|Home | About | Journals | Submit | Contact Us | Français|
Approximately half of all infertility cases can be attributed to male reproductive dysfunction for which low sperm count is a major contributing factor. The current study identified receptor-mediated lysophosphatidic acid (LPA) signaling as a new molecular component influencing male fertility. LPA is a small signaling phospholipid, the effects of which are mediated through at least five G protein-coupled receptors, named LPA 1–5. LPA1/2/3, but not LPA4/5, show high expression in mouse testis. Mice deficient in LPA1/2/3 showed a testosterone-independent reduction of mating activity and sperm production, with an increased prevalence of azoospermia in aging animals. A significant increase of germ cell apoptosis also was observed in testes. Germ cell apoptosis led to a reduction in germ cell proliferation. These data demonstrate a novel in vivo function for LPA signaling as a germ cell survival factor during spermatogenesis.
The small lipid signaling molecule lysophosphatidic acid (LPA; 1-acyl-2-sn-glycerol-3-phosphate) can be found at micromolar concentrations in serum and can be produced by many cell types, such as activated platelets, erythrocytes, leukocytes, postmitotic neurons, adipocytes, and ovarian cancer cells [1–7]. LPA is produced by the enzymes ectonucleotide lysophosphatase/phosphodiesterase 2 (ENPP2; lysophospholipase D, autotaxin/lysoPLD) and/or phospholipases A1 and A2 (PLA1 and PLA2, respectively) [6, 8]. Extracellular signaling by LPA is mediated predominantly by at least five cognate G protein-coupled receptors, known as LPA 1–5 (LPA1/EDG2/vzg-1, LPA2/EDG4, LPA3/EDG7, LPA4/GPR23, and LPA5/GPR92) [1, 9, 10], to produce a range of cellular effects, such as proliferation, survival, differentiation, and morphological changes.
Signaling by LPA has important roles in many organ systems, including the female reproductive system [1, 8, 11–16], which uses LPA3 signaling . Previous studies have reported LPA biosynthetic enzymes, including PLA1, PLA2, and ENPP2 [6, 18–22], as well as LPA receptor gene expression in the testis . Single receptor-null mutants for Lpar1, Lpar2, or Lpar3 have not shown obvious male reproductive defects [17, 24, 25]. On closer inspection, however, small reductions in sperm production have been documented, suggesting that deletion of multiple LPA receptors might unmask compensatory mechanisms. To address this possibility, we generated a mouse mutant deficient for three LPA receptors, LPA1/2/3.
Lpar1(−/−) (Lpar1tm1Jch), Lpar2 (Lpar2tm1Jch)(−/−), Lpar1(−/−)Lpar2(−/−) (Lpar1tm1JchLpar2tm1Jch), and Lpar3(−/−) (Lpar3tm1Jch) knockout mice were generated as described previously [17, 24, 25]. Lpar1(−/−)Lpar3(−/−) (Lpar1tm1JchLpar3tm1Jch), Lpar2(−/−)Lpar3(−/−) (Lpar2tm1JchLpar3tm1Jch), and Lpar1/2/3 (Lpar1tm1JchLpar2tm1JchLpar3tm1Jch) triple-knockout (TKO) mice were generated by crossing single- and/or double-knockout mice. Each knockout mouse was subjected to genotyping for Lpar1, Lpar2, and Lpar3 alleles as described previously [17, 24, 25]. All analyses reported here regarding knockout and wild-type (WT) mice were from animals with a mixed 129/SvJ and C57BL/6 background unless otherwise specified. All the animals were generated and maintained in the laboratory of one of the authors (J.C.) at the University of California, San Diego, and later at The Scripps Research Institute. All animal research conducted for the present study was approved by the Animal Subjects Committee of The Scripps Research Institute and conformed to National Institutes of Health guidelines and public law.
