|Home | About | Journals | Submit | Contact Us | Français|
The transcription factor ets variant gene 5 (ETV5; also known as ERM) is essential for self-renewal of spermatogonial stem cells (SSCs). Mice with targeted disruption of Etv5 (Etv5−/−) undergo the first wave of spermatogenesis, but all SSCs are lost during this time, causing a Sertoli cell-only phenotype. This study examined body and testis growth and the time course of SSC loss in Etv5−/− mice to understand how loss of ETV5 impacts testicular and somatic development. Body weights were reduced in postnatal Etv5−/− males, indicating a role of ETV5 in growth. Testis weights and histology in Etv5−/− and wild-type (WT) males were similar at Postnatal Day 4, but testis weights were reduced at d8 and subsequently, indicating that ETV5 impacts postnatal testis growth. SSC density (SSCs per μm2 of seminiferous tubule), estimated using an antibody against GFRA1, was similar in 4d WT and Etv5−/− mice. By 8 and 12d, GFRA1-positive cell density in Etv5−/− mice was decreased 17% and 32%, respectively, vs. WT. By 28d, GFRA1-positive cell density in Etv5−/− was reduced 95%, and GFRA1-positive cells were absent in 36d Etv5−/− males. In contrast to WT, 35- to 56-day-old Etv5−/− mice were infertile as assessed by natural breeding, artificial insemination, and in vitro fertilization, although motile sperm were present in epididymides of Etv5−/− mice during this time. In summary, initial testis development is normal in Etv5−/− mice despite decreased body weight, but SSC loss begins between 4 and 8d of age, indicating that ETV5 has effects beginning in the early neonatal period. Etv5−/− mice are infertile even when sperm is produced, indicating that ETV5 loss has other effects besides lack of SSC self-renewal that impair fertility.
Spermatogonial stem cells (SSCs) lie along the basement membrane of the testicular seminiferous tubules surrounded by Sertoli cells, which are the somatic cell component of the seminiferous epithelium. Maintenance of the SSC population and continued spermatogenesis normally involves a balance between SSC self-renewal through proliferation and loss of SSCs through differentiation into cells that have lost their “stemness” and ultimately give rise to functional spermatozoa.
The recent development of a variety of knockout mice in which SSCs are lost during development [1–8] and in vitro culture methodologies where the factors that regulate SSC development can be examined  provide useful model systems to address factors regulating SSC development. A number of factors that regulate SSC self-renewal and differentiation have been identified, including glial cell line-derived neurotrophic factor (GDNF), rearranged during transfection (Ret), GDNF family receptor α1 (GFRA1), B cell CLL/lymphoma 6 member B (BCL6), promyelocytic leukemia zinc finger protein (PLZF), TATA box binding proteinassociated factor (Taf4b), bone morphogenetic protein 4 (BMP4), and activin. Another molecule shown to be important in SSC development is ets variant gene 5 (ETV5; also known as ets-related molecule, or ERM), a member of the large ETS family of transcription factors. Mice with a targeted disruption of Etv5 (Etv5−/−) undergo a first wave of spermatogenesis during juvenile life , but show a progressive loss of spermatogenesis with age, culminating in a loss of all germ cells and a Sertoli cell-only phenotype by 10 wk of age.
Abnormalities in spermatogenesis were seen at 6 wk of age in Etv5−/− mice , when some tubules had normal spermatogenesis, some had loss of spermatogonia or spermatogonia and some later stages of spermatogenesis, and some consisted of only Sertoli cells. However, it was not clear when SSC loss begins or becomes total. The former question is of critical significance for establishing the mechanism of this phenomenon because identification of the critical temporal window when SSCs are lost is an essential first step toward establishing the mechanism of this phenomenon.
