Physiological importance of the connective tissue in the human amnion. Role of magnesium

ARTICLE

Auteur(s) : Michel Bara¹, Philippe Moretto², Jean Durlach³, Andrée Guiet-Bara¹

¹ Physiology and Physiopathology Laboratory, University Pierre et Marie Curie, 4, Place Jussieu, 75252 Paris cedex 05, France; ² Centre d'Etudes Nucléaires de Bordeaux-Gradignan, Interface Physique-Biologie, Le Haut Vrigneau, BP 130, 33175 Gradignan cedex, France; ³ SDRM, 64 rue de Longchamp, 92200 Neuilly/Seine, France

Introduction

The human amniotic membrane is a multilayered membrane, with principally an epithelial cell layer and a connective tissue (compact layer + fibroblast layer) (Figure 1).
The connective tissues of the amnion and of the chorion include two layers containing mesenchymal cells, the fibroblast layer of the amnion and the reticular layer of the chorion. The exact cellular composition of the connective tissues of the amnion and chorion has been controversial. They have been reported to consist primarily of macrophages, based upon possession of specific cell markers [1, 2] or HLA-DR/leucocyte common antigen positive dendritic cells [1, 3]. However, they have also been reported to consist exclusively of fibroblasts [4] and to include myofibroblasts, as evidenced by the location of numerous cytoplasmic bundles of filaments of the amnion [5] and chorion [6].
More recent studies have reported that the amnio-chorion is the center of important activities. For example: – in the human amnion, there is an inhibitor of metalloproteinase (TIMP), a glycoprotein, which inhibits type I and IV-specific collagenases and inhibits amniotic membrane invasion [7, 8]; – type V collagen is found in the chorion, in the reticular and in the trophoblast layers. Type VI collagen is present in the amnion and the reticular layer. The extensive and continuous networks formed by these collagens may be a major factor responsible for the mechanical integrity of the fetal membranes [9]; – a restricted zone of extreme altered morphology, characterized by marked swelling and disruption of the connective tissue, thinning of the trophoblast layer has been identified in the rupture site of all patients. Morphometric analysis of the thickness of membrane layers showed that these changes and the ratio between the thickness of the connective tissue layers and that of trophoblast and decidua were significant between the zone of extreme altered morphology and the rest of the membranes [10]; – integrins localized laterally may play a role in cell-cell interactions and cell-extracellular matrix interactions. Beta 1 (alpha 1 beta 1, alpha 5 beta 1) integrins are probably involved in cell-matrix interactions in the connective layers which are rich in collagenes and fibronectin [11]; – human chorionic and amniotic membranes are a source of interleukin-8 mRNA and peptide [12]; – human amnion contains a novel laminin variant, laminin 7, which like laminin 6, covalently associates with laminin 5 to promote stable epithelial-stromal attachment [13]; – exposed connective tissue of amniotic membranes is demonstrated to trigger platelet adhesion, aggregation and activation and platelets are shown to seal a strandardized puncture in fetal membranes [14]; – the demonstration of expression of large tenascin mRNA isoforms supports the concept that fetal membranes at term are a site of active tissue remodeling [15]; – an area of the fetal membranes, within the rupture tear is characterized by marked swelling of the amniotic and chorionic connective tissue layers, consistent with structural weakness, and a marked reduction of the thickness of both the cytotrophoblast and decidual layers. Cellular activities associated with myofibroblastic differentiation in the reticular layer of the chorion may be associated with the observed connective tissue changes, fetal membrane rupture and labour [16]; – fetal membranes overlying the cervix in patients prior to and during labour, and within the rupture tear after spontaneous delivery at term, exhibit altered morphology. In comparison to mid-zone fetal membrane biopsies, these regions are characterized by increased expression of the matricellular protein osteonectin or SPARC (secreted Protein Acidic and Rich in Cysteine). Osteonectin has been implicated in the regulation of extracellular matrix turnover, and its pattern of expression suggests a role in the regional connective tissue and cytotrophoblastic changes proposed to be involved in the cleavage and rupture of fetal membranes [17].
The morphological features in the connective tissue layers of the amnion and chorion in the “Zone of Altered Morphology” are consistent with extracellular matrix degradation in this zone. Mesenchymal extracellular matrix degradation, detected during situations of tissue wounding, is associated with the activation or differentiation of resident mesenchymal fibroblastic cells into myofibroblasts. The data of McParland et al. [16] suggest that cellular activities associated with myofibroblastic differentiation in the reticular layer of the chorion may be associated with the observed connective tissue changes, fetal membrane rupture and labour.
This intensive activity is localized in the epithelial layer and in the connective tissue. Moreover, in the human amniotic membrane, the ionic transfer is assumed largely by the epithelial layer [18-21] and traditionally, the connective tissue (compact layer + fibroblast layer) has a structural support role but there are few indications of the ionic distribution in the layers. Micro PIXE (Particle Induced X-ray Emission) is one of the few methods of microanalysis which permit a simultaneous detection of most minerals in various tissue layers. It provides unique possibilities to reveal directly the distributions of these elements at the cell scale. For this purpose, the concentrations of elemental ions in the epithelial layer and in the connective tissue are examined either in untreated reference samples or after incubation in physiological medium (composition near to amniotic fluid composition) and in high magnesium physiological fluids (magnesium salts used in case of tocolysis) to determine the role of each layer in the ionic transfert and distribution.

