Sterols in spermatogenesis and sperm maturation.

Mammalian spermatogenesis is a complex developmental program in which a diploid progenitor germ cell transforms into highly specialized spermatozoa. One intriguing aspect of sperm production is the dynamic change in membrane lipid composition that occurs throughout spermatogenesis. Cholesterol content, as well as its intermediates, differs vastly between the male reproductive system and nongonadal tissues. Accumulation of cholesterol precursors such as testis meiosis-activating sterol and desmosterol is observed in testes and spermatozoa from several mammalian species. Moreover, cholesterogenic genes, especially meiosis-activating sterol-producing enzyme cytochrome P450 lanosterol 14α-demethylase, display stage-specific expression patterns during spermatogenesis. Discrepancies in gene expression patterns suggest a complex temporal and cell-type specific regulation of sterol compounds during spermatogenesis, which also involves dynamic interactions between germ and Sertoli cells. The functional importance of sterol compounds in sperm production is further supported by the modulation of sterol composition in spermatozoal membranes during epididymal transit and in the female reproductive tract, which is a prerequisite for successful fertilization. However, the exact role of sterols in male reproduction is unknown. This review discusses sterol dynamics in sperm maturation and describes recent methodological advances that will help to illuminate the complexity of sperm formation and function.

FF-MAS and related MAS analogs possess an obvious dosedependent meiosis-activating effect in vitro. In addition, MAS compounds were demonstrated to have a positive infl uence on cytoplasmic maturation, an important process occurring during the fi nal steps of oocyte development that enables completion of nuclear maturation, fertilization, and proper early embryo development. In contrast to the initially proposed role of MAS in the resumption of meiosis, it was suggested that MAS might have another role in meiotic nuclear processes and cytoplasmic maturation ( 6 ); however, the exact mechanism of MAS in female reproduction remains controversial. early hypothesis suggested that MAS produced by cumulus cells in response to a gonadotropin surge during ovulation serve as paracrine factors to stimulate the resumption of oocyte meiosis ( 4 ). MAS were initially considered to be important in the treatment of infertility and in the development of novel contraceptive drugs that would not be based on sex steroids or sex steroid-like compounds ( 4 ); however, contradictory data indicated that MAS are not essential for the resumption of meiosis mediated by gonadotropin hormones ( 5 ). Subsequently, eight endogenous MAS molecules with similar structures and equipotent meiosis-activating activities were identifi ed ( 6 ) and demonstrated that  ( 9 ). Discordant regulation of cholesterogenic enzymes is a possible reason for the accumulation of T-MAS in the testis of rat (a), guinea pig (b) ( 12 ), and mouse (c) ( 13 ), compared with the quantity of T-MAS in the liver. Three-dimensional structures of lanosterol, FF-MAS, desmosterol, and cholesterol obtained with a Pc3D molecule viewer display the spatial conformation of individual intermediates, which might affect sperm membrane properties. testes and sperm from various mammalian species, providing an excellent argument for the potentially important role of MAS in sperm development.

UNIQUE TESTICULAR EXPRESSION PATTERN OF CYP51 ENCODING THE MAS-PRODUCING ENZYME
Substantial support for the role of MAS in spermatogenesis came from studies on lanosterol 14 ␣ -demethylase, which encodes the CYP51 enzyme that is known to catalyze the conversion of lanosterol to FF-MAS. CYP51 is a unique member of the cytochrome P450 (CYP) superfamily. Of the more than 10,000 cytochrome P450s in existence, it is the most widely distributed and the only one present in all biological kingdoms (http://drnelson. uthsc.edu/P450.statsfi le.html). From an evolutionary standpoint, Cyp51 is the most conserved gene within the P450 superfamily ( 14 ). Mammalian CYP51 from different species displays over 90% sequence identity at the protein level ( 15 ). Moreover, CYP51 from distant eukaryotic organisms has been found to be strikingly similar at the structural level (16)(17)(18)(19). These data indicate that CYP51 function has remained highly conserved throughout evolution, suggesting an essential role of this gene in organisms of various taxa. In mammals, CYP51 catalyzes an essential late step in cholesterol biosynthesis, the demethylation of lanosterol and 24,25-dihydrolanosterol into the intermediate FF-MAS, which is further converted into T-MAS by one of the enzymes with sterol-⌬ 14-reductase activity, transmembrane 7 superfamily member 2, or lamin B receptor. At least seven additional enzymatic steps are required to synthesize cholesterol ( Fig. 1 ). The essential role of Cyp51 in de novo cholesterol synthesis and embryo development in vivo has recently been demonstrated by our group ( 20 ). In addition to its role in cholesterol synthesis, Cyp51 is thought to have an essential role in reproduction as an enzyme producing the intermediate FF-MAS. This function was fi rst suggested when MAS was found to have meiosisactivating potency ( 3 ). To investigate the proposed role of Cyp51 in reproduction, the expression pattern of Cyp51 in testes from several mammalian species was examined. Northern analyses using various human tissues revealed high levels of Cyp51 transcripts in the testis primarily due to the synthesis of additional shorter testisspecifi c transcript ( 21 ). Accumulation of this testis-specifi c Cyp51 transcript was later confi rmed in sexually mature rats, in contrast to prepubertal animals ( 22 ). Interestingly, in situ hybridization and