| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Population Council, Center for Biomedical Research, New York, New York 10021
Correspondence: Address all correspondence and requests for reprints to: Dolores D. Mruk, Ph.D., or C. Yan Cheng, Ph.D., Population Council, Center for Biomedical Research, 1230 York Avenue, New York, New York 10021. E-mail: D-Mruk{at}popcbr.rockefeller.edu or Y-Cheng{at}popcbr.rockefeller.edu
 |
Abstract |
|---|
 Top Abstract I. IntroductionII. Sertoli CellsIII. Sertoli-Sertoli and Sertoli...IV. Mechanism of Germ...V. Concluding RemarksReferences |
|---|
|
I. Introduction |
|---|
|
TopAbstractI. IntroductionII. Sertoli CellsIII. Sertoli-Sertoli and Sertoli...IV. Mechanism of Germ...V. Concluding RemarksReferences |
|---|
On this note, we attempt to provide a balanced but informative treatment of Sertoli and germ cells in the testis, in particular the functions of Sertoli cells, and the interactions taking place that confer germ cell movement (e.g., cell junction restructuring) during spermatogenesis. Although brief accounts are given in many subject areas, readers are encouraged to refer to the excellent review articles cited throughout the text and to references cited therein, in particular the role of Sertoli and germ cell proteins in cell-cell interactions and the regulation of spermatogenesis (for reviews, see Refs.4 and 8, 9, 10, 11, 12). We focus our discussion largely on Sertoli-Sertoli and Sertoli-germ cell interactions at the level of cell junctions rather than the paracrine and/or autocrine factors used by these cells for communication because this latter subject area has recently been reviewed (for reviews, see Refs.8 , 9 , and 11). Furthermore, recent advances in the field have clearly shown that most, if not all, of the junction types found in the testis are nonstatic structural entities. In other words, they are dynamic structures that serve as platforms for signal transduction events that regulate the opening and closing of these junctions (see Section III). As such, we first briefly introduce the different types of junctions found in the testis (Table 1
). A detailed discussion on the structure and functions of Sertoli cells is found in Section II, which is followed by in-depth discussions on the different types of junctions, their constituent proteins, and regulation. These are followed by a description of a current biochemical model that regulates the movement of germ cells not only across the blood-testis barrier but also the remaining seminiferous epithelium during spermatogenesis.
|
). As such, the term adherens junction refers only to one of the four types of anchoring junctions. In addition, the ectoplasmic specialization and tubulobulbar complex are two of the best-studied adherens junction types in the testis (Table 1). Although their presence in the testis has been demonstrated, they remain to be poorly characterized (Table 1) (for reviews, see Refs.15, 16, 17, 18). For example, it is known that germ cells, such as pachytene spermatocytes and round spermatids, must remain attached to Sertoli cells throughout the 14 stages of the epithelial cycle in the rat, yet the molecules that confer this adhesiveness are not all known (for reviews, see Refs.15 and 18).
In the following sections, we will limit our discussion to the biology of Sertoli-Sertoli and Sertoli-germ cell interactions at the level of cell junctions, emphasizing recent advances in the field. We will also highlight specific areas that deserve additional investigation. We begin by summarizing the three junction types found in the testis and the proteins that constitute these junctions, which will be discussed in greater detail (see Section III) after a discussion on the structure and function of Sertoli cells. This is done because until recently much of the information on junction dynamics in the testis was derived from studies using Sertoli cells (for reviews, see Refs.4 , 6 , 19 , and 20). Figure 1 is a schematic drawing of the seminiferous epithelium in the rat testis illustrating the relative locations of the different types of junctions found between Sertoli cells, as well as between Sertoli and germ cells. Figure 2 is an electron micrograph showing the intimacy between the seminiferous epithelium and the basement membrane (21). As shown in Fig. 1, Sertoli cell tight junctions that create the blood-testis barrier also anatomically divide the seminiferous epithelium into basal and adluminal compartments (for review, see Ref.3). Although it is obvious that the current biochemical architecture of tight and adherens junctions in the testis will be rapidly updated, it is our intention to summarize recent advances in this rapidly evolving field so that this review can serve as a guide for future studies. Every effort was made to credit original investigators with their findings, but because of the large number of references, we often chose to cite reviews rather than original articles. As such, we apologize to investigators in the field for possible oversights on our part.
|
|
0.5–1.2 x 106 cells/cm2) on Matrigel-coated bicameral units or dishes (22, 23, 24, 25, 26, 27, 28, 29, 30). This model has been extensively characterized by different laboratories (30, 31, 32, 33, 34, 35, 36, 37). Studies on the regulation of blood-testis barrier dynamics are based on in vivo models of blood-testis barrier damage induced by cadmium chloride, glycerol, or occludin peptide (38, 39, 40, 41, 42, 43). These have also been previously characterized (for review, see Ref.44). For studies relating to the regulation of adherens junction dynamics, such as the ectoplasmic specialization (5, 45, 46, 47, 48, 49), an in vitro Sertoli-germ cell coculture system was used (5, 50, 51, 52). Unfortunately, no in vivo models to study adherens junction dynamics in the testis were available until recently (45, 49, 53, 54); these now include the use of 1-(2,4-dichlorobenzyl)-indazole-3-carbohydrazide (AF-2364; for review, see Ref.4) and androgen/estrogen implants to suppress the endogenous testosterone level (for review, see Ref.55). Both induce the loss of cell adhesion between Sertoli and germ cells, in particular spermatids. Detailed discussions on these in vitro and in vivo models are found in Sections III and IV. |
II. Sertoli Cells |
|---|
|
TopAbstractI. IntroductionII. Sertoli CellsIII. Sertoli-Sertoli and Sertoli...IV. Mechanism of Germ...V. Concluding RemarksReferences |
|---|
), the somatic constituents that extend from the base to the apex of the seminiferous epithelium, interact directly with developing germ cells throughout spermatogenesis. To this end, several studies have shown that these cells have an indispensable role in the development and movement of germ cells (for reviews, see Refs.61, 62, 63, 64, 65, 66, 67, 68, 69). The following discussion on Sertoli cell structure and function will provide the foundation for the discussion of Sertoli-germ cell interactions.
A. Sertoli cell structure
Since their discovery in the late 1800s, Sertoli cells have been extremely difficult to study morphologically because they have a constantly changing, three-dimensional relationship with developing germ cells throughout the 14 stages of the epithelial cycle (for reviews, see Refs.6 , 67 , and 70), yet Sertoli cells continue to be one of the most-studied cell types in the seminiferous tubule. They are irregularly shaped, columnar cells that extend from the basal to the adluminal compartment and occupy a volume of approximately 17–19% in the seminiferous epithelium of adult rats (71, 72) (Fig. 2). Sertoli cells are in direct contact with the basement membrane, a modified form of the extracellular matrix in the testis (21) (Figs. 1 and 2). They have an enormous surface area (Fig. 2), which allows them to sustain a vast number of developing germ cells at a Sertoli:germ cell ratio of approximately 1:50 in the adult rat testis (71, 73). This attribute, in itself, is crucial not only to spermatogenesis, but also to germ cell movement.
From a morphological standpoint, Sertoli cells at any one time can be classified into one of two categories: 1) type A, or 2) type B (for reviews, see Refs.18 , 67 , and 70). In type A Sertoli cells, mature spermatids ready for release into the tubule lumen are lodged deep within cytoplasmic crypts, whereas in type B Sertoli cells these cytoplasmic crypts are either barely visible or completely absent. Therefore, during the 14 stages of the epithelial cycle, type A Sertoli cells must transform themselves into type B Sertoli cells to adapt to the cellular changes that take place during germ cell development and movement. These observations have been supported by eloquent morphological studies in the rat, which have demonstrated that there are drastic changes in cell shape in stages II, VII, VIII, IX–XI, and XIII–XIV, illustrating that Sertoli cells function together with germ cells in forming the epithelial cycle (for reviews, see Refs.3 , 74 , and 75).
B. Sertoli cell functions
To date, several Sertoli cell functions, most of which are directly related to germ cell development and movement, have been described. These include: 1) providing structural support; 2) creating an impermeable and immunological barrier; 3) participating in germ cell movement and spermiation; 4) nourishing germ cells via their secretory products, and several others (for reviews, see Refs.62 , 66 , 69 , and 74), all of which are discussed in the following sections. It must be noted, however, that several of these functions were determined by studies using Sertoli cells cultured in vitro, and the physiological relevance of some of these functions in Sertoli cells in vivo remains to be determined.
1. Structural support.
Sertoli cells provide support for developing germ cells within the seminiferous epithelium by participating directly in the deposition of extracellular matrix components and permitting the formation of specialized cell junctions. Another striking feature of Sertoli cells is their well-developed cytoskeleton, which is responsible for the collective organization of the seminiferous epithelium (for reviews, see Refs.3 , 7 , 15 , 16 , 18 , and 70). Morphological studies have shown that the Sertoli cell cytoskeleton: 1) maintains shape; 2) positions and transports organelles within the cell; 3) forms and stabilizes the cell membrane at sites of cell-cell and cell-extracellular matrix contact; 4) positions, anchors, and aids in the movement of developing germ cells; and 5) participates in the release of mature spermatids from the seminiferous epithelium at spermiation. It consists of three major components, namely actin, intermediate filaments, and microtubules, with each component having a unique pattern of distribution that changes in association with the 14 stages of the epithelial cycle (for reviews, see Refs.7 , 67 , and 76). Collectively, these components play a crucial role in facilitating germ cell movement.
a. Actin.
Actin filaments are composed of actin monomers of molecular mass (Mr) 42 kDa, each of which polymerizes into a linear chain with a helical twist approximately 8-nm wide (for reviews, see Refs.77 and 78) and exists in six different isoforms in mammalian tissues (79, 80). In the Sertoli cell, actin plays an extremely important role in maintaining structure and conferring cell contractility, but its presence at the cell periphery at sites of cell contact also suggests that it participates in motility-related processes. There are two pools of actin, namely monomeric G-actin and polymeric F-actin, that coexist in equilibrium in all eukaryotic cells, including Sertoli cells. The predominance of one pool affects the overall status of the cytoskeleton by causing polymerization or depolymerization of actin filaments. This, in turn, affects cell shape and movement (for reviews, see Refs.81 and 82). Figure 3A depicts a typical Sertoli cell cultured in vitro. Notice the F-actin filaments (green) that were visualized by using Oregon Green 488 phalloidin. Also shown in this figure is immunoreactive vinculin (red), an adaptor protein (Section III.A.5.b), and the Sertoli cell nucleus, which was stained with DAPI (blue). Figure 3B shows two typical Sertoli cells cultured in vitro for approximately 7 d and immunostained for zyxin (green), another adaptor protein found at the focal contact and leading edge of cytoplasmic processes. Recently, zyxin was demonstrated to associate with actin in Sertoli cells by immunoprecipitation (83). Actin filaments are also major components of ectoplasmic specializations and tubulobulbar complexes, modified anchoring junction types present between Sertoli cells, as well as between Sertoli and germ cells, illustrating that both of these junctions participate in the movement of developing germ cells (for reviews, see Refs.7 and 67). It is also of interest that mouse haploid germ cells express two unique actins designated T-actin 1 and T-actin 2, which share 40% homology with other actins (84). T-actin 1 is confined to the spermatid cytoplasm, whereas T-actin 2 is restricted to the sperm head and tail (84), suggesting that these germ cell-specific actins are important to the morphogenesis of germ cells during spermiogenesis.
|
) because they confer desmosome-like junctions between Sertoli cells and between Sertoli cells-spermatocytes/round spermatids (for review, see Ref.87). Although the exact role of intermediate filaments in Sertoli cells is not entirely known, their distribution within these cells suggests that they participate in maintaining the integrity of the seminiferous epithelium (88, 89, 90).
|
, ß, 
, 
, 
, and 
(104), but the major building blocks of the microtubule cytoskeleton are - and ß-tubulins (105). In general, microtubules appear to have a much clearer role in Sertoli cell function (for reviews, see Refs.7 and 106, 107, 108). They are known to: 1) maintain a columnar shape; 2) position and transport intracellular organelles; 3) translocate spermatids within the seminiferous epithelium; and 4) adjust the contour of the Sertoli cell membrane to adapt to the irregularly shaped spermatid heads that are lodged within crypts during spermatogenesis (for reviews, see Refs.6 and 7). It is not surprising, therefore, that the most striking changes in microtubule organization occur when mature spermatids associate with Sertoli cells (109). Microtubules are also found at the site of the ectoplasmic specialization (110, 111). Although the presence of actin, intermediate filaments, and microtubules in the testis has been implicated in the events of germ cell movement, the mechanisms underlying the regulation of the Sertoli cell cytoskeleton during different stages of the epithelial cycle and germ cell movement remain to be investigated.
2. Create an impermeable barrier.
The blood-testis barrier (also known as the seminiferous epithelial barrier), a modified occluding junction, is formed by approximately 15–18 d of age in the rat as constituted by Sertoli cell tight junctions located in the basal third of the seminiferous epithelium (Refs.2 and 112, 113, 114, 115 ; for reviews, see Refs.116, 117, 118) (Fig. 1). The blood-testis barrier compartmentalizes the epithelium into two (112, 119), rather than three, compartments as previously postulated (2, 114). One is a basal compartment in which spermatogonia, preleptotene, and leptotene spermatocytes reside. The other is an adluminal compartment in which meiotic spermatocytes and spermatids in various stages of spermatogenesis and spermiogenesis exist.
a. Concept of the blood-testis barrier.
The concept of the blood-testis barrier was first developed in the 1970s based on two important findings (for reviews, see Refs.112 , 120 , and 121). First, there were drastic differences in the composition of fluids and proteins obtained from the rete testis and seminiferous tubule lumen, and the testicular lymph and blood plasma. For example, it was demonstrated that fluid collected from the rete testis was higher in potassium and chloride than that collected from the blood plasma (Ref.122 ; for review, see Ref.123). In addition, the total protein content of the rete testis fluid was only approximately 1/20 to 1/50 that of the plasma (124). Second, there were variations in the rate in which radioactive tracers or dyes could pass from the blood plasma into testicular fluids (114, 125, 126, 127, 128). Although these results suggested that a barrier existed, concrete evidence, not to mention the physiological significance of such a barrier, was for the most part lacking until the 1980s when cell biology and morphology studies confirmed earlier observations.
b. Functions of the blood-testis barrier.
Three main functions are ascribed to the blood-testis barrier. The blood-testis barrier: 1) creates a specialized environment; 2) regulates the passage of molecules; and 3) serves as an immunological barrier. First, the blood-testis barrier contributes directly to the complex structural organization of the testis by creating a specialized environment necessary for germ cell development and movement. As such, Sertoli cells are obliged to synthesize, secrete, and efficiently deliver products that are essential for the growth and differentiation of developing germ cells (Refs.129 and 130 ; for reviews, see Refs.131, 132, 133, 134, 135, 136). For instance, the synthesis of testicular transferrin, which was first reported by Thorbecke et al. (137) and later by Skinner and Griswold (138), is the classic example of this Sertoli cell characteristic (for reviews, see Refs.136 and 139). Because germ cells residing in the adluminal compartment do not have access to serum iron, a transport mechanism orchestrated by Sertoli cells must exist. Nevertheless, recent in vitro studies have demonstrated that germ cells can indeed stimulate Sertoli cell transferrin mRNA and protein levels, suggesting that germ cells govern their own iron requirements via their interactions with Sertoli cells (140, 141, 142). One must be cautioned, however, that some of the earlier reported effects of germ cells on Sertoli cell function may indeed be artifacts because residual trypsin, which was used to isolate germ cells from the testis, can mimic some of the reported effects on Sertoli cells (143).
Tight junctions that constitute the blood-testis barrier also regulate the movement of products, such as nutrients and wastes, both into and out of the seminiferous epithelium (for review, see Ref.144). Although studies that have specifically focused on this tight junction attribute in the testis are by far limited, the regulated movement of solutes across tight junctions continues to be well studied in other organs, such as the kidney and intestine in normal (e.g., food absorption in the small intestine) and pathological (e.g., migration of neutrophils across the tight junction barrier during inflammation) responses. In the testis, the best example of regulated movement of products lies in the biology of FSH, whose major site of action is on Sertoli cells. However, FSH cannot enter the seminiferous tubule fluid readily (for reviews, see Refs.145, 146, 147). Therefore, it is not unexpected that receptors for FSH are found in the basal compartment residing on the Sertoli cell membrane (146, 148, 149, 150, 151, 152).
The third main function of the blood-testis barrier is to create an immunological barrier, ensuring that the organism’s immune system does not recognize unique antigens present on the cell surfaces of haploid germ cells. This would result in the organism being immunized against its own spermatozoa. In addition, recent studies have shown that Sertoli and germ cells maintain a unique antiviral defense system via the production of interferons (IFNs), ILs, and cytokines (153, 154). The blood-testis barrier also prevents the entrance of Igs and lymphocytes into the adluminal compartment (for review, see Ref.155).
c. Regulation of the blood-testis barrier.
The blood-testis barrier is one of the tightest barriers in the mammalian body comparable to the blood-brain barrier, and tight junction barriers in the endothelium of blood vessels and the epidermis. However, the blood-testis barrier is different in that it must open periodically to accommodate the passage of germ cells during spermatogenesis. For instance, developing preleptotene/leptotene spermatocytes must traverse the blood-testis barrier, which is known to take place in late stage VIII–early IX of the epithelial cycle (2). This process requires that Sertoli-Sertoli and Sertoli-germ cell junctions be disassembled and reassembled. This also illustrates the unusual dynamic nature of the blood-testis barrier. However, the mechanism(s) by which germ cell movement takes place remains to be elucidated. Additionally, the factor(s) that regulates the formation and maintenance of the blood-testis barrier is currently not known. For instance, gonadotropins have been implicated, but this is largely a speculation that has remained relatively unexplored. Until recently, the lack of information on the biology of the blood-testis barrier has largely stemmed from the lack of suitable in vitro and in vivo models that could be used to study Sertoli cell tight junction dynamics. These will be discussed below.
i. An in vitro model.
Sertoli cells cultured in vitro on Matrigel-coated dishes or bicameral units at a cell density ranging between 0.5 and 1.2 x 106 cells/cm2 are known to establish a functional tight junction barrier (24, 31, 32, 35, 36, 37). The assembly and maintenance of the tight junction barrier was assessed by transepithelial electrical resistance (TER); polarized secretion of Sertoli cell proteins; restricted diffusion of [3H]inulin, [125I]BSA, or fluorescein isothiocyanate (FITC)-dextran (Mr, 4.4 and 35.6 kDa); or maintenance of nonequilibrium of media across the Sertoli cell epithelium in bicameral units (24, 27, 31, 32, 35, 36, 37, 48). Using this system, several laboratories, including ours, have shown that Sertoli cell tight junction barrier function is regulated by an array of molecules, including intracellular calcium, cAMP, cytokines, protease inhibitors, protein kinases and phosphatases, and others (for reviews, see Refs.4 and 44). However, this in vitro system is not flawless. For instance, the Sertoli cell tight junction that constitutes the blood-testis barrier is one of the tightest barriers in the mammalian body (Section II.B.2). However, the tightness of the Sertoli cell tight junction barrier in vitro is only approximately 100 ohm·cm2 (22, 24, 27, 28, 29, 34, 57) vs. 1000–2000 ohm·cm2 in Madin-Darby canine kidney (MDCK) cells and keratinocytes (156, 157, 158) when barrier function was assessed by TER measurement. These results seemingly suggest that Sertoli cell tight junctions in vitro are only moderately tight when compared with other cell types. Recent studies have shown that the inclusion of androgens (27, 28), cAMP (34), and protease inhibitors (29) into Sertoli cell cultures during tight junction assembly can increase the tightness of the tight junction permeability barrier in vitro, illustrating that many biological factors present in vivo contribute to the tightness of this barrier in the testis. In short, these additional data demonstrate that this in vitro model is useful for studying tight junction dynamics.
Indeed, this in vitro model was used to verify some of the in vivo observations regarding Sertoli cell tight junction barrier function. For instance, recent studies have shown that extracellular matrix components, such as laminin and heparan sulfate, can promote the formation of the blood-testis barrier in vivo (Ref.159 ; for review, see Ref.160), whereas another report illustrated that laminin can also regulate the Sertoli cell tight junction barrier in vitro (161). To determine whether a relationship between extracellular matrix proteins and the assembly of the Sertoli cell tight junction permeability barrier does indeed exist (162, 163), this in vitro model was also used. Interestingly, the inclusion of an anticollagen 3(IV) antibody into Sertoli cells cultured at high density (1.2 x 106 cells/cm2) on Matrigel-coated bicameral units was able to perturb the assembly of the tight junction barrier, indicating that this protein is essential to tight junction functionality (163). This result also brings on a rare significance all on its own, because tight junctions in the testis are located in the basal third of the seminiferous epithelium closest to the basement membrane (21) [comprised of laminin, type I and IV collagen, heparan sulfate, and entactin (164)]. This is unlike other epithelia where tight junctions are located apically, farthest away from the extracellular matrix. What is not known, however, is whether this anticollagen-induced effect on Sertoli cell tight junction assembly is mediated via integrins or other yet-to-be identified molecules that constitute the focal adhesion or hemidesmosome. It has also been suggested that peritubular myoid and endothelial cells contribute, at least in part, to the functionality of the blood-testis barrier (97, 165), but the significance of this is presently unknown because there are no in vitro or in vivo studies examining the effects of these two cell types on the Sertoli cell tight junction barrier.
