Single closed-tube quantitative real-time PCR assay with dual-labelled probes for improved sex determination of equine embryos.

In addition to fulfilling many breeders' curiosity, equine embryonic sex determination can have a profound commercial impact. However, the application of currently described assays for equine embryonic sexing has rendered variable diagnosis and validation rates, with sensitivity being the main problem. In addition, while pregnancy results of in vivo-flushed equine embryos following a needle aspiration biopsy equal those of non-biopsied embryos, the effect on in vitro-produced embryos is unknown. Here, we aimed to develop a highly sensitive and specific assay for equine sex determination that can be directly performed on few embryonic cells, and to test the effect of a needle aspiration biopsy on the viability of the in vitro-produced embryo. To this end, a multiplex quantitative real-time PCR (qPCR) assay with dual-labelled probes was designed to allow the simultaneous generation of both male-specific and control fragments in a single closed-tube reaction, avoiding potential sample loss or contamination. To improve sensitivity, multicopy and polymeric genes were chosen to be specifically amplified, i.e., eight copies of Y-chromosomal ETSTY5 as male-specific and four autosomal UBC monomers as control fragment. Specificity was enhanced by the equine-specific character of ETSTY5 and by using dual-labelled probes. The assay was optimised with equine male and female genomic DNA and demonstrated a 100% accuracy and a >95% qPCR efficiency down to 10 pg of DNA. The assay was subsequently applied to determine the sex of 44 in vitro-produced embryos, collecting trophectoderm biopsies by means of a needle aspiration biopsy and herniating cells. Of all trophectoderm biopsies and herniating cell samples (n = 54), 87% could be diagnosed. Assay results were validated on a second sample obtained from the biopsied embryo (n = 18) or, by ultrasound-based sex determination of the foetus (n = 7) following the transfer of the biopsied embryo to a recipient mare, with about half of the embryos being fillies and colts. The needle aspiration biopsy procedure did not impair initial pregnancy rate or early pregnancy losses as compared to non-biopsied embryos. In conclusion, we report a safe, reliable, fast, and cost-effective assay for equine sex determination which was validated for the sex determination of in vitro-produced embryos based on few embryonic cells, and needle aspiration biopsy did not impair the embryo's viability. The assay and safe biopsy strategy hold potential for other applications.

Validation of multiparametric panels for bovine mesenchymal stromal cell phenotyping.

Bovine mesenchymal stromal cells (MSCs) display important features that render them valuable for cell therapy and tissue engineering strategies, such as self-renewal, multi-lineage differentiation, as well as immunomodulatory properties. These cells are also promising candidates to produce cultured meat. For all these applications, it is imperative to unequivocally identify this cell population. The isolation and in vitro tri-lineage differentiation of bovine MSCs is already described, but data on their immunophenotypic characterization is not yet complete. The currently limited availability of monoclonal antibodies (mAbs) specific for bovine MSC markers strongly hampers this research. Following the minimal criteria defined for human MSCs, bovine MSCs should express CD73, CD90, and CD105 and lack expression of CD14 or CD11b, CD34, CD45, CD79α, or CD19, and MHC-II. Additional surface proteins which have been reported to be expressed include CD29, CD44, and CD106. In this study, we aimed to immunophenotype bovine adipose tissue (AT)-derived MSCs using multi-color flow cytometry. To this end, 13 commercial Abs were screened for recognizing bovine epitopes using the appropriate positive controls. Using flow cytometry and immunofluorescence microscopy, cross-reactivity was confirmed for CD34, CD73, CD79α, and CD90. Unfortunately, none of the evaluated CD105 and CD106 Abs cross-reacted with bovine cells. Subsequently, AT-derived bovine MSCs were characterized using multi-color flow cytometry based on their expression of nine markers. Bovine MSCs clearly expressed CD29 and CD44, and lacked expression of CD14, CD45, CD73, CD79α, and MHCII, while a variable expression was observed for CD34 and CD90. In addition, the mRNA transcription level of different markers was analyzed using reverse transcription quantitative polymerase chain reaction. Using these panels, bovine MSCs can be properly immunophenotyped which allows a better characterization of this heterogenous cell population.

Effect of overconditioning on the hepatic global gene expression pattern of dairy cows at the end of pregnancy.