Three WT males (age, 4 wk; C57BL/6) were anesthetized by isoflurane inhalation. Testes were immediately removed and snap-frozen in liquid nitrogen. Total RNA was isolated using Trizol (Gibco BRL) following the manufacturer's instructions except that each sample was extracted twice with chloroform. The RNA samples were analyzed by gel electrophoresis to confirm their integrity and the absence of chromosomal DNA contamination. Complementary DNA was transcribed from 1 μg of total RNA using Superscript II reverse transcriptase (Invitrogen) with random primers. Real-time PCR reactions were performed using SYBR Green intercalating dye (Sigma) on a Rotor-Gene 3000 (Corbett Research). The relative transcript number of each gene was quantified and normalized to actin, beta, cytoplasmic (Actb). Each primer pair crosses intron(s) in genomic DNA. The following primer pairs, with the expected product size indicated, were used for each gene: LPA1e3F1, 5′-TCT TCT GGG CCA TTT TCA AC-3′, and LPA1e4R1, 5′-TGC CTG AAG GTG GCG CTC AT-3′ (Lpar1, 350 bp); LPA2e2F1, 5′-ACC ACA CTC AGC CTA GTC AAG AC-3′, and LPA2e3R1, 5′-CTG AGT AAC GGG CAG ACT TG-3′ (Lpar2, 277 bp); LPA3e2F2, 5′-ACA CCA GTG GCT CCA TCA G-3′, and LPA3e3R2, 5′-GTT CAT GAC GGA GTT GAG CAG-3′ (Lpar3, 201 bp); LPA4e1F1, 5′-AGG CAT GAG CAC ATT CTC TC-3′, and LPA4e2R1, 5′-CAA CCT GGG TCT GAG ACT TG-3′ (Lpar4, 292 bp); LPA5e1F1, 5′-AGG AAG AGC AAC CGA TCA TCA CAG-3′, and LPA5e2R1, 5′-ACC ACC ATA TGC AAA CGA TGT G-3′ (Lpar5, 335 bp); Actbe5F1, 5′-TGG AAT CCT GTG GCA TCC ATG AAA C-3′, and Actbe6R1, 5′-TAA AAC GCA GCT CAG TAA CAG TCC G-3′ (actin, beta, cytoplasmic, 349 bp).
Testes were removed from anesthetized WT males at the ages of 15 days, 4 wk, and 4 mo and immediately frozen in Tissue-Tek optimal cutting temperature (OCT) compound with dry ice (Miles, Elkhart, IN). The frozen testes were kept at −80°C. Cross-sections (thickness, 20 μm) were cut and processed as described previously [26, 27]. The sections were hybridized with digoxigenin-labeled sense or antisense riboprobes, which were transcribed from linearized plasmids containing coding regions of mouse Lpar1/2/3 using T3 or T7 RNA polymerases. The hybridization was visualized by an alkaline phosphatase-conjugated anti-digoxigenin antibody (Roche).
Mice at 0, 15, and 28 days as well as 6 mo of age were anesthetized by isoflurane inhalation. Testes were immediately removed and fixed in Bouin solution (Sigma) overnight. Each testis was cut perpendicular to the long axis of the seminiferous tubules. Paraffin sections (thickness, 5 μm) were cut, processed, and stained with hematoxylin and eosin.
Wild-type or Lpar1/2/3 TKO males (age, 10–12 wk) and C57BL/6 virgin females (age, 8 wk) were used in the natural mating study. Three females were placed with each male and were checked for vaginal plugs every morning. Once a vaginal plug was found, the female was removed and placed in a separate cage. All nonplugged females were separated from males after 1 mo of cohabitation and were observed for at least 3 wk. In addition, Lpar1/2/3 TKO females (age, 2–6 mo) were mated with WT males (age, 2–4 mo) to determine litter sizes. Litter size was calculated as the total number of pups found at Postnatal Day 0.
Males were housed alone for 1 wk before they were killed. Both cauda epididymes were removed, weighed, and finely diced in 1 ml of 37°C Ringer buffer with 0.05% BSA in a 1.5-ml microcentrifuge tube, then shaken at 37°C for 30 min. Each sample was counted three times using a hemocytometer and quantified. No difference in the weight of cauda epididymes was observed between WT and Lpar1/2/3 TKO mice.
Serum testosterone levels were measured by standard radioimmunoassay according to the manufacturer's instructions (Diagnostic Systems Laboratory, Inc.).