Other studies have demonstrated that the formation of the blood-testis barrier is not dependent upon germ cells because the disappearance of these cells from the seminiferous epithelium in vivo did not prevent the formation of the barrier (166, 167). Additional studies have demonstrated that the formation of the Sertoli cell tight junction permeability barrier is dependent on temperature and hormones (27, 33, 34, 168). For instance, the addition of testosterone (0.001–10 µM) into Sertoli cells cultured on bicameral units in vitro was able to enhance TER across the Sertoli cell epithelium (33, 34). This is not entirely unexpected because the testis is an androgen-dependent organ. On the other hand, treatment of rats with estradiol valerate did not affect the passive diffusion of [3H]inulin into the tubule lumen, although testicular weight and the serum testosterone level were shown to be altered (169), illustrating that estrogens do not regulate Sertoli cell tight junction barrier function.
There is also limited evidence to suggest that protease inhibitors can affect Sertoli cell tight junctions. A recent report by Okanlawon and Dym (29) demonstrated that chloroquine (an antimalarial drug and a protease inhibitor that has the property of increasing the pH within lysosomes and endocytic vesicles) was able to dose-dependently increase TER, making the Sertoli cell tight junction barrier tighter, when compared with Sertoli cells cultured in the absence of chloroquine. Unfortunately, the effects of protease inhibitors on Sertoli cell tight junction dynamics remain limited to this single report, and there is not a single study in which the effects of proteases are examined. In studies in which other cell types and cell lines such as MDCK cells were used, proteases were found to induce the aberrant formation of tight junctions. For example, the application of trypsin, a serine protease, together with calcium to the basolateral surface of confluent MDCK monolayers resulted in the formation of basally situated tight junction strands, which was accompanied by a transient increase in TER (170). These results seemingly suggest that this increase in tight junction barrier function is a normal physiological response of an epithelium to a potential cellular hazard. What is not known is whether the cellular distributions of occludin and claudin were altered in response to trypsin treatment. The participation of protease inhibitors was also studied in the human adenocarcinoma cell line, Caco-2, when these cells were cultured in the presence of nontoxic concentrations of either leupeptin or antipain (Ref.171 ; for review, see Ref.172). It was shown that these protease inhibitors could effectively block tight junction barrier assembly in vitro. These results, taken collectively, demonstrate the involvement of proteases and protease inhibitors in tight junction dynamics, yet similar studies in the testis await to be performed.
ii. In vivo models.
Once formed, the blood-testis barrier cannot be easily disrupted, such as by heat or irradiation (166, 173). However, ip administration of cadmium salts (e.g., cadmium chloride) can cause the blood-testis barrier to break down, resulting in complete degeneration of the seminiferous epithelium (28, 37, 40, 41, 174). This in vivo model of blood-testis barrier disruption is irreversible, and an explanation that would help to understand the destructive effects of cadmium chloride in the testis is still lacking. In studies using cell lines, the effects of cadmium chloride were shown to be concentration- and time-dependent, decreasing the tight junction barrier by TER, increasing the permeability of mannitol, and altering the localizations of occludin, zonula occludens (ZO)-1, and E-cadherin (175). In another recent in vitro study from our laboratory, Sertoli cells were cultured in the presence of cadmium chloride (noncytotoxic concentrations), which induced a disruption of the tight junction permeability barrier as assessed by TER. As anticipated, the tight junction barrier failed to reseal upon removal of cadmium chloride (27). However, the addition of testosterone (1 x 10–9 M) allowed reassembly of the Sertoli cell tight junction permeability barrier after cadmium chloride was removed from cultures, and a 200-fold increase in testosterone even protected Sertoli cells from cadmium chloride-induced damage (27). This suggests that this androgen can protect Sertoli cell tight junctions from being damaged.
Another in vivo model of blood-testis barrier destruction is the glycerol model (Refs.176, 177, 178, 179, 180 ; for review, see Ref.181). Similar to cadmium chloride, intratesticular injection of glycerol irreversibly disrupts the blood-testis barrier and, consequently so, spermatogenesis. However, glycerol does not affect Leydig cell steroidogenesis or serum FSH and LH levels (176). In a recently published article by Wiebe et al. (42), the effects of glycerol on the testis were expanded when the localizations of cytoskeletal and tight junctional proteins such as F-actin, tubulin, and occludin were examined by immunofluorescence. It was shown that glycerol, in addition to upsetting the localization of occludin, was capable of perturbing microfilament and microtubule organization in the testis. Unfortunately, comparable studies on the effects of glycerol on Sertoli cells in vitro are lacking. Moreover, results obtained from the study of these two models suggest that a primary disruption of the blood-testis barrier can also lead to a loss of adherens junction function since the treatment of rats with cadmium chloride or glycerol results in the loss of germ cells from the seminiferous epithelium, causing aspermatogenesis and infertility (for review, see Ref.44).
A third in vivo model to study blood-testis barrier dynamics is based on a recent study from this laboratory. When a single dose of a synthetic occludin peptide (NH2-GSQIYTICSQFYTPGGTGLYVD-COOH, corresponding to amino acid residues 209–230 in the second extracellular loop of rat occludin) (Fig. 4) was administered intratesticularly at 1.5–10 mg/testis, it caused reversible germ cell loss from the seminiferous epithelium (43). Morphological analysis of treated testes revealed that more advanced germ cells, such as elongate spermatids, began to deplete from the epithelium 1–2 wk after administration of the peptide (43). By 4 wk post treatment, there was a massive depletion of germ cells from the seminiferous epithelium in virtually all of the tubules examined. This event also associated with a disruption of the integrity of the blood-testis barrier when barrier function was monitored by micropuncture technique, detecting leakage of [125I]BSA from the systemic circulation into the lumens of the rete testis and seminiferous tubules (43). Figure 5 shows recent results using immunofluorescent microscopy to monitor the integrity of the blood-testis barrier by localizing occludin to this site. Damage of blood-testis barrier function was associated with an extensive disruption (Fig. 5, A—E) and loss of occludin immunostaining intensity (Fig. 5F) at the site of the blood-testis barrier. In addition, tubules in occludin peptide-treated testes shrunk significantly, with the tubule diameter reducing by as much as 20–30% when compared with control testes receiving either vehicle or a myotubularin synthetic peptide (43). Thereafter, germ cells began to repopulate the epithelium. By 7–8 wk, some spermatocytes were clearly visible in all of the tubules examined. By 10–16 wk (depending on the dose of occludin peptide used), the morphology of the seminiferous epithelium was indistinguishable from that of control rats, showing full recovery from occludin peptide-induced damage. The fact that testes recovered almost completely illustrates that spermatogonia were not destroyed. Occludin peptide-induced germ cell loss was also accompanied by a reversible disruption of the blood-testis barrier when assessed by micropuncture technique that quantified [125I]BSA in rete testis and seminiferous tubule fluids after administration of [125I]BSA through the jugular vein (43). Moreover, recovery of the blood-testis barrier coincided with the recovery of the seminiferous epithelium when examined histologically (43). It is likely that this peptide exerts its effects by interfering with the homotypic interactions that exist between two occludin molecules in apposing Sertoli cell membranes, thereby disrupting blood-testis barrier function. Once the peptide is metabolized and/or cleared, the blood-testis barrier reseals. Thus, these results illustrate that occludin peptide-induced reversible damage of the blood-testis barrier can be a useful in vivo model to study the dynamics of the blood-testis barrier if adequately characterized.
|
|
) (for reviews, see Refs.182, 183, 184). In Sertoli cells cultured in vitro, for instance, the mRNA level of TGF-ß3 was found to decrease approximately 2-fold during the assembly of the tight junction barrier from d 1–3, which coincided with a transient but significant increase in the expression of occludin and ZO-1 (22). However, during the maintenance phase of the tight junction permeability barrier from d 5–8, the level of TGF-ß3 increased. The reason for this increase is not immediately known, but it is likely that TGF-ß3 is not needed for the maintenance of the Sertoli cell tight junction barrier.
|
). These observations were verified when an in vivo model of blood-testis barrier disruption induced by cadmium chloride was used in which 4-(4-fluorophenyl)-2-(4-hydroxyphenyl)-5-(4-pyridyl)1H-imidazole (SB202109, a specific p38 MAPK inhibitor) could effectively block the cadmium chloride-mediated loss of occludin from the site of the blood-testis barrier (38). This illustrates that the dynamics of the blood-testis barrier are regulated, at least in part, by TGF-ß3 via the p38 MAPK pathway.
|
can also perturb Sertoli cell tight junction barrier function by affecting the production of occludin, collagen, matrix metalloprotease-9 (MMP-9), and tissue inhibitor of metalloproteases-1 (TIMP-1) (163). It was postulated that TNF- induces the production of Sertoli cell collagen 3(IV), MMP-9, and TIMP-1, as well as promoting the activation of pro-MMP-9. Because activated MMP-9 cleaves the collagen network in the basement membrane, the Sertoli cell tight junction barrier is perturbed. This creates a negative feedback loop that causes TNF- to induce the synthesis of collagen and TIMP-1, which are needed to reassemble the disrupted collagen network and tight junction barrier (163). TNF- may also limit the activity of MMP-9. As such, this cytokine likely regulates collagen, MMP-9, and TIMP-1 which collectively maintain the integrity of the blood-testis barrier in vivo (163).
Many other cytokines and growth factors have also been investigated with regard to their effects on tight junction barrier assembly and disassembly. However, these studies were all performed in nongonadal cell types and cell lines. For instance, cytokines such as TGF-, epidermal growth factor (EGF), and IL-1 have all been documented to affect tight junction barrier function in different epithelial cell types and cell lines in vitro (Table 3). Such a disturbance may be mediated via: 1) a redistribution and/or shut-down in the production of tight junction proteins, or 2) remodeling of the actin cytoskeleton (for reviews, see Refs.182 and 184). If adequately expanded, these studies should shed light on the regulation of tight junction dynamics in the testis.
Interestingly, a recent study has shown that nitric oxide (NO) also plays an important role in the regulation of Sertoli cell tight junction dynamics via its effects on soluble guanylate cyclase and protein kinase G (PKG). For instance, significant decreases in the expression of inducible (iNOS) and endothelial NO synthase (eNOS) were detected during the assembly of the Sertoli cell tight junction permeability barrier in vitro, seemingly suggesting that NO can down-regulate tight junction barrier function (30). Indeed, the inclusion of zinc II protoporphyrin-IX (ZnPP, a specific inhibitor of NOS that can also inhibit soluble guanylate cyclase) into Sertoli cell cultures facilitated tight junction barrier assembly, whereas the presence of 8-bromo-cyclic GMP (cGMP) perturbed tight junction barrier function (30). Furthermore, C29H25N3O5 (KT-5823, a specific inhibitor of PKG) could also facilitate Sertoli cell tight junction barrier assembly (30). Lastly, iNOS and eNOS were both shown to physically interact with occludin, but not with N-cadherin and nectin-3 as demonstrated by coimmunoprecipitation (30). Collectively, these data have demonstrated that NO regulates Sertoli cell tight junction dynamics via the NOS/soluble guanylate cyclase/PKG signaling pathway. Needless to say, these data also illustrate that blood-testis barrier dynamics are under the regulation of multiple variables, some operating independently and others cooperatively.
e. Additional comments.
After nearly three decades of study by reproductive biologists, a much finer physiological distinction has arisen between the blood-testis barrier (e.g., occluding zonule) and Sertoli cell tight junction (e.g., occluding macule) based on the fact that the assembly of the blood-testis barrier and Sertoli cell tight junction are not synchronous events in vivo (213, 214, 215). The occluding zonule, which separates the seminiferous epithelium into basal and adluminal compartments, is the basis of the blood-testis barrier. The fibrils that compose the occluding zonule, or blood-testis barrier, are continuous, encircling each Sertoli cell and forming an impermeable barrier. In the rat, the blood-testis barrier forms by approximately 15–18 d of age (for reviews, see Refs.16 and 116). The occluding macule, on the other hand, is constituted by a focal seal through which interstitial fluids or tracers can readily pass. Morphological studies have shown that the fibrils that constitute the occluding macule in the testis are discontinuous and situated immediately above and below the occluding zonule (213, 214, 215). It must be noted that in other epithelial cell types and cell lines, such as MDCK cells, the occluding zonule is synonymous with the tight junction.
3. Assisting in germ cell translocation.
Sertoli cells are directly involved in the movement of developing germ cells because these cells lack an architecture characteristic of migrating cells, such as fibroblasts (for review, see Ref.3). A possible mechanism underlying germ cell movement may involve members of the GTPase superfamily, as shown in a series of recent studies from our laboratory (45, 50). Moreover, other studies have shown that the translocation of elongating/elongate spermatids across the seminiferous epithelium is largely conferred by the ectoplasmic specialization (Section III.A.3) (for reviews, see Refs.20 , 61 , and 65). It has been shown that microtubules, together with several motor proteins such as dynein (216, 217) and myosin VIIa (218), ATPases, and GTPases (for review, see Ref.219) present at the site of the ectoplasmic specialization assist in the migration of elongating/elongate spermatids across the seminiferous epithelium. This is analogous to a train (e.g., elongating/elongate spermatid) on a railroad track (e.g., microtubule) whose motion is driven by an engine (e.g., dynein, myosin VIIa). Because this hypothesis has been recently reviewed (for reviews, see Refs.20 and 216), it will not be discussed any further. However, many questions still remain unanswered. For instance, if the movement of germ cells must coincide with their developmental status, then what signal(s) dictates the timely movement of spermatids across the epithelium during the epithelial cycle? What is the mechanism(s) that regulates the restructuring of the ectoplasmic specialization? Do germ cells signal to Sertoli cells when they need to be lifted upward? To answer these questions, the biochemical architecture of the ectoplasmic specialization must first be completely elucidated.
Sertoli cells also assist in the translocation of early meiotic spermatocytes from the basal to the adluminal compartment during the epithelial cycle in the seminiferous epithelium. This process, which likely requires the disassembly of Sertoli-Sertoli and Sertoli-germ cell adherens junctions, is a highly selective process. For instance, type A spermatogonia cannot be lifted but must remain attached to the basement membrane, suggesting that an intriguing mechanism must exist to determine which specific cell types can be lifted from the basement membrane to the basal compartment.
4. Spermiation.
Sertoli cells are indispensable to the release of mature spermatids from the seminiferous epithelium. This process involves a cascade of events, some of which include: 1) encapsulation of spermatid heads by Sertoli cell cytoplasmic processes; 2) expulsion of spermatids from Sertoli cell crypts; and 3) release of spermatid heads (for reviews, see Refs.3 and 6).
5. Phagocytosis.
After the release of mature spermatids from Sertoli cell cytoplasmic crypts, Sertoli cells phagocytose the residual bodies that were released from spermatids and any germ cells that may have degenerated during spermatogenesis (220, 221, 222, 223). In this respect, Sertoli cells function as macrophages by maintaining the integrity of the seminiferous epithelium.
6. Pinocytosis and receptor-mediated endocytosis.
In addition to phagocytosis, studies have shown that Sertoli cells undergo active pinocytosis and receptor-mediated endocytosis in the adluminal and basal compartments of the epithelium, respectively (222, 223, 224).
7. Secretory function.
Sertoli cells are secretory cells. This is backed by the observation that approximately 15% of the total proteins produced by Sertoli cells are glycoproteins that are destined to be secreted (for reviews, see Refs.9 , 35 , 62 , 63 , 132 , 133 , 225 , and 226). As part of an attempt by many investigators to understand the many functions of Sertoli cells, in particular the relationship between these cells and germ cells, numerous research articles were published that identified and characterized an array of proteins; these include 1) proteases, 2) protease inhibitors, 3) hormones, 4) energy substrates, 5) growth factors, 6) paracrine factors, and 7) extracellular matrix components, as Sertoli cell secretory products (for reviews, see Refs.9 , 11 , 62 , 63 , 132 , 133 , 134 , 225 , and 227, 228, 229, 230). Brief descriptions on some of these proteins as they relate to germ cell movement are found below.
a. Proteases and protease inhibitors.
Sertoli cells synthesize and secrete proteases and protease inhibitors, which participate in nearly all cellular processes involving tissue maintenance, repair, growth, and development. Proteases and their inhibitors have also been implicated in the events of germ cell movement (for reviews, see Refs.225 , 231 , and 232). For example, it was postulated that cathepsin L plays a role in spermiation because its expression was highest in stage VII before germ cell release when compared with other stages of the epithelial cycle (47, 233, 234). However, subsequent studies have shown that the mRNA levels of cathepsins D (an aspartic protease) and S (a cysteine protease) peaked in stages VII–IX and VII–VIII, respectively (47). Thus, if cathepsin L is indeed required to cleave Sertoli cell-elongate spermatid junctions at spermiation, then what are the roles of cathepsins D and S? Needless to say, each protease or protease inhibitor likely works with other molecules so that a unique physiological effect relating to spermatogenesis and germ cell movement can be executed. Parvinen’s group also showed enhanced secretion of plasminogen activator (PA, a serine protease) in stages VII–VIII of the epithelial cycle (235), suggesting that this protease may also be required to cleave Sertoli-germ cell junctions to permit for the release of elongate spermatids into the tubule lumen. In yet another study from our laboratory, it was demonstrated that there was a transient but significant increase in overall serine protease activity during the adhesion of germ cells to Sertoli cells in vitro, but preceding the formation of Sertoli-germ cell junctions (5). Taken collectively, these results suggest that proteases need to be activated before junction disassembly (47, 233, 234) and assembly (5).
Studies in which changes in protease inhibitors correlated with junction assembly and/or disassembly in the testis include cystatin C (236) and 2-macroglobulin [2-MG (237)]. In both cases, the localizations of these protease inhibitors were reduced in stage VII when proteases such as urokinase-type PA (u-PA) and cathepsin L were highest (Refs.5 and 233 ; for review, see Ref.225), illustrating the existence of a ying-yang relationship between proteases and protease inhibitors. Furthermore, it was shown that 2-MG associated tightly with elongating spermatids in stages I–VI of the epithelial cycle, apparently protecting the epithelium from unwanted proteolysis (237). These results also illustrate the significance of protease inhibitors in germ cell movement in the seminiferous epithelium.
b. Growth, autocrine, and paracrine factors.
Growth, autocrine, and paracrine factors are secretory molecules that bind to cell surface receptors to induce a cascade of signal transduction events that affect cell growth, differentiation, and function (for reviews, see Refs.13 , 14 , 238 , and 239). In the testis, these factors play a crucial role in the maintenance of spermatogenesis and germ cell movement (for reviews, see Refs.240, 241, 242, 243, 244, 245, 246, 247). For instance, TGF-, a member of the EGF family (for reviews, see Refs.248, 249, 250) and putative Sertoli, germ, and peritubular myoid cell product (251), stimulates the proliferation of prepubertal Sertoli cells (252). On the other hand, several different TGF-ß isoforms are known to exist (for reviews, see Refs.253, 254, 255, 256, 257), and the presence of TGF-ß1, -ß2, and -ß3 (23, 258, 259) and their receptors (260, 261, 262) have all been demonstrated in the testis. Interestingly, the TGF-ßIII receptor, also known as betaglycan (for reviews, see Refs.253 and 255), was recently shown by Vale and colleagues to bind inhibin (Refs.263 and 264 ; for review, see Ref.265). Moreover, it appears that TGF- and -ß elicit opposite effects on cell division and differentiation in several different cell types, such as granulosa (266) and Leydig (267, 268) cells, suggesting that these cytokines may work synergistically but have opposite effects in regulating specific cellular events.
c. Extracellular matrix components.
It has been known for several decades that the extracellular matrix plays an extremely important role in cellular movement, such as in the invasion and metastasis of malignant cells. However, its participation in germ cell movement is less clear. In the testis, the extracellular matrix at the basal lamina surrounding seminiferous tubules constitutes the basement membrane, which ultimately associates with Sertoli cells and developing germ cells (Fig. 1). It is largely composed of laminin, type IV collagen, heparan sulfate proteoglycan, and entactin (for review, see Ref.21). Sertoli cells are known to secrete components of the extracellular matrix, such as collagen (269, 270, 271, 272) and laminin (270, 272), that contribute to the structural integrity of the seminiferous epithelium, whereas peritubular myoid cells produce fibronectin (269). Indeed, a previous study has shown that laminin is crucial in maintaining the integrity of the Sertoli cell tight junction and blood-testis barriers (161). Recently, collagen was also shown to play an important role in regulating the Sertoli cell tight junction barrier in vitro, possibly via an interaction with TNF-, MMP-9, and TIMP-1 (163). Indeed, it has been shown that TNF- can stimulate Sertoli cell collagen 3(IV) production by as much as 12-fold vs. controls without TNF- (163) and that germ cells and testicular macrophages are the major source of TNF- in the testis (273). Collectively, these data illustrate that Sertoli, germ, and peritubular myoid cells contribute to the making and maintenance of the extracellular matrix, which in turn affects junction dynamics.
