Male Sexual Dysfunction in Mice Bearing Targeted Mutant Alleles of the PEA3 ets Gene

Abstract

PEA3, a member of the Ets family of transcriptional regulatory proteins, is expressed in a unique spatial and temporal pattern during mouse embryogenesis; its overexpression is positively correlated with HER2-mediated breast tumorigenesis in both humans and mice. To determine whether PEA3 plays a part in development and oncogenesis and to uncover its normal physiological role, we generated mice lacking functional PEA3 by gene targeting in embryonic stem cells. PEA3^−/− mice arose from heterozygous crosses with the expected Mendelian frequency, revealing that PEA3 is dispensable for embryogenesis. PEA3 mutant mice displayed no overt phenotype and lived a normal life span. However, PEA3-deficient males failed to reproduce. PEA3 is expressed in several male sexual organs, but gross and histological analyses of the organs from PEA3^−/− mice revealed no abnormalities. Spermatogenesis and spermiogenesis also appeared normal in mice homozygous for the PEA3 mutation, and their sperm were capable of fertilizing eggs in vitro. PEA3^−/− males engaged in normal mating behavior, but they did not set copulatory plugs and sperm could not be detected in the uteri of females that had mated with PEA3^−/− males. Erections could be evoked by abdominal pressure in PEA3-deficient male mice, and the results of in vitro experiments revealed that the corpus cavernosum isolated from PEA3 mutant males relaxed in response to acetylcholine. Therefore, the infertility of PEA3 mutant males involves either mechanisms proximal to the cavernosal smooth muscle or an ejaculatory dysfunction. However, PEA3 mutant mice are phenotypically distinguishable from other knockout mice with such deficits and thus provide a unique model for further investigation of male sexual dysfunction.

The ets genes, which currently comprise nearly 30 paralogs in mammals, encode transcription factors bearing conserved sequence-related DNA binding domains (the ETS domain) (16). Ets proteins are capable of regulating transcription by binding to sites in the promoters of their cognate target genes. DNA binding is achieved by interaction between the ETS domain and an ∼10-bp sequence element termed the Ets binding site (EBS) comprising a highly conserved core sequence, 5′-GGA(A/T)-3′. Individual Ets proteins demonstrate specificity for sequences flanking this core, but it is not uncommon for different Ets proteins to bind to the same EBS.

PEA3 (36) is the founding member of a subfamily of Ets proteins, which also includes ER81 (6) and ERM (11, 26). Members of this subfamily possess nearly identical ETS domains and harbor additional regions of sequence similarity. Analyses of the transcriptional properties of individual PEA3 subfamily members reveal that they commonly activate transcription (6, 11, 26, 36). Whereas few bona fide PEA3 target genes have been identified, transient-transfection studies suggest that PEA3 is capable of regulating the transcription of genes whose products facilitate cell motility and invasion (reviewed in reference 13).

PEA3 is expressed in a spatially and temporally restricted pattern during mouse development in cells derived from each of the three germ layers and in regions of the embryo undergoing cellular proliferation and migration (11). PEA3 appears to be preferentially expressed at sites of epithelium-mesenchyma interactions. In the developing chick, PEA3 is selectively expressed in specific classes of motor neurons and corresponding muscle afferent sensory neurons at limb levels of the spinal cord (24). PEA3 expression by both motor and sensory neurons is governed by signals derived from the limb muscle. In adult mice PEA3 RNA is most abundant in the brain and epididymis (36).

Overexpression of PEA3 is associated with breast cancer in both humans and mice, suggesting a role for PEA3 in this malignancy (3, 34). Seventy-six percent of all human breast tumors contain elevated levels of PEA3 RNA; 93% of the c-ERB-B2 (also known as HER2)-positive subclass of these tumors overexpress PEA3 (3). PEA3 is also overexpressed in all mouse mammary tumors arising in transgenic mice engineered to overexpress murine c-Erb-B2 (also known as Her2) in their mammary glands (34). These findings suggest the possibility that PEA3 plays a role in mammary oncogenesis or its progression.

To assess the role of the PEA3 gene in embryonic development, adult physiology, and oncogenesis, we introduced a loss-of-function mutation in this gene in the mouse germ line. PEA3 mutant mice were viable, but analyses of males revealed an unexpected role for the protein in male sexual function.

MATERIALS AND METHODS

Construction of a PEA3 targeting vector.

Bacteriophage lambda recombinants bearing mouse genomic PEA3 DNA were isolated from a 129/sv library using a full-length murine PEA3 cDNA as a probe (36). The entire mouse PEA3 gene was sequenced, and this information was used to construct a targeting construct whose structure is illustrated in Fig. 1A. The targeting vector comprised 6.0 kb of PEA3 DNA on one side and 10.8 kb on the other; these sequences were separated by a PGKneoPA expression cassette (25). The targeting vector was linearized with NotI prior to its introduction into embryonic stem (ES) cells.

FIG. 1.