Testes from 2-mo-old males (n = 3–4 mice/group) were removed and placed in a Petri dish containing serum-free Dulbecco Modified Eagle Medium (DMEM)/F12 medium (Gibco). The tunica albuginea was quickly removed, and the seminiferous tubules were immediately transferred to a 1.5-ml centrifuge tube in which they were minced. The cell suspension was filtered through nylon mesh (pore size, 100 μm; Beckman) into a 50-ml Falcon tube and centrifuged (900 × g, 5 min, room temperature). The cells were washed again with serum-free DMEM/F12 medium and cultured under serum-free conditions overnight at 32°C. The medium was then replaced with fresh medium containing variable concentrations of LPA (Avanti Polar Lipids) or sphingosine 1-phosphate (Avanti Polar Lipids) dissolved in 0.1% fatty acid-free BSA. Cells were incubated at 32°C for 10–60 min before they were harvested for protein extraction. Protein preparations and kinase assays were done as described previously [25, 28]. Primary antibodies to MAPK3/1 (extracellular regulated kinases 2/1, also known as ERK2/1) and phospho-MAPK3/1 (1:1,000; Cell Signal) were used in the present study.
Wild-type, Lpar3(+/−) (as a control), and Lpar1/2/3 TKO mice at 15 days, 3 mo, and 8 mo of age were injected intraperitoneally with 5-bromo-3-deoxyuridine (BrdU; 20 μl/g body weight, 10 mM in 1× PBS) 2 h before death. Testes were immediately removed and freshly frozen in M-1 embedding compound (Thermo Shandon). Cross-sections (thickness, 20 μm) were processed for BrdU immunolabeling and in situ end-labeling plus (ISEL+), performed as described previously [29–31], using a Cy5-conjugated anti-BrdU (red) primary antibody. Semiquantification of germ cell proliferation was performed in seminiferous tubules arbitrarily divided into the following five groups based on the percentage of BrdU-labeled germ cells: 0%, no labeling; less than 5%, sparsely labeled; 5–30%, low level of labeling; 30–70%, medium level of labeling; and greater than 70%, high level of labeling. All the tubules in the whole-testis section were given a score and tallied.
Data are expressed as the mean ± SD or SEM as indicated. Statistical analyses were done using the Student t-test, two-sample rank testing, or chi-square test. The significance level was set at P < 0.05.
It has been demonstrated previously by Northern blot analysis that the transcripts of Lpar1, Lpar2, and Lpar3, but not those of Lpar4 and Lpar5, are highly detectable in the testis [9, 23, 32]. We quantified mRNA expression levels of the five LPA receptors at 4 wk of age, which demonstrated Lpar1, Lpar2, and Lpar3 levels of 3,200- to more than 10000-fold higher than those of Lpar4 or Lpar5 in the testis (Fig. 1a); therefore, we focused our analyses on LPA1, LPA2, and LPA3. In situ hybridization studies (Fig. 1b) revealed gene expression of these three receptors in seminiferous tubules at all ages examined. At 15 days of age, Lpar1 mRNA was detected in the periphery of all seminiferous tubules, overlapping with a similar pattern for Lpar2 and Lpar3. At later ages (4 wk and 4 mo), Lpar1 mRNA was detected in the inner tubules, tracking with waves of spermatogenesis, whereas Lpar2 and Lpar3 mRNA continued to be detected in early stage germ cells on the periphery of all seminiferous tubules (Fig. 1b). LPA receptor gene expression in male germ cells suggested that LPA signaling might affect spermatogenesis.
Although previous studies with Lpar1, Lpar2, or Lpar3 single-knockout males did not reveal functional defects in male reproduction [17, 24, 25], we documented mild germ cell degeneration in the testes at 6 mo of age and a slight, but significant, reduction of sperm count from these single knockout males. Lpar1/2, Lpar1/3, and Lpar2/3 double-knockout males had a further reduction of sperm count (Fig. 2a and data not shown). These results indicated that all three LPA receptors are involved in sperm production. Therefore, we focused on the Lpar1/2/3 TKO testes.
