The Background of the Invention Spermatogenesis is a complex sequence of events involving both hormonal regulators and local interactions between Sertoli and germ cells (Bellve, A. R. & Zhang, W. J. , Reprod. Fertil. 85, 771-793 (1989) ; Skinner, M. K. , Endocr. Rev. 12,45-72 (1991) ; Pescovitz, 0. H. , et al. , Trends Endocrinol. Metab. 5,126-131 (1994). Parvinen, M. , et al. , J. Cell Biol. 117,629-41 (1992), Djakiew, D. , et al., Biol. Reprod. 51,214-21 (1994), Cancilla, B. & Risbridger, G. P. , Biol. Reprod. 58,1138-45 (1998) ; Bitgood, M. J. , et al. , Curr. Biol. 6,298-304 (1996) ; Zhao, G. Q. , et al. , Genes Dev. 10,1657-1669 (1996) ; Zhao, G. Q. , et al. , Development 125, 1103-1112 (1998) ). The diffe- rentiation of sperm cells or spermatogenesis can be divided in three phases. In the first phase of the spermatogenesis undifferentiated stem cells from the germ line, the spermato- gonia, are divided mitotically and differentiate into sper- matocytes. Spermatogonia are very resistant against different types of damages and after damage spermatogenesis starts from said cells. At present, it is not known by which mechanism the spermatogonia are maintained and secondly by which signals they are differentiated into spermatocytes. Spermatogonia are located in the peripheral parts of the seminiferous tubules and are in contact with the Sertoli cells, which represent the regulating cells of the seminiferous tubules and are classi- fied as somatic cells of the organism.

In the next phase the spermatogonia are at first differenti- ated into primary and thereafter into secondary spermatocytes, whereby their chromosomes become haploid by two meiotic cell divisions after which they are called spermatides. In the third phase the spermatides mature into spermazoa or mature sperm cells.

The luteinizing hormone, the follicle stimulating hormone and testosterone are the most important hormonal regulators of spermatogenesis. The cells of the germ line in the semini- ferous tubules of testis are not direct targets of the hor- monal regulation but the hormonal regulation is transmitted by the Sertoli cells. They modulate the hormonal signals into paracrine signals which are believed to be signal molecules and growth factors, such as stem cell factor (c-kit ligand), insulin like growth factor-1 and three members of the trans- forming growth factor-S family (transforming growth factor-S, activin and inhibin) (Bellve A. R. & Zhang W. J. , Reprod.

Fertil. 85,771-793 (1989) ; Skinner, M. K. , Endocr. Rev. 12, 45-72 (1991) ; Pescovitz, O. H. , et al., Trends Endocrinol.

Metab. 5,126-131 (1994) ). All said ligands are expressed by Sertoli cells and their receptors are located in the cells of the germ line. Functional proof of said regulating system is scare. The probably best experimental evidence of the para- crine regulation of spermatogenesis is obtained with desert hedgehog-deficient transgenic mice (Bitgood, M. J. , et al., Curr. Biol. 6,298-304 (1996) ). Desert hedgehog is a signal molecule, which is expressed in the Sertoli cells. The patched-receptor thereof is located in the Leydig cells, i. e. the testosterone producing interstitial cells of the testis.

When the desert hedgehog molecule is missing the differentia- tion of sperm stops, sperm cells are not formed and the male mice are all sterile.

The spermatogenetic cells can also send the signals back to the Sertoli cells (Bellve, A. R. Zhang W. J. , Reprod.

Fertil. 85,771-793 (1989) ; Skinner, M. K. , Endocr. Rev. 12, 45-72 (1991) ; Pescovitz, O. H. , et al. , Trends Endocrinol.

Metab. 5,126-131 (1994) ). This is indicated for example by the fact that nerve growth factor and fibroblast growth factor-1 are expressed in the cells of the germ line and the p75 nerve growth factor receptor and fibroblast growth factor receptor are located in the Sertoli cells. Hints of auto- crine regulation of spermatogenesis also exist. Bone morphoge- netic protein 8a and 8b and their receptors are expressed in the same spermatogenetic cells (Zhao, G. Q. , et al. , Gene Dev.

10, 1657-1669 (1996) ; Zhao, G. Q. , et al. , Dev. 125,1103-1112 (1998) ). The functional evidence for the autocrine regulation is still lacking.

The glial cell line-derived neurotrophic factor (GDNF) is a member of the transforming growth factor- superfamily (Lin, L. F. , et al., Science 260,1130-1132 (1993) ). GDNF maintains the dopaminergic, noradrenergic, cholinergic and motor neurons in the nervous system and also protects peripheral parasympa- thetic ciliary and sensory nerve cells (Lin, L. F., Neural Notes. 2,3-7, (1996). During the development of mammals, GDNF is expressed also in other tissues, such as embryonic kidney and testis (Suvanto, P. , et al. , European Journal of Neuro- science 8,816-822, (1996) ; Trupp M, et al. , Journal of Cell Biology 130,137-148, (1996) ). The role of GDNF as a regula- tor of the differentiation of the kidney has been shown with gene knockout or gene deletion techniques in mice (Pichel, J. G., et al., Nature 382,73-75 (1996). Because said mice die as newborn, it has not been possible to study with said deletion gene model - or other experimental models - the function of GDNF in spermatogenesis, which is initiated much later, only after birth.

Studies with different model systems indicate that there is a multitude of ways for interfering with spermatogenesis and for developing male contraceptives. So far the methods suggested have been very rough with severe side effects for the male subject. Thus, the problem yet to be solved is how to select from the multitude of hormonal growth regulators one which can sufficiently specifically be targeted to the testis and at the same time provides a system for regulating spermato- genesis without disturbing side-effects and which system does not interfere negatively with the male mating activity.

The problem was solved by the present inventors when studying transgenic mice with human GDNF incorporated into their testis. It was shown that mice with an elevated level of human GDNF were capable of mating but not capable of fertilizing the female. Thus, the solution to the problem is to use glial cell line-derived neurotrophic factor family-related compounds as defined in the claims of the present invention for regulating spermatogenesis and inhibiting or preventing spermatogenesis followed by infertility and as active in- gredients for manufacturing compositions useful as male contraceptives, which are both effective and without severe side-effects.

Thus, the first objective of the present invention is to provide an effective male contraceptive without severe, undesired side-effects by using GDNF-family-related compounds as male contraceptives.

Another objective of the present invention is to use the GDNF-family-related compounds as active ingredients for manufacturing compositions useful as male contraceptives.

One other objective of the present invention is to provide tools for regulating and studying spermatogenesis. This is facilitated by inhibiting the differentiation of sperm cells. A third objective of the present invention is to provide tools, which have an effect on infertility caused by disturbed differentiation of sperm cells.

The objective of the present invention is also to provide tools for regulating and studying spermatogenesis as well as for inhibiting the differentiation of sperm cells.

Further the objective of the present invention is to provide transgenic animals, especially mice, which due to the testis targeted GDNF-family-related compounds provide a useful tool for studying spermatogenesis.

