Carbonic Anhydrase IV Deficiency Causes Intrauterine Embryonic Loss in Mice

The implantation of the mammalian embryo is a complex process [1]. During the initial phases of pregnancy in mammals, the interaction between the uterus and the embryo is of high importance for ensuring the embryo’s successful implantation and subsequent development. Morphological and physiological alterations within the endometrium, including processes like the epithelial-mesenchymal transition and decidualization, are under control of ovarian steroids, namely, estrogen and progesterone [2]. This interaction is very susceptible to disruption and about 20% of pregnancies in human end at the time of implantation [3]. To maintain the nutrition of the developing embryo, development of a placenta is of outmost importance. The specialized cells required for placental development such as trophoblast and extraembryonic mesoderm are formed during the early stages of embryonic development. Therefore, the precise understanding of the molecular interactions leading to successful embryo implantation and placentation is considered essential from a clinical point of view [4].

The mouse is widely used as an animal model as it is well suited for reproductive studies. As in humans, uterine receptivity for implantation lasts for a limited time, and the decidual reaction only occurs in a uterus appropriately primed with progesterone and estrogen. However, in contrast to humans, physiological decidualization in mice is only induced in the presence of a blastocyst. The feto-maternal interaction of the embryonic trophoblast with the maternal decidua is indispensable for regular embryogenesis as well as adequate placenta formation. This interaction is orchestrated by highly complex and finely tuned molecular and cellular processes, which are regulated by multiple factors [5, 6]. Extending our knowledge of these complex events is indispensable to prevent infertility caused by implantation failure.

Carbonic anhydrases (CAs) are zinc enzymes that catalyze the conversion of CO₂ into the soluble hydrogen carbonate HCO₃⁻ [7] and play a crucial role in regulation of carbon dioxide concentrations throughout the body. The activity of the enzymes is dependent on both their abundance and the pH level [8]. At least 16 different isozymes of CAs have been found in mammals [9]. Cellular localization can be divided into cytosolic (CA I, II, III, VII, and XIII), membrane-bound (CA IV, IX, XII, XIV, and XV), mitochondrial (CA VA and VB), and salivary secreted (CAVI) forms [8]. The majority of these isozymes are generated by various members of a gene family that appears to be evolved from a common precursor through gene duplication. CAIV has been identified previously as a key enzyme for the early activation of sperm motility [10]. The function of CAs includes pH-regulating processes that are essential for sperm storage, motility, and capacitation during fertilization [11]. Moreover, CAIV expression has been shown in various tissues including lung, heart, muscles, liver, brain, kidney, gallbladder, distal intestine, placenta, and specialized capillaries [12‒19]. Unlike other isoenzymes, CAIV is attached extracellularly to the plasma membrane by a glycosyl-phosphatidylinositol anchor, in contrast to a membrane-spanning domain, and plays an active role in mediating the transport of CO₂ and HCO₃⁻ [18]. Thus, it may be involved in maintaining proper homeostasis and may be also associated with the blood-brain barrier in cerebral capillaries [20]. Because of these properties, CAIV may also play an essential role in placental function and embryo development. During embryo development, regulation of CAIV previously has been analyzed in late murine pregnancy on gestational days 11, 15, and 19. CAIV protein has been localized in the endodermal layer of the yolk sac (both intra- and extraplacental) and in the uterine epithelium but was mostly expressed in the labyrinth, at the interface between fetal capillaries and maternal blood, and increased with gestation [21]. Yet there is no information on the distribution of CAIV at the time of implantation and during early embryonic and placental development. Moreover, deletion of CAIV resulted in reduced litter size, which preferentially affected female offspring [22]. However, it was not clear from this study at what time the loss of the embryos, mainly female embryos, took place. The aim of the present study thus was to analyze the distribution of the CAIV protein in early stages of pregnancy and placenta development from 4.5 to 9.5 dpc and to determine the time span in which intrauterine loss of the embryos occurs.

