The testis lies within the scrotum and is covered on all surfaces except its posterior border by a serous membrane called the tunica vaginalis. This structure forms a closed cavity representing the remnants of the processus vaginalis into which the testis descended during fetal development (Figure 1). Along its posterior border, the testis is loosely linked to the epididymis which at its lower pole gives rise to the vas deferens (1).
Figure 1. The relationships of the tunica vaginalis to the testis and epididymis is illustrated from the lateral view and two cross sections at the level of the head and mid-body of the epididymis. The large arrows indicate the sinus of the epididymis posteriorly. Reproduced with permission from de Kretser et.al.1982 in 'Disturbances in Male Fertility' Eds K Bandhauer and J Frick, Springer - Verlag Berlin.
The testis is covered by a thick fibrous connective tissue capsule called the tunica albuginea. From this structure, thin imperfect septa run in a posterior direction to join a fibrous thickening of the posterior part of the tunica albuginea called the mediastinum of the testis. The testis is thus incompletely divided into a series of lobules.
Within these lobules, the seminiferous tubules form loops, the terminal ends of which extend as straight tubular extensions, called tubuli recti, which pass into the mediastinum of the testis and join an anastomosing network of tubules called the rete testis. From the rete testis, in the human a series of six to twelve fine efferent ducts joined to form the duct of the epididymis. This duct is extensively coiled and forms the structure of the epididymis that can be divided into the head, body and tail of the epididymis (1). At its distal pole, the tail of the epididymis gives rise the vas deferens (Figure 2).
Figure 2. The arrangement of the efferent ducts and the subdivisions of the epididymis and vas are shown. Reproduced with permission from de Kretser et.al.1982 in 'Disturbances in Male Fertility' Eds K Bandhauer and J Frick, Springer - Verlag Berlin.
The arterial supply to the testis arises at the level of the second lumbar vertebra from the aorta on the right and the renal artery on the left and these vessels descend retroperitoneally to descend through the inguinal canal forming part of the spermatic cord. The testicular artery enters the testis on its posterior surface sending a network of branches that run deep to the tunica albuginea before entering the substance of the testis (2). The venous drainage passes posteriorly and emerges at the upper pole of the testis as a plexus of veins termed the pampiniform plexus (Figure 3). As these veins ascend they surround the testicular artery, forming the basis of a countercurrent heat exchange system which assists in the maintenance of a temperature differential between the scrotally placed testis and the intra-abdominal temperature (3).
Figure 3. The arrangement of the vasculature of the testis in the region of the distal spermatic cord and testis is shown. Reproduced with permission from de Kretser et.al.1982 in 'Disturbances in Male Fertility' Eds K Bandhauer and J Frick, Springer - Verlag Berlin.
The vas deferens ascends from the testis on its posterior surface as a component of the spermatic cord passing through the inguinal canal and descends on the posterolateral wall of the pelvis to reach the posterior aspect of the bladder where its distal end is dilated forming the ampulla of the vas (Figure 4). At this site it is joined by the duct of the seminal vesicle, on each side, to form an ejaculatory duct that passes through the substance of the prostate to enter the prostatic urethra. Together with seminal vesicles, the prostate which opens by a series of small ducts into the prostatic urethra, contribute approximately 90-95% of the volume of the ejaculate. During the process of ejaculation, these contents, together with sperm transported through the vas, are discharged through the prostatic and penile urethra. Retrograde ejaculation is prevented by contraction of the internal sphincter of the bladder during ejaculation. Failure of this sphincter to contract results in retrograde ejaculation and a low semen volume.
Figure 4. The diagram depicts the relationship between the vas deferens, the seminal vesicles, the posterior aspect of the bladder and the prostate gland. The cytological features of the epithelium of the seminal vesicles is shown: this tissue is androgen dependent. Reproduced with permission from de Kretser et.al.1982 in 'Disturbances in Male Fertility' Eds K Bandhauer and J Frick, Springer - Verlag Berlin.
Spermatogenesis represents the process by which precursors termed spermatogonia undergo a complex series of divisions to give rise to spermatozoa (4,5). This process takes place within the seminiferous epithelium which is a complex structure composed of germ cells and radially oriented supporting cells called Sertoli cells (Figure 5). The latter cells extend from the basement membrane of the seminiferous tubules to reach the lumen. The cytoplasmic profiles of the Sertoli cells are extremely complex as this cell extends a series of processes that surround the adjacent germ cells in an arboreal pattern (5-7).
Figure 5. This photomicrograph illustrates the typical structure of the testis showing the seminiferous tubules containing the germ cells and Sertoli cells. The tubules are surrounded by thin plate-like contractile cells called peritubular myoid cells. The Leydig cells and blood vessels lie within the intertubular tissue.
Spermatogenesis can be divided into three major phases (i) proliferation and differentiation of spermatogonia, (ii) meiosis, (iii) spermiogenesis which represents a complex metamorphosis involved in the transformation of round spermatids arising from the final division of meiosis into the complex structure of the spermatozoon.
These cells represent a population that divide by mitosis providing both a renewing stem cell population as well as spermatogonia that are committed to enter the meiotic process. The identification of different types of spermatogonia is complex due to a lack of definitive markers that can identify specific stages. To date, classification of these cells has depended on the features of their nuclei and, in particular, their chromatin patterns (5). The identification of the latter, which represents a crucial step in classification, is often obscured by poor fixation especially if the tissue is fixed in formalin. The ideal fixatives for testicular tissue are Bouin's or Cleland's solution. In general two main classes of spermatogonia can be identified in all mammals: Type A exhibiting fine pale staining nuclear chromatin and Type B with coarse chromatin collections found close to the nuclear membrane(8). In many mammals, the Type A spermatogonia can be divided into several subtypes that may represent different phases of proliferation and progression towards Type B spermatogonia. The Type B spermatogonia are generally agreed to represent the cells which differentiate and enter into the process of meiosis where they are called primary spermatocytes(9).
In the human and other primates, the Type A spermatogonia can be further divided into A dark (Ad) and A pale (Ap)(9) Some investigators have proposed that the Ad spermatogonia represent the reserve or non-proliferative spermatogonial population which can give rise to Ap(10-12) whereas others have suggested that the Ap spermatogonia are the true stem cell of the testis(13). Some of these differences arise from the difficulties in identifying spermatogonial cell types.
Spermatogonia do not separate completely after meiosis due to incomplete cytokinesis and remain joined by intercellular bridges (14). These intercellular bridges persist throughout all stages of spermatogenesis and are thought to facilitate biochemical interactions allowing synchrony of germ cell maturation.
Following damage to the seminiferous epithelium, some investigators have suggested new criteria which may facilitate the identification of the true spermatogonial stem cell within the epithelium. These criteria have emerged from studies of investigators engaged in transplantation of germ cells into the testis. Following the induction of cryptorchidism, the surviving spermatogonia, from which restoration of spermatogenesis is possible, show the presence of α6β1 integrin (15). The studies of Brinster and colleagues have shown that transplantation of spermatogonial populations can restore spermatogenesis in infertile recipients (15a-d). Further our understanding of some of the factors controlling spermatogonial stem cell proliferation have emerged and involve such proteins as Glial-derived neurotropic factor (15e-i).
