Severely Reduced Female Fertility in CD9-Deficient Mice

See allHide authors and affiliations

Science  14 Jan 2000:
Vol. 287, Issue 5451, pp. 319-321
DOI: 10.1126/science.287.5451.319


CD9 is a widely expressed cell surface molecule that belongs to the tetraspanin superfamily of proteins. The tetraspanins CD9, KAI-1/CD82, and CD63 are involved in metastasis suppression, an effect that may be related to their association with β1 integrins. Knockout mice lacking CD9 were created to evaluate the physiological importance of CD9. CD9−/− females displayed a severe reduction of fertility. Oocytes were ovulated but were not successfully fertilized because sperm did not fuse with the oocytes from CD9−/− females. Thus, CD9 appears to be essential for sperm-egg fusion, a process involving the CD9-associated integrin α6β1.

Sequence analysis of CD9 predicts a structure with four transmembrane domains, two extracellular loops, and short intracytoplasmic amino and carboxyl tails (1). This structure is shared by all members of the tetraspanin superfamily of proteins (2). CD9 is expressed in multiple tissues but is not ubiquitous (3). The tetraspanins CD9, KAI-1/CD82, and CD63 act as metastasis suppressor molecules (4). Low expression of these molecules is correlated with an increased invasive and metastatic potential (2, 5). In vitro studies show that tetraspanins function in cell adhesion, motility, proliferation, differentiation, and signal transduction (2,6). The tetraspanins are physically associated with each other and with several cell surface molecules, including a subset of β1 integrins, which are receptors for extracellular matrix proteins (7). Thus, tetraspanins may function as surface organizers or facilitators that group and interconnect specific cell surface proteins in macromolecular complexes and thus increase the formation or the stability of functional signaling complexes, or both (2, 7).

To investigate the role of CD9 in vivo, we generated mice in which the CD9 gene was disrupted by gene targeting in embryonic stem (ES) cells (8, 9). The murine CD9 gene consists of eight exons spanning more than 20 kb (10, 11). The promoter and exon 1 were replaced by the neomycin resistance gene (Web fig. 1) (12). The recombination was verified by Southern (DNA) blot (13). Splenocytes from homozygous mice displayed no detectable CD9 surface expression as determined by flow cytometry analysis or by immunohistochemistry of kidney sections (10).

Heterozygous CD9+/− mice appeared normal and, when interbred, yielded litters of normal size with a Mendelian genetic distribution 25:50:25(%) and with an equal distribution between males and females. Homozygous adult CD9−/− mice showed no obvious abnormalities and appeared healthy. However, when the CD9−/− mice were intercrossed, fertility was severely reduced. Normal fertility (93 to 100% of pregnancies) was observed when CD9−/− males were mated with wild-type or heterozygous females. In contrast, only 50 to 60% of CD9−/− females produced litters after being maintained in the presence of fertile wild-type, heterozygous, or CD9−/− males for up to 2 months. The delay before the beginning of a successful pregnancy was 19 to 30 days on average for CD9−/− females, as compared with 4.5 days for wild-type animals. In addition, the litter size was reduced (2 ± 0.6 pups in CD9−/− females versus 8 ± 2.3 pups in CD9+/+ females) and the initial mortality rate was increased to 32 to 55% as compared with less than 2% for wild-type females (Web table 1) (12).

No difference in the frequency of vaginal plugs was observed between wild-type and CD9-deficient mice, indicating normal mating behavior (16.9%; n = 46 for CD9−/− versus 17.4%;n = 42 in CD9+/+). The infertility was also not due to the absence of sperm at the site of fertilization, as shown by the presence of numerous sperm in the oviduct. Histological examination of ovaries of 6-week-old CD9−/− mice showed that there was no difference when compared to wild-type females (10) and that the number of ovulated oocytes was normal (Table 1). However, the naturally ovulated oocytes from CD9−/− mated females collected at day 0.5 did not divide and became fragmented if maintained in cell culture (Fig. 1A), whereas oocytes from CD9+/+ mated females divided and 2 days later reached stage 4 (Fig. 1B). Similar fragmented oocytes (Fig. 1C), instead of blastocysts (Fig. 1D), were present in the uterus of CD9−/− females 3.5 days after mating.

Figure 1

Morphology of oocytes or embryos recovered after natural ovulation and mating of either CD9+/+ or CD9−/− knockout females. Oocytes from CD9−/−(A) and CD9+/+ (stage 2 or 4) (B) females were kept in culture for 2.5 days. Embryos or oocytes were recovered from the uterine horns 3.5 days after mating. (C) Fragmented oocytes obtained from CD9−/−females. (D) Embryos at blastocyst stage obtained from wild-type females. Scale bar, 30 μm.

Table 1

Effect of CD9 deficiency on ovulation rate and oocyte fertilization in vivo. Ovulated oocytes were recovered from the oviduct ampulla and examined by phase contrast microscopy. Granulosa cells were removed with type III hyaluronidase (0.1 mg/ml; Sigma). Oocytes from naturally ovulated CD9+/+ females had one or two polar bodies: a residual first polar body in some and a second polar body in all cases. Values are the mean ± SEM. NA, not applicable.

