| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Other Original Articles |
Pediatric Endocrinology, University Children’s Hospital (M.W., I.M., B.S.S., D.D., H.A.W., M.B.R.), D-72076 Tubingen, Germany; Department of Internal Medicine, University of Bonn (D.S., G.R., H.-U.S.), 53111 Bonn, Germany; and Department of General Zoology and Endocrinology, University of Ulm (W.W., K.-D.S.), 89069 Ulm, Germany
Address all correspondence and requests for reprints to: Martin W. Elmlinger, Ph.D., Pediatric Endocrinology, University Children’s Hospital, Hoppe Seyler Strasse 1, D-72076 Tubingen, Germany. E-mail: martin.elmlinger{at}med.uni-tuebingen.de
Abstract
The action of androgen by way of the AR is required for the development of male gonads and external genitalia. The interplay between androgens and the somatotropic axis, in particular the IGFs in sexual development, is currently under thorough investigation. The IGF system is thought to mediate the androgen action in androgen-responsive cells.
To investigate the interaction of androgens with the IGF system, we compared the expression of IGFs and IGF-binding proteins in cultured genital skin fibroblasts from nine patients with the syndrome of complete androgen insensitivity with that in genital skin fibroblasts from 10 normally virilized males. Mutations in the AR gene and/or abnormalities of the AR protein in the immunoblot were detected in all complete androgen insensitivity genital skin fibroblast strains. They caused a complete failure of DHT binding. RIA and RT-PCR demonstrated that the genital skin fibroblast strains expressed IGF-II, IGF-binding protein-2, and IGF-binding protein-3, but no IGF-I. Most strikingly, complete androgen insensitivity genital skin fibroblast strains produced significantly lower IGF-II (P < 0.001; 42.2 ± 9.7 vs. 106.9 ± 11.8 ng/mg protein) and IGF-II mRNA (P < 0.01, by RT-PCR) than control genital skin fibroblast strains. The production of IGF-binding protein-2 was also decreased (P < 0.03) in complete androgen insensitivity genital skin fibroblasts, whereas that of IGF-binding protein-3 did not differ. Furthermore, high levels of IGF-binding protein-5 mRNA were detected in all genital skin fibroblast strains, whereby the 28-kDa band in the ligand blot, probably representing IGF-binding protein-5, was more abundant in complete androgen insensitivity genital skin fibroblasts. Exposure of the genital skin fibroblasts to T (5 x 10-8 M) had only weak effects on the expression of IGFs and IGF-binding proteins.
In conclusion, although the mechanism underlying these differences requires further study, it is conceivable that in addition to the endocrine actions of IGF-I, IGF-II and IGF-binding protein-2, as local growth factors, are involved in the mediation of androgen action and growth of genital tissues.
CLINICALLY, THE COMPLETE androgen insensitivity syndrome (CAIS) is primarily indicated by a 46,XY karyotype, normal serum androgen levels, and bilateral testes in the pure female phenotype (1). In this form of pseudohermaphroditism, androgen resistance of the target tissue could in many cases be attributed to mutations in the AR gene (2, 3, 4). If androgen is absent or the AR is defective, neither internal nor external male genitalia are developed. Genital skin fibroblasts (GF) are valuable to study the interaction of the IGF system and androgens (5). As a new approach, the interactions of androgens and the IGF system were investigated quantitatively by comparison of GF from 9 CAIS patients and 10 healthy males in the present study.
In androgen-responsive cells, the AR undergoes a conformational change
upon binding of T or 5
DHT. This enables the AR to directly bind to
androgen-responsive DNA elements and thereby regulate
androgen-dependent genes. The subsequent molecular steps leading to an
androgen response, e.g. interaction with the IGF-I receptor
signaling, are largely unknown. It is well established that the IGFs
via the IGF-I receptor and their binding proteins (IGFBPs) regulate
growth, differentiation, and apoptosis in many cell types
(6, 7).
