Abstract
The prostate gland develops from the fetal urogenital sinus at the base of the urinary bladder. It finally differentiates into three lobes; ventral, lateral and dorsal lobes of the prostate. In spite of their common developmental origin and similar glandular structure, these lobes show the different biochemical characteristics, for example, in the proteins they secrete. In the present study, we investigate the involvement of the epithelial–mesenchymal interaction in the lobe-specific differentiation of the prostatic epithelium by means of epithelial–mesenchymal recombination experiments. We have used a prostatic steroid-binding protein (PSBP) as a specific differentiation marker for the ventral prostate. PSBP is a tetramer which consists of 2 subunits, one containing the polypeptides Cl and C3 and the other containing the polypeptides C2 and C3. Northern analysis with a complementary DNA probe encoding Cl peptide (PSBP-C1) revealed that the mRNAs were detected exclusively in the ventral prostate but not in the dorsal prostate or in other organs such as urinary bladder and kidney. In situ hybridization with a complementary (anti-sense) RNA probe demonstrated that the transcripts were found only in the epithelium, not in the mesenchyme of the ventral prostate. In situ hybridization also showed that, in normal development, the mRNAs for PSBP-C1 in the ventral epithelium were first detectable at day 14 after birth, coinciding with the onset of its cytodifferentiation, and that they reached mature levels by day 21.
We then carried out tissue-recombination experiments to examine whether the transcription of the PSBP-C1 gene in the epithelium is affected by the surrounding mesenchyme. Fetal urogenital sinuses were subdivided into ventral and dorsal halves. Following collagenase treatment, both halves were separated into their epithelial and mesenchymal compartments. Homotypic (ventral epithelium plus ventral mesenchyme [Ev/ Mv] and dorsal epithelium plus dorsal mesenchyme [Ed/Md]) and heterotypic (ventral epithelium plus dorsal mesenchyme [Ev/Md] and dorsal epithelium plus ventral mesenchyme [Ed/Mv]) recombinations were carried out. After 4–5 weeks of growth in male host, the glandular structures characteristic for prostate glands were formed in all explants. However, in situ hybridization revealed the transcripts of the PSBP-C1 gene only in the epithelium associated with the ventral mesenchyme (Ev/Mv and Ed/Mv). In contrast, the transcripts were never found in the epithelium associated with the dorsal mesenchyme (Ev/Md and Ed/Md). Moreover, we found that the urinary bladder epithelium has limited ability to transcribe the PSBP-C1 gene under the influence of the ventral mesenchyme of the urogenital sinus.
These results indicate that the ventral mesenchyme is a potent inducer for ventral prostate-specific cytodifferentiation and suggest that the lobe-specific differentiation in the prostatic epithelium could be directed by its surrounding mesenchyme through epithelial–mesenchymal interactions.
Introduction
Epithelial–mesenchymal interactions play an important role in organogenesis (reviewed by Wessels, 1977; Kratochwil, 1892; Sawyer and Fallon, 1983). Many studies have demonstrated that the mesenchyme directs or influences the developmental fate of the epithelium. For instance, lens crystalline is induced in lentoids formed in the trunk ectoderm associated with the lens cup (Karkinen-Jaaskelainen, 1978) and embryonic chick pepsinogen (a marker protein for glandular stomach epithelium) is induced in the glands formed in esophagus and gizzard epithelium associated with the mesenchyme of glandular stomach (Takiguchi et al. 1986).