This process commences when Type B spermatogonia lose their contact with the basement membrane to form preleptotene primary spermatocytes. The preleptotene primary spermatocytes engage in DNA synthesis and condensation of individual chromosomes providing the appearance of thin filaments in the nucleus which identify the leptotene stage (15). At this stage, each chromosome consists of a pair of chromatids (Figure 6). As the cells move into the zygotene stage, there is further thickening of these chromatids and the pairing of homologous chromosomes. The further enlargement of the nucleus and condensation of the pairs of homologous chromosomes termed bivalents, provides the nuclear characteristics of the pachytene stage primary spermatocyte. During this stage, there is exchange of genetic material between homologous chromosomes derived from maternal and paternal sources. The sites of exchange of genetic material are marked by the appearance of chiasmata and these become visible when the homologous chromosomes separate slightly during diplotene. The exchange of genetic material involves DNA strand breakage and repair (16).
Figure 6. The diagrammatic representation of the events occurring between homologous chromosomes during the prophase of the first meiotic division shows the period of DNA synthesis, the formation of the synaptonemal complex and the processes involved in recombination. Reproduced with permission from de Kretser and Kerr (1994) in "The Physiology of Reproduction" Ed E Knobil & J D Neill, Lippincott, Williams & Wilkins.
The diplotene stage is recognised by partial separation of the homologous pairs of chromosomes that still remain joined at their chiasmata and each is still composed of a pair of chromatids. With dissolution of the nuclear membrane, the chromosomes align on a spindle and each member of the homologous pair moves to opposite poles of the spindle during anaphase. The resultant daughter cells are called secondary spermatocytes and contain the haploid number of chromosomes but, since each chromosome is composed of a pair of chromatids, the DNA content is still diploid. After a short interphase, which in the human represents approximately six hours, the secondary spermatocytes commence a second meiotic division during which the chromatids of each chromosome move to opposite poles of the spindle forming daughter cells that are known as round spermatids (17,18). Meiotic maturation in the human takes about 24 days to proceed from the preleptotene stage to the formation of round spermatids.
Several studies have identified key molecules that are necessary to allow meiosis to be completed. These include the synaptonemal complex protein (SCP1) and the chromosomal core protein (COR1)(19,20). Further, the heat shock protein, HSP 70-2, is required for desynapsis of the synaptonemal complexes and the completion of the first meiotic division (21). Targeted gene disruption approaches have identified a number of proteins that are critical to the meiotic process showing differences between the sexes and also emphasizing the important role of proteins involved in DNA repair mechanisms (For review see 21a).
The transformation of a round spermatid into a spermatozoon represents a complex sequence of events that constitute the process of spermatogenesis. No cell division occurs but a conventional round cell becomes converted into a spermatozoon with the capacity for motility. The basic steps in this process are consistent between all species and consists (a) the formation of the acrosome (b) nuclear changes (c) the development of the flagellum or sperm tail (d) the reorganisation of the cytoplasm and cell organelles and (e) the process of release from the Sertoli cell termed spermiation (Figure 7) (5, 21,22).
Figure 7. The changes during spermiogenesis involving the transformation of a round spermatid to a mature spermatozoon are shown. Redrawn with permission from de Kretser and Kerr (1994) in "The Physiology of Reproduction" Ed E Knobil & J D Neill, Lippincott, Williams & Wilkins.
The formation of the acrosome commences by the coalescence of a series of granules from the Golgi complex which migrates to come into contact with the nuclear membrane where it becomes applied as a cap-like structure over approximately 30-50% of the nuclear surface (22).
The nuclear changes involve a reorganisation of the nucleus and cytoplasm such that the nucleus comes into close apposition with the cell membrane in a region where it is covered by the acrosomal cap. Subsequently, there is a progressive condensation of chromatin to form progressively larger and more electron dense granules together with a change in the shape of the condensed nucleus. This change in shape varies significantly between species. The condensation of chromatin represents the morphological appearances of significant biochemical changes which result in the stabilization of DNA. These changes include the replacement of lysine-rich histones with transitional proteins which in turn are subsequently replaced by arginine-rich protamines (24,25). The resultant DNA becomes resistant to digestion by the enzyme DNase. Associated with these changes there is a marked decrease in nuclear volume and a cessation of gene transcription (26).
The formation of the tail commences early in spermiogenesis when a filamentous structure emerges from one of the pair of centrioles which lie close to the Golgi complex. Associated with the changing nuclear cytoplasmic relationships, the developing flagellum and the pair of centrioles become lodged in a fossa in the nucleus at the opposite pole to the acrosome. The central core of the axial filament, called the axoneme, consists of nine doublet microtubules surrounding two single central microtubules, which represents a common pattern found in cilia. This basic structure is modified at the region of its articulation with the nucleus through the formation of a complex structure known as the connecting piece (27). As spermiogenesis proceeds, the outer dense fibres and fibrous sheath which characterize the regions of the sperm known as the mid and principle-piece are developed. The biochemical characteristics of these components of the sperm tail are emerging (27a-f). While they provide some structural stability to the tail, recent evidence suggests that they may serve as a molecular scaffold to position key enzymes critical to successful sperm motility. For instance, CatSper 1 is an ion channel plasma membrane-located protein that has been shown to regulate calcium ion fluxes critical for the process of hyperactivation of sperm motility associated with capacitation (27g) which is located on the principal-piece. It is one of a family of four related proteins that are located within the testis (27h). Further studies have shown that plasma membrane calcium-ATPase 1 is also located to the principal-piece and has been shown to be critical for the process of hyperactivation of sperm motility (27i). While these are plasma membrane located complexes, tpx1 (also called CRISP2), a protein localised to the outer dense fibres of the tail and acrosome (27j) has been shown to regulate ryanodine receptor calcium signalling (27k).
The formation of the mitochondrial sheath occurs at the time of the final reorganisation of the cytoplasm and organelles of the spermatid (5,22,23). The mitochondria that had remained around the periphery of the spermatid aggregate around the proximal part of the flagellum to form a complex helical structure (Figure 8). In the process of spermiation, the cytoplasm of the spermatid which has now migrated to a caudal position around the tail, is shed, most likely involving a process which requires the active participation of the Sertoli cell. Some observations suggest that prolongations of Sertoli cell cytoplasm send finger-like projections which invaginate the cell membrane of the spermatid cytoplasm and literally 'pull' the residual cytoplasm off the spermatid (22). These cytoplasmic collections, termed residual bodies, which contain mitochondria, lipid and ribosomal particles are phagocytosed and moved to the base of the Sertoli cell where they are broken down by lysosomal mechanisms. Recent studies on the process of spermiation reveal that this is biochemically regulated and that inhibition of FSH and LH secretion by androgen/gestagen combinations may lead to retention of sperm (27l,27m).
Figure 8. A cross-section through the developing mid-piece of the sperm tail shows the aggregation of mitochondria (arrows) surrounding the outer dense fibres (labelled 1-9) which in turn surround the axoneme composed of 9 doublet microtubules surrounding two central microtubules. Reproduced with permission from "Visual atlas of human sperm structure and function for assisted reproductive technology" Ed A.H. Sathanathan 1996.
Within the seminiferous epithelium, the cell types that constitute the process of spermatogenesis are highly organized to form a series of cell associations or stages. These cell associations, or stages of spermatogenesis, result from the fact that a particular spermatogonial cell type when it appears in the epithelium is always associated with a specific stage of meiosis and spermatid development. The cycle of the seminiferous epithelium was defined by LeBlond and Clermont (28), as the series of changes in a given area of the seminiferous tubule between two appearances of the same developmental stage or cell association. They defined 14 stages in the rat cycle based on the 19 phases of spermiogenesis as identified by the periodic acid Schiff (PAS) stain (Figure 9). In effect, if it was possible to observe the same region of the seminiferous epithelium by phase contrast microscopy over time, the appearance would progress through the 14 stages before stage I reappeared. They also demonstrated that the duration of any one stage was proportional to the frequency with which it was observed in the testis. As type A spermatogonia in any one area of the epithelium progress through meiosis and spermiogenesis to become spermatozoa, the specific area of the tubule would pass through the 14 stages four times. In each progression, the progeny of the spermatogonia progressively move toward the lumen of the tubule.