northern analysis using testis cross-sections and different testicular cell fractions, respectively, revealed stage-specifi c expression of Cyp51 in germ cells. Cyp51 expression was lowest in pachytene spermatocytes but increased, reaching its highest level in elongating spermatids; however, Cyp51 mRNA transcripts were not detected in most elongated spermatids that line the luminal edge of the seminiferous epithelium. With increasing accumulation of the shorter testis-specifi c transcript, the expression of the ACCUMULATION OF MEIOSIS-ACTIVATING STEROLS IN THE TESTIS AND SPERM Although MAS function in oocytes has been examined extensively, studies investigating its function in spermatogenesis are scarce. Early in vitro experiments demonstrated that a diffusible component present in adult mouse testes triggers nondifferentiated fetal male germ cells to enter meiosis ( 7,8 ). After its identifi cation in 1995, the role of MAS in male germ cell development was often suggested, but its causal relationship was never demonstrated conclusively in vivo or in vitro . Most studies aimed to quantify MAS in the male reproductive tract or the expression levels of MAS-producing enzyme cytochrome P450 14 ␣demethylase (CYP51) in testes from different mammalian species and to correlate these parameters with reproductive function. First reports of MAS in adult testes of bull, mouse, and horse demonstrated that MAS accumulate at concentrations above 30 g/g (parts per million [ppm]) ( 3,4 ); however, lower concentrations of MAS were detected in isolated rat seminiferous tubules (15 ppm) ( 4 ). Nevertheless, T-MAS was found to be the predominant sterol intermediate in the rat testis, with a 40-fold higher concentration than in the liver ( Fig. 1 ). During sexual maturation (i.e . , 19-70 days of age), the concentration of T-MAS in the rat testis also increased 8-fold (i.e . , 6.1-49 ppm) ( 9 ). Another group demonstrated a 4-fold increase of T-MAS in the rat testis during puberty ( 10 ). No other sterol precursor has shown comparable changes in testis during puberty, implying the important role of T-MAS in spermatogenesis.
Accumulation of T-MAS was further demonstrated in sexually maturing stallions. A study on 16 stallions aged 1-12 years demonstrated a positive correlation between testicular T-MAS concentration and testicular weight as well as a negative association between testicular T-MAS concentration and cryptorchidism after puberty. Testis of mature studs contained a 7-fold higher concentration of T-MAS than that of immature males (19.3 vs. 2.8 ppm) ( 11 ). The concentration of T-MAS in isolated spermatozoa has been only measured in humans, where it was 4-fold higher (2 ppm) than in the ejaculate (< 0.5 ppm), illustrating accumulation of MAS in sperm cells and not in seminal plasma ( 4 ). However, measurements of MAS and several other sterol intermediates in guinea pig testes revealed that, by quantity, T-MAS ranked third (15.6 ppm) after lathosterol (23.5 ppm) and desmosterol (21.4 ppm). The concentration of T-MAS in the testis was at least 15fold higher than in the liver ( Fig. 1 ) ( 12 ). In the adult mouse testis (10-12 weeks), T-MAS was shown to be a predominant sterol (71 ppm), followed by desmosterol (15 ppm). The concentration of T-MAS in testis was at least 70-fold higher than in the liver ( Fig. 1 ) ( 13 ). These results obtained by HPLC are not completely in line with recent gas chromatography-mass spectrometry-based sterol profi ling studies that used mice testes of the same age because desmosterol was found to be higher than T-MAS (52 vs An additional argument to support the role of MAS in spermatogenesis comes from the unique testicular expression pattern of other cholesterogenic genes upstream of CYP51. Besides Cyp51 , testis-specifi c expression was observed also for farnesyl diphosphate synthetase ( Fdps ) and farnesyl diphosphate farnesyl transferase 1 or squalene synthase ( Fdft1 ). Fdps is expressed as a longer, developmentally regulated testis-specifi c transcript that localizes specifi cally to postmeiotic round spermatids in rat ( 29 ). Similarly, a longer testis-specifi c transcript was described for Fdft1 ( 30 ). Fdft1 mRNA levels were highest in round spermatids and decreased in elongated spermatids ( 22 ). During spermatogenesis, Fdft1 expression peaked earlier than that of Cyp51 ( Fig. 2 ), which is in line with the fact that Fdft1 acts upstream of Cyp51 . A protein analysis study showed that Fdft1 translation is moderately delayed, similar to Cyp51 ( 24 ). As mentioned previously, POR is expressed exclusively in late stages of spermatogenesis. Taken collectively, these results demonstrate that cholesterogenic proteins are not constantly and coordinately expressed during spermatogenesis, as is the case in other tissues where they provide a housekeeping role in cholesterol synthesis. Indeed, the transcriptional regulation of cholesterogenic genes in the testis is different from cholesteroldependent regulation by the sterol regulatory element binding protein pathway in the liver ( 23 ). One example is the unique germ cell-specifi c regulation of Cyp51 by cAMP-responsive element modulator (CREM ). CREM acts as a master switch that is responsible for the cAMPdependent transcriptional activation of specifi c genes in postmeiotic male germ cells, including Cyp51 ( 31 ). This regulatory mechanism is responsible for the presence of the shorter Cyp51 transcript in the testis ( 23 ). Crem Ϫ / Ϫ mice lacked germ cell-specifi c Cyp51 mRNA in testis, whereas the expression of somatic Cyp51 and testis-specifi c Fdft1 remained unaffected. In a recent largescale expression analysis study of Crem Ϫ / Ϫ mice, Crem was found to be involved in several aspects of spermatogenesis ( 32 ). Based on these results, one could speculate that the role of Cyp51 in germ cell development is indispensable.