View this table:

The oocytes recovered from nonmated superovulated CD9−/−mice (Fig. 2A) or CD9+/+ mice (Fig. 2C) 12 hours after human chorionic gonadotropin (post-hCG) injection had a similar response: 70 and 74%, respectively, showed a polar body (Web table 1) (12), which indicates completion of meiosis I. Labeling of these oocytes with a monoclonal antibody (mAb) to tubulin YL1/2 and propidium iodide showed that they were blocked in metaphase of the second meiotic division (Fig. 2, B and D). The first polar body is unstable because less than 25% of the oocytes collected at 17 hours post-hCG had a polar body (10). The oocytes recovered from mated CD9−/−females at day 0.5 frequently had several sperm (up to 10) in the perivitelline space (Fig. 2E) (Web table 1) (12), suggesting that they were not competent for fertilization. Few of these oocytes (17%) had a residual first polar body, but DNA-tubulin labeling showed that CD9−/− oocytes [12 out of 13 (12/13)] remained blocked in metaphase of the second meiotic division (Fig. 2F). One or two polar bodies were observed in 94% of oocytes from CD9+/+ mated females (Table 1) (Fig. 2G) in which DNA-tubulin labeling showed that most oocytes (16/19) were fertilized and had exited the second meiotic metaphase. This was confirmed by the metaphase plate disappearance and by the extrusion of the second polar body (Fig. 2H).

Figure 2

Oocyte DNA-tubulin labeling. Oocytes were fixed (23) and incubated with mAb YL1/2 (24), followed by fluorescein isothiocyanate (FITC)–labeled goat F(ab′)2antibody to mouse. Oocytes from nonmated superovulated CD9−/− (A and B) or CD9+/+ (C and D) females were identically blocked in metaphase of meiosis II as confirmed by the presence of a metaphase spindle (*) and a first polar body containing residual DNA-tubulin material (solid arrowheads). (E andF) An oocyte from a CD9−/− naturally ovulated mated female blocked in metaphase of meiosis II with sperm in the perivitelline space (open arrowheads). (G and H) An oocyte from a CD9+/+ naturally ovulated mated female that has progressed to telophase and has extruded the second polar body (arrow), whereas the first one remains visible. Inset shows partially decondensed sperm DNA in a different focal plane. Scale bar, 10 μm.

CD9 was strongly expressed on ovulated oocytes (Fig. 3, A and B) and blastocysts (Fig. 3, C and D) from CD9+/+ but not on oocytes from CD9−/− females. The expression of CD9 was also observed on follicular oocytes in the ovary, whereas no expression was found on follicular cells (10). Next, in vitro fertilization experiments (14) were performed with the oocytes of five superovulated C57BL/6 females. For each female, oocytes collected from the two oviducts were kept separate and were fertilized in the presence of either the mAb to mouse CD9 4.1F12 (15) or a control rat mAb. The fertilization rates, determined by the number of stage 2 to 4 embryos at 24 to 48 hours after fertilization, were 4% (n = 42) and 49% (n = 43), respectively, showing an inhibition of fertilization by the CD9 mAb.

Figure 3

CD9 expression on oocytes and embryos from wild-type mice. Staining was performed with a rat mAb to mouse CD9 4.1F12 (15) followed by FITC-coupled goat F(ab')2 antibody to rat, and was then examined under a fluorescence microscope. CD9+/+ oocytes denuded of cumulus cells (A and B) and CD9+/+blastocysts (3.5-day-old embryos) (C and D) were labeled and washed in microdrops of phosphate-buffered saline without fixation. Strong labeling of the cell surface was observed. The labeling of CD9−/− oocytes was identical to background, which was negligible and is not shown here. Scale bar, 30 μm.

To distinguish between a defect in fusion or in subsequent events, we performed a transfer test with the dye Hoechst 33342 (16). In 35 zona pellucida (ZP)–free oocytes of four superovulated CD9−/− females, we observed no sperm-egg fusion, whereas in five superovulated CD9+/+females (68 oocytes), there was a mean sperm-egg fusion of 75 ± 19% (10). No difference in sperm-egg binding was observed. This shows that CD9 is involved in sperm-egg fusion.

Female infertility has previously been shown to be linked to oocyte growth and maturation or ovulation (17). A defect of fertilization was also observed in females lacking the ZP glycoprotein ZP3 (18). Here we have shown that in the absence of the tetraspanin CD9, the ovaries and ovulation rate are normal. However, most of the ovulated oocytes are not competent for fusion with sperm. As in other cells, CD9 may participate in oocytes in multimolecular complexes and may even play a role in the formation of these complexes. It should be noted that the integrins α3β1, α5β1, and α6β1, which are associated with CD9 (7,19), are expressed on oocytes (20). Because the integrin α6β1 has been shown to bind to sperm fertilin and could function as a sperm receptor (21), it is possible that CD9 plays a role in sperm-egg fusion by modulating α6β1 receptor function.

The effect of the CD9 knockout on fertility contrasts with the lack of apparent phenotype in other tissues expressing high levels of this molecule (22). In tissues outside of the ovary, CD9 may be functionally replaced by one or several tetraspanins. Our data show the involvement of CD9 in oocyte fertilization and reveal a biological function of a tetraspanin. It now must be questioned whether certain human infertilities are caused by a defect of this molecule.

  • * To whom correspondence should be addressed. E-mail: boucheix{at}


View Abstract

Navigate This Article