Single-cell ATAC and RNA sequencing reveal pre-existing and persistent cells associated with prostate cancer relapse.

Prostate cancer is heterogeneous and patients would benefit from methods that stratify those who are likely to respond to systemic therapy. Here, we employ single-cell assays for transposase-accessible chromatin (ATAC) and RNA sequencing in models of early treatment response and resistance to enzalutamide. In doing so, we identify pre-existing and treatment-persistent cell subpopulations that possess regenerative potential when subjected to treatment. We find distinct chromatin landscapes associated with enzalutamide treatment and resistance that are linked to alternative transcriptional programs. Transcriptional profiles characteristic of persistent cells are able to stratify the treatment response of patients. Ultimately, we show that defining changes in chromatin and gene expression in single-cell populations from pre-clinical models can reveal as yet unrecognized molecular predictors of treatment response. This suggests that the application of single-cell methods with high analytical resolution in pre-clinical models may powerfully inform clinical decision-making.

A novel tissue-slice culture model for non-malignant human prostate.

A novel tissue culture system was established for modeling the non-neoplastic human prostate in vitro. Precision-cut prostate slices were cultivated in culture plates with a gas-permeable base in a novel serum-free mixture. Cultivated specimens was evaluated by an immunohistochemical analysis of cytokeratins 18 and 14, androgen receptor (AR), prostate specific antigen (PSA), prostate acid phosphatase (PAP), and the endothelial cell marker von Willebrand factor. Epithelial viability in the presence and absence of dihydrotestosterone (DHT) was also assessed. Satisfactory maintenance of glandular cytoarchitecture was observed in the presence of DHT with approximately half of the glands displaying a columnar or cuboidal phenotype and an intact layer of basal cells. In the absence of DHT, the corresponding percentage was significantly lower. The occurrence of involutive changes and epithelial cell death was significantly higher in the absence of DHT. Glandular and stromal cells maintained their capacity to express AR. PSA and PAP were expressed throughout the culture period, albeit at a lower level than in uncultured tissue. The viability of endothelial cells differed markedly between individual samples. During culture, the tissue slices became covered with epithelial cells originating from glands that were cut open during tissue slicing. This cell layer consisted of a stratified basal compartment overlaid by cells with a luminal phenotype. The present culture system provides a novel in vitro setting in which to study normal human prostate biology and pathobiology and may help to obviate problems related to the use of established cancer cell lines and animal models.

Altered levels of Smad2 and Smad4 are associated with human prostate carcinogenesis.

Alterations have been demonstrated in ligand and cognate receptor system of the transforming growth factor beta (TGF-beta) pathway in prostate cancer (PC). Still, little is known about changes in the activity of the intracellular Smad cascade of TGF-beta signaling during prostate carcinogenesis. We used immunohistochemistry to analyze phosphorylated Smad2 (p-Smad2), nuclear Smad4 and inhibitory-Smad7 in epithelial cells of normal, hyperplastic and malignant prostate. Specimens comprised 49 tissue cores of PC, 10 benign prostate hypertrophies and three normal prostates. Nuclear p-Smad2 (P<0.001) and nuclear Smad4 (P=0.023) were significantly decreased in PC with remarkable variations in cytoplasmic Smad7 levels. Substantial decreases in p-Smad2 and Smad4 levels were found in specimens with primary Gleason grades 3 and 4, whereas in grade 5, levels were markedly higher. Our results provide the first evidence for changes and reversible attenuation in the Smad system of the TGF-beta pathway during prostate carcinogenesis.

A novel organotypic culture model for normal human endometrium: regulation of epithelial cell proliferation by estradiol and medroxyprogesterone acetate.

BACKGROUND: A novel organotypic culture system was established for modelling the hormonal responses of the normal human endometrium in vitro. METHODS: Endometrial epithelial cells were cultured as glandular organoids within reconstituted extracellular matrix (Matrigel) in tissue culture inserts and stromal cells on plastic below the epithelial compartment. The effects of estradiol (E2) and E2 together with medroxyprogesterone acetate (MPA) on cell proliferation and the expression of estrogen receptor alpha (ERalpha) and progesterone receptor (PR) were studied in 10 epithelial-stromal co-cultures and in three parallel monocultures of epithelial organoids. RESULTS: In co-cultures, E2 was shown to increase the percentage of Ki67-positive cells by approximately 2-fold relative to untreated controls. In the presence of MPA, a significant decrease in cell proliferation was detected. Similar results were obtained when the corresponding percentages of Ki67-positive organoids were calculated instead of individual cells. In the absence of stromal fibroblasts, Ki67 epithelial labelling remained below the control value after both hormonal treatments. Epithelial organoids retained their capacity to express estrogen and progesterone receptors in culture. E2 was shown to markedly increase and MPA to down-regulate the expression of PR. The expression of ERalpha was only slightly affected by either hormonal treatment. CONCLUSIONS: The present organotypic model provides a novel in vitro system in which to study the effects of steroids in the normal human endometrium both in terms of cell proliferation and gene expression. The culture system holds promise as a useful method to screen novel steroid compounds and may help to circumvent problems related to the use of animal models.

