In vitro differentiation of germ cells from nonobstructive azoospermic patients using threedimensional culture in a collagen gel matrix

In vitro differentiation of germ cells from nonobstructive azoospermic patients using threedimensional culture in a collagen gel matrix Jae-Ho Lee, Ph.D., a Myung C. Gye, Ph.D., b Kyoo Wan Choi, Ph.D., a Jae Yup Hong, M.D., Ph.D., c Yong Bok Lee, M.D., c Dong-Wook Park, Ph.D., d Seung Jae Lee, M.D., c and Churl K. Min, Ph.D. d a Laboratory of IVF, MDplus LIS; b Department of Life Sciences, Hanyang University; c Mirae and Heemang Ob/Gyn Clinic; and d Department of Molecular Science and Technology, Ajou University, Suwon, South Korea Objective: To assess the effectiveness of the three-dimensional culture of spermatogenic cells in a collagen gel matrix from nonobstructive azoospermic patients and examine the relation between the success rate of in vitro spermatogenesis and serum FSH level as a diagnostic prediction. Design: Prospective study using radioimmunoassay, immunocytochemistry, and flow cytometry with primary cultured cells. Setting: Gynecologic clinics and human reproduction research laboratory. Patient(s): Primary culture of spermatogenic cells established from 18 nonobstructive azoospermic patients who underwent histologic diagnoses. Intervention(s): Primary culture of spermatogenic cells in a collagen-based gel matrix, subjected to immunological and flow cytometric analyses. Main Outcome Measure(s): In vitro culture of spermatogenic cells was established in an extracellular milieu that more closely resembled the in vivo condition. The number of chromosomes in newly generated cells during culture was determined by fluorescence-activated cell sorter (FACS) and immunocytochemical analysis. Effects of FSH on the differentiation of the spermatogenic cells were measured. Result(s): Results of histologic studies indicated that 8 of 18 patients showed the spermatocyte arrest. Immunocytochemical and FACS analysis indicated that after 12 days in culture, haploid cells comprised 11% 37% of the cultured cell population with a characteristic expression of a cellular marker for spermatids. The serum level of FSH appeared to be closely correlated with an increase in the number of haploid cells in culture. Conclusion(s): The present three-dimensional culture in a collagen gel matrix provides a suitable means by which spermatocytes could be induced to differentiate into presumptive spermatids in vitro. In addition, the plasma FSH level could be a good indicator for the success of differentiation of cultured spermatogenic cells obtained from patients with spermatogenic arrest. (Fertil Steril 2007;87:824 33. 2007 by American Society for Reproductive Medicine.) Key Words: In vitro spermatogenesis, three-dimensional culture, collagen gel matrix Successful in vitro differentiation of spermatogenic cells into spermatids appears to have an extremely attractive potential for the treatment of male infertility, particularly caused by spermatogenic arrest (1 3). Various methods for the in vitro spermatogenesis have been developed, mainly focusing on testicular cell cultures that are directed at studying male germ cell differentiation. Among them, tissue culture, organ culture, and co-culture systems have been attempted extensively to achieve meiotic or postmeiotic differentiation of cultured male germ cells (1, 4 12). Despite a few clinical reports that a normal child was born after fertilization with germ cells from a man with maturation arrest at the primary spermatocyte stage (1), effective procedure for the completion of spermatogenesis, including meiosis, still remains to be seen. Received December 2, 2005; revised and accepted September 4, 2006. Supported by grants No. R01-2003-000-10672-0 from KOSEF and No. KRF-2002-041-C00249. Reprint requests: Churl K. Min, Ph.D., Department of Molecular Science and Technology, Ajou University, Suwon 443-749, S. Korea (FAX: 82-31-219-1885; E-mail: minc@ajou.ac.kr Ever since three-dimensional (3-D) cell culture, a new mode of cell culture system that is considered to authentically represent a cell s environment in a living organism, was proposed (13, 14), the 3-D culture system has drawn much attention in the field of cell biology (3, 15). Spermatogenesis is the process of germ cell proliferation and differentiation in the testis, which eventually leads to the production of haploid sperms. Spermatogenesis requires complex endocrine and autocrine regulation, as well as direct cell-tocell interactions (16). In support of