8. Provision of nutrients.
Sertoli cell products, many of which are important for germ cell survival, need to be synthesized and efficiently delivered in sufficient quantities to spermatocytes and spermatids residing behind the blood-testis barrier. These products include amino acids, carbohydrates, lipids, vitamins, and metal ions. This transfer of nutritive products from Sertoli to germ cells is largely made possible because an intricate relationship exists between these two cell types (for review, see Ref.62). This is exemplified best when one isolates and then attempts to culture germ cells for an extended period of time because these cells are incapable of survival past 24 h (for review, see Ref.274). The finding that stem cell factor, a putative Sertoli cell product, can prevent germ cell apoptosis in vitro (275), suggests that germ cell survival is largely dependent on Sertoli cells. Indeed, earlier studies have shown that germ cells can survive up to approximately 8 d in vitro when cocultured with Sertoli cells (276). Other factors that can enhance the viability of germ cells both in vivo and in vitro include hormones [e.g., testosterone (277), dihydrotestosterone, and estradiol (278)] and cytokines [e.g., TNF- (279)].
a. Sertoli cell metabolism.
The metabolism of Sertoli cells continues to be poorly understood because many of the investigations that were conducted thus far are in vitro studies, and they may not reflect the actual in vivo condition (for reviews, see Refs.62 , 227 , and 228). For instance, the hormonal environment of Sertoli cells cultured in vitro is drastically different from that of in vivo because these cells are cultured in serum-free, chemically defined media. If serum is added into Sertoli cell cultures, results obtained from these experiments become misleading because the presence of the blood-testis barrier in vivo restricts access of serum proteins into the adluminal compartment. As such, the interpretation of these results in the context of the testis in vivo is often difficult, if not impossible. Nevertheless, morphological studies of Sertoli cells have shown that these cells have a large surface area, which remains in close contact with developing germ cells in all stages of the epithelial cycle. This is of great importance to the biology of the testis when one considers that a single Sertoli cell can sustain approximately 50 developing germ cells (71, 73). As such, the net output of energy is enormous, which is important not only to spermatogenesis, but also to germ cell development and movement (for review, see Ref.133). This is best exemplified in the observation that Sertoli cells can efficiently metabolize glucose into lactate (the preferred energy source of germ cells) and pyruvate (280, 281), which is likely channeled via gap junctions to germ cells to maintain their survival.
|
III. Sertoli-Sertoli and Sertoli-Germ Cell Interactions |
|---|
|
TopAbstractI. IntroductionII. Sertoli CellsIII. Sertoli-Sertoli and Sertoli...IV. Mechanism of Germ...V. Concluding RemarksReferences |
|---|
. In this section, these junctions and their constituent proteins as they relate to junction dynamics and germ cell movement are discussed in greater detail.
A. Cell junctions present in the testis
1. Occluding junctions.
Tight junctions are the only known example of occluding junctions (for reviews, see Refs.282, 283, 284, 285) (Figs. 1 and 6 and Table 1). In the testis, the tight junction is an area of close contact between plasma membranes of neighboring cells composed of fibrils, usually 50–100 in number, that completely encircle the basal domain of each cell. This divides the Sertoli cell, forming a basal and an adluminal compartment. Although the biochemical composition of these strands is not entirely known, it is believed that each of these fibrils represents a row of tightly arranged transmembrane proteins, such as occludin and claudin, that seal the paracellular space (for reviews, see Refs.282, 283, 284, 285). Unlike tight junctions found in other epithelia, in the testis these junctions have a unique location in that they are closest to the basement membrane, which is a modified form of the extracellular matrix (for reviews, see Refs.13 , 14 , and 16). In addition, the blood-testis barrier as constituted by Sertoli cell tight junctions is not assembled until 15–18 d of age in the rat testis, at which time a specialized microenvironment is created. This microenvironment is essential for germ cell development and movement.
a. Functions of occluding junctions.
Tight junctions (for reviews, see Refs.13 and 14) have two main functions; they 1) create a barrier, and 2) form a boundary. First, tight junctions create a semipermeable barrier, which restricts the passage of molecules. This is known as the barrier function of tight junctions. However, the passage of molecules depends largely on their molecular weight and chemical nature (for reviews, see Refs.144 and 286, 287, 288). In fact, tight junction permeabilities are known to vary among tissues so as to adjust to the physiological needs and functions of different organs (for reviews, see Refs.118 , 184 , and 289, 290, 291). Studies have also shown that tight junction permeability can be regulated by different factors and physiological conditions. For instance, sodium orthovanadate, a protein tyrosine phosphatase inhibitor, was able to perturb tight junction permeability in MDCK cells when the diffusion of inulin across the epithelium was examined (292), suggesting that an increase in protein tyrosine phosphorylation can perturb tight junction barrier function. These results were corroborated and expanded in a recent report from our laboratory which illustrated that inhibitors of protein tyrosine and serine/threonine phosphatases could perturb the assembly and maintenance of the Sertoli cell tight junction barrier (293). However, these observations warrant further investigation to complement earlier in vivo studies (125, 126, 127, 128).
Second, tight junctions in the testis form a boundary, dividing the Sertoli cell into a basal and an adluminal compartment, which are different from each other. This is known as the fence function of tight junctions (for reviews, see Refs.287 , and 294, 295, 296). This function also confers cell polarity in epithelia. Basically, the fence function of tight junctions prevents the intermixing of molecules, such as proteins and lipids, between the two domains. This is also an important feature for Sertoli cell function because these cells secrete several of their products such as androgen binding protein, testin, and transferrin in a polarized fashion (for reviews, see Refs.62 , 133 , 136 , and 226).
b. Models explaining the chemical nature of occluding junction strands.
Presently, there are two models that explain the chemical nature of occluding junction strands. In the first model, known as the lipid model, tight junction fibrils are proposed to be formed by micelles with the polar groups of the lipids directed inward and the hydrophobic tails immersed in the lipid matrix of the two plasma membranes (297, 298, 299). However, the lipid model appears to be inconsistent with currently available data. For instance, when the total composition of phospholipids, sphingolipids, and cholesterol in epithelial cells was altered, the appearance of tight junctions did not change. The barrier and fence functions of tight junctions were also not changed (300). However, another study reported that a rapid reduction in total cell cholesterol by methyl-ß-cyclodextrin resulted in a decline in TER, an increase in the influx of mannitol, and an increase in the number of tight junction particles associating with the E face [in freeze-fracture technique, the cytoplasmic face of the membrane is known as the protoplasmic or P face, and the exoplasmic is the E face (301)]. The validity of the lipid model has not yet been a subject of investigation in the testis.
In the second model, known as the protein model, integral membrane proteins associate with other integral membrane proteins (e.g., occludin + occludin, claudin 1 + claudin 2, claudin 1 + claudin 3) in the apposing plasma membrane of the adjacent cell. It is also possible that microscopic pores are present in these strands to permit for the passage of molecules across the tight junction barrier (302). Collectively, these results suggest that tight junction fibrils are composed of integral membrane proteins, but that their structure and functions are sensitive to rapid changes in the lipidic environment.
c. Proteins of occluding junctions.
The membrane-associated guanylate kinase (MAGUK) homolog protein family is composed of several members, including: 1) ZO-1; 2) ZO-2; and 3) ZO-3 (Fig. 6). They are known to: 1) regulate cell proliferation and differentiation; 2) control membrane organization; 3) manage signal transduction pathways; and 4) regulate cell polarity (for reviews, see Refs.285 and 303). All members of the MAGUK family possess three distinct domains: 1) a yeast enzyme guanylate kinase; 2) a src-homology region 3 (SH3); and 3) one or more postsynaptic density-95/Discs-large/ZO-1 (PDZ) domains, which are all important for protein-protein interactions (for reviews, see Refs.285 and 304, 305, 306). Proteins containing PDZ domains 1) coordinate the assembly of multiprotein complexes, 2) cluster tight junction proteins, and 3) recruit signaling molecules. Non-PDZ domain-containing proteins, on the other hand, can be recruited to the site of the tight junction by PDZ domain-containing proteins or interact directly with tight junction integral membrane proteins. For example, the recently identified multi-PDZ domain protein 1 (MUPP1) is known to have 13 PDZ domains (307, 308). This protein, which binds to claudin-1 and junctional adhesion molecule (JAM, another tight junction molecule) and colocalizes with ZO-1 in MDCK cells, may function by recruiting molecules to the site of tight junctions (308). These studies did not determine whether MUPP1 is present in the testis, but this is possible given the fact that MUPP1’s interacting partners, namely claudin-1 (309), JAM-1, JAM-2 (310), and ZO-1 (311, 312), are all found in the testis. Moreover, the SH3 domain of MAGUK proteins is known to interact with actin (for review, see Ref.313), illustrating that a relationship exists between membrane-associated proteins belonging to the MAGUK homolog family and the actin cytoskeleton.
The most recently identified MAGUK proteins are the MAGUKs with inverted orientation (MAGI) proteins. Studies on the first MAGI protein, MAGI 1, have revealed that it interacts with K-Ras (314), whereas MAGI 2 and 3 both relate with a protein known as phosphatase and tensin homolog deleted on chromosome 10 (PTEN) [a phosphatase and tumor suppressor; PTEN 2, a homolog of PTEN, is found only in the testis (315, 316, 317, 318)]. Furthermore, the expression of MAGI 3 was recently detected in the testis by PCR (315), whereas a study to determine whether MAGI 1 and 2 are found in the testis has not yet been performed. The significance of this observation as it relates to germ cell movement is currently not known, but studies that investigated the interplay of MAGI 3 with PTEN in other systems suggest that these proteins work coordinately to inhibit cell movement (for reviews, see Refs.319, 320, 321). This postulate is not proposed without ground because deletion of PTEN in a mouse fibroblast cell line led to increased cell motility likely mediated by Rac1 and Cdc42 (322), two small GTPases that regulate the dynamics of the actin cytoskeleton. What is also known is that the PDZ domain of MAGI 1 interacts with ß-catenin (323), suggesting that there is cross-talk between constituents of the tight and adherens junction via MAGIs. Although MAGIs appear to be potentially important for the function of tight, and possibly adherens junctions, little is understood regarding their role in junction dynamics.
i. ZO-1.
ZO-1, a 210- to 225-kDa peripheral membrane protein that associates with tight junctions in epithelial and endothelial cells, is believed to have scaffolding and signaling functions (for reviews, see Refs.285 and 324) (Fig. 6). ZO-1 is most notably known for its direct interaction with the C terminus of occludin at an occludin:ZO-1 stoichiometric ratio of 1:1 (325, 326). ZO-1 also associates with actin (326, 327) and -catenin (328, 329) in epithelial and nonepithelial cells, illustrating that ZO-1 can associate with adherens junctions. In another recent report by Gonzalez-Mariscal’s group (330), ZO-1 shuttled away from the site of the tight junction and moved into the nucleus when the tight junction in epithelial cells was compromised. Not surprisingly, this was mediated via the actin cytoskeleton. Thus, it is possible that ZO-1 has other cell functions, possibly of a signaling nature (e.g., involved in transcriptional regulation). In the testis, ZO-1 is localized almost exclusively to Sertoli cell tight junctions at the site of the blood-testis barrier consistent with an earlier report (312) (Fig. 7). Immunoreactive ZO-1 was also reported to be found at the site of Sertoli cell ectoplasmic specializations adjacent to early, but not late, elongating spermatids just before spermiation (312), suggesting its involvement in germ cell movement.
|
–, and 2) +, which are the result of alternative splicing of its mRNA (331, 332). These two isoforms play different roles in the assembly, maintenance, and regulation of tight junctions because they associate differently with tight junctions (331, 332, 333). For instance, in the testis ZO-1– and ZO-1+ are expressed exclusively during adulthood and puberty, respectively, suggesting that these isoforms are crucial to junction disassembly (e.g., germ cell movement) and assembly (e.g., blood-testis barrier formation) (332, 333). [The homolog of ZO-1 in Drosophila, Polychaetoid, sometimes referred to as Tamou or Tam, is also alternatively spliced, generating two isoforms that differ in localization and function, similar to that seen for ZO-1 in the testis. When antibodies specific for each isoform were generated, it was found that one isoform localized to the adherens junction, whereas the other one had a much broader distribution (334).]
In addition to occludin, ZO-1 can also associate with other proteins (Fig. 6). ZO-1 was found to interact with ZO-2 (335, 336), ZO-3 (335), ZO-1 associated protein kinase [a serine protein kinase (337)], ZO-1-associated nucleic acid-binding protein [ZONAB, a nucleic-acid binding protein/transcription factor (338, 339)], and connexin43 [a gap junction protein (340)]. Other proteins interacting with ZO-1, which were largely determined by coimmunoprecipitation and/or immunolocalization experiments, include F-actin (336), JAM (341), claudin-1 (342), cingulin (343), spectrin (344), partitioning-defective-3 [PAR-3, a cell polarity protein (345)], ALL-1 fusion partner from chromosome 6 [AF-6 (346)], Rab8 [a Rab GTPase (50, 347)], Rab13 (348), 19B1 (349), atypical protein kinase C (PKC) isotype-specific interacting protein [ASIP (350)], Go (351, 352), G2 (351), Gi2 (351, 352), Gs (352), G12 (353), absent, small, or homeotic 1 [ASH1 (354)], protein 4.1R (344), MAGI 1 (355), junction-enriched and -associated protein [JEAP; this protein was not found in the testis by immunoblotting (356)], peripheral myelin protein 22 (357), vasodilator-stimulated phosphoprotein (VASP) (358), proteins associated with Lin-7 [Pals; consists of two variants, Pals1 and Pals2, but thus far only Pals2 has been positively identified in the testis (359)], Crumbs homolog 1 [CRB1, which determines apical polarity (360)], Sec6, and Sec8 (361). It must be noted that the association of ZO-1 with many of these proteins in the testis has not yet been shown or may not even exist, but recent studies have shown that some of these proteins, such as Rab8B and Rab13, are likely to play a crucial role in regulating junction dynamics in the testis (50).
ii. ZO-2.
ZO-2, a 160-kDa protein first identified as a ZO-1 binding partner in MDCK cells (Ref.362 ; for reviews, see Refs.285 and 324), is present in the testis (363) (Fig. 6), but has not been studied extensively in this organ. A detailed characterization of ZO-2 has demonstrated that it shares a 51% homology with ZO-1 in the three MAGUK domains (364, 365), but only a 25% homology in the C terminus, suggesting that the latter region participates in determining the functional uniqueness of these two molecules. ZO-2 can be phosphorylated by PKC-ß, -, -
, and - isoforms (366), illustrating that it can be activated by protein kinases. Other in vitro and in vivo studies have demonstrated that ZO-2 binds to occludin (336) and claudin (342). Several reports have also described the association of ZO-2 with F-actin (336) and -catenin (367), alluding to the fact that this signaling molecule may also participate in the regulation of the adherens junction, similar to ZO-1.
iii. ZO-3.
ZO-3, first identified as a 130-kDa polypeptide in MDCK cells and designated as p130, was subsequently renamed as ZO-3 when it was found to share a high homology with both ZO-1 and ZO-2 (Ref.335 ; for reviews, see Refs.285 and 324). In MDCK cells it was shown that ZO-3, similar to ZO-1 and ZO-2, is a putative substrate of protein kinases and can be phosphorylated on serine and tyrosine residues (Refs.332 and 368 ; for review, see Ref.369). By immunofluorescence and immunoelectron microscopy, ZO-3 was shown to colocalize with ZO-1 at the site of tight junctions (335). It was also found to interact with occludin (335), claudin (342), actin (336), and Pals1 associated tight junction protein (360), but not with ZO-2 (335). Although its cDNA sequence has been available since 1998 (335), its presence in the testis has not yet been demonstrated.
iv. Occludin.
To date, the best-studied tight junction transmembrane protein is occludin (for reviews, see Refs.285 and 370, 371, 372, 373) (Figs. 4, 6, and 8). Occludin is a 60- to 65-kDa protein [often migrates on SDS-polyacrylamide gels as multiple bands as a result of its different phosphorylation states (374)] containing four transmembrane domains, two extracellular loops, and one intracellular loop. Both N and C termini exit into the cell cytoplasm (Figs. 4 and 6). The first extracellular loop of occludin is comprised of approximately 60% tyrosine and glycine residues, and it participates in cell-cell adhesion (375). On the other hand, the second extracellular loop participates in the formation of the paracellular barrier because the tight junction permeability barrier was shown to be altered when a synthetic peptide corresponding to the second extracellular domain was added into epithelial (376, 377) or Sertoli (43) cell cultures. Peptides corresponding to the first extracellular loop exhibited no effects on the tight junction barrier in epithelial cells (376, 377). Moreover, when a truncated form of occludin lacking the second extracellular loop was generated and expressed in occludin-null Rat-1 fibroblasts and MDCK cells, it failed to localize to the site of tight junctions (378), illustrating the significance of the second extracellular loop in targeting occludin to tight junctions. In line with the above functional studies, when chicken occludin was overexpressed in MDCK cells, the number of tight junction strands increased (379). This occurred concomitantly with a surge in TER (379), whereas in another important study, the depletion of occludin from tight junctions did not significantly affect cell morphology (380), suggesting that other tight junction proteins can substitute its structural role. In a follow-up study by the same group, occludin–/– male mice failed to produce litters with wild-type females, indicating that males were sterile (381). This observation is consistent with a recently published report from our laboratory which demonstrated that the intratesticular injection of a 22-amino acid synthetic peptide corresponding to the second extracellular loop of occludin resulted in reversible infertility (43). This is not entirely unexpected because this peptide induced aspermatogenesis via a disruption of the blood-testis barrier (Fig. 5) by interfering with the homotypic interactions between two occludins in apposing Sertoli cell membranes. This subsequently resulted in germ cell loss from the seminiferous epithelium because it is possible that Sertoli-germ cell adherens junctions were also perturbed. When the peptide was metabolized or cleared, the blood-testis barrier reassembled and spermatogenesis resumed (43). Taken collectively, these results illustrate that occludin is required for testis function, namely spermatogenesis and germ cell movement.
|
) (390). Interestingly, its localization in either species was not stage-specific (390), suggesting the importance of this protein throughout the entire epithelial cycle during spermatogenesis. On the other hand, an irregular staining pattern of occludin in Sertoli cells in 1-wk-old testes when the blood-testis barrier had not yet been formed was demonstrated by Cyr et al. (390). This is consistent with our results when immunofluorescence was performed using approximately 2-wk-old testes (our unpublished observations). Recently, a variant of occludin, occludin 1B, with an apparent Mr of 70 kDa was identified and found to be present in several mouse tissues, including the testis (391). Occludin 1B is different from occludin in at least three respects. Occludin 1B has a unique N-terminal sequence of 56 amino acids, and a 193-bp insertion that codes for a larger occludin molecule (391). Moreover, occludin 1B does not appear to be phosphorylated (391), suggesting that phosphorylation, although important, is not the only mechanism regulating tight junction dynamics. The precise roles of occludin and occludin 1B in germ cell movement are not yet known.
v. Claudins.
Presently, the claudin superfamily of tight junction transmembrane proteins of Mr 20–24 kDa consists of at least 24 members (for reviews, see Refs.58 , 370 , 371 , and 392, 393, 394, 395, 396). Claudins have been shown to localize precisely to the site of tight junctions (Fig. 6) as demonstrated by conventional freeze-fracture technique (397). This observation, together with the fact that tight junction strands could be reconstituted when the expression of a single claudin was induced in cultured fibroblasts (398), illustrates that claudins are one of the main building blocks of tight junctions. Although the topology of claudin has been shown to be similar to that of occludin, except that the cytoplasmic domain and second extracellular loop of claudin are significantly smaller, different claudin proteins do not exhibit any significant sequence homologies with occludin (for reviews, see Refs.370 and 393). In addition, the expression pattern of each claudin varies considerably among different tissues (for reviews, see Refs.58 , 392 , and 394). For example, claudin 11 is restricted largely to the testis and brain (397, 399). A 2.3-kb claudin 11 transcript was also barely detectable in the kidney (399). In addition, it is not known whether claudins are phosphorylated, yet the C termini of claudins are known to contain putative phosphorylation sites for PKC, casein kinase 2, and cAMP-dependent protein kinase (for review, see Ref.394). In fact, the greatest differences between the various claudins lie in their C termini, suggesting that their participation in tight junction dynamics in different organs may be related to their association with different cytoplasmic proteins (for reviews, see Refs.58 , 370 , 371 , and 392, 393, 394, 395). However, the last two amino acid residues at the C terminus, namely Tyr-Val, present in most claudin species are needed for their interaction with PDZ domains (342). Tsukita and Furuse (393) have also put forth a hypothesis stating that aqueous pores (or paracellular channels) that control ion selectivity exist within paired tight junction strands as was shown in transfection experiments using MDCK cells (400, 401). These studies illustrated not only that at any one time a cell can express more than two claudins, but also that the combination of different claudins between two cells determines the nature and possibly the physical size of the aqueous pores (for review, see Ref.58). This assigns unique properties to tight junctions in different cells and tissues. However, this model cannot explain how cells, such as preleptotene/leptotene spermatocytes, can pass through the Sertoli cell tight junction barrier.
