| | Production and characterization of a novel monoclonal antibody against progesterone-induced blocking factor (PIBF)Received 4 September 2007; received in revised form 5 December 2007; accepted 11 December 2007. published online 13 February 2008. Abstract Progesterone-induced blocking factor (PIBF) has been described as an active factor intimately involved in regulation of the immune response in pregnancy. It has been shown that PIBF biased the cytokine balance to Th2-type in pregnancy and inhibited the activity of NK cells. The biological roles of PIBF would be better defined if methods for its detection and measurement in biological fluids are available. However, so far, reliable antibodies have not been developed to be used as specific probes. A monoclonal antibody designated as MAB 3A6 was produced and characterized. MAB 3A6 reacts specifically with PIBF. It can detect this protein in biological fluids when tested by immunoblot and recognizes PIBF expressed on the surface of lymphocytes of pregnant women stimulated in vitro with progesterone. The characteristics of MAB 3A6 makes it the possible basis for development of a clinically applicable assay to assess the presence and concentration of PIBF in biological samples. 1. Introduction  Progesterone-induced blocking factor (PIBF) was described initially by Szekeres-Bartho et al., 1985, Szekeres-Bartho et al., 1989a as an active factor intimately involved in regulation of the immune response in pregnancy. Subsequently, it has been shown that binding of antibodies to the γδ T-cell receptor expressed by γδ CD8+ T cells down-regulated the production of PIBF and increased NK cell activity (Polgar et al., 1999). This would suggest that γδ T cells are either the source of this immunomodulatory factor or in some way control its secretion. The hypothesis suggested by the authors is that activated γ/δ CD8+ lymphocytes express de novo functional progesterone receptors (PR) on their surface. Detailed investigations have shown that PIBF enhanced the secretion of IL-3, IL-4 and IL-10 by spleen cells of mice when cultured in vitro, while the production of INF-γ was not modified in comparison to controls (Szekeres-Bartho and Wegmann, 1996). Kozma et al. (2006) reported that the effect of PIBF is mediated via a novel IL-4 receptor which consists most probably of IL-4R α-chain and the PIBF receptor which is a GPI-anchored protein. The activation signal induced phosphorylation and nuclear translocation of STAT-6 and inhibition of the STAT-4 pathway. Since it has been proven that PIBF induces dominance of the Th2 immune response, it is tempting to hypothesize that the protein may have a role in anti-tumor immunity and this has attracted the attention of a number of researchers. Srivastava et al. (2004) detected the presence of PIBF mRNA in tumor cell lines, thus suggesting that PIBF might be a novel target for immunotherapy. It is necessary to use specific antibodies as probes to detect the expression of PIBF by normal and/or malignant cells, or to detect its presence in biological fluids. To this end, using polyclonal antibody against lymphocyte-secreted PIBF to test the product of an identified cDNA clone Polgar et al. (2003) have shown that PIBF can be detected in at least two forms—a full-length protein (∼90 kDa) and located in the nucleus, and a short-length form of 34kDa detected in the cytoplasm. The biological activity seems to be mediated by the N-terminal part of the protein molecule, as shown by experiments of the same research group. In these studies, the authors have used polyclonal antibodies against secreted 34kDa protein, or fusion 89-recombinant PIBF-GST protein, and monoclonal antibodies produced against 48kDa recombinant polypeptide. However, it is not clearly stated whether these monoclonal antibodies reacted against the secreted PIBF detected in human biological fluids. Despite numerous attempts, there is no antibody produced so far which can be applied in development of diagnostic tests for detection of PIBF in biological fluids. The aim of our experiments was to produce a monoclonal antibody which would react specifically with PIBF in tissue lysates and biological fluids. The results presented here demonstrate that the monoclonal antibody designated as MAB 3A6 reacts with PIBF in different tests, such as immunoenzyme assay, immunochemistry, Western blot and flow cytometry, which makes it suitable as a basis for the development of clinically applicable tests for assessment of PIBF in patients. 2. Experimental procedures 2.3. GST-PIBF fusion protein Glutathione S transferase-PIBF fusion protein was isolated from E.coli strain cells transformed with a GST-PIBF vector kindly donated by Prof. J. Szekeres-Bartho (Pecs University Medical School, Pecs, Hungary). 