The Summary of the Invention The characterstics of the present invention are as defined in the claims. The present invention describes a method, whereby it is possible to study sterility caused by disturbed differen- tiation of sperm cells and secondly it provides a system for inhibiting the differentiation of sperm cells in a contra- ceptive sense, whereby the male (man) is sterile (infertile). The GDNF-transgene carrying transgenic mouse strain, the GDNF t-mouse strain, is a model for infertility caused by distur- bances in the differentiation of sperm cells. It is character- ized in that GDNF, compounds acting like, said GDNF, another cRet receptor, another GDNF-receptor activating compound or a compound, which activates the cRet receptor signal transmit- ting tubules in spermatogonia, are suitable as a male contra- ceptive and it is also characterized in that the GDNF t-mouse model is suitable as an animal model for studying the regula- tion of spermatogenesis and spermatogenesis.

A Short Description of the Drawings Figure la. Testicular morphology in wild-type mice at 3-weeks of age. Scale bar 100 m. Figure lb. Testicular morphology in transgenic mice at 3-weeks of age. Scale bar 100 m. Note the undifferentiated spermatogo- nia formed cell cluster.

Figure lc. Testicular morphology in wild-type mice at 8-weeks of age. Scale bar 100 m.

Figure ld. Testicular morphology in transgenic mice at 8-weeks of age. Note in d the remnants of type A spermatogonia clusters that are abundant in the young mice. Scale bar 100 Am.

Figure le. Testicular morphology in wild-type mice at 6-months of age. Scale bar 100 m.

Figure If. Testicular morphology in transgenic mice at 6- months of age. Note the advanced germ cell atrophy and Leydig cell hyperplasia (star). Scale bar 100 gm.

Figure Ig. Testicular morphology in wild-type epididymal ducts at 8-weeks. Scale bar 100 m.

Figure lh. Testicular morphology in transgenic epididymal ducts at 8-weeks. Note the lack of sperm. Scale bar 100m.

Figure 2a. Northern blotting for GDNF mRNAs in wild-type mice at various ages. GDNF levels are strongly downregulated after the first or second postnatal week.

Figure 2b. Northern blotting for Ret mRNAs in wild-type mice testes at various ages. GDNF mRNA levels are strongly down- regulated after the first or second postnatal week.

Figure 2c. Northern blotting for GFRa1 mRNAs in wild-type mice at various ages. GFRal mRNA levels are strongly downregulated after the first or second postnatal week. Figure 2d. Northern blotting for GDNF mRNAs in transgenic testes at various ages. The hGDNF levels remain high after the first or second postnatal week.

Figure 2e. Northern blotting for Ret mRNAs in transgenic mouse testis at various ages.

Figure 2f. Northern blotting for GFRa1 mRNAs in transgenic testes at various ages. GFRa1 mRNA levels are high in the transgenic mice at all stages analysed.

Figure 2g depicts immunoprecipitation of GDNF from wild-type (WT) and transgenic (TG) testes at 3 and 6 weeks of age. In wild-type testes, the GDNF protein is not detectable. The control, 25 ng of recombinant human GDNF protein is shown on the right side.

Figure 3a. The distribution of GDNF transcripts in testes of wild-type (WT) mice by cRNA in situ hybridisation. Sense controls did not show grain density above the background. 1- week-old wild-type testis. Scale bar 100 m.

Figure 3b. The distribution of GDNF transcripts in 8-weeks-old testis of wild-type (WT) mouse by cRNA in situ hybridisation.

Sense controls did not show grain density above the back- ground. Scale bar 100 m.

Figure 3c. The distribution of GDNF transcripts in testes of transgenic mice (TG) by cRNA in situ hybridisation. Sense controls did not show grain density above the background. 8-weeks-old transgenic testis. Note that the spermatogonia in the transgenic mice continuously express human GDNF at high levels, but the number of these cells is clearly diminishing at 8 weeks of age. Scale bar 100 m.

Figure 3d. The distribution of Ret transcripts in testes of wild-type (WT) mice by cRNA in situ hybridisation. 2-weeks-old wild-type testis. At two weeks of age, Ret expression is localised to some spermatogonia in the basal layer of the seminiferous tubules (arrows).

Figure 3e. The distribution of Ret transcripts in testes of wild-type (WT) mice by cRNA in situ hybridisation. 8-weeks-old wild-type testis. At 8 weeks of age, Ret is expressed in both some spermatogonia and spermatides.

Figure 3f. The distribution of Ret transcripts in testes of transgenic mice (TG) by cRNA in situ hybridisation. 8-weeks- old transgenic testis. Note that the spermatogonia in the transgenic mice continuously express Ret at high levels, but the number of these cells is clearly diminishing at 8 weeks of age. In transgenic mice, Ret is expressed in the clusters of spermatogonia. Scale bar 100 m.

Figure 3g. The distribution of GFRa1 transcripts in testes of wild-type mice (WT). 1- to 2-weeks-old wild-type testis. GFRa1 is expressed by some spermatogonia, morphologically similar to those with Ret (arrows).

Figure 3h. The distribution of GFRal transcripts in testes of wild-type by cRNA in situ hybridisation. 8-weeks-old wild- type testis. GFRa1 mRNA distribution follows that of Ret, but is also highly expressed by spermatocytes. Scale bar 100 m.

Figure 3i. The distribution of GFRa1 transcripts in testes of transgenic mice (TG) by cRNA in situ hybridisation. 8-weeks- old transgenic testis. Note that the spermatogonia in the transgenic mice continuously express human GDNF at high levels, but the number of these cells is clearly diminishing at 8 weeks of age. In transgenic mice, GFRal is expressed in the clusters of spermatogonia. Scale bar 100 m.

Figure 4a. Cell proliferation in wild-type mice testes. In 3-weeks-old testis, BrdU labelling of S-phase cells is seen only in the periphery of wild-type seminiferous tubules. Note the absence of labelling in some transsections of tubules, reflecting the segmented distribution of S-phase cells during spermatogenesis (stars). Scale bar 100 m.

Figure 4b. Cell proliferation in transgenic mice testes. In a 3-week-old transgenic mouse testis, many BrdU-labelled cells are seen in the luminal area of the seminiferous tubules, reflecting the abnormal distribution of spermatogonia (arrow).

The segmental distribution of S-phase cells is not seen in the transgenic tubules. Scale bar 100 m.

Figure 4c. Distribution of cell line proliferation indexes in wild type and transgenic testes. The number of BrdU labeled cells/100 spermatogonia was counted in 100 transections of wild type and transgenic seminiferous tubules.

Figure 4d. Apoptosis in wild-type mice testes. TUNEL-labelling for apoptosis in 4-week-old mice testes. Only a few apoptotic cells are seen. Scale bar (c) 100 m.

Figure 4e. Apoptosis in transgenic mouse testis. TUNEL-label- ling for apoptosis in 4-week-old testes. Note the increase of labelling in spermatogonia clusters in (Figure 4e), as com- pared to wild-type (Figure 4d) seminiferous tubules. Scale bar 100 ym.

Figure 4f. Live morphology in both a wild-type and transgenic seminiferous tubule. Live preparation of seminiferous tubules from 15-week-old wild-type (WT) and transgenic (TG) mice. Note the reduced diameter of the transgenic tubule and the increase of Leydig cells (arrows) around it. Scale bar 200 m.

Figure 4g. Live morphology in transgenic testes. High magni- fication of living cells from a transgenic seminiferous tubule in an unfixed squash preparation showing large numbers of dead cells (arrow), some of which resemble type A spermatogonia, in a Sertoli cell (arrow head). Scale bar 10 m.

The Detailed Description of the Invention Definitions In the present invention the terms used have the meaning they generally have in the fields of biochemistry, pharmacology, recombinant DNA technology, including transgenic animal production, but some terms are used with a somewhat deviating or broader meaning than in the normal context. Accordingly, in order to avoid uncertainty caused by terms with unclear meaning some of the terms used in this specification and in the claims are defined in more detail below.

