Uncoupling protein 2, but not uncoupling protein 1, is expressed in the female mouse reproductive tract.

Uncoupling proteins (UCPs) are transporters of the inner mitochondrial membrane. Whereas UCP1 is uniquely present in brown adipose tissue where it uncouples respiration from ATP synthesis and activates respiration and heat production, UCP2 is present in numerous tissues, and its exact function remains to be clarified. Two sets of data provided the rationale for this study: (i) the intriguing report that UCP1 is present in uterus of mice (Nibbelink, M., Moulin, K., Arnaud, E., Duval, C., Penicaud, L., and Casteilla, L. (2001) J. Biol. Chem. 276, 47291-47295); and (ii) an observation that Ucp2(-/-) female mice (homozygous matings) have smaller litters compared with Ucp2(+/+) animals (S. Rousset and A.-M. Cassard-Doulcier, unpublished observations). These data prompted us to examine the expression of UCP1 and UCP2 in the reproductive tract of female mice. Using wild type, Ucp1(-/-) mice, and Ucp2(-/-) mice, we were unable to detect UCP1 in uterus of mice with appropriate antibodies, and we conclude that the signal assigned to UCP1 by others was neither UCP1 nor UCP2. Using a polyclonal antibody against UCP2 and tissues from Ucp2(-/-) mice as controls, UCP2 was detected in ovary, oviduct, and uterus. Expression of Ucp2 mRNA was also observed in ovary and uterus using in situ hybridization analysis. Bone marrow transplantation experiments revealed that the UCP2 signal of the ovary was restricted to ovarian cells. UCP2 level in ovary decreased during follicular growth and increased during the pre-ovulatory period, during which aspects of an inflammatory process are known to exist. Because UCP2 down-regulates reactive oxygen species, a role in the regulation of inflammatory events linked to the preparation of ovulation is suggested.

the first member identified, is uniquely expressed in brown adipocytes where it stimulates heat production by uncoupling oxidative phosphorylation (2)(3)(4). The thermogenic potential of BAT is strictly related to UCP1 and is proportional to UCP1 amount. Genes encoding homologues of the brown fat UCP1 have been identified in mammals, birds, marsupials and also in plants or fungi (see reviews in Refs. [5][6][7][8]. Despite the significant homology of these proteins with UCP1 and their apparent uncoupling activity in heterologous expression systems (9), there are no conclusive results about their in vivo function (10,11).
Contrary to Ucp1, Ucp2 is widely expressed in mammalian tissues (12,13), and the presence of the corresponding protein was demonstrated in macrophages, spleen, gut, lung, and white adipose tissue mitochondria (14) as well as in pancreatic islets (15,16) and thymocytes (17). The phenotypic analysis of Ucp2(Ϫ/Ϫ) mice indicates that this gene does not play a role in either thermogenesis or the regulation of body fat content but is strongly implicated in the regulation of ROS production (18), as was proposed previously for proteins able to uncouple respiration (19,20). It was also reported that UCP2 acts as a negative regulator of insulin secretion in pancreatic islets by decreasing ATP level (15). In addition, very recent studies support a role of UCP2 in immunity and inflammatory responsiveness (14,21,22).
