| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
C Inhibin1
a URA CNRS 1291, Laboratoire de Physiologie de la Reproduction des Mammifères Domestiques, 37380 Nouzilly, France
b CRBM-CNRS (ERS 155 et INSERM U249), BP 5051 34033 Montpellier, Cedex, France
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
C subunit of inhibin. Thus, rete testis fluid (RTF) proteins were purified by immunoaffinity chromatography using antibodies against residues 1–7 plus 7–30 (experiment 1, A-fractions) and 14–28 of the C inhibin subunit (experiment 2, B-fractions), and the LHRH Statin activity of the fractions was examined by intracerebroventricular administration in castrated rams followed by RIA of plasma LH levels in 15-min blood samples. Fractions that bound to the immunoaffinity column with low affinity were eluted with 0.5 M NaCl, pH 7.4 (-F2); then highly bound fractions were eluted sequentially in acidic (pH 2.5, -F3) followed by basic conditions (pH 11.5, -F4). In experiment 1, RTF (40 µg, n = 4) and highly bound fractions (A-F3, 30 ng, n = 8, 150 ng, n = 3; A-F4, 120 ng, n = 5) decreased LH mean plasma levels between 4 and 6 h after injection by 39%, 29%, 43%, and 37%, respectively (P < 0.001 to 0.01), while the weakly bound fractions (A-F2, 180 ng, n = 4) and albumin control (40 µg, n = 4) had no activity. In experiment 2, RTF (100 µg, n = 4) and B-F3 (100 ng, n = 3) decreased plasma LH levels by 48% and 38%, respectively (P < 0.001 to 0.05), whereas B-F4 (100 ng, n = 4) and albumin control (100 µg, n = 4) had no effect. A fraction obtained from B-F3 by gel filtration had significant LHRH Statin activity (63%, n = 6, P < 0.001). PAGE with colloidal gold staining revealed 3 high molecular weight bands and 5 low molecular weight bands in B-F3. The 3 high molecular weight bands were shown to belong to the clusterin family and did not appear to have LHRH Statin activity. The 5 low molecular weight bands were all labeled by anti-C inhibin antibodies. Collectively, these results strongly suggest that LHRH Statin has some homology with the 14–28 C inhibin sequence.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Inhibins are heterodimers made of and ß subunits associated by disulfide bridges. The subunit is processed from a precursor that may be cleaved to give pro, N, C, proNC, NC proteins (for reviews, see [5–9]). Beta subunits are processed from either ßA or ßB precursors. The inhibin dimers (mainly C-ßA/ßB and NC-ßA/ßB) are biologically active in suppressing FSH release by decreasing the biosynthesis of FSHß subunit. Moreover, the C-ßA/ßB dimer can decrease the specific binding of LHRH to rat pituitary cells [10]. Free inhibin subunits do not have such biological actions. However, they have been shown to have other activities. Indeed, precursor is able to compete with FSH for binding to its receptor [11]. Furthermore, experiments involving inhibin knockout mice implicated inhibin as a tumor suppressor protein [12]. Finally, immunoneutralization experiments involving administration of antibodies against N-specific peptides or recombinant molecules suggested that N or N-related proteins might facilitate ovulation in the ewe [13–15].
As LHRH Statin activity is present in ovine rete testis fluid (RTF) proteins [1] and 32-kDa inhibin was isolated from the same source [16, 17], the question arose whether LHRH Statin and 32-kDa inhibin were the same entity. To address this question, we have previously administered rams or ewes via the ICV route either charcoal-treated RTF, bovine follicular fluid (bFF), or purified bovine 32-kDa inhibin and analyzed LH pulsatility before and after treatment [18]. In conditions under which RTF proteins suppress LH pulses, neither 32-kDa inhibin (in an amount known to be present in RTF) nor bFF was able to modify LH release, whereas they both decreased FSH release in a well-characterized rat pituitary cell culture system. Thus, it was concluded that LHRH Statin activity is not borne by 32-kDa inhibin or bioactive inhibin molecules, or by inhibin immunologically related molecules known to be present in bFF [19–26].
In the present study we tested the hypothesis that LHRH Statin is at least partly constituted by inhibin subunit. Why such a hypothesis? First, in the testis or ovary, inhibin mRNA was shown to be expressed at levels some 10 times greater than that for either ßA or ßB mRNA [27, 28]. Thus the possibility cannot be excluded that inhibin subunit-related or -containing proteins have a range of biological activities that have yet to be identified. Second, rats passively immunized against an N-terminal peptide of C inhibin subunit show a clear-cut increase of LH pulse frequency 1–5 h after treatment [29]. This can be interpreted as evidence that a blood-borne molecule immunologically related to C inhibin spontaneously decreases LH pulse frequency. As it is established unequivocally that LH pulsatility has a hypothalamic source [2], this action would take place at the hypothalamic level and alter LHRH pulse frequency.