In the testis, seven different claudins have been positively identified, namely claudins 1, 3, 4, 5, 7, 8, and 11 (for reviews, see Refs.58 , 392 , and 394). Claudin 11, the only claudin in the testis that has been studied in some detail, was shown to be restricted to the Sertoli cell at the site of the blood-testis barrier (399, 402). In the mouse testis, its expression coincided with the assembly of the blood-testis barrier at 6–16 d after birth (402), and its expression in Sertoli cells in vitro could be inhibited by FSH and TNF- (402). More importantly, gene knock-out experiments demonstrated that mice lacking claudin 11 are sterile (403), illustrating that the integrity of the blood-testis barrier is crucial for normal spermatogenesis.
vi. JAMs.
JAM, a member of the newly described Ig superfamily, was found to localize at the site of the tight junction in epithelial and endothelial cells (404). Unfortunately, it is not yet known whether JAM is an actual component of the tight junction. To date, three JAMs, namely JAM 1, 2, and 3, have been identified, and the expression of JAM 1 and 2 in the testis has been demonstrated by Northern blots (310). Studies have shown that JAM colocalizes with occludin (341, 405), ZO-1 (341, 406), and cingulin (341) (Fig. 6). Other reports have demonstrated that JAM associates with AF-6 (407), ASIP/PAR-3 (406, 408), MUPP1 (308), and calcium/calmodulin-dependent serine protein kinase (409). When an anti-JAM antibody was used in a functional study, the migration of leukocytes across endothelial cells was inhibited (404), whereas in another report, the inclusion of TNF- and IFN- resulted in the redistribution of JAM (410). This enhanced the migration of leukocytes across umbilical vein endothelial cells (410). JAM may also participate in homotypic cell adhesion (404), but a detailed investigation examining this characteristic has not yet been performed. Taken collectively, these results assign JAM a crucial role in cell movement, but it remains to be determined whether JAM facilitates or retards the movement of developing germ cells in the seminiferous epithelium. In view of recent findings that implicate JAM in cell movement, this family of tight junction integral membrane proteins should be vigorously studied in the testis.
vii. Coxsackievirus and adenovirus receptor (CAR).
Another recently identified tight junction protein (Table 1) is CAR, a 46-kDa integral membrane protein possessing a transmembrane, long cytoplasmic, and extracellular domain with two Ig-like regions. Although it is too premature to speculate on the function of CAR in tight junction dynamics, it is known that CAR mediates viral attachment and infection (411, 412). For instance, coxsackievirus and adenovirus are two enteroviruses that traverse the tight junction in epithelia (either the gastrointestinal or respiratory tract) during the course of infection, resulting in the loss of tight junction integrity, before disseminating to target organs such as the heart and brain. It is likely that these viruses gain entry into cells via CAR, which is similar to JAM [recently identified as a receptor for mammalian reoviruses (413)] and some claudins [can also function as receptors for Clostridium perfringens enterotoxin (414, 415)]. In nonpolarized cells, CAR localized to intercellular contacts and recruited ZO-1 to these sites, illustrating that CAR can promote cell adhesion (416). This observation was supported by additional results from the same group that illustrated the colocalization of CAR with N-cadherin in intercalated discs in cardiac myocytes (416), where ZO-1 was also found to be present (417). In polarized epithelial cells, CAR and ZO-1 colocalized at the site of the tight junction. This association was subsequently verified in coimmunoprecipitation experiments (416). The overexpression of CAR was even shown to increase barrier function (416). Surprisingly, CAR has been found to be expressed in germ cells in the mouse (418), but its presence in Sertoli cells remains to be determined.
viii. Others.
Several other tight junction-associated proteins have also been identified at the site of tight junctions, including those in the testis (Fig. 6 and Table 1). These include cingulin (343, 419), symplekin (420), spectrin/fodrin (344), 7H6 [also known as barmotin (421, 422, 423)], AF-6 (346, 424), PAR-6 (425), atypical PKC (350, 426), Sec 6, Sec 8 (361), Rab13 (348), Rab3B (427), ZONAB (338, 339), ASH1 [a transcription factor colocalizing with ZO-1 and cingulin (354)], Tax [a viral oncoprotein and transcription factor (428)], protein 4.1R (344), BG9.1 antigen (429), 220-kDa protein (430), 19B1 (349), ASIP (350, 431), MAGI 1 (314, 323), and WNK1 kinase (432). It must be mentioned that ZO-1 (433), symplekin (420), and cingulin (434) have all been localized to the nucleus as well. The presence of many of these proteins in the testis has not yet been determined.
Cingulin is a well-characterized, peripheral membrane protein that has two molecular variants of 108 and 140 kDa (419, 434, 435, 436). Studies have shown the 108-kDa protein to be a coiled dimer, sharing a high homology with cytoskeletal proteins and belonging to the myosin family. Indeed, cingulin was shown to interact with myosin II in pull-down assays (343). Cingulin is phosphorylated on serine residues in MDCK cells (437), and it is present in the testis (for review, see Ref.16). Symplekin is another tight junction-associated protein found in the testis at the site of Sertoli cell tight junctions (420). However, its expression, was also detected in other tissues, some of which do not even possess functional tight junctions. As such, it was labeled as a housekeeping gene in tissues lacking tight junctions (420). Both cingulin and symplekin are non-PDZ domain-containing proteins (for review, see Ref.324).
Another well-studied tight junction-associated protein is 7H6. The localization of this protein was found to be related to its phosphorylation status because the depletion of cellular ATP resulted in the dissociation of 7H6 from the tight junction (422). Moreover, hepatocyte growth factor (HGF)/scatter factor (SF), which induced the migration of MDCK cells, reduced the expression of 7H6, illustrating once again the participation of cytokines/growth factors in tight junction dynamics via their effects on tight junction-associated proteins (423).
Reports dealing with the role of AF-6 at the tight junction have thus far been arguable. A previous investigation has shown AF-6 to be a Ras effector protein, as determined by a study that characterized mutants of the Caenorhabditis elegans homolog Ce-AF-6 (438), yet AF-6 was also shown to associate with actin, occludin, and ZO-1 (346, 439). Furthermore, when a null mutation in the murine AF-6 locus was generated, homozygous mutant embryos died at 10 d post coitum, resulting from a loss of epithelial cell polarity and junctions (440). In addition, AF-6 interacts with nectin, a homophilic cell adhesion molecule belonging to the Ig superfamily (441). Preliminary observations from our laboratory have demonstrated the presence of AF-6 in the testis (83). Collectively, these results illustrate the potential significance of AF-6 in the function of the blood-testis barrier.
ix. Additional comments.
Less studied (at least, in mammals), but equally important, is cell polarity as it relates to the dynamics of tight junctions (for review, see Ref.442). It was recently shown that the polarity of a typical epithelial cell is initiated when that cell becomes integrated into an epithelium and that these events are largely mediated by cell adhesion. In the Drosophila embryo, removal of E-cadherin or ß-catenin resulted in not only the loss of cell adhesion but also cell polarity (443, 444, 445), suggesting that an intimate relationship between tight and adherens junctions exists. It must also be noted that cell polarity is not only a manifestation of the tight junction, but of the cytoskeleton as well, because forces that are responsible for cell shape must be transduced throughout the entire cell (for review, see Ref.442). Thus, the cytoskeleton also participates in the formation of spatially and functionally distinct domains within cells, which is an area of research receiving much attention lately through the study of GTPases, in particular Rabs and ADP-ribosylation factors (for reviews, see Refs.446, 447, 448, 449).
Investigators studying tight junction dynamics, in particular epithelial cell polarity, in the Drosophila are all familiar with the recently identified Crumbs, Stardust, Discs Lost, Scribble, Bazooka/Par-3, and DmPar-6 proteins (for reviews, see Refs.442 and 450). We will not discuss these in detail because they are outside the scope of this review, but will briefly comment because it is possible that counterparts of these tight junction proteins are likely to be found in mammalian tight junctions. This is based on a recent study that reported the existence of a human homolog of Drosophila Crumbs, designated CRB1, a tight junction integral membrane protein found in many vertebrate epithelia and proposed to function in the maintenance of cell polarity (Ref.451 ; for review, see Ref.452). By Northern blots, CRB1 was detected in the brain, but not in the testis (453). The human homolog of Drosophila Scribble (a member of the leucine-rich repeats and PDZ domains protein family functioning in the maintenance of cell polarity) has also been identified and designated Vartul (454). In MDCK cells, Vartul was found to colocalize with ZO-1 at the site of tight junctions (454), yet it is not known whether Vartul is present in mammalian tight junctions, such as those found in the brain, mammary gland, intestine, testis, and epididymis. Needless to say, if some of these tight junction polarity-conferring proteins are present at the site of the blood-testis barrier, which is likely the case, this area of research should be vigorously expanded in future studies.
x. Future perspectives.
The study of junction dynamics has matured in the sense that investigators in the field are interested in determining how the junctional complex is assembled into functional tight and adherens junctions. Several questions relating to the interrelationship between tight junctions and germ cell movement have arisen for which answers are needed. For instance, how does a typical epithelial cell perceive polarity/orientation so that tight and adherens junctions can be assembled properly? Is there a link between cell junction assembly/disassembly and endocytic machinery? To what extent is the underlying actin cytoskeleton involved? For the most part, cell junctions assemble in such a way that they can inherently differentiate between the apical and basolateral domain of an epithelial cell. Stated simply but eloquently by Lecuit and Wieschaus, "junctions can organize polarity" (for review, see Ref.455). The assembly of tight junctions in vitro also requires a drastic increase in membrane surface area, which ultimately results in a tall columnar epithelium (for review, see Ref.455). Moreover, it is known that functional adherens junctions must be assembled first before tight junctions can be assembled. This conclusion is supported based on an analysis of the spatial and temporal patterns of expression of tight junction proteins in early mouse embryonic development. For instance, in the compacting eight-cell stage embryo, immature apicolateral junctions contain ZO-1 and catenins (456). However, from the 32-cell stage and onward, ZO-1 is found in association with occludin and cingulin at tight junctions, separated from the adjacent cadherin-catenin adherens junction complex (456). This illustrates that adherens junctions are formed before tight junctions. How this sorting of proteins occurs remains unclear, but it is postulated that the recruitment of proteins into different complexes depends on the different binding affinities of each component protein.
2. Anchoring junctions.
There are four types of anchoring junctions: 1) adherens junctions (also known as zonula adherens); 2) focal contacts or adhesions; 3) desmosomes; and 4) hemidesmosomes (for reviews, see Refs.13 and 14), all of which are structurally and biochemically different from each other (Fig. 1 and Table 1). The major difference among these four junction types is that the adherens junction and focal adhesion use actin filaments as their attachment sites, whereas the desmosome and hemidesmosome utilize intermediate filaments. In the testis, two modified adherens junction types are also present, namely the ectoplasmic specialization and tubulobulbar complex, which are also discussed below.
a. Functions of anchoring junctions.
Although anchoring junctions differ from each other, they all have one main function. They link a cell’s internal machinery or cytoskeleton either to another cell or to the extracellular matrix, creating a network that maintains tissue integrity. Additional reports have suggested that anchoring junctions may also function in signal transduction events because some of the component proteins of these junctions are signal transducers that relay bidirectional signals to regulate basic cellular processes, such as cell proliferation, differentiation, and movement (for reviews, see Refs.457, 458, 459).
b. Proteins of adherens junctions.
Presently, four adherens junction functional units have been identified conferring actin-based cell adhesion. These are: 1) cadherin-catenin, 2) nectin-afadin-ponsin, 3) integrin-laminin, and 4) vezatin-myosin cell adhesion units (for review, see Ref.460) (Fig. 9). Only 1) to 3) have been positively identified in the testis, because it is not known whether vezatin is present in this organ.
|
). More than six dozen cadherin family members have been identified thus far. Most cadherins are glycoproteins composed of several domains: 1) two cytoplasmic, 2) one transmembrane, and 3) five calcium-binding domains (designated EC1–5). The most highly conserved region found in classical cadherins, the cell adhesion recognition sequence, namely His-Ala-Val (468), is harbored within the EC1 domain. This sequence is important to adherens junction function because synthetic peptides of His-Ala-Val or antibodies against this tripeptide could perturb cell adhesion in a variety of cell types, including Sertoli and germ cells in the testis (468, 469, 470). Specifically, adhesion conferred by cadherins is via calcium-dependent [calcium is known to rigidify the multidomain structure of cadherins (471, 472)], homotypic interactions with cadherins residing in the adjacent cell membrane. In the absence of calcium, cadherins are inactive and susceptible to proteolysis (473, 474). On the other hand, the importance of the cadherin cytoplasmic domain lies in its interaction with intracellular proteins. These proteins participate in the clustering of cadherins to form functional adherens junctions, as well as to relay information from the cell surface to the nucleus, in turn activating specific gene(s) so that the integrity of the epithelium can be maintained. Moreover, several studies have demonstrated that E-cadherin can mediate the assembly of desmosomes, tight and gap junctions (475, 476, 477), but the mechanism by which this is accomplished remains to be elucidated.
The function of cadherin also depends on the binding of its cytoplasmic domain to either ß- or -catenin (Ref.478 ; for reviews, see Refs.462 , 467 , and 479). -Catenin, a close relative of ß-catenin, is also known as plakoglobin (480, 481, 482) (Fig. 9). ß-Catenin has an additional function in that it can also behave as a transcriptional cofactor when stimulated by the Wnt signal transduction pathway (for reviews, see Refs.483, 484, 485). The cadherin-catenin complex is linked indirectly to the actin cytoskeleton by -catenin, illustrating that this catenin specifically recruits molecules that participate in actin dynamics to the site of the adherens junction (for review, see Ref.486). -Catenin, which is related to vinculin (487, 488, 489), also interacts with spectrin (490). Moreover, ß-catenin was able to colocalize with ezrin, a protein linking the cell membrane with the actin cytoskeleton (491) and TGF-ß type II receptor (TGF-ß1 RII) (492). The latter interaction was demonstrated in a recent study which showed that TGF-ß1 RII dissociated from ß-catenin when HK-2 cells were stimulated by TGF-ß1 (492), suggesting a possible link between TGF-ß1 and adherens junction dynamics.
Another recently identified catenin, p120ctn [formerly known as p120cas (cadherin-associated Src substrate); for review, see Ref.493 ], is a putative Src-substrate associating with cadherins (494) but not with -catenin (494), suggesting that p120ctn has a unique function in the cadherin-catenin cell adhesion unit. Another study has shown that p120ctn can induce the clustering of E-cadherin (495). p120ctn can be phosphorylated on both tyrosine and serine residues in response to a variety of growth factors such as EGF and platelet-derived growth factor (PDGF) (496, 497). It can also enter the nucleus, where it interacts with a newly identified transcription factor, Kaiso (498). Recent studies have shown that an initial interaction between E-cadherin and ß-catenin is required for eliciting the cascade of events leading to cell adhesion, but that an interaction between E-cadherin and p120ctn is necessary for the formation of stable adherens junctions (for review, see Ref.499). Moreover, changes in the phosphorylation status of p120ctn can affect cell adhesion (500). In general, a decrease in phosphorylation is associated with tighter cell adhesion, whereas an increase results in the loss of adhesiveness (for reviews, see Refs.500 and 501). Lastly, p120ctn also interacts with Fujinami poultry sarcoma/feline sarcoma kinase [a tyrosine kinase (54, 502)], BP180 [type XVII collagen (503)], and connexin431 (504).
Other proteins that have been shown to localize to the adherens junction include VASP [a protein that regulates actin polymerization and directly binds to vinculin (358, 505)] and mammalian Lin-7 [also known as Veli; a PDZ domain-containing protein that binds directly to ß-catenin (506) and localizes at the nectin-afadin-ponsin-based adherens junction (507)]. Mammalian Lin-7, which was shown to associate with several other proteins such as Pals1 and Pals2 (359), is also found at the tight junction in MDCK cells (508).
The key to cadherin and catenin regulation may be found in the study of GTPases. For example, the overexpression of p120ctn increased cell motility via its action on Rho GTPases (for reviews, see Refs.509 and 510). Several members of this family such as RhoA, Rac1, and Cdc42 have recently been shown to localize to the cadherin-catenin-based cell adhesion unit (for reviews, see Refs.511 and 512). Other studies have shown that during junction disassembly E-cadherin and ß-catenin are first internalized and then recycled back to the plasma membrane by Rab5 (513, 514), another member of the GTPase superfamily. Phosphorylation also appears to play a significant role in adherens junction dynamics as was shown in experiments in which tyrosine phosphorylation of ß- and/or -catenin resulted in the loss of cadherin-mediated cell adhesion (Refs.515, 516, 517 ; for review, see Ref.500).
In the testis, at least two dozen cadherins are known to be present, and all exhibit unique expression patterns (518, 519). For example, N-cadherin associated with the heads of elongating spermatids in a stage-specific manner (111) (Fig. 10), suggesting its involvement in germ cell movement. However, most of the N-cadherin was confined to the basal compartment of the seminiferous epithelium (50, 53, 111). In addition, the inclusion of an anti-N-cadherin antibody perturbed the adhesion of germ cells to Sertoli cells in vitro (469, 520). Moreover, a series of coimmunoprecipitation and chemical cross-linking experiments demonstrated that Sertoli cell N-cadherin interacts with actin and vimentin (to a lesser extent) (Ref.53 and our unpublished observations), indicating that the testis utilizes the cadherin-catenin actin-based cell adhesion unit. Other studies have shown that Sertoli-germ cell adherens junctions are also comprised of E-cadherin, -, ß-, and -catenins, p120ctn, Src [a protein tyrosine kinase (53)], Rho (45) and Rab GTPases (50). Taken collectively, these results suggest that the germ cell is equipped with similar adhesion and signaling proteins as its cellular counterpart, the Sertoli cell, in the seminiferous epithelium.
|
). It is important to recognize that the nectin-afadin-ponsin cell adhesion unit does associate with the cadherin-catenin system and that the two units function cooperatively to form functional adherens junctions, such as those found in the testis (Ref.521 ; for review, see Ref.460). Nectin [originally known as poliovirus receptor-related protein, but later renamed as nectin (441)] is a calcium-independent cell adhesion protein of the Ig superfamily consisting of three domains: 1) a transmembrane, 2) cytoplasmic, and 3) three extracellular Ig-like domains (522, 523). Thus far, at least five nectins, namely nectins-1, -2, -3, -4, and poliovirus receptor (some nectins such as nectin-1 can also act as poliovirus or -herpesvirus receptors) have been identified (460, 524, 525). Assembly of the nectin-afadin-ponsin cell adhesion unit is mediated in part by each nectin family member interacting homo- or heterotypically with another nectin molecule on the apposing cell membrane (441, 524, 526, 527, 528). In fact, the heterotypic interaction of nectin-3 (e.g., nectin-3 + nectin-1) was shown to be stronger than its homotypic interaction (524).
In the testis, the presence of nectin-1 (weak expression), -2 (moderate expression), -3 (strong expression), and -4 (weak expression) has been demonstrated by Northern blots (524, 528). Specifically, nectin-2 was found in Sertoli, germ, and Leydig cells, whereas nectin-3 was restricted to spermatids (529). In the mouse testis, nectin-2 associated with Sertoli cells and elongate spermatids colocalizing with F-actin, which was consistent with its localization at the site of apical ectoplasmic specializations. Its level was also highest in stages VI–VIII of the epithelial cycle (530), suggesting a role in germ cell movement. More importantly, nectin-2–/– mice displayed a defective sperm morphology and were infertile (530).
Afadin has two variants: 1) l-afadin (large splicing variant of AF-6 protein localized at adherens junctions), and 2) s-afadin (small splicing variant of l-afadin). The expression of l-afadin is ubiquitous, whereas s-afadin is found only in neural tissues (439). l-Afadin is an F-actin binding protein whose PDZ domain binds to the cytoplasmic domain of nectin (439, 441). In another recent report, -catenin was shown to contain the entire binding site for l-afadin, illustrating that there is cross-talk between the cadherin-catenin and nectin-afadin-ponsin cell adhesion units (for review, see Ref.525). In fact, -catenin is required for the nectin-afadin-ponsin complex to localize to the adherens junction (531). Moreover, when MDCK cells were cultured in calcium-depleted media, E-cadherin, but not l-afadin, was rapidly endocytosed possibly via Rab GTPases (532). Mouse embryos that lack l-afadin die at approximately 10 d post coitum, resulting from a disruption of tight and adherens junctions in neuroectodermal cells (440, 533), suggesting that this protein is important in tight and adherens junction dynamics. l-Afadin is also known to associate with the cytoplasmic protein ponsin (534). On the other hand, s-afadin lacks the actin binding domain that is characteristic of l-afadin (439, 441). Human s-afadin has also been shown to be identical to AF-6 (535). Presently, there is not a single report in the literature examining the function of afadin in the testis.
