; 147.3 ± 1.1 X 103 cells in the hSCCM, P < 0.002). Clearly, the growth factor(s) present in SCCM is quite potent by virtue of the ability to stimulate proliferation significantly, even in the presence of 5% FBS. We also compared the growth factor activity of SCCM from various sources (Table 1). Using the protein concentration in the SCCM as a measure of secretory activity, we found that the degree of mitogenicity roughly correlated with changes in protein secretion. Based upon this correlation, when SCCM from 18- to 20-day-old rats is compared with SCCM from humans 6, 7, 10 (CAP patients), and 9 (CRP patient), the growth factor appears to be equally potent from both species. The lower levels of protein and mitogenicity in SCCM from humans 1 and 2 (SRS patients) apparently relate to a lower plating density in these primary cultures. Of further interest are the data relative to EGFLA detected in SCCM by RRA (Table 1). Taking into account the age discrepancies between the SRS patients (mean of 30 years) and the CAPpatients (mean of 65 years), it appears that only these Buch et al. Human Sertoli cell secreted growth factor

661

Table 1 Comparison of Mitogenicity and EGF-Like Activity (EGF-LA) from Various Sources of SCCM A431 Cell proliferation fold increase above control

Protein concentration

Conditioned media

EGF-LA

p.g/ml

ng/ml

rSCCM (18-20-day-old) rSCCM (35-day-old)

35 71

3.1 4.1

1.1 0.0

hSCCM 1 (SRS patient) hSCCM 2 (SRS patient)

20 20

1.9 2.0

0.9 1.3

hSCCM 6 (CAP patient) hSCCM 7 (CAP patient) hSCCM 10 (CAP patient)

35 35 30

3.4 3.4 2.9

0.0 0.0 0.0

hSCCM 9 (CRP patient)

32

3.1

0.0

lial cell growth factor (ECGF)/ 5 SCSGF did not bind to heparin agarose. Size exclusion HPLC then was employed in efforts to partially purify the growth factor. Prior to HPLC, hSCCM was concentrated with an Amicon filtration unit using a > 1000 daltons molecular weight retention filter. After Amicon filtration, 1-ml samples of hSCCM concentrate (approximately 500 1-lg protein/ml) were applied to HPLC size exclusion columns and fractions were collected and tested for mitogenicity upon A431 cells (Fig. 2). One major peak of activity at 8000 daltons was observed. This observation correlates well with that of Holmes et al. 9 noted for rSCCM. DISCUSSION

younger patients and the younger rats exhibit EGF-LA. Parallelism with the EGF standard curve in the RRA was noted for SCCM, as previously described. 9 Further experiments were conducted to characterize human SCSGF for comparison to that of the rat and to other known growth factors (Table 2). The mitogen was stable to freeze-thaw, heat, acid (pH 3.0), and trypsin. Although many growth factors are known to bind to heparin, such as endothe-

Table 2 Characterization of the Growth Factor Present in hSCCM as a Function of A431 Proliferationa Final cell no.

Treatment

Cell proliferation as% of hSCCM control

This report extends our initial observations on SCSGF in rats 9 by documenting its presence in human Sertoli cell cultures. Human SCSGF was found to be a potent mitogen for Swiss 3T3 cells and, especially, for A431 cells, which are not stimulated by any growth factors previously reported. 16 The potency of SCSGF was further exemplified by its ability to stimulate A431 proliferation greater than 2-fold over control in the presence of 5% FBS. Another unique property of SCSGF is its ability to displace radiolabeled EGF from its receptor, thus exhibiting EGF-like activity. This does not appear to be EGF for two reasons. When we tested for hEGF in an RIA that is sensitive to 0.2 ng, no hEGF immunoreactivity was noted in hSCCM (data not shown). Also, as our laboratory has documented previously, 9 EGF at the concentrations

XJ(i'

Experiment A DME + 1/FSH 15.0 ± 0.6 hSCCM 51.3 ± 0.2b hSCCM/freeze-thaw 49.7±1.7b 47.0 ± 2.1b hSCCM at 100°C for 2 min hSCCM at 100°C for 10 min 42.7 ± 3.5b DME + IIFSH +trypsin inhibitor 24.3 ± 1.7 DME + IIFSH + trypsin 22.3 ± 1.0 hSCCM + trypsin inhibitor 47.1 ± 0.7b 46.9 ± 1.9b hSCCM + trypsin DME + IIFSH, heparin agarose 19.6 ± 0.4 49.7 ± 2.2b hSCCM, heparin agarose Experiment B DME + 1/FSH hSCCM hSCCM + acid control hSCCM +acid a b

100 97 92 83

92 91

97 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

32.6 94.3 85.4 80.9

± ± ± ±

3.0 2.4b 2.8b 2.2b

Fraction Number

100 91 86

Values are the mean of triplicates ± SD. P < 0.0005 for all conditions relative to internal controls.