Overconditioning is a risk factor for upregulated pre- and postpartum fat mobilization. Therefore, we hypothesized that overconditioning at the end of pregnancy leads to the accumulation of lipids in the liver and modifications of the hepatic gene expression pattern. The aim of this study was to evaluate the effect of normal- versus overconditioning on the hepatic transcriptomic profile of dairy cows at the end of pregnancy. Ten dry multiparous Holstein cows were killed 2 wk before expected calving. Body condition score (BCS) and backfat thickness (BFT) were evaluated, and blood samples for nonesterified fatty acids (NEFA) were taken before cows were killed. After cows were killed, liver biopsy samples were collected for further assessment of total lipids and RNA sequencing. Five cows were classified as normal-conditioned (median BCS = 3, range 2.75-3.5) and 5 as overconditioned (median BCS = 4, range 4-5). Regression models confirmed that normal-conditioned cows had lower BFT (1.29 ± 0.29 cm; least squares means ± standard error) and serum NEFA (0.16 ± 0.04 mmol/L) in comparison to overconditioned cows (3.14 ± 0.43 cm and 0.38 ± 0.07 mmol/L for BFT and NEFA, respectively). Total liver lipid percentage tended to be lower in normal- versus overconditioned cows (4.63 ± 0.40% and 6.06 ± 0.44%, respectively). In comparison to the mean liver lipid percentage of the normal- and overconditioned cows, 1 overconditioned cow had a relatively low (5.21%) and 1 normal-conditioned cow had a relatively high (6.07%) liver lipid percentage. Differentially expressed genes analysis (edgeR quasi-likelihood method) showed that normal-conditioned cows presented 11 upregulated and 12 downregulated genes in comparison to overconditioned cows. Linear discriminant analysis effects size revealed 133 differentially expressed genes between normal- versus overconditioned cows. Notably, the liver of normal-conditioned cows had upregulated genes associated with liver functionality (ALB, SELENOP, IGF1, and IGF2). On the other hand, overconditioned cows had upregulated genes associated with the acute-phase response (C3, HPX, and, LBP). High basal lipolysis in overconditioned cows at the end of pregnancy increased liver lipid content, and this may alter the hepatic gene expression pattern to a pro-inflammatory state.

Distinct neutrophil C5a receptor inflammatory events in cows initiated by chemoattractant C5a and lipopolysaccharide around parturition and in mid lactation.

In neutrophils, toll-like receptor and complement component 5a (C5a) signaling are critical pathways regulating innate immunity. In cows, not much is known about the second C5a receptor, complement component 5a receptor 2 (C5AR2). It is an interesting player in sepsis treatment because it is considered to have an anti-inflammatory effect during normal inflammation. Periparturient cows are prone to severe infections, and the objectives of this study were to investigate the expression and functionality of C5AR2 during peripartum. We investigated the effect of 2 major inflammatory stimuli, C5a and lipopolysaccharide (LPS), on the expression of a selected number of genes (C5AR1, C5AR2, TLR4, ITGAM, COX2, and CXCL8) and functions linked to these receptors. Overall, TLR4, ITGAM, and C5AR2, all of which are involved in early inflammation, showed a lower expression in periparturient cows. However, an overall lower expression seems not to be the only explanation for the increased risk of sepsis in periparturient cows. Normally, in response to inflammation and as seen in the mid-lactation group, the expression of these genes increases after stimulation with LPS. However, in periparturient cows, stimulation with LPS led to a decrease in expression of these receptors, indicating a different response of neutrophils in response to LPS during this period. A decrease in ITGAM (coding for CD11b) expression complicates correct neutrophil localization and phagocytosis. Its downregulation upon stimulation might be detrimental for adequate eradication of the pathogen and might increase the risk of an imbalanced inflammation; C5AR2 seems to play a central role in this altered response. In addition, myeloperoxidase (MPO) activity in periparturient cows is lower in response to C5a stimulation. It has been suggested that MPO plays an important role in neutrophil shutdown and, thereby, timely resolution of inflammation. A decreased MPO activity might thus prolong the inflammatory reaction of the neutrophils. This finding was supported by the increased viability of the neutrophils obtained from periparturient cows. Even after stimulation, we found a lower caspase-3 activity in this group, indicating that they might be activated for a longer time compared with the neutrophils from mid-lactation cows. Accordingly, these alterations might contribute to a temporal mismatch in inflammatory responses, as often seen in severe periparturient infections.

Influence of adipocyte size and adipose depot on the number of adipose tissue macrophages and the expression of adipokines in dairy cows at the end of pregnancy.

