| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Published online before print December 18, 2008
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
From the Centre for Urological Research,* Monash Institute of Medical Research, and the Australian Research Council Centre of Excellence in Biotechnology and Development, Monash University, Clayton, Australia; Prince Henry’s Institute, Clayton, Australia; and the Department of Neurobiology and Physiology, Northwestern University, Evanston, Illinois
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
Activin bioactivity, or bioavailability, can be regulated in several ways including assembly of intracellular heterodimers such as inhibin A (
βA) that once secreted block activin binding to its receptors.10
Follistatin-315 can bind to, and inactivate, activin A in the circulation, whereas cell surface-bound follistatin-288 diverts activin A to a lysosomal pathway for degradation.11
It is generally thought that the activin-βC subunit is not a significant regulator of activin bioactivity. This is because of limited expression of activin-βC mRNA and a lack of abnormalities in activin-βC-null mice.12 However, in the context of the null mouse, there may be functional redundancy with other transforming growth factor-β family members. If true, overexpression rather than underexpression is more likely to have physiological consequences.
We propose that activin C is an antagonist of activin A bioactivity. To test our hypothesis, we produced recombinant activin C and several lines of mice overexpressing the human activin-βC subunit. Because activin A has tissue-specific patterns of expression and biological action, we examined tissues known to be responsive to activin A: testis, liver, and prostate.13-15 In the testis, activin A is a paracrine regulator of Sertoli cell proliferation and spermatogonial maturation.5 In the liver, activin A inhibits hepatocyte DNA synthesis7 and induces apoptosis.6 In the prostate, activin A is a negative growth regulator and involved in branching morphogenesis during development.3
Our results demonstrated activin C antagonized the effects of activin A in vitro, and testis, liver, and prostate phenotypes consistent with our hypothesis were observed in vivo. Additionally, we show an increase in activin-βC immunoreactivity in human cancer tissue microarrays, thus demonstrating an analogous situation might occur in human pathology.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Human activin-βC (under the control of a CMV promoter)-overexpressing mice were produced at Mouseworks, Department of Physiology, Monash University, by standard methods. Three independent lines were established and crossed with wild-type (WT) C57BL/6 mice and heterozygous littermates to obtain single-heterozygous (SH) haploid, and double-heterozygous (DH) diploid, transgenes. Southern blot and semiquantitative polymerase chain reaction (PCR) were used to determine transgene copy number. SH1 had 
2 to 5 copies of the transgene, SH2, 5 to 10 copies, and, SH3, 20 to 30 copies (data not shown). A competitive genomic PCR screening strategy with primers specific for the incorporated human activin-βC and endogenous mouse activin-βC was used to identify SH- or DH-positive progeny and results were confirmed by breeding studies.
Tissue Collection
All animal handling and procedures were performed in accordance with National Health and Medical Research Council guidelines for the Care and Use of Laboratory Animal Act and according to the Animal Experimentation and Ethics Committee at Monash Medical Centre, Clayton, Australia. WT and TG mice were obtained from the same litters at age 14 to 16 weeks. Animals were euthanized by cervical dislocation after blood was collected by cardiac puncture. Organs were removed, wet weight recorded, and a portion immersion-fixed in Bouin’s for histology or stored at –80°C for RNA and protein extraction.
Histology
Fixed tissues were processed and embedded in paraffin for histological analysis. Serial sections were cut and dried onto Superfrost Plus-slides (Menzel-Glazer, Braunschweig, Germany) before examination by investigators blind as to genotype and overall hypothesis.
Expression of Activin C
CHO cells were stably transfected with full-length human activin-βC cDNA16 in the pCI-neo vector or empty vector as a control. Serum-free conditioned media from these stable cells were collected and concentrated and 0.1% bovine serum albumin was added for stability. CHO cell-conditioned media was assessed under nonreducing conditions for activin C expression by Western blot. Conditioned media was also analyzed under reducing and nonreducing conditions with a specific activin A subunit antibody (clone E4; Oxford Bio-Innovations, Oxford, UK) and a follistatin antibody (clone H1017 ) to determine whether activin-βA subunit proteins (βAβA, βAβC) or follistatin were detectable in the conditioned media. We could not detect activin-βA subunit proteins or follistatin in the conditioned media (data not shown). Concentration of activin C in conditioned media was determined by running samples run under reducing and nonreducing conditions on 12.5% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) along with known amounts of activin A (R&D Systems, Minneapolis, MN) followed by SYPRO Ruby staining (Bio-Rad, Hercules, CA).