Generation of a mouse line bearing a targeted deletion at the PEA3 locus. (A) Schematic diagram of the PEA3 gene and the targeting construct (pBSneo9331) used to generate the mutant allele. The location of the probe (840-bp ClaI/KpnI fragment) used in Southern blot analyses is marked on the schematic. A naturally occurring NcoI site in exon 6 was converted to an EcoRI site to facilitate molecular cloning of the targeting construct. All other restriction enzyme cleavage sites shown on the schematic occur naturally in the PEA3 gene. The length of the expected EcoRI digestion products (RI) representative of the WT PEA3 gene (8.2 kb) and targeted PEA3 allele (7.6 kb) that would be detected in Southern analyses using the appropriate hybridization probe are indicated by horizontal lines. mut, mutant. (B) Southern analysis of the genomic DNA of four G418-resistant ES cell clones recovered after transfection of the targeting construct. The BE8 ES cell line did not undergo a homologous recombination event and is included here for comparative purposes. The odd-numbered lanes contain the products homologous to the probe in EcoRI/ClaI double digestions of cellular DNA, whereas the even-numbered lanes contain products of EagI-cleaved cellular DNA that were homologous to the probe. (C) Southern analysis of ClaI/EcoRI-cleaved cellular DNA from the tails of mice of the F₂ generation, illustrating germ-line transmission of the targeted PEA3 allele. Results for three genotypes (PEA3^+/+ [+/+], PEA3^+/− [+/−], and PEA3^−/− [−/−]) (D) RNase protection products from brain RNA of PEA3^+/+ (+/+), PEA3^+/− (+/−), and PEA3^−/− (−/−) mice. The sizes of the observed RNase-protected products as well as that of the phosphoglycerate kinase (PGK-1) internal control are shown on the right of the autoradiogram. Marker DNAs of the indicated size (in base pairs) were electrophoresed in lane 1. The products of digestion of the riboprobes following hybridization to tRNA are shown in lane 2. Lanes 3 to 5 contain the products of digesting hybrids between the probes and total cellular RNA from the brains of mice of the three genotypes. The asterisk denotes the expected RNase digestion product from the mutant PEA3 gene. (E) Immunoblot analysis of PEA3 protein in nuclear lysates of mouse embryo fibroblast (MEF) cell lines derived from WT (+/+) and PEA3^−/− (−/−) mice. PEA3 protein was detected with a mixture of two different monoclonal antibodies. Lane 1, MEF-4; lane 2, MEF-D; lane 3, MEF-1; lane 4, MEF-H. The various PEA3 protein species migrate at apparent molecular masses of 69, 64, and 63 kDa.

Isolation of PEA3 mutant mice.

ES cells of the J1 line were cultured and electroporated as described previously (5). Homologous recombination events at the PEA3 locus were identified by simultaneous cleavage of ES cell DNA with ClaI and EcoRI, followed by Southern blotting (31) and hybridization with a radiolabeled probe corresponding to an 840-bp ClaI/KpnI DNA fragment (Fig. 1A). The structure of the targeted locus was confirmed by EagI digestion of ES cell DNA and Southern analysis using the same DNA probe. Chimeras were generated by blastocyst injection as described previously (5). A male chimera was mated with BALB/c females to generate an outbred mouse line and with 129/sv females to generate an inbred 129/sv line.

RNase protection analysis of tissue RNA.

RNA was isolated from tissues and organs using the guanidinium thiocyanate method (10). RNase protection experiments were performed as described previously (36).

Analysis of PEA3 protein in mouse embryonic fibroblast cell lines.

Fibroblast cell lines were established from individual mouse embryos harvested at 13.5 days postcoitus using a 3T3 protocol (33). Nuclear extracts were prepared from these cells as described previously (23). Protein samples (25 μg) were electrophoresed through a sodium dodecyl sulfate–10% polyacrylamide gel and transferred to a membrane. Membranes were incubated with a mixture of two PEA3-specific monoclonal antibodies (MP-13–MP-16) followed by a goat anti-mouse secondary antibody. The Western blots were developed by chemiluminescence.

Histological analyses.

Mice were euthanized and perfused with 4% paraformaldehyde prior to dissection. Organs were fixed overnight in 4% paraformaldehyde, dehydrated through a graded series of ethanol baths, and embedded in paraffin wax. The paraffin blocks were cut into 8.0-μm-thick sections, adhered to glass slides, dewaxed, counterstained with hematoxylin and eosin, and mounted with Permount. Slides were photographed using a Zeiss Axioskop microscope.

Analysis of male fertility.

In vitro fertilization of mouse oocytes was performed as described previously (4). Eggs that had cleaved were scored as fertilized; one-celled oocytes were recorded as unfertilized. Observations were made using an Olympus S-30 dissecting microscope. In vivo fertilization was performed as described previously (18). BALB/c mice were superovulated and paired with a male overnight. The following morning, females were examined for the presence of a copulatory plug and for the presence of spermatozoa in their uterine horns and vagina. Pronucleation was scored as described above.

Determination of hormone concentrations by RIA.

Blood was collected from the inferior vena cava using a 21-gauge needle attached to a 3-ml syringe containing 0.1 ml of heparin. Blood samples were transferred to 1.5-ml tubes and centrifuged for 15 min at 4°C. The plasma was collected and stored at −20°C. The levels of luteinizing hormone (LH) and follicle-stimulating hormone (FSH) in plasma were measured by radioimmunoassay (RIA) with reagents provided by the National Hormone and Pituitary Program of the National Institute of Health and Human Development. Anti-rat LH (S-11), anti-rat FSH (S-11), and reference preparations RP-3 (LH) and RP-2 (FSH) were used in these assays. The intra-assay coefficient of variation for the quality controls ranged from 6.65 to 13.99% for LH and 6.2 to 14.3% for FSH. Testosterone levels were measured by RIA using antibody-coated tubes manufactured by ICN Biomedicals. The assay had a sensitivity of 0.22 ng/ml and less than 7.8% reactivity with other relevant hormones including dihydrotestosterone. The intra-assay coefficient of variation for the quality controls ranged from 2.5 to 8.2%.