Ponsin, the third member of the nectin-afadin-ponsin cell adhesion unit, interacts with either l-afadin or vinculin (489, 536, 537, 538, 539) to form a complex. It has even been proposed that vinculin is the protein linking the cadherin-catenin and nectin-afadin-ponsin cell adhesion units together (534). Ponsin has remained relatively uncharacterized since being identified in 1999, but it has been found to be present in the testis by Northern blots (534). However, the significance of the nectin-afadin-ponsin cell adhesion unit as it relates to germ cell movement in the testis remains to be determined.
iii. Vezatin-myosin VII.
The vezatin-myosin VII complex also confers actin-based cell adhesion. Vezatin is a novel transmembrane protein that bridges either myosin VIIA or the cadherin-catenin complex to the actin cytoskeleton (540, 541). Vezatin has three domains: 1) a short extracellular, 2) transmembrane, and 3) long intracellular domain that binds to myosin VIIA. Myosin VIIA, a 250-kDa motor protein, was shown to be present in a variety of organs such as the brain, intestine, liver, kidney, and testis (542, 543). In the testis, myosin VIIA is highly expressed, associating with Sertoli-germ cell ectoplasmic specializations (544). Additional reports have implicated motor proteins in other processes, such as in vesicle motility and signal transduction (for reviews, see Refs.545, 546, 547). Due to the inherent property of myosin VIIA in regulating actin dynamics, it is likely that this protein participates in the turnover of actin filaments at the site of the ectoplasmic specialization. Myosin VIIA also associates with protein kinase A (PKA) (540), suggesting that it may be regulated by protein kinases. Furthermore, vezatin and myosin VIIA were both recruited to the adherens junction during its assembly in MDCK cells (540, 541). Vezatin was also shown to coimmunoprecipitate with - and ß-catenin (540, 541). Thus, it was postulated that vezatin is recruited to adherens junctions via its interaction with -catenin, which is followed by the enlisting of myosin VIIA to these sites. The presence of vezatin in the testis has not yet been determined.
iv. Integrin-laminin.
Two studies by Salanova et al. (548, 549) have linked 6ß1 integrin to ectoplasmic specializations between Sertoli cells and between Sertoli cells-elongate spermatids, illustrating the existence of yet another actin-based adherens junction functional unit (Fig. 9). However, these findings presented a biological dilemma in that integrins, traditionally known to be cell adhesion molecules between cells and the extracellular matrix, interact with collagens and laminins (for reviews, see Refs.463 and 550) in the extracellular matrix at the site of the focal adhesion and hemidesmosome. If 6ß1 integrin is found in Sertoli cells, then what is the binding partner of this adhesion molecule in germ cells? Interestingly, a recent study has shown the presence of the 3 laminin chain of laminin 12 [laminins are glycoproteins composed of an -, ß-, and -chain (551, 552)] in the testis, localized to the adluminal compartment of the seminiferous epithelium, but not to the basement membrane (553). This observation represents the first nonbasement membrane laminin chain found in any epithelium. Taken collectively, these results suggest that the 3 laminin chain may be the binding partner of 6ß1 integrin in the ectoplasmic specialization. However, this is largely a speculation that requires further investigation. Moreover, the identities of the other - and ß-chains needed to contribute to a functional laminin receptor that can interact with 6ß1 integrin are still not known.
3. Ectoplasmic specializations.
Specifically, the ectoplasmic specialization is a modified adherens junction type found only in the testis. It uses actin filaments as its attachment site (for reviews, see Refs.3 , 18 , 20 , 61 , and 87) (Fig. 9). The ectoplasmic specialization is important to spermiation, the translocation of spermatids, and possibly in the stabilization of other junction types present between Sertoli and Sertoli-germ cells, such as tight and desmosome-like junctions (for reviews, see Refs.3 , 7 , 16 , 18 , and 20). In the rat testis, the ectoplasmic specialization is first seen facing spermatids when they begin to elongate [step 8 of spermiogenesis (554)], and they disappear from the seminiferous epithelium at spermiation. It must also be noted that this junction type confers adhesion not only between Sertoli cells and elongating/elongate spermatids, but also between Sertoli cells (3, 20, 554, 555). In addition, the separation of elongating/elongate spermatids (step 8 and beyond of spermiogenesis in the rat) from Sertoli cells requires the use of trypsin (554) or cytochalasin D (18), illustrating the stability of this testis-specific anchoring junction. Basically, ectoplasmic specializations prevent the premature release of elongating/ elongate spermatids into the tubule lumen because desmosome-like junctions are no longer present between Sertoli cells and spermatids during this time (18). It has even been suggested that the ectoplasmic specialization replaces the desmosome-like junction as the anchoring junction type between Sertoli cells and elongate spermatids (18), yet it is not physiologically known why one junction type would be replaced by another with essentially the same function. It is possible that ectoplasmic specializations also participate in positioning elongating/elongate spermatids with an identical orientation in the epithelium because elongating/elongate, but not round, spermatids are highly polarized cells. Moreover, when elongate spermatids are embedded within Sertoli cell crypts, which is accompanied by drastic changes in the position and shape of spermatid heads, ectoplasmic specializations constantly remodel themselves to adapt to these morphological changes (Refs.2 , 556 , and 557 ; for reviews, see Refs.6 , 15 , and 16).
Although the precise biochemical constituents of the ectoplasmic specialization are not entirely known, recent studies have shown that there are two types of ectoplasmic specializations, designated apical and basal. The best-studied apical ectoplasmic specialization protein is 6ß1 integrin (548, 558, 559). Another recent study has shown that 4ß1 integrin (560) is another likely molecule present at the apical ectoplasmic specialization, but it must be noted that 4 integrin is also found at various sites in the seminiferous epithelium including the basal ectoplasmic specialization (560). Although 6ß4 integrin has been postulated to be a constituent protein of hemidesmosome junctions between Sertoli cells and the basement membrane (559), the constituent proteins at basal ectoplasmic specializations between Sertoli cells at the blood-testis barrier in the seminiferous epithelium of the rat testis are not characterized as well as those of apical ectoplasmic specializations. In addition, the nectin-afadin-ponsin complex is apparently restricted to apical ectoplasmic specializations (529), whereas the cadherin-catenin complex is more predominant in basal ectoplasmic specializations (53, 560, 561). Indeed, recent studies using both semiquantitative RT-PCR and immunoblotting have confirmed the presence of E-cadherin and ß-catenin in highly purified germ cells at a 1:1 equimolar ratio (53), whereas N-cadherin was shown to be more predominant in Sertoli cells (53). This was followed by studies that used immunofluorescent microscopy to confirm the presence of the cadherin-catenin complex in the seminiferous epithelium at the site of the apical ectoplasmic specialization (83, 561). This is also consistent with an earlier report that demonstrated the presence of N-cadherin at sites of Sertoli-Sertoli and Sertoli cell-spermatid contact in the rat seminiferous epithelium (561). Moreover, the N-cadherin-catenin complex likely associates with desmosome-like junctions because N-cadherin was shown by immunofluorescent microscopy to colocalize with plectin, an intermediate filament protein (561). This also strengthens another earlier morphological study which illustrated that there is cross talk between ectoplasmic specializations and desmosome-like junctions (559), since studies by coimmunoprecipitation have repeatedly demonstrated that the cadherin-catenin complex in the seminiferous epithelium is similar to cadherin-catenin complexes found in other tissues in that it uses actin filaments as its attachment site (53, 83). However, approximately 5% of cadherin associated with vimentin-based intermediate filaments (83).
Germ cells were also shown to express specific forms of nectin, such as nectin-3, which heterotypically associate with nectin-2 in the apposing germ or Sertoli cell plasma membrane (529). Collectively, these data illustrate that germ cells contribute to the protein complexes that confer cell adhesion to the ectoplasmic specialization. This is not entirely unexpected because ectoplasmic specializations between Sertoli and germ cells are constantly changing during spermatogenesis and spermiogenesis because of the continuous movement of germ cells across the seminiferous epithelium. As such, ectoplasmic specializations are probably using additional cell adhesion proteins to regulate their restructuring, which may not be found in adherens junctions. For instance, integrins are restricted to cell-matrix anchoring junctions in other epithelia, yet 6ß1 integrin is a putative structural protein at the site of apical ectoplasmic specializations. More recent studies have identified proteins, such as focal adhesion kinase (FAK), at the site of the ectoplasmic specialization (49, 559). However FAK regulates cell movement and is usually restricted to the focal contact (562, 563, 564). These results illustrate that the proteins that constitute apical and basal ectoplasmic specializations are different.
Other proteins that confer functionality to the ectoplasmic specialization include actin (for review, see Ref.7), -actinin (for review, see Ref.565), myosin VIIA (544), fimbrin (555), espin (566), vinculin (567, 568), paxillin (559), gelsolin (569), integrin-linked kinase [ILK (559 ], nectin (530), testin (Ref.57 ; for review, see Ref.4), and Kelch-like neurofilament (NF)-E2-related molecule (ECH)-associating protein 1 [Keap1, a negative regulator of the NF-E2-related factor 2 transcription factor (218, 570)] (Fig. 9). Keap1, an interacting partner of myosin VIIA (544) was found to be targeted to ectoplasmic specializations even in the absence of myosin VIIA, suggesting that Keap1 can associate with other molecules at the site of the ectoplasmic specialization (218). On the other hand, nectin was restricted to sites of contact between Sertoli cells and elongating/elongate spermatids in stages VII–VIII, consistent with its localization at the ectoplasmic specialization (530). In this context, it is also important to note that other junction types, such as tight and desmosome-like junctions, are present within the vicinity of basal ectoplasmic specializations (for reviews, see Refs.7 , 16 , and 18). Moreover, "free" ectoplasmic specialization structures (ectoplasmic specializations that are found within Sertoli cell plasma membranes but not in contact with another cell) are often found in the seminiferous epithelium (3). Although the significance of "free" ectoplasmic specialization structures is not immediately known, it is tempting to speculate that they are participating in the recycling ectoplasmic specialization protein complexes. Alternatively, they may be attachment sites that were leftover after germ movement. This illustrates not only the dynamic nature, but also the multifunctional role of the ectoplasmic specialization in the seminiferous epithelium. However, much research is needed to define the biochemical architecture of ectoplasmic specializations found between Sertoli-Sertoli and between Sertoli cell-elongating/elongate spermatids, as well as to understand their regulation.
Espin, consisting of a single polypeptide chain of Mr 110 kDa, is present in many organs including the brain, heart, liver, kidney, small intestine, spleen, thymus, and others, but it is found most abundantly in the testis, restricted to basal and apical ectoplasmic specializations (566, 571) (Fig. 11). Each espin molecule consists of three putative actin binding sites with an espin:actin stoichiometric ratio of approximately 1:20 (566, 572). Furthermore, it was estimated that each ectoplasmic specialization is composed of approximately 4.5 x 106 molecules of espin (572). By immunofluorescence microscopy, espin was found to localize to both basal and apical ectoplasmic specializations in the seminiferous epithelium, but in stages VI—VII, espin was found primarily at the site of the apical ectoplasmic specialization (Fig. 11), illustrating that it is a stage-specific protein. This is also consistent with earlier published reports (566, 571) that demonstrated espin to be an apical and basal ectoplasmic specialization-associated protein.
|
It is believed that the function of the tubulobulbar complex is to prevent the premature release of elongate spermatids into the tubule lumen (556). It is also possible that these structures facilitate the removal of cytoplasm from spermatids (573) because the volume of cytoplasm was reduced by as much as 70% when tubulobulbar complexes were present in the epithelium (573). Indeed, a recent study has shown that MN7, an acrosomal glycoprotein having periodic acid-Schiff reactivity, is incorporated into the tubulobulbar complex, supporting the notion that this junction type facilitates the removal of excessive acrosomal contents at spermiation (576). Furthermore, tubulobulbar complexes may internalize junctions [basal tubulobulbar complexes are intermingled between tight and gap junctions (574)] during germ cell movement (574, 577, 578).
5. Focal contacts.
Focal contacts, also known as focal adhesions or adhesion plaques, are actin-based cell-extracellular matrix anchoring junctions (Table 1).
a. Functions of focal adhesions.
Focal adhesions, comprised largely of integrins, connect a cell’s actin filaments to the extracellular matrix. They are found in virtually all epithelia. Focal adhesions are dynamic structures that regulate cell movement, such as in fibroblasts and macrophages, and are also known to function in signal transduction events (for reviews, see Refs.510 and 579, 580, 581, 582, 583). Although this junction type has yet to be investigated in the testis, a review of the focal adhesion is included here. This is because proteins that are usually found at the focal contact in other epithelia have now been shown to be present at the site of the ectoplasmic specialization. This suggests that these proteins may participate in cell adhesive function at the ectoplasmic specialization. As such, the following sections will serve as a guide for future investigators in the field.
b. Proteins of focal adhesions.
Integrins belong to a large family of ß heterodimeric transmembrane proteins (for reviews, see Refs.550 , 584 , and 585). Thus far, 18 -subunits and eight ß-subunits have been identified, and they can combine to form 24 different functional integrin receptor proteins, which fall into three basic categories: 1) ß1, 2) ß2, and 3) v. Some integrins only bind to one ligand (components of the extracellular matrix such as collagen, laminin, fibronectin, or vitronectin), whereas others bind to more than one ligand. In addition to providing a physical link between the cytoskeleton of a cell and the extracellular matrix, integrins also transduce bidirectional signals across the plasma membrane to regulate many aspects of cell behavior such as differentiation, proliferation, migration, and death. For instance, integrin cytoplasmic-domain-associated protein-1, which associates exclusively with ß1 integrin, was shown to act as a negative regulator of cell adhesion and migration (586, 587, 588). Moreover, integrin cytoplasmic-domain-associated protein-1 contains phosphorylation sites for PKC, cAMP-/cGMP-dependent kinases, and calcium/calmodulin-dependent protein kinase II and can be phosphorylated in response to cell adhesion (for review, see Ref.589).
i. Vinculin.
Vinculin, an adaptor protein localized to the focal adhesion in other epithelia, is functionally related to -catenin (Ref.590 ; for reviews, see Refs.591 and 592). Although -catenin is a constituent of the cell-cell adherens junction, vinculin can be found in both cell-cell and cell-extracellular matrix anchoring junctions. For instance, vinculin is localized largely to the E-cadherin-catenin complex, but not the N-cadherin-catenin complex in multiple epithelia (593). Vinculin and -catenin also function as tumor suppressors. More importantly, when the expression of vinculin was suppressed in 3T3 cells, adhesion was perturbed and motility increased (594). On the other hand, the overexpression of vinculin decreased cell movement (595, 596). Furthermore, vinculin appears to be important in the early phases of focal adhesion assembly because it was found within this structure 15 min after the binding of cells to the extracellular matrix (597, 598). It is also a substrate for tyrosine and serine/threonine protein kinases (599, 600). Vinculin has been shown to interact with -/ß-catenin (in cell-cell adherens junctions only; for review, see Ref.592), actin (537, 601), talin (537, 602), -actinin (603), paxillin (604), vinexin (605), and VASP (505, 606). Vinculin can also substitute -catenin to form functional adherens junctions (593). In the testis, vinculin is found in both apical and basal ectoplasmic specializations (567, 568) (Fig. 9). It is also detected in the cytoplasm of Sertoli cells (111, 571) (Fig. 3A).
ii. Talin.
Talin (Mr, 235 kDa) is another major structural component of the focal adhesion (for review, see Ref.589). Talin contains binding sites for actin [both G- and F-actin (607, 608)], vinculin (602), FAK (609), phospholipids (610), TES [shares a high homology with zyxin (611, 612, 613)], and layilin [a hyaluronan receptor (614)]. In addition, talin binds to the cytoplasmic tails of ß1, ß2, ß3, and ß7 (weakly) integrins (Refs.615, 616, 617, 618 ; for review, see Ref.589). Both talin and ß integrin are required for focal adhesion assembly (619). Moreover, microinjection of an antitalin antibody or talin antisense RNA perturbed adhesion, spreading, and migration of fibroblasts and HeLa cells (Ref.620 ; for review, see Ref.591), suggesting that talin and integrins facilitate cell movement. Talin has been found in the testis (621).
iii. Paxillin.
Paxillin (Mr, 68 kDa), an adaptor protein, associates exclusively with the cytoplasmic domain of 4 integrin (Ref.622 ; for reviews, see Refs.623 and 624). (Adaptor proteins recruit proteins to form a multiprotein complex at the site of the tight or adherens junction). Likewise, paxillin participates in cell spreading and movement (Ref.622 ; for reviews, see Refs.625 and 626). It can be tyrosine and serine/threonine phosphorylated in response to cell adhesion or growth factor stimulation (627, 628, 629, 630). Paxillin has been shown to interact with other proteins, such as FAK (631, 632, 633, 634), Src (635, 636), Crk (636, 637), C-terminal Src-kinase (Csk) (636), actopaxin [a newly identified focal adhesion protein that also interacts with actin (638)], paxillin kinase-linker with a Mr of 95 kDa (639, 640), p21 GTPase-activated kinase-interacting exchange factor [a guanine nucleotide exchange factor (639)], p21 GTPase-activated kinase (639, 641), protein-tyrosine phosphatase-Pro-Glu-Ser-Thr [a nonreceptor protein tyrosine phosphatase that dephosphorylates p130 Cas (Crk- and Src-associated substrate having a Mr of 130 kDa, Ref.639)], proline-rich tyrosine kinase-2 [a nonreceptor tyrosine kinase related to FAK and found in Sertoli and germ cells (642)], E6 oncoprotein of bovine papillomavirus type 1 (643), and E6 protein of the cancer-associated HPV-16 (644). In the testis, paxillin is an adaptor protein (49, 111, 559, 560) localized to the site of the ectoplasmic specialization (111). It associates with actin, p120ctn, cortactin, FAK, 6ß1 integrin, tubulin, and vinculin (559, 560).
iv. Others.
The transmembrane 4 (also known as TM4, TM4SF, or tetraspanins because they span the plasma membrane four times) superfamily is a group of ubiquitously expressed proteins associating with selected integrins (e.g., 3ß1, 4ß1, 5ß1, 6ß1, and 6ß4 integrins) (for reviews, see Refs.584 , 645 , and 646). At least 25 members have been identified to date. Tetraspanins have been implicated in the regulation of diverse cellular functions, such as cell proliferation and movement (for reviews, see Refs.584 and 645). For instance, tetraspanins have been localized to the periphery (e.g., lamellipodia) of migrating cells (647, 648), suggesting that this superfamily of proteins participates in cell movement. It is not yet known whether members of this protein family are present in the testis.
Other proteins that participate in the function of focal adhesions include VASP, which was found at the site of lamellipodia (Refs.505 and 649 ; for reviews, see Refs.650 and 651). VASP interacts with profilin [G-actin binding protein (652)] and zyxin (653). Zyxin (Mr, 82 kDa) is a phosphoprotein that localized not only to the adherens junction (Fig. 3B) [interacts with -actinin (486, 654, 655, 656)], but also to the nucleus [interacts with Vav, a nucleotide exchange factor (657)], illustrating that zyxin has other functions in addition to its role as an adaptor protein. Lastly, calpain, a calcium-dependent cysteine protease has been shown to localize to focal adhesions (658), where it can cleave integrins, talin, and FAK (659, 660). Studies using calpain inhibitors have suggested that this protein participates in the assembly (661) and disassembly of focal adhesions during cell migration (662).
v. Additional comments.
It is now well established that integrins transmit bidirectional signals that trigger events such as tyrosine phosphorylation (for reviews, see Refs.663, 664, 665). This activates downstream signal transducers (e.g., Rho GTPase, FAK), triggering a cascade of events to affect Sertoli-germ cell adherens junction dynamics (45, 49). For instance, it is known that Sertoli-germ cell adherens junction dynamics at the site of apical ectoplasmic specializations are likely regulated, at least in part, via the integrin/Rho-associated protein kinase (ROCK)/Lin-11, Isl-1, and Mec-3 (LIM) kinase/cofilin signaling pathway (45) because these signaling molecules were shown to be induced concomitantly with changes in their phosphorylation status when cell adhesion in the seminiferous epithelium was perturbed by AF-2364 (666, 667). This conclusion is further strengthened by the fact that pretreatment of rats with (R)-(+)-trans-N-(4-pyridyl)-4-(1-aminoethyl)cyclohexane-carboxamide, 2HCl (Y-27632, a specific ROCK inhibitor) could delay AF-2364-induced germ cell loss from the epithelium (45). Furthermore, AF-2364 was also shown to perturb Sertoli-germ cell adhesion at the site of the apical ectoplasmic specialization via the integrin/FAK/phosphatidylinositol 3 (PI 3)-kinase/p130 Cas/ERK signaling pathway (49), downstream of which is likely MAPK. Taken collectively, these results suggest that the integrin/ROCK/LIM kinase/cofilin and integrin/FAK/PI 3-kinase/p130 Cas/ERK signaling pathways are two putative pathways that regulate ectoplasmic specialization restructuring. Work is now in progress to investigate whether these two pathways merge via one of the crucial signaling molecules, such as PI 3-kinase, a specific GTPase, or an adaptor, such as p130 Cas.