2.4. Antibodies A sample of polyclonal anti-PIBF antibody was kindly donated by Prof. Szekeres-Bartho and used as control antibody in all experiments to characterize the selected monoclonal antibody. Additionally, a rabbit polyclonal antibody was produced after injection of rabbits with GST-PIBF following the procedure used in the laboratory and used as a positive control antibody in the experiments. Monoclonal antibodies were produced after immunization of mice with GST-PIBF. Each animal received 5s.c. injections of 10μg GST-PIBF each; the first was in complete Freund's adjuvant and the next injections were with incomplete Freund's adjuvant (Sigma–Aldrich, St Louis, MO, USA). Animals were bled through puncture of the retro-orbital plexus and the sera obtained were tested for reactivity against GST-PIBF. Splenocytes from positively reacting animals were fused with P3U1 cells as described by Kyurkchiev et al. (1988). Supernatants from wells with growing hybridoma cells were simultaneously tested against PIBF-GST or GST by ELISA. Selected 3A6 hybridoma was found to be reacting against GST-PIBF, but was negative against GST. This hybridoma was further cloned by limiting dilution, grown in mass culture and aliquots frozen for storage in liquid nitrogen. The antibody secreted by hybridoma 3A6 was shown to be of the IgG isotype. 2.5. Peripheral blood mononuclear cells (PBMC) PBMC were isolated from blood collected by vein puncture with 7ml vacutainer with LH 119 I.U. (Becton Dikenson, San Jose, CA, USA) as an anticoagulant agent. After centrifugation the cells were resuspended in 9ml serum-free RPMI 1640 (PAA Laboratories GmbH, Pashing, Austria) and overlayered carefully onto 3ml Ficoll-Paque™ PLUS (GE Healthcare, Bio-Sciences AB, Uppsala, Sweden) and centrifuged at 1750rpm for 30min. The layer containing the mononuclear cells was collected with Pasteur pipette and washed in serum-free RPMI 1640. Separated PBMC were analyzed by flow cytometry as described below. In another series of experiments, isolated PBMC were adjusted to a concentration of 1×106cell/ml in complete RPMI medium containing 10% fetal bovine serum (PAA Laboratories GmbH, Pashing, Austria), 2mM l-glutamine (PAA Laboratories GmbH, Pashing, Austria), sodium pyruvate (PAA Laboratories GmbH, Pashing, Austria) and antibiotics, and distributed in a 6-well plate with 3ml added to each well (Orange Scientific, Braine-l’Alleud, Belgium). Progesterone at final concentration of 20μg/ml was added to some of the wells containing PBMC. The cells were cultured for 48h at 37°C, 5% CO2 and the supernatant was collected and cells processed for testing by flow cytometry. 2.6. Enzyme-linked immunosorbent assay (ELISA) ELISA was performed as routinely performed in this laboratory. Briefly, 96-well microwell plates were coated with 1μg/ml GST-PBS in coating buffer, pH 9.0 overnight at 4°C and washed with T-PBS (phosphate buffer, pH 7.4 containing 0.05% Tween 20). Unsaturated binding sites of the plate were blocked with 5% non-fat dry milk in PBS, pH 7.4, for 1h at room temperature (RT) and then the plate was washed as before. Supernatant of hybridoma 3A6 was added to the wells, kept for 2h at RT and washed 3× 5min each with T-PBS. The second antibody was goat anti-mouse IgG antibody labeled with horseradish peroxidase (Sigma–Aldrich, St Louis, MO, USA), which was added at dilution 1/5000 for 1h at RT. After thorough washing in T-PBS, the enzyme reaction was developed with ortho-phenylendiamine (OPD; Sigma–Aldrich, St Louis, MO, USA) in citrate buffer pH 5.0 containing 0.01% hydrogen peroxide, and stopped with 10% H2SO4 after 10min in the dark. The intensity of the color reaction was read at 492nm on an ELISA reader (Labsystems Multiscan Plus, Helsinki, Finland). 2.8. SDS-PAGE and immunoblotting This was performed on 5–20% gradient gels under reducing conditions in the presence of 2-mercaptoethanol (Sigma–Aldrich, St Louis, MO, USA), as done routinely following the basic protocol of Towbin et al. (1979). Following the electrophoresis, proteins were electro-transferred onto a nitrocellulose sheet which had been blocked with 2% non-fat dry milk for 1h at RT and were then incubated for 2h at room temperature in MAB 3A6 supernatant. The nitrocellulose sheet was further incubated in anti-mouse IgG labeled with peroxidase (Sigma–Aldrich, St Louis, MO,USA) for 1h at RT and washed three times for 5min in T-PBS. The blots were visualized by enhanced chemiluminescence (ECL), as recommended by the manufacturer (Amersham Biosciences, Amersham, UK). 