662

Buch et al.

Human Sertoli cell secreted growth factor

Figure 2 HPLC profile of hSCCM as a function of A431 proliferation. Pooled fractions from three successive HPLC runs were collected and tested in triplicate in the A431 cell growth assay at concentrations of 33% (total assay volume, 1.5 ml/35mm plate). Control level of A431 proliferation is 19 X 103 • Fertility and Sterility

present in SCCM is inhibitory to A431 proliferation; this is in direct contrast to the stimulatory action of SCSGF. Offurther interest is the fact that EGF-LA was only found in the SCCM from the younger estrogenized (SRS) humans and the young (18- to 20-day-old) rats. The obvious implication is that the secretion of EGF-LA may be a maturation- related phenomenon. Efforts to further characterize and purify SCSGF employed Amicon filtration and size-exclusion HPLC. Based upon the HPLC data, hSCSGF appears to have a molecular weight of 8000, which is consistent with the data from HPLC of rSCSGF previously reported by our laboratory.9'17 Like the rat SCSGF, human SCSGF is heat-, acid-, and trypsin-stable, and does not bind heparin agarose, 17 as do some other reported growth factors. 15·18 The present finding of trypsin stability is in contrast to the trypsin sensitivity of SCSGF reported previously by Holmes et al. 9 This discrepancy was caused by the previous use of a less-pure trypsin in which contaminants affected A431 proliferation. These contaminants were inactivated by heat treatment in the previously reported controls. Present testing employs a purified trypsin, which is subsequently inactivated by trypsin inhibitor, and trypsin inhibitor alone was compared to provide a complete set of controls. The stability of SCSGF to trypsin indicates that the factor does not have a trypsin-sensitive site or that the factor maintains activity despite trypsin cleavage. Of immediate relevance to these data is the recent report of a prostate-derived growth factor that is similar to SCSGF in that it is stable to trypsin and heat treatment, but differs significantly in molecular weight at 25,000 daltons. 19 It is possible that SCSGF is not a protein or may be a lower molecular weight factor bound to a protein. There have been several recent reports on the identification of polypeptide growth factors originating in the male reproductive tract. Somatomedin-e (Sm-C) immunoreactivity has been detected in human semen by Baxter et al. 20 More recently, Smith et al. 12 have partially characterized a somatomedin-like peptide from rat Sertoli cell-conditioned medium that reacts like human Sm-C in both RIA and RRA, but differs from Sm-C in molecular weight. Brown et al. 10 have reported on the presence of a growth factor in the rete testis fluid of the ram. A growth factor also has been noted in extracts from human prostatic tissue. 18·19 Finally, Feig et al. 11 have partially purified and characterized a growth factor from the cytosolic fraction of