Thus, in the present paper, we have examined the LHRH Statin activity of RTF proteins purified using immunoaffinity chromatography with antibodies against C inhibin-related peptide sequences. The LHRH Statin activity of the fractions obtained was analyzed through ICV administration to castrated rams and analysis of LH release before and after treatment as previously described [1, 18].
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Four peptides (1–7, 7–30, 1–28, 14–28) corresponding to the amino acid sequence of the ovine C subunit of inhibin ([16], S1TPPLPWPWSPAALRLLQRPPEEPAAHADC) were synthesized by a solid-phase procedure using polyacrylamide resin (Expansin from Expansia, 30390 Aramon, France) and Fmoc strategy on a Milligen 9050 Peptide Synthesizer (Waters, Milford, MA). The purity of the peptides was determined by amino acid analysis. These peptides were used after deprotection by piperidine (20% in dimethylformamide) and eventually after cleavage with trifluoroacetic acid according to Hanin et al. [30].
Antisera
Antibodies against the 1–7 and 7–30 sequences were raised in rabbits using the peptide-resin complex according to Hanin et al. [30]. Slight modifications of the technique included multiple intradermal injections instead of the subcutaneous route and different amounts of immunogen (250 µg of complex containing 125 µg peptide per rabbit per injection). A minimum of 5 immunizations was necessary to obtain antibodies. Antibodies against the 14–28 or 1–28 sequences were obtained in rams (respectively, n = 4 and n = 2) using the peptide-resin complex linked to ovalbumin with glutaraldehyde and the same protocol as applied to rabbits.
Coupling Peptide-Resin Complex to Ovalbumin
This ovalbumin complex was made because the peptide-resin complexes were found to be poor immunogens in rabbits as opposed to previous observations for other peptides [30]. Ten milligrams of 14–28 or 1–28 peptide-resin complex was suspended in 1 ml of PBS (20 mM PO4, pH 7.4; 150 mM NaCl) and sonicated. Ovalbumin (Sigma, St. Louis, MO; 4.3 mg) diluted in 1 ml PBS was added, and, under constant stirring, 2 ml of freshly prepared glutaraldehyde (0.2%) was then added drop by drop. This solution was incubated for 1 h at room temperature (RT). To stop the reaction, 1 ml of glycine (1 M, pH 7.2) was added and incubation continued (1 h, RT). The peptide-resin-ovalbumin complex was recovered by centrifugation (500 x g, 15 min, RT). The pellet was washed twice with 2 ml of PBS and centrifuged. Finally, the pellet was suspended in 5 ml PBS, aliquoted (250 µl or 125 µl according to the number of rams), and stored frozen until immunization. Coupling efficiency was checked by measurement of ovalbumin in the initial solution and in the 3 supernatants using the protein assay of Bradford [31].
Screening of the Antisera
The antisera were systematically screened using ELISA as described in detail previously [30]. Briefly, 96-well plates (Maxisorp; Nunc, Naperville, IL) were coated with the corresponding immunogen (10 ng in NaHCO3 0.5 M, pH 9.6, per well) or resin as a control for rabbit antiserum or 1–28 or 14–28 peptide for ovine antiserum. After coating, washing, and incubation of the antisera, goat anti-rabbit gamma globulin or rabbit anti-sheep gamma globulin, both labeled with alkaline phosphatase (Sigma), were, respectively, added for rabbit and sheep antisera. After reaction with paranitrophenyl phosphate (Sigma), absorbance readings were made at 405 nm. For a given antiserum the specific reading was considered as the difference between immunogen and resin control. Only high-titer antisera were considered for the next step.
Unexpectedly, the titers of all ovine antisera (including anti-14–28 sequence) assessed using 14–28 peptide-coated wells were around half those using 1–28 peptide coated wells. As this is probably attributable to a higher percentage of coating for the longer peptide, this latter coating was subsequently used to titer all antibodies. Because of low titers, only anti-14–28 sequence was further used.
Purified Antibodies: Immunoaffinity Chromatography of Antisera
Gamma globulin-enriched fractions were obtained by two successive precipitations with ammonium sulfate (50% saturation), dialyzed (twice against 50-volume minimum of water, 8500 cut-off tubing), and lyophilized (yield: 20 and 46 mg/ml serum for rabbit and sheep, respectively). Gamma globulins were solubilized (Tris/HCl, 1 M, pH 7.4; NaCl, 0.5 M), centrifuged (2000 x g, 30 min, 4°C) to remove insoluble proteins, and diluted to a final concentration (200 mg in 300 ml and 100 mg in 100 ml Tris/HCl, 10 mM, pH 7.4, for rabbit and sheep antisera, respectively), and finally incubated in batch (overnight, 4°C under mild agitation) with 500 mg peptide-resin complex (either 1–7 or 7–30 sequence accordingly for rabbit antiserum and 7–30 sequence for sheep antiserum). The slurry was then introduced into a polypropylene syringe.