cell-to-cell communication, co-culture of germ cells with somatic cells such as Sertoli cells or Vero cells has been extensively studied to increase the developmental potential of spermatogenic cells in vitro (4, 6, 7). Likewise, 3-D cell culture has been attempted with reconstructed testis tissues to gain more insights into germ somatic cell interactions or germ cell extracellular matrix (ECM) interactions during in vitro spermatogenesis (17 19). The purpose of this study was to estimate the efficacy of 3-D culture in a collagen gel matrix in supporting germ cell 824 Fertility and Sterility Vol. 87, No. 4, April 2007 0015-0282/07/$32.00 Copyright 2007 American Society for Reproductive Medicine, Published by Elsevier Inc. doi:10.1016/j.fertnstert.2006.09.015

survival in a long-term in vitro culture and in resumption of meiosis and spermatid differentiation of human immature spermatogenic cells isolated from nonobstructive azoospermic men. In addition, patients characteristic serum FSH levels and the extent to which spermatocytes undergo differentiation during culture were carefully compared to find any significant correlation between them. MATERIALS AND METHODS Testicular Cell Preparation Eighteen patients with nonobstructive azoospermia but normal karyotypes participated in the present experiment, which took place between January 2004 and January 2005. Eight of the participants suffered from premeiotic or early meiotic maturation arrest, thus enrolled in the present in vitro culture studies. Testicular tissue was obtained by open testicular biopsy from multiple sites in each testis. The collected fresh testis tissues were washed and minced to derive single cells. The mechanically dissociated testicular cells were washed twice in Dulbecco s minimum essential medium (DMEM)/ F12 medium (GIBCO-BRL, Grand Island, NY) containing 10% bovine calf serum (BCS) (GIBCO-BRL), and the resulting germ cells were examined for stage determination in spermatogenesis before culture. Collagen Preparation Tails of sacrificed Sprague Dawley rats (44 48 days old) were removed, skinned, and placed in distilled water. A dissecting probe was used to pull individual tendon fibers through the surrounding fascia out from the tail. The collagen tendons were dried and sterilized overnight in 70% ethanol. Individual tendons ( 2 mg/ml) were placed in diluted acetic acid (0.01%) for 72 hours at 4 C for full dissolution. The viscous mixture was centrifuged at 15,000 g for 30 minutes and supernatant was stored at 4 C as collagen solution. Reconstitution of Testicular Cells in a Collagen Gel Matrix Two milliliters of collagen solution was poured into a bacteriological dish (50 mm in diameter; Nunc, Roskile, Denmark) to which collagen gel matrix and somatic cells adhered poorly. The collagen solution was mixed with 0.4 ml of fetal bovine serum (GIBCO-BRL), 0.8 ml of 5 concentrated DMEM/F12, and 0.8 ml of testicular cell suspension in modified DMEM/F12 containing 5% Matrigel (BD Biosci., San Jose, CA) before allowed for gelation by gentle agitation. The resulting cell collagen gel matrix was incubated in a culture medium consisting of DMEM/F12 medium supplemented with 10 mg/ml insulin transferrin selenium solution (BD Biosci.), 10 4 M vitamin C (Sigma- Aldrich, Saint Louis, MO), 10 mg/ml vitamin E (Sigma- Aldrich), 3.3 10 7 M retinoic acid (Sigma-Aldrich), 3.3 10 7 M retinol (Sigma-Aldrich), 1 mm pyruvate (Sigma- Aldrich), 100 miu recombinant FSH (Organon-Korea, Seoul, S. Korea), 10 7 M T (Sigma-Aldrich), 10 7 M dihydrotestosterone (Sigma-Aldrich), 10% antibiotic antimycotic solution (GIBCO-BRL), and 10% fetal bovine serum (GIBCO-BRL) at 32 C with gentle shaking (45 rpm/min) in a humidified atmosphere of 5% CO 2. Culture medium was replaced on every other day. Cytological and Immunohistochemical Visualization The cell collagen gel mixture after 12 days in culture was dissolved by type 1 collagenase (Sigma-Aldrich) to release embedded cells from the collagen gel matrix. The mixture was rinsed with phosphate-buffered saline (PBS) and aliquoted. One aliquot was observed under an inverted microscope (Axioscope, Carl Zeiss, Oberkochen, Germany) for cytological analysis. Anther aliquot was smeared onto L- lysine-coated microscope slides and fixed by 100% cold acetone for immunocytochemical staining. Immunocytochemical staining was performed by using an avidin biotin immunoperoxidase technique according to the manufacturer s manual (Dako, Glostrup, Denmark). Briefly, rabbit polyclonal anti-prm2 antibody, which was kindly given