The epithelial–mesenchymal interaction is also required for the hormone-dependent development of the prostate gland. It develops from the male fetal urogenital sinus as epithelial buds projecting from the epithelium into surrounding mesenchyme. The epithelial buds branch and differentiate into a group of secretory epithelial cells lining a large lumen. These processes are androgen-dependent (Lasnitzki and Mizuno, 1977 and 1979) and castration or anti-androgen treatment cause the female type development of the urogenital sinus, which finally forms a caudal part of vagina and urethra (Jost, 1965; Neumann et al. 1970). Previous tissue recombination experiments and steroid autoradiography have revealed that the mesenchyme is an actual target for androgens and that the androgen-activated mesenchyme induces prostatic morphology of the epithelium. TTie incubation of fetal sinuses with [3H]androgen revealed that only the mesenchyme uptakes androgens (Takeda et al. 1985). The sinus epithelium obtained from androgen-insensitive Tfm (testicular feminization) mice, which are completely deficient in androgen receptors (Lyon and Hawkes, 1970; Ohno, 1977; Takeda et al. 1987) can form prostate-like glandular structures when recombined with the androgen-sensitive wild-type mesenchyme (Cunha and Lung, 1978; Lasnitzki and Mizuno, 1980; Mizuno et al. 1988). This concept has been supported by the heterotypic recombination between adult urinary bladder epithelium and the fetal sinus mesenchyme where the sinus mesenchyme could induce glandular structures in the bladder epithelium (Cunha et al. 1983). However, it has also been reported that glandular epithelium induced in urinary bladder epithelium does not show full prostatic cytodifferentiation, suggesting that the morphology and the cytodifferentiation are controlled by different mechanisms (Suematsu et al. 1988). Then the question has been raised as to whether or not the sinus mesenchyme can control the functional differentiation of the sinus epithelium as well as its morphogenesis. Recently, we found that the lobespecific differentiation of the rat prostate gland could be a good experimental model to answer this question.
The adult rat prostate consists of ventral, lateral and dorsal lobes. They all develop from the fetal urogenital sinus and form similar glandular structures but they show different biochemical characteristics, in their secretory proteins, for instance. Among the lobespecific secretory proteins, prostatic steroid-binding protein (PSBP or a-protein) has been extensively studied and its cDNA has been cloned (Chen et al. 1982; Parker et al. 1982). PSBP is a major protein synthesized and secreted by the rat ventral prostate. It is a tetramer, which consists of 2 subunits, one containing the polypeptides Cl and C3 and the other containing the polypeptides C2 and C3. Previous biochemical analyses have demonstrated that it is expressed exclusively in the rat ventral prostate and that androgens stimulate its expression. (Heyns and De Moor, 1977; Zhang and Parker, 1985). The lobe-specific differentiation of the prostatic epithelium can be analyzed with this marker. In the present study, the fetal urogenital sinus was divided into the dorsal and ventral halves and the reciprocal epithelial–mesenchymal recombinations were carried out. We then examined by in situ hybridization technique whether the ventral mesenchyme induces both the transcription of the PSBP gene and prostatic morphology in the recombinants.
Materials and methods
Animals and tissues
Inbred rats (Fischer 344, Charles River Japan Inc., Kanagawa) were mated during the night and copulation was confirmed by the presence of spermatozoa in the vaginal smears on the following morning. We consider that 0.5 days of gestation have elapsed at 12:00 noon of this day.
Urogenital sinuses and urinary bladder were isolated from 16.5-day fetuses. At this stage no prostatic bud is formed. Prostatic buds normally appear between 17.5 and 18.5 days of gestation. Adult urinary bladder was obtained from pregnant rats.
Tissue recombination and grafting
Precise tissue recombination and grafting methods were described previously by Suematsu et al. (1988). Fig. 1 shows the experimental design used in the present study. Urogenital sinuses were divided into ventral and dorsal halves. Fetal tissues were treated with 0.06% collagenase (Worthington Biochemical Corp., Type CLS) for 30min at 37°C. Adult bladders were treated with 0.03 % collagenase in Tyrode’s solution for 2h at 37 °C. The collagenase-treated tissues were washed in 50 % fetal bovine serum (FBS) in Tyrode’s solution for 2h with three changes, and the epithelium was separated from the mesenchyme with watch-maker’s forceps. The separated epithelium was recombined with the mesenchyme and cultured overnight in Medium 199 supplemented with 10% FBS to allow the epithelium and the mesenchyme to adhere, and the recombinant was grafted beneath the kidney capsule of syngenic adult male host. Grafts were harvested after 4-5 weeks of in vivo culture and were processed for in situ hybridization. The number of the recombinants used in the present study is shown in Fig. 6.
Scheme of recombination experiments. Urogenital sinuses were subdivided into ventral and dorsal halves. They were then separated into epithelial and mesenchymal parts. Homotypic (Ev/Mv and Ed/Md) and heterotypic (Ev/Md and Ed/Mv) recombinations were carried out. The fetal and adult urinary bladder epithelia were also associated with the dorsal or ventral sinus mesenchymes (not shown in the scheme). The recombinants were cultured for 4–5 weeks underneath the kidney capsule of male hosts. Ev, epithelium from the ventral half; Mv, mesenchyme form the ventral half; Ed, epithelium from the dorsal half; Md; mesenchyme form the dorsal half.