Additionally, integrins mediate signaling events that induce mechanical forces across the plasma membrane, enabling cells to generate traction during movement and exert tension during extracellular matrix remodeling (668). As such, cell movement involves cycles of adhesion and de-adhesion (for reviews, see Refs.669 and 670). Several kinases, many of which have been shown to associate with ectoplasmic specializations, in particular FAK (49, 559), Src (49, 111), Csk (111), and ILK (559), have been shown to participate in integrin-mediated signaling. For instance, fibroblasts obtained from FAK–/– embryos migrate more slowly than those from FAK+/+ embryos. FAK–/– fibroblasts also had an increased number of focal adhesions (671). These results were corroborated in another study which demonstrated that the dephosphorylation of FAK by PTEN inhibited the assembly of focal adhesions and cell migration (672). Other proteins that were also shown to be required for the turnover of focal adhesions were vinculin, talin, and -actinin (562, 673, 674).
On the other hand, several extracellular factors such as thrombospondin, tenascin, and secreted protein, acidic and rich in cysteine (SPARC, participates in cell adhesion) have been described for their antiadhesive properties (for reviews, see Refs.675 and 676). These proteins promote the disassembly of focal adhesions, in turn stimulating cell migration (677, 678). Many of these proteins, such as SPARC (679, 680) and sertolin [sertolin is related to thrombospondin (26)] have been positively identified in the testis and implicated in the regulation of junction dynamics in the seminiferous epithelium. Growth factors have also been demonstrated to induce de-adhesion. For example, EGF promotes cell rounding, resulting in a decrease in adhesive strength and the disassembly of focal adhesions (681, 682, 683). This was apparently mediated via the MAPK pathway because a mitogen-activated protein/extracellular signal-regulated kinase kinase (MEKK) inhibitor decreased EGF-induced de-adhesion (683). HGF/SF and PDGF have also been demonstrated to negatively regulate adhesive strength (684, 685).
vi. Future perspectives.
As reviewed herein, studies on the biology of adherens junction dynamics in the testis lag far behind those investigating the blood-testis barrier, largely because of the lack of suitable in vitro and in vivo models that can be used to study adherens junction dynamics. For instance, in vivo models that investigate the dynamics of the blood-testis barrier, such as the cadmium chloride (40, 41, 174) and glycerol (176, 177, 179, 180) models, have illustrated that a loss of blood-testis barrier function can also compromise cell adhesion. This postulate is further supported by another recently developed in vivo model of blood-testis barrier dynamics using an occludin peptide (43). Moreover, the biochemical constituents of the adherens junction have not been known until recently (Fig. 9). Thus, it was virtually impossible to study adherens junction dynamics as they relate to germ cell movement. Moreover, the biochemical composition of desmosome-like junctions, hemidesmosomes, and focal adhesions in the testis remain largely uncharacterized.
Nonetheless, other laboratories, including ours, have established an in vitro model to study Sertoli-germ cell adherens junction dynamics with some success (for reviews, see Refs.4 , 5 , 48 , and 50, 51, 52). In two earlier studies by electron microscopy, both desmosome-like junctions and ectoplasmic specializations were shown to form by 24–48 h after the addition of germ cells to the Sertoli cell epithelium (51, 52). In brief, Sertoli cells were cultured alone for 5 d on Matrigel-coated dishes at a density of 0.5 x 106 cells/cm2. This allowed cells to form an epithelium with intact tight and adherens junctions. On d 6, total germ cells were isolated from adult rat testes and added to the Sertoli cell epithelium to initiate Sertoli-germ cell adherens junction assembly (Fig. 12). In Fig. 12A, a germ cell [still connected to another germ cell by cytoplasmic intercellular bridges (arrowheads in Fig. 12A) has assembled functional adherens junctions with a Sertoli cell (arrows in Fig. 12A). A microvillus, commonly found in Sertoli cells cultured in vitro, was also detected (asterisk in Fig. 12A). Figure 12B is the boxed area shown in Fig. 12A that shows an ectoplasmic specialization, a testis-specific adherens junction type, formed between a Sertoli and germ cell in vitro, which is typified by the presence of packed actin filaments sandwiched in between cisternae of endoplasmic reticulum and the plasma membranes of two opposing Sertoli and germ cells (arrowheads in Fig. 12B). Figure 12C shows the presence of desmosome-like junctions between a Sertoli and germ cell. This is typified by the presence of electron-dense patches (arrowheads in Fig. 12D), similar to those present in other epithelia. Figure 12, E and F, shows the corresponding magnified views of an ectoplasmic specialization and desmosome-like junctions between Sertoli and germ cell cultured in vitro (also see Fig. 12, B and D). These results illustrate that functional anchoring junctions are present between Sertoli and germ cells in vitro, consistent with two previously published reports (51, 52). Using this in vitro model, we have demonstrated that the assembly of Sertoli-germ cell adherens junctions requires the participation of numerous biological factors, including structural and signaling proteins, proteases, protease inhibitors, and GTPases (5, 45, 47, 48, 49, 50).
|
More importantly, cadherins and integrins are cell adhesion molecules present in other epithelia. Why wouldn’t AF-2364 damage anchoring junctions in other organs? First, adult male rats treated with AF-2364 at doses effective in inducing transient infertility did not display any damage in the kidney, liver, and epididymis when these organs were examined morphologically. This was in contrast to the testis (667). AF-2364 also failed to affect liver and kidney function based on serum microchemistry analysis (666, 667). Although AF-2364 induced germ cell loss via an initial activation of RhoB, this GTPase was not induced in the liver and kidney (45). This was also true for cadherin and catenin; both proteins were activated by AF-2364 in the testis (53, 54) but not in the liver and kidney (our unpublished observations). Taken collectively, these data suggest that cadherin/catenin and integrin/laminin complexes at the site of the ectoplasmic specialization require unique, testis-specific proteins such as testin (for review, see Ref.4) and other yet-to-be identified molecules to act as a receptor protein complex for AF-2364. In summary, these data suggest that the primary targets of AF-2364 are Sertoli cell-germ cell adherens junctions, such as the apical ectoplasmic specialization.
There is yet another question that remains to be addressed. Does AF-2364 limit its action only on the adherens junction or does it also affect the integrity of the blood-testis barrier? Shown in Fig. 13 are results of a recent study using immunofluorescent microscopy to monitor the integrity of the blood-testis barrier after treatment with AF-2364 (50 mg/kg body weight by gavage) (Fig. 13, D–F) vs. control rat testes (Fig. 13, A–C). Routine histological analysis has shown that virtually all of the tubules were devoid of elongating/elongate and round spermatids, and more than 70% of spermatocytes (Fig. 13, H vs. G) on d 12 after AF-2364 treatment, consistent with three earlier reports (54, 666, 667). However, the immunostaining intensity of occludin and ZO-1 at the site of the blood-testis barrier remained unaltered (Fig. 13, D–F vs. A–C). By d 120, germ cells repopulated the seminiferous epithelium, and more than 40% of the tubules were indistinguishable from control rats (Fig. 13, I vs. G and H). These observations are consistent with our postulate that AF-2364 exerts its effects on the cadherin/catenin/testin/integrin multiprotein complex at the site of the apical ectoplasmic specialization (for review, see Ref.4). Taken collectively, these results illustrate that a disruption of adherens junctions in the seminiferous epithelium by AF-2364 does not compromise the integrity of the blood-testis barrier. In light of these findings, we have named AF-2364 as Adjudin (an inducer of adherens junction disruption). Recently completed acute toxicity (in both mice and rats) and genotoxicity studies (micronucleus test in CHO cells and mutation test in bacteria) performed by licensed toxicologists have shown that AF-2364 is not toxic and mutagenic at doses effective to induce infertility in male rats and up to 10 times higher than the effective antifertility dose (our unpublished observations). The results discussed in this section have also permitted us to identify two putative signaling pathways that regulate cell adhesion in the testis.
|
6. Desmosomes.
Desmosomes, cell-cell anchoring junctions that use intermediate filaments as their attachment sites (Fig. 1; Fig. 12, C, D, and F; and Table 1), have been studied most extensively in the epidermis, where they have important functions in the maintenance of tissue integrity. Although the presence of this type of junction in the testis has been known for almost three decades, its biochemical architecture remains unexplored (for reviews, see Refs.18 and 87) (Table 1).
a. Functions of desmosomes.
Studies have demonstrated that desmosomes are highly organized spots between adjacent cells sandwiched in between two plaques (for reviews, see Refs.688, 689, 690, 691, 692, 693). They consist of two principal domains: 1) the extracellular core or desmoglea; and 2) symmetrical dense cytoplasmic plaques that lie parallel to the plasma membrane. Intermediate filaments attach to the cytoplasmic plaques to form a continuous network throughout the entire cell.
b. Proteins of desmosomes.
Desmosomes are composed of members from three protein families: 1) the cadherins, 2) armadillo proteins, and 3) plakins (for review, see Ref.690). Desmosomal cadherins can be further categorized into the desmocollins and desmogleins. Desmocollin and desmoglein each have three isoforms, namely desmocollins 1, 2, and 3 and desmogleins 1, 2, and 3. Each is the product of a different gene, having distinct patterns of expression among different tissues. These major desmosomal cadherins are connected to the cytoskeleton via cytoplasmic proteins, such as plakoglobin (also known as -catenin), desmoplakin, and plakophilin. It must be noted that plakoglobin is the only protein that is common to both desmosomes and adherens junctions (694). Minor components of desmosomes, such as IFAP 300 (695), pinin (696), desmocalmin (697), plectin (698), envoplakin (691), and periplakin (691) may also contribute to plaque structure. Except for plakoglobin (53), it is not known whether any of these proteins are found in the testis.
7. Hemidesmosomes.
Hemidesmosomes differ from desmosomes both structurally and functionally (for reviews, see Refs.13 , 14 , 699 , and 700) (Table 1). In the testis, hemidesmosomes are restricted to the basal lamina between the basal portion of the Sertoli cell and the underlying basement membrane (701, 702) (Fig. 1).
a. Functions of hemidesmosomes.
Hemidesmosomes are cell junctions that connect a cell’s cytoskeleton to the underlying basement membrane. In essence, their role is similar to that of focal adhesions, except that hemidesmosomal constituent proteins are linked to the cytoskeleton via intermediate filaments, not via actin.
b. Proteins of hemidesmosomes.
Except for integrins, the proteins that constitute the hemidesmosome have not been extensively characterized. It is known, however, that ß4 integrin confers hemidesmosome function by linking to intermediate filaments (for review, see Ref.703). Moreover, Imamura and colleagues (704) demonstrated that a monoclonal antibody designated 1-2B7B, which was obtained by immunizing mice with human prostate epithelial tissue, reacted with a protein that localized to the basement membrane of the testis. However, the identity and function of this protein have yet to be determined. In light of the closeness of the hemidesmosome to the blood-testis barrier, it is important that this junction type be investigated in the testis.
8. Communicating junctions.
There are two types of communicating junctions: 1) gap, and 2) chemical synapses (for reviews, see Refs.13 , 14 , and 705, 706, 707, 708, 709). Our discussion on communicating junctions will be limited only to gap junctions as they relate to germ cell movement in the testis (Fig. 1 and Table 1). Readers are encouraged to refer to relevant review articles cited in this section for more information.
a. Functions of gap junctions.
Gap junctions are the means by which ions and small molecules are exchanged between two adjacent cells. They are intercellular channels formed by the noncovalent interaction of two hemichannels, or connexons, with each cell contributing a connexon. Each connexon (1.6–2.0 nm in diameter with the two connexons separated by a space of 2–4 nm) is commonly found clustered together with other connexons to form a gap junctional plaque. Each connexon is further broken down into six integral membrane subunits, or connexins, surrounding a central hydrophilic pore (for reviews, see Refs.13 , 14 , and 710). Because gap junctions mediate signals between Sertoli and germ cells, they likely play a crucial role in coordinating events pertinent to germ cell movement in the seminiferous epithelium.
b. Proteins of gap junctions.
Gap junctions are comprised of connexin proteins. Each connexin traverses the plasma membrane four times, and both the N and C termini exit into the cytoplasm. The transmembrane and extracellular domains of different connexins are highly conserved, but regions located between the transmembrane domains and C termini are divergent. At least 20 connexins have been identified in mammalian tissues (for reviews, see Refs.710 and 711).
Morphological evidence for the presence of gap junctions between Sertoli cells has been known since the 1970s (112). More recent in vivo studies have demonstrated that dye coupling between Sertoli cells took place coordinately and associated with defined stages of the epithelial cycle (712). For instance, connexin43, the best-studied connexin, was found to be present at the site of Sertoli cell gap junctions, with the most intense staining detected in stages I–VIII (713). It was also shown that Sertoli cell connexin43 increased during testicular development (713). Other studies illustrated that gap junctions formed between Sertoli cells and pachytene spermatocytes in vitro as well (52, 714). Moreover, connexin33 is testis-specific and expressed largely by germ cells (25). To date, at least 13 connexins have been found to be present in the testis, namely connexins26, 31, 31.1, 32, 33, 36, 37, 40, 43, 45, 46, 50, and 57 (715).
Presently, it is not known how gap junctions contribute to germ cell movement in the testis. In a recent report, treatment of Novikoff hepatoma cells with cytochalasin B (disrupts actin filaments), colchicine (blocks microtubule assembly), or nocodazole (disrupts microtubules) did not appear to interfere with gap junction assembly (716). However, transfectants of 3T3 fibroblasts, which expressed a connexin43 mutant protein lacking the C-terminal cytoplasmic domain, were able to exhibit decreased cell motility as determined in Boyden chamber wounding assays (717). These results provide a link between gap junctions and cell movement, yet similar studies in the testis wait to be performed.
|
IV. Mechanism of Germ Cell Movement |
|---|
|
TopAbstractI. IntroductionII. Sertoli CellsIII. Sertoli-Sertoli and Sertoli...IV. Mechanism of Germ...V. Concluding RemarksReferences |
|---|
), how they coordinate these events requires further investigation. Nevertheless, several possible theories have advanced in recent years based largely on morphological studies. These theories are elaborated on below.
|
3. Intermediate cellular compartment theory.
The intermediate cellular compartment theory, which asserts the existence of a unique compartment inhabited by germ cells in transit from the basal to the adluminal compartment (2), also falls short. This is because studies by two different groups, Pelletier and Friend (213) and Cavicchia and Sacerdote (119), have unquestionably demonstrated the existence of only one occluding zonule per Sertoli cell. Thus, the existence of a compartment in which a developing germ cell is trapped in between two occluding zonules is arguable.
4. Repetitive removal of membrane segments theory.
Lastly, the repetitive removal of membrane segments theory contends that the continuous, upward migration of a large number of germ cells creates a massive amount of stress on the Sertoli cell junctional complex (for review, see Ref.118). As a consequence, junctions proliferate to form numerous loops in which developing germ cells, except elongate spermatids, are lodged. Each intercellular pocket, sealed at both ends by tight, adherens, and gap junctions, progressively traverses the seminiferous epithelium until it is internalized in autophagic vesicles (for review, see Ref.118). Once again, this hypothesis is intriguing, but studies have failed to demonstrate the localization of tight junction proteins, such as occludin, throughout the epithelium. Needless to say, other yet-to-be identified tight and adherens junction integral membrane proteins may exist to substantiate this hypothesis. As such, it is important that future resources be allocated to the identification and characterization of novel proteins that constitute the various junction types in the testis.
5. Junction restructuring theory.
On this note, we hereby propose the junction restructuring theory (Fig. 14). This hypothesis, which is based on recently published and ongoing biochemical and molecular studies from this and other laboratories (for review, see Ref.4 and references cited therein), is not an attempt to discredit the aforementioned theories. On the contrary, this postulate complements the existing pool of knowledge on germ cell movement in the testis.
|
), which allow for the timely passage of germ cells from one site to another in the epithelium (for review, see Ref.11). For example, tight junctions comprised of transmembrane proteins, such as occludin and claudin (371, 718) (Tables 1 and 4), must gradually disassemble (Fig. 14) one fibril at a time to allow for the passage of preleptotene/leptotene spermatocytes across the blood-testis barrier. Adherens and gap junctions also need to be disassembled in due course. It is envisioned that cytokines such as IL-1 and -1ß; basic fibroblast growth factor (bFGF); TNF-; and TGF- and -ß produced by Sertoli and/or germ cells (Table 4) trigger the cascade of events leading to the activation of proteases (719, 720, 721), which in turn cleave cell junctions (Fig. 14). Cytokines can also stimulate or inhibit the production of junctional component proteins (for reviews, see Refs.183 and 722) (Fig. 14). Indeed, a recent study from our laboratory has shown that TGF-ß3 can perturb the Sertoli cell tight junction barrier in vitro by blocking the transient induction of occludin and ZO-1 (22), which are needed for tight junction assembly (24, 25, 27, 43). Likewise, TNF- was shown to inhibit the production of occludin, but not ZO-1, by Sertoli cells and to modulate tight junction dynamics by affecting collagen 3(IV), MMP-9, and TIMP-1 (49). However, disassembled junctions must be quickly reassembled to maintain the integrity of the seminiferous epithelium. The second phase of germ cell movement, cell junction assembly, requires that cytokines activate protease inhibitors (48, 723, 724, 725). This triggers the synthesis of proteins required for assembling disrupted junctions (Fig. 14). This is supported by recent studies which demonstrated that cytokines, such as TGF-ß3, can also affect the production of adherens junction-associated proteins [e.g., E-cadherin, ß-catenin, p120ctn, and Src (39)] and protease inhibitors [e.g., 2-MG (39)], besides their effects on tight junction proteins (22, 23, 38). It is also our belief that cytokines, proteases, and proteases inhibitors are involved in the restructuring of the ectoplasmic specialization in light of several reports which illustrated that these molecules are activated during the assembly of functional anchoring junctions between Sertoli and germ cells (5, 48). Additionally, one must remember that the mechanical force that lifts germ cells upward during the epithelial cycle is not entirely generated by Sertoli cells. This driving force may also be produced by the continuous mitotic divisions of spermatogonia in the basal compartment adjacent to the basement membrane (Figs. 2 and 14). As such, junction disassembly and reassembly during spermatogenesis may be the result of the upward movement of germ cells and possibly not the cause of it (for review, see Ref.16). To reduce the complexity of this model, we have not included information on the use of motor proteins, ATPases, and GTPases by the ectoplasmic specialization for the translocation of elongating/elongate spermatids across the epithelium (for reviews, see Refs.20 and 87). All in all, this model will continue to be updated as advances in cell and molecular biology help us to better understand this cellular phenomenon in the seminiferous epithelium.
B. Additional comments
1. Cell movement in other cell types.
This hypothesis (Fig. 14) attempts to bring together the biochemical and molecular events of germ cell movement. However, it fails to take into account that germ cells cannot move from one site to another because they lack the cytoskeletal machinery necessary to do so. Numerous studies using other cell types, such as fibroblasts, have shown that cell movement involves cell adhesion via cycles of detachment and attachment (for reviews, see Refs.773 and 774). This is similar to the translocation of germ cells across the epithelium, which requires phases of Sertoli-germ cell junction disassembly (detachment) and assembly (attachment). It is possible that the polymerization and depolymerization of actin in Sertoli cells can generate the needed mechanical force to lift germ cells from one site to another similar to the mechanical force generated by structures known as filopodia and lamellipodia at the leading edge of migrating fibroblasts (for reviews, see Refs.773 and 774). These processes are mediated, at least in part, by members of the GTPase superfamily. However, germ cells are dissimilar to migrating fibroblasts in at least two aspects. First, germ cells do not possess an extensive cytoskeletal network, such as that found in fibroblasts. Second, germ cells in different stages of their development do not assume amoeboid cell shapes. Thus, it is believed that the cytoplasmic processes of Sertoli cells generate the protrusive force that is required for the movement of germ cells from one site to another in the seminiferous epithelium.
C. Regulation of cell junction dynamics in the testis by cytokines
1. Cytokines in the testis.
Sertoli and germ cells produce cytokines in the seminiferous epithelium. It is known that these two cell types synthesize several cytokines such as TGF- and -ß; TNF-; nerve growth factor; FGF; and IFN-, -ß, and -, which regulate cell growth, differentiation, and function (for reviews, see Refs.9 , 11 , 229 , 230 , and 244). For instance, EGF stimulates meiosis in spermatocytes but has no reported effects on spermatogonia and spermatids (for reviews, see Refs.8 , 9 , 775 , and 776). Specific cytokines, such as IFN-, can even participate in the defense mechanism of cells in the testis (Refs.154 , 733 , and 734 ; for review, see Ref.777).