sonicated seminiferous epithelium, which they have termed seminiferous growth factor (SGF). SCSGF has unique characteristics that differentiate it from all of these factors. Both SCSGF and mouse EGF are heat- and acid-stable with similar molecular weights (about 6000), 21 but differ in their ability to stimulate A431 cell proliferation. Somatomedins do not displace radiolabeled EGF from binding to its receptor, 9 and since they do not stimulate A431 proliferation, 17 they are clearly distinct from SCSGF. Furthermore, the Sm-C-like peptide reported by Smith et al. 12 in rat SCCM has a molecular weight of 25,000 daltons. As regards the growth factor noted in the rete testis fluid of rams, it has an apparent molecular weight of 45,000 daltons and is sensitive to treatment with trypsin/ chymotrypsin. 10 This contrasts to SCSGF, with an apparent molecular weight of 8000 and trypsin stability. The SGF characterized by Feig et alP is sensitive to treatment with acid (pH 4.0), trypsin, and heat, and has a molecular weight of 15,700. It has been detected in homogenates of the seminiferous epithelium from prepubertal mouse and calf, adult mouse, rat, and guinea pig. The properties of SGF contrast with those of SCSGF in that the secreted factor is stable to treatment with acid (pH 3.0), trypsin, and heat, and has an apparent molecular weight of 8000 daltons. Additionally, SCSGF does not bind to heparin agarose, in contrast to SGF, which does bind heparin agarose, as reported by Story et al. 18 Furthermore, Story et al. suggest that SGF may be identical to the growth factor they have noted in extracts of prostate tissue called prostate growth factor or PrGF. 18 Certainly, SGF should be classified within the family of ECGF, which all bind to heparin. 15 Feig et al. 11 noted other physiochemical similarities between SGF and ECGF. The finding of EGF-LA associated with SCSGF is of special interest for several reasons. No other growth factors have been reported that displace radiolabeled EGF from binding to its receptor. 22 Elson et al. 23 have identified EGF-LA in human seminal fluid and in the cytosol fraction from tissue homogenates of prostate, testis, epididymis, and seminal vesicle. However, the EGF-LA in human seminal fluid has been identified as EGF in our laboratory by a highly specific RIA (data not shown). Seemingly of great relevance is the recent report by Tsutsumi et al., 13 in which EGF was found to play a role in normal spermatogenesis of the mouse. In their report, Tsutsumi et alY re-

Vol. 49, No.4, April1988

Buch et al.

Human Sertoli cell secreted growth factor

663

moved the submandibular glands of adult mice and thus reduced the serum levels of EGF to zero. Four weeks after this treatment, spermatogenesis was significantly reduced. Supplementation with exogenous EGF to restore normal serum levels also restored normal spermatogenesis in these mice. 13 In contrast, however, are the findings of Welsh and Hsueh, 24 who note a role for EGF as an inhibitor of gonadotropin-stimulated testosterone biosynthesis by rat interstitial cells in culture. Furthermore, high-affinity binding sites for 125 I-EGF were identified on the Leydig cells, and were believed by the authors to indicate a possible direct action of EGF upon the Leydig cells. 24 Based on this data, one could postulate an inhibitory role for EGF in the spermatogenesis of the rat in direct contrast to the findings of EGF-stimulated spermatogenesis in the mouse, as reported by Tsutsumi et al. 13 Polypeptide growth factors are hormone-like in their structure and function and may act in an endocrine, paracrine, or autocrine manner. An autocrine function for SCSGF is highly improbable because only fetal Sertoli cells are known to proliferate in the human. Of greater probability is a paracrine role for SCSGF as a regulator of germ cell proliferation and maturation, as well as a regulator of Leydig cell function. Indeed, Verhoeven and Cailleau 25 have noted a factor(s) in spent media from rat Sertoli cells that stimulates steroidogenesis by Leydig cells in vitro. However, preliminary results indicate that their factor has a molecular weight > 10,000 daltons and is sensitive to both heat and trypsin (32), thus differentiating it from SCSGF. In summary, human SCSGF is a potent mitogen that stimulates cell proliferation even in the presence of 5% FBS. It has unique physiochemical properties which distinguish it from other known growth factors. We have now identified SCSGF in conditioned media from human and rat Sertoli cells. SCSGF may play a key role in normal spermatogenesis as a paracrine factor that regulates stem cell renewal, germ cell proliferation and maturation, or Leydig cell function. The true role of SCSGF in normal spermatogenesis can only be clarified after its purification to homogeneity. Acknowledgment. We thank Peter T. Scardino, M.D., and David R. Roth, M.D., for providing some of the surgical specimens necessary for this study. We would also like to thank Mr. Gerald Spotts for his excellent technical assistance. Finally, we are indebted to Ms. Elizabeth Barrick for her outstanding work in preparing this manuscript.

664

Buch et al.