2.9. Indirect Immunofluorescence Formalin-fixed paraffin-embedded blocks from term placentae obtained after normal delivery or early placenta (8–10 gestational weeks) were used. The paraffin sections were briefly heated up to 57°C in a thermostat, immersed in xylene three times for 10min each and then hydrated through a series of graded alcohols and then in PBS, pH 7.2. In order to block autoflorescence, the sections were treated with NaBH4 for 40min on ice and then washed with PBS. Sections treated in that way were incubated with MAB 3A6 overnight at 4°C. After washing with PBS, specific binding was detected by anti-mouse FITC (Sigma–Aldrich, St Louis, MO, USA) diluted 1:100 with PBS and incubated for 1h in the dark. Some sections were then countersterstained with hematoxylin (Sigma–Aldrich, St Louis, MO, USA), and all were embedded with Fluoromount-G (Southern Biotech, Birmingham, UK). Serial sections from the same samples with no first antibody or treated with irrelevant monoclonal antibody (anti-IFNgamma MAB 2A9) were used as negative controls. 2.11. Statistical methods The non-parametric two-tailed Mann–Whitney U-test for statistical analysis was used to analyze the changes in expression of PIBF by lymphocytes cultured with or without progesterone. Statistically significant differences were considered at P-value <0.05. 3. Results About 15 days after the fusion, supernatants from wells with growing hybridomas were tested by ELISA and a hybridoma designated MAB 3A6 selected which did not react against control antigens tested (human serum albumin, GST, bovine serum albumin, other GST constructs like GST-NH2) but gave a positive reaction (OD492nm 0.805) against GST-PIBF. When the antigen was coated in serial dilutions, the intensity of the reaction of MAB 3A6 at a constant concentration gradually decreased forming a well-defined dose-dependent curve. In contrast, the anti-GST-PIBF polyclonal serum reacted with rather constant intensity to antigen dilutions in the ranges of 0.13–33ng/ml forming a plateau. Contrary to that when tested in serial dilutions, MAB 3A6 showed a dose-dependent reactivity against a constant concentration of the coated GST-PIBF (Fig. 1A). To test whether MAB 3A6 reacted with the antigen in liquid phase, a sandwich ELISA was developed and the results from these experiments clearly showed that MAB 3A6 bound to the plastic could capture PIBF in the liquid phase and a dose-dependant curve was outlined which demonstrated the specific character of the binding. Well-defined and reliable positive reaction was recorded repeatedly with concentrations of PIBF in the ranges of 0.156–0.312ng/ml of the fusion GST-PIBF protein (Fig. 1B). Further, MAB 3A6 was tested in immunoblot experiments which showed that it reacted with the PIBF moiety of the fusion protein, staining a band with molecular mass of 78kDa. This was expected to be the molecular mass of the fusion protein and additional positive stain bands were seen in the regions of 150kDa (Fig. 2A). The control polyclonal anti-PIBF antibody gave a similar staining pattern, but reacted with GST protein in the region of 30kDa. These results confirm the specific reaction of MAB 3A6 to PIBF. In experiments to develop a reliable assay for detection and measurement of PIBF in biological fluids for clinical purposes, we tested urine samples from pregnant women or patients with prostate carcinoma. These samples were tested by ELISA, sandwich ELISA and immunoblot. The sandwich ELISA worked well with GST-PIBF diluted in PBS, pH 7.4 but, with urine samples, the results were contradictory as a high background was observed, and good results obtained could not be confirmed. Similarly, in the immunoblot experiments, the results were inconsistent because, in some samples, positively stained bands were registered at approximately 34kDa (Fig. 2B) while, in other samples, no reaction could be recorded; importantly, the results were not confirmed in the next experiments when the same samples were stored at −70°C and tested again. Just the same type of inconsistency was observed when polyclonal serum was used as a positive control. It was concluded that urine samples should be tested when fresh and further efforts are needed to develop a reliable assay for detection and measurement of PIBF in biological fluids. The next step in our experiments was to use MAB 3A6 in indirect immunofluorescence. In attempts to determine the localization of PIBF in trophoblast, formalin-fixed paraffin-embedded human term placenta or early placenta (8–10 weeks) sections were incubated with MAB 3A6. A bright green specific fluorescence was observed mainly in cytotrophoblast cells and cells in the mesenchyme of placental villi of early placenta (Fig. 3). Staining of human term placenta was rather homogeneous and it was difficult to distinguish the trophoblast layers or separate cells being stained. Further, MAB 3A6 was used in flow cytometry experiments to detect the expression of PIBF on the surface of PBMC from pregnant donors. In preliminary