The aim of this study was to determine the number of adipose tissue macrophages (ATM) and the mRNA expression of adipokines [adiponectin (ADIPOQ), leptin (LEP), interleukin 6 (IL6), tumor necrosis factor (TNF), and interleukin 10 (IL10)] in different adipose depots from cows with a variable body condition score (BCS) at the end of the dry period. We hypothesized that the number of ATM and the expression of these adipokines depend on adipocyte size and the anatomical location of the adipose depot. Subcutaneous, omental, mesenteric, perirenal, and intrapelvic adipose tissue samples were taken immediately after euthanasia of 10 Holstein Friesian dairy cows (upcoming parity 2 to 5, age 3.9 ± 1.4 yr; mean ± standard deviation) at the end of pregnancy (actual days of pregnancy at the moment of euthanasia: 269 ± 5 d). During the dry period, all animals received similar diets to meet but not exceed requirements. Five animals were considered to have a normal BCS (2.5-3.5) and 5 animals were considered to be over-conditioned (BCS = 3.75-5). Body weight of the animals at the moment of euthanasia was 717 ± 77 kg. Expression of the different genes was determined by reverse transcription quantitative real-time PCR. Adipocyte size was determined by measuring the area of 100 adipocytes on histological sections. Average adipocyte area was 10,475 ± 1,019, 8,500 ± 780, 10,383 ± 1,227, 11,466 ± 1,039, and 11,087 ± 1,632 µm2 for the subcutaneous, mesenteric, omental, intrapelvic, and perirenal adipose depot, respectively. Immunohistochemistry using anti-bovine CD172a antibodies was performed to determine the proportion of ATM (the number of CD172a-positive cells per 100 adipocytes, given as a percentage). Expression of LEP, IL6, and TNF was positively associated with adipocyte size, whereas no association could be detected between ADIPOQ and IL10 with the size of the adipocytes. The omental adipose depot was especially infiltrated with ATM (1.92 ± 0.55, 1.10 ± 0.33, and 8.28 ± 2.24% for the subcutaneous, mesenteric, and omental adipose depot, respectively). The proportion of ATM was positively associated with the size of the adipocytes in the omental and mesenteric adipose depot. Expression of ADIPOQ, LEP, IL6, TNF, and IL10 differed among depots, which suggests differences in inflammatory characteristics depending on the anatomical location of depots. In conclusion, the results of the present study confirm the adipose tissue as a potential source of inflammatory mediators and demonstrate ATM infiltration, especially in the omental adipose depot.

F4-related mutation and expression analysis of the aminopeptidase N gene in pigs.

Intestinal infections with F4 enterotoxigenic Escherichia coli (ETEC) are worldwide an important cause of diarrhea in neonatal and recently weaned pigs. Adherence of F4 ETEC to the small intestine by binding to specific receptors is mediated by F4 fimbriae. Porcine aminopeptidase N (ANPEP) was recently identified as a new F4 receptor. In this study, 7 coding mutations and 1 mutation in the 3' untranslated region (3' UTR)were identified in ANPEP by reverse transcriptase (RT-) PCR and sequencing using 3 F4 receptor-positive (F4R+) and 2 F4 receptor-negative (F4R-) pigs, which were F4 phenotyped based on the MUC4 TaqMan, oral immunization, and the in vitro villous adhesion assay. Three potential differential mutations (g.2615C > T, g.8214A > G, and g.16875C > G) identified by comparative analysis between the 3 F4R+ and 2 F4R- pigs were genotyped in 41 additional F4 phenotyped pigs. However, none of these 3 mutations could be associated with F4 ETEC susceptibility. In addition, the RT-PCR experiments did not reveal any differential expression or alternative splicing in the small intestine of F4R+ and F4R- pigs. In conclusion, we hypothesize that the difference in F4 binding to ANPEP is due to modifications in its carbohydrate moieties.

Differentiation of F4 receptor profiles in pigs based on their mucin 4 polymorphism, responsiveness to oral F4 immunization and in vitro binding of F4 to villi.