C57Bl/6J wild-type mice were bred in the animal facility of the University Hospital Essen. CAIV knockout B6.129S1-Car4tm1Sly/J mice were provided by the laboratory of William S. Sly (Department of Biochemistry and Molecular Biology, St. Louis University School of Medicine, St. Louis, MO, USA) [22]. Mice were maintained at a 12-h light/dark cycle and had free access to tap water and food. As only organs were removed for the investigations and no animal experiments were carried out, approval by the local Ethics Committee was not necessary (ethics approval waiver was granted). The mice were treated in accordance with guidelines approved by the University of Duisburg-Essen Animal Care and Use Committees as well as the State Office for Nature, Environment and Consumer Protection of North Rhine-Westphalia (LANUV, protocol 84-02.04.2014.A219; address: Leibnizstrasse 10, 45659 Recklinghausen, Germany) before the start of the investigations. Institutional animal care facility was the Zentrales Tierlaboratorium (ZTL; Universitätsklinikum Essen, Hufelandstraße 55, 45147 Essen, Germany). Animals were sacrificed by cervical dislocation.

Female reproductive organs of nonpregnant mice were obtained from the estrous phase of the ovarian cycle. Cycle monitoring was performed by microscopic evaluation of vaginal smear. For analysis of implantation sites and embryos, female mice were mated overnight with males, and the day of vaginal plug finding was defined as day 0.5 of pregnancy (dpc). Mice were euthanized by cervical dislocation. After dissection, tissues were fixed overnight with formaldehyde (4%), dehydrated, and embedded in paraffin. In total, 29 mice were sacrificed in this study. For immunohistochemical evaluation, organs from three different female nonpregnant wild-type mice and implantation sites of at least three different mice of each of the six pregnancy stages investigated (4.5–9.5 dpc) were analyzed. For assignment of number of implantation chambers and sex determination of embryos, uteri of three pregnant heterozygous and five pregnant CAIV knockout mice were dissected at 9.5 dpc.

C57Bl/6J wild-type mice were used to analyze the localization of CAIV in the female reproductive tract and in embryos by immunoperoxidase method. Briefly, 7-μm-thick paraffin sections were deparaffinized in xylol and rehydrated in ethanol of decreasing concentrations. After rinsing in distilled water, slices were blocked with PBS-glucose-oxidase buffer and incubated with a polyclonal goat IgG antibody against CAIV (AF2414; R&D Systems, Minneapolis, MN, USA) (1:100 in PBS/5% BSA) overnight at 4°C, followed by incubation with biotinylated anti-goat IgG (1:200 in PBS/5% BSA) for 30 min at room temperature. The VECTASTAIN^® Elite ABC kit for peroxidase (Vector Laboratories, Burlingame, CA, USA) was used for signal enhancement according to the protocol of the manufacturer. Detection was carried out with 3,3′diaminobenzidine (Sigma, St. Louis, MO, USA) as chromogen. Nuclear staining was performed with hematoxylin (Roth, Karlsruhe, Germany). After staining, slides were dehydrated in ethanol of raising concentrations and embedded in Eukitt^® (Sigma). Negative controls were performed by using an unspecific serum instead of the primary antibody on consecutive sections. Murine renal tissue was used as a positive control. Morphology of implantation sites was examined using hematoxylin and eosin-stained paraffin section. Hematoxylin-eosin staining was performed according to standard protocols. All slides were examined with a Leica DM4000 (Leica, Nussloch, Germany) microscope.

For sex determination of implanted embryos, 26 embryos of three different pregnant WT mice and 21 embryos of 5 different pregnant CAIV knockout mice were analyzed at day 9.5 pc. Three separated gene combinations were analyzed: X-chromosomal Xlr (X-linked lymphocyte-regulated complex) and Y-chromosomal Sly (Sycp3-like Y-linked), the homologous genes Uba1 and Ube1y1 on the X and Y chromosome, and Y-chromosomal Zfy (zinc finger protein Y-linked) have been amplified by PCR. In brief, 30-μm sections of embedded uteri from pregnant mice were prepared and microdissected to exclude maternal uterus tissue. For this purpose, embryonic tissue was isolated under microscopic control from 10 paraffin sections (thickness 30 μm) of each embryo using a scalpel and was transferred to an Eppendorf vial. Genomic DNA then was isolated from the embryonic tissues using ReliaPrep FFPE gDNA Miniprep System (Promega, Mannheim, Germany) according to the manufacturer’s instructions. PCRs were performed in a final volume of 25 μL containing 500 ng genomic DNA using standard PCR protocol as follows: 94°C for 2 min followed by 35 cycles of 94°C for 30 s, 57°C for 30 s, and 72°C for 45 s. PCR products were subsequently separated on a 2% agarose gel containing GelRed^® Nucleic Acid Gel Stain (Biotium Inc., Fremont, CA, USA) and were visualized by ultraviolet light. PCR products: Sly/Xlr PCR showed amplicons of 480/650 bp for XX and 280 bp for XY, Uba/Ube PCR 200 bp for XX and 180 bp for XY, Zfy PCR 190 bp for XY. Primer sequences are listed in online supplementary Table S1 (for all online suppl. material, see https://doi.org/10.1159/000544000) and one representative agarose gel is shown in online supplementary Figure SF1.