experiments, it was shown that MAB 3A6 could stain positively lymphocytes treated with the monoclonal antibody and anti-mouse IgG labeled with FITC as a second antibody (Fig. 4A). PBMC isolated from women with normal pregnancy (4–6l.m.) were cultured for 48h in the presence of progesterone (20μg/ml) and then analyzed by flow cytometry for the expression of PIBF. When CD3+ lymphocytes were gated in a side-scattering plot, it was found that, without progesterone, the average percentage of PIBF-expressing lymphocytes was 3.23%, σ=1.24, and, after stimulation with progesterone, the percentage raises to 30.52%, σ=11.52, positive cells (Fig. 4B). Contrary to that, PBMC from 10 healthy women treated in the same way did not change the percentage of the PIBF-expressing PBMC as the average percentage in the absence of progesterone was 1.77%, σ=0.7%, and in the presence of progesterone was 1.23%, σ=0.96. In a separate series of experiments, PBMC from non-pregnant women were analyzed by FACS without pre-culturing with progesterone in vitro and very low percentages of PIBF-positive lymphocytes were detected (n=14; 1.23±0.65%). 4. Discussion Szekeres-Bartho et al., 1996, Szekeres-Bartho et al., 1989a have reported that PIBF is a major factor for a shift of Th1- to Th2-type immune response and suppression of NK cell activity in successful pregnancy. Recently, it was speculated that PIBF might be an important player in tumor defense against the immune response of the host (Check et al., 2001). Detection and quantitative measurement of PIBF in human biological fluids and/or tissue lysates would seem to be a very significant parameter in monitoring pregnancy, recurrent abortion, IVF failure or patients with different carcinomas. However, a reliable and clinically applicable method for assessment of PIBF has not yet been developed. Indeed, Szekeres-Bartho et al. (1989b) have described an ELISA using polyclonal antibody against PIBF which was used to test sera from healthy pregnant women and reported that the concentration of PIBF might have a predictive meaning for pregnancy pathologies such as miscarriage and/or pre-term deliveries. The authors have not published further results with this assay or refined the technique, which would suppose that there might have been some problems with consistency other problems which did not allow the wide application of this method. Check et al. (1996) tried another approach in attempts to look for a method to quantify PIBF in peripheral blood. The authors used immunochemistry staining of lymphocytes in peripheral blood of IVF patients 9 days after the embryo transfer. Patients were distributed in three groups—non-IVF, IVF-ET and frozen ET patients and, when analyzing the results the patients were in two groups—95 with negative and 139 with positive β-hCG levels in serum. PIBF was detected in 49.5% of non-pregnant women and 70.5% of pregnant women, and there was a significant difference between non-pregnant and all pregnant women. Obviously, counting stained cells is a rather laborious method prone to individual variations. More recently, Polgar et al. (2004) has described a competitive ELISA to assess PIBF concentration in urine samples which is quite sophisticated. The test designed seems to be working correctly with concentrations higher than 14ng/ml, and this concentration was determined as a cut-off value to distinguish between non-pregnant and pregnant urine. The assay is not suitable for lower concentrations. Still, the authors claim that it can be used to detect differences in PIBF concentration in normal pregnant women and women with different forms of pregnancy pathologies. However, the major finding was the fact the PIBF can be detected in urine from healthy donors. All together, these data once again demonstrate the need to search for a good reliable assay for measurement of PIBF in urine and serum because the PIBF concentration is considered to be of predictive meaning in pregnancy and cancer patients. Our experiments to measure PIBF concentrations in urine using a sandwich ELISA with MAB 3A6 gave inconsistent results. We have proven that, when MAB 3A6 is used in affinity chromatography it can bind the GST-PIBF (data not presented). This means that the cognate epitope is accessible for the antibody when the protein is in its native confirmation, which is a pre-requisite for the development of a widely applicable laboratory test. Polgar et al. (2004) reported finding a positively stained band with a molecular mass of 34kDa when lysates from term placenta were treated with anti-PIBF polyclonal antibody in immunoblot. That fact prompted us to check the expression of PIBF by trophoblast cells although it has been accepted that PIBF is secreted by lymphocytes after stimulation by the fetus. Staining