The term"glial cell line-derived neurotrophic factor (GDNF) family-related compounds"i. e. GDNF-family-related compounds comprises glial cell line-derived neurotrophic factor (GDNF), neurturin, persephin, artemin and other similar growth fac- tors, but also other compounds which have the same effect as GDNF-like compounds on the GDNF-receptor (s) or co-receptors.

Said receptor (s) or co-receptor (s) comprises Ret tyrosine kinase and GDNF-family-receptor a : s, respectively.

The term"glial cell line-derived neurotrophic factor (GDNF)" means the glial cell line-derived neurotrophic factor (GDNF), which is a member of the transforming growth factor-6 family (Lin, L. F. , et al. , Science 260,1130-1132 (1993). GDNF maintains the dopaminergic, noradrenergic, cholinergic and motor neurons in the nervous system and also protects peri- pheral, parasympathetic, ciliary and sensory nerve cells (Lin, L. F., Neural Notes. 2,3-7 (1996) ). During the development of mammals, GDNF is expressed also in other tissues, such as embryonic kidney and testis (Suvanto P, et al. , European Journal of Neuroscience 8,816-822, (1996) ; Trupp M, et al., Journal of Cell Biology 130,137-148, (1996) ). The properties and structure, amino acid sequence is disclosed by Lin, L. F., et al. (Science 260,1130-1132 (1993). Said protein is coded by a gene characterized by the GDNF cDNA sequence deposited as Genebank accession number L15306.

The term"GDNF-like factors"or GDNF-like compounds mean growth factors which have a similar structure or act like GDNF on its receptor or co-receptors. The term"GDNF-like factor" or "GDNF-like compound" above all relates to three GDNF-like growth factors, persephin, artemin and neurturin, which are described in (Milbrandt, J. , et al. , Neuron 20,1-20 (1998) ; Kotzbauer P. T. , et al., Nature 384,467-470 (1996) ).

All four growth factors of the GDNF-family are structurally similar and cRet receptor tyrosine kinase (Durbec, P. , et al. , Nature 381,789-793 (1996) ; Buj-Bello, A. , et al., Nature. 387,721-724, (1997) ) acts as the common signal transmitting receptor of GDNF, artemin, persephin and neur- turin.

The"compounds which act like said GDNF"means compounds which act as GDNF on the receptor transmitting signals of GDNF or co-receptors thereof.

The term"receptor"means GDNF-family binding substances, i. e. substances or compounds belonging to the GDNF-family-related compounds which are capable of transmitting the signals of GDNF-family-related compounds. The term"receptor"above all means the GDNF-family signal mediatory receptor is cRet receptor tyrosine kinase.

The term"co-receptors"means receptors, which do not transmit the signals of GDNF-family related compounds, but activate the receptor transmitting the signals of the GDNF-family-related compounds. Such compounds are above all the GDNF-family receptor a : s (GFRa), which is a new receptor class, the members of which are called GDNF-family-receptor a:s (GFRa).

These also belong to the GDNF, persephin, artemin and neur- turin signalling receptor complex. At present four members (GFRa1-4) (Jing, S. , et al. , Cell 85,9-10 (1996), Jing, S.

Q. , et al. , J. Biol. Chem. 272,33111-3311 7 (1997) ; Treanor, J. J. , et al. , Nature. 382,80-83, (1996) ; Suvanto, P. , et al., Human Molecular Genetics. 6,1267-1273, (1997) ) of said family are known. They may transmit signals independently, but they are essential for ligand binding and cRet activation.

The term"concentration of GDNF-family-related compounds in such amounts that the differentiation of sperm cells is prevented, spermatogenesis can be regulated and infertility caused by disturbed differentiation of sperm cells or that the male is rendered unable to fertilize the female"means a concentration of GDNF-family-related compounds, which is approximately 50 - 5000, preferably 50 - 500, most preferably 50 - 100 times higher than the concentration of endogenous GDNF in the testis of said male. Said amount can be calculat- ed from the test results obtained when measuring the GDNF mRNA levels of transgenic mice shown in the Northern blots (Figures 2) and which indicate an elevated amount of mRNA from which the expressed protein levels can be calculated. From these results it can be concluded that the amount of GDNF in trans- genic mice is 50-500 times higher than the endogenic level in three to six week old mice, but only a little bit higher than in new born mice, when the GDNF levels are at their highest and no spermatogenesis is presented. From these results it is possible for those skilled in the art to calculate the amounts needed for obtaining the desired effects in man.

The term"derivatives"means isomers of GDNF-family-related compounds. It also includes hybrid molecules constructed based on the structures of the compounds GDNF, persephin, neurturin and artemin, as well as truncated forms, complexes and fragments of said compounds. The only prerequisite for said compounds is that they still have substantially the same properties and effects as the compounds listed above, deter- mined by the same methods as described in the present inven- tion. Naturally, the term"derivatives"includes derivatives of the compounds listed above in a more conventional sense, e. g. the free groups in the listed compounds, such as car- boxy-, amino- and/or hydroxy- groups may be substituted by alkylation, esterification, etherification and amidation with e. g. alkyl groups, including methyl, ethyl, propyl, etc. The only prerequisite in this case, too is that the substitutions do not substantially alter the properties or effectivity of said molecules or compounds listed above in this paragraph.

General Description of the Invention Glial cell line-derived neurotrophic factor (GDNF) is a potent survival factor for various sets of neuronal cells, it regu- lates kidney differentiation (Lin, L. F. , et al. , Science 260, 1130-1132 (1993) ; Tomac, A., et al. , Nature 373,335-339 (1995) ; Winkler, C. , et al. , J. Neurosci. 16,7206-7215 (1996), Oppenheim, R. W. , et al. , Nature 373,344-346 (1995), Nguyen, Q. T. , et al., Science 279,1725-1729 (1998) ; Trupp, M. , et al. , J. Cell Biol. 130,137-148 (1995) ; Hellmich, H.

L. , et al. , Mech. Dev. 54,95-105 (1996) ; Schuchardt, A. , et al. , Nature 367,380-383 (1994) ; Pichel, J. G. , et al., Nature 382,73-75 (1996) and is expressed in testis (Trupp, M. , et al. , J. Cell Biol. 130,137-148 (1995) ; Hellmich, H.

L. , et al. , Mech. Dev. 54,95-105 (1996) ; Suvanto, P. , et al. , European Journal of Neuroscience 8,816-822 (1996) ) where its function, however, remains unknown.

GDNF, a distant member of the transforming growth factor- (TGF-S) superfamily (Lin, L. F. , et al. , Science 260,1130- 1132 (1993) ) is a potent survival factor for dopaminergic neurones of the substantia nigra, and some other sets of neurones in the central and peripheral nervous systems (Lin, L. F. , et al. , Science 260,1130-1132 (1993) ; Tomac, A. , et al. , Nature 373,335-339 (1995) ; Winkler, C. , et al. , J.

Neurosci. 16,7206-7215 (1996), Oppenheim, R. W. , et al., Nature 373,344-346 (1995), Nguyen, Q. T. , et al. , Science 279,1725-1729 (1998) ; Trupp, M. , et al., J. Cell Biol. 130, 137-148 (1995) ; Hellmich, H. L. , et al. , Mech. Dev. 54, 95-105 (1996) ; Schuchardt, A. , et al. , Nature 367,380-383 (1994) ; Pichel, J. G. , et al. , Nature 382,73-75 (1996) ).