Induced erections.

The procedure used to evoke erections in male mice has been described previously by Sachs (30). Briefly, male mice were allowed to run partway into a plastic cylinder (56 mm long by 26 mm wide) and then placed in a supine position with their hind legs held by the tester. The penile sheath was retracted and held in this position using a cotton-tipped applicator held at the base of the penis. Gentle pressure was applied to the mouse's abdomen with the tester's finger for 10 to 15 s, and the occurrence of penile erection was noted. Erections were scored as either strong (engorgement, color change, and change in length and circumference) or weak (poor engorgement and little change in dimensions).

In vitro relaxation of corpora cavernosa.

Mice were sacrificed by cervical dislocation, and the corpora cavernosa was dissected in situ. One corpus cavernosum from each mouse was suspended in a jacketed organ bath (1 ml, 37°C) containing a modified Krebs-Henseleit solution (118 mM NaCl, 4.7 mM KCl, 1.2 mM MgSO₄, 1.2 mM KH₃PO₄, 25 mM NaHCO, 2.5 mM CaCl₂, 5.6 mM glucose). A mixture of 95% oxygen and 5% carbon dioxide was bubbled through the solution. The tissues were subjected to 0.2 g of tension and allowed to equilibrate for 30 min. Isometric relaxation responses to single concentrations of acetylcholine were measured after contracting the corpus with phenylephrine (10 μM). The tissues were then washed by overflow for 10 min. The tissue was equilibrated for 5 min and then contracted again with phenylephrine before another concentration of acetylcholine was tested. No more than five different concentrations of acetylcholine were used per experiment. Drugs were prepared freshly on the day of the experiment. Phenylephrine (Sigma) was dissolved in Krebs-Henseleit solution; acetylcholine (Sigma) stock solution was prepared in 0.1 mN HCl and diluted in Krebs-Henseleit solution.

RESULTS

Isolation of mice with targeted loss-of-function PEA3 alleles.

We constructed a gene targeting vector by deleting PEA3 gene sequences spanning exons 6 through 11 and replacing these sequences with an expression cassette bearing the neo gene (Fig. 1A). We selected this region for disruption because it includes exon 11, which encodes an essential region of the ETS DNA binding domain (36). The linearized targeting vector was electroporated into J1 ES cells, and G418-resistant clones were screened by Southern analysis for homologous recombination events. The wild-type (WT) and targeted PEA3 alleles were distinguished by cleavage of ES cell DNA with ClaI and EcoRI and hybridization with an appropriate DNA probe (Fig. 1A). An 8.2-kb DNA fragment or a 7.6-kb fragment is expected to result from cleavage of the cellular DNA of WT and PEA3 mutant mice, respectively. Three ES cell clones (CF11, DD1, and DE3) of the 210 that were screened yielded both 8.2- and 7.6-kb fragments, suggesting that they contained the targeted allele (Fig. 1B, lanes 3, 5, and 7). Cleavage of the DNA of a representative G418-resistant ES cell clone (BE8) that had not undergone homologous recombination is also shown (Fig. 1B, lane 1). Homologous recombination was confirmed by cleavage of genomic DNA from these same ES cell lines with EagI, which cleaves once within exon 2 in the WT PEA3 gene (Fig. 1A). A second EagI site (present in the PGKneoPA cassette) should be present in the appropriately targeted mutant PEA3 allele. Southern analyses revealed the presence of the predicted 10-kb fragment in EagI-cleaved ES cell DNA from the three clones bearing the targeted PEA3 allele (Fig. 1B, lanes 4, 6, and 8); this fragment was not detected in the EagI-cleaved DNA from the B8 ES cell line (lane 2).

The three ES cell clones were independently injected into BALB/c blastocysts, but only one, DD1, produced chimeric mice. A male chimera was backcrossed to BALB/c females, and heterozygous F₁ mice were interbred to produce the F₂ generation. Southern analysis of ClaI/EcoRI-cleaved tail DNA from these mice revealed the occurrence of all three genotypes in the F₂ generation, illustrating germ line transmission of the targeted locus (Fig. 1C).

To determine whether the targeted allele was expressed, RNase protection experiments were performed with brain RNA, a relatively abundant source of PEA3 transcripts (36). The antisense riboprobe spanned exons 5 through 7, which included the region of insertion of the neo expression cassette in the targeting vector (Fig. 1A). Full-length PEA3 mRNA should protect a 323-bp fragment following RNase digestion. This fragment and three others (∼176, ∼166, and ∼147 bp) were detected following analysis of RNA from the brains of PEA3^+/+ mice (Fig. 1D, lane 3). The 176-bp protected species very likely resulted from hybridization of mRNA lacking exon 7 to the riboprobe. It is noteworthy that the differential splicing of exon 7 does not alter the reading frame of the PEA3 mRNA. Furthermore, we have cloned a PEA3 cDNA that lacks only this exon (data not shown). These observations suggest that the 176-bp protected species likely resulted from the hybridization of a bona fide differentially expressed PEA3 mRNA to the riboprobe. The structures of the transcripts that gave rise to other protected species are not known with certainty, but preliminary analyses suggest that they too are alternatively spliced mRNAs.

