This study was the first to visualize the changes in immunoreactive GDNF expression in the adult testes in a species-specific as well as spermatogenic activity- and seminiferous cycle-dependent manner. As anticipated, GDNF expression was specifically observed in Sertoli cells of the seminiferous epithelia, and its expression levels appear to be tightly regulated by the spermatogenic activity of the testes in both mice and hamsters. In mice, higher levels of GDNF expression were observed in seminiferous epithelium lacking germ cells than in seminiferous tubules colonized by donor germ cells (), which may reflect a positive response to compensate for the reduced germ cell number in the basal compartment. In contrast, lower levels of GDNF expression were noted in almost all Sertoli cells in photoregressed and hibernating hamster testes (). Moreover, during the subsequent testicular recrudescence, GDNF expression was shown to be clearly up-regulated at the initial phases which coincide with the resumption of spermatogenesis (). Since a rapid recovery in serum FSH/LH levels occurs before testicular function in the adult photoinhibited hamster 
, this is clearly consistent with the previous data that GDNF expression is tightly regulated immediately downstream of the gonadotropin-gonadal axis 
. Moreover, in “active” testes in hamsters, the stages II–VII of the high levels of GDNF-positive signals in the Sertoli cells () roughly coincide with those of the highest levels of FSH-induced cAMP production in the seminiferous epithelia (stages II–VI) 
Anti-GDNF immunostaining showed high levels of immunoreactive GDNF expression at the timing of spermiation in a seminiferous cycle-dependent pattern in hamsters. It is well known that, at the same seminiferous cycle stage as spermiation (stages VII/VIII in hamsters or stages VIII/IX in mice), preleptotene spermatocytes move across the blood-testis barrier from the basal to the adluminal compartment of the seminiferous epithelium 
. It is likely that the transition of preleptotene spermatocytes from the basal to adluminal compartment also leads to a transient increase in “spare room” for spermatogonia within the basal compartment of the seminiferous epithelium. Since higher levels of GDNF expression were observed in seminiferous epithelia lacking germ cells than in seminiferous epithelia colonized by germ cells in SSC-transplanted W/Wv
testes, these data suggest that some other signals which are derived from the presence or absence of “spare room” and/or advanced spermatogenic cells within the basal compartment may partially contribute to cyclical changes in GDNF expression in mammalian spermatogenesis.
The present anti-GDNF staining of whole seminiferous tubules successfully visualized the cyclical and patch-like distribution patterns of GDNF-positive granular deposits along the basal surface of Sertoli cells in both species. Double-staining of GDNF and its receptor, GFRα1, showed close co-localization of GDNF deposits and a subpopulation of GFRα1-positive spermatogonia in the basal region. Moreover, the present quantitative analysis revealed that GFRα1-positive cells showed a slender bipolar shape as well as a tendency for increased cell numbers in the GDNF-enriched area, as compared with those in the GDNF-low/negative area of the seminiferous tubules. For morphometric determination, further studies are required to generate hard data which can be statistically verified by using more accurate quantification of the GDNF signal levels around each GFRα1-positive cell. However, these findings suggest that GDNF-positive deposits along the basal surface of Sertoli cells have a “niche” function in the in vivo maintenance of SSCs in mammalian testes. These GDNF-positive deposits may not be static, but unstable dependently on the functional states and seminiferous cycle stages, although this does not fit the classical niche for the SSCs as a fixed and static structure in several lower vertebrates and invertebrate species 
The cyclical and patch-like distribution of GDNF deposits along the basal surface of Sertoli cells possibly leads to the asymmetric interaction of GDNF signals with some Apaired
GFRα1-positive cells (arrow in ; double-arrowhead in ), which may be consistent with recent suggestion showing asymmetric selection of SSCs from Aaligned
spermatogonia after fragmentation in vivo 
. Moreover, the present regionalized GDNF regulation in a small subpopulation of GFRα1-positive cells would explain the findings of a recent clone-fate study which showed that SSCs have an unexpectedly short life-span (average: ≤2 weeks) in the seminiferous epithelia 
. This is because, in both hamster and mouse, many GFRα1-positive cells do not appear to physically associate and co-localize with GDNF deposits (see , ), possibly leading to their eventual removal from a potential SSCs pool. Taken together, it is reasonable to speculate that such regionalized GDNF regulation may define the size of a pool of GFRα1-positive spermatogonia, especially in hamsters, possibly leading to the finely-tuned control between the self-renewal/survival and differentiation of the SSCs in the basal compartment of seminiferous tubules. Moreover, the present study demonstrated that the GDNF-positive signals are accumulated largely on the c-kit-positive spermatogonia along the basal surface of the seminiferous tubules (). This in turn suggests that the dynamics of a c-kit-positive population of Aaligned
spermatogonia clearly affects the size of a pool of GFRα1-positive spermatogonia (mostly Asingle
) in a positive feedback fashion. The components of GDNF-positive granular deposits, their association with the blood vessels, interstitial cells, peritubular myoid cells, and the molecular mechanisms underlying their distribution and turnover within the basal compartment of seminiferous tubules, could be a focus for future studies.