The role of the IGFs and its interaction with the functional AR in sexual development are currently under thorough investigation. From earlier studies there are indications that IGF-I is an endocrine modulator of androgen activity (8, 9, 10). For example, androgenic effects in males increase during puberty, probably because of elevated IGF-I production (11). In pubertal rats, blockade of the AR with the analog flutamide, a drug used in the therapy of prostate cancer, caused significant decreases in GH receptor mRNA, serum IGF-I levels, and body weight gain (12). Interestingly, prostate growth during fetal development of mice and development of prostate cancer are related to the interaction of IGF-I with the AR and the ER (13, 14). However, IGF-I via the IGF-I receptor was also found to mediate androgen-induced ovarian follicular growth (15) in female monkeys. Furthermore, the breakdown of sex steroid action and its interaction with the somatotropic axis in aging men and women is a prime target to study the pathology of degenerative disorders (16). IGF-II is known to have autocrine/paracrine effects in human fetal development and cell differentiation (17). Only recently was IGF-II shown to be a major regulator of the AR expression in prostate tissue (18). Furthermore, to date little is known about the role of the IGFBPs in mediating androgen’s effects (5).
In the present study the interplay between the action of androgen through the AR and the IGF system was investigated in cultured human GF as target cells of androgen action. To elucidate the roles of IGF-I, IGF-II, and IGFBP-2, -3, and -5 in mediating the androgen action there, we compared their expression and production quantitatively in 9 GF strains from patients with CAIS and 10 GF strains from normally virilized, healthy male controls. In an additional experiment we investigated the effects of T treatment on these parameters in the GF.
Materials and Methods
Source of human tissue
The GF used here were derived from foreskin samples of 10 normal males who had undergone circumcision for correction of phimosis and from labial biopsies of 9 patients with male pseudohermaphroditism (performed for diagnostic purposes), who were later diagnosed as having CAIS. The study was approved by the federal ethics committee in Germany.
Cell culture
The GF cultures were established from biopsies essentially as previously described (19, 20) and were used between passages 4 and 10. In brief, cells from stock dishes were seeded on d 0 at a concentration of approximately 2 x 105 cells in dishes 10 cm in diameter and cultured at 37 C in a humidified atmosphere of 95% air and 5% CO2. After formation of confluent monolayers (d 6), the cell culture medium [MEM (Life Technologies, Inc., Eggenstein, Germany) plus 10% FCS (Cytogen, Berlin, Germany)] was removed and the cells were washed twice with PBS (Biochrom, Berlin, Germany) and incubated for 24 h with serum-free medium. On d 7 the medium was removed, and the monolayers were cultured for 48 h in serum-free MEM containing 5 x 10-8 M T (Serva, Heidelberg, Germany). The controls were treated with the MEM solution containing an equal amount of ethanol, the solvent used to dissolve T.
Analysis of AR
Specific binding of
[1,2,4,5,6,7-3H]5DHT to the AR was assessed
in confluent GF monolayers as previously described (20).
Immunoprecipitation and Western immunoblot analysis were carried out as
described previously (21), and mutations in the AR gene of
the CAIS GF were analyzed (4).
Measurements of IGFs and IGFBPs
The amounts of IGF-I, IGF-II, IGFBP-2, and IGFBP-3 secreted by each cell line within 48 h in culture in conditioned medium (CM) were analyzed by specific RIAs (Mediagnost, Tubingen, Germany). IGFBP-1 was measured by a validated in-house assay with a detection limit of 0.2 ng/ml. IGFBP-1 from human amniotic fluid was used as the standard. As no IGFBP-1 was detectable, data are not shown. The precision value, expressed as the coefficient of variation of the IGFBP-1 and -2 (22) and IGFBP-3 RIAs was below 8% for intraassay and below 9.4% for interassay variations. The sensitivity of all the IGFBP assays was less than 0.2 ng/ml.
Western ligand blotting
The pattern of IGFBPs was examined semiquantitatively by Western ligand blotting as previously described (23). After nonreducing 12% SDS-PAGE, proteins were blotted on nitrocellulose and incubated with 30,000 cpm/ml [125I]IGF-II. IGF-II-binding bands were made visible by autoradiography, and their molecular masses were estimated by comparison with mol wt protein standards (Rainbow marker, Bio-Rad Laboratories, Inc., Munich, Germany). The IGFBP-1 band was expected to occur at about 25 kDa, the IGFBP-2 band at about 31 kDa, the IGFBP-3 band at 42–44 kDa, and the IGFBP-5 band at about 28 kDa.