The targeted allele contains a partial deletion of exon 6 and lacks all the other downstream exons present in the antisense RNA probe. Assuming normal transcription of the targeted allele and processing of its nuclear transcript, a protected RNA species of 153 bp should be detected in RNA from PEA3^+/− and PEA3^−/− mice. A protected fragment of this size was observed following analysis of the brain RNA samples from PEA3^+/− and PEA3^−/− mice, suggesting that the targeted locus is transcribed and that the resulting transcript is stable (Fig. 1D, lanes 4 and 5). As expected, the 323-bp species, representative of the WT PEA3 gene transcript, was not present in the RNA samples from the PEA3^−/− mice (lane 5). It is noteworthy that none of the other protected species, presumably representative of alternatively spliced mRNAs, were observed in RNA samples from the PEA3^−/− mice. The failure to detect these species is likely due to the absence of the appropriate donor and acceptor splice sites in the targeted allele.

To determine whether expression of the PEA3 protein occurred in the organs of mice representative of each genotype, we carried out immunoblot analyses of extracts prepared from their brains and epididymides using PEA3-specific monoclonal antibodies. We were invariably unsuccessful in detecting the PEA3 protein in these samples even after immunoprecipitation of lysates followed by immunoblot analysis of the resulting immunocomplexes (M. A. Laing and J. A. Hassell, unpublished data). Our inability to detect the protein likely resulted from the fact that PEA3 is expressed in only a small fraction of the cells within most tissues and organs.

In an attempt to circumvent this problem, we established mouse embryo fibroblasts from each genotype and assayed nuclear extracts prepared from these cells for the PEA3 protein, because previous analyses of 3T3 mouse fibroblasts revealed the occurrence of relatively abundant PEA3 transcripts in these cells (36). We used two monoclonal antibodies that recognize different antigenic determinants encoded by exon 8; all of our monoclonal antibodies bind to epitopes that are encoded by exons that are missing from the targeted PEA3 gene. Two independent PEA3^+/+ mouse embryo cell lines expressed the three species of PEA3 protein commonly detected in other mouse cell lines (Fig. 1E, lanes 1 and 2). These PEA3 proteins were not detected in two independent PEA3^−/− cell lines (lanes 3 and 4). Because we do not have antibodies capable of detecting amino-terminal epitopes in the PEA3 protein, we have been unable to determine whether the targeted locus encodes a truncated amino-terminal fragment of the PEA3 protein. However, as noted below, the phenotype conferred by mutation of the PEA3 gene is discernible only in PEA3^−/− mice and not in PEA3^+/− mice. Hence, it is unlikely that the targeted locus encodes a dominant-acting, truncated PEA3 protein.

PEA3 mutant mice are viable, but males are sterile.

To learn whether the loss of PEA3 affected embryogenesis, we measured the occurrence of the PEA3 mutation among the progeny of heterozygous crosses. These crosses yielded offspring that segregated PEA3 alleles in the expected Mendelian ratio (33+/+:57+/−:26−/−), as determined by Southern analyses, suggesting that loss of PEA3 is not associated with embryonic lethality. Analyses of the sex of these mice revealed that PEA3^−/− mice of both sexes were viable. Moreover, there were no overt differences in the appearance, growth rate, behavior, state of health, or life span of PEA3^−/− mice compared to their PEA3^+/− or WT littermates (Laing and Hassell, unpublished).

To determine whether loss of PEA3 affected reproductive ability, mating pairs of the various genotypes of both sexes were bred and scored for offspring. All mating pairs that included a PEA3^−/− male failed to produce litters, even after pairing more than 30 different PEA3^−/− males with multiple females over a 12-month period (Laing and Hassell, unpublished). Heterozygous males fathered pups at the same frequency as WT males did, and females of all three genotypes reproduced normally. Moreover, the average litter size of PEA3^−/− females was similar to those of their PEA3^+/+ and PEA3^+/− counterparts. PEA3^−/− females nursed their pups, and they developed at the same rate as the offspring of PEA3^+/− and PEA3^+/+ females did (Laing and Hassell, unpublished).

PEA3 is expressed in male reproductive organs.

To aid elucidation of the cellular and molecular bases for the reproductive failure of PEA3^−/− males, we determined the expression profile of PEA3 by examining a larger repertoire of organs than we had previously analyzed (36). We measured PEA3 RNA levels by RNase protection analysis from the organs of male mice at weekly intervals from 3 to 8 weeks of age, a period during which males achieve sexual maturity. PEA3 RNA was expressed in the brain, spinal cord, kidney, small intestine, skeletal muscle, testis, and epididymis of 3- to 8-week-old males (a representative example from a 4-week-old male mouse is shown in Fig. 2A). The highest levels of expression occurred in the brain (lanes 2 and 3), spinal cord (lane 4), thymus (lane 6), small intestine (lane 12), skeletal muscle (lane 13), testis (lane 14), and epididymis (lane 15). Much lower levels of PEA3 RNA were also detected in the lung (lane 7), kidney (lane 11), and salivary gland (16) after prolonged exposure of the autoradiogram. We were unable to detect PEA3 RNA in the liver, spleen, and pancreas after this or later exposure times (Laing and Hassell, unpublished). The only change in the PEA3 expression profile between 3 and 8 weeks of age was in the heart. PEA3 RNA was expressed in the heart at low but detectable levels in the 3-week sample, but not in samples prepared thereafter (Laing and Hassell, unpublished).