Scheme of recombination experiments. Urogenital sinuses were subdivided into ventral and dorsal halves. They were then separated into epithelial and mesenchymal parts. Homotypic (Ev/Mv and Ed/Md) and heterotypic (Ev/Md and Ed/Mv) recombinations were carried out. The fetal and adult urinary bladder epithelia were also associated with the dorsal or ventral sinus mesenchymes (not shown in the scheme). The recombinants were cultured for 4–5 weeks underneath the kidney capsule of male hosts. Ev, epithelium from the ventral half; Mv, mesenchyme form the ventral half; Ed, epithelium from the dorsal half; Md; mesenchyme form the dorsal half.
Some separated tissues were fixed and processed for histological observations to confirm that separated epithelium and stroma are free of contaminating cells. Previously we also recombined mouse and rat tissues and stained them with Hoechst dye No. 33258 (Calbiochem-Behring, La Jolla, CA). By this staining, we can distinguish mouse from rat tissues (Cunha et al. 1984: Mizuno et al. 1988). So far, no contamination was observed in any recombinants.
Northern blot analysis of PSBP-C1mRNA
pSP64 plasmid carrying cDNA for Cl peptide of prostatic steroid-binding protein (450 base pairs; PSBP-C1) was kindly given by Dr M. G. Parker (Imperial Cancer Research Fund Laboratories, London).
Total RNA was isolated from adult rat male tissues by the guanidinium thiocyanate method. Total RNA was treated with formaldehyde/formamide, separated by electrophoresis on a 1.5% agarose gel and transferred to a nylon filter
(Hybond-N, Amersham). Equal amount of loading was confirmed by ethidium bromide staining of the gel after electrophoresis, which sho.wed nearly the same intensity of 28S and 18S ribosomal RNAs. Radiolabelling of the probes was performed by the random primer method (Random primer labelling kit, Boehringer Manheim) with [32P]dCTP (Amersham). The specific activity of this random primed probe is 7 × 108 disints min-1 μg-1. The hybridization condition was 5 × SSPE (1 × SSPE is 180mM NaCl, 1mM EDTA, 10mM Na2PO4), 50 μ gml-1 salmon sperm DNA and 1 × 106 disints min-1 probe/ml hybridization solution at 68 °C for 16 h. The filter was washed with a solution containing 1 × SSC (150mM NaCl, 15mM sodium citrate), 1% SDS at room temperature for 3 × 15 min, and 2 × SSC, 1% SDS at 65°C for 2 × 30min. Hybrids were detected by autoradiography with an intensifying screen at–70°C.
Radiolabelling of cRNA probe
To make an RNA probe for in situ hybridization, PSBP-C1 cDNA was cut out from pSP64 plasmid and subcloned at the PstI site of pBluescriptll SK(+) (Stratagene, San Diego, CA). For 35S-UTP (Amersham) labelled sense and anti-sense RNA probes, the plasmids were linearized with BamHI or EcoRI and transcribed in vitro with T7 or T3 RNA polymerase (Stratagene), respectively (Sambrook et al. 1989). The transcribed RNAs were separated by gel electrophoresis and reduced to an average size of 100–150 base pairs by limited alkaline hydrolysis.
In situ hybridization
The tissues were fixed overnight at 4 °C with 2 % paraformaldehyde, 2% glutaraldehyde and 1% saturated picric acid in 0.1M phosphate buffer (pH7.5), dehydrated and embedded in paraffin. 7 μ m sections were made and mounted onto gelatin-coated slides. After deparaffinization and rehydration, tissue sections were treated with proteinase K (1 μ gml-1) at 37°C for 30min, fixed again with 4% paraformaldehyde, washed in phosphate-buffered saline (PBS), dehydrated through graded-ethanol series, and air dried. Hybridization conditions were as follows: 0.3 M NaCl, 1 mgml-1E. coli tRNA, 125 mM dithiothreitol (DTT), 20mM Tris-HCl (pH8.0), 2.5mM EDTA, 0.02% ficoll, 0.02 polyvinyl-pyrrolidone, 50% formamide, 10% dextran sulfate. Hybridization was carried out at 50 °C for 16 h. After brief washing with 2 × SSC containing 50% formamide and 0.1% β-mercapto-ethanol (BME), tissue sections were treated with RNase A (20 μ gml-1) at 37°C for 30min to remove non-specifically hybridized probes and washed again with 2 × SSC, 50% formamide, 0.1 % BME at 50°C for 1h and finally washed with 1 × SSC, 50 % formamide, 0.1 % BME at 50°C for 1.5 h. After washing, slides were dehydrated, dipped in NTB-3 Kodak emulsion (1:1 dilution; Eastman Kodak Co., Rochester, NY, USA) and exposed for one week at 4°C in a desiccated black box. The slides were developed with D-19 Kodak developer (1:1 dilution) at 15 °C and stained with haematoxylin.