Lpar1/2/3 TKO mice have a number of phenotypes consistent with previously reported single-receptor deletants, such as reduced body weight, hematomas, and perinatal lethality, as seen in Lpar1 mutants, as well as delayed implantation, prolonged pregnancy, and reduced litter size from Lpar1/2/3 TKO females mated with WT males, as seen in Lpar3 mutants (Fig. 2, b–d, and data not shown). Although Lpar1/2/3 TKO females had phenotypes similar to those of Lpar3 mutants in reproduction, more Lpar1/2/3 TKO mice had hematomas compared with the LPA1 mutants: 41.2% (28 of 68) Lpar1/2/3 TKO embryos (Embryonic Days 11.5–18.5) had hematomas, compared with 25.2% (31 of 123) in Lpar1/2 double-mutant embryos (Embryonic Days 11.5–18.5), 26.5% (9 of 34) in Lpar1/2 double-mutant neonatal pups, and 2.5% (4 of 160) in Lpar1 mutant neonatal pups . Interestingly, Lpar2/3 double-mutant embryos (Embryonic Days 11.5–18.5) did not show hematomas (0 of 49). These results suggest that a receptor subtype compensation occurs via different LPA receptor combinations on hematomas, with a predominant predisposition produced by LPA1 loss. In addition, Lpar1/2/3 TKO mice showed more dramatic embryonic and postnatal lethality than did any of the single-null mice [17, 24, 25]. When Lpar1/2/3 TKO males and females were crossed, they produced small litters (~2.8 pups/litter vs. 8 pups/litter in WT animals). Approximately 36% of the Lpar1/2/3 TKO pups were born dead, compared with 2% in WT animals. In addition, 48% postnatal lethality in the Lpar1/2/3 TKO pups, compared with 1% in WT animals, was observed. On average, only one Lpar1/2/3 TKO pup per litter survived beyond 15 days of age (Fig. 2d). The TKO survivors gave us the opportunity to study the in vivo function of LPA signaling in male reproduction, with a focus on spermatogenesis.
Lpar1/2/3 TKO young males (age 10–12 wk) were mated with 8-wk-old virgin WT females. These young TKO males only mated with two thirds of the females as indicated by plugs, whereas WT controls mated with 100% of females within the 1-mo mating period. These young TKO males also had a plugging latency nearly twice as long as that of WT controls. Plugging latency was the average time when the plug was found after cohabitation (Fig. 3a, b). As a result, these young TKO males had one third of the mating activity of WT controls. The females that developed plugs by mating with WT or TKO males, however, had similar pregnancy rates (Fig. 3c). In addition, no significant reduction of litter sizes was observed (Fig. 2c), indicating that the sperm from these young TKO males were functional and that the sperm numbers were not low enough to cause reduced litter size.
Lpar1/2/3 TKO testes showed age-related degenerative changes in the seminiferous tubules. At birth, the results of histological analyses of Lpar1/2/3 TKO testes appeared to be grossly normal, although interstitial regions between tubules were discernibly larger (Fig. 4a). At P15, abnormal tubular lumena and fewer spermatogonia and spermatocytes were observed in all Lpar1/2/3 TKO testes examined (Fig. 4b). By 4 wk of age, all Lpar1/2/3 TKO testes showed further reduction in cell numbers, along with increased vacuoles in seminiferous tubules (Fig. 4c). This phenotype became even more severe by 6 mo of age as the progressive accumulation of vacuoles destroyed the tubular architecture (Fig. 4d). Despite these changes, the Lpar1/2/3 TKO adult testes did not show reduced weight. Instead, a slight but significantly higher weight increase compared with WT controls was found (Fig. 5a). Age-related degenerative changes in the testes of Lpar1/2/3 TKO mice were paralleled by marked reductions in sperm count and increased rates of azoospermia. Sperm count from Lpar1/2/3 TKO males was approximately 50% of controls at 2 mo and reduced to 18% of controls at 8 mo (Fig. 5b). This progressive reduction in sperm count occurred concomitantly with a progressive increase in azoospermic animals that reached 65% by 8 mo (Fig. 5c), the oldest age examined. LPA1/2/3-deficient sperm showed normal motility (data not shown), consistent with the results of mating studies.