FIG. 2.

RNase protection analyses of RNA from organs of 4-week-old male mice. (A) RNase protection products using total RNA from various male mouse organs. Only the principal 323-nucleotide PEA3 protected species and the 124-nucleotide protected product of the PGK-1 transcript used as an internal control is shown. Thirty micrograms of RNA was analyzed from each source. Lane 1, mouse FM3A cells; lane 2, cerebrum; lane 3, cerebellum; lane 4, spinal chord; lane 5, heart; lane 6, thymus; lane 7, lung; lane 8, liver; lane 9, pancreas; lane 10, spleen; lane 11, kidney; lane 12, small intestine; lane 13, skeletal muscle; lane 14, testes; lane 15, epididymis; lane 16, salivary gland. (B) RNase protection products of RNA isolated from the various regions of the epididymis of an adult male mouse. Lane 1, PEA3 and PGK-1 riboprobes used for the analysis; lane 2, products of RNase T₂ digestion of the riboprobes following hybridization with tRNA; lane 3, products of digestion of the hybrids between the probes and RNA from the initial segment and caput epididymis; lane 4, products of digestion of the duplexes between the probes and RNA isolated from the corpus epididymis; lane 5, products of digestion of the hybrids between the probes and RNA from the cauda epididymis. The numbers to the left of the gel indicate the sizes (in base pairs) of the products.

These results and our previous analysis (36) revealed relatively high levels of PEA3 transcripts in the epididymis, a male sexual organ. The epididymis is a convoluted tubular epithelium that connects the testis with the vas deferens (1). Sperm originates in the testis and matures during passage through the epididymis where it is also stored. The epididymis is organized into three morphologically and functionally distinct segments: the proximal region closest to the testis comprising the initial segment and caput epididymis; the middle region or corpus epididymis; and the most distal region, the cauda epididymis. Each epididymal region is thought to be functionally distinct and expresses a unique set of proteins (12).

To learn whether PEA3 is expressed in a distinct segment of the mouse epididymis, we surgically dissected the organ and analyzed RNA isolated from each of the three major segments by RNase protection assays (Fig. 2B). PEA3 transcripts were restricted principally to the initial segment and caput epididymis (lane 3). Much lower levels of PEA3 RNA were detected in the corpus (lane 4) and cauda epididymis (lane 5) that could be accounted for by contamination of these regions with tissue from the initial segment and caput epididymis. Analyses of the rat epididymis also revealed that PEA3 transcripts and protein are expressed in the initial segment of this organ likely in the principal cells (22). Hence, PEA3 expression occurs in several male reproductive organs of the mouse, including the testes and epididymides, which are required for spermatogenesis, spermiogenesis, and sperm maturation.

Histological analyses of the testes and epididymides.

The expression of PEA3 in these male reproductive organs raised the possibility that male infertility in PEA3^−/− mice might be associated with a deficit in one or both of these tissues. To learn whether the development and cellularity of the testis and epididymis were affected by disruption of the PEA3 gene, we performed histological analyses of these organs from WT and PEA3^−/− littermates. The gross structure of the testes appeared to be the same for mice of both genotypes (Fig. 3A and B). The architecture of the testicular seminiferous epithelium also appeared to be the same in mice of both genotypes; Sertoli cells, spermatogonia, primary spermatocytes, spermatids, and spermatozoa were all present in approximately equivalent numbers in sections of testes from each genotype (Laing and Hassell, unpublished). The sections also revealed the presence of Leydig cells and stromal cells outside of the basal laminae surrounding the seminiferous tubules. Hence, there appeared to be no discernible morphological or cellular differences between the testes of PEA3^+/+ and PEA3^−/− mice.

FIG. 3.

Histology of the testes and the initial segment of the epididymis. (A and B) Seminiferous epithelium of a PEA3^+/+ tubule (A) and a PEA3^−/− tubule (B). Original magnification of panels A and B, ×50. Bar, 250 μm. (C and D) Transverse sections of the initial segment from PEA3^+/+ and PEA3^−/− epididymides. Original magnification, ×200. Bar, 100 μm. Lu, lumen.

Histological analyses of the three principal segments of the epididymides of these mice, including the epithelium of the initial segment where PEA3 is expressed, also revealed no obvious abnormalities in any of these segments (Fig. 3C and D). Both basal cells and principal cells containing stereocilia were present, forming the high columnar epithelium characteristic of the initial segment (32). Furthermore, these analyses revealed the storage of similar amounts of sperm in the cauda segment of the epididymides of mice of both genotypes (Laing and Hassell, unpublished). Hence, histological analyses of male sexual organs that express PEA3 did not provide an explanation for the infertility of PEA3^−/− males.

Sperm from PEA3^−/− males is capable of fertilizing eggs in vitro.

In vitro fertilization analyses were performed to address whether the spermatozoa from PEA3^−/− males are capable of fertilizing oocytes. A series of six experiments were performed comparing sperm samples from an outbred WT mouse strain (ICR) (as a positive control) with those of PEA3^+/+ and PEA3^−/− mice. This analysis revealed that spermatozoa from the PEA3^−/− mice fertilized oocytes as efficiently in vitro as did the sperm samples from the PEA3^+/+ and ICR control mice (Table 1). Statistical analyses showed that the mean fertilization rates were not significantly different between the sperm samples of the two genotypes. Hence, under conditions of these in vitro fertilization experiments, sperm from PEA3^−/− males did not appear to harbor any intrinsic defects.