After two washings controlled under UV absorbance at 280 nm (first, Tris/HCl, 10 mM, pH 7.4; second, the same plus NaCl, 0.5 M), purified antibodies were eluted using acid (glycine/HCl, 25 mM, pH 2.5; yield: around 0.3% and 1.1% of total gamma globulins for rabbit or sheep antiserum, respectively) followed by basic conditions (Na2HPO4, 0.1 M, pH 11.5, yield: around 0.3% and 2% of total gamma globulins for rabbit or sheep antiserum, respectively) and immediately neutralized (phosphate buffer, 0.1 M, pH 8.1 and 6.8 for rabbit or sheep antiserum, respectively). Before neutralization, aliquots were sampled for protein quantification using the Bradford assay [31] adapted for 96-well microplates according to Bio-Rad's specifications. Purified antibodies (pooled acidic and basic and only acid-eluted proteins for rabbit or sheep antiserum, respectively) were used to obtain immunoaffinity matrices after antibody titer measurement by ELISA as earlier described.
Coupling Purified Antibodies to Cyanogen Bromide (CNBr)-Activated Sepharose
CNBr-activated Sepharose (Amersham Pharmacia Biotech, Piscataway, NJ; rabbit: 1 g in 3.5 ml; ovine: 3 g in 10-ml gel) was reacted with purified antibodies (rabbit: 2 mg mixed anti-1–7 + 7–30 sequences in equal amounts; ovine: 7.5 mg anti-14–28 sequence) according to the manufacturer's recommendations. Observed coupling efficiency was 50% and 90% for rabbit and ovine proteins, respectively. These affinity materials/columns will be further referred to as A and B for rabbit and sheep material, respectively. A and B column efficiency was analyzed using 32-kDa inhibin (Peninsula #IP1095; Belmont, CA) labeled with 125I by the choramine T method as described by Blanc and Poirier [32]. This evidenced their ability to bind 125I-inhibin dimer and to release it in acid (pH 2.5) and basic (pH 11.5) conditions.
Obtaining RTF
RTF was obtained by cannulation of the extratesticular rete testis of adult rams as previously described [33]. The collected fluid was centrifuged to separate spermatozoa (1500 x g, 20 min, 4°C). Charcoal treatment (10 mg/ml, 30 min, RT, according to Blanc et al. [1]) was used to remove steroids, and material (around 400 µg protein per milliliter) was stored either at -196 or at -20°C.
Immunoaffinity Chromatography of RTF Proteins
Various amounts of RTF proteins (from 50 ml to 50 ml 10 times concentrated on an Amicon YM10 membrane [Danvers, MA] for A column; from 25 ml to 200 ml concentrated to 50 ml for B column) in different conditions (batch or closed circuit; 18 h to 75 h) were incubated with the matrix at 4°C. Washing and elution were performed in conditions identical to those used to obtain purified antibodies. Three fractions were obtained and considered (F2: 0.5 M NaCl; F3: pH 2.5; F4: pH 11.5). Each fraction was individually pooled, partially concentrated on a Speed-Vac Savant (Farmingdale, NY) centrifuge or on Amicon YM10 membranes (cut-off 10 000) according to the volume, dialyzed against deionized water (Amicon YM10, or Pierce [Rockford, IL] 500 µl dialysis system with 8500 cut-off membranes) before injection, and further concentrated on a Speed-Vac centrifuge for PAGE analysis. According to the yield obtained for highly bound F3 and F4 fractions, 50-ml RTF volumes were subsequently used for routine chromatography.
Gel Filtration
Fractions from B column were recovered in sufficient amounts to be further purified by fast protein liquid chromatography (FPLC) gel filtration (HR 75, 10 x 300 mm; Amersham Pharmacia Biotech). Elution buffer was ammonium acetate (0.4 M, pH 6.0) and elution rate 0.5 ml/min. Eluted proteins were analyzed by PAGE and bioassayed.
PAGE of Fractions [34]
Samples were diluted (1:3, v:v) in denaturing buffer (Tris/HCl 250 mM, pH 6.8, containing 40% glycerol, 8% SDS, 0.004% bromophenol blue) and boiled for 2 min. Samples were electrophoresed in a 12% constant or 6–16% polyacrylamide gradient SDS slab gel with a 4% stacking gel, 0.75-mm thickness, using a Miniprotean II apparatus (Bio-Rad, Richmond, CA). PAGE was run in buffer (Tris/HCl, 25 mM; glycine, 200 mM; 0.1% SDS) at constant voltage (200 V) during approximately 1 h.