Another support for the role of MAS in spermatogenesis comes from the comparison of expression patterns between preMAS and postMAS genes ( 9 ). This study showed that during sexual development of the rat testis, preMAS genes were up-regulated, whereas postMAS genes were expressed at low levels, unchanged, or downregulated with increasing age. These results also implied that other, yet-to-be-defi ned transcription factors besides cAMP/CREM might be involved in the control of cholesterogenic genes or sets of genes during spermatogenesis. Microarray data on the expression of cholesterogenic genes in mouse spermatogenic fractions (http://mrg.genetics.washington.edu) confi rmed the high accumulation of presqualene Fdft1 and squalene longer somatic transcripts declined ( 22 ). Moreover, a similar stage-specific expression pattern of Cyp51 was later confirmed in mouse ( 23 ) and human testes ( 9 ). Only background levels of Cyp51 mRNA were detected in steroidogenic Leydig cells. Studies showing high evolutionary conservation of the Cyp51 gene and a unique mRNA expression profile in testis suggested an important role of Cyp51 (and its product MAS) in reproduction.
An immunolocalization study revealed CYP51 to be highest in Leydig cells and round and elongated spermatids in the rat testis ( Fig. 2 ) ( 24 ). When these findings are interpreted together with the aforementioned expression results, we can propose that translation of Cyp51 mRNA most likely occurs without delay and continues in later stages of the epithelial cycle that lack mRNA synthesis. Translation of the testis-specific Cyp51 mRNA might be less efficient to supply a template for its protein synthesis during the final stages of spermatogenesis. On the other hand, CYP51 translation in Leydig cells where low expression resulted in an abundance of protein appears to be highly efficient, in contrast to germ cells. CYP51 protein was not restricted to the cytoplasm of the developing germ cells where the endoplasmic reticulum resides but was also seen in the acrosomal regions of round and elongated spermatids and in residual bodies ( 25 ). The presence of CYP51 protein in residual bodies has led to the hypothesis that MAS synthesis in haploid germ cells might be important for the initiation of meiosis in premeiotic stages. If the synthesis of MAS in postmeiotic germ cells is important for the regulation of meiosis during early spermatogenesis, MAS would need to be reversely transported to premeiotic germ cells. One possible transport route would be via residual bodies that are formed during spermiogenesis and that contain excess cytoplasm and organelles. Residual bodies are phagocytosed by surrounding Sertoli cells and transported toward the basal membrane ( 26,27 ). Transport of compounds contained in residual bodies from the adluminal to the basal cell compartment has been reported previously ( 28 ). In addition to CYP51, the expression of cytochrome P450 oxidoreductase (POR), which serves as an obligate CYP51 electron donor, was examined in the testis. POR protein was present in Leydig cells, residual bodies, and late-stage elongated spermatids ( Fig. 2 ). The entire lanosterol 14 ␣ -demethylase complex (CYP51 and POR) has also been detected in the acrosomal membranes of round and elongated spermatids in the mouse ( 25 ). In addition, conversion of lanosterol to FF-MAS was observed in isolated acrosomal membranes from ejaculated bull sperm, suggesting that spermatozoa might produce FF-MAS in situ ( 25 ). It was therefore speculated that spermatid-produced MAS sterols might trigger the resumption of meiosis in quiescent oocytes upon fertilization. This hypothesis was also supported by the fact that the highest concentration of MAS was not found in the ovary but in the testis and spermatozoa ( 4 ).

SYNTHESIS OF CHOLESTEROL IN GERM CELLS DURING SPERMATOGENESIS
In contrast to MAS intermediates, the role of cholesterol in sperm maturation and male fertility has been well studied. The underlying imperative was that one could obtain insight into the role of cholesterol in sperm development by studying the dynamic cellular events that control its level during spermatogenesis and after spermiation. Due to the extensive production of germ cells throughout spermatogenesis, the requirement for cholesterol is signifi cant. During sperm release from the seminiferous epithelium, sperm plasma membranes epoxidase ( Sqle ) RNA in round spermatids ( 33 ). Based on the lack of a coordinated transcriptional regulation and a discrepancy between preMAS and postMAS gene expression, it was suggested that the primary role of male germ cells may not be to synthesize cholesterol but to produce MAS ( 9 ). This hypothesis is well supported by an 8-fold increase in T-MAS and a 43% decrease of cholesterol during sexual maturation in rat. The stagespecifi c expression pattern of Cyp51 and the differential expression of other cholesterogenic genes in germ cells suggest that a specifi c regulatory mechanism might exist to facilitate the accumulation of MAS intermediates in the testis.  ( 34 ). The stage-specifi c expression of cholesterogenic enzymes farnesyl diphosphate farnesyl transferase 1 (FDFT1), CYP51, and POR during spermatogenesis is displayed in gray boxes (left) ( 24 ). Supporting Sertoli cells provide an additional source of cholesterol for spermatogenesis. Sertoli cells acquire cholesterol by de novo synthesis from acetate ( 35 ) or import external cholesterol from HDL ( 37 ) by specialized cholesterol transporters ( 36 ). Another important source of cholesterol in Sertoli cells might be from the recycling of lipid-rich residual bodies and apoptotic germ cells ( 39 ). Excess cholesterol can be esterifi ed to cholesterol esters (CE) and stored in lipid droplets, serving as cholesterol reservoirs ( 40 ). Some cholesterol can be effl uxed to HDL by reverse cholesterol transport ( 41 ). Spermatozoa formed in the testis enter the caput epididymis and progress to the caudal region. The epididymis possesses the ability for de novo cholesterol synthesis ( 51 ), or it can import cholesterol from the circulation ( 59 ). Cholesterol might be effl uxed into the epididymal lumen by ABCA1 and ATP-binding cassette sub-family G member 1 (ABCG1) ( 53 ). The principal cells of the epididymis secrete small membranous vesicles known as epididymosomes, which could serve as a source for cholesterol exchange with maturing sperm cells ( 128 ). The cholesterol content of sperm membranes is decreased during epididymal transit in several species, resulting in the decreased ratio of cholesterol (CH) to phospholipids (PL) ( 48 ). The loss of cholesterol results in an increase in sperm membrane fl uidity and sperm motility ( 71 ). The content of cholesterol in the sperm membrane is further decreased in the female reproductive tract, mostly by the process of capacitation. Albumin serves as a cholesterol acceptor during capacitation. Prostasomes present in the ejaculate are able to fuse with the membranes of spermatozoa, to increase motility, and to prevent early acrosome reaction ( 73 ). Individual data were obtained from in vitro and in vivo experiments on different mammalian species and used to collate this scheme. in a process called "reverse cholesterol transport." This process relies on the presence of cholesterol transporters of the ATP-binding cassette superfamily, mainly ATP-binding cassette transporter 1 (ABCA1) ( 41 ). Perturbations in Sertoli cell cholesterol homeostasis resulted in complete infertility or subfertility caused by the excessive accumulation of cholesterol esters. This was demonstrated by knockout mouse models lacking functional nuclear receptors involved in cellular cholesterol homeostasis, such as retinoid X receptor ␤ (gene symbol Rxrb ) ( 42,43 ), and by Abca1 knockout mice ( 41 ). In the case of liver X receptor (synonym symbol Lxr , offi cial symbol Nr1h ) double knockout mice ( Lxr ␣ ; ␤ Ϫ / Ϫ ), the accumulation of cholesterol esters was due to an ABCA1-dependent decrease in cholesterol effl ux from Sertoli cells, ABCA1 being a well-known LXR target gene ( 44 ). Therefore, the role of Sertoli cells in the regulation of cholesterol homeostasis seems to be very dynamic and to involve mechanisms for cholesterol uptake, storage, recycling, and effl ux, which we schematically summarize in Fig. 2 . To conclude, Sertoli cells should be considered in the interpretation of results relating to spermatogenesis, including the potential effects of MAS intermediates on germ cell development.

IMPORTANCE OF CHOLESTEROL DURING SPERM MATURATION IN THE EPIDIDYMIS
When sperm is released from the seminiferous epithelium, it is immotile and unable to fertilize an oocyte. Full fertilizing capacity of spermatozoa is acquired during their transport through the epididymis and in the female reproductive tract by capacitation. Capacitation is a multistep process whereby activation of bicarbonate-dependent adenylyl cyclase results in the elevation of cAMP and protein kinase A-mediated tyrosine phosphorylation of a subset of fl agellar proteins, leading to changes in sperm motility and acrosomal responsiveness (45)(46)(47). These events are highly dependent upon changes in plasma membrane cholesterol ( 48 ). Spermatozoa formed in the testis enter the caput epididymis, progress to the corpus, and fi nally reach the caudal region, where they are stored. Only sperm cells collected from the cauda epididymis of different species have full fertilizing capacity ( 49 ). An accepted hypothesis is that cholesterol excreted from the epididymal epithelium contributes to the maturation of transiting sperm ( 50 ). Furthermore, the epididymis possesses the ability to synthesize de novo cholesterol ( 51,52 ). Similar to Sertoli cells, cholesterol traffi cking was shown to involve an LXR-regulated mechanism in the mouse epididymis. Effl ux of cholesterol into the lumen is likely to be mediated by the complementary action of ABCA1 and ABCG1, which are well known LXR targets ( 53 ). It is interesting to note that epididymal epithelial cells use a unique apocrine mode of secretion to produce small membranous vesicles called epididymosomes ( Fig. 2 ) ( 54-56 ), mainly composed of sphingomyelin and various amounts of cholesterol ( 57,58 ). Recently, an important role for dietary cholesterol in are believed to be highly loaded with cholesterol. Early experiments examining the role of de novo cholesterol synthesis revealed elevated rates of 14 C acetate incorporation into cholesterol during the leptotene, zygotene, and pachytene spermatocyte stages of development ( 34 ). This implies that increased de novo cholesterol synthesis is ongoing in these germ cells, which correlates well with the increase in germ cell diameter and surface area during development, requiring the synthesis of new plasma membranes. The capacity to synthesize cholesterol decreased in late pachynema and remained low in all subsequent stages of spermatogenesis, including mature sperm. From the end of the pachytene stage, spermatocytes do not increase signifi cantly in size but differentiate to diplotene spermatocytes, which undergo meiotic divisions to form round spermatids. In contrast to cholesterol, the rate of acetate incorporation into dolichol remained elevated in late pachytene spermatocytes and round spermatids and then decreased and remained low in mature sperm ( 34 ). This study suggests that the early stages of cholesterol synthesis in pachytene spermatocytes and round spermatids might precede the synthesis of dolichol, which is a crucial component in protein glycosylation and the production of specifi c membrane glycoproteins. However, the exact role of de novo cholesterol synthesis in germ cells has not been determined conclusively, mainly due to the complex relationship between the germ cells and supporting Sertoli cells.