Semiquantitative RT-PCR analysis
For analysis of the levels of IGF-II, IGFBP-2, IGFBP-3, and IGFBP-5 mRNA, total RNA was extracted from approximately 106 cells and analyzed using semiquantitative RT-PCR (21). As a protein corresponding approximately to the molecular mass of IGFBP-5 was observed in the ligand blot, and no immunoassay to measure IGFBP-5 was available, we measured IGFBP-5 mRNA. The level of glyceraldehyde phosphate dehydrogenase (GAPDH) mRNA was used as internal standard for gene expression. IGF-II primers spanned from exons 7–8, which are comprised in all IGF-II transcripts. Primers were: sense, 5'-CTGGAGACGTACTGTGCTACCCCC-3'; and antisense, 5'-GTGTCATATTGGA AGAACTTGCCC-3' (amplicon 115 bp). The IGFBP-2 and GAPDH primers used were previously described (21). The IGFBP-3 primers were: sense, 5'-AGTGAGTCGGAGGAAGACC-3'; and antisense, 5'-GAGAACTTCTGGGTATCTGTGC-3' (amplicon 192 bp). The IGFBP-5 primers were: sense, 5'-CTCAACGTTGCTGCTGTCGAAG-3'; and antisense, 5'-CTAAGAGAAGATGGTGTTGCTC-3' (amplicon 168 bp). After PCR, amplicons were electrophoresed and quantitated densitometrically as previously described (22).
Northern blot analysis
Total RNA was isolated from approximately 2.5 x 106 fibroblasts as previously described (22). Twenty micrograms of each RNA preparation were incubated (30 min at 37 C) with deoxyribonuclease I (Roche, Mannheim, Germany) separated through a denaturing formaldehyde (2%)/agarose (1.2%) gel, and blotted onto a positively charged nylon membrane by capillary transfer. The amount of RNA in all preparations was estimated photometrically at 260 nm and normalized by the amount of rRNA. A digoxigenin-labeled RNA probe for IGFBP-2 was obtained by RT of a 321-bp cDNA fragment coding for the C-terminal part of the peptide. For construction of the RNA probe for IGF-II, a 296-bp cDNA fragment of IGF-II using a 5'-primer CCCAATGGGGAAGTCGATGC, a 3'-primer AAGCACGGTCGGAGGGGTCG, and reversely transcribed RNA from human brain tissue as a template was amplified. As, due to different active promoters, various IGF-II transcripts were seen, results for IGF-II are not shown. After cloning the cDNA fragments into a pGEM-Teasy vector (Promega Corp., Mannheim, Germany), the corresponding RNA probe was obtained by RT, using digoxigenin-labeled nucleotides. Following the manufacturer’s protocol (Roche, Basel, Switzerland) for Northern blot analysis with the digoxigenin system, hybridization was performed with 60 ng/ml RNA probe each at 67.5 C for 16 h. The 1.6-kb IGFBP-2 bands were visualized photochemically with CDP-Star as substrate (Roche, Mannheim, Germany) and analyzed with a video imaging system, using the Aida 2.1 software package (Raytest, Straubenhardt, Germany).
Statistical analysis
The mRNA and protein values for IGF-II and IGFBPs are given as the mean ± SEM from the cultured GF of the 10 controls and 9 patients, respectively. The Mann-Whitney U test was used for statistical analysis, and significance was assigned for P < 0.05.
Results
AR defects in GF from CAIS patients
All patients who donated fibroblasts displayed the typical clinical features of the CAIS syndrome, i.e. the 46,XY karyotype, bilateral testes with undisturbed T synthesis, and an unambiguously female phenotype.
Androgen binding was absent in all CAIS fibroblasts (not shown). Table 1
lists the results of the AR gene
analysis. In three cases (cell strains GS-176, GS-480, and GS-641),
point mutations in exon 4, 6, or 7, associated with amino acid
substitutions, were responsible for the AR defect. A nonsense mutation
in exon 1 of the AR gene could be detected in cell strain GS-613. In
cell strain GS-540, molecular analysis revealed a 2-bp deletion in exon
1 that was associated with a frameshift and the introduction of a
premature stop codon. Both of the latter mutations result in the
expression of C-terminal truncated AR proteins lacking both the
DNA-binding domain and the ligand-binding domain. In cell strain GS-26,
a mutation in the splice donor site in intron 4 of the AR gene was
found to be the cause of aberrant splicing and deletion of 123
nucleotides. In three of the cell strains (GS-396, GS-423, and GS-612),
no genomic irregularities have been identified to date, but in these
cases an immunodetectable AR was absent in the Western blot analysis of
the AR (lanes 2, 3, and 4 in Fig. 1).
|
|
IGF concentrations were measured by RIA in the CM after 48 h of culture of the GF strains in serum-free medium. IGF-I was not detectable in the CM of any cell line (<0.2 ng/mg protein), even after T treatment (5 x 10-8 M) of the cells. Accordingly, there was also no IGF-I mRNA detectable in GF by means of RT-PCR (not shown).
