ARTICLES Generation of pluripotent stem cells from adult human testis Sabine Conrad1, Markus Renninger3, Jorg �� Hennenlotter3, Tina Wiesner1, Lothar Just1, Michael Bonin4, Wilhelm Aicher5,6, Hans-Jorg �� Buhring7, �� Ulrich Mattheus2, Andreas Mack2, Hans-Joachim Wagner2, Stephen Minger8, Matthias Matzkies9, Michael Reppel9, Jurgen �� Hescheler9, Karl-Dietrich Sievert3, Arnulf Stenzl3 & Thomas Skutella1,6 Human primordial germ cells and mouse neonatal and adult germline stem cells are pluripotent and show similar properties to embryonic stem cells. Here we report the successful establishment of human adult germline stem cells derived from spermatogonial cells of adult human testis. Cellular and molecular characterization of these cells revealed many similarities to human embryonic stem cells, and the germline stem cells produced teratomas after transplantation into immunodeficient mice. The human adult germline stem cells differentiated into various types of somatic cells of all three germ layers when grown under conditions used to induce the differentiation of human embryonic stem cells. We conclude that the generation of human adult germline stem cells from testicular biopsies may provide simple and non-controversial access to individual cell-based therapy without the ethical and immunological problems associated with human embryonic stem cells. Theability toderivepluripotentstem cellsfrom theadulthuman testis has important implications for biotechnology and regenerative medi- cine1. Although these cells are unipotently restricted to the generation of gametes in the course of normal development2,3, several lines of evidence suggest that under certain circumstances, cells of the germ line havetheability togiverise tocellsthat are pluripotent4���6. Theterm of pluripotency is differently defined in research with mouse and human stem cells. The NIH and the ISSCR guidelines and criteria for human pluripotency include teratoma formation in addition to microarray assays for transcription factors and other gene activity associated with pluripotency. Teratomas, which are tumours contain- ing different kinds of cells and tissues from all three germ layers at various stages of maturation, occur almost exclusively in the gonads7. Furthermore, primordial germ cells (PGCs) give rise to pluripotent cells when cultured under appropriate conditions4,8. PGCs have dif- ferentiation properties similar to those of embryonic stem (ES) cells isolated from the inner cell mass9. Recently the successful establish- ment of germline stem cells from neonatal mouse testis was reported5. In addition, one study6 successfully generated mouse adult germline stem cells (GSCs) with pluripotency from spermatogonial stem cells from adult mouse testis. As in the experiments reported previously5, these cells were able to differentiate into derivatives of all germ layers in vitro, generated teratomas in immunodeficient mice and, when injected into an early blastocyst, contributed to the development of various organs. Similar results with GPR1251 germline progenitor cells have been reported by another study10. Generation of pluripotent human adult GSCs Weusedintotal22differenthumantesticularparenchymasto generate human adult GSCs. The obtained tissues were mechanically and enzy- matically dissociated and filtered to obtain a single-cell suspension containing cells of varying sizes and shapes (Supplementary Fig. 1a). Ina nextstepthesingle cells werecultured for 4 days in uncoated dishes with knockout culture medium with glia-derived neurotrophic factor (GDNF), a growth factor essential for the self-renewing division of spermatogonial stem cells11, or cultured directly in leukaemia inhib- itory factor (LIF ref. 12)-supplemented medium (basic medium), which is sufficient to maintain mouse ES cells or embryonic germ cells13,14 in an undifferentiated state (Supplementary Fig. 1a). Under these conditions most of the single testis cells attached to the culture plate. For pre-selection of spermatogonial cells with magnetic- activated cell separation (MACS) we used CD49f (a6 integrin)15,16, a marker selected by us from the different tested surface antigens. By using other antibodies like CD90 (Thy-1) or GDNFR-a1, which have been described in the literature for mouse spermatogonial stem cell enrichment (see for example refs 17, 18), or CD133 (a marker for human ES and precursor cells) we achieved comparable but not better selection (data not shown). An important tool to gain a highly pure spermatogonial cell population is the subsequent matrix selection pro- cedure with collagen and laminin19 to extract spermatogonial cells for furthercultivationwithbasicmediumandLIFtogeneratehumanadult GSCs. With this procedure we were able to obtain a pure population of spermatogonial cells (VASA1 vimentin2) and completely deplete so- matic cells (VASA2 vimentin1) (Fig. 1). After this selection and puri- fication, colonies of spermatogonial cells appeared (Fig. 2a, panel 1) and increased in size (Fig. 2a, panel 2). After 10���15 days, these colonies changed their morphology (Fig. 2a, panel 3), became multilayered and clearly demarcated colonies with boundaries appeared (Fig. 2a, panel 4). These colonies continued to increase in number and size (Fig. 2a, far right panel). Functional proof is provided by the fact that the negative fraction of somatic cells (VASA2 vimentin1) did not form stem cell colonies under LIF supplementation and were negative for stem cells markers, and even more importantly did not form any ter- atomas (Supplementary Fig. 1). In contrast, even after long-term 1 Institute of Anatomy, Department of Experimental Embryology, 2 Institute of Anatomy, Department of Cellular Neurobiology, Osterbergstra��e �� 3, 72074 Tu ��bingen, Germany. 