- Email: [email protected]
Secretion and transfer of the thyroid hormone binding protein transthyretin by human placenta
Placenta 33 (2012) 252e256
Contents lists available at SciVerse ScienceDirect
Placenta journal homepage: www.elsevier.com/locate/placenta
Secretion and transfer of the thyroid hormone binding protein transthyretin by human placenta R.H. Mortimer a, b, c, *, K.A. Landers a, B. Balakrishnan e, H. Li a, M.D. Mitchell d, e, J. Patel a, b, K. Richard a, b a
Conjoint Endocrine Laboratory, Royal Brisbane and Women’s Hospital, Herston 4029, Brisbane, Australia School of Medicine, The University of Queensland, Herston 4006, Brisbane, Australia c Disciplines of Medicine, Obstetrics and Gynaecology, The University of Queensland, Herston 4006, Brisbane, Australia d University of Queensland Centre for Clinical Research, Building 71/918, Royal Brisbane & Women’s Hospital Campus, Herston QLD 4029, Australia e Liggins Institute, The University of Auckland, Private Bag 92019, Victoria Street West, Auckland 1142, New Zealand b
a r t i c l e i n f o
a b s t r a c t
Article history: Accepted 3 January 2012
Context: The thyroid hormone and retinol binding protein transthyretin (TTR) is synthesised by human trophoblasts. Polarised JEG-3 choriocarcinoma cells grown in bicameral chambers secrete TTR predominantly apically but also basally and these cells and human trophoblasts also take up TTR suggesting that there may be a placental TTR shuttle that participates in materno-fetal transfer of thyroid hormones and retinol. Objectives and methods: Our objective was to investigate TTR secretion into the maternal and fetal circuits of the ex vivo dually perfused placental lobule to confirm that placenta secretes TTR into the fetal circulation. We also investigated translocation of Alexa Fluor-594 labelled TTR from incubation medium into the fetal placental capillaries in early (14e15 weeks) and term placental villus explants. Results: The perfused placental lobule secretes TTR into the maternal and fetal circuits. Secretion in both circuits is linear with time and is predominantly into the maternal circuit (mean maternal/fetal ratio 99.4 25.6). The mean data fitted well to a three compartment mathematical model (maternal circuit, placenta and fetal circuit, constant secretion of TTR and return of maternal circuit TTR to the placental compartment). Explants from early (14e15 weeks) and late (38e40 weeks) placentas translocated fluorescently labelled TTR from medium to villus (fetal) capillaries. Conclusions: Our results confirm that human placenta secretes TTR into maternal and fetal circulations and supports the hypothesis that placental TTR secreted into the maternal placental circulation can be taken up by trophoblasts and translocated to the fetal circulation, forming a TTR shuttle system. This may have important implications for materno-fetal transfer of thyroid hormones, retinol/retinol binding protein and xenobiotics (such as polychlorinated biphenyls) all of which bind to TTR. Crown Copyright Ó 2012 Published by Elsevier Ltd. All rights reserved.
Keywords: Placenta Transthyretin Secretion Transport
1. Introduction Transthyretin (TTR), a 56 KD homotetrameric protein that binds thyroid hormone (TH) and retinol binding protein, is synthesised and secreted by liver, choroid plexus, pancreas and retina [1,2]. Hepatic TTR, thyroxine binding globulin and albumin serve as important binding proteins for circulating TH [3], so that only about 0.03% of circulating T4 and 0.3% of circulating T3 are free [4]. The choroid plexus is a rich source of TTR which is secreted into the
* Corresponding author. Office of Health and Medical Research, Level 13, Queensland Health Building, 147-163 Charlotte St, Brisbane Q4001, Australia. Tel.: þ61 7 3234 1373; fax: þ61 7 3234 0107. E-mail addresses: [email protected], [email protected] (R.H. Mortimer).
cerebro-spinal fluid (CSF), perhaps contributing to TH delivery to brain [5]. The retina also makes TTR and this may be involved in TH transport within the eye [6]. Although early studies reported the presence of TTR in human yolk sac [7] the synthesis and secretion of TTR by human trophoblast was not reported until 2005 [8]. It was subsequently shown that human trophoblast and the human choriocarcinoma cell line JEG-3 internalise radiolabelled TTR [9]. Internalisation is enhanced by the presence of unlabelled T4 which induces tetramerization of TTR. T4 cross linked to TTR is also internalised. Using polarised JEG3 cells grown in bicameral chambers we demonstrated secretion of TTR predominantly via apical but also by basal cell membranes. These findings suggested that TTR might participate in a shuttle system in which trophoblast TTR is secreted into the maternal placental circulation and internalised back into the trophoblast.