F4(+) enterotoxigenic Escherichia coli (F4(+) ETEC) are an important cause of diarrhoea and mortality in piglets. F4(+) ETEC use their F4 fimbriae to adhere to specific receptors (F4Rs) on small intestinal brush borders, resulting in colonization of the small intestine. To prevent pigs from post-weaning diarrhoea, pigs should be vaccinated during the suckling period. Previously, we demonstrated that F4acR(+), but not F4acR(-) piglets could be orally immunized with purified F4 fimbriae resulting in a protective immunity against F4(+) ETEC infections, indicating that this immune response was F4R dependent. Recently, aminopeptidase N has been identified as a glycoprotein receptor important for this oral immune response. However, in some oral immunization experiments, a few F4acR(+) piglets did not show an antibody response upon oral immunization, suggesting additional receptors. Therefore, the binding profile of F4 to brush border membrane (glyco)proteins was determined for pigs differing in F4-specific antibody response upon oral immunization, in in vitro adhesion of F4(+)E. coli to small intestinal villi, and in Muc4 genotype. Six groups of pigs could be identified. Only two groups positive in all three assays showed two high molecular weight (MW) glycoprotein bands (>250kDa) suggesting that these high MW bands are linked to the MUC4 susceptible genotype. The fact that these bands were absent in the MUC4 resistant group which showed a positive immune response against F4 and was positive in the adhesion test confirm that at least one or perhaps more other F4Rs exist. Interestingly, two pigs that were positive in the villous adhesion assay did not show an immune response against F4 fimbriae. This suggests that a third receptor category might exist which allows the bacteria to adhere but does not allow effective immunization with soluble F4 fimbriae. Future research will be necessary to confirm or reveal the identity of these receptors.

Anaphylatoxin C5a-induced toll-like receptor 4 signaling in bovine neutrophils.

It is well known that signaling in neutrophils through both the complement component 5a (C5a) and C5a receptor (C5aR) and the toll-like receptor 4 (TLR4) pathways plays an essential role in innate defense. Neutrophil dysfunction, as seen during sepsis in severe mastitis during the periparturient period, is correlated with elevated concentrations of anaphylatoxin C5a. The aim of the current study was to elucidate the effect of C5a on TLR4 signaling in bovine neutrophils. Neutrophils were incubated with a high (but physiological) dose of purified C5a, and mRNA was extracted from neutrophils at different time points postincubation (PI). The incubation with C5a resulted in a biphasic C5aR expression profile, a phenomenon that might be explained by internalization (at 10 min PI) with subsequent reconstitution (starting at 40 min PI) of this receptor. The expression of TLR4, as well as its coreceptor, CD14, showed a similar biphasic change as observed with C5aR. In addition, changes in the mRNA expression levels of several genes belonging to the TLR4 pathway, such as TICAM-1, IKKα, and MAP3K7 were noted. The maximal expression of TLR4, CD14, and C5aR mRNA at 80 min PI was accompanied by a peak in IL8 mRNA, indicating that C5a is able to induce IL-8 production in neutrophils in vitro without the need of a costimulatory factor such as lipopolysaccharide. Moreover, a relatively constant expression of RELA was accompanied by increased expression of ATF3, an endogenous inhibitor of nuclear factor-κB mediated transcription, implying that C5a regulates TLR4 signaling and IL-8 synthesis independently. A significant time-dependent correlation was found between C5aR and TLR4, with the majority of the selected TLR4-dependent genes showing a significant correlation with C5aR at 80 min PI, when C5aR and TLR4 mRNA expression reached its maximum, suggesting crosstalk between both receptors. Taken together, this study showed that C5a is able to (1) alter the expression of genes belonging to the TLR4 pathway and (2) induce IL8 gene expression in bovine neutrophils. In addition, indications for cross-talk between complement activation and TLR4 signaling were found in the present study.

Distribution of the Shadoo protein in the ovine brain assessed by immunohistochemistry.

Shadow of prion protein is a gene potentially involved in the pathogenesis of prion diseases. However, the Shadoo protein encoded by this gene has not yet been studied in sheep, an important species in prion matters. Therefore, we developed a polyclonal antibody against ovine Shadoo and assessed the presence and distribution of this protein in the ovine brain by immunohistochemistry. The strongest staining level was found in the cerebellum (especially in the Purkinje cells) and in the pons, but cerebrum, hippocampus, pituitary gland, medulla oblongata, thalamus and hypothalamus were also immunopositive. Remarkably, a typical granular pattern was seen in most of the tested brain tissues, which might indicate that Shadoo is primarily expressed at synapses. The results of this study and the availability of an ovine anti-Shadoo antibody can contribute to future research on the function of Shadoo and on its potential involvement in prion diseases.

Identification of polymorphisms in the ovine Shadow of prion protein (SPRN) gene and assessment of their effect on promoter activity and susceptibility for classical scrapie.