of term placenta sections was difficult to assess because it was weak and homogeneously located in the cytoplasm. The staining of early placenta by Mab 3A6 is new fact and should be interpreted cautiously. Positively stained are cytotrophoblast cells and cells in the villous mesenchyme, and the staining is perinuclearly located. This finding differs from that of Lachmann et al. (2004) who reported that full-length 90kDa PIBF shows a centrosomal localization in rapidly proliferating cells while the 35kDa form was found diffusely localized in the cytoplasm of the cells. Another aspect of our experiments concerns the expression of PIBF by cells in peripheral blood. Results from these experiments showed that PIBF-positive cells could be detected in about 1.43% of peripheral blood cells of healthy non-pregnant women. Since this is the first report, it is difficult to interpret but it may be speculated that PIBF+ cells represent a specific cell population whose functions are still unknown. In another series of experiments PBMC from pregnant women were cultured in vitro in the presence or absence of progesterone, which is a well-established model for induction of PIBF expression and can be used as a positive control for characterization of the reactions of MAB 3A6. Under these conditions, binding of MAB 3A6 to lymphocytes from PBMC was analyzed by flow cytometry. Gating the CD3+ lymphocyte fraction of PBMC for analysis, it was recorded that stimulation with progesterone for 48h caused a significant increase in the percentage of lymphocytes stained positively and the difference was statistically significant (P<0.05). It is obvious that MAB 3A6 can be used for phenotyping the lymphocytes in peripheral blood by flow cytometry which appears to be a significant asset of this antibody. The data from flow cytometry analysis are different from the findings of Check et al. (1996), who reported that PIBF could be detected in 70.5% of pregnant women, while we found PIBF expression in all pregnant women tested after stimulation with progesterone in vitro. The differences in results can be explained by the different antibodies and methods used in these two studies. In conclusion, a hybridoma was constructed which secretes a monoclonal antibody MAB 3A6 which reacts with PIBF both in liquid phase and coated on plastic or in immunoblot. The antibody can recognize PIBF expressed on the surface of lymphocytes from pregnant women stimulated with progesterone. MAB 3A6 can be used as a basis for development of a clinically applicable assay for detection and measurement of PIBF in biological fluids, or cells expressing this protein, for diagnostics and prognostics of both pregnancy and tumors in humans. Acknowledgements This work was partially supported by Grant No. Genomika 4.1./2005 and Grant VU-L-201/06 by the National Science Fund, Ministry of Education and Science, Sofia, Bulgaria References Check et al., 1996. 1.Check J, Szekeres-Bartho J, O'Slaughnessy A. Progesterone induced blocking factor seen in pregnancy lymphocytes soon after implantation. Am. J. Reprod. Immunol. 1996;35:277–280. Check et al., 2001. 2.Check K, Nazari P, Goldberg J, et al. A model for potential tumor immunotherapy based on knowledge of immune mechanisms responsible for spontaneous abortion. Med. Hypotheses. 2001;57:337–343. Abstract |
Full-Text PDF (79 KB)
|
CrossRef
Kozma et al., 2006. 3.Kozma N, Halasz M, Polgar B, et al. Progesterone-induced blocking factor activates STAT6 via binding to a novel IL-4 receptor. J. Immunol. 2006;176:819–826. MEDLINE Kyurkchiev et al., 1988. 4.Kyurkchiev S, Surneva-Nakova Ts, Ivanova M, et al. Monoclonal antibodies to porcine zona pellucida that block the initial stages of fertilization. Am. J. Reprod. Immunol. Microbiol. 1988;18:11–18. MEDLINE Lachmann et al., 2004. 5.Lachmann M, Gelbmann D, Kálmán E, Polgár , et al. PIBF (progesterone induced blocking factor) is overexpressed in highly proliferating cells and associated with the centrosome. Int. J. Cancer. 2004;112:51–60. MEDLINE |
CrossRef
Polgar et al., 1999. 6.Polgar B, Barakonyi A, Xynos I, Szekeres-Bartho J. The role of γδ TCR positive cells in pregnancy. Am. J. Reprod. Immunol. 1999;41:239–244. Polgar et al., 2003. 7.Polgar B, Kispal G, Lachmann M, et al. Molecular cloning and immunologic characterization of a novel cDNA coding for progesterone-induced blocking factor. J. Immunol. 2003;171:5956–5963. MEDLINE Polgar et al., 2004. 8.Polgar B, Nagy E, Miko E. Urinary progesterone-induced blocking factor concentration is related to pregnancy outcome. Biol. Reprod. 2004;71:1699–1705. MEDLINE |
CrossRef
Srivastava et al., 2004. 9.Srivastava MD, Check JH, Szekeres-Bartho J. Expression of progesterone-induced blocking factor (PIBF) by human cell lines. J. Clin. Oncol. 2004;22:3191;.