After electrophoresis, proteins were first detected in the gel using the silver staining technique of Morissey [35]. When it was shown that this technique had too high a detection limit for inhibin-related molecules, samples were electrotransferred to nitrocellulose membranes (Hybond C; Amersham Pharmacia Biotech) and labeled with colloidal gold (Protogold, Tebu, France) according to the supplier's instructions. The semidry transfer system (Milliblot electroblotter; Millipore, Bedford, MA) was run for 2 h at 120 mA with two gels in a Tris/glycine buffer, pH 8.3, with 20% methanol. Western blots were made on nitrocellulose membranes as previously described [30]. After checking for the transfer using Ponceau Red, the membrane was blocked overnight with 10% (w:v) skimmed dry milk, washed in TBS (Tris/HCl, 50 mM, pH 7.4; 200 mM NaCl; 3 times), and incubated with sheep anti- inhibin antiserum (SAIP2, 1:10 000 in TBS containing 3% BSA, 3 h, 37°C). After 3 washings in TBS, membranes were incubated with rabbit anti-sheep gamma globulin linked to peroxidase (Sigma; 1:1600 in TBS containing 3% albumin, 3 h, 37°C). Membranes were then washed (3 times in TBS, then twice in TBS containing 0.3% Tween 20, 30 min). Peroxidase was then revealed using luminol (Amersham Pharmacia Biotech) as indicated by the supplier. The sheep anti- inhibin antiserum SAIP2 was an anti-14–28 sequence. Its specificity was analyzed through comparisons of the bands evidenced with SAIP2 before and after incubation with 7–30 peptide-resin complex (5 mg, overnight, RT) and centrifugation. Using the conditions mentioned above, the only band stained after incubation with 7–30 peptide-saturated SAIP2 (120 kDa) was also present when SAIP2 was omitted; this band was thus considered to be aspecific.
Measurement of the LHRH Statin Activity of the Fractions
Animals were cared for in full accordance with the NIH guidelines on the handling and care of laboratory animals.
LHRH Statin activity was measured in castrated rams using the ICV route as previously described [1, 18]. Briefly, under deep anesthesia, adult rams were fitted with a permanent cannula into the third ventricle using stereotaxical techniques, as described elsewhere [36], and at the same time they were castrated. Seven days to 9 mo later, the animals were fitted with an indwelling jugular catheter. On the next day, rams were blood-sampled every 15 min for 12–15 h. Five hours after the beginning of blood sampling, fractions in a total volume of 200 µl were slowly (2–4 min) injected into the ventricle. Just before the injection, fractions were diluted with pyrogen-free saline containing 50 µg pyrogen-free human serum albumin per injected dose and sterilized using 0.22-µm filters. Plasma samples were stored either at -20°C until LH and FSH assays or at 4°C when the assays were to begin the next day.
RIAS
Plasma LH levels were determined by a specific RIA [37] and the results expressed as nanograms LH CY 1051 (obtained from Dr. Y. Combarnous, INRA, Nouzilly, France, equivalent to 2.5 LH NIH S1) per milliliter plasma. The minimal detection level (measured at B/B0 = 95%) was 0.2 ng/ml. Intra- and interassay coefficients of variation (CV) at B/B0 = 50% were 7–10% and 10–14%, respectively. All samples from the same sampling session (7–9 rams) were assayed in duplicate and in the same assay. This was also the case for FSH assays.
Plasma FSH levels were measured by a highly specific RIA [32] and the results expressed as nanograms FSH HG 225 per milliliter plasma. This standard was provided by Dr. H. Grimek (University of Wisconsin, Madison, WI) and is equivalent to 4.3 NIH FSH S3. The minimal detection level, measured at B/B0 = 95%, was 1.1 ng/ml. Intra- and interassay CV at B/B0 = 50% were 8–11% and 13–15%, respectively.
Statistical Analysis
Clear pulsatile LH patterns were not observed in all rams, particularly in long term-castrated (more than 1 mo) rams, as previously observed [2]. Indeed, in these rams, the one-to-one relationship between an LH pulse and an LHRH pulse that is observed in intact or short term-castrated rams is seen only in, at the most, 50% of the LHRH pulses, because LH pulses are ill defined (erratic pulse [2]). In those conditions a mathematical LH pulse lacks physiological significance. This holds whatever algorithm is used to define an LH pulse when sampling interval is 10 min or more (A. Caraty, personal communication). Thus, only mean plasma LH levels were evaluated for statistical purposes. Previous as well as the present results showed that the effects of RTF proteins on plasma LH levels took place between 1 and 2 h after ICV injection. For this reason, sampling periods were divided into 2-h periods: control period, P0 (-2 h–0), P1 (0–+2 h), P2 (+2 h–+4 h), and P3 (+4 h–+6 h). To assess the effects of one treatment, control versus experimental hormone levels during these periods were subjected to ANOVA with repeated measures. If a treatment x period interaction was associated with a significant P level (P = 0.05), this was followed by an ANOVA with factorial analysis within one treatment. If a significant effect of period was observed, interperiod comparisons were made with P0 (i.e., preinjection) values using post hoc Scheffé F test. A logarithmic transformation was applied to the data in order to homogenize variances before ANOVA analyses.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