Spermatogenesis is controlled by both endocrine hormones and locally produced autocrine and paracrine factors [33, 34]. Testosterone is the major hormone regulating spermatogenesis; therefore, we assessed serum testosterone levels in Lpar1/2/3 TKO males. No significant differences were observed between 3 and 6 mo (median, 0.43 ng/ml; range, 0.15–13.69 ng/ml; n = 19) compared with age-matched WT controls (median, 0.39 ng/ml; range, 0.22–11.76 ng/ml; n = 24). These data indicate that spermatogenic defects in Lpar1/2/3 TKO males likely are the result of local LPA receptor-mediated signaling influences in the seminiferous tubules rather than systemic changes in endocrine function.
It has been shown previously that LPA can promote cell survival and proliferation through receptor activation of the heterotrimeric G protein, Gi/o [1, 35–37]. Figure 6a showed the time course and dose–response of LPA-induced MAPK3/1 phosphorylation in the control testicular primary cells. The LPA-induced, Gi/o-mediated phosphorylation of MAPK3/1 was completely abolished in testicular primary cells from Lpar1/2/3 TKO mice, whereas MAPK3/1 phosphorylation induced by another lysophospholipid, sphingosine 1-phosphate, was not affected (Fig. 6b).
Loss of Gi/o-mediated LPA signaling was accompanied by increased germ cell apoptosis in the Lpar1/2/3 TKO testes. Apoptotic germ cells detected by ISEL+ were present throughout the seminiferous tubules of both WT and Lpar1/2/3 TKO testes (Fig. 7a) . The percentages of ISEL+-labeled germ cells, however, were significantly increased in Lpar1/2/3 TKO testes at all ages examined (15 days, 3 mo, and 8 mo) (Fig. 7b), indicating increased germ cell apoptosis in the Lpar1/2/3 TKO testes.
The same testes also were examined for germ cell proliferation as assessed by BrdU incorporation and immunolabeling. The BrdU-labeled germ cells were detected throughout the seminiferous tubules at 15 days of age and were restricted to the proliferating spermatogonia (in the periphery of the seminiferous tubules) in adult testes (Fig. 8a). Semiquantitative analyses demonstrated a reduced percentage of tubules with BrdU labeling in the 3- and 8-mo Lpar1/2/3 TKO testes, indicating that germ cell proliferation was significantly reduced in adult Lpar1/2/3 TKO testes. No reduction of germ cell proliferation, however, was observed in 15-day-old Lpar1/2/3 TKO testes (Fig. 8b). These data suggest that the reduction of germ cell proliferation in Lpar1/2/3 TKO seminiferous tubules is a secondary effect of increased germ cell apoptosis.
The three closely related LPA receptors, Lpar1, Lpar2, and Lpar3, have similar expression patterns in the testes before puberty. The expression of Lpar1, however, changes between puberty and adulthood, resembling the expression pattern of the growth factor bone morphogenetic proteins (BMPs) Bmp7 and Bmp8, the expression patterns of which are stage specific . Lpar2 and Lpar3 are constitutively expressed in the basal regions of seminiferous tubules, primarily in immature germ cells, such as spermatogonia and spermatocytes, and are undetectable in elongating spermatids or mature spermatozoa (Fig. 1b). This expression pattern is similar to that of Bcl2l2 (Bcl-w) . Interestingly, disruption of BMP7 and BMP8 or BCL2L2 also results in impaired spermatogenesis [38, 39]. The stage-specific expression of Lpar1 in germ cells and the apparent constitutive expression of the Lpar2 and Lpar3 in spermatogonial and early spermatocytes suggest that LPA may play multiple and convergent roles through different receptors in promoting germ cell development.
The potential roles of LPA in spermatogenesis are manifested by the histological changes in the testes of LPA receptor(s) KO mice (Fig. 4 and data not shown). The tubules are not affected at the same time and rate; however, this heterogeneity in spermatogenic disruption also is apparent in other knockout models [39–41]. Despite this heterogeneity, vacuolar degeneration was seen in all the TKO testes. This degeneration is progressive, and by 8 mo of age, the degeneration results in azoospermia in most males. Defects in spermatogenesis appear to be the result of a direct effect on germ cell development and are distinct from indirect effects, as seen in Esr1 (ERα) knockout mice, in which male infertility was caused by the interruption of luminal fluid in the head of epididymis . Loss of germ cells usually leads to testicular atrophy [43, 44], yet interestingly, the TKO testes had an average mass comparable to that of age-matched WT testes. This may reflect a defect in luminal fluid absorption in the TKO mice, resulting in retained fluid and maintained testicular mass.