Kanatsu-Shinohara et al. (2008) 
reported that the general characteristics of hamster germline stem (GS) cells are similar to those of mouse and rat GS cells, indicating a conserved GDNF action of self-renewal and maintenance of the SSCs pool between seasonal and non-seasonal breeding rodents 
. Interestingly, we noticed the following species-specific differences in the expression profiles of GDNF and GFRα1 between mouse and hamster testes: 1) Hamster GFRα1-positive spermatogonia are more slender in shape and lower in cell density than those in mice, 2) GDNF expression in hamsters is more cyclical, is restricted to a narrower area along the longitudinal seminiferous tubule (i.e., only at stages II~VII), and consists of patch-like deposits. In contrast, GNDF expression in mice is ubiquitous/less cyclical, with granular GDNF deposits in a wider area along the longitudinal seminiferous tubule. These findings imply that the less cyclical and ubiquitous GDNF distribution in mice is closely associated with the maintenance of a large number of GFRα1-positive cells. On the other hand, the more restricted GDNF distribution would explain the relatively small number of GFRα1-positive spermatogonia in hamsters, as compared with that in mice. Interestingly, hamster GFRα1-positive spermatogonia are significantly more slender in shape than those in mice, which might possibly reflect the high migratory activity in hamsters. This small number of GFRα1-positive cells with a high migratory activity may have advantages over a SSCs pool which is rapidly changing in size during the transition between inactive and active states in seasonal breeding animals. This is because the up- and down-regulation of GDNF expression is directly transmitted to the rapid expansion of, and/or reduction in, the SSCs pool throughout the longitudinal seminiferous tubule. This observation is consistent with the present data which demonstrated the ubiquitous and widespread nature of GDNF expression in most seminiferous tubules in the initial phases of spontaneous testicular recrudescence in hamsters (“D6”, “C13” in ). Both GDNF and GFRα1 may be highly conserved molecules between mice and hamsters 
, reflecting the successful maintenance and colonization of hamster SSCs in mouse testicular soma 
and the higher cross-species reactivity of anti-GDNF and anti-GFRα1 antibodies (this study). Taken together, these findings indicate that the hamster testes in photoregressed, hibernating and subsequent recrudescent states are very useful in a comparative animal approach to understand the seasonal regulation and evolution of the SSCs and their niche in mammalian spermatogenesis.
In conclusion, the present study was the first to demonstrate the dynamic changes in immunoreactive GDNF expression and its close association with a small subpopulation of GFRα1-positive spermatogonia in the basal compartment of seminiferous epithelia. The unexpectedly cyclical and patch-like distribution of GDNF deposits implicates a novel hypothesis for in vivo maintenance of SSCs based on highly regionalized association between GFRα1-positive cells and extracellular GDNF signals in the basal compartment of the seminiferous epithelia of mammalian testes.