A recent publication has presented evidence for Ucp1 expression in the longitudinal smooth muscle of uterus (1). This observation challenges the prevailing view that Ucp1 is expressed exclusively in brown adipocytes. In addition, it has already led to other publications in which a role for UCP1 has been proposed in mechanisms of smooth muscle relaxation in intestine (23). Accordingly, we decided to reinvestigate the putative presence of UCP1 in this organ with UCP1-deficient mice and also to investigate the presence of UCP2 in the reproductive tract of female mice. This particular study was, to an extent, justified by our observations that the Ucp2(Ϫ/Ϫ) mice raised on a C57BL/6J genetic background have litters significantly lower than the Ucp2(ϩ/ϩ) mice (5.6 Ϯ 0.4 and 7.9 Ϯ 0.9 pups, respectively, p Ͻ 0.01). 2
Animals and Treatments-Studies on mice were performed in agreement with the institutional CNRS guidelines defined by the European Community guiding principles and by the French decree number 87/848 of October 19, 1987. Authorization to perform animal experiments was given by the French ministry of Agriculture, Fisheries, and Food (A92580 issued February 2 1994 and 92-148 issued May 14, 2002). C57BL/6J, OF1, and B6D2F1 female mice were obtained from Elevage Janvier (Orléans) and maintained under a 12-h light, 12-h dark schedule with food and water ad libitum. Rats from the Wistar strain were used. The Ucp1(Ϫ/Ϫ) and the Ucp2(Ϫ/Ϫ) mice were described previously (18,24). Ucp2(Ϫ/Ϫ) and Ucp1(Ϫ/Ϫ) mice were raised on a C57BL/6J genetic background. To cycle mice, mature female were intraperitoneally injected with 5 IU of PMSG to stimulate follicular growth. They were also intraperitoneally injected 48 h later with 5 IU of hCG to trigger ovulation and luteinization. Timed pregnancies of mature C57BL/6J female mice were determined by the presence of a vaginal plug. Plugged females were isolated, and this day was taken as day 0 of pregnancy. Mice were euthanized by cervical dislocation, and tissues were removed immediately and processed for mitochondria extraction as described previously (14). To collect tissues for in situ hybridization analysis, mice were anesthetized with 0.3-0.5 ml of a mixture containing 2 mg/ml ketamine and 2.5 mg/ml xylazine immediately and perfused intracardially with 20 ml of ice-cold isotonic saline followed by 100 ml of paraformaldehyde (4%) solution. At the end of perfusion, the reproductive tract was removed. To carry out bone marrow transplantation in mice, medullar aplasia was induced by 9.5-gray total body irradiation of Ucp2(ϩ/ϩ) or Ucp2(Ϫ/Ϫ) mice. Mice were reconstituted intravenously under anesthesia with 1.2 ϫ 10 6 bone marrow cells extracted from the femur and tibia of either Ucp2(ϩ/ϩ) or Ucp2(Ϫ/Ϫ) mice. Mice were allowed to recover for 4 months.
Western Blot Analysis-Mitochondrial protein extracts were loaded onto polyacrylamide gels, separated by electrophoresis, and transferred onto nitrocellulose membrane as described previously (14). Western blots were hybridized with the appropriate antibodies (see above) according to the following conditions: UCP1 detection, 11A antibody (at 1/4000 dilution) and NA934 rabbit anti-IgG (at 1/3000 dilution); UCP2, hUCP2-605 antibody (at 1/10,000 dilution) and rabbit anti-IgG (at 1/3000 dilution); and COX, anti-COX antibody (at 1/4000 dilution) and mouse anti-IgG (at 1/3000 dilution). For analysis of UCP1, Western blots were developed by autoradiography using Hyperfilm ECL-TM as described in Nibbelink et al. (1). The amount of UCP2 was normalized by using the antibody against COX. Direct recording of the chemiluminescence corresponding to UCP2 and COX was performed using the CCD camera of the GeneGnome analyzer, and quantification was achieved using the Genetool software (Ozyme, Saint-Quentin en Yvelines, France).
In Situ Hybridization Analysis-The fixative of collected tissues was washed out of the samples, which were dehydrated in graded concentrations of ethanol. The ethanol was then replaced by toluene, and the samples were embedded in paraffin wax. 5-m sections were prepared and mounted on glass slides. The protocol used was largely adapted from the technique described by Simmons et al. (25) and used previously, comprising the labeling of the antisense 35 S-labeled cRNA probes (26). After being developed, tissues were rehydrated and stained with Harri's hematoxylin and eosin solutions for histological examination.
Statistical Analysis-Values were expressed as mean Ϯ S.E. Comparisons among the different groups were assessed by one-way analysis of variance. The level of significance (p value; shown in Fig. 4) were as follows: *, p Ͻ 0.05; **, p Ͻ 0.01; and ***, p Ͻ 0.001.