In addition to determining when SSCs are lost in Etv5−/− mice, it is also important to examine early testicular development in these animals. If neonatal testicular morphology is normal in Etv5−/− mice, this would suggest that Etv5 did not play a critical role in testicular organogenesis and early development, and Etv5 may only affect postnatal SSC self-renewal and/or differentiation. Conversely, if early testicular development and/or establishment of the initial SSC population are altered in Etv5−/− mice, this could reflect Etv5 effects on initial differentiation, migration, and/or proliferation of Sertoli cells and germ cells, as well as possibly other testicular cell types. Determining whether there are developmental changes in the seminiferous epithelium and testis that contribute to the eventual testicular phenotype of the Etv5−/− mice would provide valuable clues for ultimately establishing the overall role of Etv5 in testis.
The first wave of spermatogenesis during juvenile life in Etv5−/− mice  suggests that these animals may be transiently fertile, a phenomenon previously observed in Taf4b−/− mice , which, like the Etv5−/− mice, initially undergo spermatogenesis then gradually lose their germ cells. However, ETV5 is expressed at high levels in juvenile Sertoli cells , and Etv5 mRNA has also been reported in cultured murine SSCs [3, 7]. Therefore, lack of Etv5 may produce Sertoli and/or germ cell changes that preclude sperm fertility. To understand the gamut of Etv5 effects in testis, it is necessary to establish whether loss of Etv5 has effects on fertility that are distinct from the loss of SSCs that eventually renders these animals aspermic.
This study sought to determine whether initial testicular and SSC development was normal in Etv5−/− mice and to identify the time frame in which SSC loss occurs in these males. Our studies also sought to determine whether the Etv5−/− males were fertile during the transient period in early development when they produce sperm.
Etv5−/− mice on a 129Sv/Ev background were developed as described previously . Heterozygous 129Sv/Ev animals were bred to generate wild-type (WT), heterozygous, and Etv5−/− animals; only WT and Etv5−/− mice were used in these studies. Animals were genotyped neonatally using DNA from ear punches. The primers used to identify the WT mice were 5′-GACTCCTGGATGCTCCTTAGCAC-3′ and 5′-AGACAAGTCTTCTGGACTAACCCTG-3′. The primers used to identify the Etv5−/− mice were 5′-AGACAAGTCTTCTGGACTAACCCTG-3′ and 5′-GGGAAACTCGAGGGACCTAATAAC-3′.
Mice were housed at 25°C with a 12L:12D photoperiod and were given water and a standard rodent diet ad libitum. All experiments involving animals were approved by the Institutional Animal Care and Use Committee of the University of Illinois and conducted in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals.
Body and testis weights of WT and Etv5−/− males were measured at 4-day intervals from 4 to 24 days of age. Body and testis weights were also determined in older males (Days 36–90) used for fertilization experiments.
To determine the time course of SSC loss, testis tubule whole-mount immunohistochemistry for a marker of Asingle (As) and Apaired (Apr) spermatogonia, GFRA1, was performed on WT and Etv5−/− tubules . Some data suggest only As spermatogonia are the true stem cells , so using GFRA1 as a stem cell marker will give an overestimation of the total number of SSCs. Conversely, other data suggest As, Apr, and Aaligned (Aal) spermatogonia all make up the true stem cell population . In this case, defining stem cells as GFRA1-positive will lead to an underestimation of true stem cell number. For the purposes of this study, GFRA1-positive cells are regarded as the true SSC in order to determine relative SSC number in tubules. Tubules were collected from WT and Etv5−/− males at 4, 8, 12, 28, and 36 days of age (n = 4), then immersed in Dent fixative (1:4 dimethyl-sulfoxide-methanol) overnight at 4°C. A goat anti-mouse GFRA1 antibody (Santa Cruz Biotechnology, Santa Cruz, CA) that specifically recognizes GFRA1  was used at a 1:500 dilution for immunohistochemistry. The secondary antibody was a donkey anti-goat IgG conjugated to biotin (1:1000). Tubules were incubated with streptavidin-peroxidase, and the reaction product was visualized with the AEC kit from Zymed (San Francisco, CA), which produces a red precipitate following reaction with streptavidin-peroxidase. As a negative control, tubules were incubated in a similar fashion, but the primary antibody was omitted. Tubules were mounted on slides and then viewed under a light microscope. GFRA1-positive cells in which the whole nucleus was visible were counted by focusing on all planes of the tubule. Individual tubule diameter and length were measured using NIH software program ImageJ. Values for positive staining cells are reported as number of cells in 1000 μm2 of seminiferous tubule.