Vitamin D and prostate cancer.

Our recent epidemiological study (Ahonen et al., Cancer Causes Control 11(2000) (847-852)) suggests that vitamin D deficiency may increase the risk of initiation and progression of prostate cancer. The nested case-control study was based on a 13-year follow-up of about 19000 middle-aged men free of clinically verified prostate cancer. More than one-half of the serum samples had 25OH-vitamin D (25-VD) levels below 50 nmol/l, suggesting VD deficiency. Prostate cancer risk was highest among the group of younger men (40-51 years) with low serum 25-VD, whereas low serum 25-VD appeared not to increase the risk of prostate cancer in older men (>51 years). This suggests that VD has a protective role against prostate cancer only before the andropause, when serum androgen concentrations are higher. The lowest 25-VD concentrations in the younger men were associated with more aggressive prostate cancer. Furthermore, the high 25-VD levels delayed the appearance of clinically verified prostate cancer by 1.8 years. Since these results suggest that vitamin D has a protective role against prostate cancer, we tried to determine whether full spectrum lighting (FSL) during working hours could increase serum 25-VD concentrations. After 1-month exposure, there was no significant increase in the serum 25-VD level, although there was a bias towards slightly increasing values in the test group as opposed to decreasing values in controls. There was no significant change in the skin urocanic acid production. The possibility to use FSL in cancer prevention is discussed. In order to clarify the mechanism of VD action on cell proliferation and differentiation, we performed studies with the rat and human prostates as well prostate cancer cell lines. It is possible that 25-VD may have a direct role in the host anticancer defence activity, but the metabolism of vitamin D in the prostate may also play an important role in its action. We raised antibodies against human 1alpha-hydroxylase and 24-hydroxylase. Our preliminary results suggest that vitamin D is actively metabolised in the prostate. Vitamin D appears to upregulate androgen receptor expression, whereas androgens seem to upregulate vitamin D receptor (VDR). This may at least partially explain the androgen dependence of VD action. VD alone or administered with androgen causes a suppression of epithelial cell proliferation. VD can activate mitogen-activated kinases, erk-1 and erk-2, within minutes and p38 within hours. Also, auto/paracrine regulation might be involved, since keratinocyte growth factor (mRNA and protein) was clearly induced by VD. Based on these studies, a putative model for VD action on cell proliferation and differentiation is presented.

Correlation between serum inhibin B and other indicators of the ovarian function.

OBJECTIVE: To investigate the correlation between the early follicular phase serum inhibin B levels and other indicators of ovarian reserve. STUDY DESIGN: Seventy-four women aged 24-40 years (mean 32) with different infertility etiologies were investigated in the early follicular phase of a spontaneous mentrual cycle. The volume of the ovaries was measured and the total number of follicles <5 mm in size counted by ultrasound. Serum levels of FSH, estradiol (E2) and inhibin B were measured on the same day. In stepwise regression analysis inhibin B levels were correlated with age, body-mass-index, the ultrasound measurements, cause of infertility, parity, FSH and E2. RESULTS: FSH, BMI and the number of follicles proved to be statistically significant independent predictive factors for the inhibin B levels, FSH and BMI correlating negatively and the number of follicles positively with inhibin B serum concentrations. CONCLUSION: The number of small follicles reflect the inhibin B production of the ovaries. BMI being as strong predictive factor of inhibin B levels as FSH could in part explain the impaired likelihood of conceiving in obese patients.

Inhibin A, B and pro-alphaC in serum and peritoneal fluid in postmenopausal patients with ovarian tumors.