Human Sertoli cell secreted growth factor

REFERENCES 1. Dym M, Fawcett DW: Observations on the blood-testis

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

barrier of the rat and on the physiological compartmentation of the seminiferous epithelium. Bioi Reprod 3:308, 1970 Fawcett DW: Ultrastructure and function of the Sertoli cell. In Handbook of Physiology, Endocrinology, Vol 5, Male Reproductive System, Edited by RO Greep, EB Astwood. Washington, D.C., American Physiological Society, 1975, p 21 Black VH: Gonocytes in fetal guinea pig testis: phagocytosis of degenerating gonocytes by Sertoli cells. Am J Anat 131:415, 1971 Louis BG, Fritz IB: Follicle-stimulating hormone and testosterone independently increase the production of androgen-binding protein by Sertoli cells in culture. Endocrinology 104:454, 1979 Steinberger A, Steinberger E: Secretion of an FSH-inhibiting factor by cultured Sertoli cells. Endocrinology 99:918, 1976 Lipshultz Ll, Murthy L, Tindall DJ: Characterization of human Sertoli cells in vitro. J Clin Endocrinol Metab 55:228, 1982 Holmes S, Lipshultz Ll, Smith, RG: Regulation of transferrin secretion by human Sertoli cells cultures in the presence or absence of human peritubular cells. J Clin Endocrinol Metab 59:1, 1984 Waites GMH, Setchell BP: Some physiological aspects of the function of the testis. In The Gonads, Edited by KW McKern. New York, Appleton, 1969, p 649 Holmes SD, Spotts G, Smith RG: Rat Sertoli cells secrete a growth factor that blocks epidermal growth factor (EGF) binding to its receptor. J Bioi Chern 261:4076, 1986 Brown KD, Blakely DM, Renville A, Setchell BP: Rete testis fluid contains a growth factor for cultured fibroblasts. Biochem Biophys Res Commun 105:391, 1982 Feig LA, Klagsbrun M, Bellve AR: Mitogenic polypeptide of the mammalian seminiferous epithelium: biochemical characterization and partial purification. J Cell Bioi 97:1435, 1983 Smith EP, Svoboda ME, VanWyk JJ, Kierszenbaum AL, Tres LL: Partial characterization of a somatomedin-like peptide from the medium of cultured rat Sertoli cells. Endocrinology 120:186, 1987 Tsutsumi 0, Kurachi H, Oka T: A physiological role of epidermal growth factor in male reproductive function. Science 233:975, 1986 Bradford MM: A rapid and sensitive method for the quantitation of microgram quantities of protein using the principle of protein-dye binding. Anal Biochem 72:248, 1976 Burgess WH, Mehlman T, Friese! R, Johnson WV, Maciag T: Multiple forms of endothelial cell growth factor: rapid isolation and biological and chemical characterization. J Bioi Chern 260:11389, 1985 Barnes DW: Epidermal growth factor inhibits growth of A431 human epidermoid carcinoma in serum-free cell culture. J Cell Bioi 93:1, 1982 Lamb DJ, Spotts G, Holmes S, Smith RG: Unique characteristics of rat Sertoli cell secreted growth factor. In Cell Biology of the Testis and Epididymis, Edited by M -C Orgebin-Crist. New York, New York Academy of Science. In press Fertility and Sterility

18. Story MT, Sommer B, Jacobs SC, Lawson RK: Comparison of the physiochemical properties of prostatic growth factor and seminiferous growth factor (Abstr 20). Presented at the 81st annual meeting of the American Urological Association, New York, May 18-22, 1986 19. Maehama S, LiD, Nanri H, Leykam JF, Deuel TF: Purification and partial characterization of prostate-derived growth factor. Proc Natl Acad Sci USA 83:8162, 1986 20. Baxter RC, Martin JL, Handelsman DJ: Identification of human semen insulin-like growth factor-I!somatomedin-C immunoreactivity and binding protein. Acta Endocrinol (Copenh) 106:420, 1984 21. Cohen S: Isolation of a mouse submaxillary gland protein

Vol. 49, No.4, April1988

22.

23.

24.

25.

accelerating incisor eruption and cycled opening in the newborn animal. J Biol Chern 237:1555, 1962 DeLarco JE, Todaro GJ: Growth factors from murine sarcoma virus-transformed cells. Proc Natl Acad Sci USA 75:4001, 1978 Elson SD, Browne CA, Thorburn GD: Identification of epidermal growth factor-like activity in human male reproductive tissues and fluids. J Clin Endocrinol Metab 58:1, 1984 Welsh TH, Hsueh AJW: Mechanism of the inhibitory action of epidermal growth factor on testicular androgen biosynthesis in vitro. Endocrinology 110:1498, 1982 Verhoeven G, Cailleau J: A factor in spent media from Sertoli cell-enriched cultures that stimulates steroidogenesis in Leydig cells. Mol Cell Endocrinol40:57, 1985

Bucb et al. Human Sertoli cell secreted growth factor

665