To assess the presence and motility of epididymal sperm, cauda epididymides from young adult (35 to 56 days old) WT and Etv5−/− mice were removed and triturated in warmed human tubal fluid (HTF) media (Fisher, Waltham, MA). Motility of sperm was assessed by diluting sperm 1:10 into HTF media and transferring 15 μl into each of two compartments on a glass cannula for computer-assisted sperm analysis (CASA) using the integrated visual optical system (IVOS) motility analyzer (Hamilton-Thorne Research, Beverly, MA). The operational settings of the IVOS were the standard mouse parameters as recommended by the manufacturer. For each sample, one slide, with 10–15 scans per slide, was analyzed.
Tests were conducted to determine whether Etv5−/− males were transiently fertile during development. WT and Etv5−/− males (n = 6) were housed with two adult WT females for 3 wk consecutively, starting at 35 days of age. Each week, males were placed with two different females. Vaginal plugs were checked each morning. To determine pregnancy status, females were killed 7 days after the mating pair was separated.
Behavioral response to pheromones present in estrus urine was assessed. Urine was collected from females in estrus, which was determined by examining vaginal swabs. Water and estrus urine were placed on opposite ends of a cage in equal volumes (250 μl). Males were then placed in the center of the cage and monitored . If males displayed no sign of interest the test was terminated after 5 min. Interest in urine was assessed as extensive sniffing and pawing at the bedding by the urine. Animals failing to exhibit these signs were classified as lacking interest in estrus urine .
The ability to produce an erection was tested in WT and Etv5−/− males as described . An erection was scored as strong if the penis became fully engorged with blood, weak if the penis became slightly engorged but a full erection was not obtained, and absent if no engorgement was noted and an erection was not obtained.
Sperm for artificial insemination and in vitro fertilization experiments was collected as described . Sperm were collected in HTF media. Females were hormonally primed for both artificial insemination and in vitro fertilization experiments . Artificial insemination experiments were performed as previously described  to test Etv5−/− and WT male fertility. Females were killed 1 wk following artificial insemination to determine pregnancy. In vitro fertilization was carried out using published protocols ; mouse oocytes were collected in HTF media, and 8–12 eggs were fertilized with sperm from each of three Etv5−/− and WT males. Oocytes were co-incubated with sperm at 37°C in a 95% air/5% CO2 environment. Twenty-four hours following in vitro fertilization, cleaved oocytes were scored as fertilized, and one-celled oocytes were recorded as unfertilized.
Testes from 44-day-old WT and Etv5−/− mice were collected and fixed in neutral buffered formalin (Sigma-Aldrich, St. Louis, MO) overnight and then transferred to 70% ethanol. Testes were embedded in paraffin and sectioned at 5 μm. Some sections were stained with hematoxylin and eosin and examined under an Olympus light microscope (Optical Analysis Corp., Nashua, NH). Other sections were used to assess apoptosis by TUNEL staining using an In Situ Cell Death Detection Kit according to the manufacturer's directions (Roche Diagnostics, Indianapolis, IN).
The General Linear Models procedure of SAS (Cary, NC) was used to calculate significance and standard error for body weights, testis weights, and whole-mount immunohistochemistry data. Differences were considered significant at P < 0.05.
Body weights in Etv5−/− mice were reduced compared to WT controls throughout life (Fig. 1). The differences between Etv5−/− and WT mouse body weights ranged from 12% to 31% and were significant at all ages except 4d, 12d, 24d, and 28d (Fig. 1).
Testes weights were also not different in 4-day-old WT and Etv5−/− mice (Fig. 2). By 8 days of age, testes weights in Etv5−/− mice were 27% smaller (P < 0.05) than WT controls. These differences persisted during subsequent development, and became more pronounced (41% to 75%) at later ages such as 36 and 56 days, when spermatogenesis is well underway in the WT mice but progressive losses of spermatogenesis occur in Etv5−/− mice . By 90 days of age, testis weights in Etv5−/− mice were reduced by 86% compared to WT, reflecting total loss of spermatogenesis by this age.