Figure 9. This is a diagrammatic representation of the stages of the seminiferous cycle in the rat and shows the types of germ cell associations which form the stages. The most mature spermatids are shed at stage VIII and this is reflected by the alteration in the transillumination patterns seen in the tubules (central representation of the tube). The stages at which proliferative events occur are shown as well as some key physiological events.
Studies in many mammalian species demonstrated that the cycle of spermatogenesis could be identified for each species but showed that the duration of the cycle varied for each species (17). In many species, especially the rat, the same stage of spermatogenesis extends over several millimetres of the adjacent tubule and it is possible, by observation under transillumination, to dissect lengths of seminiferous tubules at the same phase of spermatogenesis (29). Such observations amply confirmed the earlier studies of Perey and colleagues (30), that the stages of spermatogenesis were sequentially arranged along the length of the tubule (Figure 10). As the cycle progress, this arrangement resulted in a "wave of spermatogenesis" along the tubule. Regaud (30) interpreted his observations correctly by the statement "the wave is in space what the cycle is in time".
Figure 10. The pattern of the stages of spermatogenesis as they occur along the tubule are shown. Data based on Perey et. al. (30). Reproduced with permission from de Kretser and Kerr (1994) in "The Physiology of Reproduction" Ed E Knobil & J D Neill, Lippincott, Williams & Wilkins.
For many years, investigators believed that such a cycle did not occur in the human testis but the careful studies of Clermont (32) showed that human spermatogenesis could be subdivided into 6 stages. However unlike the rat, each stage often only occupied one quadrant of a tubule giving the disorganized appearance. By careful studies using tritiated thymidine injections into the testis, Clermont and Heller (18) demonstrated that the duration of the cycle in the human took 16 days and the progression from spermatogonia to sperm took 70 days or four and a half cycles of the seminiferous cycle. Other studies showed that the cycle length was specific for each species (eg rat 49 days) and the progression of each cell type in spermatogenesis involved a defined duration (17).
The mechanisms that defined these temporal constraints on spermatogenesis have been the subject of speculation as to whether these were intrinsic or were imposed by the Sertoli cells. The recent demonstration that when rat germ cells were transplanted into the mouse testis, spermatogenesis proceeded at the normal rate for the rat indicated that this rate is determined by intrinsic mechanisms within germ cells (33). Nevertheless, the Sertoli cell, because it extends from the base of the tubule to the lumen, restricts entry to the tubule because of the blood-testis barrier and is intimately associated with several generations of germ cells from spermatogonia to spermatids, is in a powerful position to influence the radially directed organization of cell associations. Given that the germ cells influence the metabolism of the Sertoli cells at each specific stage of the seminiferous cycle, the cell associations representing the stages of the seminiferous cycle may well be essential to allow successful completion of spermatogenesis.
As indicated earlier, the Sertoli cells have an intimate physical relationship with the germ cells (Figure 11) during the process of spermatogenesis (5,7,34). The cytoplasmic extensions that pass between the germ cell populations surrounding the Sertoli cell provides structural support through a microfilament and microtubular network present in the cytoplasm of the Sertoli cell (35). This architecture is not static but changes in the tubule depending on the stage of the spermatogenic process. Additionally, the Sertoli cells have a role in regulating the internal environment of the seminiferous tubule. This arises from specialized inter-Sertoli cell junctions which are formed at the sites where processes of Sertoli cell cytoplasm from adjacent cells meet (36). The junctions are predominantly located at the baso-lateral regions of the cell and form occluding-type junctions. As a result of these junctional specializations, intercellular transport between the Sertoli cell and spermatogonia is possible but does not extend to the more centrally placed germ cells that are effectively sequestered from the extra-tubular environment by the tight occluding junctions. The physiological counterpart of these anatomical structures is known as the blood testis barrier that regulates the entry of a variety of substances into the central compartment of the seminiferous tubule(37). The siting of the cell junctions effectively divides the seminiferous epithelium into a basal compartment which contains the bases of the Sertoli cells and spermatogonia and an adluminal compartment containing the central regions of the Sertoli cell and the other germ cell types. It is clear that as Type B spermatogonia migrate from the basement membrane of the tubule into the adluminal compartment, these tight junctions must open up to allow this cellular migration to take place (Figure 12) and reform beneath the Type B spermatogonia which have now left the basement membrane to form preleptotene spermatocytes. Recent studies have indicated that the formation and dissolution of these junctional specializations are under the control of physiological regulators such as tumor necrosis factor (TNFα) (See review Mruk & Cheng 37a)
Figure 11. The general architecture of the Sertoli cell is shown. Note the thin cytoplasmic processes that extend between the germ cells. The Sertoli cell is in contact with a variety of germ cells and adjacent Sertoli cells when three dimensional perspectives are considered.
Figure 12. The relationship between the Sertoli cells and spermatogonia as they proceed through the leptotene and zygotene stages of meiosis is shown. Note that the tight inter-Sertoli cell junctions "open up" to allow germ cells to move from the basal to the adluminal compartments.
These physical and biological relationships between Sertoli cells and germ cells is reflected in observations that have led to the conclusion that the number of Sertoli cells in a testis determines the ultimate spermatogenic output(38-41). There is considerable data to indicate that experimental manipulation of Sertoli cell number by such agents as FSH or the induction of neonatal hypothyroidism can increase Sertoli cell number and increase the spermatogenic output of the testis. These observations suggest that there are a maximum number of germ cells that can be supported by an individual Sertoli cell. Whether this is a functional or physical relationship is unclear. As discussed later in this chapter, the hormonal regulation of spermatogenesis is also dependent on the Sertoli cell. Germ cells do not contain receptors for testosterone or FSH (42-45). These receptors are present on Sertoli cells and the hormonal signals must be transduced through molecular mechanisms involving the Sertoli cell which remain to be elucidated.
External to the basement membrane of the seminiferous tubule, there are several layers of modified myofibroblastic cells termed myoid cells (46,47). These cellular layers are responsible for the irregular contractions of the seminiferous tubules which propel fluid secreted by the Sertoli cells, together with testicular spermatids into the lumen and through the tubular network to the region of the rete testis. These cells also participate in regulating processes within the seminiferous tubules by the production of growth factors such as activin A and platelet derived growth factors (47a, 47b).
The Leydig cells lie within the intertubular regions of the testis and are found adjacent to blood vessels and the seminiferous tubules (5,48). They are the cell type responsible for testosterone production which is essential for the maintenance of spermatogenesis. There are very significant organizational differences in the intertubular tissue betweens species reflecting the number of Leydig cells and differing architecture involving blood vessels and lymphatic sinusoids(49). Additionally, fibroblasts and macrophages and small numbers of mast cells are found in the intertubular regions of the testis (51,52).
In most species there are two generations of Leydig cells (53,54). A fetal generation under the stimulation of hCG results in the production of testosterone during gestation (55). These cells in the human, decrease in number towards term and degenerate and are lost from the intertubular region at about twelve months of age (56). The adult generation of Leydig cells in the human results from LH stimulation commencing at the time of puberty. This generation arises by division and differentiation of mesenchymal precursors under the influence of LH (57).