0143-4004/$ e see front matter Crown Copyright Ó 2012 Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.placenta.2012.01.006
R.H. Mortimer et al. / Placenta 33 (2012) 252e256
Such a system would ensure high concentrations of TTR at the maternal surface of chorionic villi and, as TTR has greater affinity for T4 than T3, could provide a mechanism for preferential transfer of maternal T4 into trophoblast and, possibly, then into the fetal circulation. To further explore the hypotheses that the human placenta secretes TTR into maternal and fetal circuits and that TTR taken up into trophoblast via the apical membrane can be translocated into the fetal placental circulation we examined secretion of TTR into maternal and fetal circulations using dually perfused isolated human placental lobules. In vitro uptake of Alexa-labelled TTR by human term placental explants was also examined. The results indicate that while TTR secretion by human placenta is predominantly into the maternal circuit significant amounts appear in the fetal circuit. Incubation of early (14e15 weeks) and late (36e38 weeks) placental explants with Alexa-labelled TTR demonstrate translocation of TTR from incubation medium to villous capillaries, part of the fetal circulation. 2. Methods
253
tetramethyl-benzidine liquid substrate (Sigma) and absorbance read at 370 nm with a VersaMaxÔ Microplate reader. 2.3. ELISA validation 2.3.1. Standard curves and sensitivity The working ranges for the calibration curves were 5.5e545.5 pmol/L (300e3000 pg/mL). The curves generated were linear with a correlation coefficient of r 0.975. The lowest TTR concentration measurable outside the range of two standard deviations was determined as 35.5 pmol/L (CV 9%). 2.3.2. Precision data Assay variation was determined by its intra- and inter-assay reproducibility. Intra-assay reproducibility was determined using TTR concentrations in a range of 18.12e181.2 pmol/L (1000e10,000 pg/mL). Intra-assay reproducibility (coefficients of variation) was 2.28%. A total of nine replicates were measured on the same microplate on the same day using the ELISA method. To calculate inter-assay reproducibility, the ELISA assay was run on five different days in nonuplicate using the TTR concentrations mentioned above. Inter-assay coefficients of variation were between 8 and 21%. 2.3.3. Accuracy TTR recovery was determined in duplicate and triplicate on two different days (n ¼ 5) with 7 different concentrations of TTR (5.5e545.5 pmol/L). The recovery of added TTR ranged from 59 to 140% with coefficients of variation from 8 to 22%.
2.1. Placental perfusion Seven term placentas were obtained with informed consent from healthy nonsmoking women undergoing elective caesarean section in Auckland City Hospital (Auckland, New Zealand). This use was approved by the regional ethics committee. Ex vivo dual perfusion (maternal and fetal circuits) was done by a well-validated method [10e12]. In brief, perfusions were established within 30 min of delivery by catheterisation of a fetal (chorionic) vessel and perfusate circulated through the fetal side using a digitally controlled pump. The perfused cotyledon was then excised from the placenta and mounted fetal side up in the perfusion chamber. The maternal circuit was then accessed by cannulas inserted through the maternal placental surface. Initially perfusates in both compartments were recirculated for 1 h then replaced by fresh perfusate. The perfusate consisted of phenol red-free medium-199 (M-3769; SigmaeAldrich, St Louis, MO) cell culture media supplemented with 25 g/L polyvinylpyrrolidone PVP-40 (SigmaeAldrich), 1 g/L bovine serum albuminefraction V (SigmaeAldrich), 2 g/L glucose (SigmaeAldrich), 20,000 IU/L heparin (Multiparin, CP Pharmaceuticals Ltd, Wexham, UK), and 48 mg/L gentamicin reagent solution (GIBCO; Invitrogen, Auckland, New Zealand). Maternal perfusate was gassed with 95% O2 and 5% CO2 while fetal perfusate was gassed with (5% N2 and 5% CO2). After equilibration, perfusate was replaced with fresh perfusate containing antipyrine (SigmaeAldrich, St Louis, MO) at 40 mg/mL and fluorescein isothiocyanate (FITC)-dextran, (SigmaeAldrich, St Louis, MO) 12.5 mg/mL. The perfusions were then carried on for up to 3 h in recirculating mode. 7.5 mL of perfusate was collected at each time interval from both circuits. Antipyrine was measured by HPLC and FITCdextran was measured spectrofluorometrically [10]. Perfusions were continued for up to 180 min. Viability and metabolic activity of perfused cotyledons was assessed by measuring glucose utilization, lactate production (Hitachi 902 autoanalyzer; Hitachi High Technologies Corp, Tokyo, Japan), and human chorionic gonadotropin (hCG) secretion (enzyme-linked immunosorbent assay). Stringent quality control measures were observed, namely a fetal flow rate of 4 mL/min; a maternal flow rate of 10 mL/min; a fetal pressure 30e60 mmHg; a pH 7.2e7.4; and a feto-maternal fluid shift of less than 3 mL/h. The mean (SD) feto/ maternal ratio of antipyrine rose to 0.64 0.17 by 180 min, indicating satisfactory perfusion. No FITC-dextran was detected in the fetal circuit and this together with absence of significant volume change in the fetal circuit excludes significant leakage from the maternal to fetal circuits. 