During neonatal life, the ratio of body weight to testis weight is comparable in WT and Etv5−/− mice (1.56 × 103 and 1.44 × 103, respectively, at 4 days of age). As the animal matures and begins to undergo spermatogenesis, the ratio of body weight to testis weight decreases in both WT and Etv5−/− mice. For example, at 24 days of age, the ratio of body weight to testis weight is 0.40 × 103 and 0.29 × 103 in WT and Etv5−/− animals, respectively. In WT animals this trend continues, and by 90 days the body weight:testis weight ratio is 0.25 × 103. However, in Etv5−/− animals, the ratio of body weight to testis weight begins to increase as these animals lose spermatogenesis. By 90 days of age, the ratio of body weight to testis weight in Etv5−/− animals has increased to 1.35 × 103 and is comparable to that seen at 4 days.
At 4 days of age there was no difference in GFRA1-positive SSC density between WT (0.22 ± 0.06 SSCs per 1000 μm2 of seminiferous tubule) and Etv5−/− (0.24 ± 0.04 SSCs per 1000 μm2) mice (Figs. 3 and and4).4). This, in conjunction with the normal testis weights in the Etv5−/− mice at this age, indicated that initial overall testis growth and development and spermatogonial stem cell development were not impaired in Etv5−/− mice. At 8 days, Etv5−/− males had a 17% decrease in SSC density (Figs. 3 and and4)4) compared to WT controls (WT and Etv5−/− = 1.04 ± 0.08 × 10−1 and 0.86 ± 0.03 × 10−1 SSCs per 1000 μm2, respectively; P < 0.05). By 12 days, spermatogonial stem cell density was decreased 32% in Etv5−/− versus WT mice (0.73 ± 0.09 × 10−1 and 0.50 ± 0.08 × 10−1 SSCs per 1000 μm2, respectively; P < 0.05; Figs. 3 and and4).4). The Etv5−/− testis at day 28 had 95% fewer SSCs per area of seminiferous tubule than WT (WT and Etv5−/− tubules contained 0.50 ± 0.10 × 10−2 and 0.02 ± 0.02 × 10−2 SSCs per 1000 μm2, respectively; Figs. 3 and and4).4). At this age, 86% of Etv5−/− tubules observed were devoid of SSCs, whereas 100% of WT tubules had SSCs. At later ages (36 and 44 days), SSCs were absent in Etv5−/− mice. Total seminiferous tubule length in a 10-day-old mouse is approximately 1 m . Therefore, based on the SSC density we reported for 8-day-old mice, this suggests that there are about 15 000 stem cells present in WT mice at this age. It is not clear when SSC number in mice plateaus, but SSC numbers are still increasing neonatally . Thus, our estimate of 15 000 SSCs at day 8 is consistent with previous reports of adult SSC populations of 35 000 .
The epididymis of both WT and Etv5−/− mice contained sperm at 44–56 days of age (n = 3 for both groups), although the amount of sperm seen in Etv5−/− mice was reduced compared to that in WT (Fig. 5). The concentration of sperm seen in the cauda epididymis, determined by CASA, of Etv5−/− males averaged 2.47 ± 0.67 million/ml, whereas the concentration of WT sperm averaged 14.90 ± 2.72 million/ml (P < 0.01).
Histological examination of Etv5−/− testes revealed that the decrease in sperm concentration in these juvenile mice is due in part to impaired spermiation and subsequent phagocytosis of the elongated spermatids (Fig. 6). Epididymal cross sections also showed sloughed round spermatids, indicating that in some cases these cells were not progressing to mature stages of spermatogenesis before leaving the testis (Fig. 5). Due to the decrease in epididymal sperm, we examined the adult testis for evidence of increased germ cell apoptosis among the maturing spermatids that might account for the lower sperm number in the epididymis compared to the testis. TUNEL staining of seminiferous tubules showed an increase in apoptosis of round spermatids in 44d Etv5−/− compared to 44d WT mice (Fig. 7). Sperm obtained from the cauda epididymis of Etv5−/− mice were also less motile than WT. Etv5−/− sperm motility averaged 7.0% ± 3.2% compared to 41.0% ± 4.9% in WT males (n = 3 for both groups). The decrease in motility seen in Etv5−/− may be due to a pooling of live and dead sperm in the cauda epididymis.