Cell Culture and Activin Assays
The activin-responsive human prostate tumor cell line, LNCaP, was routinely cultured in RPMI (Gibco, Invitrogen, Grand Island, NY) with 10% heat-inactivated fetal calf serum (Hyclone, Logan, UT) and antibiotics (100 IU/ml penicillin and 10 µg/ml streptomycin) (Gibco) in 75-cm2 culture flasks at 37°C in a humidified atmosphere of 5% CO2 in air. For growth assays LNCaP cells were plated at a density of 5000 cells/well, in RPMI/5% fetal calf serum into 96-well plates for 24 hours and allowed to attach. Medium was removed and replaced with RPMI/2% fetal calf serum containing 50 ng/ml of activin C, 40 ng/ml of follistatin, buffer, or empty vector controls. After 6 hours, 10 ng/ml of activin A or buffer controls were added and incubated for a total of 72 hours. The CellTiter Aqueous One Solution cell proliferation assay (Promega Corp, Sydney, Australia) was used to determine the number of viable cells after 72 hours of treatment and percent change in cell number was calculated relative to media control.
Transient transfections were performed in the pituitary gonadotrope cell-line, LβT2. Cells were plated 1 day before transfection in 24-well plates and transiently transfected with LipoFectamine Plus (Invitrogen, Carlsbad, CA) for a 338rFSHβ promoter construct18 or in Opti-MEM (Invitrogen) with a CAGA (12)-lux, a luciferase reporter encoding 12 copies of the CAGA canonical Smad DNA-binding sequence.19 All experimental treatments were performed in phenol-free, serum-free media. Empty vectors of reporter and expression plasmids were used as controls and to balance DNA where necessary. Cells were treated with control media, activin A (10 ng/ml), increasing concentrations of CHO cell-expressed activin C (25 to 200 ng/ml), or matched volumes of empty vector control CHO cell-conditioned media (62.5 to 500 µl). Cells were lysed and luciferase activity measured for 30 seconds using an AutoLumat (Berthold Technologies Co., Oak Ridge TN) as previously described.18,19 At least three independent experiments were performed in triplicate.
RNA Purification and RT-PCR
Total RNA was extracted from testis and liver using the RNeasy mini kit (Qiagen Pty. Ltd., Victoria, Australia). Genomic DNA contamination was removed and total RNA reverse-transcribed using Superscript II reverse transcriptase (Roche Diagnostics, Basel, Switzerland). Real-time PCR was performed using the Roche LightCycler and SYBR Green protocol (Roche Diagnostics). Transgene expression was quantified by expressing cDNA (human activin-βC, mouse activin-βC, mouse activin-βA) as a ratio of β-actin.
Serum Assays
Circulating activin A concentration was measured using a specific enzyme-linked immunosorbent assay (Oxford Bio-Innovations) as previously described.20
The intra- and interassay coefficients of variation were 5.0% and 10.9%, respectively, and the limit of detection for the assay was 10 pg/ml (four assays). Plasma concentrations of follicle-stimulating hormone (FSH) were measured by radioimmunoassay using mouse FSH as the standard. The assay sensitivity was 1.3 ng/ml, the mean ED50 was 6.4 ng/ml, and the intra- and interassay coefficients of variation were 7.2% and 5.4%, respectively (five assays). Follistatin was measured by radioimmunoassay, as previously described.21
The assay sensitivity was 1.2 ng/ml, the mean ED50 was 10.1 ng/ml, and the intra- and interassay coefficients of variation were 9.2% and 7.9%, respectively (four assays). The inhibin radioimmunoassay used iodinated 31-kDa bovine inhibin (no. 1989) as tracer and a rat ovarian extract as a standard.22
No significant cross-reactivity is detected with activin A, however, the assay cross-reacts with pro-c. Consequently, inhibin levels are reported as immunoreactive total inhibin. Therefore, concentrations need to be interpreted as including cross-reacting inhibin species, which may not be biologically active. The assay sensitivity was 0.12 ng/ml, the mean ED50 was 1.13 ng/ml, and the intra- and interassay coefficients of variation were 10.6% and 6.0%, respectively (four assays).