by Dr. Braun in the University of Washington, was diluted to 1:100 before incubation with cells. The PRM2 immunoreactivity was visualized by using biotinylated polyvalent antibody and avidin horseradish peroxidase followed by chromogenic development with a solution of AEC substrate-chromogen (Dako). All slides were counterstained with hematoxylin. As a negative control, cells were incubated with preimmune serum. Flow Cytometric Analysis Flow cytometry was performed as described previously (20). Briefly, ethanol-fixed cells (1 2 10 6 cells/ml) were washed twice in PBS. After centrifugation at 500 g, cells were incubated in a staining solution containing 25 g/ml propidium iodide, 40 g/ml RNase, and 0.3% Tween-20 in PBS at room temperature for 20 minutes. After further washing twice in PBS, cells were analyzed by FACSCaliber flow cytometer (BD Biosci.). The fluorescent signals from propidium iodide were recorded, and a cytogram of DNA area versus cell count was used to select cell populations based on their DNA contents. A total of 10,000 events were recorded for each histogram. The relative numbers of each testicular cell type such as round spermatids (1C, haploid), secondary spermatocytes or spermatogonia (2C, diploid), and primary spermatocytes (4C, tetraploid) were calculated using software Summit (Cytomation, Fort Collins, CO). Quantitative Determination of Hormones Serum concentrations of FSH, LH, PRL, and T were analyzed by RIA using a commercial assay kit (T, Orion Diagnostica, Espoo, Finland; FSH, LH, and PRL, IRMA-mat Kits, Byk-Sangtec Diagnostica, Dietzenbach, Germany) according to the manufacturer s protocol. Fertility and Sterility 825

FIGURE 1 Histologic examinations of testicular biopsy samples from men with nonobstructive azoospermia. Fresh testicular biopsy samples from 18 patients with nonobstructive azoospermia were taken for histologic examinations by the hematoxylin and eosin (H & E) staining method (A, B) or by immunostaining with anti-prm2 antibodies (C, D). A and C represent a testicular tissue of spermatogenic arrest, whereas B and D represent a normal spermatogenic tissue. The arrow in D indicates PRM2-positive cells. Magnification, 400. Statistical Analysis Data were analyzed by means of two-sided Student s t-test for independent samples, and all analyses were performed with SPSS version 11.5. Differences of P values.05 were considered statistically significant. RESULTS Histologic and Immunohitochemical Evaluation of Testicular Tissues Fresh biopsy testicular tissues were isolated from 18 patients with symptoms of nonobstructive azoospermia and were subjected to a diagnostic histologic examination 826 Lee et al. by hematoxylin and eosin (H & E) staining or by immunostaining for protamine 2 (PRM2), a cellular marker for mature spermatozoa. Of these, 10 patients showed a normal spermatogenesis, 7 patients of germ cell arrest at spermatocyte stage, and 1 patient of germ cell arrest at round spermatid stage. Figure 1 illustrates representative seminiferous tubules of a patient with immature germ cell arrest at spermatocyte stage (Fig. 1A) and of a patient with normal spermatogenesis (Fig. 1B). Note the round spermatids in the middle of the seminiferous tubule from the normal spermatogenic tissue, whereas very few spermatids were found in the tissue of spermatogenic arrest. An immunohistochemical staining for RPM2 confirmed Three-dimensional culture of spermatogenic cells Vol. 87, No. 4, April 2007

FIGURE 2 Photographs of a collagen gel matrix and testicular cells within the collagen gel mixture before and after culture in vitro. Testicular cells prepared from testicular biopsy tissues were mixed with a collagen-rich solution that was previously isolated from rat tails according to the methods described in the Materials and Methods section. The resulting cell collagen mixture was solidified and cultured in the culture medium at 32 C for 12 days. The overall morphologies of the solidified collagen gel matrix before and after 12 days in culture are compared in A and B, respectively, whereas testicular cells in the gel mixture before and after 12 days in culture are shown in C and D, respectively. Magnification in C and D, 200. the presence of normal spermatozoa in the normal seminiferous tubule (Fig. 1D, indicated by an arrow) but not in the seminiferous tubule of a patient with spermatogenic arrest (Fig. 1C). Characterization of 3-D Culture in a Collagen Gel Matrix Testicular cell suspensions were mixed with a collagen-rich solution to reconstitute testicular tissues that might help testicular cells grow and differentiate in a long-term culture. The solidified collagen gel matrix was reduced in size by about 75% after 12 days in culture in a humidified incubator at 32 C, resulting in an increased testicular cell-to-cell interaction or cell-to-ecm interaction (Fig. 2A,B). In addition, the testicular cells, both germ cells and somatic cells, within the gel appeared to remain alive and homogeneous throughout the culture period (Fig. 2C,D). Fertility and Sterility 827