For the quantitative analysis, the exposure time was reduced to 4 days so that each epithelial nucleus was easily identified after haematoxylin staining. More than ten sections were randomly selected from each explant and the silver grains over the glandular epithelial cells were counted. The epithelial cells possessing more than 30 silver grains were considered to be labelled. PSBP-C1 transcription was expressed as the percentage of labelled cells.
Results
Tissue-specific expression of PSBP-C1 mRNA
Northern blot analysis clearly showed that mRNAs encoding a PSBP-C1 peptide were expressed exclusively in the ventral prostate. In contrast, PSBP-C1 mRNA were not detected in the dorsal prostate, kidney, bladder or liver (Fig. 2). Even,in the longer exposure conditions, no sign of accumulation of the RNA was observed (data not shown). To investigate the tissue localization of the transcript, in situ hybridization was carried out on the tissue sections using an in vitro transcribed RNA probe. The autoradiograph showed that the expression of PSBP-C1 mRNAs was restricted to the epithelium of the ventral prostate but not detectable in the mesenchyme (Fig. 3A). Control experiments with sense-strand RNA showed only background levels of hybridization (Fig. 3B). By contrast to the ventral prostate, no specific hybridization signal was observed over the sections from other organs such as dorsal prostate and urinary bladder (Fig. 3C and 3D).
Tissue specificity of PSBP-C1 mRNA expression. 5 μ g of total RNA extracted from ventral (V) prostate, dorsal (D) prostate, kidney, urinary bladder and liver of adult rats were subjected to Northern blot analysis with 32P-labelled cDNA probes. The position of 18S, 28S ribosomal RNA and tRNA used as a size marker is shown on both sides.
Tissue specificity of PSBP-C1 mRNA expression. 5 μ g of total RNA extracted from ventral (V) prostate, dorsal (D) prostate, kidney, urinary bladder and liver of adult rats were subjected to Northern blot analysis with 32P-labelled cDNA probes. The position of 18S, 28S ribosomal RNA and tRNA used as a size marker is shown on both sides.
Tissue specificity of PSBP-C1 mRNA expression revealed by in situ hybridization. An RNA probe synthesized from PSBP-C1 cDNA was hybridized to a cross-section of rat adult ventral prostate (A,B), dorsal prostate (C) and urinary bladder (D). B shows an autoradiograph with sense RNA probe and the others with anti-sense RNA probe. The epithelium of the ventral prostate shows a high concentration of silver grains while dorsal prostate and urinary bladder do not show a specific signal. Prostatic fluids which sometimes remain in lumens tend to show relatively high back ground of silver grains (C). A section hybridized with sense RNA probe (B) shows no specific signal above background level, e, epithelium; 1, lumen; m, mesenchyme. Exposure period: 1 week. Bar, 50 μm.
Tissue specificity of PSBP-C1 mRNA expression revealed by in situ hybridization. An RNA probe synthesized from PSBP-C1 cDNA was hybridized to a cross-section of rat adult ventral prostate (A,B), dorsal prostate (C) and urinary bladder (D). B shows an autoradiograph with sense RNA probe and the others with anti-sense RNA probe. The epithelium of the ventral prostate shows a high concentration of silver grains while dorsal prostate and urinary bladder do not show a specific signal. Prostatic fluids which sometimes remain in lumens tend to show relatively high back ground of silver grains (C). A section hybridized with sense RNA probe (B) shows no specific signal above background level, e, epithelium; 1, lumen; m, mesenchyme. Exposure period: 1 week. Bar, 50 μm.