OBJECTIVE: To compare serum and peritoneal fluid concentrations of inhibin A, B, and pro-alphaC in women with ovarian tumors. METHODS: Serum and peritoneal fluid samples were taken from 41 postmenopausal women operated on for an ovarian tumor. Twenty-one patients with endometrial cancer formed a control group. Serum and peritoneal fluid inhibin A, B, and pro-alphaC concentrations, and serum FSH and tumor marker CA 125 (study group only) concentrations were analyzed. RESULTS: Inhibin A was found in low concentrations (median 4.1pg/ml, range <2-29pg/ml) in serum in most postmenopausal patients with epithelial ovarian carcinoma, whereas inhibin B was not measurable. Inhibin pro-alphaC circulated in high concentrations (median 125pg/ml, range 37->1000pg/ml). All inhibins were found in clearly greater concentrations in the peritoneal fluid than in serum. International Federation of Gynecology and Obstetrics (FIGO) stage III-IV and poor differentiation grade were associated with significantly lower concentrations of inhibin A and pro-alphaC in the peritoneal fluid compared with stages I-II or low grade. This correlation was not found in the serum concentrations of inhibin A or pro-alphaC. In the control group, no dimeric inhibins were found in serum, and pro-alphaC circulated in median concentrations of 47pg/ml (range 12-174pg/ml). CONCLUSIONS: Postmenopausal patients with epithelial ovarian tumors had low concentrations of inhibin A and relatively high concentrations of inhibin pro-alphaC in serum. The peritoneal fluid concentrations of all inhibins far exceeded those in the serum. Relatively low concentrations of inhibin A and pro-alphaC in the peritoneal fluid of patients with ovarian cancer seem to be associated with high stage and grade and, to a lesser degree, with positive peritoneal cytology.

Prognostic factors in controlled ovarian hyperstimulation.

OBJECTIVE: To determine whether the number of retrieved oocytes and the required amount of gonadotropins per oocyte in IVF treatment can be predicted with use of the following independent predictive variables: age, parity, cause of infertility, body mass index, day 3-5 FSH, E2, inhibin B, ovarian volume, the number of follicles, and intraovarian and uterine artery vascular resistance measured by ultrasonography before ovarian hyperstimulation. DESIGN: A retrospective analysis. SETTING: University hospital infertility clinic. PATIENT(S): Seventy-four consecutive women attending the university hospital infertility clinic for IVF treatment. INTERVENTION(S): The investigated factors were measured on day 3-5 of the cycle, in which luteal phase suppression was begun before ovarian hyperstimulation preparatory to IVF. MAIN OUTCOME MEASURE(S): The amount of gonadotropins required per oocyte and the number of retrieved oocytes were correlated with the predictive factors in stepwise regression analysis. RESULT(S): The best predictive factors for the number of oocytes retrieved were FSH, inhibin B, and parity, explaining 25% of the ovarian response. Intraovarian vascular resistance, parity, FSH, and inhibin B best predicted the amount of gonadotropins needed, explaining 44% of the variation. CONCLUSION(S): FSH, inhibin B, and parity were the independent predictive factors for the number of retrieved oocytes. The same factors and intraovarian vascular resistance predicted the required amount of gonadotropins per oocyte. The main part of the ovarian response cannot be predicted using the factors investigated.

Identification of a nuclear localization signal in activin/inhibin betaA subunit; intranuclear betaA in rat spermatogenic cells.

Activin is a dimeric glycoprotein hormone that was initially characterized by its ability to stimulate pituitary FSH secretion and was subsequently recognized as a growth factor with diverse biological functions in a large variety of tissues. In the testis, activin has been implicated in the auto/paracrine regulation of spermatogenesis through its cognate cell membrane receptors on Sertoli and germ cells. In this study we provide evidence for intranuclear activin/inhibin betaA subunit and show its distribution in the rat seminiferous epithelium. We have shown by transient expression in HeLa cells of beta-galactosidase fusion proteins that the betaA subunit precursor contains a functional nuclear localization signal within the lysine-rich sequence corresponding to amino acids 231-244. In all stages of the rat seminiferous epithelial cycle, an intense immunohistochemical staining of nuclear betaA was demonstrated in intermediate or type B spermatogonia or primary spermatocytes in their initial stages of the first meiotic prophase, as well as in pachytene spermatocytes and elongating spermatids primarily in stages IX-XII. In some pachytene spermatocytes, the pattern of betaA immunoreactivity was consistent with the characteristic distribution of pachytene chromosomes. In the nuclei of round spermatids, betaA immunoreactivity was less intense, and in late spermatids it was localized in the residual cytoplasm, suggesting disposal of betaA before spermatozoal maturation. Immunoblot analysis of a protein extract from isolated testicular nuclei revealed a nuclear betaA species with a molecular mass of approximately 24 kDa, which is more than 1.5 times that of the mature activin betaA subunit present in activin dimers. These results suggest that activin/inhibin betaA may elicit its biological functions through two parallel signal transduction pathways, one involving the dimeric molecule and cell surface receptors and the other an alternately processed betaA sequence acting directly within the nucleus. According to our immunohistochemical data, betaA may play a significant role in the regulation of nuclear functions during meiosis and spermiogenesis.