For the examination of significances, Student’s t test was used. Probability values of ≤0.05 were considered as statistically significant.

To show whether CAIV is consecutively expressed in the female reproductive tract, immunostaining with an antibody directed against CAIV was performed on mice ovaries (Fig. 1a–d), fallopian tubes (Fig. 1e, f), uteri (Fig. 1g), and vaginal tissue (Fig. 1h) of three different wild-type mice. Murine renal tissue served as a positive control (Fig. 1i). In the nonpregnant state, no immunostaining for CAIV is visible in the ovary (different follicle states and corpus luteum), the fallopian tube, and the vagina. In the uterus, we detect a slight staining in small vessels of the endometrium, whereas the uterine epithelium and the stromal cells exhibit no immunoreactivity.

At developmental stage 4.5 dpc, murine blastocysts from wild-type mating show a strong immunostaining for CAIV (Fig. 2a). In the pregnancy stages up to 7.5, the CAIV signal is shown in the visceral endoderm (VE) and the embryonic ectoderm (IE) (Fig. 2b–d). There is no immunoreactivity in the extraembryonic ectoderm and the ectoplacental cone. No staining of the endometrial tissue is observed. At stage 8.5 dpc (Fig. 3a–c), a strong CAIV immunostaining is visible in the amniotic epithelium lining the amniotic sac and the yolk sac epithelium (Fig. 3b). Additionally, CAIV is induced in the endometrial epithelium of the maternal compartment (Fig. 3c). At stage 9.5 dpc (Fig. 4a–h), strong immunostainings are present within the developing embryonic body, e.g., in the gut tube, the floor plate of the neural. Furthermore, the yolk sac epithelium, the trophoblast giant cells (TGC), and the endometrial epithelium, which form together the yolk sac placenta, show a strong immunostaining for CAIV. Moreover, also the labyrinthine compartment of the developing hemochorial placenta is CAIV positive. These results are summarized in Table 1.

In comparison to heterozygous mating of wild-type and CAIV knockout mice, homozygous mating of CAIV knockout mice results in a significant reduction of litter size (Fig. 5a). In heterozygous mating, the CAIV genotype of the offspring does not follow the mendelian rules. While genotyping of pups show an enhanced number of wild-type (31% instead of 25% expected) and heterozygous genotype (62% instead of 50% expected), the number of CAIV knockout pups is significantly reduced (7% instead of 25% expected) (Fig. 5c).

Interestingly, in CAIV knockout mice, this reduction of litter size is mainly due to a reduced number of female mice born. The sex distribution proved to be non-mendelian as 74% male and 26% female are born instead of the expected 50% each (Fig. 5b). In addition, the sex distribution from heterozygous mating (wild-type x CAIV knockout) turns out to be non-mendelian. While in both wild-type (36% male vs. 64% female) and CAIV heterozygous (46% male vs. 54% female) pups there is a slight preponderance of female pups, CAIV knockout pups resulting from heterozygous mating are 67% male versus 33% female (Fig. 5d). This reduction of litter size takes place after mid-stage gestation since on 9.5 dpc with the start of placenta development, the number of implantation chambers is similar in heterozygous and CAIV knockout mating (Fig. 5e). At this early stage on 9.5 doc, the sex distribution of embryos is rather shifted to female in both experimental groups, showing a higher percentage of female embryos from heterozygous mating in total (40% male vs. 60% female) as well as in CAIV knockout embryos resulting from heterozygous mating (40% male vs. 60% female). Likewise, CAIV knockout mating results in a higher percentage of female embryos on day 9.5 of pregnancy (43% male vs. 57% female) (Fig. 5f). Thus, the observed reduction of litter size and loss of female embryos appears to occur during the phase of placentation after 9.5 dpc.