Expression of PSBP-C1 mRNA during development of the rat ventral prostate
The developmental change of PSBP-C1 mRNA expression in the ventral prostate was examined by in situ hybridization. Prostate glands develop from the fetal urogenital sinus as epithelial buds projecting from the sinus epithelium into the surrounding mesenchyme. At 8 –10 days after birth, lumina begin to develop in the prostatic buds and by 3–4 weeks of age the prostatic epithelium differentiates into a columnar secretory epithelium lining a large lumen, which is the typical morphology for the adult prostate. The fetal prostate from 20-day fetuses and the postnatal prostate from 4-, 9-, 14- and 21-day-old male rats were removed and processed for in situ hybridization. The PSBP-C1 mRNAs were undetectable until day 9 after birth. They were first detectable at day 14 and reached mature levels at 21 days (Fig. 4).
Developmental control of mRNA encoding PSBP-C1. Antisense (A–C) and sense (D) RNA probe were hybridized to a cross-section of the ventral prostate from 9-day (A), 14-day (B) and 21-day (C, D) rats. The specific signal is first detectable on day 14 after birth when the lumen starts to develop in the epithelial buds (arrows) (B). The transcripts reaches mature level by day 21 (C). A section hybridized with sense RNA probe shows no specific signal above background level, b, epithelial bud; e, epithelium; 1, lumen; m, mesenchyme. Exposure period: 1 week. Bar, 50 μ m.
Developmental control of mRNA encoding PSBP-C1. Antisense (A–C) and sense (D) RNA probe were hybridized to a cross-section of the ventral prostate from 9-day (A), 14-day (B) and 21-day (C, D) rats. The specific signal is first detectable on day 14 after birth when the lumen starts to develop in the epithelial buds (arrows) (B). The transcripts reaches mature level by day 21 (C). A section hybridized with sense RNA probe shows no specific signal above background level, b, epithelial bud; e, epithelium; 1, lumen; m, mesenchyme. Exposure period: 1 week. Bar, 50 μ m.
PSBP-C1 mRNA transcription induced in the sinus epithelium
Although both ventral and dorsal prostate develop from the urogenital sinus and show glandular struc-tures, PSBP-C1 mRNA is transcribed only in the epithelium of the ventral prostate. To investigate whether its transcription is under the control of the surrounding mesenchyme, we performed the epithelial–mesenchymal recombination experiments (Fig. 1). The fetal urogenital sinuses were subdivided into the ventral and dorsal halves and each half was separated into the epithelium and mesenchyme by collagenase treatment. The homotypic recombinations (ventral epithelium plus ventral mesenchyme [Ev/Mv] and dorsal epithelium plus dorsal mesenchyme [Ed/ Md]) and heterotypic recombinations (ventral epithelium plus dorsal mesenchyme [Ev/Md] and dorsal epithelium plus ventral mesenchyme [Ed/Mv]) were carried out and the recombinants were grafted beneath the kidney capsule of male hosts. After 4–5 weeks’ culture, they were fixed and processed for histological observations or for in situ hybridization. In all recombinants, the epithelium formed glandular structures characteristic for the adult prostate. In situ hybridization showed that the transcription of PSBP-C1 mRNAs was induced both in the ventral epithelium and dorsal epithelium when they were associated with the ventral mesenchyme (Ev/Mv and Ed/Mv, Fig. 5A and B). In contrast, neither the ventral epithelium nor the dorsal epithelium expressed detectable amount of PSBP-C1 mRNAs when they were associated with the dorsal mesenchyme (Ev/Md and Ed/Md, Fig. 5C and D). The percentage of the epithelial cells positive for PSBP-C1 mRNA expression was determined in every explants. About 65 % and 45 % of the epithelial cells were positive in the recombinants of Ev/Mv and Ed/Mv, respectively, while none of the cells were positive in the recombinants of Ev/Md and Ed/Md (Fig. 6).
Induction of PSBP-C1 mRNA in recombinants of fetal urogenital sinuses. (A) A ventral epithelium plus a ventral mesenchyme, (B) a dorsal epithelium plus a ventral mesenchyme, (C) a ventral epithelium plus a dorsal mesenchyme, (D) a dorsal epithelium plus a dorsal mesenchyme. The recombinants were grown for 4 to 5 weeks and processed for in situ hybridization. PSBP-C1 mRNA was expressed in the epithelium associated with the ventral mesenchyme (A, B) while no expression was observed in the epithelium associated with the dorsal mesenchyme (C, D). e, epithelium; 1, lumen; m, mesenchyme. Exposure period: 1 week. Bar, 50 μ m.