SDS-PAGE and Western Blot
Serum activin-βC subunit protein was detected by reducing SDS-PAGE and Western blotting with a specific monoclonal activin-βC subunit antibody (clone 1)13-15 at 0.5 µg/ml. Equal volumes of urea-treated (8 mol/L), albumin-stripped (Blue Sepharose; Amersham Biosciences, Uppsala, Sweden) serum were loaded onto 12.5% gels and Western blot was performed as previously described.14,23 Treated LNCaP cells were lysed in RIPA buffer containing protease and phosphatase inhibitors, and the BCA assay was used to determine total protein concentration. Thirty µg of treated LNCaP supernatant were assessed for phosphorylated Smad-2 activity (no. 3101; Cell Signaling Inc., Beverly, MA), Smad-2 (no. 3122, Cell Signaling), and Smad-4 (no. 9515, Cell Signaling), Total Smad-2 or GAPDH (ab9484; Abcam, Cambridge, UK) was assessed as a loading control and intensity of bands was assessed with Scion Image (National Institutes of Health, Bethesda, MD).
Liver Enzymes
Serum albumin was assayed in duplicate using a 96 well protocol (Bio-assay Systems CA, USA) and serum L-alanine:2-oxogluterate aminotransferase (ALT) was analyzed using an Infinity ALT liquid stable reagent (Thermo Electron Corp., Waltham, MA).
Immunohistochemistry
Proliferating cell nuclear antigen (PCNA) (PC10; DAKO, Kingsgrove, Australia), activated caspase-3 (Asp175, Cell Signaling), activin-βC (clone-115 ), and Smad-2 (no. 3122, Cell Signaling) were detected using the DAKO Autostainer universal staining system (DAKO A/S, Glostrup, Denmark). Microwave antigen retrieval for PCNA was performed in 0.01 mol/L citrate buffer (pH 6.0), activin-βC in 0.1 mol/L glycine buffer (pH 4.5), caspase-3, and Smad-2 in 0.01 mol/L sodium citrate buffer (pH 6.0). Sections were treated with CAS blocking reagent (DAKO, Kingsgrove, Australia), antibodies detected with avidin-biotin complex (ABC) and color-reacted with diaminobenzidine (DAKO, Kingsgrove Australia).
Fertility Assessment
Male mice (WT, SH, and DH) of varying ages (6 to 30 weeks) were cohabited with 10- to 12-week female C57BL/6 WT mice (1:1). Females were checked daily for the presence of copulatory plugs. Females with positive plugs were separated from males, and the number of pups per plug was recorded. Daily sperm production (DSP) was determined using a frozen left testis from WT, SH, and DH mice by a procedure that has been previously described.24
Sperm Motility Analyses
Sperm were collected and capacitated in modified Tyrode’s medium without sodium pyruvate or sodium lactate. Sperm from 23 mice were collected at 14 to 16 weeks (n = 10 WT, 5 SH2, and 9 DH2). Both caudae epididymidis were dissected from the mice and a small slit was made in each before transfer to a 1.5-ml tube containing 1 ml of pre-equilibrated modified Tyrode’s medium and incubation in 5% CO2 in air at 37°C for 10 minutes to allow the sperm to swim out from the tissue. After incubation the caudae were removed and a 10-µl aliquot of the sperm suspension placed into an 80-µm chamber (Hamilton-Thorne Biosciences, Beverly, MA). Replicates (two per mouse) were analyzed using a Hamilton-Thorne IVOS computer-assisted sperm analyzer (Hamilton-Thorne Biosciences). At least 1000 sperm were analyzed per replicate. Analyses were performed using the recommended settings for mouse sperm. Data were arcsin-transformed and then subjected to general linear analysis, and the difference between means was determined by Tukey’s HSD test (SPSS for Windows; SPSS, Chicago, IL).
In Situ Detection of DNA Fragmentation in Testis Sections
Apoptosis in testis sections was analyzed by the ApopTag in situ apoptosis detection kit (Chemicon Int., Temecula, CA).25
Testis Stereology
Slides were masked before quantitation to facilitate unbiased counting. PCNA-positive cell types were identified based on their location within the tubule, their size, and the shape of the cell nucleus. Apoptotic cells were identified by deep brown nuclear staining and included spermatogonia, spermatocytes, and spermatids. Two sections per mouse were examined. Each tubule cross section was classified in one of three stage groupings (XII to IV, V to VIII, and IX to XI).