The Leydig cells have the capacity to synthesize cholesterol from acetate or to take up this substrate for steroidogenesis from lipoproteins (48,58). Typical of any steroid secreting cell, the Leydig cell contains abundant smooth endoplasmic reticulum and mitochondria which have tubular cristae which are unique to steroidogenic cells. The enzymes required for steroidogenesis are located in the mitochondria and in endoplasmic reticulum requiring intracellular transport of substrates between these organelles to achieve successful androgen production.
The spermatogenic process is controlled by both classic endocrine mechanisms as well as by intrinsic mechanisms that are mediated by growth factors, cytokines and other molecules (5,6,59). It is not possible within this chapter to provide a full description of all of these processes, some of which are still unclear. This section provides the reader with an understanding of some examples of these intrinsic processes over which external hormonal mechanisms exert important control. In some cases the manner in which hormones influence these intrinsic mechanisms remain unclear.
Pivotal in our understanding of intrinsic mechanisms is the central role played by the Sertoli cell. As discussed earlier, the Sertoli cell, through the formation of the blood-testis barrier divides the seminiferous epithelium into a basal and adluminal compartment and limits intercellular transport to the centrally placed germ cells (36). These cells therefore control the environment in which all germ cells other than spermatogonia develop. By modulating entry of substances into the adluminal compartment, the Sertoli cells are required to provide such factors as substrates for glycolysis (60-62). Several studies have established that lactate rather than glucose is the preferred substrate for glycolysis in primary spermatocytes and lactate is generated from glucose by Sertoli cells under the influence of FSH. The production of lactate leads to changes in the pH resulting in the alterations of the processing of stem cell factor from a soluble to a cell bound form, thereby influencing its action on spermatogonial stimulation (62a)
Further examples of the involvement of Sertoli cells arise from their ability to produce locally, a range of proteins which are essential for spermatogenesis but cannot gain access to the seminiferous epithelium because of the blood-testis barrier. For instance, testicular transferrin is an iron binding protein which is a secretory product of the Sertoli cell that is regulated by FSH and delivers iron to primary spermatocytes through a receptor-mediated endocytotic process (63,64). Another example is the copper binding protein, ceruloplasmin that is involved in the delivery of copper to the germ cells (65). All cells require iron to maintain respiration and cytochrome function, and copper is required as a coenzyme for a number of proteins.
There is ample evidence to document the importance of vitamin A in spermatogenesis but the mechanism may be an increased synthesis of secretory products such as transferrin. While the retinoid binding proteins are localised in Sertoli cells themselves, some evidence suggests that there may be a direct effect on the germ cells (66) further supported by the disrupted sperm production in mice with targeted inactivation of the retinoic acid receptor(67).
There are many more examples of the cooperative linkages between Sertoli cells and germ cells and numerous reviews can provide detailed information (6,7,59,68). The close association between germ cells and Sertoli cells is reflected by the capacity for germ cells to influence the metabolic activity of the Sertoli cells. This view can be illustrated by the cyclic changes in Sertoli cell function that occur during the stages of the cycle of spermatogenesis as shown by studies, on rats, involving transillumination of the seminiferous tubules to enable stage specific segments of tubules (Figure 9, see above) to be dissected for metabolic experiments(29). The data indicate that the metabolic activity of the Sertoli cell is influenced by the populations of germ cells that are found at a particular stage of the seminiferous cycle (69-71). For example maximal transferrin production occurs at stages X to XIV (72), while the greatest concentrations of androgen binding protein is found at stages VII-VIII (73). Some of these data can be linked to pathophysiological studies; for instance the greatest concentrations of androgen receptors are found at stages VII-VIII(74), which are the stages that show maximal apoptosis when the testis is exposed to androgen withdrawal induced by treating rats with the Leydig cell cytotoxin, ethane dimethane sulphonate (75).
Earlier in this chapter, the evidence for the concept that the number of Sertoli cells in the testis determined the total sperm output of the testis was discussed, emphasizing the studies that showed that the perinatal induction of hypothyroidism extended the duration Sertoli cell proliferation((40,41). Other Sertoli cell mitogens such as FSH and activin (76,77), together with thyroxine can also exert significant changes in the number to Sertoli cells in the testis depending on the temporal pattern of their secretion. The latter must occur before the cessation of Sertoli cell proliferation. In the rat, this occurs at about 20 days whereas in the human, Sertoli cells cease to divide during the pubertal process (78). It is possible that the failure of many men with hypogonadotropic hypogonadism to achieve normal testicular size and normal sperm counts, when treated by gonadotropic stimulation, may result from abnormal Sertoli cell proliferation during fetal and prepubertal life resulting in a decreased Sertoli cell complement (79).
The delay in germ cell maturation that occurs when perinatal hypothyroidism is induced to extend the duration of Sertoli cell proliferation suggests that there is a link between the degree of "maturity" of the Sertoli cell and its capacity to support spermatogenesis (41,42, 80). Alternatively, the continuing proliferation of Sertoli cells in this state prevents the formation of the occluding tight junctions between processes of adjacent Sertoli cells and the consequent development of the blood testis barrier. Despite the profound delay in germ cell development during the hypothyroid state, when the hypothyroidism is reversed, spermatogenesis resumes and produces an augmented total sperm output.
Given that there is increasing evidence that a range of cytokines and growth factors are important regulators of spermatogenesis through a variety of mechanisms (see reviews, 59,81), the delay in spermatogenic development in the hypothyroid model may be a reflection of the inability of the Sertoli cells in that state to supply the necessary cytokines and growth factors at specified times. The production of these proteins by the Sertoli cells is closely linked to specific stages of the spermatogenic cycle, implying but not proving that they are required to modulate yet to be defined mechanisms during spermatogenesis.
These mechanisms are likely to be involved in the expansion and survival of germ cell populations since the successful production of a normal sperm output, several germ cell populations must proliferate during mitosis and meiosis, and differentiate during spermiogenesis. There is no doubt that these events are influenced by the hormones discussed earlier but emerging data also implicate locally produced growth factors and cell survival molecules that influence the process of apoptosis.
The migration and proliferation of the primordial germ cells and their daughter cells termed gonocytes represents a crucial step in the establishment of spermatogenesis (82,83). This process is dependent on the interaction of stem cell factor produced by the Sertoli cells and the receptor, c-kit, which is located on germ cells particularly spermatogonia. Recent studies suggest that the production of a membrane bound form of stem cell factor by alternative splicing may be essential for spermatogonial multiplication (83). The production of the membrane-bound form is stimulated by a decrease in pH facilitated by the generation of lactate by Sertoli cells, an essential element in facilitating the survival of primary spermatocytes whose preferred substrate for glycolysis is lactate (62) This is an example of the close interaction between Sertoli cells and germ cells both in the production of growth factors and the interplay of their metabolic pathways.