TABLE 1.

In vitro fertilization analysis of sperm samples from ICR (control), PEA3^+/+, and PEA3^−/− mice

Expt	Control (ICR) mice (n = 6)		PEA3+/+ mice (n = 5)		PEA3−/− mice (n = 7)
	No. of eggs	Fertilization rate (%)	No. of eggs	Fertilization rate (%)	No. of eggs	Fertilization rate (%)
1	33	51.0	29	24.0	30	50.0
2	28	71.0	ND	ND	26	80.0
3	NDa	ND	ND	ND	29	68.9
4	19	68.0	18	61.0	20	80.0
5	30	43.3	29	44.8	27	44.0
6	26	62.0	27	56.0	30	60.0
7	48	65.0	38	55.2	37	67.5
Total no. of eggs	184		141		199
Fertilization rate
Meanb		60.05		48.20		64.34
SD		10.713		14.755		13.901
SEM		4.37		6.60		5.25

^aND, not done. 

^bMean percentage of the total number of eggs fertilized for each genotype in the seven experiments. The mean fertilization rates were analyzed by one-way analysis of variance followed by Tukey's multiple-range test. Tukey's test was used rather than Duncan's test because of the unequal number of mice (n) for each group. None of the three means were significantly different. 

In vivo fertilization analyses.

To further investigate the fertility deficiency in male PEA3 mutant mice, we determined whether these males successfully copulated with females. To this end, PEA3^+/+, PEA3^+/−, and PEA3^−/− males were independently paired overnight with superovulated female BALB/c mice. The following morning, females were examined for the presence of a copulatory plug. Copulatory plugs were found in approximately 70% of the females paired with either PEA3^+/+ or PEA3^+/− males (Table 2). However, not a single copulatory plug was detected in 113 pairings of females with 36 PEA3^−/− males. The uteri of females were routinely flushed after mating and examined for the presence of spermatozoa. Spermatozoa were detected in the uteri of females mated with either PEA3^+/+ or PEA3^+/− males, but they were not found in the uteri of females mated with PEA3^−/− males (Laing and Hassell, unpublished). The pronucleation rate, assessed by cleavage of isolated oocytes cultured in vitro, was ∼80% in one-cell embryos recovered from matings of females with PEA3^+/+ and PEA3^+/− males, whereas it was less than 2% in embryos isolated from females following mating with PEA3^−/− males (Table 2). Two eggs from females that had mated with PEA3^−/− males did undergo one cleavage. However, the first cleavage was unequal and both eggs failed to develop further in culture. This suggests that these cleavages were not likely the result of a true fertilization event but were parthenogenic in nature (4).

TABLE 2.

Assessment of fertility of male mice of different genotypes

Characteristic	PEA3+/+	PEA3+/−	PEA3−/−
No. of males mated	24	30	36
No. of mating trialsa	70	90	113
No. of plugs set (%)	46 (66)	65 (72)	0 (0)
Pronucleation rate (%)b	78	81	<2c

^aEach mating trial was set up with one superovulated female. 

^bPronucleation was scored by observing the cleavage of isolated zygotes. 

^cDue to parthenogenesis. 

Circulating levels of male sex hormones and male sexual behavior are normal in PEA3 mutant mice.

The levels of various circulating hormones affect male sexual function (37). To learn whether the concentrations of sex hormones in the blood of the PEA3^−/− male mice were affected by loss of functional PEA3, we measured the levels of FSH, LH, and testosterone by RIA. The concentrations of all of these hormones in the PEA3^−/− male mice were within normal limits and did not differ significantly from those of the PEA3^+/+ and ICR control mice (Table 3). This observation suggests that male infertility is not a consequence of a neuroendocrine disorder.

TABLE 3.

Mean hormone concentrations in male mice^a

Hormone	Mean hormone concn (ng/ml) in mice
	Control (ICR)	PEA3+/+	PEA3−/−
Testosterone	12.69	16.38	18.97
FSH	12.43	10.00	10.48
LH	2.52	1.18	2.60

^aThe mean concentrations of testosterone, FSH, and LH were determined by RIA for the male mice sacrificed for the in vitro fertilization analysis. The data were analyzed by one-way analysis of variance followed by Tukey's multiple-range test. The mean hormone concentrations for the three groups of mice were not significantly different. 

We also observed the sexual behavior of PEA3^−/− males and compared it to the behavior of age-matched PEA3^+/+ littermates that had been paired with superovulated females. PEA3^−/− males, like their WT counterparts, displayed normal grooming behavior and mounted females, suggesting that their sexual behavior was not affected by mutation of the PEA3 gene (Laing and Hassell, unpublished).

Erections could be induced in male PEA3 mutant mice.

The transfer of spermatozoa into the female reproductive tract requires penile erection and subsequent ejaculation. Penile erection is a complex neurovascular event. Within the penis it involves relaxation of the corporal smooth muscle and subsequent engorgement with blood of the paired corpora cavernosa due to compression of the emissary veins against the connective tissue sheath surrounding the corpora. This process is regulated by the central and peripheral nervous systems.