Induction of PSBP-C1 mRNA in recombinants of fetal urogenital sinuses. (A) A ventral epithelium plus a ventral mesenchyme, (B) a dorsal epithelium plus a ventral mesenchyme, (C) a ventral epithelium plus a dorsal mesenchyme, (D) a dorsal epithelium plus a dorsal mesenchyme. The recombinants were grown for 4 to 5 weeks and processed for in situ hybridization. PSBP-C1 mRNA was expressed in the epithelium associated with the ventral mesenchyme (A, B) while no expression was observed in the epithelium associated with the dorsal mesenchyme (C, D). e, epithelium; 1, lumen; m, mesenchyme. Exposure period: 1 week. Bar, 50 μ m.
Proportion of the cells expressing PSBP-C1 mRNA in recombinants. The number of silver grains was counted over the epithelial cells in a 4-day exposed autoradiograph. The cells possessing more than 30 silver grains were considered as positive. Values are expressed as means±standard deviations. The number of the explants used for cell count was shown in parenthesis above a column. Abbreviations for the recombinants were as follows; Ev/Mv, ventral epithelium plus ventral mesenchyme; Ed/Mv, dorsal epithelium plus ventral mesenchyme; Ev/Md, ventral epithelium plus dorsal mesenchyme; Ed/Md, dorsal epithelium plus dorsal mesenchyme; BE/Mv, adult bladder epithelium plus ventral mesenchyme; BE/Md, adult bladder epithelium plus dorsal mesenchyme; eBE/Mv, embryonic bladder epithelium plus ventral mesenchyme; eBE/Md, embryonic bladder epithelium plus dorsal mesenchyme.
Proportion of the cells expressing PSBP-C1 mRNA in recombinants. The number of silver grains was counted over the epithelial cells in a 4-day exposed autoradiograph. The cells possessing more than 30 silver grains were considered as positive. Values are expressed as means±standard deviations. The number of the explants used for cell count was shown in parenthesis above a column. Abbreviations for the recombinants were as follows; Ev/Mv, ventral epithelium plus ventral mesenchyme; Ed/Mv, dorsal epithelium plus ventral mesenchyme; Ev/Md, ventral epithelium plus dorsal mesenchyme; Ed/Md, dorsal epithelium plus dorsal mesenchyme; BE/Mv, adult bladder epithelium plus ventral mesenchyme; BE/Md, adult bladder epithelium plus dorsal mesenchyme; eBE/Mv, embryonic bladder epithelium plus ventral mesenchyme; eBE/Md, embryonic bladder epithelium plus dorsal mesenchyme.
PSBP-C1 mRNA transcription induced in the urinary bladder epithelium
It has been demonstrated that the fetal and adult urinary bladder epithelium (urothelium) could form prostate-like glands when combined with fetal urogenital mesenchyme (Cunha et al. 1980; Cunha et al. 1983; Suematsu et al. 1988). We investigated the transcription of PSBP-C1 mRNAs in the glandular epithelium induced from the urothelium. The epithelia separated from adult and fetal (day 16.5 of gestation) urinary bladder were associated with the ventral or dorsal mesenchyme of fetal urogenital sinuses (day 16.5 of gestation) and grafted underneath the renal capsule of male rat host. After 4–5 weeks’ in vivo culture, prostate-like glands were induced in all explants. In situ hybridization demonstrated that the glandular epithelium induced by the ventral mesenchyme showed the transcription of PSBP-C1 mRNAs (Fig. 7A, B and D).
Induction of PSBP-C1 mRNA in urinary bladder epithelium. (A) Adult bladder epithelium plus ventral mesenchyme; (B) higher magnification of A (indicated by arrow); (C) adult bladder epithelium plus dorsal mesenchyme; (D) fetal urinary bladder epithelium plus ventral mesenchyme. The recombinants were grown for 4 to 5 weeks and processed for in situ hybridization. The ventral mesenchyme can induce the mRNA transcription both in the adult and fetal bladder epithelium. However, only a small portion of the cells is positive (arrows in A, B, and right top in D). The dorsal mesenchyme never induces the transcription (C). e, epithelium; 1, lumen; m, mesenchyme. Exposure period: 1 week. Bar, 50 μm.