Material and Methods

Human amnion sampling

Specimens of human amnion, isolated from the placental zone of the amniotic sac, were obtained after 8 normal deliveries at term. To minimise the diffusion of physiologically active ions and structural deterioration of the tissue once it has been extracted from its natural environment, the tissue must be fixed. Artefacts will inevitably be induced according to the particular fixation process chosen, so that the minimisation of process of induced artefacts is the aim at all stages of specimen preparation. An entirely physical process which fixes the tissue in its current physiological state and restricts diffusion of unbound elements is cryofixation [22]. For each specimen, three strips were isolated, the first one being immediately quench frozen in isopentane cooled with liquid nitrogen, and this, without any rinsing procedure. This sample was considered as a reference sample. The second one was transferred in Hanks'solution, a physiological fluid (composition in mmol/l: NaCl 150, KCl 6, MgSO₄ 0.5, MgCl₂ 0.5, CaCl₂ 1, glucose 5.5, NaH₂PO₄-KH₂PO₄-NaHCO ₃ 1) at 37 °C and pH 7.4. For the third one, magnesium (MgCl₂, 2 mM) was added to normal Hanks'solution. After one hour incubation, the two strips were cryofixed as previously. The three strips were kept in liquid nitrogen until sectioned. Sectioning was performed at – 30 °C using a cryomicrotome (Reichert-jung frigocut 2 800). Thin frozen sections (thickness in the range of 20 to 30 µm) were collected on the knife, placed directly on thin formvar^R foils of about 20 µg/cm² and kept in the cryostat for several hours until complete freeze-drying. The slides were then stored in a dessicator over silica-gel prior to analysis. The morphology of the amnion sections was elucidated using light microscopy of adjoining sections mounted on glass slides and stained with haematocylin and eosin.