Cell Number Estimates
The optical dissector (sic) stereological method26 was used to determine the total number of cell nuclei per testis. All measurements were performed using a x100 objective on an Olympus BX-50 microscope (Olympus, Tokyo, Japan). A microcator (D 8225; Heideinhain, Traunreut, Germany) that monitored scanned depth was attached to the microscope stage. Fields were selected by a systematic uniform random sampling scheme as previously described26 with the use of a motorized stage (Multicontrol 2000; ITK, Lahnau, Germany).
Liver and Prostate Proliferation and Apoptosis
The incidence of proliferation (PCNA)- and apoptosis (activated caspase-3)-positive cells in tissue sections was estimated based on a method that allowed an unbiased semiquantitation of the percentage of positive cells in TG and WT samples. Masked tissue sections were mapped at x10 magnification to define tissue boundaries. Random fields were systematically selected by CAST V1.10 software and sampling was conducted using an unbiased counting frame. Frame counting was performed on sections uniformly spaced throughout the tissue, 150 frames, and x100 magnification per section, n = five triplicate sections with an average of 1000 cells counted per section.
Localization of Smad-2
Tissue sections were masked and the incidence of nuclear localization of total Smad-2 in testis, liver, and prostate sections was estimated as described above. Frame counting was performed on five to eight duplicate sections, 150 frames, x40 magnification, with an average of 1000 cells counted per section.
Cancer Tissue Microarrays
Activin-βC subunit protein was assessed in normal human and cancer tissue arrays with one example of each tissue and tumor type on each array (n = 2; SuperBioChips Laboratory, Seoul, Korea) using a specific monoclonal antibody (clone 1) as previously described.15
Statistical Analysis
TG and WT littermate controls were compared using analysis of variance with Dunnett’s posthoc test and the significance threshold used at a level of 5% (GraphPad Software, Inc., San Diego, CA).
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
CHO cells were stably transfected with full-length human activin-βC cDNA in the pCI-neo vector or empty vector (EV) control. CHO cell-conditioned media containing activin C (50 ng/ml) was added to the activin-responsive LNCaP prostate cell line and cell growth-assessed (Figure 1A)
. Empty vector conditioned media or media only served as controls. Empty vector or activin C alone did not alter cell number compared to media control, indicating that activin C alone had no effect on cell growth. As expected, activin A (10 ng/ml) decreased cell number by 60% in both media and EV controls (P < 0.001), whereas activin A only reduced growth by 30% in the presence of activin C-conditioned media (P < 0.01 versus media and EV + activin A controls), indicating that activin C antagonized the growth inhibitory effects of activin A. Again, as expected, follistatin, a well-characterized activin binding protein, antagonized the growth inhibitory effects of activin A with cell numbers returning to 80% of control. Addition of follistatin and activin C together attenuated this effect with values rising to 110% of media control (P < 0.01 versus media + activin A + follistatin and EV + activin A + follistatin), which implies that antagonism of activin A is likely to be via different mechanisms.
|
. Activin C- conditioned media (50 ng/ml) decreased activin A-induced Smad-2 phosphorylation by 50% (P < 0.001 versus media and EV + activin A), indicating that activin C antagonism of activin A-induced signaling is as potent as follistatin. Results are presented relative to total Smad-2 (Figure 1B)
. There was no significant difference in total Smad-2 protein levels with treatment, indicating that the decreases in phosphorylated Smad-2 were not related to total Smad-2 levels. Activin A (10 ng/ml) increased Smad-4 protein expression (P < 0.01 versus media and empty vector controls. Activin C (50 ng/ml) or empty vector-conditioned media did not alter Smad-4, however, activin C-conditioned media (50 ng/ml) decreased activin A-mediated Smad-4 protein expression (P < 0.01) as effectively as follistatin (40 ng/ml, Figure 1C
). Activin C Antagonized Activin A-Stimulated Activity of the Rat FSH-β Promoter
To confirm results in an independent activin A assay, we assessed the effects of activin C- or empty vector-conditioned media on a rat FSH-β promoter. Activin C-conditioned media (25 to 200 ng/ml) or an equal volume of empty vector control (62.5 to 500 µl) had no effect in the FSH-β promoter assay, whereas activin A (10 ng/ml) increased rat-FSH-β promoter activity (P < 0.001, Figure 1D
). Addition of activin C-conditioned media (25 to 200 ng/ml) plus activin A (10 ng/ml) significantly decreased activin A-stimulated rat-FSH-β promoter activity (50 ng/ml, P < 0.05; 100 to 200 ng/ml, P < 0.01), whereas an equal volume (62.5 to 500 µl) of empty vector control had no effect (Figure 1E)
.