CHOLESTEROL SYNTHESIS AND TRANSPORT IN SERTOLI CELLS
The majority of nutrients, including lipids, needed for spermatogenesis are provided by supporting Sertoli cells, which have the capacity to synthesize cholesterol from acetate in vitro ( 35 ). Because the amount of cholesterol required to support spermatogenesis far exceeds the biosynthetic capacity of Sertoli cells, specialized cholesterol transporters are likely to facilitate the infl ux of cholesterol from the circulation ( 36 ). The basal membrane, which separates seminiferous tubules from blood capillaries present in the interstitial space, blocks the entry of LDLs but allows the entry of HDLs, the major source of cholesterol for Sertoli cells ( 37,38 ). Rat Sertoli cells have the ability to uptake HDL cholesterol mainly by APOE-dependent pathways ( 38 ). Another important source of cholesterol might be the recycling of lipid-rich residual bodies and apoptotic germ cells that are constantly phagocytosed by Sertoli cells ( 39 ). To avoid toxicity, excess cholesterol is esterifi ed and stored in lipid droplets, which are highly mobile and dynamic structures involved in the maintenance of cholesterol equilibrium. Lipid droplets are known to communicate with intracellular organelles and cell junctions through stable or transient surface contacts, enabling the effi cient transfer of cholesterol from the site of synthesis and uptake to the site of usage or elimination ( 40 ). Another important aspect of cholesterol homeostasis in the Sertoli cell is the effl ux of excess cholesterol to HDL proteins. During capacitation, ZP-binding proteins cluster within lipid rafts present in the apical ridge area of the sperm head plasma membrane ( 75 ). An obligatory event during capacitation is the loss of cholesterol from the sperm plasma membrane ( 76 ). Cholesterol effl ux is promoted by albumin ( Fig. 2 ) and high bicarbonate levels, which facilitate lipoprotein-mediated cholesterol effl ux (77)(78)(79). Cholesterol removal was restricted only to the nonraft membrane fraction of the sperm plasma membrane, whereas the content of cholesterol in the raft membrane fraction remained unchanged ( 75 ). These data suggest that cholesterol depletion is important in the remodeling of lipid rafts on the sperm surface ( 80 ). Simultaneous with cholesterol effl ux, phospholipids inserted into the sperm membrane and decreased the cholesterol to phospholipid ratio ( Fig. 2 ) ( 48 ). The fatty acid composition of sperm during maturation is also highly modifi ed, similar to sterols ( 81 ). These combined events increase membrane fl uidity and facilitate Ca 2+ infl ux into sperm, triggering the acrosomal reaction and fusion with the oocyte ( 82 ).
Cholesterol content within sperm plasma membranes varies signifi cantly among different species. In humans, the amount of cholesterol in sperm varies considerably even among ejaculates ( 83 ). However, current evidence suggests that the proportion of cholesterol present within sperm membranes is directly related to human sperm morphology ( 84 ) and fertility potential ( 83 ). The role of cholesterol in reproduction was also demonstrated in a mouse knockout model with complete ablation of the 24-dehydrocholesterol reductase gene ( Dhcr24 ), which is involved in the conversion of desmosterol to cholesterol. These animals contained only trace amounts of cholesterol and were unable to reproduce ( 85 ).

OTHER STEROL INTERMEDIATES IN MALE REPRODUCTION
T-MAS was suggested to have an important biological role in male reproduction because it was found to be a predominant intermediate in the testis, but only in a few cases ( 3,9,13 ). By contrast, desmosterol (cholesta-5,24dien-3 ␤ -ol) was identifi ed as a major sterol intermediate in guinea pig ( 12 ) and mouse (Keber et al., unpublished data) testes as well as in human ( 86 ) and boar ( 70 ) spermatozoa. Furthermore, monkey ( 87 ) and rabbit ( 88 ) sperm membranes were found to be signifi cantly richer in desmosterol than in cholesterol. Desmosterol was the predominant sterol found in all subfractions of rabbit semen (56.7% of total sterols in prostatic secretory granules, 63.8% in the seminal plasma, and 60% in spermatozoa) ( 88 ). The proportion of desmosterol in sperm varies depending on the species, representing 59% of total sterols in rhesus monkey and rabbit ( 87,88 ) and 25% of total sterols in human ( 89 ). In rhesus monkey testes, the desmosterol concentration was found to increase with animal age ( 90 ). High levels of desmosterol might be the result epididymal sperm maturation was suggested ( 59 ). Dietary cholesterol overload in an LXR-defi cient mouse model resulted in complete infertility caused by the accumulation of cholesterol ester lipid droplets in smooth muscle cells lining the epididymal duct. Spermatogenesis was not affected, as shown by normal sperm count, testicular weight, and histology. It should also be noted that a high cholesterol diet signifi cantly reduced testosterone levels in double knockout Nr1h2 Ϫ / Ϫ ;Nr1h3 Ϫ / Ϫ animals ( 59 ). Unfortunately, the direct effect of dietary cholesterol intake on the endogenous cholesterol concentration in sperm plasma membranes of wild-type animals has not been demonstrated, as it was for fatty acids (60)(61)(62). Cholesterol content in sperm traveling through the epididymal tract may be regulated by human epididymisspecifi c proteins (HE1-HE6), initially described in the human epididymis ( 63 ). The best studied member of this group is HE1 or Niemman-Pick C2, a secreted sterol-binding glycoprotein involved in sterol traffi cking ( 64 ). The HE1 homolog has been suggested to be one of the important regulators of cholesterol content in sperm membranes at least in some species ( 65 ). During epididymal transit, the content of cholesterol in sperm membranes decreased by approximately 50% in several species ( 57,(66)(67)(68). In some species, cholesterol content increased during sperm maturation ( 69 ) or remained unchanged ( 70 ). The loss of cholesterol is known to increase the fl uidity of sperm membranes, as demonstrated in human spermatozoa ( 71 ), which might be necessary for the fi nal steps of sperm maturation in the female reproductive tract. The importance of the stage-specifi c sterol content of sperm membranes is additionally demonstrated by the presence of vesicular formations produced and secreted by the prostate, known as prostatic secretory granules or prostasomes. Prostasomes have a different lipid composition from sperm plasma membranes ( 72 ). In particular, they contain large amounts of sterols and sphingomyelin. However, their physiological function has been thoroughly debated by many investigators. Prostasomes present within sperm are able to fuse with the plasma membrane ( Fig. 2 ), where they are thought to increase motility and prevent early acrosome reaction ( 73 ). The various mechanisms that regulate cholesterol content in sperm membranes during their epididymal transit indicate the functional relevance of cholesterol in sperm function.

DYNAMICS OF SPERM MEMBRANE CHOLESTEROL IN THE FEMALE REPRODUCTIVE TRACT
Ejaculated sperm is motile but remains unable to fertilize the oocyte. Sperm is activated during the process of capacitation, which occurs in the female reproductive tract ( 74 ). During capacitation, the surface of the sperm head is remodeled, priming it to bind to the zona pellucida (ZP) and to undergo the ZP-induced acrosomal reaction. An important fi rst step enabling capacitation is an organized destabilization of the plasma membrane, which also involves the redistribution of cholesterol, lipids, and

FUNCTIONAL ROLE OF STEROLS IN SPERM DEVELOPMENT
The development of highly specialized spermatozoa requires extensive remodeling of membrane structure, involving stage-specifi c changes in membrane permeability and fl uidity. During the maturation process, a sperm cell needs to maintain integrity and to acquire a high level of motility. In the female reproductive tract, before fusion with the oocyte, a drastic change in membrane structure enables the extrusion of acrosomal content and eventually fusion with the oocyte. Highly specialized membrane function is mostly acquired with precise regulation of sperm membrane lipid content at distinct stages of the maturation process. The functional importance of sterol composition during sperm maturation was clearly demonstrated by unique molecular events that resulted in stage-specifi c modifi cations of sperm membrane sterol content. In addition to cholesterol, high levels of different sterol intermediates are known to accumulate in the testis. Their exact function is unknown, but they are speculated to have important roles in the modifi cation of membrane properties. The biological basis for selecting cholesterol as the major sterol in the majority of animal cells is unclear; however, it is generally accepted that cholesterol is the most effective sterol at modulating membrane properties. Its structure allows the tight packing of cholesterol alongside the fatty acid tails of neighboring phospholipids. This can increase order within liquid membranes through the effects of its rigid ring system and the ability to fi ll interstitial spaces. A higher order of lipid packing results in a specialized type of membrane organization that provides a functional permeability barrier ( 99 ). Generally, the highest concentration of cholesterol is present within immature and immotile sperm, collected from the testis or caput epididymis. At this point, cholesterol is believed to function as a membrane stabilizer to provide protection against the premature release of proteolytic enzymes from the acrosome. In the majority of species, cholesterol content during sperm maturation decreased, whereas the content of other sterols, such as desmosterol, increased. Depletion of cholesterol is crucial for normal sperm function, as demonstrated in patients with idiopathic infertility that display a higher cholesterol to phospholipid ratio ( 83 ). High cholesterol levels in sperm are negatively correlated with sperm quality in humans ( 86 ). Moreover, cholesterol is believed to have an important specifi c role in sperm raft function. During capacitation, a preferential effl ux of cholesterol from the nonraft pool might serve as a stimulus to promote coalescence of separate microdomains into larger rafts over the anterior sperm head ( 100 ).
The second most common sterol in the testis is desmosterol. One basic function of desmosterol and cholesterol in the sperm membrane is the inhibition of sperm capacitation. A suffi cient level of both sterols within the sperm membrane is therefore constantly maintained by the presence of epididymosomes in the epididymis and prostasomes in ejaculated sperm. Both types of vesicles contain high levels of cholesterol and desmosterol and may of low levels or low activity of 24-dehydrocholesterol reductase, which converts desmosterol to cholesterol, or the presence of an inhibitor of this enzyme in the testis and epididymis. Microarray studies using male mouse germ cell fractions (http://mrg.genetics.washington. edu) show that Dhcr24 is indeed expressed at very low levels compared with other genes involved in earlier steps of cholesterol synthesis ( 33 ). A possible reason for the accumulation of desmosterol in the testis might be the specifi c progestin-mediated inhibition of 24-dehydrocholesterol reductase, as demonstrated by Lindenthal et al. ( 10 ). Similar to cholesterol, the proportion of desmosterol changed during the epididymal transit of sperm. The amount of desmosterol in sperm membranes increases as sperm travel from the caput to the cauda epididymis in boar ( 70 ), hamster ( 67 ), and rat ( 10 ), resulting in a higher desmosterol to cholesterol ratio. By contrast, desmosterol is present in ram spermatozoa entering the epididymis but is barely detected in mature sperm collected from the cauda epididymis ( 68 ). Based on the differential expression of cholesterogenic genes in the epididymis ( 91 ) and the presence of sterol-loaded epididymosomes, it is very unlikely that synthesis within sperm alone is responsible for changes in sterol composition during epididymal transit.