Microanalysis and data processing

The analysis was performed using the microprobe facility in Bordeaux [23]. Well defined parts of amnion slides were chosen including both epithelial cells and connective tissue. These regions were irradiated with a 1 MeV proton beam focalized down to a 2 µm spot diameter. The beam current, measured on the target was 150 pA and the total collected charge was 0.5 µC. The extension of the scan ranging, according to the sample structure, was chosen between 50 × 50 µm². PIXE and RBS (Rutherford Backscattering Spectrometry) analysis were simultaneously carried out in order to insure mass standardization. Sodium, potassium, calcium, magnesium, chlorine and phosphore were determined. X-rays were detected using a 80 mm² Si(Li) solid state detector (Link system) fitted with a thin beryllium window (8 µm), which allowed us to measure the NaK line with low attenuation. The backscattered particles were detected at 135° of the beam with a Si 20 mm² detector, thus allowing carbon, nitrogen and oxygen measurements and beam current monitoring. Elemental mapping of Na, K, Mg, Ca, Cl and P revealed a high compartmentalization of ionic species allowing thus a precise delimitation of the epithelial layer (EL) and of the connective tissue (CT) on the whole scale area. An off-line specific treatment of data permitted us to extract X-rays and backscattered particles spectra issuing from the previously defined subregions [24].

Data reduction

Quantitative results expressed in term of dry mass were obtained using the following scheme: all PIXE spectra were fitted with GUPIX (guelf pixe software package) software [25] in order to obtain absolute elemental masses. The organic mass of the analyzed specimen was then assessed using backscattered particles. For that prupose, RBS spectra were treated with an extension of the RUMP (Rutherford Universival Manipulation Program) code [26], a program developed by Moretto and Razafindrabe [27], taking into account the heterogeneity of the sample thickness, non-Rutheford backscattering cross-sections and the autoabsorption of low X-rays. Unfortunately, this program was not available at the beginning of this study. In order, to include all experimental data in the reported results, we therefore expressed quantitative values using elemental ratios (Na/S, K/S, Cl/S, Ca/S, Mg/S, P/S). Sulfur was estimated to be the best reference element because of its unvarying concentration in EL and CT, whether the amnion was incubated or not. This point was checked using the mass normalization procedure above described.
The statistical comparisons between reference and incubated specimens were carried out with Student's test. The values of the significance level (p) of 0.05 and less were considered significant.

Results

Ionic distribution in epithelial layer

Figure 2 indicates the ionic distribution in the human amniotic epithelial layer.
– Control sample: in control, the ionic concentration distribution was ranged in the followed sequence: K > Na > Cl = Ca > Mg.
– Hanks'solution: in a physiological medium, the Na⁺, Cl ^–  and Mg²⁺ concentrations were significantly (p < 0.05) increased with regard to control sample, while the K⁺ and Ca²⁺ concentrations remained constant (p = 0.16 and p = 0.15).
– Hanks'solution + MgCl₂: in a physiological medium + 2 mM MgCl₂, the Na⁺ and Ca²⁺ concentrations remained constant (p = 0.1 and 0.07) with regard to control sample and Hanks sample, while K⁺ and Cl ^–  concentrations were significantly decreased and Mg²⁺ concentration significantly increased (p < 0.01).

Ionic distribution in connective tissue

Figure 3 indicates the ionic distribution in the compact layer and the fibroblast layer of the amniotic connective tissue.
– Control sample: in control, the ionic concentration distribution was ranged in the followed sequence: K > Na > Cl > Ca > Mg.
– Hanks'solution: in a physiological medium, the Na⁺, Cl ^– , Ca²⁺ and Mg²⁺ concentration was significantly increased (p < 0.05), while the K⁺ concentration remained constant with regard to control values (p = 0.3).
– Hanks'solution + MgCl₂: in a physiological medium + 2 mM MgCl₂, the Ca ²⁺ concentration remained constant (p = 0.2) with regard to control sample and Hanks sample, while the Na⁺, K⁺ and Cl ^–  concentrations were significantly decreased and Mg²⁺ concentration significantly increased (p < 0.01) with regard to Hanks samples.