The two cell divisions that constitute meiosis involve a number of events that require explicit control. For instance, the pairing of homologous chromosomes to form synaptonemal complexes is an essential step in the process of genetic recombination allowing transfer of DNA between paternal and maternal chromosomes. Data emerging from "gene knockouts" have identified several key mechanisms that are crucial for the progression of spermatocytes through meiosis. For example, inactivation of the gene encoding heat shock protein 70-2 results in the homozygous males failing to complete the first meiotic division (84). Genetic inactivation of several DNA repair processes also has resulted in a testicular phenotype which shows a disruption of spermatogenesis with abnormal meiosis (85). They showed that disruption of the DNA mismatch repair gene PMS 2 resulted in abnormal chromosomal synapsis in meiosis. However, small numbers of abnormally-shaped sperm were formed. A less severe phenotype was noted in mice with inactivation of HR 6B, the ubiquitin-conjugating DNA repair enzyme (86). As the molecular machinery of meiosis is progressively unravelled, patients with azoospermia resulting from mutations in the genes encoding the proteins which are key steps in this process will emerge.
As discussed earlier in this chapter, the steps in the formation of a sperm from its precursor, the haploid round spermatid, represent a fascinating process in cell biology. The development of the sperm tail, the remarkable nuclear changes involving the condensation and complexing of DNA and the changes in the relative positions of the nucleus, cell organelles and the cytoplasm, all pose innumerable questions as to how these events are controlled. Little is known of the mechanisms which control these steps. However the importance of precise structural configuration of sperm can be illustrated by the failure of motility in patients with the condition of the immotile cilia syndrome where the sperm tail and cilia lack specific proteins such as dynein (87,88). Even structurally normal sperm can fail to move as shown by the genetic inactivation of the gene encoding for a sperm calcium ion channel (89). This knowledge promises better diagnostic and potential therapeutic approaches to the problems of male infertility.
It has been recognized for some time that germ cells degenerate at under various stages during spermatogenesis but this occurs by apoptosis rather than necrosis. These apoptotic processes result in the removal of germ cells from the testis during a variety of hormonal manipulations (see review 90). The mechanisms and regulation of apoptosis in the testis are the subject of many ongoing studies but an example of the importance of this process in spermatogenesis emerged from recent studies of the inactivation of the gene encoding bcl-w, a cell survival molecule (91-93). These studies showed that the pubertal wave of spermatogenesis almost reached completion but the entire process collapsed rendering homozygous male mice infertile with an ultimate phenotype of Sertoli cell only tubules. Further studies of the reasons for the successful completion of the pubertal wave of spermatogenesis showed that while bcl-w was one of several pro-survival molecules expressed in the seminiferous epithelium during the onset of spermatogenesis, it was the only pro-survival protein expressed in the adult (94).
In a recent review, Young and Nelson (95) emphasized the role of apoptosis in the mediation of seasonal testicular regression. This data is consistent with the increase in apoptosis demonstrated in the testis following FSH withdrawal by passive immunization or in hypophysectomized rats (96,97). In their summary of potential mechanisms, Young and Nelson(95) proposed two options (a) Withdrawal of support by Sertoli cells for germ cells leading to alteration in the balance of Bcl-2 family members which in turn activates the caspase cascade (b) Activation of death pathways through the binding of Fas ligand to its receptor Fas (98,99).
In the earlier part of this Chapter, the process of spermatogenesis and the organization of the seminiferous epithelium have been discussed together with intrinsic mechanisms that are crucial to establish successful spermatogenesis. These intrinsic mechanisms are subjected to external influences exerted through the pituitary gonadotrophins follicle stimulating hormone (FSH) and luteinizing hormone (LH). These two hormones together with testosterone, produced by the Leydig cells in the intertubular regions of the testis are important for successful spermatogenesis. The secretion of the gonadotrophins FSH and LH are regulated by the episodic secretion of gonadotrophin releasing hormone produced in the hypothalamus and details of this system are considered elsewhere in this text. Through the stimulation of the Leydig cells by LH, testosterone is produced locally within the intertubular regions of the testis in high concentrations and exerts a very important influence on spermatogenesis. This section of the Chapter is divided into two parts (1) the production of testosterone and its control, (2) the control of spermatogenesis.
Testosterone is the major androgen secreted by the testis from its site of production within the Leydig cells. A normal male produces approximately 7 mg testosterone daily but also produces lesser amounts of weaker androgens such as androstenedione and dihydroepiandrosterone. In addition to testosterone, through the actions of the enzyme 5α reductase, dihydrotestosterone is produced by the testis in smaller amounts. The testis also contributes approximately 25% of the total daily production of 17βestradiol through the local action of the enzyme aromatase which converts androgenic substrates to this estrogen (100). The remainder of the circulating estradiol is produced by the adrenal and peripheral tissues through the actions of aromatase.
Cholesterol represents the major substrate for androgen production by the Leydig cells and is derived by an uptake mechanism involving the binding of circulating low density lipoprotein to specific receptors on Leydig cells which, following internalisation provides a significant source of cholesterol (101,102). In addition, the Leydig cells are able to undertake de novo synthesis of cholesterol from acetate and relative contributions of these two sources is partly dependent on species and the state of stimulation of the Leydig cells. The conversion of cholesterol to testosterone involves a number of steps that are catalyzed by enzymes, predominantly belonging to cytochrome P450 family (Figure 13).
The mobilization of cellular sources of cholesterol is achieved through the action of cholesterol ester hydrolase and subsequently, this is converted to pregnenolone by the enzyme cholesterol side-chain cleavage termed cytochrome P450SCC (103). The conversion of cholesterol to pregnenolone is a key step at which regulation of androgen production within the Leydig cells occurs. Availability of cholesterol substrate can be rate-limiting and the intracellular trafficking of cholesterol across mitochondrial membranes is dependent on the steroidogenic acute regulatory protein (STAR)(104-106). The role of this protein has been well demonstrated in patients with mutations in the gene encoding STAR in the disorder termed congenital lipoid adrenal hyperplasia wherein the mitochondria from the adrenals and gonads of these patients are unable to convert cholesterol to pregnenolone (107). Further, the results of studies involving targeted disruption of the mouse gene encoding STAR support the data derived from human studies (108).
Pregnenolone may progress to testosterone production through two pathways. It can be converted to progesterone through the enzyme 3β hydroxysteroid dehydrogenase (the D4 pathway) or can be hydroxylated at the 17α position by the enzyme 17ahydroxylase to form 17α hydroxypregnenolone (the D5 pathway). The relative importance of these two pathways vary with the species and the physiological status of the male (109). The further conversion of 17α hydroxypregnenolone through the D5 pathway involves the formation of the C19 steroid dehydroepiandrosterone catalyzed by the enzyme 17,20 lyase and both steps appear to be catalyzed by a single microsomal enzyme cytochrome P450 c17 encoded by a single copy gene on chromosome 10(110-111). The conversion of dehydroepiandrosterone to androstenediol is mediated by a microsomal enzyme 17β hydroxysteroid dehydrogenase encoded by a single gene(112,113).
The conversion of substrates from the D5 to the D4 pathway are catalyzed by the enzyme 3β hydroxysteroid dehydrogenase(114). In the D4 pathway 17α hydroxyprogesterone proceeds through the action of cytochrome P450 c17 to androstenedione and testosterone. Testosterone can be converted to a dihydrotestosterone by the enzyme 5α reductase(115) or can be metabolised to 17β estradiol by the enzyme aromatase(116).
It is important to recognize that intracellular transport of steroid substrates involved in androgen production is important with the transport of cholesterol into the mitochondrion to form pregnenolone and the transport of pregnenolone to smooth endoplasmic reticulum for the remainder of the steps in the production of testosterone.
LH, through specific receptors found on the surface of Leydig cells, controls the production and secretion of testosterone (117-118). The structure of the LH receptor is that of a member of the seven transmembrane domain G protein coupled receptor super family and mutations of this receptor are the cause of familial testicular resistance and male pseudohermaphroditism (119-120). Some men have constitutively activating mutations in the LH receptor and this has resulted in the onset of precocious puberty (121-123).