All WT males (n = 6) bred to adult WT females starting at 35 days of age produced viable offspring and successfully impregnated an average of 3.7 ± 0.2 out of the 6 test females used for natural breeding. In contrast, breeding Etv5−/− males (n = 6) to adult WT females starting at 35 days of age never resulted in pregnancies. Etv5−/− males failed to produce copulation plugs, suggesting that these males might not be capable of or interested in mating. There were no morphological differences between secondary sex organs of WT and Etv5−/− mice. Therefore, the lack of a copulatory plug in the Etv5−/− mice does not result from obvious abnormalities in secondary sex organs. These animals appear to be capable of mating due to the presence of strong erections in all males of both the WT and Etv5−/− groups (n = 6 for each group). These results indicate that Etv5−/− males are capable of showing normal penile erection; therefore, their inability to impregnate WT females is not due to erectile problems. However, Etv5−/− males appeared to show reduced interest in breeding. Therefore, we tested their response by placing the males in cages where one drop of urine from estrous females had been placed on one wall and the bedding immediately under where the drop was added. Stimulation of Etv5−/− males with estrus urine occurred in only 1 of 5 males tested. However, 100% of WT males (5/5) showed a response following placement in a cage with estrous urine. These results indicate that Etv5−/− males may have impaired ability to sense and/or respond to female pheromones, and this may contribute to the lack of mating, as judged by absence of copulation plugs. To further confirm this phenomenon, nasal septum isolated from both WT and Etv5−/− males was analyzed by quantitative real-time PCR. Nasal septum was isolated because it contains vomeronasal receptors, which are responsible for sensing female pheromones. Male Etv5−/− mice had a 60% decrease in the mRNA level of the vomeronasal receptor (V1RC3; n = 4 for both groups, P < 0.001).
All adult WT females (n = 6) artificially inseminated with sperm from WT males (35 to 56 days old; n = 2) were pregnant when killed, 1 wk following insemination (Table 1). In contrast, no pregnancies were obtained in the six WT females using sperm from Etv5−/− males (n = 2).
WT sperm incubated with 10–12 oocytes collected from WT females produced on average 8 ± 0.6 two-cell embryos by 24 h following in vitro fertilization (n = 3). In contrast, when sperm from Etv5−/− males (n = 3) was incubated with oocytes derived from WT females 24 h following the in vitro fertilization, only one-celled oocytes were observed, indicating the Etv5−/− sperm did not fertilize WT oocytes (Table 1).
ETV5 is a transcription factor that is necessary for self-renewal of SSCs, as Etv5−/− mice lose their SSCs and become aspermic in the juvenile period even though they undergo a first wave of spermatogenesis. The results of the present study reveal the kinetics of GFRA1-positive SSC loss in Etv5−/− mice and shed light on the role of ETV5 in both testicular and overall body development, and these results will be useful for understanding the mechanism of overall ETV5 effects in seminiferous epithelium.
The reduced body weights in Etv5−/− mice are the first indication that ETV5 has significant effects on overall growth. These findings are consistent with previous reports that Etv5 mRNA is ubiquitously expressed in the developing and adult mouse and human, and has been detected in a variety of tissues, including lung, thymus, kidney, large intestine, mammary gland, lymphocytes, heart, salivary gland, and skeletal muscle [21–24]. The widespread expression of this transcription factor during both development and adulthood indicates that growth deficits seen in Etv5−/− mice likely result from effects on many organs, although the viability of Etv5−/− mice argues that effects on other organs may not be as pronounced as the testicular effects, even though Etv5 mRNA is more abundant in tissues such as brain, colon, and lung than in testis [21, 22].