A unique sterol composition of spermatozoa was demonstrated in some mammalian species. The predominant sterol intermediate in hamster spermatozoa is cholesta-7,24-dien-3 ␤ -ol, a direct precursor of 7-dehydrocholesterol ( 92,93 ). The amount of cholesta-7,24-dien-3 ␤ -ol increased during the transit of spermatozoa through the epididymis, similar to desmosterol ( 67 ). The predominant sterol in the testis of guinea pig is lathosterol, with concentrations similar to that of desmosterol ( 12 ). Other sterol compounds found in the semen are sterol sulfates. The sterol sulfate fraction of human ejaculated spermatozoa contains 85% cholesterol sulfate ( 94 ), which represents about 6% of total cholesterol in spermatozoa ( 89 ). By contrast, the sterol sulfate fraction of hamster spermatozoa is composed almost exclusively of desmosterol sulfate ( 95 ). Both sterol sulfates were also present in monkey spermatozoa ( 96 ). Another modification of sterol compounds during sperm transit through epididymis is esterifi cation, which was shown to occur in ram ( 97 ). Extreme diversity of sterol content during sperm maturation in different species is further demonstrated by high levels of 7-and 8-dehydrocholesterol in the rat caput epididymis. This intermediate is undetectable in mature sperm collected from the cauda epididymis ( 98 ). An ongoing study examining the expression of cholesterogenic genes and an array of sterol intermediates in Crem -deficient mice reported that the profi le of sterol intermediates in the testis could not be explained by an established model of cholesterol biosynthesis (Ačimovič et al., unpublished data). The existence of new sterol elimination pathways was suggested based on mathematical model analyses. This study indicates that cholesterol biosynthesis in germ cells might involve another yet-to-be-defi ned regulatory mechanism. cholesterol transport suggest that the transport of cholesterol may play an essential role in spermatogenesis. Mice with complete ablation of type 2 apolipoprotein E receptor ( Lrp8 ) ( 107 ) and mice with partial ablation of apolipoprotein B ( Apob ) ( 108 ) exhibited severely compromised fertility. Based on systemic knockout models, we cannot conclude with certainty that infertility is solely due to perturbations in spermatogenesis or sperm maturation because the underlying cause could also be of systemic origin. However, infertility in the Lrp8 knockout was ascribed to defects in sperm maturation, mainly due to a high expression of Lrp8 in the epididymis and the specifi c nature of the sperm phenotype ( 107 ). A list of genes involved in different cholesterol-related biological processes that are also implicated in male fertility as demonstrated by mouse models is shown in Table 1 . In addition to knockout models, the role of cholesterol transport in sperm development is further demonstrated by the negative effects of dietary cholesterol overload on fertility in rats ( 109,110 ). The importance of de novo cholesterol synthesis during spermatogenesis is currently being studied by our group by examining conditional knockouts of Cyp51 in different cells of the testis. Characterizing testis and fertility phenotypes in these conditional knockout models is likely to provide insight on the importance of de novo cholesterol synthesis and/or accumulating intermediates in germ, Sertoli, and Leydig cells as well as in male reproductive function.
Based on various accumulation patterns of sterol intermediates in the testis of different mammalian species, one could speculate that the main function of sterols is stageand species-specifi c regulation of membrane function. Studies on cell-based systems suggest that sterol synthesis might be accompanied by the accumulation of several sterol intermediates that do not remain confi ned to the endoplasmic reticulum but rather are incorporated into the outer leafl et of the plasma membrane along with cholesterol ( 103 ). Lange and coworkers reported that the transport of newly synthesized zymosterol to the plasma membrane in fi broblasts was two times faster than that of newly synthesized cholesterol ( 111 ). The mobilization rate of different sterol precursors from the cell membrane differs not only from that of cholesterol but also from each other, with more polar intermediates being more avidly effl uxed ( 112 ). Based on these experiments, we conclude that sterol intermediates can be readily incorporated into the membrane. Different sterol intermediates differ in the number of methyl groups, number and position of double bonds, and conformation of the sterol ring system ( Fig. 1 ). Each sterol intermediate therefore possesses unique structural features and can affect membrane structure in a specifi c way. Different membrane properties might be especially benefi cial during sperm development. Additionally, species-specifi c differences in sterol composition during sperm development could be important for unique reproductive characteristics of individual species.