On average, 50 ml RTF (20 mg) yielded 180, 30, and 120 µg for fractions A-F2, A-F3, and A-F4, respectively. Twenty-eight rams were injected with either serum albumin (control, 40 µg, n = 4), RTF proteins (40 µg, n = 4), A-F2 (180 ng, n = 4), A-F3 (30 ng, n = 8, 150 ng n = 3), or A-F4 (120 ng, n = 5). The effects of these materials are shown in representative individual rams (Fig. 1), and group data are summarized in Figure 2. Unpurified RTF, as well as highly bound proteins (A-F3 and A-F4), were able to depress plasma LH levels (treatment x period interaction: P = 0.0001, 0.01, 0.003, 0.001; within treatment period effect: P = 0.0001, 0.0001, 0.001; 0.0007, respectively, for RTF, 30 ng A-F3, 150 ng A-F3, and A-F4), while low-affinity bound proteins (A-F2) had no effect (treatment x period interaction: P = 0.1). This indicated a ~1000-fold increase in the specific activity between unpurified RTF and A-F3. With all active fractions, the effect was not present during the first 2 h after administration. Examination of individual profiles indicated that variability in plasma LH levels was decreased during P2 and P3, suggesting that LH pulse frequency and/or amplitude was diminished or suppressed between 2 and 6 h after injection (Fig. 1, D–F). However, it was not possible to ascribe this suppression to either LH pulse frequency or amplitude because the LH patterns were those of long term-castrated rams. An individual variability of the response was observed among the 8 animals injected with A-F3 (30 ng). Indeed, 4 rams did not respond, while others did (P0, P1, P2, P3: 9.8 ± 1.5, 9.2 ± 1.5, 9.5 ± 1.2, 9.3 ± 1.4, and 9.8 ± 1.5, 7.8 ± 1.6, 5.5 ± 1.1, 4.4 ± 0.9 for nonresponders and responders, respectively). Although only 3 rams were examined in the 150 ng A-F3 group, the magnitude of the decrease ([P0-P3]/P0 ratio) was higher than that of the 30 ng A-F3 group (43% vs. 29%, respectively). The magnitudes of the decrease for RTF and A-F4 groups were 39% and 37%, respectively. Plasma FSH levels were not modified by any of the treatments (data not shown).
|
|
With use of silver staining, active fractions A-F3 and A-F4 contained 3 main proteins (44, 62, and 120 kDa), while 6 were revealed in fraction A-F2 (22, 44, 60, 74, 100, and 120 kDa, not shown).
Because of low amounts of protein recovered, partly due to the low capacity of the immunoaffinity column A and to considerable losses during the dialysis step, it was not possible to apply a further purification step with the present immunoaffinity chromatography procedure.
Experiment 2: Purification of RTF Proteins on Anti-14–28 Immunoaffinity Chromatography (Column B) and Their Bioactivity
Because of low titers, only antisera against 14–28 peptide-resin-ovalbumin complex could be used to generate an immunoaffinity chromatography column.
On average, 50 ml RTF (20 mg) yielded 119 ± 55, 132 ± 24, and 141 ± 50 µg proteins for B-F2, B-F3, and B-F4 fractions, respectively. Increasing the amount of RTF proteins added to the column above 20 mg did not significantly affect the recovery of the highly bound fractions (B-F3 and B-F4).
Fifteen rams were injected with either serum albumin (control, 100 µg, n = 4), RTF proteins (100 µg, n = 4), B-F3 (100 ng, n = 3), or B-F4 (100 ng, n = 4). Individual profiles for representative rams are shown in Figure 3, and combined results are summarized in Fig. 4. Injection of RTF proteins depressed plasma LH levels in all 4 rams, while no effect was observed in control serum albumin-injected rams (treatment x period interaction: P = 0.04; within treatment period effect P = 0.0001). Fraction B-F4 was without effect in all 4 rams (treatment x period interaction P = 0.4), while B-F3 significantly decreased plasma LH levels (treatment x period interaction P = 0.04; within treatment period effect P = 0.0095). As observed in experiment 1, the effect induced by active fractions was not observed during the first 2 h after injection but only afterward (Figs. 3 and 4). The magnitude of the decrease was 48% and 38% for RTF and B-F3, respectively. Furthermore, plasma FSH levels were not affected by any of the treatments (not shown).