After 48 h, however, the CM contained high concentrations of
IGF-II (Table 2), which had been secreted
by the GF. In CMs derived from the GF of the nine CAIS patients
(47.2 ± 9.7 ng/mg protein) we measured, on the average, 61%
(P < 0.001) less IGF-II compared with that in GF from
the 10 male controls (106.9 ± 11.8 ng/mg protein). Treatment of
GF with T did not cause any difference in the IGF-II concentration in
CM (Table 2). A similar difference between the IGF-II of CAIS patients
and controls was observed for the mRNA levels in GF by means of
semiquantitative RT-PCR (Table 3 and Fig. 2). Namely, the CAIS GF accumulated, on
the average, approximately 31% less (P < 0.01) IGF-II
mRNA than controls, that is 53.8 ± 2.4 vs. 37.2
± 5.6 relative units (Table 3). As various transcripts of IGF-II were
expected in the Northern blot (not shown), quantitative analysis could
not be done. Again, T treatment did not influence the IGF-II mRNA
levels.
|
|
|
The Western ligand blot of CM (Fig. 4) of six representative cell
strains demonstrated that CAIS GF (GS-26, GS-176, GS-164, and GS-540)
gave a banding pattern of IGFBPs that is both qualitatively and
semiquantitatively altered compared to that of the control GF (GS-517
and GS-614). The noticeably fainter IGFBP band at 31 kDa in CAIS GF was
most striking. This band most probably corresponds to IGFBP-2. This
finding is in accordance with the lower IGFBP-2 concentrations measured
by RIA of 39.4 ± 13.1 ng/mg protein for CAIS GF vs.
82.3 ± 37.7 ng/mg (P < 0.03) protein for
controls (Table 2
). Concerning IGFBP-2 mRNA levels, no difference
between CAIS and controls was observed (Table 3 and Figs. 2 and 3). IGFBP-1, with an expected molecular
mass of 25.6 kDa, was not detectable in either the ligand blot or the
RIA. In the molecular mass range between 42 and 44 kDa, the double band
that is typical for glycosylated IGFBP-3 forms was recognizable. These
band was very prominent in the ligand blot in accordance with the high
RIA IGFBP-3 values of approximately 300-1200 ng/mg protein in GF CM
(Table 2 and Fig. 4). The IGFBP-3
concentration in the culture medium and the IGFBP-3 mRNA levels showed
no significant differences between CAIS GF and control GF. T slightly
stimulated the expression of IGFBP-3 mRNA (Table 3), but not the
IGFBP-3 concentration in control GF (Table 2). In the range of 28 kDa,
an IGFBP band was detected that was stronger in CAIS GF (Fig. 4). This
band corresponds approximately to the mass of IGFBP-5. The band was
weaker in the CM of control GF that had been treated with T, but was
hardly changed in CAIS GF after T treatment. On the mRNA level, a
reduction of IGFBP-5 mRNA by 36% (P < 0.001) due to T
treatment was measured using RT-PCR, but was not found in control GF
(Table 3 and Fig. 2). In addition, an IGFBP band of 21 kDa that was
particularly prominent in control GF, which has the same mol wt as
IGFBP-6, may represent a proteolytic IGFBP fragment. This unidentified
band appeared to be weaker in the CM of CAIS GF and in this case was
not influenced by T (Fig. 4).
|
|
Discussion
In addition to the prostate (13, 14, 18), GF represent a suitable model for androgen-responsive tissues (5, 19, 20, 23, 24). As a new approach to elucidate the interplay between the androgen action via the AR and that via the IGF system, normal androgen-responsive GF strains were compared with completely androgen-insensitive GF strains. The latter were established from nine clinically well documented patients with CAIS. Normal GF strains (controls) were established from foreskin tissue of 10 healthy, age-matched, normally virilized males undergoing circumcision.
Mutations in the AR gene and/or biochemical alterations of the AR
protein, and complete absence of 5DHT binding, i.e. AR
function, were detected in all CAIS GF strains. The loss of the
functional AR had considerable consequences to the IGF system in the GF
strains. In particular, the amounts of secreted IGF-II
(P < 0.001) and IGFBP-2 (P < 0.03)
were significantly decreased in CAIS GF strains. In addition, the
IGF-II mRNA level was lower (P < 0.01) in CAIS GF.