Shadow of prion protein (SPRN) is an interesting candidate gene thought to be involved in prion pathogenesis. In humans, an association has already been discovered between mutations in SPRN and the incidence of variant and sporadic Creutzfeldt-Jakob disease. However, in sheep, the effect of mutations in SPRN is largely unknown. Therefore, we analysed the presence of mutations in the entire ovine SPRN gene, their association with scrapie susceptibility and their effect on SPRN promoter activity. In total, 26 mutations were found: seven in the promoter region, four in intron 1, seven in the coding sequence and eight in the 3' untranslated region. The mutations detected in the coding sequence and the promoter region were subsequently analysed in more detail. In the coding sequence, a polymorphism causing a deletion of two alanines was found to be associated with susceptibility for classical scrapie in sheep. Furthermore, a functional analysis of deletion constructs of the ovine SPRN promoter revealed that the region 464 to 230 bp upstream of exon 1 (containing a putative AP-2 and putative Sp1 binding sites) is of functional importance for SPRN transcription. Six mutations in the SPRN promoter were also found to alter the promoter activity in vitro. However, no association between any of these promoter mutations and susceptibility for classical scrapie was found.

High prevalence of bovine papillomaviral DNA in the normal skin of equine sarcoid-affected and healthy horses.

Bovine papillomavirus (BPV), the causative agent of papillomas in cattle, has been shown to play a major role in the pathogenesis of equine sarcoids in horses. BPV has also been detected occasionally in normal equine skin. In this study, presence and activity of BPV in normal skin and peripheral blood of 4 groups of horses were evaluated: sarcoid-affected horses, horses living in contact with sarcoid-affected horses, horses living in contact with papilloma-affected cattle and control horses. From each horse, 3 samples on 4 locations were collected: a swab of the intact skin surface and both a swab and a biopsy after decontamination. BPV DNA was found in the normal skin of 24 of 42 horses (57%). Mainly sarcoid-affected horses and horses living in contact with cattle were carriers (73%), but BPV DNA was also detected in 50% of the horses living in contact with sarcoid-affected horses and in 30% of the control population. BPV mRNA was detected in 58% of the samples positive for BPV DNA, although in a much lower quantity compared to sarcoids. In most of the BPV DNA positive samples mild acanthosis, slight basophilic cytoplasmic swelling of the epidermal layers and/or thickening of the basal membrane were noticed, but these observations were also present in several BPV DNA negative normal skin samples. BPV DNA could not be detected in peripheral blood. These findings suggest latent infection and a wide-spread occurrence of BPV in the horse population.

Bovine papillomavirus load and mRNA expression, cell proliferation and p53 expression in four clinical types of equine sarcoid.

Equine sarcoids, the most common skin tumours in horses, are induced by bovine papillomavirus (BPV). Their clinical appearance varies from small stable patches to aggressively growing masses. Differences in BPV load and mRNA expression and Ki67 and p53 immunostaining among four clinical types (fibroblastic, occult, nodular and verrucous sarcoids) were evaluated to test the hypothesis that the clinical behaviour of equine sarcoids correlates with BPV activity. Viral load and expression of the BPV E2, E5, E6 and E7 genes were determined using quantitative real-time PCR. The proliferative fraction (PF) of the tumours was determined by Ki67 immunostaining and expression of p53 was analysed by immunohistochemistry. Nodular sarcoids showed a significantly higher viral load than the other types. A significant overall difference among the four types was observed for E2, E5, E6 and E7 mRNA expression. Nodular sarcoids showed the highest expression level for each BPV gene examined, followed by verrucous, fibroblastic and occult tumours. Viral DNA and mRNA outcomes correlated with each other, indicating a similar transcription pattern in each type of sarcoid. The PF was significantly higher in the superficial layers of verrucous and fibroblastic sarcoids compared with occult and nodular types. No significant difference was observed for the PF in the deep layers and for p53 expression. These results clearly demonstrate the omnipresence and active transcription of BPV in equine sarcoids. However, the hypothesis that the clinical behaviour of an equine sarcoid can be explained on the basis of differences in BPV activity could not be demonstrated.

Porcine OGN and ASPN: mapping, polymorphisms and use for quantitative trait loci identification for growth and carcass traits in a Meishan x Piétrain intercross.