3 Department of Urology, University Clinic Tu ��bingen, Hoppe-Seyler-Stra��e 3, Tu ��bingen 72076, Germany. 4 Institute of Anthropology and Human Genetics, Microarray Facility, University Clinic, Calwerstra��e 7, 72076 Tu ��bingen, Germany. 5ZMF Research Laboratories, University Clinic Tu ��bingen, Waldhornlestra��e �� 22, 72072 Tu ��bingen, Germany. 6Center for Regenerative Biology and Medicine (ZRM), Paul-Ehrlich-Stra��e 15, 72076 Tu ��bingen, Germany. 7Department of Internal Medicine II, University Clinic Tu ��bingen, Otfried-Mu��ller-Stra��e 10, 72076 Tu ��bingen, Germany. 8Stem Cell Biology Laboratory, Wolfson Centre for Age-Related Diseases, King���s College London, King���s College, London SE1 1UL, UK. 9Institute of Neurophysiology, University of Cologne, Robert-Koch-Stra��e 39, 50931 Cologne, Germany. doi:10.1038/nature07404 1 ��2008 Macmillan Publishers Limited. All rights reserved

cultivation the VASA1 vimentin2 human adult GSC colonies behaved more like human ES cells in their molecular profile and differentiation capacity, and formation of teratomas. Electron microscopy of purified spermatogonial cells showed typi- cal morphological characteristics of spermatogonia (Supplementary Fig. 1b). A more detailed immunohistochemical characterization revealed that these cells were positive for VASA, SSEA4, OCT4, TSPYL2, DAZL, CD133 and CD49f, but negative for NANOG and e-cadherin and the somatic marker vimentin (Figs 1, 3a and Supplementary Fig. 1a). Generally the cells were passaged 3 to 4 weeks after initiation of the culture. After passaging the multilayered colonies reappeared with constant doubling times up to high passages (over 40 and higher). We also tested different medium conditions for their ability to induce the generation of human adult GSCs (Fig. 2d). To do this, we plated spermatogonial cells for 14 days under different conditions and observed the formed colonies over another 42-day period (Fig. 2d). Without supplementation with GDNF and LIF no clusters were formed at all. In contrast, both LIF alone and GDNF followed by LIF resulted in a constant rate of cluster formation over 42 days with DAPI/e-cadherin DAPI/NANOG Neg. control DAPI/CD49f DAPI/ CD49f/ VASA DAPI/CD133 DAPI/SSEA4 DAPI/TSPYL2 DAPI/OCT4 DAPI/DAZL a b c d e f g h i Figure 1 | Selection of spermatogonial cells from adult human testis. a���i, Immunohistochemistry with the germ and stem cell markers NANOG, SSEA4, OCT4, e-cadherin, TSPYL2, DAZL, CD133 and the selection marker CD49f in spermatogonial cells. Co-expression of CD49f and VASA in spermatogonial cells is shown. Nuclei were stained with 4,6-diamidino-2- phenylindole (DAPI) and magnifications (original magnification 340) show nuclear or cytoplasmic staining. haG SC neural dif f. c Human ES cells (H1) haGSC P7+LIFhaGSC haGSC P3+LIF P14 +LIF pSTAT3 STAT3 80 kDa 81 kDa Spermatogonial cells 1 4 haGSC P16 P0 haGSC 3 2 Spermatogonial cluster a Total cell number Human ES cells (H1) haGSC P3 700 600 500 400 300 200 100 0 haGSC P36 0 24 48 72 96 120 Hours e 7 14 21 28 35 42 Days 0 Clusters (%) 0 20 40 60 80 100 d LIF GDNF/LIF GDNF Without GDNF or LIF GDNF/FGF2 LIF/FGF2 Hum an ES cell s (H1 ) Wat er haGSC P1+ GDNF/LIF haGSC P1+LIF haG SC P7 + GDNF/LIF haGSC P7 +LIF gp130 LIF-R b Figure 2 | Generation of human adult GSCs from spermatogonial cells. a, Panel 1 shows colonies of spermatogonial cells on laminin (LamB cells). Panel 2 shows a proliferating spermatogonial cell colony. Panel 3 shows an early human adult GSC cluster (haGSC P0). A human adult GSC cluster at higher passage (P16 panel 4) and an overview of typical human adult GSC clusters (far right panel) is also shown. b, RT���PCR of LIF receptor complex in human adult GSCs cultured under GDNF/LIF or LIF alone from lower (P1) and higher (P7) passage in comparison to FGF2-cultivated H1 cells. A 1-kilobase DNA ladder is shown. c, Western blot analysis of human adult GSCs under LIF from passage 3, 7 and 14, H1 cells and neural differentiated human adult GSCs. d, Percentage of formed clusters after plating spermatogonial cells for 14 days under different growth factors over a further cultivation time period of 42 days expressed. e, Comparison of the doubling times of human ES cells (H1) and human adult GSCs from passage 3 and 36 over a 120-h period. Error bars in d and e show standard deviations (n 5 3). ARTICLES NATURE 2 ��2008 Macmillan Publishers Limited. All rights reserved