Induction of PSBP-C1 mRNA in urinary bladder epithelium. (A) Adult bladder epithelium plus ventral mesenchyme; (B) higher magnification of A (indicated by arrow); (C) adult bladder epithelium plus dorsal mesenchyme; (D) fetal urinary bladder epithelium plus ventral mesenchyme. The recombinants were grown for 4 to 5 weeks and processed for in situ hybridization. The ventral mesenchyme can induce the mRNA transcription both in the adult and fetal bladder epithelium. However, only a small portion of the cells is positive (arrows in A, B, and right top in D). The dorsal mesenchyme never induces the transcription (C). e, epithelium; 1, lumen; m, mesenchyme. Exposure period: 1 week. Bar, 50 μm.
However, the proportion of the positive cells was greatly reduced as compared with the recombinants with the sinus epithelium (Fig. 6); 2% for those with adult urothelium (Fig. 7A and B) and 15 % for those with fetal urothelium (Fig. 7D). Those positive cells tended to be clustered into groups rather than dispersed as single cells. The dorsal mesenchyme never induced PSBP-C1 mRNA transcripion in the adult or fetal bladder epithelium (Fig. 7C).
Discussion
PSBP-C1 mRNA expression in the ventral prostate
Our Northern analysis showed that PSBP-C1 mRNA was transcribed exclusively in the ventral prostate. The dorsal prostate, though it also develops from the fetal urogenital sinus and shows glandular structures similar to the ventral prostate, never expressed PSBP-C1 mRNA. Moreover, in situ hybridization analysis using cRNA of PSBP-Cl demonstrated that the mRNA for PSBP-Cl was transcribed in the epithelium but not in the mesenchyme. These results are in agreement with the previous works (Heyns and De Moor, 1977; Aumuller et al. 1982) and indicate that the transcription of PSBP-Cl gene can be used as a specific marker for the epithelium of the ventral prostate.
We then investigated the time course of PSBP-Cl mRNA production during the normal development of the ventral prostate by in situ hybridization. The transcription of PSBP-Cl gene was initiated between days 9 and 14 after birth, and reached maximum level by day 21. It is now believed that the prostatic buds start to differentiate into the functional secretory epithelium between days 10 and 14; the appearance of androgen receptors (Takeda et al. 1984; Zhang et al. 1988), the lumen formation within the buds and the establishment of secretory cytoplasmic organelles (Brandes, 1966), all happen at this stage. It can, therefore, be said that the onset of PSBP-Cl mRNA gene expression correlates with the cytodifferentiation of the epithelium of the ventral prostate and that PSBP-Cl can be used as a specific differentiation marker for the epithelium of the ventral prostate.
Epithelial–mesenchymal interaction induces the PSBP-Cl gene transcription in the prostatic epithelium and urothelium
Heterotypic morphogenesis has been reported in most cases but in some cases it was not accompanied by heterotypic cytodifferentiation. In the recombinant of embryonic mammary epithelium and salivary gland mesenchyme, the mammary epithelium shows the branching pattern characteristic of that of salivary gland but still synthesized milk proteins (Sakakura et al. 1976). Similarly, Yasugi et al. (1985) and Hayashi et al. (1988) have demonstrated that chick embryonic endoderm of the small intestine fails to synthesize embryonic chick pepsinogen under the influence of glandular stomach mesenchyme, despite its glandular-stomachlike morphogenesis. These results support the idea that the mesenchyme controls the morphogenesis of the epithelium, while the epithelium is the determinant of the type of cytodifferentiation to be expressed. In the present study, however, we showed that the cytodifferentiation of the prostatic epithelium is also controlled by its surrounding mesenchyme.
The mesenchyme from both dorsal and ventral part of the urogenital sinus could induce the glandular structures in the sinus epithelium. However, the transcription of PSBP-C1 gene, a specific marker for the ventral prostatic epithelium, was observed only when the epithelium was recombined with the ventral mesenchyme (Ev/Mv and Ed/Mv). The epithelium from both the ventral and dorsal parts of the sinus expressed PSBP-C1 under the influence of the ventral mesenchyme. By contrast, the ventral epithelium, which synthesized PSBP during normal development, failed to express PSBP-C1 mRNA when recombined with the dorsal mesenchyme (Ev/Md). These data demonstrate that the lobe-specific cytodifferentiation is determined by the surrounding mesenchyme and strongly suggest that, during the development of the prostate gland, the mesenchyme directs the cytodifferentiation as well as the morphogenesis of the epithelium.