The interaction of LH with its receptor initiates signalling through the cyclic AMP pathway through GTP binding proteins (124-125). Signal transduction occurs through the protein kinase A pathway as its principal signal transduction mechanism. Some data suggests that intracellular calcium concentration can be induced by the action of LH by activating phospholipases in the lipoxygenase pathway (126). In addition, the changes in calcium can also regulate adenylate cyclase through the protein kinase C pathway.
An in vivo injection or an episode of LH secretion induced by GnRH, results in stimulation of the side-chain cleavage enzyme with the subsequent release of testosterone within 30-60 minutes of LH stimulation((58). The acute response to an injection of LH is dramatic in some species such as the rat and the ram but is much more attenuated in the human. This testosterone response lasts approximately 24-48 hours (127). If human chorionic gonadotrophin is used as an LH substitute, the kinetics of the initial stimulation are similar to LH but a second peak of testosterone secretion is evidence with hCG and occurs 48-72 hours after the initial injection(128). This biphasic pattern has been attributed to the observation that between 24 and 48 hours after an LH or hCG injection, the Leydig cells are refractory to further stimulation by either hormone (129-130). The second phase of testosterone secretion after hCG but not LH is associated with the longer half-life of hCG in comparison to LH. The hCG levels persist in the circulation and, following recovery from the refractoriness, testosterone levels increase. This observation has significant clinical importance since, in many men, a single weekly injection of hCG will suffice to maintain optimum testosterone responses rather than the frequent practice of giving injections of hCG two to three times per week.
It is important to recognise that LH enhances the transcription of genes that encode a range of enzymes in the steroidogenic pathway and that continued LH stimulation results in Leydig cell hypertrophy and hyperplasia (57,131,132). In the normal male, the episodic nature of LH stimulation is likely to avoid prolonged periods of Leydig cell refractoriness to LH stimulation (133).
A significant body of evidence has accumulated from studies in rats to suggest that the seminiferous tubules can exert an influence on Leydig cell testosterone production (134-135). This data emerges from a number of studies where changes in Leydig cell function have been demonstrated in association with temporary disruption of spermatogenesis such as the application of single episode of heat to the rodent testis (see Figure 14)(136). The most convincing data emerges from unilateral testicular damage such as that induced by cryptorchidism or efferent duct ligation, where the Leydig cells from the testis with spermatogenic damage shows an increased capacity for testosterone biosynthesis and a decrease in LH receptor number (137-138). The nature of the molecular mechanisms involved in this regulation is yet to be elucidated. While similar mechanisms are difficult to identify in the human, it is recognized that elevated LH and low to low normal testosterone concentrations, indicative of compromised Leydig cell function are found in 15-20% of men with severe seminiferous tubule failure.
Further support for the concept that the state of spermatogenesis can affect the function of the Leydig cells in men has emerged from the studies of Andersson et al (132a), who showed that lower testosterone and higher estradiol concentrations were present, and accompanied by higher LH levels in infertile. They concluded that this may reflect an extension of testicular dysgenesis to affect steroidogenesis or alternatively may result from inter-compartment interactions in the testis.
Figure 14. The changes in the seminiferous epithelium after a single episode of heat to the testis is shown. A depletion of heat sensitive germ cells is seen at 14 days and is associated with an increase in testosterone production by the testis in vitro and a decline in seminiferous tubule fluid production and androgen binding protein (ABP) levels in the testis. Data from Jégou et. al. 1984 (reference 136).
It is important to emphasize again, the relationship between the germ cells and the Sertoli cells that were discussed earlier in this chapter. The hormones that are essential for the successful development of normal spermatogenesis, FSH and testosterone, act through the Sertoli cells since the receptors for these hormones are located on these cells and are not located on germ cells. The processes by which the Sertoli cells are involved in the expression of hormone action are more readily apparent for FSH than testosterone, since in the latter, the molecular mechanisms are still obscure.
There have been a large number of studies in a variety of species using a range of experimental models often generating conflicting results. Some of these issues have emerged due to the use of impure hormone preparations and inadequate rigor in the analysis of germ cell types. More recent studies have paid attention to these issues and have utilised morphometric techniques to evaluate germ cell numbers resulting in the emergence of more coherent concepts. These conclusions suggest that both hormones, FSH and testosterone, are important for the achievement of full quantitative spermatogenesis, they exert independent actions but are likely to exhibit synergism particularly during experimental or pathological manipulations of testicular function (139-140). There may however be subtle species differences between species.
The role of testosterone: There is general agreement that testosterone is essential for spermatogenesis in all species but as discussed below, there is some debate as to the relative levels required. The androgen receptors are located on Sertoli cells (44,45) and the peritubular myoid cells(74) and, since they are not expressed on germ cells, the signal must be transduced by these cells, particularly the Sertoli cells.
It is unclear as to how these receptors function within the high intra-testicular concentrations of testosterone found in most mammalian species. As a result of local Leydig cell production of testosterone, in man, non-human primates and rodents, intratesticular T levels are approximately 100 fold greater than peripheral circulating levels (141-143). Consequently, androgen receptors are exposed to concentrations at which they should be saturated. Further, for spermatogenesis to be maintained in models wherein intratesticular T concentrations are experimentally manipulated, concentrations of T, well in excess of those needed to maintain androgen effects in other regions of the body, are required (144-146). Given that to date only a single androgen receptor gene has been identified, it is unclear how these differential concentrations of T required for normal androgen effects can be explained.
The relative roles of T and DHT in the maintenance of spermatogenesis are an area of interest since the affinity of DHT for the androgen receptor is greater than for T and forms a more stable complex (147,148). However in the presence of high T concentrations such as those found within the testis, the receptor-T complex is more comparable to the complex formed by DHT (147). The potential action of DHT on spermatogenesis has gained importance from the observations in several studies that, while the intratesticular T concentrations fall to about 2% of normal during gonadotrophin deprivation (142), the levels of DHT in the testis do not change significantly (142,149,150) Further, the activity of the enzyme 5α reductase, that is required to convert T to DHT, is negatively regulated by T (151). These observations may explain the relative resistance of spermatogenesis to complete suppression in men treated with T-progestin combinations for contraceptive development.
T acts at several sites during spermatogenesis but at some sites it is not possible to conclusively separate the actions of T and FSH. Several studies indicate that high testicular T concentrations impair spermatogonial proliferation in models where spermatogenesis has been damaged by agents such as irradiation (152). Additionally, there is no evidence that T can restore spermatogonial numbers in chronically gonadotrophin suppressed rats (153). In view of these data, the decrease in spermatogonial proliferation observed during gonadotrophin suppression is attributed to the lowering of FSH levels rather than T(for review 140).
During meiosis, T inhibits apoptosis and maintains spermatocyte populations in a variety of experimental models designed to investigate the control of spermatogenesis (154-156). However, during spermiogenesis, the withdrawal of T results in a premature detachment of round spermatids from the Sertoli cells leading to a marked lowering of elongated spermatid numbers within the rat testis (146,157). Emerging evidence indicates that this detachment is due to a loss of cell adhesion molecules from the junctional complexes between Sertoli cells and spermatids. Restoration of T levels rapidly reverses this process pointing to androgen action but, FSH by maintaining normal Sertoli cell function, may exert an action in this process (140). There may be some species differences in the expression of this action as in primates, including man, there is not a rapid loss of spermatid numbers associated with partial lowering of intratesticular T due to gonadotrophin suppression (158-161).