Comparison between the ionic distribution in the epithelial layer and connective tissue

– Sodium: In control samples, the Na⁺ concentration was identical in EL and CT (p = 0.07). Concentrations were strongly increased in Hanks' solution, without modifications between EL and CT and the addition of MgCl₂ had no significant effet on the Na⁺ concentration in EL (p = 0.06), but the decreased one in CT (p < 0.05).
– Potassium: In control samples, the K⁺ concentration was identical in EL and CT (p = 0.57). The K⁺ concentration remained identical in EL and CT after incubation in Hanks'solution (p = 0.16 and 0.60), and after the addition of MgCl₂.
– Chlorine: In control samples and in experimental medium, the Cl ^–  concentration was identical in EL and CT (p = 0.13).
– Calcium: In control samples and in experimental medium, the Ca²⁺ concentration was higher in EL than in CT (p < 0.001).
– Magnesium: In control samples and in experimental medium, the Mg²⁺ concentration was higher in EL than in CT (p < 0.001).

Discussion

McParland et al. [16] have identified a specific alteration in the phenotype of mesenchymal cells in connective tissue layers of the fetal membranes, consistent with the differentiation of myofibroblasts in an anatomically defined region of the fetal membranes, i.e. membranes lying in the lower uterine pole and associated with the process of parturition at term. The definition of a myofibroblast is essentially one dependent upon the identification of certain ultrastructural features such as a well developed fibrillar cytoplasm, a cell to matrix fibronectus, gap junctions and in incomplete encapsulating basal lamina [29]. Cells with such features have been described in the amniochorion [5-6]. A number of functions have been ascribed to these cells, including turnover of the extracellular matrix. Moreover, regional changes in thickness of the constituent layers of the membranes have been reported, in particular, the thickness of the connective tissue layers of the amnion was significantly thicker in biopsies associated with the rupture line compared to the mid-zones [16].
The major cellular capacity for fibrillar collagen synthesis which provides the major tensile strength of the membranes appears to tie within the connective tissue cells of the fetal membranes [30].
The qualitative results indicate that, in control samples, the common monovalent ions (Na⁺, K⁺, Cl ^– ) are localized in the same ratio in the “exchanging layer” (epithelial layer) and in the “supporting layer” (connective tissue, CT, i.e. compact layer + fibroblast layer). The divalent cations Ca²⁺ and Mg²⁺ are significantly higher in EL than in CT. This data seems to be normal for Mg which is an intracellular cation which is bound to cellular constituents such as membrane or nucleic acids [28].
In incubated specimens in Hanks'solution, the Na⁺, K⁺ and Cl ^–  levels are identical in EL and CT but strongly higher than in control samples. These data indicate that the EL, a layer which is responsible for the major part of ionic transfert, and CT stock the ions and that CT acts as a buffer which can fix minerals. The Ca²⁺ level increased in CT, without difference with regard to EL level, this data implies that Ca supplementation (1 mM in Hanks'solution) induces a storage of Ca ions in the connective tissue, particularly in the compact layer.
The addition of MgCl₂ to Hanks'solution induces an Mg level increase in the two layers and a decrease of the monovalent ion concentrations with regard to Hanks'solution, indicating an ionic storage in the EL and CT. These data indicate that Mg²⁺ ions are an important element of the ionic storage in the EL and in the CT.
The fact that the connective tissue fixes the ions is of great interest in the case of pathological pregnancies, particularly those complicated by polyhydramnios and oligohydramnios [31], associated with congenital abnormalities [32]. Indeed, the amniotic epithelial cell layer from normal pregnancies was 8-12 µm thick. In polyhydramnic pregnancies, the cell layer varied widely from the normal thickness to as much as 18 to 56 µm in diabetic patients. Other ultrastructural changes observed in pregnancies complicated by polyhydramnios were abnormal microvilli, diminished intercellular channels, increased tonofilaments and decreased rough-surfaced endoplasmic reticulum. In pregnancies complicated by oligohydramnios, the thickness of the cell layer was 3 to 6 µm. The amniotic epithelial cells had increased tonofilaments, decreased desmosomes, collapsed and fused intercellular canals and caused a marked decrease of the Golgi apparatus and the rough-surfaced endoplasmic reticulum. The sparse microvilli were short, plump and had bizarre shapes [31]. Our data indicates that the connective tissue stocks the ions present in the external medium and it would be interesting to study the ionic distribution in the case of pathological pregnancies.
The data obtained in this study are also fundamental in the case of spontaneous rupture in labor. Indeed, the changes and the ratio between the thickness of the connective tissue layers and that of the trophoblast and decidua were significant between the zone of extreme altered morphology and the rest of the membrane [33]. Following spontaneous term birth, an area of the fetal membrane overlying the cervix was characterized by an increased thickness of the connective tissue layer [34] and marked swelling of the amniotic and connective tissue layers was consistent with structural weakness [16]. The data of the stockage of ions in connective tissue may be in relation with the increase thickness. This point might be developped in further studies.