UCP1 Is Not Expressed in the Uterus of Mice-Mitochondria
were isolated from tissues of mice, and a Western blot analysis was carried out as described previously (14). The 11A antibody used previously by Nibbelink et al. (1) was used to detect UCP1. UCP1 was detected in 0.1 g of BAT mitochondria protein isolated from rat or C57BL/6J (Fig. 1A, lanes 1 and 2). For this amount of protein, no signal was obtained in BAT mitochondrial protein obtained from the Ucp1(Ϫ/Ϫ) mice (lane 3), undoubtedly indicating that the signal detected using the 11A antibody was UCP1. However, when a large amount (30 g) of BAT mitochondrial protein from Ucp1(Ϫ/Ϫ) mice was loaded onto the gel, a single and strong band corresponding to a protein with a smaller molecular weight was observed. This result indicates that the 11A antibody also cross-reacted with a BAT mitochondrial protein other than UCP1 when a large amount of protein was used. The 11A antibody was able to detect the same signal in uterus mitochondria from either C57BL/6J or OF1 mice (Fig. 1B, lanes 1 and 4). This signal was clearly not UCP1, because it was present in uterus mitochon- dria isolated from Ucp1(Ϫ/Ϫ) mice (Fig. 1B, lane 2). We tested the possibility that the UCP1-antibody could detect UCP2 in uterus, but Fig. 1B, lane 3 shows that the same signal was detected in mitochondria extracted from Ucp2(Ϫ/Ϫ) mice. At this stage, we concluded that a mitochondrial protein, detected by the 11A antibody, was present at low levels in a large amount of uterus mitochondria, and it was neither UCP1 nor UCP2 (Fig. 1A). Because Nibbelink et al. also reported the detection of UCP1 in smooth muscle of some tissues comprising gut and stomach of mice (1), mitochondria were isolated from several organs of Ucp1(Ϫ/Ϫ) mice. Actually, the 11A antibody detected a mitochondrial protein in many organs of C57BL/6J mice comprising the stomach (Fig. 1C), but, again, this protein was distinct from UCP1, because a similar signal was detected in mitochondria prepared from the testis, ovary, kidney, spleen, brain, stomach, and heart. No signal was recorded in liver, lung, or gut mitochondria.
Expression of the UCP2 in the Genital Tract of Female Mice-Tissues from the reproductive tract of wild type or Ucp2(Ϫ/Ϫ) female mice were dissected. Tissues were also collected from B6D2F1 mice. Mitochondria were isolated and analyzed for the expression of UCP2 in Western blot. Spleen mitochondria were also prepared, because this organ has the highest UCP2 content (14). UCP2 was detected in the ovary, uterus, and oviduct of Ucp2(ϩ/ϩ) mice ( Fig. 2A). As expected, UCP2 protein was undetectable in mitochondria from the tissues of Ucp2-deficient mice ( Fig. 2A). In mammary gland mitochondria, the picture was more complex, because an abundant and unknown protein other than UCP2 was also observed. The UCP2 amount was estimated by using the COX as an indicator of the mitochondria enrichment in two strains of mice. The level of UCP2 in the ovary, uterus, and lung was estimated to be ϳ20% of the level in the spleen in both strains of mice (Fig. 2B). In the stomach, the level of UCP2 was 20 or 60% of the level in the spleen for C57Bl/6J and B6D2/F1 mice, respectively.
Analysis of UCP2 Expression in Cells of Ovary and Uterus-To identify cells expressing Ucp2 mRNA in ovary and uterus of mice, in situ hybridization analysis was performed. In the ovary, the Ucp2 mRNA was strongly expressed in granulosa cells of primordial and primary ovarian follicles (Fig. 3, A and  B). In contrast, the hybridization signal was weak in theca cells immediately surrounding the follicle (Fig. 3, B and C). At the level of the early antral follicle (Fig. 3C), the marked expression of Ucp2 mRNA was not only found in healthy granulosa cells but also in some granulosa cells that showed necrobiotic changes, as it occurs during a degenerative process known as atresia (arrow, in Fig. 3C). In the uterus, the hybridization signal was uniformly detected at a strong level in the endometrium, including glandular epithelium cells as well as uterine glands (Fig. 3, D-F). In contrast, the hybridization signal was low in connective tissue (stroma) and completely absent in myometrium (data not shown). In the oviduct, the hybridization signal was present through the length of the mucosa epithelial cells (data not shown). All sections are bright field photomicrographs. Panels D-F are uterine cross-sections that were treated as described for panels A-C. The hybridization signal (silver grains) was strong at the level of the endometrium (En), including glandular epithelium cells (EpC) and the uterine glands (Gl). L, lumen. G, 30 g of mitochondrial protein were analyzed in Western blot as described in the Fig. 2 legend. Ucp2(ϩ/ϩ) mice were irradiated and transplanted with bone marrow from Ucp2(Ϫ/Ϫ) mice, whereas Ucp2(Ϫ/Ϫ) mice were transplanted with bone marrow from Ucp2(ϩ/ϩ) mice. The Ucp2(ϩ/ϩ) mice that received bone marrow from Ucp2(Ϫ/Ϫ) mice were named Ucp2(ϩ/ϩ)/t(Ϫ/Ϫ), and the Ucp2(Ϫ/Ϫ) mice that received bone marrow from Ucp2(ϩ/ϩ) were named Ucp2(Ϫ/Ϫ)/t(ϩ/ϩ).