|
|
Fractions were analyzed by PAGE, and total protein was visualized by silver staining of gels or with colloidal gold after transfer to nitrocellulose membranes (Fig. 5, B and C). The gold staining was chosen as it proved to be more sensitive than silver, especially for the 14- to 30-kDa range proteins (for instances, see lane B-F3 in Fig. 5B, silver staining; and Fig. 5C, colloidal gold staining). Using colloidal gold staining and nonreducing conditions, PAGE analysis of fraction B-F3 identified 3 main (55–57 doublet and 67 kDa) and 5 minor proteins (15–16, 18–20, 22–23, 28, and 31 kDa, Fig. 5C). Using Western blotting with an inhibin antibody, 3 specific bands (44, 28, and 22–23 kDa) were identified in B-F3 (Fig. 5D). However, with another preparation and using a higher amount of material, all minor proteins identified with gold staining were also labeled with anti- inhibin (Fig. 5D, lane 5). Furthermore, the 44-kDa band was revealed in Western blots but not with colloidal gold, probably due to the higher detection limit of the latter technique. Thus, in the B-F3 fraction, the 3 main proteins (55–57 doublet and 67 kDa) were the only bands not labeled in Western blots. These bands may be identical to those that appeared as main bands in unpurified RTF stained with either Coomassie blue or silver (Fig. 5A). Western blotting did not reveal any inhibin-related proteins in B-F2 and RTF: the approximate 120-kDa protein was also present when the first antibody was omitted (Fig. 5D) and must be considered as nonspecifically labeled.
|
Fractions B-F3 and B-F4 were further purified by gel filtration on an HR 75 column (FPLC; Amersham Pharmacia Biotech; 10 x 300 mm) eluted with 0.4 M ammonium acetate, pH 6.0 (Fig. 6). As the inactive fraction B-F4 resolved into two main absorbance peaks that were also present in active B-F3 fraction, the remaining fraction (B-F3-HRII) seemed particularly interesting and was therefore selected for ICV injection.
|
The effects of B-F3-HRII are shown in all rams injected (Fig. 7), and combined data are summarized in Figure 8. In all 6 rams, LH plasma level was decreased (treatment x period interaction P = 0.015; within treatment period effect P = 0.0001). Although plasma LH levels during the first period after injection were not decreased (P = 0.2, Scheffé F test), we could have considered them as depressed using other post hoc tests (P < 0.05 for Duncan, Student-Neuman-Keuls, or Dunnett one-tailed). The magnitude of the decrease observed between 4 and 6 h after injection (63%) was the highest inhibition observed among all treatment groups analyzed in the present work. Although clear LH pulses were not evident in all rams during control periods, it appears that at the time of maximal effect, no LH pulses were present (Fig. 7). Furthermore, nonspecific toxic effects of carryover reagents do not appear to explain the effects observed with B-F3-HRII, as plasma cortisol levels were modified similarly in B-F3-HRII- and human serum albumin-injected rams as compared with preinjection levels, according to the time of the day (not shown). Finally, plasma FSH levels were not changed in any of the treated animals (not shown).
|
|
PAGE analysis of the fractions obtained from B-F3 and identified with colloidal gold showed that the two high molecular weight bands (55 and 67 kDa) were still present in B-F3-HRII, although in less relative abundance than in B-F3, as B-F3-HRI only contains these proteins. B-F3-HRII also includes 15–16, 18–20, 22–23, and 28-kDa proteins, the latter being one of the more abundant components of this fraction (Fig. 5E). All these low molecular weight bands were also identified with anti-C inhibin antiserum.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
C subunit of inhibin have LHRH Statin activity while fractions bound with low affinity do not have such activity (experiment 1), strongly suggest, for the first time, that LHRH Statin is immunologically related to the N-terminal region of the C inhibin subunit. This result is reinforced by the ratio between active doses of the source of LHRH Statin (RTF: 20–100 µg, present data; and Blanc et al. [1]) and those of highly bound fractions (30–150 ng A-F3, 120 ng A-F4, present data), i.e., around 1000. Furthermore, the use of an immunoaffinity column differing with respect to both immunogen and the species used to raise antibodies also gave a highly bound fraction with high LHRH Statin activity both before and after further purification (B-F3 or B-F3-HRII, experiment 2), confirming the results of the first experiment. The observation that the basic-eluted fraction eluted after the acid-eluted fraction was (A-F4, experiment 1) or was not active (B-F4, experiment 2) can most likely be explained on the basis of the higher affinity of antibodies to LHRH Statin in the first experiment. However, these affinities have not been analyzed, as, to our knowledge, it is not possible to measure affinity of polyclonal antibodies both before and after binding to an insoluble matrix. Although LH pulses could not be identified in all sampling sessions and have not been considered in the present statistical analyses, various lines of evidence strongly suggested that the proteins highly bound to the immunoaffinity matrices suppressed LHRH pulses. Indeed, we have previously shown that, with this animal paradigm, RTF total proteins suppressed LH pulses quite likely through a suppression of the LHRH pulses as the ICV-injected material was not available to the pituitary cells [1, 18]. Furthermore, the patterns of LH plasma levels in rams injected with active fractions, and particularly the more purified one (see Fig. 7), evidenced that the variability was almost completely abolished from 3 to 5 h after administration when compared with preinjection variability. This suggested that LH pulses had been suppressed. However, it still remains to be determined whether this likely suppression of LHRH pulses observed after administration of purified fractions is due to an effect on frequency or amplitude of the pulses.