Cytological Evaluation of Testicular Cells in 3-D Culture The evaluation of development of spermatocytes into spermatids or spermatozoa during the 3-D culture was performed by a cytological analysis of the cultured testicular cells. At day 0, pachytene spermatocytes or primary spermatocytes were present in the testicular cell population, but no spermatid was observed (Fig. 3A,C). To the contrary, at day 12 in culture, round spermatids emerged among the cultured cells. Overall, fewer pachytene spermatocytes, but more round and elongating spermatids, were observed at day 12 compared to at day 0 (Fig. 3B,D). The presence of mature spermatids in the testicular cell culture was further confirmed by immunoreactivity against PRM2. Before culture, the testicular cells from patients with maturing arrest at the spermatocyte stage showed very few cells that were positive for PRM2 expression (Fig. 3E). In vitro 3-D culture for 12 days, however, led to an in vitro maturation of arrested spermatocytes as evidenced by PRM2 positivities (Fig. 3F). As a control, few PRM2-positive cells were found when cells were stained with preimmune serum (Fig. 3G). Flow Cytometric Analysis of the Cultured Testicular Cells For a more quantitative analysis, we further characterized a shift in the testicular cell population, which could be due to a release from maturation arrest or differentiation into mature spermatids, on the basis of the DNA content and the RMP2 immunoreactivity. The round or elongated spermatids should be haploid (1C) due to two consecutive meiotic divisions, whereas spermatogonia, secondary spermatocytes, and somatic cells are all diploid (2C), and primary spermatocytes are tetraploid (4C). After propidium iodide staining, FACS analysis was performed with a testicular cell suspension before and after 12 days in culture in a collagen gel matrix. As a control, normal fresh testicular cell suspension was analyzed. The distribution pattern of a normal testicular cell population features a dominant haploid subpopulation with lesser amount of presumptive diploid and tetraploid cells (Fig. 4A). To the contrary, the testicular cells obtained from men with spermatogenic arrest largely consisted of 2C and 4C subpopulations with a negligible amount of 1C cells (Fig. 4B to D), indicating that the arrest was either at the primary (4C) or at secondary (2C) spermatocyte stage. Interestingly, testicular cells of spermatogenic arrest tissues could be further divided into three groups based on the ability of the arrested spermatocytes to differentiate into spermatids during the in vitro culture. The first group, consisting of patients nos. 1 and 2, represents arrested spermatocytes with a high differentiation capability during in vitro culture (Fig. 4B). Likewise, patients nos. 3 through 6 comprise a group of medium differentiation capability, and patients nos. 7 and 8 belong to a low differentiation group, as evidenced by an apparent lack of shift in the subpopulation proportion during the culture. In the high differentiation group, at day 0, the testicular cells consisted of roughly equal amount of 2C cells and 4C cells, with a negligible presence of 1C cells. At day 12 in culture, however, the proportion of 1C cell population was drastically increased with a concomitant decrease in the 4C population, indicating that many germ cells were released from the block at spermatocyte stage, thus developed into spermatids (Fig. 4B). The transition from 4C or 2C to 1C during 12 days in culture became less significant in groups with medium and low differentiation (Fig. 4C,D). For a quantitative evaluation of haploid cells or spermatids, which are RPM2-positive, testicular cells were also analyzed flow cytometrically after immunostaining for RPM2. Results are summarized in Table 1, revealing the proportions of haploid cell from patients nos. 1 and 2 were 4.43% and 4.93%, respectively, but the respective values were increased to 28.02% and 30.87% after 12 days in culture. To the contrary, the proportions of