The normal testicular weight and GFRA1-positive SSC density in 4-day-old Etv5−/− compared to WT mice, despite a trend toward decreased body weight in Etv5−/− mice, indicates that loss of ETV5 does not cause substantial alterations in initial testicular development. Thus, testicular organogenesis, primordial germ cell ontogeny and migration, gonocyte proliferation and differentiation into spermatogonia, and other key processes in early testicular development appear not to be significantly dependent on ETV5. In contrast, at Day 8 and subsequently, testis weight is decreased in Etv5−/− compared to WT mice due to loss of SSCs and subsequent stages of germ cell maturation, as discussed below, although the testis:body weight ratio is comparable in WT vs. Etv5−/− mice during early postnatal life.
Initial results showing that absence of ETV5 resulted in a loss of spermatogenesis clearly indicated that it was essential for normal self-renewal and maintenance of SSCs during juvenile life. However, it was unclear whether initial germ cell development was qualitatively and quantitatively normal and then SSCs were lost postnatally, or whether the spermatogenic deficits appearing during juvenile life reflected impairments in some early stage(s) of germ cell development that decreased SSC number, in combination with a lack of SSC maintenance after spermatogenesis begins. The comparable initial testicular development in Etv5−/− and WT mice clearly indicates that the major effects of ETV5 deficiency occur postnatally.
GDNF is a secreted Sertoli cell protein that is a distant member of the transforming growth factor β family. In vitro [7, 10, 25, 26] and in vivo [1, 2] evidence has indicated that GDNF promotes SSC self-renewal. GDNF signals through GFRA1, an extracellular protein bound to the plasma membrane that contains the ligand-binding region that interacts with GDNF . Since GFRA1 lacks transmembrane or intracellular regions, transmission of signaling into the cell requires a tyrosine kinase receptor, RET. In the seminiferous epithelium, GFRA1 is highly expressed in As spermatogonia and Apr spermatogonia. The latter arise from the division of As spermatogonia and represent the first step of spermatogonial differentiation. The As spermatogonia have previously been considered to be the true SSC, but recent elegant work by Nakagawa et al.  indicated that this may be an oversimplification and the true spematogonial stem cell still remains to be established. However, for the purposes of this study, cells that can be specifically identified using anti-GFRA1 antibodies  will be referred to as SSCs because this is presently the most accurate marker of stem cells available in the seminiferous epithelium, and later stages of germ cell development and Sertoli cells have low or undetectable GFRA1 expression. GFRA1 immunostaining thus provides a specific and sensitive method for quantitating SSCs in whole-mount preparations of seminiferous tubules .
Our results indicate that GFRA1-positive SSC density in WT seminiferous tubule decreases dramatically between Days 4 and 28 postnatal. This finding is consistent with previous results showing high neonatal SSC density, but decreases thereof during subsequent development, resulting in SSCs comprising only 0.03% of the cells in the adult seminiferous tubule [19, 28]. The postnatal decreases in SSC density result from the onset of spermatogenesis and consequent dilution of the SSCs by more advanced stages of spermatogenesis. The seminiferous epithelium at Day 4 postnatal consists of only Sertoli cells and spermatogonia that have just differentiated from gonocytes and will form the SSCs. As subsequent stages of spermatogenesis appear with advancing age, the large increase in the numbers of differentiating germ cells produces a striking decrease in SSC density, despite increases in total number of SSCs between the neonatal and pubertal periods .
Comparison of GFRA1-positive SSC density in Etv5−/− and WT mice indicates that decreased GFRA1-positive SSC density is demonstrable as early as Day 8, consistent with our data indicating that a decrease in testicular weight is first observed in Etv5−/− mice at this age. Thus, initial SSC loss begins in Etv5−/− mice between Days 4 and 8, and this is the critical period for subsequent investigations to determine the mechanism by which lack of ETV5 impairs SSC self-renewal and maintenance.
Previous results have indicated high expression of ETV5 in Sertoli cells during juvenile life , when progressive loss of spermatogenesis occurs. However, the current data indicate that the initial loss of SSCs occurs much earlier than originally anticipated. Therefore, it is critical that future studies examine the temporal and cell-specific pattern of ETV5 production during the perinatal period.