Erections can be induced artificially in mice by a combination of retraction of the penile sheath and gentle abdominal pressure (30). The mechanism of the erection is uncertain but may be reflexic (30) or due to obstruction of the venous outflow from the penis. To determine whether the penile tissue of PEA3 mutant mice was capable of supporting an erection, the abdominal pressure test was applied to male mice of each genotype. Almost all of the male mice tested (n = 34) by this method achieved strong erections; only four males achieved weak erections, which were characterized by poor engorgement and little increase in the length of the penis (Table 4). Importantly, 11 of 13 homozygous PEA3 mutant male mice achieved strong erections by this method. These findings suggest that the PEA3 mutant male mice do not harbor any physical impediments to achieving penile erections. This conclusion was sustained by histological analyses of penis sections from mice of the various genotypes. These analyses failed to reveal any differences in the tissue architecture among males of all three genotypes (Laing and Hassell, unpublished). Taken together, these findings suggest that the penile structures required for erection were normal in PEA3 mutant males.

TABLE 4.

Incidence of erections evoked by abdominal pressure in male mice of different genotypes

Characteristic	PEA3+/+	PEA3+/−	PEA3−/−
No. of males tested	10	11	13
No. of strong erectionsa	10	9	11
No. of weak erectionsb	0	2	2

^aA strong erection is a normal erection, characterized by engorgement. 

^bA weak erection is characterized by poor engorgement with little change in dimensions. 

Erectile tissue isolated from PEA3 mutant males was functional in vitro.

Parasympathetic nerves, via cholinergic and noncholinergic neurotransmission, are responsible for the corporal smooth muscle relaxation underlying erection, but the major part of their effect is due to the liberation of nitric oxide (NO) and subsequent activation of biochemical events in the muscle (2, 8, 19, 21). Deficits in erection could occur at the level of the corpus cavernosum smooth muscle (17) or in the neural organization of relaxation. We directed our attention initially at the level of the corpus, because other attempts to produce genetic models of impotence targeted this level (9, 17).

In order to determine whether the erectile tissue of the PEA3 mutant mouse has the biochemical machinery required to support relaxation, we measured the capacity of this tissue to relax in response to acetylcholine in vitro (2). Corpora cavernosa isolated from male mice of each PEA3 genotype did not display spontaneous contractile activity in vitro. The α-adrenoreceptor agonist phenylephrine contracted isolated corpora cavernosa (average increase in tension, 39.6 ± 5.3 mg) (n = 20). There was no difference between the magnitude of contraction elicited in tissue from PEA3^−/− males (41.4 ± 9.0 mg) (n = 11) and that in tissue from WT and PEA3^+/− males combined (37.4 ± 4.6) (n = 9). Acetylcholine added in the presence of phenylephrine-induced contracted corpora cavernosa, produced a concentration-dependent relaxation of this tissue (Fig. 4A). Dose-response curves constructed for acetylcholine-induced relaxation were similar for corporal tissue isolated from mice of all three genotypes (Fig. 4B). Hence, the erectile tissue from the PEA3 mutant mice demonstrated a normal relaxation response to acetylcholine in vitro.

FIG. 4.

Effect of acetylcholine on mouse corpus cavernosum in vitro. (A) Representative traces are from PEA3^−/− mice. The tissue was contracted with 10 μM phenylephrine (presence of phenylephrine shown by the horizontal bars) (mean response 39.6 ± 5.3 mg). Arrows indicate administration of acetylcholine to give the final concentrations shown to the right of the arrows (in micromolar). (B) Mean concentration-response relationships for acetylcholine (ACh) in the three genotypes (PEA3^−/− [−/−], PEA3^+/− [+/−], and PEA3^+/+ [+/+]). Data points are means for 4 to 12 tissue samples. Bars representing standard errors of the means are shown only in one direction and only on two curves to improve clarity.

DISCUSSION

PEA3 appears to be required for normal male sexual function. All PEA3 mutant males that we have characterized are unable to impregnate females, suggesting that the penetrance of this phenotype is absolute. Moreover, introgression of the targeted PEA3 allele to other mouse genetic backgrounds (129/sv and FVB) did not alter the nature or penetrance of the described phenotype of male PEA3^−/− mice (Laing and Hassell, unpublished). Gross and histological analyses of male reproductive organs including the epididymis, which expresses relatively high levels of PEA3 RNA in its initial segment, revealed no discernible differences between WT and PEA3^−/− males. Hence, the inability of PEA3^−/− males to impregnate females does not appear to result from morphological aberrations or cellular deficits in these organs.

The sperm of PEA3^−/− males proved capable of fertilizing eggs in vitro, suggesting that spermatogenesis, spermiogenesis, and sperm maturation occur normally in PEA3 mutant males. However, we have not assessed PEA3^−/− sperm motility or the rate with which these sperm bind to and penetrate eggs. Moreover, it is noteworthy that the sperm/egg ratios used in these in vitro fertilization studies greatly exceed the ratios common in vivo. Hence, it is formally possible that some aspect of sperm function is compromised by loss of functional PEA3.

PEA3 mutant males displayed normal grooming and mating behavior. These males, like their age-matched WT counterparts, repeatedly mounted superovulated females shortly after being paired with them. However, PEA3^−/− males did not produce vaginal plugs in females. This finding is consistent with our inability to detect sperm in the reproductive tracts of females shortly after mating with PEA3^−/− males; by contrast, sperm was readily detected in the reproductive tracts of females that mated with PEA3^+/+ or PEA3^+/− males. These findings are compatible with an erectile and/or ejaculatory dysfunction in male PEA3 mutant mice.