haploid cells from patients nos. 7 and 8 were 13.21% and 11.51% before culture, respectively, which remained similar with 18.03% and 15.31%, respectively, at day 12. Therefore, this striking difference between patients nos. 1 through 6 and patients nos. 7 and 8, in particular, with respect to their ability of in vitro differentiation into mature spermatids from spermatocytes could be an interesting issue to be answered. Also noted in Table 1 is differentiation rate or ability to differentiate into mature spermatids or spermatozoa. For instance, in patient no. 1, a representative in the high differentiation group, the percentage of haploid cells was 4.43%, which was then increased up to 28.02% (6.33-fold increase) after 12 days in culture. In patient no. 3, a representative in the medium differentiation group, the corresponding values were 13.92% and 36.53% (2.62-fold increase), whereas in patient no. 7, who belongs to the low differentiation group, those values were 13.21% and 18.03%, corresponding to a 1.36-fold increase. The fold increase in the percentage of the haploid cells during the culture period was referred to as differentiation rate. Based on the differentiation rate, the testicular cell populations can be clearly classified into three groups: high differentiation rate group in which the differentiation rate is 6.0 or larger; medium differentiation rate group in which the differentiation rate is about 2.0; low differentiation rate group, where the differentiation rate is 2.0 or less. Correlation Between the Serum FSH Level and the Differentiation Rate Eight patients with the germ cell maturation arrest at the spermatocyte stage were further subjected to endocrinological measurement for the serum level of FSH, LH, T, and PRL. As shown in Table 2, the serum concentrations of LH, PRL, and T were all in the normal range, which were 1 6 IU/L, 3.2 12.0 ng/ml, 2.8 6.8 ng/ml, respectively. However, the serum FSH concentrations of patients nos.7 and 8 828 Lee et al. Three-dimensional culture of spermatogenic cells Vol. 87, No. 4, April 2007

FIGURE 3 Morphological and immunohistologic analyses of testicular cells before and after culture in vitro. Testicular cells were photographed before (A, C) and after (B, D) 12 days in culture at magnification 200 (A, B) and 400 (C, D). Arrows in A to D indicate primary spermatocyte (PS) and round spermatid (RS). Testicular cells were photographed after immunostaining with anti-prm2 antibody before (E) and after (F) 12 days in culture at magnification 200. Arrows in F indicate PRM2-positive cells. Testicular cells were immunostained with preimmune serum as a control and photographed at magnification 200 (G). Fertility and Sterility 829

FIGURE 4 Flow cytometric analysis of the DNA content in testicular cells. Testicular cells were analyzed flow cytometrically after propidium iodide staining based on the nuclear DNA content. Histograms represent results of FACS analysis of a representative testicular cell suspension obtained from a normal man (A), from groups with high meiotic differentiation rate before and after 12 days in culture (B), from groups with medium differentiation rate (C), and from groups with low differentiation rate (D). 1C haploid cells; 2C diploid cells; 4C tetraploid cells. were significantly higher than the rest when compared against patients nos. 1 through 6 by using two-sided Student s t-test (P.05). In an attempt to search for a diagnostic clue as to in vitro amenability of spermatogenic arrest, we compared pairwisely the serum FSH level and differentiation rates that are calculated in Table 1. As illustrated in Figure 5, two patients (patients nos. 7 and 8) whose differentiation rates were less than 2.0 were expected to show a low recovery of mature spermatids through in vitro culture revealed high serum FSH levels, whereas the patients whose differentiation rates were moderate to high (2.0 or higher, patients nos. 1 through 6) revealed normal serum FSH levels (1 6 IU/L). The patient s serum FSH level, therefore, could be a good diagnostic indicator for the success of in vitro germ maturation procedure in that the serum FSH level and differentiation rates appear to be reversely correlated: the higher the serum FSH level, the lower the success rate for in vitro germ cell maturation. DISCUSSION The completion of spermatogenesis in vitro remains a daunting task in reproductive biology. The present studies define a new 3-D culture system in an attempt to devise a long-term culture that might help result in an improvement of