CAGA Smad DNA-Binding Sequence
To confirm that activin C had no independent effect on Smad signaling we assessed if activin C-conditioned media activated or antagonized an extremely sensitive CAGA Smad DNA-binding sequence.18
As expected, activin A (10 ng/ml) activated the CAGA reporter (P < 0.001), whereas increasing concentrations (25 to 200 ng/ml) of activin C or matched volumes of empty vector control-conditioned media (62.5 to 500 µl) had no effect (Figure 1F)
, thus indicating that activin C had no independent effect on the CAGA reporter.
Overexpression of Activin-βC in Vivo Causes an Age-Related Decrease in Male Fertility
Based on previous reports on the expression of activin-βC in the testis13,14
and the known importance of activin A for male fertility,5
we examined testis histology and fertility of TG mice compared to age-matched WT littermates. Hypospermatogenesis, characterized by the presence of normal and abnormal testis tubules in close proximity, was evident in TG testis sections (Figure 2A)
. Similar histology was recorded in all TG lines.
|
, shows a significant decrease in litter size when double-heterozygous (DH) or single-heterozygous (SH) TG male mice were mated with WT females, compared with WT-WT matings. Significant decreases in number of pups per plugged mating were evident in all DH lines (P < 0.01 within 10 to 12 weeks). SH1 showed a significant decrease in litter size compared with WT age-matched males from 19 to 21 weeks (P < 0.01, Figure 2B
); SH2 from 13 to 15 weeks (P < 0.01, Figure 2C
) and SH3 from 16 to 18 weeks (P < 0.01, Figure 2D
). Mechanisms Underlying Decreased Litter Size DSP
To examine the mechanism underlying decreased litter size DSP was assessed in mice 14 to 16 weeks of age. Overexpression of activin-βC decreased DSP by 31%, 59%, and 39% in DH1, DH2, and DH3 and by 58%, 63%, and 19% in SH lines 1, 2, and 3, respectively (Figure 3A)
.
|
To determine whether reduced sperm output was a result of alterations in germ cell proliferation and/or survival we quantified apoptosis and proliferation in stage-specific cross-sections of testis tubules in DH2 mice 14 to 16 weeks of age compared to age-matched WT littermates. DH2 mice were chosen because there were no confounding serum FSH effects (see serum proteins below). The percentage of tubules containing apoptotic germ cells (TUNEL-positive) was elevated by 180% and 130% in stages V to VIII (P = 0.001) and IX to XI (P = 0.04) (Figure 3B)
in DH2 males, respectively, compared with age-matched WT littermates. In contrast to apoptosis, there was no significant change in percentage of proliferating (PCNA-positive) germ cells (data not shown).
To determine whether the alteration in apoptosis affected a specific germ cell type, we used stereological methods to quantitate the number of germ cells and Sertoli cells in tubular cross sections from DH2 mice compared to age-matched WT littermates. The number of Sertoli and germ cells remained at control levels; although it was noted that pachytene spermatocytes were lower compared to WT controls (66%, P = 0.069; see Supplementary Table S1 at http://ajp.amjpathol.org).
Compromised Sperm Motility and Function
Sperm was collected from the caudae epididymides and motility characterized using a computer-assisted sperm analyses system. Sperm from TG mice showed a significant reduced percentage of motile sperm compared to age-matched WT littermates (DH2, 54.2%; SH2, 49.3%; WT, 69.2%; P < 0.05 versus WT). Further, of those sperm that were motile, a significant percentage displayed reduced sperm velocity and beat cross frequency (Figure 3C
; see Supplementary Table S2 at http://ajp.amjpathol.org).