The circulating levels of various sex hormones can affect erectile function (37). However, the concentrations of LH, FSH, and testosterone were the same in WT and PEA3 mutant males. Hence, if PEA3^−/− males have an erectile dysfunction, it is unlikely to be of neuroendocrine origin. Commonly, sterility resulting from neuroendocrine malfunction is manifested in the form of defects in the maturation of secondary sexual characteristics, such as hypogonadism, a phenotype that was not apparent in PEA3^−/− male mice (37). Indeed, all of the male sexual organs, including the genitalia, of PEA3-deficient males were identical to those of their WT littermates.

There was no obvious relationship between loss of PEA3 function and lack of fecundity in male mice. Therefore, we initially chose to test two very basic requirements for erectile function, namely, that the penile tissue had the required structure to engorge and become erect when venous outflow was blocked and that a fundamental biochemical step in the mechanism underlying erection was present in corporal tissue.

Abdominal compression produced erection in the majority of male mice of all three genotypes. Although the mechanistic basis of abdominal pressure-induced erections is unclear, it probably impairs venous outflow from the penis. It is an artificial procedure and bypasses the muscular pumping that underlies erections in rodents (15, 30). Whatever its shortcomings, this test demonstrated that the penile structures required for erection were present in PEA3^−/− males. This was confirmed by histological analyses of the penises of PEA3^−/− males, which also did not reveal any obvious structural deficits of this organ.

A critical biochemical step for producing the corporal relaxation underlying erection is the release of NO and subsequent intracellular generation of cyclic GMP (cGMP) (2, 8, 19). Cholinergic agonists cause relaxation through release of NO in vascular tissue including corpus cavernosum (17, 21). In our experiments, corporal relaxation induced by acetylcholine in tissue from PEA3^−/− mice was indistinguishable from that observed in tissue from PEA3^+/+ or PEA3^+/− mice. This result implies that the molecular targets and biochemical steps through which acetylcholine causes relaxation in corporal tissue are active and intact in PEA3^−/− mice.

This conclusion is supported by the phenotype of mice bearing targeted deletions in the downstream effectors of acetylcholine action in the corpus cavernosum. Targeted disruption of the gene for cGMP-dependent kinase I (cGK-I), the likely molecular target for the cGMP generated by NO action, prevents cholinergically mediated corporal relaxation (17). Unlike PEA3^−/− male mice, cGK-I^−/− male mice have a substantially reduced but finite reproductive capacity. Similarly, specific disruption of the neuronal isoform of the NO-synthesizing enzyme, nitric oxide synthase (nNOS), does not eliminate male potency (9, 20, 28). Whereas alternative splicing of nNOS transcripts (8) as well as other mediators and genetic sources of NOS activity may compensate for nNOS gene disruption in these mutant mice (14), the sexual phenotype of PEA3^−/− males is clearly distinguishable from that of nNOS mutant males.

An alternative interpretation of the observation that PEA3 mutant males are unable to set copulatory plugs in timed matings is that they harbor an ejaculatory dysfunction. However, the phenotype of PEA3 mutant male mice is also distinguishable from that of other knockout mice, which display clear ejaculatory defects. The neural messenger, carbon monoxide (CO), like NO, is implicated in neurotransmission (7). CO is synthesized by heme oxygenase 2 (HO2), whose expression is localized to neurons that mediate ejaculation. HO2 knockout male mice are fertile, despite the fact that they mount females less frequently and display reduced intromission activity compared to their WT counterparts (7). Moreover, the reflex activity of the bulbospongiosus muscle, a muscle that plays a significant role in promoting fertility in the male mouse (15), is substantially reduced in HO2 mutant male mice (7).

Similarly, male mice with a targeted disruption of the P2X1 ATP receptor also display ejaculatory abnormalities that result in a 90% reduction in fertility (27). ATP is released with noradrenaline from sympathetic neurons and acts through P2X1 receptors on smooth muscle to effect contraction. Disruption of the P2X1 gene reduces the contractile response within the smooth muscle of the vas deferens, leading to a reduced sperm count in the ejaculate and resulting in reduced male fertility (27). Importantly, P2X1 null male mice copulate normally and consistently set plugs in timed matings, characteristics that distinguish them from PEA3 mutant male mice (27). Hence, whatever the underlying cellular and molecular bases for the sexual dysfunction of PEA3 mutant males, these mice are clearly phenotypically distinguishable from other mouse mutants with erectile and ejaculatory deficits described so far.

We suspect that the infertility of PEA3^−/− males has an underlying neuronal basis. PEA3 is expressed in specific bundles of motor neurons that innervate limb muscles and in afferent sensory neurons of these same muscles (24). Hence, it is conceivable that PEA3 is also expressed in neurons that innervate the penis. To address this, we are now deriving new PEA3 mutant mice by homologous recombination in ES cells that carry bacterial beta-galactosidase in place of PEA3 coding sequences. Analysis of the pattern of beta-galactosidase activity in such male mice may help to uncover the cellular basis and ultimately the molecular basis of male infertility in PEA3^−/− mice. Nerve stimulation experiments to determine whether there is a peripheral neurotransmission defect in these mice are also under way. The possibility that there is an ejaculatory defect is also being considered. Whatever the deficit, these mutant mice afford a unique model for studies of male sexual dysfunction and its treatment. Moreover, contingent on the nature of the lesion in PEA3^−/− mice, our findings suggest the potential that PEA3 antagonists may act as effective male contraceptives.