in vitro spermatogenesis of human germ cells. Several reports in recent years have demonstrated the importance of somatic cells in stimulating germ cell progression during culture. Therefore, this study will provide a stepping stone for future efforts aimed at optimizing in vitro culture conditions, including the 3-D co-culture system. We established a culture system in which dissociated biopsy testicular cells were embedded, thus allowing for aggregation within a collagen gel matrix. Reaggregation in a collagen gel is likely to re-establish Sertoli cell germ cell contacts that might stimulate the progression of immature germ cells during the period of culture. Moreover, embedding in a collagen gel will provide a structure that mimics the ECM of structural proteins and other biological molecules found in living tissues. It has been previously reported that reaggregates of male germ cells and Sertoli cells on a filter lacking a collagen matrix flatten and no spermatocyte is formed, indicating that a 3-D structure of the cell aggregates is required for further differentiation in vitro (18). We should also consider the possibility that the collagen matrix retains growth factors or other humoral factors that are secreted by Sertoli cells in close proximity of germ cells for spermatogonial proliferation. The exact mechanism whereby ECMs support spermatogenesis in vitro remains to be explained. Nonetheless, it is perhaps no surprise that ECM will play an instructive role for a variety of cellular activities, including germ cell differentiation, given the fact that the cell surface contains receptors to respond to extracellular sig- 830 Lee et al. Three-dimensional culture of spermatogenic cells Vol. 87, No. 4, April 2007

TABLE 1 The percentage of haploid cells before and after 12 days in culture. Patient no. Haploid cell population at day 0 (%) Haploid cell population at day 12 (%) Differentiation rate 1 4.43 28.02 6.33 2 4.93 30.87 6.26 3 13.92 36.53 2.62 4 9.62 24.44 2.54 5 14.83 37.07 2.50 6 13.58 27.85 2.05 7 13.21 18.03 1.36 8 11.51 15.31 1.33 Note: The testicular cells were isolated from eight nonobstructive azoospermic men, subjected to immunostaining for RPM2 and sorted by FACS before and after 12 days in culture in a collagen matrix. The in vitro differentiation rate was calculated as the fold-increase in percentage of haploid cells during 12 days in culture. nals. Sertoli cells cultured on top of ECM components assume a phenotype and morphology more characteristic of in vivo differentiated cells (19, 21). When grown on ECMs, Sertoli cells secrete markedly greater amounts of total protein, androgen-binding protein, transferrin, and type I collagen than when grown on plastics (22). In addition, Sertoli cells growing within reconstituted basement membrane gels induce morphogenesis of the cells into cords, which closely resemble the organ from which the cells were dissociated. They apparently provide an environment permissive for germ cell differentiation (22). Given this complex mechanical and biochemical interplay, it is possible that ECM materials extracted from rat tails will be beneficiary for male germ cell differentiation during in vitro culture. TABLE 2 The serum levels of T, FSH, LH, and PRL in eight patients showing spermatogenic arrest. Patient no. T (ng/ml) LH (IU/L) FSH (IU/L) PRL (ng/ml) 1 4.10 1.00 4.41 4.42 2 4.79 5.14 5.19 5.43 3 4.47 3.07 6.55 6.03 4 6.79 4.39 8.46 3.23 5 3.13 2.29 8.49 8.99 6 5.31 5.82 8.91 12.00 7 2.82 3.71 13.41 4.52 8 3.90 2.52 17.02 7.81 Note: The concentrations of T, FSH, LH, and PRL present in the serum were determined by RIA as described in the. In native testis, Sertoli cells are directly nursing developing germinal cells (16, 23). Among four somatic cells, including Sertoli cell, Leydig cell, extrasomatic cells, and myoid cell, Sertoli cells are the primary somatic cells to directly interact with the developing germinal cells (7, 23, 24). The cytoarchitectural arrangement between Sertoli cells and the developing germinal cells provide one of the most complex examples of an environmental cell-to-cell interaction (23, 25, 26). In view of this, direct association of germ cells and Sertoli cells was re-established by aggregation with lectins (27) or by phytohemagglutinin (15), followed by encapsulation with calcium alginate. Alginate-encapsulated cells were found to aggregate approximately the cell density and linear arrangement of native tubules that maintained close intercellular association, leading to a successful in vitro production