GNDF also acts as a signalling molecule in ureteric branching during kidney morphogenesis (Schuchardt, A. , et al., Nature 367,380-383 (1994) ; Pichel, J. G. , et al. , Nature 382, 73-75 (1996) ). It is synthesised by several neuronal and non-neu- ronal embryonic organs including the embryonic and postnatal testis (Trupp, M. , et al. , J. Cell Biol. 130,137-148 (1995) ; Hellmich, H. L. , et al. , Mech. Dev. 54,95-105 (1996) ; Suvan- to, P. , et al. , European Journal of Neuroscience 8,816 - 822 (1996) ) The signalling receptor-complex for GDNF includes Ret receptor tyrosine kinase (Durbec, P. , et al. , Nature 381, 789-793 (1996) ; Trupp, M. , et al. , Nature 381,785-789 (1996) ) and a member of the glycosylphosphatidylinositol- linked co-receptors, GDNF-family-receptor a:s (GFRa1-4) (Jing, S. , et al. , Cell 85,9-10 (1996) ; Jing, S. Q. , et al. , J.

Biol. Chem. 272,33111-3311 7 (1997) ; Enokido, Y. , et al., Curr. Biol. 18,1019-1022 (1998) ). The GDNF-, GFRal and Ret-null mice exhibit remarkably similar phenotypes with absent or hypodysplastic kidneys and severe defects in enteric innervation (Schuchardt, A. , et al., Nature 367,380-383 (1994) ; Pichel, J. G. , et al. , Nature 382,73-75 (1996) ). As they die during the first postnatal day, spermatogenesis could not be analysed in these mutant mice.

In order to approach the role of GDNF in spermatogenesis, the present inventors studied transgenic mice with targeted overexpression of GDNF in testicular germ cells. The trans- genic males were infertile and showed an early arrest in the differentiation of spermatogenic cells, the type A sperma- togonia. They formed clusters within seminiferous tubules, finally resulting in testicular atrophy. The data obtained showed a novel function for GDNF as a paracrine inhibitory signal during early spermatogenesis. In wild-type mice, GDNF is believed to regulate the pool of spermatogonia that is gradually committed to sperm diffe- rentiation. These infertile transgenic mice serve as a model for male infertility with spermatogenic deficiency. The activation of the GDNF signalling cascade possibly provides us with means to develop contraceptives for man. The effects of GDNF on spermatogenesis should also be taken into conside- ration, when GDNF-like molecules are used in clinical trials of neurodegenerative disorders, such as Parkinson's disease.

When studying transgenic mice overexpressing human GDNF under human translation elongation factor-1 (EF-1) promoter (Mizu- shima, S. & Nagata, S. , Nucleic Acids Res. 18,5322 (1990), which targets the transgene expression specifically to testi- cular germ cells (Furuchi, T. , et al., Development 122, 1703-1709 (1996)), a novel function for GDNF was found. The disturbance of spermatogenesis in transgenic testis, targeted to overexpress GDNF, suggested that GDNF acts as a local inhibitory signal in spermatogonia differentiation. The gradual degeneration of germ cells and the hyperplasia of Leydig cells in the transgenic mice are apparently secondary to the spermatogenic deficiency, because Sertoli cells elimi- nate poorly differentiated and abnormal germ cells, and interplay between the cell types is disrupted. Therefore, GDNF targeting to testis provides a new animal model for male infertility. The activation of the GDNF signalling cascade by GDNF, Ret agonists or downstream substrates possibly provides us with means whereby sperm production in man could be pre- vented and male contraceptives developed. The effects of GDNF on spermatogenesis should also be taken into consideration, when GDNF-like molecules are designed for and used in clinical trials of neurodegenerative disorders, such as Parkinson's disease.

To study the function of GDNF in testis, the inventors gene- rated transgenic mice overexpressing human GDNF under human translation elongation factor-1 (EF-1) promoter (Mizushima, S. & Nagata, S., Nucleic Acids Res. 18,5322 (1990)), which targets the transgene expression specifically to testicular germ cells (Furuchi, T. , et al., Development 122,1703-1709 (1996) ). Other promoters targetting the transgene expression specifically to testicular germ cells can be used as well. Four independent transgenic founders were analysed, two males C10, C12 and two females S6, E19 with transgene copy numbers 10,3, 20, and 4, respectively. The male founders C10 and C12 were infertile and further analysis of the male phenotype was done with the offspring of the two female founders S6 and E19. The transgenic mice developed normally to adulthood and showed a similar phenotype. The body weight was normal, and they did not have any defects in the weights and morphology of paren- chymal organs (data not shown), except in the testis. After four weeks of age, a reduction in weight of testes was ob- served in the transgenic mice as compared'to wild-type con- trols (Table 1). All transgenic males in the offspring of the female founders S6 and E19 were infertile. During six months of continuous breeding, three transgenic males were caged with wild-type FVB and NMRI females in three to four day intervals.

The wild-type females mated with transgenic males at normal frequency, as shown by the appearance of vaginal plugs, but none became pregnant. Further 20 transgenic males at breeding age were subjected to wild-type FVB and NMRI females in short term breeding test. During two weeks of continuing breeding, approximately 200 females from both strains were plugged, but none became pregnant. Strain dependence of the infertile phenotype was excluded by crossbreeding the transgenic FVB mice to NMRI background. Like the transgenic males of the FVB strain, those of the F1 generation were also infertile.

It was shown that the testicular histology of one-week-old transgenic mice was normal. From two to three weeks of age (Fig. la, lb), they showed clusters of cells morphologically similar to type A spermatogonia, partially detached from the basal lamina of the seminiferous tubules. These cells under- went gradual degeneration during subsequent weeks so that by ten weeks the tubules showed already advanced atrophy with Sertoli cells and only spermatogonia (Fig. lc-lf). The Leydig cells, the testicular interstitial cells, showed progressive hyperplasia from the fourth postnatal week (Fig. ld, lf). Occasional germ cells outside the spermatogonial clusters developed further into spermatocytes and rarely to spermatids, but they arrested postmeiotically and degenerated. Mature sperm was not observed in seminiferous tubules or epididymal ducts, although no obstruction was found in the ductuli efferens, which conduct sperm from the testis to the epidi- dymis (Fig. 1g, 1h).

It was also shown that the levels of mouse and human GDNF and its receptor mRNAs were elevated in the transgenic mice as determined by Northern blotting in wild-type and transgenic testes at various ages (Fig. 2). In wild-type testis, GDNF, Ret and GFRa1 transcripts were highly expressed at birth and during the first postnatal week (Fig. 2a-c). GDNF mRNA was strongly downregulated shortly before the initiation of puberty (at two weeks of age) (Fig2a). Simultaneously, the Ret (Fig. 2b) and GFRal (Fig. 2c) mRNA levels declined but were still detectable in adult testes. The forth number artemin (ART) binds to GFRa3 and also signals via Ret.