Ovarian cancer and gonadotropins in vitro: new evidence in favor of independence.

BACKGROUND: The aim of this study was to investigate whether newly established epithelial ovarian carcinoma cell lines secrete inhibins, and whether their proliferative and secretory activity can be regulated by gonadotropins. MATERIALS AND METHODS: Three recently characterized human epithelial ovarian carcinoma cell lines were exposed to human choriongonadotropin hCG and follicle stimulating hormone FSH. Cell proliferation was determined by counting. Secretory activity of the cell lines was studied by analyzing inhibin A, inhibin B, inhibin pro-aC, estradiol, progesterone, testosterone, and CA 125 concentrations from the medium. The expression of gonadotropin receptors was studied by RT-PCR. RESULTS: None of the cell lines were found to secrete any of the inhibins, progesterone or testosterone. Only UT-OC-4 cells secreted low levels of estradiol. Gonadotropin receptors were not expressed by any of the cell lines, and accordingly neither FSH nor hCG stimulated the growth of these cells. However, hCG had some dose dependent growth inhibitory effect on UT-OC-3. Passage 42 cells of UT-OC-3 secreted significantly more CA 125 than passages 8 cells (P = 0.000). CONCLUSIONS: The present results suggest that the carcinomatous epithelial cells of the ovary do not secrete inhibins. The serum inhibin levels previously detected in patients with epithelial ovarian carcinoma may therefore reflect an ovarian stromal response to carcinoma. The findings are also in favor of an independence of ovarian cancer of gonadotropins.

Localization of 1,25-dihydroxyvitamin D3 receptor (VDR) expression in human prostate.

1,25-dihydroxyvitamin D3 (1,25(OH)2D3) has been found to have a variety of physiological functions, including effects on growth and differentiation in normal and malignant cells. The antiproliferative effects of 1,25(OH)2D3 are reported to be mediated through the genomic signaling pathway by binding to a specific high affinity receptor protein, the 1,25-dihydroxyvitamin D3 receptor (VDR). VDR has been localized in a variety of tissues, but little is known about VDR distribution in human prostate. In this study, we raised an antibody against a synthetic peptide corresponding to amino acids 10-24 of human vitamin D receptor. The sequence selected for immunization is identical in human, rat and mouse VDR. Based on this antibody, we developed an immunohistochemical method suitable for studying VDR expression in paraffin-embedded tissue. The immunohistochemical staining was verified using classical target organs for vitamin D (kidney, intestine, skin). With this method, we studied VDR localization on paraffin-embedded human prostatic tissue obtained from 8 patients undergoing radical prostatectomy for urinary bladder cancer and demonstrate VDR expression in the secretory epithelial and few stromal cells of human prostate. The nuclear staining in the secretory epithelial cells was concentrated near the nuclear membrane and in discrete foci in the nucleoplasm. This suggests that effects of 1,25-dihydroxyvitamin D3 are mediated through VDR in these cells. Moreover our result indicates that there are strong variations in VDR expression between prostatic samples.

Activin B: detection by an immunoenzymometric assay in human serum during ovarian stimulation and late pregnancy.