of haploid germ cells in the long-term culture (15). Very few round spermatids were observed in freshly dissociated seminiferous tubules from patients with nonobstructive azoospermia, which were evidenced by both H& E staining and PRM2 immunostaining (Fig. 1), suggesting that spermatogenesis was arrested at most at the spermatocyte stage. However, we should very carefully interpret this result. One point comes from FACS analysis of the RPM2- positive haploid cells (Table 1). The percentage of haploid cells in the freshly dissociated testicular tissue comprises 4% 15% of the whole testicular cell population depending on the patients. Therefore, about 10% of the testicular cells are haploid cells, spermatids, or spermatozoa, even in testicular tissues obtained from men with spermatocyte arrest. This speculation might interpret the fact that in many IVF laboratories, intracytoplasmic sperm injection (ICSI) has been successful for finding mature sperms in patient with spermatogenic arrest. The discrepancy between the histologic observation and FACS analysis could be due to a small focus in the microscopy. Fertility and Sterility 831

FIGURE 5 Relationship between the serum FSH level and the in vitro meiotic differentiation rate. The level of serum FSH of nonobstructive azoospermic men was measured by RIA as described in the Materials and Methods section. In vitro meiotic differentiation rate was calculated as the percentage of the haploid cell population emerging after 12 days in culture. *P.05, each compared with patients nos. 1 through 6. Interestingly, we have shown that the serum FSH levels of patients whose germ cells are capable of resuming spermatogenesis in vitro is significantly lower compared with patients whose germ cells remain blocked during the culture. In fact, our results revealed an inverse correlation: the higher serum FSH concentration, the lower in spermatogenesis. This result is consistent with a previous report by Tesarik et al. (28) that high serum FSH concentration is associated with poor in vitro differentiation potential of germ cells. In particular, FSH concentrations exceeding 12 IU/L predict a poor prognosis, whereas concentrations between 1.0 and 6.0 IU/L are often associated with a good in vitro differentiation potential of germ cells. It remains to be determined whether the refractoriness of germ cells from men with elevated serum FSH concentrations can be overcome by increasing the concentration of FSH in the culture medium. The other points that deserve a further discussion are the etiology for the spermatogenic arrest/defect. It is well established that Sertoli cells stimulates spermatogenesis in response to FSH. Also well known is that Sertoli cells control the level of FSH by a negative feedback by a hormone inhibin produced by the Sertoli cells. Thus, malfunction of Sertoli cells is likely to be one possible explanation for the present observation: high FSH level is correlated with a low outcome of the sperm production in vitro. It remains to be seen whether defective Sertoli cells are one of the etiologies for the spermatogenic defect. In conclusion, the present 3-D collagen gel culture system confers spermatogenic arrest cells their differentiating ability in vitro in the long-term culture up to 12 days. Therefore, recovery of mature spermatids or spermatoza from arrested germ cells could be greatly facilitated by an introduction of a new culture system that provides an in vitro environment in which immature germ cells progress into mature spermatids or beyond. However, it remains to be seen whether the mature sperms obtained from the present 3-D in vitro culture lead to a successful pregnancy. Consistent with previous results, the patients with normal serum FSH levels show more successful differentiation of immature spermatocytes compared to the patients with higher serum FSH levels. It is probably important to evaluate the hormone profile of the patients before attempting in vitro culture of spermatocytes for infertility patients with spermatogenic arrest. The model could be used as a basis for clinical application of in vitro spermatogenesis for male infertility treatment. Acknowledgments: The authors thank Dr. Lee Seung Ho for his help with statistical analysis. 832 Lee et al. Three-dimensional culture of spermatogenic cells Vol. 87, No. 4, April 2007