Recent studies have identified a failure of spermiation, the release of mature testicular spermatids from the Sertoli cells in many species during periods of T and/or FSH suppression. This failure in spermiation can result in the retention of up to 50% of testicular spermatids (159,161).
The role of FSH: As described earlier in this chapter, FSH exerts an important action on Sertoli cell proliferation and differentiation consistent with the presence of receptors for this hormone on Sertoli cells. Similarly, the data obtained from the initiation of testicular function in hypogonadotrophic men, indicates that FSH is required for the induction of spermatogenesis in a significant proportion (162,163). However the results of studies in rodent models of FSH deficiency have raised the possibility that FSH is not required for the completion of spermatogenesis. In mice with the targeted disruption of the FSH β subunit, spermatogenesis proceeds to completion and the mice are fertile despite a significant reduction in testis size (164). Quantitative studies showed a marked reduction in Sertoli cell number and a decrease in germ cell numbers in this model (165). Additionally, the capacity of a Sertoli cell to support germ cells was reduced in comparison to the controls suggesting that the metabolic capacity of the Sertoli cells was impaired. The knock-out of the FSH receptor gene produced a similar phenotype to the FSH β subunit knock-out model but, also showed impaired progression in spermatogenesis and defects in sperm chromatin condensation (166,167). A third model, the hpg mouse, where failure of GnRH secretion causes failure of sexual maturation, treatment with T alone restored full spermatogenesis despite smaller than normal testes (155). However it is important to note that, in these mice, spermatogenesis has proceeded to the spermatogonial stage in the absence of any hormone treatment in contrast to men with Kallmann's syndrome, where frequently, spermatogenesis has arrested at the gonocyte stage. It is therefore important to exercise caution in translating results obtained in rodents to human physiology.
In men, inactivating mutations of the FSH receptor gene can result in a variable reduction of sperm counts some of which are associated with fertility (168). Interestingly, a mutation in the FSHβ subunit gene resulted in azoospermia (169), in contrast to the normal fertility maintained in the FSHβ knock-out mouse (164).
In other experiments in normal men, Matsumoto and colleagues (170-171) suppressed gonadotrophins by the administration of T until suppression of spermatogenesis occurred. They then introduced injections of hCG to stimulate Leydig cell function and to restore intratesticular T concentrations which increased sperm counts but not to pre-treatment levels (Figure 15). These data suggested that, in association with undetectable FSH levels, increasing intratesticular T could restore sperm output partially (171). Using the same model, they initiated hFSH treatment when sperm counts were suppressed and showed that, in the presence of low intratesticular T concentrations, FSH alone could partially restore sperm output (172). The latter study strongly suggests a role for FSH which appears to be able to synergise with low T to stimulate sperm production. This data supports the results of studies that showed a reduction in the levels of T required to maintain spermatogenesis in the presence of FSH (173-175).
Figure 15. The response in the sperm counts from normal volunteers to a suppression of FSH and LH by testosterone injections is shown. Note the recovery in sperm counts when hCG and hFSH were introduced singly into the treatment regime. Data from Matsumoto et. al. (reference 171, 172) and Bremner et. al. (reference 172).
The emerging evidence for a role of FSH in the control of spermatogenesis has been further supported by data arising from other studies particularly in human and non-human primates. In long term immunization of monkeys against FSH, a reduction of both sperm production and quality was noted (176).
Several studies have established that FSH regulates the adult spermatogonial population in rodents and primates (154,177,178). In one experiment, the passive neutralisation of FSH resulted in decreases in type A spermatogonia at stages XIV-I of the rat spermatogenic cycle (177). In normal men, gonadotrophin suppressed induction of decreased sperm counts resulted in marked decreases in type B spermatogonia (160) raising the possibility of a diminished progression from A pale to B. As discussed earlier, this effect is more likely to result from the decrease in FSH rather than the concomitant decline in T. The results of more recent studies in primates and men reviewed by McLachlan et al (140) indicate a regulation of both Ap and B spermatogonia by FSH.
While several studies have suggested that FSH can prevent the degeneration and loss of spermatocytes during meiosis, studies in men were unable to show a decrease in spermatocyte numbers over that to be expected from the decline in the spermatogonial population (160-161). During spermiogenesis, decreased FSH action does not effect progression to spermatids but FSH has been shown to be important in spermatid Sertoli cell adhesion (179,180). However, as discussed in the section on T, decreasing gonadotrophic stimulation of the testis results in a marked inhibition of spermiation (140,159,161). While this process is impaired by both T and FSH individually, they act synergistically by retaining about 50% of late testicular spermatids. This synergistic action may explain the partial restoration of sperm counts seen in the study by Bremner et al (172) when FSH treatment was given in the presence of low intratesticular T. In terms of relative gonadotrophic effects on germ cell populations, a recent study showed that exogenous FSH or hCG were similarly effective in partially maintaining germ cell number during the onset of experimentally-induced gonadotrophin suppression in normal men although improved support for pachytene spermatocyte number by FSH was found (180a).
As briefly discussed earlier in this chapter, the successful initiation of testicular function is dependent on the hypothalamic secretion of GnRH which in turn stimulates FSH and LH to act on the testis. These actions initiate spermatogenesis and testosterone production.
It is well recognised that the testis in turn, through the secretion of hormones produced in the Sertoli and Leydig cells, exerts a negative feedback control on the production of gonadotrophins. The presence of such a negative feedback control by the testis on pituitary FSH and LH secretion is best demonstrated by the rapid rise of FSH and LH after castration. The mechanisms by which the secretion of FSH and LH increases in response castration involves a rise in the hypothalamic secretion of GnRH and also involves direct actions at the pituitary level which allow an increase in pulse amplitude. Further, the fact that LH and FSH are co-secreted by the majority of gonadotrophs raises a number of questions as to how GnRH and the inhibitory signals act on the pituitary to result in differential regulation of FSH and LH secretion.
There is a substantial body of evidence to indicate that the steroid hormones testosterone, estradiol and dihydrotestosterone inhibit LH secretion (181). The demonstration that non-aromatisable androgens could inhibit LH secretion resolved the argument as to whether testosterone exerts its action directly or whether metabolism to estradiol or dihydrotestosterone was necessary (182,183). From the studies by Santen and Bardin(184), it is evident that testosterone acts at the hypothalamic level by decreasing GnRH pulse frequency without a change in pulse amplitude. However, the action of estradiol appears to be predominantly at the pituitary where it decreases LH pulse amplitude without changing pulse frequency. Further support for the action of testosterone at the hypothalamus emerged from the observation of a decrease in GnRH pulse frequency in portal blood (185). In addition, these studies demonstrated that treatment with estradiol lowered LH levels by decreasing LH pulse amplitude without altering GnRH secretory patterns in portal blood.These conclusions have been challenged by observations that a selective aromatase inhibitor, anastrozole, caused an increase in LH pulse amplitude and pulse frequency (186). These changes were seen in the presence of increased testosterone concentrations and were accompanied by an increase in LH and FSH. The investigators concluded that estradiol exerted a negative feedback by acting at the hypothalamus to decrease GnRH pulse frequency and at the pituitary to decrease the responsiveness to GnRH.
There is a substantial body of evidence to indicate that testosterone and estradiol are capable of suppressing FSH in the male (187). For many years, it was proposed that the action of the steroid hormones could account for the entire negative feedback exerted on FSH levels by the testis despite the existence of a hypothesis that a specific FSH feedback regulator named inhibin existed (188).