A recently developed immunoenzymometric assay for activin B has been characterized further by measurement during ovarian stimulation and pregnancy. The assay is based on a monoclonal anti-peptide antibody, anti-betaB(101-115). In addition to quantitative analyses, the antibody has been used for immunohistochemical localization of the activin betaB-subunit in human term placenta. Serum samples obtained from patients suffering from tubal factor infertility who were admitted for in-vitro fertilization (IVF) treatment protocols or from patients with proven fertility who were admitted for laparoscopic tubal ligation were collected. The aim was to correlate serum activin B concentrations with other parameters during IVF and with phases of the menstrual cycle. Serum samples obtained from healthy pregnant volunteers were studied to correlate activin B concentrations with clinical parameters. During the IVF treatment protocols, activin B was detectable in all patients studied, and a significant negative correlation was observed between serum activin B and oestradiol concentrations. On the other hand, no significant difference was observed in activin B concentrations when serum samples obtained from patients at different phases of the menstrual cycle were compared, and low concentrations of activin B were observed in the samples obtained from these patients. During pregnancy, a positive correlation was observed between serum activin B concentrations and gestational age. In immunohistochemical analyses of human placental tissue obtained from healthy parturients, the activin betaB-subunit was present in trophoblast, amniotic epithelial and Hofbauer cells. The results suggest a potential clinical application in female reproductive medicine for serum activin B measurements.

Inhibin A and B in peri- and postmenopause.

OBJECTIVE: The aim of the study was to investigate the origin of inhibin A and B during the last years of the reproductive age and after menopause by measuring their levels in the ovarian and peripheral venous blood. METHODS: The study population consisted of 43 women, aged 42-69 years (mean 50), who underwent hysterectomy with ovarian removal for a benign disease. A total of 24 of them were in follicular phase, 11 in luteal phase, and eight were postmenopausal. Peripheral and ovarian venous blood was collected for measurement of inhibin A and B. In addition, sex steroid hormone and gonadotropin levels were measured. RESULTS: Ovarian venous inhibin B correlated significantly with ovarian estradiol secretion (r = 0.5, P = 0.001). The levels of inhibin B were significantly higher in the ovarian vein than in the peripheral vein (P = 0.006). The highest inhibin B concentrations were detected in the mid-proliferative (mid-follicular) phase (median 31.6 pg/ml range 25.9-47.9). In postmenopausal women, inhibin B was not detectable. No correlation between FSH and ovarian inhibin B was found. Inhibin A rose rapidly in late proliferative (late follicular) phase (median 28.5, range < 2-51.8) and dominated in the circulation throughout the luteal phase (median 20.9, range 8.8-60). For inhibin A, no concentration gradient existed between the ovarian and peripheral vein. Unlike inhibin B, inhibin A was detectable in ovarian and peripheral blood in postmenopausal women. A significant negative correlation between ovarian and peripheral inhibin A and FSH was found (r = -0.386, P = 0.015; r = -0.345, P = 0.034, respectively). CONCLUSION: Inhibin B correlates with ovarian estradiol secretion and seems to reflect follicular function. Inhibin A dominates in circulation during the luteal phase but is detectable at low concentrations both in follicular phase and even in postmenopause. Our findings suggest that inhibin A may play a role in FSH suppression in the female reproduction. In addition to the ovary, there may be extragondal source(s) of inhibin A.

Reappraisal of the role of heat shock proteins as regulators of steroid receptor activity.

Almost 30 years have passed since the original demonstration that steroid receptors, comprising a subfamily of the nuclear receptor (NR) superfamily, exist as large (6-8S) non-DNA-binding complexes in hypotonic extracts (cytosol) of target cells; later such complexes were shown to correspond to a heterooligomer composed of receptor, heat shock (Hsp), and other proteins. Subsequently, an impressive number of studies have dealt with the composition of the "nonactive" complex, its dissociation and/or reassembly in vitro, possible functions of the non-receptor components, and their subcellular compartmentalization. While there is little dispute about the chaperoning role of some Hsps in such a complex, there is still no final proof of an association in vivo of NRs and Hsps in the nuclei of target cells, which is requisite for a direct regulatory involvement of Hsps in NR function. Here we critically review the various models that have been put forward to attribute a biological function to the NR-Hsp90 interaction, evaluate the corresponding experimental data, and integrate recent concepts originating from the structural and functional analyses of NRs.

Immunodetection of androgen receptor in human urinary bladder cancer.

We investigated the expression of androgen receptor (AR) protein in transitional cell carcinoma of human urinary bladder in paraffin-embedded sections of tumours obtained from nine patients with urinary bladder cancer treated by radical cystectomy. In addition, immunoblotting of AR was also performed on selected samples. Nuclear immunoreactivity of AR was found in seven of the nine urinary bladder cancers studied. AR showed variable staining intensity within a tumour. In the immunoblots, a 110 kDa AR signal was seen with anti-AR antibody, and faint bands of 90 and 60 kDa were also observed. Immunohistochemistry of p53 and c-erbB-2 was also carried out and compared with the distribution of AR. The high frequency of AR expression suggests a role for androgens in transitional cell carcinoma of human urinary bladder.