FIG. 2. Western blot analysis of mitochondria isolated from the genital tract or other tissues of mice using
anti-UCP2 antibody. 30 g of mitochondrial protein were used. The blots were incubated with the anti-UCP2 antibody described by Pecqueur et al. (14) and the antibody against COX (see "Experimental Procedures"). A, mitochondria were isolated from tissues of Ucp2(ϩ/ϩ) or Ucp2(Ϫ/Ϫ) mice as indicated. mam gl, mammary gland. B, analysis of UCP2 in several tissues of two strains of mice.
UCP2 is present in immune cells that are disseminated over many organs (12,17,18). In the absence of antibodies allowing immunohistochemical analysis of UCP2, it was decided to perform transplantation of bone marrow cells in irradiated animals, as was recently carried out to document the protective role of UCP2 in macrophages against atherosclerosis (22). Ucp2(Ϫ/Ϫ) mice were irradiated and transplanted with bone marrow from Ucp2(ϩ/ϩ) mice (and henceforth called Ucp2-(Ϫ/Ϫ)/t(ϩ/ϩ) mice), whereas Ucp2(ϩ/ϩ) mice were irradiated and transplanted with bone marrow from Ucp2(Ϫ/Ϫ) mice (and henceforth called Ucp2(ϩ/ϩ)/t(Ϫ/Ϫ) mice). Transplanted mice were allowed to recover for 4 months to ensure that the engraftment of resident macrophages was close to completion. Then, the mice were sacrificed, and expression of UCP2 in ovary and uterus was analyzed by Western blot analysis (Fig.  3G). In the ovary, expression of UCP2 was still present in Ucp2(ϩ/ϩ)/t(Ϫ/Ϫ) mice, whereas it completely disappeared in the Ucp2(Ϫ/Ϫ)/t(ϩ/ϩ) mice. This experiment demonstrated that UCP2 expression was restricted exclusively to the ovary cells. In contrast, UCP2 protein was present in the uterus of Ucp2(ϩ/ϩ)/t(Ϫ/Ϫ) and Ucp2(Ϫ/Ϫ)/t(ϩ/ϩ) mice (Fig. 3G), suggesting that it was present both in immune and endometrium cells in this organ.
Analysis of UCP2 Variation during Estrus Cycle and Gestation of Mice-Mature C57BL/6J female mice were synchronized by an intraperitoneal injection with 5 IU of PMSG to stimulate follicular growth. They were injected 48 h later with 5 IU of hCG to trigger ovulation and luteinization. Western blot analysis of the ovary showed variations in UCP2 expression along the estrus cycle. The beginning of the pro-estrus was characterized by the highest level of UCP2. The UCP2 level transiently decreased during follicular growth prior to induction before ovulation (Fig. 4A). In the uterus, a similar pattern of expression was observed (Fig. 4A). To assay UCP2 in the uterus during gestation, females were sacrificed at various stages of pregnancy or 1 day post partum. As shown in Fig. 4B, the UCP2 level in the uterus increased during the first part of the gestation and reached its maximal level at day 12. In the ovary, no significant variation occurred during gestation (data not shown). DISCUSSION A number of historical and more recent studies established the unique respiration uncoupling in brown adipocytes mitochondria and its direct relationship to the mitochondrial UCP1. Biochemical, immunological, transcriptional, and physiological studies demonstrated the specificity of expression of UCP1 in brown adipocytes (2)(3)(4)27). Recently and surprisingly, Nibbelink et al. reported that UCP1 was present in the longitudinal smooth muscle of uterus (1). In the present study, using Ucp1(Ϫ/Ϫ) mice tissues as control animals, we show that the antibody used by Nibbelink et al. (1) certainly recognizes UCP1 in BAT but also interacts with another protein having a slightly lower molecular weight in uterus mitochondria as well as in many other organs. Similar data were obtained in C57BL/6J and OF1 mice, this latter strain being the strain used by Nibbelink et al. (1). Using Ucp2(Ϫ/Ϫ) mice, we also demonstrated that the protein identified by Nibbelink et al. was not UCP2. Therefore, we confirm the specific expression of UCP1 in BAT. UCPs are members of the mitochondrial carrier protein family (12, 28 -31) and they share amino acid motifs with many mitochondrial carriers, including the ADP/ATP translocator (32). The difficulty in obtaining specific and sensitive antibodies is probably due to the possible cross reactivity with these different proteins (14,(33)(34)(35). The previously reported ectopic expression of UCP1 in skeletal muscle of mice chronically treated by a ␤3-adrenoreceptor agonist was never confirmed (36).