Over the past twenty years, a substantial body of evidence has accumulated to confirm the existence of a glycoprotein hormone termed inhibin that exerts a specific negative feedback inhibition on FSH secretion at the pituitary level (189). Two forms of inhibin have been isolated namely inhibin A and inhibin B(190-193). These proteins represent disulphide-linked dimers of an α and ß subunit. The alpha subunit is common both to inhibin A and B but the ß subunit, though closely related, are different (α βA = inhibin A: aßB = inhibin B). Both inhibin A and inhibin B have the capacity to specifically inhibit FSH secretion by pituitary cells in culture. In contrast, dimers of the ß subunit, termed activins (activin A = ßAßA: activin B = ßBßB; activin AB = ßAßB) all have the capacity to stimulate FSH secretion by pituitary cells in culture (194-195). Finally, a structurally unrelated protein termed follistatin, has the capacity to suppress FSH secretion specifically by pituitary cells in culture (196-198). This action has been demonstrated to be due to the capacity of follistatin to bind and neutralize the actions of activin thereby suppressing FSH secretion(199).
a. Role of testosterone: In men and males from other species testosterone, when administered in an amount similar or greater to its production rate can suppress FSH as well as LH (181). However, in most instances there was a parallel and often greater suppression of LH secretion in contrast to the actions of inhibin (187). Further, there appears to be a difference in the response of FSH to testosterone in primates, where the actions are totally inhibitory in contrast to rats, where following an initial suppression of FSH by testosterone, higher doses caused a return of FSH levels to baseline (144,200). Clear evidence for a physiological role of testosterone in the control of FSH can be shown in experiments in which the Leydig cells were destroyed by the cytotoxin ethane dimethane sulphonate (EDS). This treatment results in a rapid decline in testosterone levels and a concomitant increase in FSH concentrations to levels which were only 50% of those found in castrates (201). Since the inhibin levels in these experiments did not change, the maintenance of FSH levels at 50% of those seen in castrate animals was likely to be due to the continuing feedback control by inhibin (202). Further support for the dual role of testosterone and inhibin in the control of FSH emerged from the use of EDS in cryptorchid rats where baseline FSH levels were increased in association with decreased inhibin concentration. The removal of testosterone feedback in these animals with low basal inhibin levels resulted in an increase in FSH to the castrate range (203). The observation of an increase in FSH levels in men treated with a selective aromatase inhibitor raised the possibility that estradiol exerts a negative feedback action on FSH especially since the treated men experienced a concomitant significant increase in testosterone (186)
b. Role of inhibin, activin and follistatin: The predominant evidence indicates that in the male, inhibin is produced by the Sertoli cell and is secreted both basally across the basement membrane of the seminiferous tubule and also into the lumen (204,205). Several studies have now demonstrated that the predominant form of inhibin secreted by the testis is inhibin B since the predominant mRNA was ßB (206,207). The levels of inhibin B in males, measured by a specific ELISA, are inversely related to the levels of FSH (208,209). Several studies have indicated that FSH predominantly stimulates inhibin α subunit production and does not alter the ß subunit message (204,210). This action results in the testis predominantly secreting inhibin rather than activin. Further support for this concept emerges from the studies of men undergoing chemotherapy where declining inhibin B levels are associated with a rise in FSH. However, with assays that detect α subunit products, there was a clear increase in these substances under the stimulation of elevated FSH levels (211). There is also evidence that a subunit of inhibin can be produced by Leydig cells (212) and increased LH levels result in the release of a subunit products into the circulation (213,214). There is still controversy as to whether the Leydig cells can produce bioactive inhibin (212).
In man, testosterone-induced gonadotrophin suppression reduced circulating inhibin B and pro alpha C levels by only 25% and 50%, respectively, indicating that their secretion is not fully gonadotrophin-dependent (214a). In that model, exogenous FSH and LH both restorted pro alpha C levels supporting Sertoli and Leydig cell origins of alpha subunit peptides,
respectively, but only FSH restored inhibin B presumably reflecting Sertoli cell ßB synthesis.
While there is evidence that the Sertoli cells, Leydig cells and peritubular myoid cells can produce activin, castration does not result in a decrease in circulating activin A levels (215-218). Unfortunately, due to the lack of a suitable assay to measure activin B, there is no data available concerning the behaviour of this substance after castration. There is considerable evidence that activin may exert local actions within the testis such as the stimulation of spermatogonial mitosis (219) and a variable action on rat Sertoli cells dependent upon the age of the animals from which the cells were isolated (76,77). Additionally, receptors for activin are present on primary spermatocytes, round spermatids and Sertoli cells (220).
Follistatin is also produced in the Sertoli cells, spermatogonia, primary spermatocytes and round spermatids in the testis (221,222). However, castration does not result in a net decrease in follistatin levels in the circulation suggesting that the testis does not contribute significantly to circulating levels of follistatin (223).
The failure of activin and follistatin to change after castration whereas the inhibin levels in the circulation decreased to undetectable levels, strongly suggest that the gonadal feedback signal on FSH secretion is inhibin. This view has been further supported by the studies of Plant and colleagues in primates where they showed that in arcuate nucleus-lesioned monkeys maintained on a constant GnRH pulse regime, testosterone could prevent the post-castration rise in LH but not FSH (224). Further, in several species the infusion or injection of recombinant human inhibin caused a specific fall in FSH secretion commencing some six hours following administration (225-227). In further studies, inhibin A, sufficient to restore circulating levels in castrate rams to normal, suppressed FSH levels into the normal range in the absence of testosterone (228).
The role of these proteins appears to be more complex since substantial evidence exists that activin and follistatin can exert a paracrine role directly in the pituitary gland. The α and ß subunit mRNAs are present in gonadotrophs within the pituitary gland(229). The studies of Corrigan et al(230) strongly suggest that these substances exert a local action on FSH secretion since the inhibition of the action of activin B by the use of a specific monoclonal antibody when added to pituitary cells in culture, caused a suppression of endogenous FSH secretion. Follistatin mRNAs are also present in a number of different pituitary cell types including the folliculo-stellate cells (229,231). This local production of follistatin also has the capacity to regulate the actions of activin(232). Additionally, the studies of Bilizekian et al have demonstrated that GnRH and the sex steroids estradiol and testosterone can modulate the local production within the pituitary of α, ßA , ßB and follistatin mRNAs (233,234). Clearly these interactions are complex and no clear answer can be given as to the relative roles of paracrine and endocrine actions of these glycoprotein hormones.
Some correlative evidence supporting the action of inhibin on FSH secretion is the decrease in production of inhibin by Sertoli cells in parallel with the rise of FSH in a number of models of spermatogenic damage such as cryptorchidism and intra-testicular glycerol treatment (235,236). Further, the levels of circulating inhibin B appear to be inversely related to the levels of FSH to the levels of FSH following testicular damage in a number of studies (208,209,237,238). Further, even in studies of large numbers of normal men, there is an inverse relationship between serum inhibin B levels and FSH (237). It is likely therefore that the actions of inhibin are predominantly exerted through secretion from the testis and transport via the peripheral circulation whereas the actions of activins and follistatin on FSH secretion occur through paracrine actions at the level of the pituitary gland. Further evidence supporting the stimulation of FSH by activin secretion emerges from the decline in FSH levels in mice with targeted disruption of the activin type II receptor gene(239).