Neurturin (NTN) is a GDNF-related molecule that utilises the same multicomponent receptor complex as GDNF but prefers GFRa2 (Klein, R. D. , et al. , Nature 387,717-721 (1997) ) as a co-receptor, a homologue to GFRal. Both NTN and GFRa2 tran- scripts are expressed by postnatal testes (Widenfalk, J. , et al. , J. Neurosci. 17, 8506-8519 (1997) ), however preferenti- ally after puberty. The third member of GDNF-family, persephin (Milbrandt, J. et al. , Neuron 20,1-20 (1998)), binds to chicken GFRa4 (Enokido, Y. , et al. , Curr. Biol. 18,1019-1022 (1998) ), whose homologue has, to date, not been found in mammals.

Northern blot with a human GDNF mRNA specific probe showed that the transgene was expressed only in testis, and not in other organs of both sexes, such as ovary, penis, brain, liver, kidney, lung, and skin (data not shown). In contrast to the declining of the levels of GDNF mRNA in wild-type mice, the transgenic GDNF mRNA expression remained high in testis from neonate to adulthood (Fig. 2d). Also, Ret and GFRal mRNA remained high at all ages analysed (Fig. 2e, 2f). Immunoprecipi- tation of GDNF from 3 and 6-week-old wild-type and transgenic testes verified that the GDNF protein levels were strongly increased in transgenic testes as compared to wild type testes (Fig. 2g).

Previously, embryonic Sertoli cells and a Sertoli cell lineage have been shown to express GDNF (Hellmich, H. L. , et al., Mech. Dev. 54,95-105 (1996) ; Trupp, M. , et al. , J. Cell Biology 130,138 - 148, (1995) ). By in situ hybridisation, GDNF mRNA was detectable in wild-type seminiferous tubules during the first postnatal week but not thereafter (Fig. 3a, 3b) and the label for GDNF mRNA was distributed over Sertoli cells. The transgenic testes showed a continuous and high expression of the human GDNF mRNA by type A spermatogonia clusters from two weeks of age to adulthood (Fig. 3c). Ret mRNA was expressed by both spermatogonia and spermatids in wild-type testes (Fig. 3d, 3e) and in the clusters of spermato- gonia in transgenic testes (Fig. 3f). GFRa1 mRNA was expressed by spermatogonia, spermatocytes and round spermatids in wild-type testes (Fig. 3g, 3h) and also intensively by the spermatogonia clusters in transgenic testes (Fig. 3i). Thus, the signalling receptor complex for GDNF is present in the spermatogonia of both wild-type and transgenic mice. In wild-type mice, GDNF is obviously acting as a paracrine Sertoli-cell-derived regulator of spermatogenesis, but in transgenic mice this regulation is disrupted by the continuous autocrine expression of GNDF by germ cells.

Spermatogonia are the only proliferating germ cells in testis and the mitotic activity is highly segmented in normal semini- ferous tubules (Fig. 4a). In transgenic mice, the BrdU- labelled spermatogonia showed an abnormal distribution within the seminiferous tubules, reflecting the spermatogonia clust- ers in the lumina of seminiferous tubules (Fig. 4b). Because of the highly segmental nature of cell proliferation in wild-type testis, it is difficult to compare the net proli- feration rates of wild-type and transgenic testes. However, the peak cell proliferation index of spermatogonia (BrdU- labelled nuclei/100 spermatogonia) was clearly lower in transgenic than wild-type mice : 0. 26 +/-0. 11 (n=714) and 0. 72 +/-0. 14 (n=436 cells, counted from the proliferative segments of seminiferous tubules), respectively (Fig. 4c).

Apoptosis was prominent in the clusters of spermatogonia and other poorly differentiated germ cells from the second post- natal week, as shown by TUNEL-labelling (Fig. 4d, 4e). Large numbers of dead cells, some of which resembled type A sperma- togonia, were engulfed and phagocytically removed by Sertoli cells (Fig. 4f, 4g). Thus, increased apoptosis overriding cell proliferation may explain the gradual testicular atrophy in the GDNF overexpressing mice. On the other hand, the accumu- lation of spermatogonia in young mice seems to be caused by the block of their differentiation and not enhanced cell proliferation.

Three GDNF overexpression males were followed for 5 months and in that time neither macroscopic testicular tumours nor microinvasive cancer cells were observed. Furthermore, we have not seen testicular malignancies in any of the younger male mice autopsies (n=103), showing that the differentiation defect of spermatogonia in GDNF overexpressing mice does not ultimately lead to germ line malignancies but results in testicular atrophy. However, more mice should be followed in long term to exclude any increased risk of testicular cancer.

Male mice or rats are sterile in the absence of the vitamin A metabolite, retinoic acid (RA), as shown by feeding them with vitamin A deficient (VAD) diet. Differentiation of type A spermatogonia is completely blocked by VAD diet and restored by reintroduction of vitamin A or its active metabolites (Kim, K. H., et al. , Molec. Endocrinology 4 : 1679-1688,1990).

Retinoic acid receptor a (RARa) mRNA and protein levels decline upon a VAD diet (Akmal, K. , et al. , Endocrinology 139, 1239-1248,1998). Accordingly, RARa-deficient transgenic mice are infertile and develop testicular atrophy (Lufkin, T. , et al. , Devel. Biol 90,7225-7229, 1993). Therefore, the expres- sion of RAR proteins was analysed in wild type and GDNF overexpressing transgenic mice. The RAR expression practically disappeared in the spermatogonial clusters of the transgenic mice, as show by indirect immunoperoxidase staining with antibodies to RAR proteins. The data indicate that one of the mechanisms how GDNF acts as a negative regulator of spermato- gonia is downregulation of RAR expression in Ret expressing spermatogonia.

To analyse whether the other members of the GDNF family would act as negative regulators of spermatogenesis, the investi- gators constructed transgenic mice overexpressing neurturin in testis. Indeed, the transgenic mice expressing neurturin under the translation elongation factor la promoter that targets the transgene expression specifically to testis show also inferti- lity during the first four postpubertal weeks as described by mating tests identical to those with the GDNF overexpressing mice.

Sperm differentiation in mammals can be divided into three distinct phases that are regulated by both local cellular interactions and hormonal control (Russell, L. D. , et al., Histological and histopathological evaluation of the testes.

Clearwater, F. L. , Cache River, pp. 1-40 (1990) ). The first phase of spermatogenesis is the mitosis phase, when the spermatogonia give rise to the spermatocytes. Beyond the spermatogonial stage, spermatogenesis proceeds without mitotic cell divisions. Spermatogonia are further classified to subtypes, in which the type A spermatogonia subtypes include the spermatogenic stem cells. The second phase is the meiosis phase of spermatocytes resulting in haploid spermatids. The third phase is the spermiogenic phase, in which the spermatids transform into spermatozoa, the sperm cells, structurally equipped to reach and fertilise the egg. During the mitosis phase, type A spermatogonia proliferate and are triggered to differentiate. However, some of them remain undifferentiated and are considered as the reservoir of stem cells (Russell, L.

D. , et al. , Histological and histopathological evaluation of the testes. Clearwater, F. L. , Cache River, pp. 1-40 (1990)). The precise mechanisms by which the stem cells are maintained or transformed into differentiating spermatogonia have re- mained unresolved. Our Northern blotting showed that the down- regulation of GDNF coincides with the initiation of the sperm differentiation in normal mice and GDNF overexpression arrests spermatogonial differentiation. Therefore, local GDNF signal- ling may be essential for spermatogonia to maintain their undifferentiated stage.

There is increasing evidence that each step of spermatogenesis requires paracrine interactions between somatic Sertoli cells and germ cells, and these interactions may be mediated by growth factors and their receptors (Bellve A. R. & Zhang W.