REFERENCES 1. Tesarik J, Bahceci M, Ozcan C, Greco E, Mendoza C. Restoration of fertility by in-vitro spermatogenesis. Lancet 1999;353:555 6. 2. Staub C. A century of research on mammalian male germ cell meiotic differentiation in vitro. J Androl 2001;22:911 26. 3. Parks JE, Lee DR, Huang S, Kaproth MT. Prospects for spermatogenesis in vitro. Theriogenology 2003;59:73 86. 4. Cremades N, Bernabeu R, Barros A, Sousa M. In-vitro maturation of round spermatids using co-culture on Vero cells. Hum Reprod 1999; 14:1287 93. 5. Hue D, Staub C, Perrard-Sapori MH, Weiss M, Nicolle JC, Vigier M, et al. Meiotic differentiation of germinal cells in three-week cultures of whole cell population from rat seminiferous tubules. Biol Reprod 1998;59:379 87. 6. Staub C, Hue D, Nicolle JC, Perrard-Sapori MH, Segretain D, Durand P. The whole meiotic process can occur in vitro in untransformed rat spermatogenic cells. Exp Cell Res 2000;260:85 95. 7. Sousa M, Cremades N, Alves C, Silva J, Barros A. Developmental potential of human spermatogenic cells co-cultured with Sertoli cells. Hum Reprod 2002;17:161 72. 8. Tanaka A, Nagayoshi M, Awata S, Mawatari Y, Tanaka I, Kusunoki H. Completion of meiosis in human primary spermatocytes through in vitro coculture with Vero cells. Fertil Steril 2003;79 Suppl 1:795 801. 9. Tesarik J, Guido M, Mendoza C, Greco E. Human spermatogenesis in vitro: respective effects of follicle-stimulating hormone and testosterone on meiosis, spermiogenesis, and Sertoli cell apoptosis. J Clin Endocrinol Metab 1998;83:4467 73. 10. Tesarik J, Greco E, Rienzi L, Ubaldi F, Guido M, Cohen-Bacrie P, et al. Differentiation of spermatogenic cells during in-vitro culture of testicular biopsy samples from patients with obstructive azoospermia: effect of recombinant follicle stimulating hormone. Hum Reprod 1998; 10:2772 81. 11. Tesarik J, Nagy P, Abdelmassih R, Greco E, Mendoza C. Pharmacological concentrations of follicle-stimulating hormone and testosterone improve the efficacy of in vitro germ cell differentiation in men with maturation arrest. Fertil Steril 2002;77:245 51. 12. Tesarik J. Overcoming maturation arrest by in vitro spermatogenesis: search for the optimal culture system. Fertil Steril 2004;81:1417 9. 13. Weaver VM, Petersen OW, Wang F, Larabell CA, Briand P, Damsky C, et al. Reversion of the malignant phenotype of human breast cells in three-dimensional culture and in vivo by integrin blocking antibodies. J Cell Biol 1997;137:231 45. 14. Zhang S. Beyond the petri dish. Nature Biotech 2004;22:151 2. 15. Lee DR, Kaproth MT, Parks JE. In vitro production of haploid germ cells from fresh or frozen-thawed testicular cells of neonatal bulls. Biol Reprod 2001;65:873 8. 16. Saunders PT. Germ cell-somatic cell interactions during spermatogenesis. Reprod Suppl 2003;61:91 101. 17. Cukierman E, Pankov R, Stevens DR, Yamada KM. Taking cell matrix adhesion to the third dimension. Science 2001;294:1708 12. 18. Ito R, Abe SI. FSH-initiated differentiation of newt spermatogonia to primary spermatocytes in germ-somatic cell reaggregates cultured within a collagen matrix. Int J Dev Biol 1999;43:111 6. 19. Rosso F, Giordano A, Barbarisi M, Barbarisi A. From cell ECM interactions to tissue engineering. J Cell Physiol 2004;199:174 80. 20. Krishnamurthy H, Weinbauer GF, Aslam H, Yeung CH, Eberhard N. Quantification of apoptotic testicular germ cells in normal methoxyacetic acid-treated mice as determined by flow cytometry. J Androl 1998; 19:710 7. 21. Siu MK, Cheng CY. Dynamic cross-talk between cells and the extracellular matrix in the testis. Bioessays 2004;26:978 92. 22. Hadley MA, Byers SW, Suarez-Quian CA, Kleinman HK, Dym M. Extracellular matrix regulates Sertoli cell differentiation, testicular cord formation, and germ cell development in vitro. J Cell Biol 1985;101: 1511 22. 23. Skinner MK. Cell cell interaction in the testis. Endocrinol Rev 1991; 12:45 77. 24. Dym M. Basement membrane regulation of Sertoli cells. Endocrinol Rev 1994;15:102 15. 25. Griswold MD. The central role of Sertoli cells in spermatogenesis. Semin Cell Dev Biol 1998;9:411 6. 26. Sofikitis N, Pappas E, Kawatani A, Baltogiannis D, Loutradis D, Kanakas N, et al. Efforts to create an artificial testis: culture systems of male germ cells under biochemical conditions resembling the seminiferous tubular biochemical environment. Hum Reprod 2005;11:229 59. 27. Grootegoed JA, Jutte NHPM, Rommerts FFG, Roblero L, Miller IF, Zaneveld LJ. First pregnancies and live births from transfer of sodium alginate encapsulated embryos in rodent model. Fertil Steril 1993;59: 652 6. 28. Tesarik J, Mendoza C, Greco E. In-vitro maturation of immature human male germ cells. Mol Cell Endocrinol 2000;166:45 50. Fertility and Sterility 833