In our experimental conditions, the proportion of the cells showing PSBP-C1 mRNA transcription did not reach 100 % even in the homotypic recombinants (Ev/ Mv). This could be explained by the fact that the mesenchyme from the ventral side was contaminated to some degree with a lateral lobe tissue. The mesenchyme of the lateral lobe might induce a lateral lobe in the recombinants and the lobe is negative for PSBP.
Although the ventral mesenchyme has the ability to induce PSBP-C1 mRNA, the induction is also dependent on the responsiveness of the epithelium. A significant difference was observed in the proportion of the cell positive for the PSBP-C1 mRNA between the ventral (Ev/Mv) and dorsal (Ed/Mv) epithelium recombinants (65 % and 45 %, respectively). Moreover, in the recombinant of adult bladder epithelium (urothelium) and ventral sinus mesenchyme, the proportion of the cells positive for PSBP-C1 mRNA was less than 2 % while it was increased to 15 % when fetal urothelium was used. These results suggest that, in addition to the inductive activity of the mesenchyme, there exist regional and temporal differences in the epithelial responsiveness to the mesenchymal induction. Although the urinary bladder epithelium and the urogenital sinus epithelium share a same developmental origin, the responsiveness of the bladder epithelium decreases dramatically with advancing age. However, it is worth noting that a few per cent of the cells in adult urothelium still keep their responsiveness to the ventral mesenchyme of the fetal sinus to transcribe de novo the ventral prostate-specific mRNA. In other words, the fully differentiated adult urothelium still contains a population of cells which, like embryonic cells, shows a great variety of developmental options. These cells might be involved in neoplastic or malignant transformation later in life. Also, it was observed that the positive cells in the glands induced from the bladder epithelium tended to form clusters, which indicates the clonal heterogeneity within the bladder epithelium with respect to its responsiveness to the mesenchymal induction.
Cunha et al. (1983) and Neubauer et al. (1983) showed by histochemistry and 2-dimensional gel electrophoresis that the adult urothelium undergoes full prostatic functional differentiation as well as morphogenesis when associated with embryonic sinus mesenchyme. Our previous experiments (Suematsu et al. 1988)have demonstrated by means of steroid autoradiography and SDS-PAGE that the induced urothelium has neither androgen receptor activity nor any prostatespecific proteins, suggesting that the urothelium is unable to attain full prostatic cytodifferentiation even under the influence of fetal sinus mesenchyme. The present results revealed that urothelium has a limited responsiveness to the sinus mesenchyme (prostatic inducer), supporting our previous experiments. Studies with more specific markers will be required to account for these discrepancies.
The establishment of regional organization of the male urogenital tract
In the course of fetal development, the urogenital sinus produces the prostate (ventral, lateral and dorsal lobes), urethra, bulbourethral glands and urethral glands. Sugimura et al. (1985) examined the regional difference in inductive activity of the fetal sinus mesenchyme. They divided fetal mouse urogenital sinuses into two parts, cranial half and caudal half, and observed that the mesenchyme from the cranial half induces prostate and urethral gland, while one from the caudal half induces urethra and occasionally urethral glands. In addition to the differences in the cranial and caudal parts, we found in the present study that there exists a difference in inductive activity between the dorsal and ventral sides of the sinus mesenchyme. These data strongly suggest that regional heterogeneity of inducing activity of the sinus mesenchyme plays a decisive role in the regional organization of the male urogenital tract. A similar concept has been proposed in the regional differentiation of the chick digestive tract (Yasugi and Mizuno, 1974; Takiguchi-Hayashi, 1989).
In conclusion, our present data clearly show that transcription of a particular prostate-specific gene is induced by epithelial–mesenchymal interaction in the prostatic epithelium and also in urinary bladder epithelium. Since gene structures of PSBPs have been extensively studied (Parker et al. 1988), the activation of PSBP during the development can be analyzed at the molecular level, much like the muscle-specific actin gene in early amphibian induction (Gurdon et al. 1989). This analysis will be a promising approach to the clarification of the mechanisms underlying tissue interaction during organogenesis.
ACKNOWLEDGEMENTS
This work was supported in part by grants-in-aid No. 01010057 and No. 02640568 from the Ministry of Education, Science and Culture of Japan and No. A6304 from the Ministry of Public Welfare of Japan. We are grateful to Dr M. G. Parker of the Imperial Cancer Research Fund Laboratories, London for the gifts of cDNAs for prostatic steroid binding protein. We thank Masa Takeda for help in preparing the manuscript.