J. , Reprod. Fertil. 85,771-793 (1989) ; Skinner, M. K., Endocr. Rev. 12,45-72 (1991) ; Pescovitz, 0. H., et al., Trends Endocrinol. Metab. 5,126-131 (1994). Parvinen, M. , et al. , J. Cell Biol. 117,629-41 (1992), Djakiew, D. , et al., Biol. Reprod. 51,214-21 (1994), Cancilla, B. & Risbridger, G. P., Biol. Reprod. 58,1138-45 (1998) ; Bitgood, M. J. , et al. , Curr. Biol. 6,298-304 (1996) ; Zhao, G. Q. , et al., Genes Dev. 10, 1657-1669 (1996) ; Zhao, G. Q. , et al. , Develop- ment 125,1103-1112 (1998) ). However, direct functional data regarding the intercellular communication between testicular somatic cells and germ cells are still scarce. The interplay between Sertoli and germ cells may involve a number of mole- cules, such as stem cell factor (c-kit ligand), insulin-like growth factor-I, three members of the TGF-S family (Mullerian inhibiting substance, activin, and inhibin), and transferrin (Pescovitz, O. H. , et al., Trends Endocrinol. Metab. 5, 126-131 (1994) ). Sertoli cells secrete all these molecules and germ cells possess their receptors. Conversely, Sertoli cells express receptors for ligands secreted by germ cells. Sertoli cells for instance express p75 nerve growth factor receptor and fibroblast growth factor receptors, whose cognate ligands are secreted by testicular germ cells (Pescovitz, O. H. , et al., Trends Endocrinol. Metab. 5,126-131 (1994) ; Parvinen, M. , et al. , J. Cell Biol. 117,629-41 (1992) ; Djakiew, D. , et al. , Biol. Reprod. 51,214-21 (1994) ). Furthermore, Sertoli cells also interact with Leydig cells. Desert hedgehog (Dhh) gene is expressed by Sertoli cells and its putative receptor, patched, predominantly by Leydig cells (Bitgood, M. J. , et al. , Curr. Biol. 6,298-304 (1996). Indeed, Dhh null mice show a severe differentiation defect of spermatogenesis (Bitgood, M. J. , et al., Curr. Biol. 6,298-304 (1996) ). Recently, bone morphogenetic protein 8a and 8b together with their putative receptors were found in the germ cells and inactivation of these ligands resulted in abnormal spermatogenesis and male infertility (Zhao, G. Q. , et al., Genes Dev. 10,1657-1669 (1996) ; Zhao, G. Q. , et al., Development 125,1103-1112 (1998) ). The co-localisation of those growth factors and their receptors in germ cells suggests an autocrine regulation of germ cells.

Our data demonstrate a novel function for GDNF. The disturb- ance of spermatogenesis in transgenic testis, targeted to overexpress GDNF, suggests that GDNF acts as a local inhibit- ory signal in spermatogonia differentiation. The gradual testicular athrophy and the hyperplasia of Leydig cells in the transgenic mice are apparently secondary to the spermatogenic deficiency. Therefore, GDNF targeting to testis provides a new animal model for male infertility. The activation of the GDNF signalling cascade by GDNF, Ret agonists or downstream sub- strates possibly provides us with means whereby sperm pro- duction in man could be prevented and male contraceptives developed. The effects of GDNF on spermatogenesis should also be taken into consideration, when GDNF-like molecules are designed for and used in clinical trials of neurodegenerative disorders, such as Parkinson's disease.

The invention is described in more detail in the following experimental part, which discloses the material and methods used the preparation of transgenetic mice and the experiments used to demonstrate the effect of GDNF in the animal model.

Said experimental part should not be construed to limit the scope of the invention. Those skilled in the art can foresee a multitude of different application based on the results obtained in the following experimental part.

Methods GDNF transgene construction and production of transgenic mice.

The human GDNF cDNA was cloned to the pEF-BOS vector with the EF1a-promoter (Migushima, S. and Nagata, S. , Nucleic Acid Res.

18,5322 (1990) ) by replacing XbaI-XbaI stuffer fragment with entire coding sequence of human GDNF cDNA (GenBank accession number L15306). The correctness of the expression construct was verified by sequencing. Proper expression GDNF protein with the construct was verified by Cos cell transfection followed by immunoprecipitation of GDNF from lysate and culture supernatant. Transgenic mice were produced by micro- injecting the 2.7kb PvuI-HindIII fragment of the GDNF cDNA construct into the pronuclei of newly fertilised FVB inbred mouse eggs as described (Hogan, B. , et al., Manipulating the Mouse Embryo. Cold Spring Harbor, NY : Cold Spring Harbor Laboratory Press (1994) ). The founders were identified by Southern blotting with a human GDNF cDNA and the offspring of the transgene positive mice by PCR from tail DNA. Primers for human GDNF were from exon 1 (5'--TGT CGT GGC TGT CTG CCT GGT GC-31) and exon 2 (5'-AAG GCG ATG GGT CTG CAA CAT GCC-3').

Histology, cell proliferation and apoptosis. For histology, freshly dissected testes and epididymides were fixed in Bouins fixative or 4% PFA for 2-24 hours depending on the size, processed in paraffin, cut at 7 m and stained by hemato- xylin/eosin. For the cell proliferation assay, the mice were injected intraperitoneally with a BrdU and 5-fluoro2-deoxy- uridine (FrdU) cocktail (Amersham) at a concentration of 10 mg/kg in PBS. After 2 hr, the mice were sacrificed by neck dislocation, autopsy was done, testes were removed and fixed in Bouin solution. Testes were processed in paraffin, 7 m sections were cut and deparaffinised. BrdU incorporation was detected by indirect immunofluorescence with monoclonal antibodies to anti-bromodeoxyuridine (Amersham) and rhoda- mine-conjugated goat anti-mouse IgG (Jackson). Apoptotic cells were detected in deparaffinised sections with ApopTag in situ cell death detection kit (Oncor) according to manufacturer's instructions.

Living cells. Seminiferous tubules were dissected and squashed on the preparation slides with PBS buffer. Computer-aided analysis of cells within the seminiferous tubules was carried out according to Parvinen M. , et al. (Histochem. & Biol. 108, 77-81 (1997)).

Northern blotting. Total RNA of different tissues were iso- lated with TRIZOL Reagent kit (Life Technology) according to the manufacturer's instructions and 30g of total RNA was used in each lane. Northern blotting was carried out according to Maniatis protocol (Sambrook, J. , et al. , T. Molecular Cloning : A Laboratory Manual. Second Edition. Cold Spring Harbor Laboratory Press 1989 pp. 7. 43-7. 50). The cDNA probes (the same as in situ hybridisation, see below) were labelled with 322dCTP. The filters were exposed in a FujiBAS phospho- imager.