The deletion of Lpar1, Lpar2, and Lpar3 affects male reproductive function in at least two ways: reduced mating activity, and an age-related increase of sterile males. Even at the peak of fertility (age, ~3 mo), the copulation frequency of TKO mice as a population was approximately one third that of age- and background-matched WT controls (Fig. 3a, b). Reduced mating activity also has been reported in ER and Bcl6 knockout mice [41, 45]. The Lpar1/2/3 TKO males, however, do not initially produce significantly smaller litters, suggesting that early sperm reduction (~50% at ~3 mo of age) is not sufficient to affect the litter size, similar to what has been reported in Arl4a, centromere protein B, or testis-specific cytochrome c knockout mice [43, 46, 47]. A study with Fshb and activin receptor IIA knockout mice indicated that a sperm reduction to 8% of that in WT animals was needed to reduce litter sizes to 50% of those for WT animals. The litter size was not affected even when the sperm count was 27% of that in WT animals .
Spermatogenesis is controlled by endocrine and local (paracrine/autocrine) signals . With no change in testosterone level in the TKO mice, the documented role of LPA itself as a paracrine/autocrine mediator in other systems  suggests that impaired spermatogenesis in the LPA receptor(s) KO males is caused by local LPA signaling defects that affect cell proliferation and/or apoptosis. A similar effect also was seen in Schwann cells of the peripheral nervous system . It has been established that germ cell degeneration limits spermatogenesis and that survival factor deprivation leads to apoptosis . Furthermore, the phenotype is consistent with the antiapoptotic effects of LPA that involve Gi/o signaling [35, 36] and includes activation of AKT1 . Because LPA-induced Gi/o signaling was impaired in Lpar1/2/3 TKO testis primary cells (Fig. 6b), LPA1, LPA2, and/or LPA3, coupled at least to Gi/o signaling, can account for the germ cell survival effects of LPA. The potential roles of other G proteins and their downstream signaling pathways that mediate the survival effect of LPA on male germ cells require further study. The results presented here thus implicate LPA as a novel lipid factor influencing spermatogenesis through its signaling effects on germ cell survival.
The endogenous source of LPA is likely to be local rather than through the bloodstream. This interpretation is supported by studies with transgenic mice that ubiquitously overexpressed phosphatidic acid phosphatase 2a (PPAP2A; or lipid phosphate phosphatase-1 [LPP-1]), an enzyme that degrades LPA, and that showed disrupted spermatogenesis without affecting bloodborne LPA levels . The presence of LPA and LPA metabolic enzymes in human seminal fluid  suggests that LPA signaling may have similar roles in human spermatogenesis.
To our knowledge, the present study is the first to identify receptor-mediated lysophospholipid signaling as an important factor in maintaining mating activity and normal sperm production via three identified receptors, LPA1, LPA2, and LPA3. Many issues remain to be explored. These include the source(s) of signaling LPA and LPA metabolism, involvement of other LPA receptors (both known and unknown), actual downstream signaling pathways, intersections with other signaling pathways, and neural mechanisms associated with behavioral deficits. Additionally, the age-related progression of germ cell degeneration associated with loss of LPA signaling suggests a temporal window during which sperm production could be pharmaceutically altered. Thus, LPA drug-based therapies could potentially result in new treatment options for male infertility or a new contraceptive strategy. Combined with previous data regarding female reproductive functions , these results underscore pivotal roles for LPA signaling in mammalian reproduction through defined receptor subtypes.
We thank J.J.A. Contos, C. McGiffert, S. Kawamura, S. Miyamoto, J.H. Brown, I. Sadler-Riggleman, M.A. Kingsbury, C. Akita, Y. Yung, J.W. Westra, C. Paczkowski, D.R. Herr, and D. Letourneau for technical assistance and critical reading of the manuscript.
1Supported by National Institutes of Health grant R01 HD050685 to J.C.