Following the data of Nibbelink (1), Shabalina et al. reported a difference in intestinal relaxation between Ucp1(Ϫ/Ϫ) mice and wild type mice and also proposed a role for UCP1 in uterus muscle relaxation (23). However, these investigators did not attempt to independently confirm the presence or absence of UCP1 in the intestine despite that fact that they had access to Ucp1(Ϫ/Ϫ) mice. Their conclusions on a role for UCP1 in smooth muscle do not agree with our findings that UCP1 is not present in the uterus or in several other smooth muscle preparations that we have analyzed. The importance of the genetic background for the study of a protein function and, in particular, of UCP1, was supported by recent studies (37). It seems FIG. 4. Expression of UCP2 during the estrus cycle and gestation. For panels A and B, 30 g of mitochondrial protein were analyzed in Western blot as described in the Fig. 2 legend. The signals were quantified, and the data represent the ratio between UCP2 and COX. A, female C57BL/6J mice were synchronized by a hormonal treatment and sacrificed 24 or 48 h after PMSG injection. Mice were also injected with hCG 48 h after PMSG administration and sacrificed at 51, 54, 56, 58, 60, 62, 64, or 72 h (at least three mice per point). B, gestation of female C57BL/6J mice was identified by the presence of a vaginal plug, and mice were killed at different stages of gestation (three mice per point).
that Shabalina et al. compared Ucp1(Ϫ/Ϫ) mice raised on a mixed genetic background to three strains of wild type mice (23). Therefore, their data may simply reflect differences between genetic backgrounds. The present study provides clear evidence that UCP2 is expressed in the female reproductive system, both in the ovary and the uterus. An increase of the amount of UCP2 was observed in the uterus at day 12 of gestation and correlates with previous mRNA data (38). However, if the mRNA level remains high until the end of the gestation (38), the UCP2 protein appears to decline during the last third of gestation. These data underline the complexity of the UCP2 regulation due to transcriptional regulation as well as strong translational regulation (14).
Immune cells contain UCP2 (12,17,21,39), and such cells are resident in the genital tract. In the ovary, the number and types of leukocytes change during the estrus cycle, especially the number of macrophages and neutrophils, which increase around ovulation (40). Results of in situ hybridization analysis showed that the cells principally expressing the UCP2 mRNA in the ovary are granulosa cells, whereas the signal was particularly low in the theca layer, wherein the resident white blood cells are localized. However, results from Ucp2(ϩ/ϩ) mice transplanted with bone marrow of the Ucp2(Ϫ/Ϫ) led to the conclusion that the UCP2 signal of the ovary was restricted to ovarian cells. In the endometrium, macrophages represent an important mechanism of defense, including the degradation of cellular debris in endometrial shedding and repair and protection against infections. However, the mechanisms involved in recruiting, maintaining, and activating uterine macrophages are not fully defined (41)(42)(43). In situ hybridization has shown that the expression of UCP2 in the uterus is restricted to the endometrium, especially in epithelial cells and uterine glands. The bone marrow transplantation experiment suggest that immune cells contribute to UCP2 expression in the uterus.
The organs forming the reproductive tract are the site of continuous remodeling. In the ovary, follicular growth, atresia, ovulation, and formation and regression of the corpus luteum occur during cycle. Many changes also occur in the uterus, such as proliferation of the endometrium, possible embryo implantation, and endometrial shedding. The analysis of the expression profile of UCP2 during the estrus cycle showed that UCP2 expression was very low during follicular growth before increasing in the hours preceding ovulation. Based upon these results an important question yet to be answered relates to UCP2 function in the female genital tract. Several studies implicated UCP2 in decreasing ROS production (18,19,21). Whether UCP2 decrease during follicular growth increases ROS in the ovary or whether the UCP2 level decreases in response to a production of ROS was not investigated. In the ovary, the pre-ovulatory period is characterized by rapid and dramatic changes in the follicular structure comparable with an inflammatory process (44,45). It may be hypothesized that UCP2 increase during this period lowers ROS production and attenuates the inflammatory process. Because we observed a lower rate of reproduction of the Ucp2(Ϫ/Ϫ) compared with the Ucp2(ϩ/ϩ) mice, it may be proposed that the absence of UCP2 causes an excessive inflammation and maintains ROS at a high level that inhibits ovulation.