2. Effects of cytokines on tight junction dynamics.
A thorough literature search on the effects of cytokines on tight junctions in different epithelia has demonstrated that these effects are mediated by: 1) affecting the expression and/or synthesis; 2) redistribution; or 3) phosphorylation status of tight junction proteins (for review, see Ref.183). For example, TNF- and IFN-, used either alone or together in cultures of human intestinal HT-29/B6 cells, diminished occludin promoter activity (778). IL-4 and -13 were also shown to perturb the tight junction barrier by inhibiting the expression of occludin and ZO-1 in Calu-1 cells (196), which is similar to the way TGF-ß3 perturbs the Sertoli cell tight junction barrier in vitro (22). On the other hand, HGF affects the tight junction barrier by moving ZO-1 away from the site of tight junctions (187, 188, 779). Lastly, other studies have shown that IFN- and EGF can perturb tight junction functionality by inducing the phosphorylation of occludin and ZO-1 in LLC-PK1 and A431 cells, respectively (185, 191). In the testis, TGF-ß2 and -ß3, and TNF- participate in Sertoli cell tight junction dynamics by affecting occludin and/or ZO-1 expression (22, 163). The effect of TGF-ß3 on Sertoli cell tight junctions in vitro and in vivo is mediated via the mitogen-activated protein/MEKK/p38 MAPK pathway (23, 38), whereas TNF- likely affects tight junction dynamics via the ILK/p130 Cas/JNK (c-Jun N-terminal kinase) signaling pathway (163).
3. Effects of cytokines on adherens junction dynamics.
In contrast to tight junctions, the effects of cytokines on adherens junction dynamics have not been studied extensively. What is known is that cytokines can either alter the phosphorylation status of adherens junction proteins or perturb the interactions that exist between adherens junction proteins and actin filaments. For example, EGF, HGF, and vascular endothelial growth factor can all stimulate tyrosine phosphorylation of -catenin, ß-catenin, and p120ctn in different cell lines in vitro (517, 780, 781). Other cytokines, such as FGF, induced dissociation of the cadherin-ß-catenin complex and recruited ß-catenin to the nucleus (782). This is similar to TGF-ß1, which caused - and ß-catenin, cadherin-5, and plakoglobin to move away from the site of adherens junctions in endothelial cells (783). Treatment of a proximal tubular cell line, HK-2, with TGF-ß1 also elicited similar effects (492). In a rat carcinoma cell line, EGF was shown to activate Src, resulting in cell motility (784). In summary, cytokines play an important role in the regulation of Sertoli cell tight junction dynamics. Although the role of cytokines on adherens junctions in the testis remains to be investigated, studies from other epithelia have clearly implicated them in the regulation of this junction type.
D. Regulation of cell junction dynamics in the testis by proteases and protease inhibitors
In this section, we review some of the recent studies on proteases and protease inhibitors as they relate to the events of germ cell movement in the seminiferous epithelium.
1. Proteases.
Proteases, enzymes that hydrolyze peptide bonds at specific sites, are categorized into one of two groups: 1) exopeptidases, and 2) endopeptidases (for reviews, see Refs.785, 786, 787, 788, 789, 790). Exopeptidases (also known as peptidases) cleave peptide bonds exclusively near the N or C terminus of a polypeptide, usually one or two amino acid residues from its end. On the other hand, endopeptidases (also known as proteinases) cleave peptide bonds internally and are classified into one of four types: 1) serine, 2) cysteine, 3) aspartic, and 4) metalloproteases, depending on the amino acid residue at its catalytic site (e.g., 1, 2, and 3) or if its activity requires divalent cations (e.g., 4).
a. Serine proteases.
Serine proteases have been studied extensively because they are active at a neutral pH and present nearly everywhere in the body. In addition, they are synthesized as inactive precursors, known as zymogens, and require limited proteolysis for activation. Serine proteases do not require any cofactors to exert their proteolytic activity (for review, see Ref.14).
To date, the best-studied serine proteases are trypsin, PA, and plasmin. PA is present in several tissues and cells (for reviews, see Refs.791 and 792). In the testis, Sertoli cells synthesize and secrete both u-PA and tissue-type PA, but not trypsin (143, 719, 793, 794). In fact, most of the PA activity detected in testicular extracts is derived from Sertoli cells, suggesting that this protease plays an important role in the biology of the Sertoli cell (744, 793). Moreover, studies have demonstrated that Sertoli cells secrete PA preferentially into the basal compartment when cultured on bicameral units (795). When staged seminiferous tubules were cultured in vitro, this serine protease was synthesized and secreted in a cyclic manner (794, 796). For example, studies have shown that PA secretion was at its highest level in stages VII–VIII of the epithelial cycle coinciding with the event of spermiation (796, 797). Although u-PA is the predominant type of PA secreted by Sertoli cells in vitro (796), tissue-type PA becomes the predominant type secreted when FSH or cAMP derivatives are added into Sertoli cell cultures (796, 798). In addition, coculture of residual bodies obtained by elutriation from adult rat testes with Sertoli cells was able to significantly increase PA activity (721, 799). Subsequently, the phagocytosis of these residual bodies by Sertoli cells resulted in an increase in IL-1 activity, suggesting that cytokines may regulate protease activity (721, 799). Insulin, cGMP derivatives, prostaglandins, EGF, estrogens, androgens, LH, and human chorionic gonadotropin all failed to affect Sertoli cell PA activity (800, 801). However, a previously published report demonstrated that TGF-ß affected Sertoli cell PA expression (719). Low levels of secreted PA in the form of two molecular variants, which differed in Mr from those previously reported, have also been detected in concentrated peritubular myoid cell-conditioned media (for review, see Ref.225). Germ cells consisting largely of spermatocytes, spermatids, and residual bodies secrete PA, but this may be the result of Sertoli cell contamination (745, 793).
A unique 41-kDa protein from germ cell-conditioned media has been identified (740). Studies have demonstrated that this molecule, in addition to being a serine protease, was also able to affect Sertoli cell secretion in vitro. For instance, this protein inhibited Sertoli cell clusterin and testin secretion at doses between 0.5 and 5 ng/ml, whereas at doses greater than 500 ng/ml it cleaved [125I]collagen and [125I]testin (740). It remains to be determined whether Sertoli cells can synthesize this serine protease. These findings illustrate that germ cells contribute to the pool of protease activity in the testis that is required for spermatogenesis and germ cell movement.
b. Cysteine proteases.
The best-studied acid lysosomal cysteine proteases, the cathepsins, are synthesized by many cell types (for reviews, see Refs.787 , 802 , and 803). They possess a highly reactive cysteine residue at their active site and require a pH 4.0 for maximal activity. To date, several cathepsins have been identified such as cathepsins B, H, K, L, N, O, S, and T (788).
Cathepsin L, also known as cyclic protein-2, is a 39-kDa Sertoli cell secretory product (47, 233, 804, 805, 806, 807). Germ cells can also synthesize cathepsin L (47, 741, 808). Studies have reported that cyclic protein-2 is the proenzyme of cathepsin L (807). More importantly, Sertoli cell cathepsin L was highest in stages VI–VII before spermiation, suggesting that it participates in the movement of elongate spermatids toward the tubule lumen (47). Indeed, studies by immunohistochemistry have localized cathepsin L to the site of the apical ectoplasmic specialization in late stage VII–early stage VIII of the epithelial cycle in the rat, consistent with its involvement in spermiation (233). Cathepsin L also increased in Sertoli cells isolated from aging testes with highest levels of cathepsin L detected in 38- to 45-d-old rats (809). In addition, this enzyme can: 1) activate u-PA; 2) degrade proteins; 3) initiate a cascade of events involving proteases; and 4) promote cell proliferation (for review, see Ref.225). However, before cathepsin L can participate in these processes, it must first be activated. Additional studies have shown that cathepsin L can be stimulated by oncogenes and tumor promoters. Thus, it has been implicated in the metabolism and invasion of tumor cells (for review, see Ref.810). The proenzyme, but not the mature protein, apparently participates in the regulation of Leydig cell steroidogenesis by forming a complex with TIMP-1 (811).
c. Aspartic proteases.
To date, few aspartic proteases (for reviews, see Refs.812 and 813) have been identified in the testis. For instance, pepsin is the best-studied aspartic protease, yet it is not present in the testis. Conversely, cathepsin D, the only known aspartic protease present exclusively in lysosomes (for reviews, see Refs.814 and 815), is found in the testis, rete testis, epididymis, and efferent ducts (816). Specifically, studies have demonstrated the presence of cathepsin D in primary spermatocytes, as well as in Sertoli and Leydig cells (816). In addition, cathepsin D mRNA was found to be highest in stages VII–IX of the epithelial cycle, suggesting that it participates in the release of mature spermatids into the tubule lumen (47). Similar to cathepsin L, it is also secreted as a proenzyme. Second, this protease requires a pH 3.0–4.0 for maximum activity and is inactive at or above pH 7.0. This illustrates that cathepsin D may be an important lysosomal enzyme required for the scavenger function of Sertoli cells, such as in the reabsorption of apoptotic germ cells or the proteolysis of cytoplasmic droplets during spermatogenesis.
In addition to cathepsin D, other aspartic proteases that are found in the testis include angiotensin-converting enzyme (ACE) and renin (for review, see Ref.817). Developing germ cells can synthesize a specific ACE that is different from somatic ACE found in tissues such as the epididymis, lung, and kidney (818, 819). More importantly, ACE plays an important role in male fertility because mice lacking the ACE gene were sterile (820, 821).
d. Metalloproteases.
Recent studies have demonstrated the presence of metalloproteases in the testis. These include: 1) the matrix metalloproteases, 2) neutral metalloendopeptidase, and 3) N-arginine dibasic convertase.
N-arginine dibasic (NRD) convertases, zinc-dependent metalloendopeptidases, cleave basic amino acids at the N terminus of arginine residues (for review, see Ref.822). A characterization of the first NRD (NRD1) convertase cDNA from the rat testis (743, 823) illustrated that it contains a long stretch of acidic amino acids (824). Recently, a second NRD (NRD2) convertase was identified, differing from NRD1 by an insertion of 68-amino acid residues close to the active site (825). NRD convertase was found to be highest in the testis, restricted to elongate spermatids (743). To date, the function of NRD convertase remains to be known, but it possibly relates to the event of spermiation.
Members of another emerging family of metalloproteases, the M13 family (for review, see Ref.826), which is comprised of six members, namely neprilysin [NEP (742, 827)], Kell blood group protein (828), phosphate regulating gene with homologies to endopeptidases on the X chromosome (829), endothelin-converting enzyme-2 (830), and X-converting enzyme are also found in the testis. The membrane metalloendopeptidase (MME) family has also been shown to play a significant role in the homeostasis of NEP, which can regulate other biological molecules by lowering their extracellular concentration via proteolysis, thereby affecting receptor binding and signaling function (for review, see Ref.831). In fact, overexpression of cell-surface NEP inhibited the phosphorylation of FAK (832). FAK can be activated by phosphorylation at Tyr397, which in turn affects ectoplasmic specialization dynamics in the testis. However, it remains to be shown whether proteases such as MMP-9 colocalize with pFAK. Studies have demonstrated the presence of NEP in several tissues such as the kidney, intestine, adrenal glands, lung, brain, epididymis, and prostate gland (833). Specifically, NEP is present in Sertoli cell membranes (834). Recently, a new member of the MME family, MME-like 2, was cloned, and its mRNA level was found to be highest in the testis (835). However, the significance of this finding remains to be determined.
2. Protease inhibitors
a. Serine protease inhibitors.
Serine protease inhibitors, also known as serpins, are large molecules consisting of approximately 400 amino acid residues (for review, see Ref.836). In the testis, PA inhibitor (PAI) is the best-studied serpin (792). To date, three PAIs are known to exist: 1) PAI-1, 2) PAI-2, and 3) PAI-3 (837). Sertoli and peritubular myoid cells are known to synthesize PAI-1, which specifically inhibits u-PA (719, 748, 800, 837, 838, 839), but not plasmin or trypsin. Interestingly, a recent study failed to detect PAI-1 mRNA in germ cells (748), yet in other studies it was detected in spermatids and residual bodies (838, 840). The reason for such a discrepancy may be the result of cell contamination. In addition, the level of PAI-1 increased in peritubular myoid cells after they were treated with TGF-, TGF-ß, or bFGF (841), once again implicating cytokines in the homeostasis of proteases and protease inhibitors. Moreover, TGF-ß, bFGF, EGF, FSH, GnRH, and forskolin can also affect Sertoli cell PAI-1 mRNA in vitro, suggesting that PAI-1 can be regulated by various factors (724, 748, 801, 841).
A recently completed gene knockout experiment has demonstrated the significance of protease inhibitors in cell junction dynamics as they relate to germ cell movement. For instance, protein C inhibitor-1–/– (a nonspecific serine protease inhibitor) male mice are infertile with impaired spermatogenesis (842). Histological analysis of testes from protein C inhibitor-1–/– mice revealed that the seminiferous tubule lumen was filled with immature germ cells, possibly resulting from a loss of cell adhesion between Sertoli and germ cells (842). Furthermore, the Sertoli cell tight junction barrier was damaged (842). This is in concert with findings from our laboratory which demonstrated that the assembly of tight and adherens junctions in the testis is regulated by the interplay of proteases and protease inhibitors (5, 24, 27, 47). Although trypsin, the major serine protease found in the systemic circulation, is absent in the testis (143), recent studies have demonstrated the presence of tryptase in the testis (5). More importantly, the expression of tryptase, similar to u-PA, was induced by as much as 4-fold during the adhesion of germ cells to the Sertoli cell epithelium (5). Furthermore, there was a transient but significant induction of serine protease activity during Sertoli-germ cell adherens junction assembly, strengthening the notion that proteases are important in adherens junction dynamics (5). Indeed, a more recent study using an in vitro adhesion assay with fluorescently labeled germ cells has demonstrated that the interplay of proteases and protease inhibitors can affect Sertoli-germ cell adherens junction dynamics (48).
b. Cysteine protease inhibitors.
Cystatins, cysteine protease inhibitors, are important regulatory proteins (for reviews, see Refs.843, 844, 845). There are three types of cystatins: 1) cystatins, 2) stefins, and 3) kininogens, all of which are functionally related.
The cystatins are present in body fluids. They consist of cystatins C, D, E, M, and S. Specifically, cystatin C can be detected in blood plasma, and cerebrospinal and seminal vesicle fluids. It can also inhibit cathepsin B, H, and L activity. In the testis, Sertoli and Leydig, but not germ cells synthesize cystatin C (236, 747). Studies have demonstrated that in addition to the testis and seminal vesicles, the epididymis and prostate also contain cystatin C mRNA (846). Interestingly, the highest level of cystatin C was detected in seminal vesicle fluid. In addition, the synthesis of cystatin C was highest in Sertoli cells 2 and 3 d after these cells were plated on plastic or Matrigel-coated dishes, respectively (747). However, the inclusion of FSH into cell cultures failed to elicit any changes in cystatin C synthesis (747). To date, it is not known whether cystatins D and S are present in the testis.
Stefins are 11-kDa intracellular proteins (for review, see Ref.847). Cystatin A, a member of this protease inhibitor family, can inhibit cathepsin B, C, H, L, and S activity (Ref.843 ; for reviews, see Refs.844, 845). On the other hand, kininogens are large, multifunctional proteins present exclusively in the systemic circulation. They participate in the acute phase response, inflammatory reaction, and blood coagulation (for reviews, see Refs.848 and 849).
c. Aspartic protease inhibitors.
Although several aspartic proteases (for reviews, see Refs.812 and 850) such as cathepsin D, ACE, and renin are present in the testis, specific inhibitors against this protease class have not yet been identified. It must be noted, however, that 2-MG, a nonspecific protease inhibitor secreted by Sertoli cells, can inhibit aspartic proteases (746, 851, 852, 853).
d. Metalloprotease inhibitors.
TIMPs are proteins that inhibit MMP activity. To date, four members of the TIMP family have been characterized: 1) TIMP-1, 2) TIMP-2, 3) TIMP-3, and 4) TIMP-4 (for reviews, see Refs.854, 855, 856).
Both TIMP-1 and -2 inhibit MMP activity so that a balance between extracellular matrix deposition and degradation can be maintained (for reviews, see Refs.855, 856, 857, 858). Specifically, TIMP-1 and TIMP-2 can: 1) inhibit tumor cell growth, invasion, and metastasis; 2) modulate cell morphology in vitro; and 3) participate in gonadal steroidogenesis. As such, they are multifunctional proteins (for review, see Ref.855). Studies have shown that Sertoli cells synthesize TIMP-1 (749) and TIMP-2 (750, 859). Further investigation demonstrated that TIMP-2 mRNA increased during testicular development from 3–60 d of age (750). This increase in TIMP-2 expression was not due to an increase in germ cell number during testicular maturation because coculture of germ cells with Sertoli cells failed to elicit changes in TIMP-2 expression (750). Additionally, a recent study has shown that TIMP-1 was significantly induced during the assembly of Sertoli-germ cell anchoring and Sertoli cell tight junctions (48), implicating its significance in junction dynamics. A recent study from this laboratory showed that TIMP-1/MMP-9 interactions are important in determining the amount of collagen that eventually will be under the influence of TNF- in the basement membrane (163). Also, it was shown that an antibody against collagen IV or recombinant TNF- could perturb the functionality of the Sertoli cell tight junction barrier (163). Based on these observations, it was postulated that the activation of MMP-9 by TNF- leads to the cleavage of collagen, releasing biologically active collagen fragments. This, in turn, perturbs Sertoli cell tight junctions, whereas TIMP-1 limits the action of MMP-9 on collagen (163). This suggests that cytokines regulate the homeostasis of the extracellular matrix via their effects on TIMPs and MMPs, which in turn affect tight junction functionality.
On the other hand, TIMP-3 (Mr, 21 kDa) participates in morphological changes that take place during cell transformation (856). Lastly, TIMP-4 has recently been cloned, and its expression was shown to be the highest in the heart (860). Both are also present in the testis (861). These results, taken collectively, illustrate that TIMPs are important regulatory molecules.
2-MG, a nonspecific protease inhibitor, is secreted by Sertoli cells (746, 851, 852, 853). Although testicular 2-MG has the same characteristics as serum 2-MG (746), their regulation is drastically different (851, 852, 853). For instance, 2-MG is an acute-phase protein in the liver but not in the testis. It was shown that the concentration of 2-MG in the systemic circulation increased by approximately 100-fold in response to experimentally induced inflammation, but not in the testis (851, 852), possibly because the blood-testis barrier prohibited the inflammatory stimulus from entering the seminiferous epithelium. Studies have reported the presence of 2-MG in stages I–VI of the epithelial cycle adjacent to elongating/elongate spermatid heads (237), implicating it in the movement of germ cells across the epithelium. Interestingly, the incubation of germ cell-conditioned media-derived proteins with Sertoli cells increased 2-MG expression dose-dependently (852, 853), illustrating the significance of Sertoli-germ cell interactions in the regulation of cell movement. Moreover, other studies have suggested that low-density lipoprotein receptor-related protein-1, a Sertoli and Leydig cell protein that mediates the endocytosis of a variety of ligands, is the receptor for 2-MG (862). 2-MG is also found in the extracellular matrix where it binds to growth factors and cytokines such as TGF-ß, bFGF, TNF-, PDGF, and IL-1 and -6 (863, 864). This provides a pool of cytokines to be used for the regulation of cellular events, such as junction disassembly and assembly. Indeed, a recent study using coimmunoprecipitation has shown that 2-MG associates with TGF-ß3, as well as with N-cadherin and ß-catenin, but not with ß1 integrin, clearly demonstrating the pivotal role of protease inhibitors in adherens junction dynamics. Collectively, these data illustrate that 2-MG participates in the movement of germ cells from the basal to the adluminal compartment by regulating paracrine factors in the testis (for review, see Ref.225).
3. Additional comments.
Several studies have shown that the synthesis of proteases and protease inhibitors by cultured Sertoli cells is under paracrine regulation (for review, see Ref.225). This is exemplified best with 2-MG, a nonspecific protease inhibitor and putative Sertoli cell product (746, 851), whose expression in Sertoli cells was not affected by pituitary hormones, testosterone, or glucocorticoids. However, its expression could be stimulated by coculturing Sertoli cells with either peritubular myoid (865) or germ cells (5, 852, 853). Also, the assembly of adherens junctions between peritubular myoid or germ cells and Sertoli cells may be responsible for this up-regulation. The activity of another protease inhibitor, PAI-1, was also enhanced when Sertoli cells were treated with TGF-ß1, which in turn inhibited PA activity in vitro (719). It would be ideal if the effects of cytokines on proteases and protease inhibitors could be confirmed by using an in vitro model of germ cell movement. However, such a model is presently unavailable. Furthermore, although coculturing germ cells with Sertoli cells can lengthen the viability of germ cells in vitro, permitting for the study of adherens junction assembly (4), the movement of germ cells across the Sertoli cell epithelium has yet to be shown.