Liver Inflammation in Activin-βC-Overexpressing Mice
Overexpression of activin-βC increased total liver weight and percentage of body weight in DH1 to DH3 and SH3 mice (Figure 4A)
. A low-power example of WT and TG liver section is presented in Figure 4, B and C
. In all TG livers, foci of inflammatory cells were observed primarily in zones 1 and 2, which occasionally extended into zone 3. Infiltration consisting primarily of neutrophils and macrophages indicated regions of acute inflammatory response (Figure 4D)
. There was occasional Kupffer cell activation and regions of necrosis (Figure 4E)
, without evidence of hepatic stellate cell activation or fibrosis. There was increased evidence of larger hepatocyte nuclei and dividing hepatocytes in livers of all TG lines.
|
Histological assessment of PCNA- and caspase-3-stained liver sections in TG mice showed increased hepatocyte proliferation and apoptosis compared to WT littermate controls (Figure 4F)
. The combined effect of these two parameters (evidenced by analysis of the ratio of proliferation to apoptosis) indicated activin-βC overexpression influenced hepatocyte proliferation to a greater extent than apoptosis. This observation was most apparent in DH2 mice; the ratio of proliferation to apoptosis in WT was 1.3, SH2 was 2.6, and DH2 was 3.1.
Prostate Hypertrophy and Epithelial Cell Hyperplasia with Activin-βC Overexpression
A significant increase in anterior prostate weight (AP) was evident in TG mice compared to WT litter-mate controls (SH3, 130% compared to WT, P < 0.05; DH1, 148%, P < 0.01; and DH2, 130%, P < 0.05). An increase in ventral prostate (VP) weight was evident (DH1, 165%, P < 0.01; DH2, 162%, P < 0.01). Lateral prostate (LP) weight increased (SH3, 167%, P < 0.0); DH2, 157%, P < 0.01; and DH3, 155%, P < 0.01). There was no change in dorsal prostate (DP) weight. WT VP sections were mostly composed of a single layer of epithelial cells lining a lumen (Figure 5A)
. Diffuse luminal epithelial cell hyperplasia, characterized by increased stratification in the form of tufting and papillary in-folding and a reduction in luminal size, was evident in all DH lines (Figure 5B)
. No atypical nuclei were observed in SH TG mice; however, there was evidence of atypical nuclei in DH VP sections; including nuclear enlargement (one to three per high-power field), increased prominence of nucleoli (one to three per high-power field), and cribriform structures (Figure 5C)
. Epithelial cell mitosis was more frequently observed in DH VP sections (one to four cells per high-power field, Figure 5D
), whereas mitotic cells were not observed in WT VP sections. Assessment of the percentage of proliferating (PCNA-positive) and apoptotic (activated caspase-3-positive) epithelial cells in 14- to 16-week-old WT and DH2 TG mice showed a significant increase in PCNA-positive cells (Figure 5E)
and no difference in epithelial cell apoptosis (Figure 5F)
.
|
To determine whether phenotypes observed with activin-βC overexpression in the testis, liver, and prostate were related to reduced Smad-2 signaling in vivo we quantified nuclear localization of total Smad-2 in TG and WT littermate controls. Significant reductions in nuclear localization of total Smad-2 were evident in the testis, liver, and prostate of activin-βC-overexpressing mice (P < 0.01; Figure 6, A–C
). Figure 6, D to F
, shows representative examples of Smad-2 staining in WT and TG testis, liver, and ventral prostate, respectively.
|
To quantify TG gene expression and to determine whether phenotypic changes were related to changes in endogenous mouse activin-βC or activin-βA gene expression, real-time RT-PCR was performed. Transgene (human activin-βC) mRNA expression was evident in the testis, liver, and prostate of all TG mice. Founder line 2 mice showed the highest total activin-βC mRNA expression in the testis and VP and line 3 in the liver (Figure 7A)
. Overexpression of human activin-βC did not alter endogenous mouse activin-βC mRNA expression (Figure 7B)
. Transgene expression did not alter activin-βA mRNA levels in the testis or prostate; however, a significant increase in activin-βA mRNA expression was evident in the liver (Figure 7C)
. Therefore tissue-specific phenotypes were not related to changes in endogenous activin-βC mRNA expression or a reduction in activin-βA mRNA. Consistent with the elevated activin-βC mRNA levels an increase in activin-βC subunit protein was evident in the serum of TG mice (Figure 7D)
.
|
Activin A
To determine whether overexpression of activin-βC lowered synthesis of activin A (βAβA homodimer), serum levels of activin A were measured by enzyme-linked immunosorbent assay. All TG mice overexpressing activin-βC showed lower serum activin A (Table 1)
compared to age-matched WT littermates, reaching significance in SH1 (60% of WT), and DH2 (48%), and DH3 (43%).
|
We investigated if increased activin-βC subunit expression altered FSH secretion. Serum FSH levels were significantly decreased in DH1 and SH1 (DH1, 69%; and SH1, 73% of control; P < 0.05) and 3 (DH3, 69%; and SH3, 75%; P < 0.05) (Table 1)
. There was a significant increase in luteinizing hormone (LH) compared to age-matched WT controls in DH1 (186%, P < 0.001) and DH2 lines (143%, P < 0.01) (data not shown).