Immunoprecipitation and Western blotting with anti-GDNF antibody. Freshly dissected testes were homogenized and lysed in high salt lysis buffer (300 mg protein/ml ; high salt lysis buffer : 1M NaCl, 100 mM Tris HCl, pH 8,2 % BSA, 4 mM EDTA, 0. 2 01 Triton X-100,2 mM PMSF, 1 mM sodium orthovanadate and 1 tabl./10 ml of protease inhibitor cocktail tablets (Complete, Mini EDTA, Boehringer Mannheim) ). The lysates were immuno- precipitated with a polyclonal antibody to human recombinant GDNF peptide that crossreacts with mouse GDNF (R & D Systems) and Protein A-Sepharose. The immunoprecipitates were run on a 20 % SDS-PAGE and transferred to Hybond ECL nitrocellulose membranes (Amersham Life Science). The membranes were blocked with 5 % bovine serum albumin and immunoprecipitated GDNF was detected with a polyclonal anti-GDNF antibody (Santa Cruz Biotechnology) at room temperature for 2 h. Detection was accomplished by using anti-rabbit secondary antibody con- jugated to horseradish peroxidase (Sigma) and ECL chemilumine- scence (Amersham) according to the manufacturer's instructions In situ hybridization. In situ hybridisation was performed as described (Wilkinson, D. & Green, P. , In situ hybridization and the three-dimensional reconstruction of serial sections.

In : Postimplantation Mammalian Embryos. A practical approach (ed. A. Copp & D. Cockroft), pp. 155-171. Oxford University Press : London (1990) ). Antisense and sense cRNA probes (mouse GDNF 328 bp from exon 3, human GDNF 513 bp from exon 1 and exon 2, mouse Ret from 3 end 714 bp, mouse GFRa1 777 bp) were synthesized using appropriate RNA polymerases and 35S-labelled UTP. Hybridisation temperature was 52°C and autoradiography slides were exposed at +4°C for 2-4 weeks. The slides were photographed with CCR camera hooked with computer. In Photo- Shop graphics programme, the dark field images were inverted, artificially stained red and combined with the bright field images.

Example 1 Preparation of transgenic mice Targeting GDNF expression under the translation elongation factor 1 alpha-promoter to testis, wherewith the function of GDNF in testis is studied. The entire coding region of human GDNF (hGDNF) cDNA is cloned to the pEF-BOS vector by replacing the XbaI-Xbal fragment with hGDNF cDNA (Mizushima, S. & Nagata, S. , Nucleic Acids Res. 18,5322 (1990) ). The 2. 7 kb HindIII-EcoRI fragment isolated from the pEF-BOS vector contains EFla (translation elongation factor la) promoter and the 0. 7 kb Xba-EcoRI fragment having a ply (A) adenylation signal. The size of the entire coding region of human GDNF cDNA is 636 bp. Transgenic mice are produced by microinjecting the 2. 7 kb PVuI-HindIII fragment into the pronucleus of fertilized egg cells. The injection is at first made into a FVB/NIH-mice strain and later the transgene positive mice are mated with a FVB/NIH- and NMRI-mice strains. Transgenic founder mice are identified by Southern hybridization with a hGDNF-probe and the transgenic offspring is identified by polymerase chain reaction (PCR) from tail DNA (the transgene transferring mouse strain is later called GDNF t-mice).

Primers used in PCR are from exon 1 (5'-TGT CGT GGC TGT CTG CCT GGT GC-3') and exon 2 (5'-AAG GCG ATG GGT CTG CAA CAT GCC-3'). The results indicated that GDNF t-mice have a coding region of human GDNF cDNA and they transfer it to their offspring.

Example 2 Expression of the transgene is studied by Northern blot hybridization and RT-PCR Expression of the transgene is studied both by Northern blot hybridization in mice of different ages (0-13 weeks old), with which endogenic and transgenic GDNF mRNA levels at different ages are detected, and by reverse transcription polymerase chain reaction (RT-PCR) from RNA from the testis of a three week old mouse, with which the expression levels of endogenic GDNF mRNA and the human transgenic GDNF mRNA of a mouse is compared. For the reverse transcription PCR 1 g entire RNA is copied by reverse transcrition in a random-primed reaction (201) using AMV reversed copy enzyme (Finnzymes, Helsinki) according to the instructions of the manufacturer. 21 of the reaction product is used as template when copying the endogen- ic mouse GDNF cDNA or the transgenic human GDNF cDNA.

The first primer (5'GCTGAG/CCAGTGACTCA/CAATATGCC3') recognizes both the mouse and human GDNF alleles and covers the intron region in order to avoid genomic contamination. The primers of the other end were selective to endogenic and transgenic GDNF (GDNF specific region in mouse : 5'TGTTAGCCTTCTACTCCGAGACAG3') and the transgenic primer (transgenic GDNF specific region 5'CACCAGCCTTCTATTTCTGGATAA3'). These primers are used to copy in identical conditions (lM primer, lmM, dNTP and Dynazyme polymerase (Finnzymes, Helsinki) in their own buffer) frag- ments of 373 bp length from endogenic and transgenic GDNF cDNA, which has been reversed copied either from RNA from the testis of a normal or transgenic mouse. The cycle is 30s 95°, 30s 62° ja 45s 72°, and it is renewed 20-36 times after a "hot starting"95°C for 2 minutes. The copying is ended after the last cycle for 5 minutes in 72°C. The PCR product is run in an agarose gel containing ethidium bromide, is studied under UV-light and is photographed with a Polaroid-camera.

Northern blot hybridization to identify transgenic hGDNF and endogenic mouse GDNF mRNA expression is carried out with species specific a-32P-labelled GDNF cDNA probes from to- tal-RNA isolated from testis with a TRIzol Reagent-kit (Life Technologies) according to the manufacturer's instructions.

The probes are : mouse GDNF (328 bp exon 3 (NCB1 Gene Bank, accession number U37459) and human GNDF (536 bp from exon 1 ja 2, (Lin et al., Science 260,1130 - 1132, (1993) ). Northern blot hybridization is carried out according to Sambrook, J. , et al. , (Molecular Cloning : A Laboratory Manual. Second Edi- tion. Cold Spring Harbor Laboratory Press 1989 pp. 7. 43-7. 50). In each sample the total RNA used is 30 g. The hybridization filter is kept on a Fuji phosphoimager plate and exposed to light overnight and the image is recorded with a FujiBAS phosphoimager. A actin-probe is used as a probe to control even loading of RNA, which is hybridized on each filter after the hybridization of the GNDF-probe. The results showed a strong expression of GDNF mRNA in the testis of a mouse at birth (Figure 2a), which expression is weakened (declines) during the first postnatal weeks, especially when the mouse reaches puberty, and is barely detectable in an adult mouse.

Based on Northern blotting it can be estimated that in the testis of a three-week-old GDNF t-mouse (Fig. 2d) there is transgenic human GDNF mRNA 50-500 times more than endogenic GDNF mRNA (Figure 2a).

Example 3 Location of GDNF to testis by Northern blot hybridization and in situ hybridization Due to the testis targeting elongation factor 1a promoter the GDNF-transgene is not expressed in the male elsewhere than in testis (Furuchi, T. , et al., Development 122,1703-1709 (1996)), which is studied with Northern blot hybridization from different female and male tissues. Also, cRNA in situ hybridization is carried out from testis, ovary and kidney according to Wilkinson, D. & Green, P., (In situ hybridisation and the three-dimensional reconstruction of serial sections.

In : Postimplantation Mammalian Embryos. A practical approach (ed. A. Copp & D. Cockroft), pp. 155-171. Oxford University Press : London (1990) ). The probes are : mouse GDNF-probe (328 bp), human GDNF-probe (536 bp), GFRal-probe (777 bp) and GFRa2-probe (approximately 500 bp). For the in situ hybridiza- tion temperature with all probe is 52° C and exposure times are at +4°C 1. 5-4 weeks. From each probe at least three different exposures are developed. The results obtained with both Northern blot hybridization and cRNA in situ hybridization showed that the human GDNF-probe hybridizes only to RNA from testis, in which the in situ hybridization signal is located into the spermatogonia and some early spermatocytes (Figure 3).