E. Regulation of cell junction dynamics in the testis by protein kinases and phosphatases
Protein kinases and phosphatases are known regulators of junction dynamics in both epithelia and endothelia. For instance, it is known that approximately 30% of all cellular proteins are phosphorylated (866) and that phosphorylation of junction-associated proteins is important in the regulation of junction dynamics (for reviews, see Refs.467 , 664 , 867 , and 868). Indeed, proteins such as occludin, cadherin, nectin, catenin, and ß1 integrin are putative substrates of protein tyrosine and/or serine/threonine kinases, and their functions are regulated by phosphorylation at specific sites (for reviews, see Refs.4 , 467 , and 664). For instance, tyrosine phosphorylation of the cadherin-catenin complex has been implicated in the regulation of adherens junctions (for review, see Ref.500). On the other hand, Src (for reviews, see Refs.869 and 870), a putative protein tyrosine kinase, is a signaling molecule that is found in the seminiferous epithelium, including the ectoplasmic specialization (53, 111, 560). In fact, its phosphorylation causes it to become concentrated at the site of adherens junctions (871), specifically at ectoplasmic specializations between elongated spermatids and Sertoli cells before spermiation (111, 560). However, because this junction type is lost at spermiation (Ref.687 ; for reviews, see Refs.3 , 87 , and 554), the localization of Src at the ectoplasmic specialization is not necessarily indicative of its presence within the ectoplasmic specialization structure. Thus, additional studies are needed to confirm whether Src is indeed a constituent of the ectoplasmic specialization.
ß-Catenin, a signaling molecule that regulates adherens junction function, is also a putative substrate of protein tyrosine kinase (516). Apparently, tyrosine phosphorylation of ß-catenin correlates with a loss of cadherin-mediated cell adhesion (515, 516, 517). Moreover, both receptor tyrosine kinases and phosphatases coimmunoprecipitate with components of the cadherin-catenin functional unit, illustrating the participation of protein phosphorylation in the regulation of adherens junction dynamics (872, 873).
Several recent studies have reported that some myotubularins, which belong to a family of putative protein tyrosine and lipid phosphatases, can regulate the events of membrane trafficking (874) by down-regulating and degrading PI 3kinase (875). Interestingly, the expression of myotubularin related protein 2 was shown to be induced when adherens junctions in the seminiferous epithelium were perturbed by treating rats with AF-2364 (C. Y. Cheng, unpublished observations), illustrating once again the importance of protein and lipid phosphorylation in the regulation of cell adhesion in the testis (phosphoinositides are important cellular messengers). It is also known that cell adhesion depends on the phosphorylation status of the cadherin-catenin complex, which is regulated by a number of intracellular signaling molecules such as casein kinase 2 (a serine/threonine protein kinase; for review, see Ref.876), Csk (for review, see Ref.877), Src (for review, see Ref.870), and paxillin (for reviews, see Refs.625 and 878), all of which are found at the site of apical ectoplasmic specializations. Recently, Chapin and colleagues (111, 560) used immunohistochemistry to demonstrate changes in the levels of phosphorylated proteins in the seminiferous epithelium during spermiation. Other studies from our laboratory using inhibitors of protein kinases and phosphatases have shown that the Sertoli cell tight junction barrier is regulated by PKA and PKC (293, 879). Taken collectively, these data demonstrate the significance of kinases and phosphatases in junction dynamics.
Additional reports from our laboratory have clearly defined the physiological significance of kinases and phosphatases in the regulation of both adherens (45, 49) and tight (23, 38) junction dynamics in the testis. For instance, it was shown that Sertoli and/or germ cells are equipped with different kinases such as ILK (559), FAK (49), glycogen synthase kinase-3ß, c-Src, PI 3-kinase (49, 163, 559, 880), and MAPKs (49), some of which regulate Sertoli-germ cell adherens junctions by phosphorylating ß1 integrin. This leads to the activation of the pFAK/glycogen synthase kinase-3ß/p130 Cas/PI 3-kinase signaling pathway (163). Sertoli-germ cell adherens junction dynamics can also be regulated by RhoB via the ROCK/LIM kinase [both are putative protein kinases produced by Sertoli and/or germ cells (45)] signaling pathway.
Although the precise molecular mechanism(s) used by kinases and phosphatases in the seminiferous epithelium remains unclear, these studies have opened up an unprecedented opportunity to investigate how these proteins regulate junction dynamics, as well as cell movement, in the testis. For instance, specific inhibitors can be used to block key molecules in these pathways to determine whether there are any phenotypic changes at the level of cell adhesion. Indeed, the use of a specific ROCK inhibitor, Y-27632, could effectively block AF-2364-induced adherens junction disruption and the subsequent loss of germ cells from the epithelium (45). As such, this area of research needs to be expanded in future studies so that junction dynamics and germ cell movement can be better understood.
F. Regulation of cell junction dynamics in the testis by cAMP and cGMP
Although there are only a few studies that examine the roles of cAMP and cGMP on adherens junction dynamics in the testis, cAMP (the product of adenylate cyclase) has been known to regulate the Sertoli cell tight junction barrier since the 1990s (34). However, it is not known whether cGMP (the product of guanylate cyclase) also has a similar role in the testis. Once studies using protein kinase inhibitors illustrated that PKA and PKC are involved in the regulation of Sertoli cell tight junctions (293), there was some interest to explore the functional significance of these cyclic nucleotides in junction dynamics. Moreover, the role of cAMP in the regulation of epithelial and endothelial tight junction barriers, such as the blood-brain barrier, has been known for decades (for reviews, see Refs.868 and 881). Using Sertoli cells cultured in vitro, dibutyryl cAMP, a cAMP analog that is noncleavable by cAMP phosphodiesterase, was shown to have a biphasic effect on the functionality of the Sertoli cell tight junction barrier. At 4–20 µM or 100–500 µM, dibutyryl cAMP either facilitated or perturbed the tight junction barrier, respectively, suggesting that the effects of cAMP on Sertoli cell tight junctions are concentration dependent (34). Another study demonstrated a transient but significant increase in intracellular cAMP and cGMP levels in Sertoli cells during assembly of the tight junction barrier in vitro (30). Once the tight junction barrier assembled, intracellular cAMP and cGMP levels reduced to their basal levels (30). These results illustrate that whereas cAMP and cGMP may not be required for the maintenance of the Sertoli cell tight junction barrier, they are needed to assemble it. Interestingly, the presence of ZnPP (882), which significantly inhibited intracellular cGMP and NO levels, could also facilitate the assembly of the Sertoli cell tight junction barrier (30), illustrating for the first time the participation of cGMP in tight junction dynamics. Moreover, when Sertoli cells were exposed to 8-bromo-cGMP, a cell membrane-permeable cGMP analog, tight junction functionality was perturbed dose-dependently. Taken collectively, these data illustrate that cAMP and cGMP are important regulators of tight junction barrier function. These results also seemingly suggest that the upstream signaling molecule regulating intracellular cGMP, at least in part, is NO (30). Its effect is likely mediated via changes in tight junction integral membrane proteins because the presence of ZnPP was shown to induce the production of occludin in Sertoli cells (30). These results are supported by an observation from another cell culture system which demonstrated that NO can induce the removal of occludin and ZO-1 from the site of the tight junction, perturbing the functionality of the intestinal tight junction barrier (883). NO-induced disruption of tight junctions in cultured human cervical epithelial cells has also been shown to be mediated via changes in cGMP and PKG (884), which in turn disrupt the homeostasis of intracellular F- and G-actin (884). This leads to destabilization of the epithelium and increased cell permeability. Although these studies need to be expanded, these data have demonstrated that cyclic nucleotides under the regulation of NO play an important role in regulating tight junction dynamics in the testis, which is required for the movement of preleptotene/leptotene spermatocytes across the blood-testis barrier.
G. Regulation of cell junction dynamics in the testis by GTPases
1. Rho GTPases.
Rho GTPases are important regulators of the actin cytoskeleton and, thereby, the morphology and motility of eukaryotic cells (for reviews, see Refs.885, 886, 887, 888). The first report in the literature that implicated GTPases in germ cell movement appeared a decade ago. It demonstrated that the inactivation of Rho GTPases in bovine sperm tail by C3 exoenzyme could significantly inhibit sperm motility (889). More recent immunohistochemistry studies have localized Rac1 to the basal compartment of the seminiferous epithelium predominantly in stage VIII of the epithelial cycle (560), whereas Cdc42 was largely restricted to Sertoli cells, spermatocytes, and elongate spermatids. Another report has shown that RhoB is stage-specific in the rat testis, restricted to the site of basal and apical ectoplasmic specializations and associating largely with Sertoli cells, spermatocytes, and elongating but not elongate spermatids in the epithelium (45). Moreover, a novel protein kinase designated TESK1, found to be restricted almost exclusively to round spermatids (890), was shown to share an approximately 50% homology with LIM kinase (3), a downstream signaling molecule of the Rho GTPase signal transduction pathway (for review, see Ref.448). Subsequent studies have illustrated that TESK1 can stimulate the assembly of actin stress fibers and focal adhesions via phosphorylation of cofilin [an actin-severing protein (891)], another Rho GTPase downstream signal transducer (for review, see Ref.448). It is possible that the movement of germ cells across the seminiferous epithelium may utilize a similar mechanism.
Other Rho effectors (proteins that can activate Rho GTPases), such as citron kinase and protein kinase N, are also found in the testis (892, 893). Just as important, several testis-specific GTPase-activating proteins (GAPs, proteins that stimulate the intrinsic rate of GTP hydrolysis by converting a GTPase from its GTP- to GDP-bound state, thereby inactivating it; for review, see Ref.448), have now been identified. These are 2-chimaerin (894), ß-chimaerin (895), and MgcRacGAP (896). Although the precise physiological roles of these GTPase regulatory proteins in germ cell movement remain to be established, these GAPs are restricted largely to germ cells and known to be involved in the reorganization of the cytoskeleton (for review, see Ref.448). Interestingly, another report has shown that Rho GDI–/– mice (guanosine nucleotide dissociation inhibitor, proteins that sequester a GDP-bound GTPase in the cytoplasm and inhibit the dissociation of GDP from it; for review, see Ref.448) were infertile (897). Taken collectively, these studies illustrate the crucial roles of GTPases in the testis and their possible involvement in germ cell movement because the downstream signaling endpoint of these GTPases is the cytoskeleton.
2. Rab GTPases.
Unfortunately, the study of Rab GTPases in the testis continues to be a largely unexplored area of research. Although the presence of several Rab GTPases, such as Rab8B (898), has been demonstrated in the testis, not a single in-depth study exists in which Rab GTPase function is explored in this organ. Several reports have demonstrated the participation of Rab8 in vesicular traffic from the trans-Golgi network to the basolateral plasma membrane (347, 899), whereas other studies have implicated this GTPase in the events of cell movement. For example, the expression of wild-type enhanced green fluorescent protein-Rab8B promoted the reorganization of actin (899, 900). Other published results have shown Rab8 to promote cell polarization, a phenomenon that is essential for the assembly and maintenance of tight (for review, see Ref.294), and possibly adherens junctions in epithelial cells (521, 901). We are currently examining the roles of Rab GTPases in the events of junction assembly and disassembly (50) and propose that selected GTPases function as a "traffic light" (GTP-bound form, green-GO; GDP-bound form, red-STOP) by participating in the delivery of proteins, such as cadherin, to Sertoli-Sertoli and Sertoli-germ cell adherens junctions during their assembly. During disassembly, on the other hand, GTPases move proteins away from tight and adherens junctions. Such a mechanism of protein trafficking is basic to cells when junctions need to be briefly but quickly disassembled and assembled. Indeed, a mechanism responsible for regulating adherens junctions in MDCK cells, which involves the participation of Rab GTPases (for review, see Ref.902), has recently been proposed. Using cell surface biotinylation and recycling assays, Stow and colleagues (514) have shown that the internalization of E-cadherin and ß-catenin after cell adhesion was disrupted by depletion of extracellular calcium. This was followed by their recycling back to the plasma membrane after calcium was repleted in the medium. They also demonstrated that E-cadherin colocalized with Rab5 in early endosomes (514). Other reports have shown that integrins are also recycled via vesicles during cell movement (Refs.903 and 904 ; for reviews, see Refs.905, 906, 907). Moreover, a recent report has placed Rab8B in the forefront of our current research. By coimmunoprecipitation, Rab8B was shown to associate with E-cadherin and -catenin but not with ß1 integrin and paxillin (50). Rab8B was also found to localize to the site of basal and apical ectoplasmic specializations by immunohistochemistry (50), confirming results obtained from coimmunoprecipitation experiments. In addition, Rab8B increased during the assembly of Sertoli-germ cell adherens junctions (50). These observations implicate GTPases in the dynamics of adherens junctions. Future studies should address whether a disruption of adherens junctions during the movement of germ cells involves a similar mechanism.
H. Regulation of cell junction dynamics in the testis by calcium
Calcium plays an important role in junction dynamics (for reviews, see Refs.908, 909, 910). However, virtually all of the studies investigating the physiological significance of calcium in junction dynamics were performed using nontesticular cell types. For instance, depletion of calcium from the media of MDCK cultures was shown to perturb the tight junction barrier within minutes. Once calcium was replaced, the disrupted tight junction barrier rapidly resealed, illustrating the importance of calcium in tight junction dynamics (for review, see Ref.911). A similar study from our laboratory has shown that depletion of calcium from Sertoli cells cultured in vitro can also disrupt tight junction barrier function within 15 min, which was followed by its resealing within 90 min after the replacement of calcium (57). In this context, it is also of interest to note that the inclusion of an anti-E-cadherin antibody into the media of MDCK cultures prevented resealing of the disrupted tight junction barrier (156, 912). In the absence of calcium, cadherins also became susceptible to proteolysis (for reviews, see Refs.458 , 908 , and 913). These data demonstrate that calcium is an important regulator of junction dynamics, functioning with molecules that confer cell adhesion (for review, see Ref.4).
|
V. Concluding Remarks |
|---|
|
TopAbstractI. IntroductionII. Sertoli CellsIII. Sertoli-Sertoli and Sertoli...IV. Mechanism of Germ...V. Concluding RemarksReferences |
|---|
As reviewed in this article, junctions are platforms for signal transduction events (for reviews, see Refs.4 , 458 , 467 , and 914, 915, 916). Signaling events not only can alter the integrity of tight and adherens junctions, but can also be relayed to the nucleus to change transcriptional activity. More importantly, most of these studies have significantly impacted our understanding of junction restructuring during germ cell movement, in that they have provided innovative concepts for the development of safe and effective male contraceptives. AF-2364 is an example of such an effort (for review, see Ref.4). Currently, it is known that AF-2364 exerts its effects by compromising cell adhesion in the testis by targeting the multiprotein complex that comprises the ectoplasmic specialization (45, 49, 50, 54, 666, 667). Taken collectively, these results demonstrate that a thorough understanding of the biochemical architecture of tight and anchoring junctions and the signaling pathways that regulate them is crucial to the development of new male contraceptives, in addition to those that are currently being investigated (for reviews, see Refs.917 and 918).
In this review, we have discussed the biology of tight (e.g., blood-testis barrier) and adherens (e.g., ectoplasmic specialization) junctions in the testis, highlighting specific areas that deserve future investigation. We have also described several key observations that were obtained when different in vitro and in vivo models of junction disassembly and assembly were used. Lastly, we have proposed a new model of germ cell movement in the seminiferous epithelium based on recent research from other laboratories, as well as ours, which takes into account the roles of tight and adherens junction-associated proteins, cytokines, proteases and their inhibitors, GTPases, kinases and phosphatases, and other autocrine and paracrine factors. This model will likely serve as a framework for future investigators whose aim will be to delineate the biology and regulation of germ cell movement during spermatogenesis. Thus, it will be continuously updated as more information becomes available in the years to come.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Abbreviations: ACE, Angiotensin-converting enzyme; AF-2364, 1-(2,4-dichlorobenzyl)-indazole-3-carbohydrazide (also known as Adjudin); AF-6, ALL-1 fusion partner from chromosome 6; ASH1, absent, small, or homeotic 1; ASIP, atypical PKC isotype-specific interacting protein; bFGF, basic FGF; CAR, coxsackievirus and adenovirus receptor; cGMP, cyclic GMP; CRB1, Crumbs homolog 1; Csk, C-terminal Src-kinase; ECH, NF-E2 (also known as p45)-related molecule, alternative names for ECH are Nrf2 or NF-B2 related factor-2; EGF, epidermal growth factor; eNOS, endothelial NOS; FAK, focal adhesion kinase; FGF, fibroblast growth factor; FITC, fluorescein isothiocyanate; GAP, GTPase-activating protein; HGF, hepatocyte growth factor; IFN, interferon; ILK, integrin-linked kinase; iNOS, inducible NOS; JAM, junctional adhesion molecule; JEAP, junction-enriched and -associated protein; Keap1, kelch-like ECH-associating protein 1; LIM, Lin-11, Isl-1, and Mec-3; MAGI, MAGUKs with inverted orientation; MAGUK, membrane-associated guanylate kinase; MDCK, Madin-Darby canine kidney; 2-MG, 2-macroglobulin; MME, membrane metalloendopeptidase; MMP-9, matrix metalloprotease-9; Mr, molecular mass; MUPP1, multi-PDZ domain protein 1; NEP, neprilysin; NF, neurofilament; NO, nitric oxide; NOS, NO synthase; NRD, N-arginine dibasic; PA, plasminogen activator; PAI, PA inhibitor; Pals, proteins associated with Lin-7; PAR, partitioning-defective; PDGF, platelet-derived growth factor; PDZ, postsynaptic density-95/Discs-large/ZO-1; pFAK, phosphorylated FAK; PI 3-kinase, phosphoinositide 3-kinase (also known as phosphatidylinositol 3 kinase); PKA, protein kinase A; PKC, protein kinase C; PKG, protein kinase G; PTEN, phosphatase and tensin homolog deleted on chromosome 10; ROCK, Rho-associated protein kinase; SF, scatter factor; SH3, src-homology region 3; TER, transepithelial electrical resistance; TIMP-1, tissue inhibitor of metalloproteases-1; u-PA, urokinase-type PA; VASP, vasodilator-stimulated phosphoprotein; Y-27632, (R)-(+)-trans-N-(4-pyridyl)-4-(1-aminoethyl) cyclohexane-carboxamide, 2HCl; ZnPP, zinc II protoporphyrin-IX; ZO, zonula occludens.
|
References |
|---|
|
TopAbstractI. IntroductionII. Sertoli CellsIII. Sertoli-Sertoli and Sertoli...IV. Mechanism of Germ...V. Concluding RemarksReferences |
|---|
2-macroglobulin, and testins: biosynthesis and differential regulation by germ cells. Mol Cell Endocrinol 89:127–140[CrossRef][Medline]
, gelatinase B (matrix metalloprotease-9), and tissue inhibitor of metalloprotease-1 in the basal lamina regulates Sertoli cell-tight junction dynamics in the rat testis. Endocrinology 144:371–387 modulation of epithelial barrier function. Time course, reversibility, and site of cytokine binding. J Immunol 150:2356–2363[Abstract]
-2b on barrier function and junctional complexes of renal proximal tubular LLC-PK1 cells. Kidney Int 55:2178–2191[CrossRef][Medline]
directly affects barrier function of cultured intestinal epithelial monolayers. J Clin Invest 83:724–727[Medline]
decreases barrier function in T84 cells by reducing ZO-1 levels and disrupting apical actin. Am J Physiol Gastrointest Liver Physiol 276:G1279–G1288
induces dissociation of the junctional complex and paracellular leakage in filter-cultured human thyrocytes. J Clin Endocrinol Metab 83:945–952. J Cell Physiol 281:C2029–C2038
abrogates the glucocorticoid stimulation of tight junction formation and reverses the steroid-induced down-regulation of fascin in rat mammary epithelial tumor cells by a Ras-dependent pathway. Exp Cell Res 273:1–11[CrossRef][Medline]
(TNF-) regulates the epithelial barrier in the human intestinal cell line HT-29/B6. J Cell Sci 112:137–146[Abstract]
. Ann NY Acad Sci 915:193–203[Medline]
and tumour necrosis factor- on T84 cell function. Gut 45:191–1982-macroglobulin in rat testes is consistent with its role in germ cell movement and spermiation. J Androl 15:575–582: an oncodevelopmental growth factor. Cancer Cells 2:389–397[Medline]
. Biochem Soc Trans 19:259–262[Medline]
. Cell Biol Int 19:373–388[CrossRef][Medline]
and epidermal growth factor receptor during embryonic and perinatal rat testis development. J Androl 22:1019–1029[Abstract]
stimulates Sertoli cell proliferation in vitro. Andrologia 32:62–63[Medline]
in mouse spermatogenic cells. Endocrinology 133:389–396 down-regulated the Fas ligand and inhibits germ cell apoptosis in the human testis. J Clin Endocrinol Metab 86:4480–4488-catenin and actin filaments. J Cell Biol 138:181–192-catenin required for the strong state of cadherin-based cell adhesion. J Cell Biol 144:1311–1322+ and - in guinea pig Sertoli cells: a possible association with F-actin and G-actin. Biol Reprod 57:367–376[Abstract]
subunit in tight junction biogenesis. J Biol Chem 271:25750–25753i2 in the maintenance and biogenesis of epithelial cell tight junctions. J Biol Chem 273:21629–2163312 directly binds to the Src homology 3 domain and regulates paracellular permeability in epithelial cells. J Biol Chem 277:24855–24858-catenin. J Biol Chem 274:5981–5986