Western blot analysis of active B-F3 fraction revealed that all bands were labeled with anti-C inhibin antibodies with 3 exceptions (55–57 doublet and 67 kDa). If they are the same proteins, the 55–57 doublet was much more abundant in B-F2 than in B-F3 or B-F4 fractions (Fig. 5, B and C), suggesting that its affinity for the gamma globulin matrix complex was less than that of C inhibin-related proteins. The 55–57 doublet and 67-kDa bands, as isolated from the B-F3-HRI fraction, belonged to the clusterin family of proteins as evidenced by their N-terminal sequence [38]. The fact that clusterins can bind to gamma globulins [39] explains their partial co-purification with C inhibin-related molecules. Both the 55–57 doublet and the 67-kDa clusterin-related proteins were shown not to have any LHRH Statin activity (unpublished results). This is also corroborated by the observation that the B-F3-HRII fraction from which these proteins were partially removed (see Fig. 5E vs. 5C) was more active than B-F3 on the basis of the LH suppression effect obtained with the same dose (63% vs. 38% inhibition for B-F3-HRII and B-F3, respectively; see Figs. 4 and 8). Thus, taken together, these data strongly suggest that LHRH Statin has some homology with the 14–28 C inhibin sequence.
The 28-kDa protein present in B-F3-HRII as one of the more abundant components was shown in another preparation, and with a higher amount of material, to have some homology with pro-C [38]. However, this molecule was devoid of LHRH Statin activity. This corroborates our previous data [18] showing that bFF, which contains high amounts of pro-C [20–22, 25], does not have any LHRH Statin activity and with those of Li et al. [40] indicating that peripherally injected bFF did not change portal LHRH levels.
Although active purified fractions probably included 31- to 32-kDa inhibin dimer (a 31-kDa band was identified only with Western blots by PAGE under nonreducing conditions in the B-F3 fraction, Fig. 5D, lane 5), 32-kDa inhibin dimer does not appear to be responsible for the LHRH Statin activity of these fractions. Indeed, purified bovine 32-kDa inhibin and bFF injected by the ICV route did not display any LHRH Statin activity in either ram or ewe models [18]. This suggests that if LHRH Statin is one of the bands stained by colloidal gold in the B-F3 fraction, it could be one of the 3 remaining proteins (15–16, 18–20, 22–23 kDa), the fourth (31 kDa) probably being inhibin dimer. However, we have not been able to further isolate these proteins and analyze their LHRH Statin activity because of the low amounts available.
The main 55-kDa protein present in B-F3 was identified as belonging to clusterin family protein. However, it appears that a small fraction of these 55-kDa proteins are inhibin-related molecules as suggested by Western blotting with higher amounts of material (Fig. 5D, lanes 3 and 5). A 55-kDa inhibin has been previously identified in follicular fluid by various authors [19, 23, 25, 26, 41–43).
The identities of the various bands detected using both colloidal gold staining and C inhibin antibody immunodetection remain to be established. Indeed, as these proteins were not completely isolated they could not be sequenced under either reducing or nonreducing conditions. However, their molecular weights suggest that some of these bands may correspond to inhibin -subunit-containing molecules that have already been identified in follicular fluid, plasma, or RTF (44, 31, and 28 kDa [17, 19–23, 25, 26]), while this does not appear to be the case for other bands (22–23, 18–20, and 15–16 kDa). The latter proteins could be specific to RTF proteins.
The immunological relation between C inhibin and LHRH Statin fits well with the negative correlation observed between C inhibin and plasma LH levels found in men evaluated for infertility [44]. These authors found a high correlation between seminal C inhibin and serum LH (r = 0.7, P < 0.001) whereas no significant correlation was found with serum FSH.
Further indirect evidence of immunological similarities between LHRH Statin or LHRH Statin-like factor and C inhibin comes from in vivo results obtained with passive immunoneutralization with anti-C inhibin. Indeed, if 1) LHRH Statin has a predominant role in LHRH pulse frequency in the physiological situation studied and 2) LHRH Statin is present in peripheral plasma, then, owing to the immunological relationship between LHRH Statin and C-containing molecules, immunoneutralization with anti-C inhibin would induce both an increase in plasma FSH levels (as a consequence of the well-established FSH-inhibin feedback) and an increase in LHRH/LH pulse frequency. Results of passive immunoneutralization will be examined first. Immunization of diestrous 2 rats against C inhibin induces a clear increase in LH/LHRH pulse frequency 1–5 h and 18–22 h after injection [29]. At the same times, the immunization also induces an increase in plasma FSH levels. Thus, it is noteworthy that LHRH/LH pulse frequency was increased at times when plasma estradiol-17ß levels could have been raised as a consequence of increased FSH plasma levels [45]; it is generally assumed that high levels of estradiol-17ß can decrease LH pulse frequency in the short term. Therefore, a possible interpretation of the increased LH pulse frequency after immunoneutralization with anti-C inhibin is that a factor (LHRH Statin?) present in the general circulation and holding LHRH/LH pulse frequency to low levels is blocked in presence of specific C inhibin antibodies.