Inhibin and Follistatin
Total serum inhibin increased in TG lines DH1 to DH3 x 191%, 157%, and 142%, respectively, and SH3 x 134% compared to age-matched WT controls (Table 1
, P < 0.005). Follistatin levels were significantly increased in lines DH2 (162%) and SH3 (156%, P < 0.005) when compared to WT controls (Table 1)
.
Liver Enzymes
Total serum albumin, a marker of liver function, decreased by 18% in all DH lines (P < 0.05), whereas ALT levels, a marker of liver inflammation, increased twofold to threefold in DH1 to DH3 and SH3 (P < 0.005) compared to WT controls (Table 1)
.
Up-Regulation of Activin-βC Subunit Immunoreactivity in Human Testis, Liver, and Prostate Cancer
To determine whether activin-βC subunit proteins were altered in human disease, we assessed activin-βC subunit protein expression in normal and cancer tissue arrays with one example of each tissue and tumor type on each array (n = two) using a specific activin-βC monoclonal antibody.15
Increased activin-βC immunoreactivity was evident in a range of human cancers including testis (two of two), liver (two of two), and prostate (two of two) when compared to normal controls (Figure 8, A, C, and E
, normal testis, liver, and prostate, respectively; Figure 8B
, testis seminoma; Figure 8D
, hepatocellular cancer; and Figure 8F
, prostate cancer). Nuclear localization of activin-βC subunit protein was evident in normal testis, testis seminoma, and hepatocellular cancer. Nuclear localization of activin-βC has previously been shown14,27
and may indicate a novel regulatory mechanism that warrants further study.
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Activin is involved in development and is a potent growth and differentiation factor in the adult. Several means of regulating bioactivity are known to occur. Based on our findings, activin C should be regarded as another important regulator of activin A. Although there are some similarities to other means of regulating activin, there are more notable differences, thus adding further evidence of the significance of activin-βC to activin biology.
Well-characterized regulators of activin biology are the inhibin -subunit and follistatin. The testis phenotypes in mice overexpressing the inhibin- subunit or follistatin differ to that of the activin-βC-overexpressing mice. Whereas overexpression of activin-βC resulted in a progressive age-related decrease in male fertility, a stage-specific increase in germ cell apoptosis, decreased sperm production, and poor motility in those sperm that were produced, overexpression of inhibin- decreased testis size because of a reduction in seminiferous tubule volume. Inhibin--overexpressing males were fertile, despite a 50% reduction in sperm number and there were no abnormalities in sperm motility.28,29
In follistatin-overexpressing mice, Leydig cell hyperplasia, decreased Sertoli cell number and seminiferous tubular degeneration were evident.30
The BK mouse, in which the activin-βB allele is knocked into the βA locus, is another model that has been used to study the consequences of perturbing activin A levels in vivo. In this model activin-βB functions as a hypomorphic version of βA, making it similar to the mouse model described herein. BK mice survive to adulthood and display a delayed onset of fertility in males.31 Whether or not BK mice have compromised DSP or sperm motility is not known. However, all models underscore the importance of activin A for normal male fertility. The activin-βC TG mice described herein identify a novel regulatory mechanism in the testis, which has not been previously appreciated.
In liver, activin-βC overexpression induced liver inflammation, which was observed microscopically and confirmed by increases in serum ALT. Synthetic function of the liver was decreased as determined by serum albumin production. These changes were associated with hyperplastic livers attributable to an alteration in hepatocyte cellular turnover that favored proliferation. No liver phenotypes were reported in the inhibin-- or follistatin-overexpressing mice, but when follistatin is overexpressed in inhibin knockout mice, inflammation in the liver also occurs.32
This is the only feature in common with the activin-βC-overexpressing mice and concurs with our view that a decline in activin A is associated with liver inflammation.