Conclusion

This study has demonstrated the role of amniotic connective tissue in the stockage of elemental ions. This data may explained some of electrophysiological results obtained after electrophysiological studies, introducing the fact that the external medium composition may be modified by the ions present in the connective tissue and the fact that the connective tissue may intervene in the case of pathological pregnancies.

References

1. Sutton L, Mason DY, Redman CW. HLA-DR positive cells in human placenta. Immunology 1983; 49: 103-12.

2. Bulmer J, Johnson PM. Macrophage populations in the human placenta and amniochorion. Clin Exp Immunol 1984; 57: 393-403.

3. Jenkins DM, O'Neill M, Johnson PM. HLDA-DR-positive cells in the human amniochorion. Immunol Lett 1983; 6: 65-7.

4. Schmidt W. The amniotic fluid compartment: the fetal habitat. Berlin: Springer-Verlag, 1992.

5. Wang T, Schneider J. Myofibroblasten im bindegewebe des menschlichen amnions. Z. Geburtshilfe Perinatol 1982; 186: 164-9.

6. Wang T, Schneider J. Fine structure of human chorionic membrane: ultrastructural and histological examination. Arch Gynecol 1983; 233: 187-98.

7. Schultz RM, Silberman S, Persky B, Bajkowski AS, Carmichael DF. Inhibition by human recombinant tissue inhibitor of metalloproteinases of human amnion invasion and lung colonization by murine B16-F10 melanoma cells. Cancer Res 1988; 48: 5539-45.

8. Vadillo-Ortega F, Sadowsky DW, Haluska GJ, Hernandez-Guerrero C, Gevara-Silva R, Gravett MG, Novy MJ. Identification of matrix metalloproteinase-9 in amniotic fluid and amniochorion in spontaneous labor and after experimental intrauterine infection or interleukin-1 beta infusion in pregnant rhesus monkey. Am J Obstet Gynecol 2002; 186: 128-38.

9. Malak TM, Ockleford CD, Bell SC, Dalgleish R, Bright N, Macvicar J. Confocal immunofluorescence localization of collagen types I, III, IV, V and VI and their ultrastructural organization in term human fetal membranes. Placenta 1993; 14: 385-406.

10. Malak TM, Bell SC. Distribution of fibrillin-containing microfibrils and elastin in human fetal membranes: a novel molecular basis for membrane elasticity. Am J Obstet Gynecol 1994; 171: 195-205.

11. Malak TM, Bell SC. Differential expression of the integrin subunits in human fetal membranes. J Reprod Fertil 1994; 102: 269-76.

12. Fortunato SJ, Menon R, Swan KF. Amniochorion: a source of interleukin-8. Am J Reprod Immunol 1995; 34: 156-62.

13. Champliaud MF, Lunstrum GP, Rousselle P, Nishiyama T, Keene DR, Burgeson RE. Human amnion contains a novel laminin variant, laminin 7, which like laminin 6, covalently associates with malinin 5 to promote stable epithelial-stromal attachment. J Cell Biol 1996; 132: 1189-98.

14. Louis-Sylvestre C, Rand JH, Gordon RE, Salafia CM, Berkowitz RL. In vitro studies of the interactions between platelets and amniotic membranes: a potential treatment for preterm premature rupture of the membranes. Am J Obstet Gynecol 1998; 178: 287-93.