CrossRef
Szekeres-Bartho et al., 1985. 10.Szekeres-Bartho J, Kilar F, Falkay G, et al. The mechanism of the inhibitory effect of progesterone on lymphocyte cytotoxicity. I. Progesterone-treated lymphocytes release a substance inhibiting cytotoxicity and prostaglandin synthesis. Am. J. Reprod Immunol. Microbiol. 1985;9:15–18. MEDLINE Szekeres-Bartho et al., 1989a. 11.Szekeres-Bartho J, Autran B, Debre P, et al. Immunoregulatory effects of a suppressor factor from healthy pregnant women's lymphocytes after progesterone induction. Cell. Immunol. 1989;122:281–294. MEDLINE |
CrossRef
Szekeres-Bartho et al., 1989b. 12.Szekeres-Bartho J, Varga P, Pejtsik B. ELISA test for the detection of an immunological blocking factor in human pregnancy serum. J. Reprod. Immunol. 1989;16:19–29. Abstract |
Full-Text PDF (520 KB)
|
CrossRef
Szekeres-Bartho and Wegmann, 1996. 13.Szekeres-Bartho J, Wegmann TG. A progesterone-dependent immunomodulatory protein alters the Th1/Th2 balance. J. Reprod. Immunol. 1996;31:81–95. Abstract |
Full-Text PDF (1300 KB)
|
CrossRef
Szekeres-Bartho et al., 1996. 14.Szekeres-Bartho J, Faust Z, Varga P, et al. The immunological pregnancy protective effect of progesterone is manifested via controlling cytokine production. Am. J. Reprod. Immunol. 1996;35:348–351. Towbin et al., 1979. 15.Towbin H, Staehelin T, Gordon J. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci. U.S.A. 1979;76:4350–4356. MEDLINE |
CrossRef
a Department of Molecular Immunology, Institute of Biology and Immunology of Reproduction, Bulgarian Academy of Sciences, Bulgaria b Department of Clinical Laboratory and Clinical Immunology, University Hospital ‘St.I.Rilski’, Medical University Sofia, Bulgaria c University Hospital of Obstetrics and Gynecology ‘Maichin dom’, Sofia, Bulgaria d Ob/Gyn Hospital ‘Dr Shterev’, Sofia, Bulgaria  Corresponding author at: Department of Molecular Immunology, Institute of Biology and Immunology of Reproduction, Bulgarian Academy of Sciences, 73 Tzarigradsko shosse, 1113 Sofia, Bulgaria. Tel.: +359 2 723890; fax: +359 2 720925.
PII: S0165-0378(07)00271-9 doi:10.1016/j.jri.2007.12.001 © 2007 Elsevier Ireland Ltd. All rights reserved. | |
|
FULL TEXT00271-9/journalimage?img=PIIS0165037807002719.gr1.lrg.gif&fig=fig1&kwhquery=&issn=0165-0378&ishighres=false&allhighres=false&free=yes','journalimage');)
00271-9/journalimage?img=PIIS0165037807002719.gr2.lrg.gif&fig=fig2&kwhquery=&issn=0165-0378&ishighres=false&allhighres=false&free=yes','journalimage');)
00271-9/journalimage?img=PIIS0165037807002719.gr3.lrg.gif&fig=fig3&kwhquery=&issn=0165-0378&ishighres=false&allhighres=false&free=yes','journalimage');)
00271-9/journalimage?img=PIIS0165037807002719.gr4.lrg.gif&fig=fig4&kwhquery=&issn=0165-0378&ishighres=false&allhighres=false&free=yes','journalimage');)