SSC loss occurs in a progressive manner, and these cells do not totally disappear until 35 days of age. Thus, SSC loss occurs over a 4-wk period. The progressive loss of SSCs could also involve increased SSC differentiation during the first wave of spermatogenesis.
The production of morphologically normal sperm in juvenile Etv5−/− mice  suggests that even though adult Etv5−/− males are aspermic, these animals could be transiently fertile for some period during development. This possibility is consistent with our present observations that 35- to 56-day-old Etv5−/− males have sperm in the testis and motile epididymal sperm that also appear morphologically normal, despite a decreased sperm concentration and motility in the cauda epididymis. However, juvenile Etv5−/− males, in contrast to age-matched WTs, were unable to impregnate or produce copulation plugs in WT females even during extensive periods of cohabitation. These results suggested that Etv5−/− males may have reproductive tract abnormalities and/or neurobehavioral changes as a consequence of ETV5 deficiency that interfere with normal mating. The fact that the Etv5−/− males showed reduced response to estrous urine suggests that infertility in the breeding trials involved an impaired response to female pheromones. This is validated by the data showing a 60% decrease in V1RC3.
Previous studies with mice that have a targeted deletion of PEA3, which is closely related structurally to ETV5 and a member of the same subfamily of ETS transcription factors, indicated that these males could not achieve penile erection . ETV5 and PEA3 have overlapping cellular distributions in several organs  and could potentially have overlapping functions. We therefore examined penile erection in Etv5−/− males and determined that it was normal, indicating that this was not the cause of the infertility.
Artificial insemination of WT females with sperm from Etv5−/− males failed to produce offspring, and even in vitro fertilization of WT oocytes with Etv5−/− sperm did not result in fertilized eggs. Therefore, even when potential neurobehavioral and other changes that could prevent fertilization are circumvented by artificial insemination or in vitro fertilization, there appear to be abnormalities in Etv5−/− sperm that preclude fertility. These abnormalities may involve the impaired sperm motility described here, as well as other potential effects resulting from lack of ETV5 in the seminiferous epithelium during sperm development.
An evaluation of stage-specific changes in Etv5−/− testes revealed two potential mechanisms to account for the loss of epididymal sperm during the first wave of spermatogenesis: failure of spermiation and increased apoptosis of round spermatids. Spermiation is the release of mature sperm into the lumen in Stage VIII , and failure of this process results in mature sperm being retained and phagocytosed by Sertoli cells in subsequent stages . Etv5−/− testes showed impaired spermiation, with mature sperm heads located near the basement membrane in the absence of one or more layers of underlying germ cells. TUNEL staining also revealed an increase in the apoptosis of round spermatids, but not other germ cells. Decreased levels of testosterone could result in increased germ cell apoptosis. However, testosterone levels were unchanged in Etv5−/− mice . Thus, loss of epididymal sperm in the Etv5−/− males was due to effects on late-stage germ cells of testis. The epididymis expresses high levels of ETV5 and other PEA3 family members , and the large decrease in epididymal sperm number and impaired motility in Etv5−/− males could also involve changes in epididymal epithelium.
In summary, the present results demonstrate that SSC loss in Etv5−/− seminiferous epithelium begins neonatally, long before deficits in spermatogenesis are observed, and these results indicate that mechanistic experiments should focus on this neonatal time frame to explain ETV5 effects on SSCs and to determine when and where ETV5 is produced in the seminiferous epithelium at this age. Despite the critical role of ETV5 in postnatal SSC maintenance, initial testicular development is normal in Etv5−/− males despite decreased body weight, showing that the effects of ETV5 in testis may be confined to the postnatal period. Lastly, sperm from Etv5−/− males are not capable of fertilization even with assisted reproduction techniques, indicating that lack of ETV5 causes effects other than lack of stem cell maintenance that impair fertility.
1This work was supported by the Billie A. Field Endowment, University of Illinois (to P.S.C.). The work at the University of Illinois was conducted in a facility constructed with support from Research Facilities Improvement Program Grant C06 RR16515 from the National Center for Research Resources, National Institutes of Health.