Activin family members are known to significantly influence hepatic homeostasis. Activin A is an inhibitor of hepatocyte DNA synthesis6 and is involved in hepatic pathologies such as viral hepatitis33 and fibrosis.34 Activin-βC mRNA is down-regulated after partial hepatectomy in the rat, suggesting that it may be a negative regulator of liver cell growth.35,36 However, the role of activin-βC in the liver is controversial, with studies suggesting the activin-βC subunit inhibits DNA synthesis of hepatic cells in vivo37 and in vitro.38 Wada and colleagues39,40 demonstrate that the activin-βC homodimer is growth promoting in vitro and in vivo, and we found increased activin-βC subunit immunoreactivity associated with mitotic hepatocytes in regenerating rat liver.23 In our mouse model, overexpression of activin-βC increased hepatocyte cell proliferation to a greater extent than apoptosis, indicating activin-βC is a positive growth regulator in the liver, thus adding further evidence to our hypothesis that the activin-βC subunit antagonizes the growth inhibitory effects of activin A in the liver.
Prostate hypertrophy was evident in the activin-βC-overexpressing mice and has not been reported in the inhibin-- or follistatin-overexpressing mice. Activin is however an important regulator of prostate growth and differentiation, particularly during development, and is involved in branching morphogenesis.3
Thus it is perhaps not surprising that if activin-βC is a functional antagonist of activin A, that prostate pathologies develop. Prostate hypertrophy, epithelial cell hyperplasia, increased evidence of activin-βC subunit immunoreactivity in human prostate cancer, and antagonism of the growth inhibitory effects of activin A in the LNCaP prostate cell line implicate a role for the activin-βC subunit in maintenance of tissue homeostasis in the prostate.
Our results in relation to the bioactivity of activin C differ from a previous study, which showed activin C did not antagonize activin A in LNCaP cells.16 The preparations of activin C arise from different cell lines; the current study uses a mammalian cell line (CHO cells), whereas the earlier work used Noctuidae insect larvae infected with a recombinant baculovirus.41 We believe these data might demonstrate the importance of expressing proteins in a mammalian cell line that has the capability to correctly process the final product.
The activin-βC knockout mouse shows no overt phenotype even when partial hepatectomy was performed.12 Lack of phenotype in a knockout model may reflect compensation by other growth factors.12,42 This has previously been demonstrated for another transforming growth factor-β superfamily member, GDF15/MIC-1. Like activin-βC, GDF-15/MIC-1 mRNA is predominately expressed in the liver and is up-regulated after liver injury, yet homozygous null mice show no abnormalities even when subjected to liver insult. These data lead to the conclusion that GDF-15 does not appear to be essential or that loss of GDF15 may be compensated by other secreted growth factors.43 Recently, GDF15/MIC-1 was shown to be low in normal tissues and up-regulated in cancers, and an overexpression mouse model demonstrated GDF15/MIC-1 is a central regulator of appetite and weight and is a potential target of cancer-associated weight loss as well as obesity.44
We have shown in independent assays that activin C antagonizes activin A in vitro. Pathologies in activin A-responsive tissues and a reduction in nuclear localization of Smad-2 with activin-βC overexpression in vivo indicates antagonism of activin signaling also occurs in vivo. Therefore, our study shows for the first time that overexpression of activin-βC in vivo alters testis, liver, and prostate tissue homeostasis and strongly suggests that the mechanism of action is antagonism of activin A signaling. Further mechanisms by which the activin-βC subunit antagonizes activin A, for instance the production of an activin AC heterodimer,16 will be the focus of future investigations. The relevance of our results to human pathology is implied by evidence of increased activin βC-subunit immunoreactivity in human cancer arrays. The results of our study provide sufficient justification to warrant efforts to obtain new reagents to determine which activin-βC ligands form in vivo and their precise biological functions. Our results will have profound implications for activin biology and should provide stimulus for a new body of work that are relevant to the biology of mice and men.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported by the National Health and Medical Research Council (program funding to G.P.R. and M.K.O.).
Supplemental material for this article can be found on http://ajp.amjpathol.org.
Accepted for publication September 22, 2008.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  C. L. Borg, K. M. Wolski, G. M. Gibbs, and M. K. O'Bryan Phenotyping male infertility in the mouse: how to get the most out of a 'non-performer' Hum. Reprod. Update, March 1, 2010; 16(2): 205 - 224. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. A. Cowin, E. Gold, J. Aleksova, M. K. O'Bryan, P. M. D. Foster, H. S. Scott, and G. P. Risbridger Vinclozolin Exposure in Utero Induces Postpubertal Prostatitis and Reduces Sperm Production via a Reversible Hormone-Regulated Mechanism Endocrinology, February 1, 2010; 151(2): 783 - 792. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