The porcine orthologues of human chromosome HSA9q22.31 genes osteoglycin (OGN) and asporin (ASPN) were mapped to porcine chromosome SSC3 using linkage analysis and a somatic cell hybrid panel. This mapping was refined to SSC3q11 using fluorescence in situ hybridization. These results confirm the existence of a small conserved synteny group between SSC3 and HSA9. Polymorphisms were revealed in both genes, including a pentanucleotide microsatellite (SCZ003) in OGN and two single nucleotide polymorphisms (AM181682.1:g.780G>T and AM181682.1:g.825T>C) in ASPN. The two genes were included in a set of markers for quantitative trait loci (QTL) mapping on SSC3 in the Hohenheim Meishan x Piétrain F2 family. Major QTL for growth and carcass traits were centred in the ASPN-SW902 region.

SNP identification, linkage and radiation hybrid mapping of the porcine lamin A/C (LMNA) gene to chromosome 4q.

The lamins are components of nuclear lamina and they have a profound influence on nuclear structure and functions. They are encoded by three genes, LMNA, LMNB1 and LMNB2. A genomic fragment of the porcine LMNA gene (822 bp; from exons 7 to 9) was amplified by polymerase chain reaction and comparatively sequenced. Four single nucleotide polymorphisms (SNPs) were identified in intronic sequences: G162A, G208A, T367G and C618T. The SNPs are within the restriction sites for enzymes Bsh1236I, HpaII, AluI and Bsh1236I respectively. Allele frequencies at SNPs G208A, T367G and C618T were determined by using eight pig breeds. Linkage analysis in the Hohenheim Meishan x Piétrain family placed the LMNA gene in the chromosome 4q linkage group, between MEF2D and GBA (MEF2D - 3.0 cM - LMNA - 0.2 cM - GBA). In radiation hybrid mapping LMNA was most significantly linked to SW270 on chromosome 4 (39 cR; LOD = 7.86). The LMNA gene is located in the quantitative trait loci region for some carcass traits on chromosome 4q.

Porcine PPARGC1A (peroxisome proliferative activated receptor gamma coactivator 1A): coding sequence, genomic organization, polymorphisms and mapping.

We report here the characterisation of porcine PPARGC1A. Primers based on human PPARGC1A were used to isolate two porcine BAC clones. Porcine coding sequences of PPARGC1A were sequenced together with the splice site regions and the 5' and 3' regions. Using direct sequencing nine SNPs were found. Allele frequencies were determined in unrelated animals of five different pig breeds. In the MARC Meishan-White Composite resource population, the polymorphism in exon 9 was significantly associated with leaf fat weight. PPARGC1A has been mapped by FISH to SSC8p21. A (CA)n microsatellite (SGU0001) has been localised near marker SWR1101 on chromosome 8 by RH mapping and at the same position as marker KS195 (32.5 cM) by linkage mapping. The AseI (nt857, Asn/Asn489) polymorphism in exon 8 was used to perform linkage analysis in the Hohenheim pedigrees and located the gene in the same genomic region. Transcription of the gene was detected in adipose, muscle, kidney, liver, brain, heart and adrenal gland tissues, which is in agreement with the function of PPARGC1A in adaptive thermogenesis.

Cell-specific localization of oestrogen receptor beta (ESR2) mRNA within various bovine ovarian cell types using in situ hybridization.

The localization of oestrogen receptor beta (ESR2) mRNA, in this article denominated as (ERbeta) mRNA, was examined using in situ hybridization in the ovaries of randomly selected cows, irrespective of the cycle stage of the animals. A 602-bp fragment of ERbeta mRNA was cloned, sequenced and digoxigenin (DIG)-labelled. Semi-quantitative evaluation showed that the scores for ERbeta mRNA were moderate to high in the follicle cells of both primordial and primary follicles, but lower in granulosa cells of secondary follicles. In vital tertiary follicles, the total ERbeta mRNA expression was low but varied between the different animals. In both obliterative and cystic atretic follicles, high to moderate ERbeta mRNA scores were noticed in the granulosa cells. The stroma cells surrounding primordial and primary follicles and the theca cells of secondary follicles showed moderate ERbeta mRNA levels, whereas the ERbeta mRNA score in theca interna and theca externa cells of vital tertiary follicles was distinctly higher. In the theca cells of atretic follicles the score was even higher. Cells of corpora hemorrhagica and corpora lutea had moderate ERbeta mRNA scores, while higher scores were seen in cells of corpora albicantia. Cells of the surface epithelium had a moderate score for ERbeta mRNA, whereas cells of the tunica albuginea and deep stroma showed high ERbeta mRNA scores. The present findings have clearly established a cell-specific localization of ERbeta mRNA in several cell types in the bovine ovary.