15. Bell SC, Pringle JH, Taylor DJ, Malak TM. Alternatively spliced tenascin-C mRNA isoforms in human fetal membranes. Mol Hum Reprod 1999; 5: 1066-76.

16. McParland PC, Taylor DJ, Bell SC. Myofibroblast differentiation in the connective tissues of the amnion and chorion of term human fetal membranes-implications for fetal membrane rupture and labour. Placenta 2000; 21: 44-53.

17. McParland PC, Bell SC, Pringle JH, Taylor DJ. Regional and cellular localization of osteonectin/SPARC expression in connective tissue and cytotrophoblastic layers of human fetal membranes at term. Mol Hum Reprod 2001; 7: 463-74.

18. Bara M, Guiet-Bara A. Cellular and shunt conductance of human isolated amnion: standard representation of Na⁺ transport system. Med Sci Res 1987; 15: 975-6.

19. Bara M, Guiet-Bara A, Durlach J. Analysis of magnesium membraneous effects: binding and screening. Magnes Res 1988; 1: 29-33.

20. Bara M, Guiet-Bara A, Durlach J. A qualitative theory of the screening-binding effects of magnesium salts on epithelial cell membrane: a new hypothesis. Magnes Res 1989; 2: 243-7.

21. Bara M, Guiet-Bara A. Inhibitor effects on the ionic exchanges through the human amniotic epithelial cell membranes. Cell Mol Biol 1994; 40: 1165-71.

22. Kirby BJ, Legge GJF. The preparation of biological tissue for a trace element analysis on the proton microprobe. Nucl Instr Meth 1993; B77: 268-74.

23. Llabador Y, Bertault D, Gouillaud JC, Moretto P. Advantages of high speed scanning for microprobe analysis of biological samples. Nucl Instr Meth 1990; B49: 435-40.

24. Moretto P, Llabador Y, Simonoff M, Bara M, Guiet-Bara A. Quantitative mapping of intracellular cations in human amniotic membrane. Nucl Instr Meth 1993; B77: 275-81.

25. Maxwell JA., Campbell JL, Teesdale WJ. The guelph PIXE software package. Nucl Instr Meth 1989; B43: 218-30.

26. Doolittle LR. Algorithms for the rapid stimulation of Rutherford backscattering spectra. Nucl Instr Meth 1985; B9: 344-51.

27. Moretto P, Razafindrabe L. Simulation of RBS spectra for quantitative mapping of inhomogeneous biological tissue. Nucl Instr Meth 1995; B104: 171-5.

28. Durlach J, Bara M. Le magnésium en biologie et en médecine. Ed. Med Inter, 403pp, Cachan, France, 2000.

29. Schurch W, Seemayer T, Gabbiani G. The myofibroblast: a quarter century after its discovery. Am J Surg Pathol 1998; 22: 141-7.

30. Casey ML, MacDonald PC. Interstitial collagen synthesis and processing in human amnion: a property of the mesenchymal cells. Biol Reprod 1996; 55: 1253-60.

31. Hebertson RM, Hammond ME, Bryson MJ. Amniotic epithelial ultrastructure in normal, polyhydramnic and oligohydramnic pregnancies. Obstet Gynecol 1986; 68: 74-9.

32. Gutierrez R, Paramo F, Alfaro M, Huerta V. Ultrasonographic diagnosis of polyhydramnios and its association with congenital malformations. Gynecol Obstet Mex 1996; 64: 1-5.

33. Malak TM, Bell SC. Structural characteristics of term human fetal membranes: a novel zone of extreme morphological alteration within the rupture site. Br J Obstet Gynaecol 1994; 101: 375-86.

34. McLaren J, Malak TM, Bell SC. Structural characteristics of term human fetal membranes prior to labour: identification of an area of altered morphology overlying the cervix. Hum Reprod 1999; 14: 237-41.