In this study, the distribution of the CAIV protein in the female reproductive tract of nonpregnant mice and during embryo implantation and placentation from 4.5 to 9.5 dpc was analyzed. Additionally, the effect of a deletion of CAIV on these processes was evaluated in a CAIV-deficient mouse model. As to date, the distribution of the CAIV protein in the murine genital tract had not been analyzed. We here show that the CAIV protein is absent in the epithelial and the stromal cells of the organs of the female genital tract (ovary, fallopian tube, uterus, vagina) of nonpregnant mice. Only weak signals are observed in endothelial cells of the endometrium. Here, CAIV potentially could be involved in the special immune situation of the uterus since it has also been described on endothelial cells of other organs revealing an immune privilege [14, 15]. During early pregnancy, CAIV is induced in the uterine epithelium adjacent to the yolk sac from 8.5 dpc onward, complementing the observations of Rosen et al. [21] who demonstrated CAIV in the uterine epithelium of pregnancy stages from 11 dpc onward. CAIV staining of decidual cells was not observed in the early pregnancy stages analyzed in our study, thus appearing only in later pregnancy stages as described by Rosen et al. [21] for gestational day 11 and ahead. The presence of CAIV was very clearly evident in the early embryo. It was present in the embryoblast already shortly before trophoblast invasion at 4.5 dpc and extended with the progression of pregnancy to the VE and embryonic ectoderm. It was visible in the visceral layer of the yolk sac already from 8.5 dpc, supplementing the findings described before by Rosen et al. [21] from gestational day 11 onward. Thus, our results complement the studies of Rosen et al. [21] showing that CAIV is not only active in the mature placenta but also during early placental development (Table 1). We here demonstrate that during implantation, an immunostaining of CAIV is present in the blastocyst, the VE, and the embryonic ectoderm. This might be explained by the need of fine adjustment of pH in the new developing embryonic cavities filled with fluids [23]. In contrast to the findings of Rosen et al. [21], we also observed CAIV immunostaining in TGC during early pregnancy.

Yolk sac epithelium, Reichert’s membrane, TGC, and the uterine epithelium together form the yolk sac placenta, which is important to ensure the nutrition of the murine embryo during early development. The role of CAIV in this context is not yet clear and so far can only be speculated. The induction of CAIV in these compartments might support the transport function of the yolk sac placenta [21]. It is discussed that carboanhydrases like CAIV possibly can facilitate paracellular regulations by locally changing the extracellular pH value, by this controlling receptor affinities or regulating pH-dependent transporters [24, 25].

Interestingly, more CAIV-deficient male than female pups were born. This is in accordance with findings by Shah and coworkers [26], who, however, did not investigate the period during which the predominantly female embryos are lost. We here show that there is no misbalance in sex distribution during fertilization, early embryo development, and embryo implantation, but the observed misbalance in sex distribution is due to a higher death rate of female embryos during placentation after 9.5 dpc. Female-biased embryonic death has already been shown for other nonsex chromosome-associated mutations [27] but remains a largely unexplained phenomenon that requires further analysis. Taken together, we show that CAIV seems to be involved in some way in the precisely balanced signaling network of embryo development, implantation, and placentation.

The authors thank Kerstin Bahr, Natalie Knipp, and Christian von Massow for excellent technical assistance.

This study protocol was reviewed and approved by the University of Duisburg-Essen Animal Care and Use Committees as well as the State Office for Nature, Environment and Consumer Protection of North Rhine-Westphalia (LANUV, protocol 84-02.04.2014.A219; address: Leibnizstrasse 10, 45659 Recklinghausen, Germany) before the start of the investigations.

The authors have no conflicts of interest to declare.

Funding was provided by the Deutsche Forschungsgemeinschaft (DFG) to G.W. (WE 2344/9-3).

Conceptualization: G.W. and R.G. Data curation: S.S., J.T.D., G.W., R.G., and S.G. Funding acquisition: G.W. Project administration: G.W. Supervision: G.W. and R.G. Writing – original draft preparation: G.W., S.G., R.G., and J.T.D. Writing – review and editing: G.W., R.G., J.T.D., S.G., and S.S.

Current affiliation for Sven Schumann: Institute of Anatomy, Brandenburg Medical School Theodor Fontane (MHB), Neuroppin, Germany.

All data generated or analyzed during this study are included in this article and its online supplementary material files. Further inquiries can be directed to the corresponding author.