Transport and distribution of cholesterol and sterol intermediates play an important role in spermatogenesis, but these processes are not well understood. Recently prevent the acrosomal reaction by supplying endogenous sterols. This is an important aspect for the fertilizing capacity of spermatozoa in species where spermatozoa have to wait a long time to interact with the oocyte without losing membrane integrity. During this period, prostasomes might remain in contact with sperm ( 88 ). Therefore, the loss of cholesterol and desmosterol is the initial obligatory step for capacitation ( 48,101 ). Desmosterol is the predominant sterol in monkey and rabbit spermatozoa and was found to accumulate preferentially in the sperm tail. The difference in membrane desmosterol composition between sperm heads and tails might underlie the different functions of these two structures. Desmosterol has an additional double bond between carbon atoms 24 and 25 and might contribute to membrane fl uidity, which is necessary for the motility of fl agella ( 90 ). Another possible explanation for the preferential use of desmosterol that was previously not considered is the high desorption efficiency of desmosterol from the plasma membrane. Desmosterol was shown to be transferred between large unilamellar vesicles two to three times faster than cholesterol ( 102 ). This resulted in a higher effl ux of biosynthetic desmosterol from cells ( 103 ). It is plausible that a high desorption effi ciency of desmosterol, facilitating its eventual effl ux from the membrane, represents the main underlying reason for the higher accumulation of desmosterol in the testis and spermatozoa. The double bond at the C24 position in desmosterol signifi cantly weakens sterol ordering potential and the formation of lipid rafts. The weaker effect of desmosterol on bilayer properties is probably due to the different tilt of the steroid ring when compared with cholesterol ( Fig. 1 ). Because desmosterol cannot replace cholesterol within lipid rafts, it was suggested that the ability of cells to follow the desmosterol synthesis route over the cholesterol synthesis route might provide a physiological mechanism to modulate raft-dependent functions ( 104 ). Similar to cholesterol and desmosterol, cholesterol and desmosterol sulfate were shown to prevent capacitation and inhibit acrosomal enzyme acrosin ( 105,106 ). Moreover, both are present in low proportions. Sulfatase activity in the female tract suggests a key role of sterol sulfates in the process of fertilization ( 106 ).

CONCLUSIONS AND PERSPECTIVES
Based on current literature, the source of cholesterol needed for spermatogenesis has not been shown conclusively, mainly due to the complexity of the process and lack of a suitable in vitro systems. Cholesterol might originate from de novo synthesis in germ or Sertoli cells, or it might be transported from the interstitial compartment with lipoprotein particles, as discussed previously. This question can be addressed by carefully examining the phenotypes of cell conditional knockouts of genes involved in cholesterol synthesis and transport in different types of testicular cells. To our knowledge, no such model has been developed. A few studies that have examined mouse models with systemic inactivation of genes involved in mainly due to the inability to establish an in vitro spermatogenic cell culture system. To investigate the specifi c role of MAS during spermatogenesis in vivo, we will use conditional knockout models of the MAS-producing gene Cyp51 in various cell types in testis. A thorough characterization of such models is likely to reveal the function of MAS and other intermediates upstream of the CYP51 enzymatic step in male germ development. Despite the longterm availability of the conditional knockout technology, germ cell-specifi c knockout of genes involved in cholesterol synthesis has not yet been developed. Although some conditional knockout models have provided answers to important physiological questions, others have shown unexpected complexities or technical limitations that have confounded interpretation of the results in testis, such as the case of the conditional knockout of the androgen receptor. Moreover, improved and more sensitive sterol measurement methods are needed to analyze quantitatively levels of intermediates in minute samples from various cell types or testis cell fractions ( 116 ). New-generation RNA sequencing technology should also allow determination of the absolute amounts of various testis-specifi c RNA developed methods ( 113 ) for high-resolution lipid imaging in vivo or in vitro may be helpful if applied to studies in the testis. For instance, one of the fi rst studies for directly monitoring sterol movement in living cells has been achieved by fl uorescently labeling sterol derivatives ( 114 ). Additionally, fl uorescent and photoreactive sterol probes may reveal not only the location and movement of these lipids but also reveal their interactions with proteins as demonstrated in a recent study by Gimpl and Gehrig-Burger ( 115 ). Although these new approaches have not been extensively used in the testis, they hold great promise in functional studies of sterol traffi cking and distribution during spermatogenesis.
Among the several sterol intermediates shown to accumulate in the testis and spermatozoa of different species, MAS are considered unique because of their potential role in the regulation of meiosis. A thorough examination of the role of MAS in oocyte development has led to the conclusion that MAS possess a nonspecifi c receptor-independent positive effect on oocyte maturation. A similar effect was demonstrated with closely related sterols ( 6 ). The role of MAS in spermatogenesis has been very poorly studied, Targeted, spontaneous, or ENU-induced inactivation of these genes in mouse models results in reduced fertility or infertility in males. Data are adopted from Mouse Genome Informatics database. species from genes involved in cholesterol production and homeostasis. Such studies have not been possible by microarray analyses, which can only provide relative but not absolute amount comparisons. In addition, sensitive proteomic analyses are required to determine complete RNA-protein relationships of cholesterogenic genes and to discover possible protein modifi cations specifi c to testis. Diet perturbation experiments combined with conditional knockout models and sensitive high-throughput techniques should help to reveal the role of sterols in highly specifi c mechanisms of sperm development. In addition to mouse, important information might be acquired from mutant Drosophila strains, which are ideal for in vivo studies of membrane remodeling during spermatogenesis. A combination of these approaches is likely to result in a better understanding of sterol function in sperm formation. It may also reveal that some cases of unexplained male infertility are the result of cholesterol and sterol intermediate perturbations in the testis due to genetic or environmental factors.