After passive immunization against C inhibin in female rats, Arai et al. [45] observed significantly increased mean plasma LH levels at all periods, although the effect appeared to be greatest at diestrus 2 (408 ± 33% vs. control). In similar studies in the golden hamster, the higher dose of C inhibin antiserum increased mean plasma FSH, estradiol-17ß, and LH levels during the 30 h after treatment on diestrus 2 [46]. Similarly, in pregnant rats, Arai et al. [47] reported an increase in mean plasma LH levels after treatment with anti-C inhibin antiserum. Evidently, the increase of mean plasma LH levels must be interpreted through both pituitary and hypothalamic sites of action, while the increase in LH pulse frequency [29] is to be attributed to a hypothalamic site of action, direct or indirect.
In ewes and cows, no effect on plasma LH levels was reported after passive immunoneutralization against C inhibin during the luteal phase (ewe [48–51]; cow [52, 53]), follicular phase (ewe [54, 55]; cow [56]), or seasonal anestrus (ewe [57]). The same absence of effect was observed when injection of antibodies took place during delivery of progesterone by intravaginal devices either during seasonal anestrus [58] or the breeding season [55, 59, 60]. However, the results of the above studies may not be fully conclusive, as the blood sampling protocols used, with two exceptions [50, 51], were designed to analyze FSH secretion rather than LH secretion. The latter authors reported no significant effect of treatment 5 days after injection (interval between two consecutive LH pulses: 71 ± 8 vs. 45 ± 10 min, P = 0.09 for control and passively immunized groups, respectively), while Mann et al. [50] observed 2/4 vs. 5/5 pulses per ewe in control and C inhibin-immunized ewe, respectively (P not given), between 12 and 20 h after treatment. These data suggest a tendency for LH pulse frequency to increase after C inhibin immunoneutralization. Thus, further systematic studies of the effect of passive immunization against C inhibin on LH pulse frequency at different times after injection and at different phases of the estrous cycle would be necessary.
An abundant literature exists on the effect of active immunization against inhibin on gonadotropin levels in various species. These data are very difficult to interpret in terms of a potential direct effect of inhibin antibodies on an LHRH Statin-like factor. Active immunization against inhibin has generally, but not always, been shown to increase plasma FSH levels (reviewed by Findlay et al. [61]). As a consequence of increased plasma FSH levels, plasma estradiol-17ß levels may be increased, and in turn this may decrease plasma LH levels [62]. This indirect consequence of inhibin immunization could potentially mask a direct effect of C inhibin antibodies on an LHRH Statin-like factor.
In conclusion, our data on the use of anti-C inhibin immunoaffinity columns to purify LHRH Statin from RTF proteins have shown that highly bound fractions are about one thousand times more active than RTF in terms of LHRH Statin activity. This, and other evidence presented, strongly suggests that LHRH Statin shares some homology with the 14–28 C inhibin sequence. This finding provides further insights into the potential diverse activities of molecules structurally related to inhibin.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 C.C.G. was supported by a grant in aid by INRA (50%) and Région Centre (50%). 
2 Correspondence. FAX: 33 2 47 42 77 43; michel.blanc{at}tours.inra.fr
Accepted: January 21, 2000.
Received: December 28, 1998.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-inhibin modulate follicle-stimulating hormone receptor binding and biological activity. Endocrinology 1991; 129:1987–1999.[Abstract] Inhibin is a tumour suppressor gene with gonadal specificity in mice. Nature 1992; 360:313–319.[CrossRef][Medline]43) on fertility of ewes. Reprod Fertil Dev 1994; 6:265–267.[CrossRef][Medline]43-subunit (N) disrupts tissue remodeling and the increase in matrix metalloproteinase-2 during ovulation. Endocrinology 1995; 136:3657–3664.[Abstract] subunit and its isolation from bovine follicular fluid. J Mol Endocrinol 1989; 2:189–200.[Abstract]-subunit precursor proteins from bovine follicular fluid. Endocrinology 1989; 125:2141–2149.[Abstract], ßA and ßB subunits in various tissue predicts diverse functions. Proc Natl Acad Sci USA 1988; 85:247–251. inhibin immunoaffinity technique. In: Serono Symposium: Second Symposium on Inhibin and Inhibin Related Proteins; 1993; Sienna. Abstract.-subunit in human serum: implications for radioimmunoassay. J Clin Endocrinol Metab 1990; 70:1208–1212.[Abstract]-inhibin precursor proteins. J Clin Endocrinol Metab 1994; 78:433–439.[Abstract]-inhibin in human seminal plasma. J Clin Endocrinol Metab 1993; 76:476–483.[Abstract] subunit: effects on gonadotrophin concentrations and ovulation rate in sheep. J Endocrinol 1990; 124:167–176.[Abstract]
| ||||||||||||||||||||||