It was also shown that in both transgenic GDNF t-mice and normal, mice GDNF-receptor Ret is expressed in prepubertal mice spermatogonia and in adult mice also in spermatides, but not any more in differentiated spermatozoa and the co-recep- tors of GFRa1 and GFRa2 are expressed during spermatogenesis also in germ line cells (Figure 3).

Example 4 Morphology of expression shown by histology The morphological results of the expression of transgenic GDNF is followed by histological studies of the testis. Normal and transgenic mouse tissues are fixed with 4 % paraformaldehyde overnight, processed in paraffin, cut in 4-7 ßm sections, deparaffinized and stained with hematoxylineosin-stain.

In the histological studies of 1-4 week old testis a strong elevation of the amount of spermatogonia can be seen in males of GDNF t-mouse strains, which elevation is related to the fact that they do not continue their differentiation to spermatids and spermatozoa. Based on their histological structure, the cell clusters formed within the seminiferous tubules can be identified as spermatogonia of A-type (Figure 1B). After four weeks the number of said spermatogonia cell clusters in the seminiferous tubules are decreased and after 11 weeks only few spermatogonia are left. A mature and free spermatozoa is not formed in any stage and sperm cells cannot be detected in epididymal ducts.

Example 5 Demonstration of ability of GDNF to cause infertility Because GDNF can prevent the formation of sperm cells, the fertilization ability of male GDNF t-mouse strain was studied by allowing them to mate with fertile, more than 8 week old females. 20 males (more than six weeks old) were kept with females for two weeks. Calculated together they mated with 200 females. Additionally, three males (from the age of eight weeks) were kept half a year with females so that the females were changed with three to four days intervals. They were mating as often as normal mice. After mating the uterus and egg leaders of some females were studied in order to find sperm cells.

The results obtained indicated that not a single female, which had mated with a transgene carrying male from GDNF t-mouse strain became pregnant, and from their uterus or egg leaders no sperm. cells were found. The transgene carrying males of the GDNF t-mouse strain are totally infertile and are not capable of breeding litter.

The results are also discussed in the Figures as described in detail below : Figure 1. Testicular morphology in wild-type (left panel : a, c, e) and transgenic (right panel : b, d, f) mice at three different ages, and wild-type (g) and transgenic (h) epidi- dymal ducts at 8-weeks. (a and b) 3-weeks of age. (c and d) 8 weeks. (e and f) 6 months. Note in d the remnants of type A spermatogonia clusters that are abundant in the young mice (d), and the advanced germ cell atrophy and Leydig cell hyperplasia (star) in f. Epididymal ducts from a wild-type (g) and transgenic (h) mouse show the lack of sperm in the trans- genic mice. Scale bar 100 ym.

Figure 2. Northern blotting for GDNF (a, d), Ret (b, e) and GFRa1 (c,f) mRNAs in wild-type (left panel) and transgenic testes (right panel) at various ages. Mouse GDNF cDNA and human GDNF cDNA probes were used in a and d, respectively. GDNF (a), Ret (b), and GFRal (c) mRNA levels are strongly downregulated after the first or second postnatal week. In contrast to the declining endogenous GDNF mRNA levels in wild-type mice, the levels of human GDNF mRNA transgene (d) remain high to adulthood. Also, Ret (e) and GFRal (f) mRNA levels are high in the transgenic mice at all stages analysed.

(g) Immunoprecipitation of GDNF from wild-type (WT) and transgenic (TG) testes at 3 and 6, 5 weeks of age. In wild-type testes, the GDNF protein is not detectable. On the right, 25 ng of recombinant human GDNF protein as a control.

Figure 3. The distribution of GDNF (a,b,c), Ret (d, e, f) and GFRa1 (g, h, i) transcripts in testes of wild-type (WT, left and middle panel) and transgenic mice (TG, right panel) by cRNA in situ hybridisation. In (a, b) mouse GDNF cDNA template and, in (c), human GDNF cDNA template were used. All other probes were to mouse transcripts. Sense controls did not show grain density above the background. (a, d, g) 1- to 2-weeks-old wild-type testes. (b, e, h) 8-weeks-old wild-type testes.

(c, f, i) 8-weeks-old transgenic testes. Note that the spermato- gonia in the transgenic mice continuously express human GDNF at high levels, but the number of these cells is clearly diminishing at 8 weeks of age. (d) At two weeks of age, Ret expression is localised to some spermatogonia in the basal layer of the seminiferous tubules (arrows). (g) Also, GFRa1 is expressed by some spermatogonia, morphologically similar to those with Ret (arrows). (e) At 8 weeks of age, Ret is ex- pressed in both some spermatogonia and spermatids. (h) GFRa1 mRNA distribution follows that of Ret, but is also highly expressed by spermatocytes. In transgenic mice, both Ret (f) and GFRal (i) are expressed in the clusters of spermatogonia.

Scale bar 100 m.

Figure 4. Cell proliferation (a, b, c), apoptosis (d, e), and live morphology (f, g) in wild-type (a, d) and transgenic (b, e, g) testes. (f) depicts both a wild-type and transgenic seminiferous tubule. (a) In 3-weeks-old testis, BrdU labelling of S-phase cells is seen only in the periphery of wild-type seminiferous tubules. Note the absence of labelling in some transsections of tubules, reflecting the segmented distribu- tion of S-phase cells during spermatogenesis. (b) In a 3-week- old transgenic testis, many BrdU-labelled cells are seen in the luminal area of the seminiferous tubules, reflecting the abnormal distribution of spermatogonia (arrow). The segmental distribution of S-phase cells is not seen in the transgenic tubules. Figure 4c. Distribution of cell proliferation indexes in wild type and transgenic testes. The number of BrdU labeled cells/100 spermatogonia was counted in 100 transections of wild type and transgenic siminiferous tubules. The normal segmental distribution of BrdU labeled cells is pertured in transgenic testes. (d and e) TUNEL-labelling for apoptosis in 4-week-old testes. Note the increase of labelling in sperma- togonia clusters (arrow) (e), as compared to wild-type (d) seminiferous tubules. (f) Live preparation of seminiferous tubules from 15-week-old wild-type (WT) and transgenic (TG) mice. Note the reduced diameter of the transgenic tubule and the increase of Leydig cells (arrows) around it. (g) High mag- nification of living cells from a transgenic seminiferous tubule in an unfixed squash preparation showing large numbers of dead cells (arrow), some of which resemble type A spermato- gonia, in a Sertoli cell (arrow head). Scale bar (a-d) 100 m, (e) 200 ym, (f) 10 m.

Table 1 Testis weights (average # SD) in wild-type and transgenic mice at various ages n=the number of testes in each group Age Wild-type Transgenic Weeks mg n mg n 1 5. 5 +2. 6 20 4. 8 + 2. 5 20 2 27.8 # 6.0 20 22.7 # 12 20 3 50. 2 +6. 5 20 58. 7 + 8. 6 20 4 68.8 # 8.0 10 40.3 # 10.0 10 <BR> 5 79.9 # 7.3 10 30.5 # 9.3 10 <BR> 8 102.2 # 4.6 10 34.7 # 8.0 8 <BR> 13 100.9 # 6.0 10 38.3 # 8.3 6 <BR> 2490. 0 4. 61043. 2 8. 96