This is of importance because IGF-II is known to be a major
autocrine/paracrine regulator of fetal growth, i.e. also of
genital growth and differentiation (17). In accordance
with these findings, reduced autocrine/paracrine activity of IGF-II in
skin fibroblasts indicated a role in the pathogenic mechanism leading
to growth failure in patients with Turner’s syndrome
(25).
The present findings are in accordance with recent findings in the prostate. In particular, the mitogenic and antiapoptotic activities of IGF-II, but not of IGF-I, by way of the IGF-I receptor in benign prostate hyperplasia were increased by locally produced androgens (26). Furthermore, IGF’s actions via the IGF-I receptor signaling pathway are thought to play a role in androgen-dependent prostate carcinogenesis and metastasis (27, 28, 29). Apart from this, epidemiological studies have revealed that high IGF-I and androgen serum levels are risk factors for developing prostate cancer (30).
IGFBP-2 is a modulator of the actions of IGFs via the IGF-I receptor and is further presumed to have cellular effects that are independent of the receptor (31). However, exposure of the cells to near-physiological amounts of T did not stimulate the secretion of either IGF-II or IGFBP-2. In prostate tissue, androgens failed to up-regulate IGF-II, whereas antiandrogens down-regulated IGF-II (18). The decreased secretion of IGFBP-2 in androgen-insensitive GF is in agreement with two earlier findings. First, androgen stimulated IGFBP-2 expression in an immortalized human osteoblastic cell line, which displays both an osteoblastic phenotype and physiological levels of functional AR (32). Second, IGFBP-2 gene expression is probably regulated by the AR, as the IGFBP-2 gene promoter contains androgen-responsive DNA elements as potential binding sites for the activated AR (33). The slight stimulation of IGFBP-2 expression upon exposure to T (5 x 10-8 M) observed in CAIS GF, however, may be attributed to estrogen effects, as GF contain aromatase activity to convert androgens into estrogen (34). The accumulation of high levels of IGFBP-5 mRNA in all GF strains and the detection of a prominent IGFBP band of 28 kDa in the ligand blot suggest the production of this IGFBP by these cells. As the 28-kDa band was markedly thicker with the CM of the CAIS GF, androgen action may be responsible for the down-regulation of IGFBP-5, as it was seen in the rat prostate after withdrawal of androgen by castration and in the mouse androgen-dependent Shionogi tumor model (35, 36). Unfortunately, however, a reliable immunoassay RIA to measure IGFBP-5 in the medium is not presently available, and thus it could not be assessed. It is, however, assumed that elevated IGFBP-5 expression in prostate cancer cells in the absence of androgen is a mechanism to magnify the antiapoptotic and mitogenic effects of IGFs as a kind of adaptive cell survival mechanism, thereby accelerating progression to androgen independence (36, 37).
It can be concluded that in addition to the known endocrine effects of IGF-I, the autocrine/paracrine effects of the fetal growth factor IGF-II and IGFBP-2 mediate the cellular action of androgen in androgen-responsive tissues. Obviously, the functional AR stimulates the local expression of IGF-II and IGFBP-2. Hence, the strong reduction in the expression of IGF-II and of IGFBP-2 in genital tissue of CAIS patients may be one factor leading to a failure of virilization during fetal growth. The functional role of IGFBP-5 in these processes, which, as in the prostate model, appears to be regulated by androgens in the opposite way, is less clear. In future studies the involvement of the IGF-system, in particular of IGFBP-5, in mediating androgenic effects as well as the role of estrogen need to be investigated in greater detail.
Acknowledgments
We thank Karin Weber (Tubingen) and Margarete Sudmann (Bonn) for able technical assistance.
Footnotes
This work was supported by the Deutsche Forschungsgemeinschaft (Grants El 167/3-1 and SFB 351, A1) and the Growth Research Center, Tubingen.
Abbreviations: CAIS, Complete androgen insensitivity; CM, conditioned medium; GAPDH, glyceraldehyde phosphate dehydrogenase; GF, genital skin fibroblasts; IGFBP, IGF-binding protein.
Received March 29, 2001.
Accepted June 6, 2001.
References
-reductase may be mediated via
insulin-like growth factor-I. Endocrinology 133:447–450[Abstract]
-reductase or androgen receptor activity. J Androl 14:73–78This article has been cited by other articles:
 |
|
 |
|
  Y. Kashida, K. Ishikawa, K. Arai, and K. Mitsumori Morphological Characterization of Skin Ganglion-like Cells in Djungarian Hamsters (Phodopus sungorus) Vet. Pathol., September 1, 2003; 40(5): 548 - 555. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