Expression of the chicken progesterone receptor forms A and B is differentially regulated by estrogen in vivo.

The chicken progesterone receptor (cPR), like its human counterpart (hPR), exists as two isoforms, PR-A and PR-B, displaying different biological activities depending upon cellular and promoter contexts. Here we show that the ratio of PR isoforms observed in the immature chicken oviduct is changed during estrogen-induced differentiation from PR-B dominancy to that of PR-A. This is the first report describing that the expression ratio of PR isoforms is altered by upregulation of PR-A by estrogen action in vivo. This result provides a plausible explanation to the differences in oviduct's response to progesterone depending on hormonal and developmental status of the animal.

Spermatogenesis in the vitamin A-deficient rat: possible interplay between retinoic acid receptors, androgen receptor and inhibin alpha-subunit.

In order to understand the mechanisms of retinol action on the testis, testicular retinoic acid receptor alpha, beta(RAR alpha and beta), androgen receptor (AR) and inhibin alpha-subunit were studied in normal, vitamin A-deficient (VAD) and vitamin A-supplemented rats by immunohistochemistry and immunoblotting. Compared to the normal testis, expression of 110 K AR was up-regulated by vitamin A withdrawal, whereas 51 K RAR alpha remained unchanged. An additional 55 K RAR alpha signal was observed. Readministration of retinol caused a marked decrease of AR in the VAD testis. By 24 h, AR declined to below the normal level. Although the 51 K RAR alpha signal remained unchanged, the 55 K band was slightly up-regulated at 6 h after retinol administration. A 51 K RAR beta protein was seen in the VAD but in not the normal testis. The intensity of the 51 K RAR beta band remained constant before and after the administration of retinol, but it had a slight up-shift at 6 h after retinol injection, suggesting post-translational modification of the receptor. The inhibin alpha-subunit of 18 K protein was undetectable in the VAD testis and increased to above normal level at 24 h after retinol administration. Immunohistochemically, nuclear AR immunostaining was more intense in the VAD testis than in the normal testis. The intensity of immunostaining declined in all AR-positive cells after the injection of retinol, but the decrease was more evident in Sertoli than in other cells. At 24 h after retinol the immunostaining was undetectable in most Sertoli cells. The regulation of the inhibin alpha-subunit by retinol in the cytoplasm of Sertoli cells detected by immunohistochemistry was correlated to the results in immunoblotting. These results suggest a possible interplay between retinoids, androgen and inhibin signalling systems in Sertoli cells in the regulation of spermatogenesis during retinol action.

Immunochemical and in situ hybridization analyses of retinoic acid receptor alpha, beta, and gamma in murine Harderian and submandibular glands.

Retinoic acid (RA), through its cognate receptors (retinoic acid receptors, RARs), plays an important role in the ontogenesis and maintenance of the normal function of murine Harderian and submandibular glands. In the present study, autoradiography was used to study RA binding to these glands. Both glands showed high radioactive labelling after [14C]-RA administration in normal and partially vitamin A-deficient (VAD) mice. The peak uptake was at 6 h after [14C]-RA administration in normal mice and at 0.5 h in VAD mice. At 24 h, RA binding remained high in normal mice, while it decreased significantly in VAD mice. In western blots with an antibody recognizing all forms of RARs, a band of molecular weight 51 kDa was seen in homogenates of both glands. Immunohistochemically, RAR staining was found in the nuclei of the glandular cells. The Harderian gland exhibited more intense staining than the submandibular gland. In the latter, the most intense staining was seen in the acinar cells, followed by the intercalated duct cells. The granular convoluted tubule showed weak immunostaining and the striated duct was negative. In the Harderian gland, RAR immunostaining was observed in both type I and II cells, but only part of them stained with RAR antibody. The expression of RAR alpha, beta, and gamma transcripts was studied by in situ hybridization using specific oligonucleotide probes. The cell-specific expression of RAR alpha mRNA in the submandibular gland corresponded to the RAR proteins detected by immunohistochemistry, while the RAR beta transcript was mainly seen in the striated duct. The transcripts of RAR alpha and beta were evenly distributed in type I and II glandular cells of the Harderian gland. RAR gamma labelling was below detectable levels in both glands. This result suggests that RA and RARs regulate the functions of Harderian and submandibular glands in a cell-specific manner.