2.2. Transthyretin ELISA assay TTR was measured in perfusate with an ‘in-house’ enzyme-linked immunosorbent assay (ELISA). Anti-rabbit polyclonal TTR antibody (“capture antibody”) was obtained from Dako (Dako Australia, Campbellfield, VIC) and anti-sheep polyclonal TTR antibody conjugated with horseradish peroxidase (HRP) from Abcam (Cambridge, England). Maxisorp plates (Nunc, In Vitro Technologies, Sydney, NSW) were coated with capture antibody, 1 mg/well and incubated at 4 C overnight. The following day plates were washed (0.05% Tween 20 in PBS, pH 7.4), coated with blocking buffer (3% bovine serum albumin in PBS/0.05% Tween 20/0.5% Azide, pH 7.4) for 1 h and again washed. TTR (Sigma) standards ranging from 5.5 to 545.5 pmol/ L (300e3000 pg/mL) and perfusate samples were added to appropriate wells and incubated overnight at 4 C on a rocking platform. Plates were washed and incubated with HRP conjugated anti-TTR antibody (0.12 ug/well) for 1 h at room temperature. Following washing, color was developed over 30 min with 3,30 ,5,50
2.4. Statistics and mathematical modelling Pearson correlation coefficients were calculated by standard methods. p values of <0.05 were regarded as significant. Data are expressed as mean SE unless indicated otherwise. Mean TTR concentration data fitted best to a three compartment model (placenta and maternal and fetal circuits) by solving differential equations (dPconc/ dt ¼ change in placental compartment with time, dMconc/dt ¼ change in maternal compartment with time and dFconc/dt ¼ changes in fetal compartment with time), using the BMDP program AR [13]. TTR concentration data were fitted to the model (placental lobule and maternal and fetal circuits) by solving differential equations dPconc/dt, dMconc/dt and dFconc/dt. The program finds the estimated parametric values in non-linear functions which minimize the weighted sum of squares of residuals between the observed values and the estimated non-linear functions. In the equations below kS represents synthesis of TTR in the lobule during the time of perfusion. The elimination rate constants k1 and k2 represent transfer from the placental to the maternal and fetal compartments respectively and the elimination rate constant k3 represents transfer from the maternal to the placental compartment. Mconc and Fconc refer to initial concentrations of drug in the maternal and fetal circuits respectively and Pconc represents the (unknown) concentration within the placental lobule. Instantaneous net flux is the product of elimination rate constants and compartmental concentration of TTR. dPconc =dt ¼ kS ðk1 þ k2Þ*Pconc þ k3*Mconc
(1)
dMconc =dt ¼ k1*Pconc k3*Mconc
(2)
dFconc =dt ¼ k2*Pconc
(3)
Quality of fit of the two compartment model was measured by the pseudo r2 (1.0 e ratio of the weighted residual sum of squares to (N-1) times the weighted variance). 2.5. Alexa Fluor labelling of TTR and placental uptake studies Term placentas (38e40 weeks gestation) were obtained from healthy nonsmoking women undergoing repeat elective caesarean section in the Royal Brisbane and Women’s Hospital. Early pregnancy placentas were obtained from terminations at 14e16 weeks gestation at a private clinic. The use of placentas was approved by the Royal Brisbane and Women’s Hospital and Queensland Institute of Medical Research Ethics Committees and fully informed consent was obtained from each donating woman. Placental villi were dissected into about 8 mm long branches from the end of the terminal villus. After three washes in cold saline solution, the placenta villi were incubated in serum free Dulbecco’s Modified Eagle Medium (DMEM, Sigma, Castle Hill, NSW, Australia) in a humidified atmosphere of 95% O2 and 5% CO2 for 1 h. Fluorescent labelling of TTR (Alexa-TTR) was performed with an Alexa Fluor-594 Labelling Kit according to manufacturer’s instructions (Invitrogen, Mulgrave, VIC, Australia). For TTR uptake, placenta villi were incubated in serum free DMEM uptake medium containing 1 mM of Alexa-TTR plus 10 mM T4 or Alexa 594 dye only for 1 h. Tissue was counterstained with 40 ,6-diamidino-2-phenylindole (DAPI) to visualise nuclei and zona occludens (ZO-1) monoclonal antibody conjugated to Alexa Fluor 488, 1:200, (Invitrogen, Mulgrave, VIC, Australia) to mark the fetal capillary
254
R.H. Mortimer et al. / Placenta 33 (2012) 252e256
endothelium. Villi were mounted on slides with ProLong Gold antifade reagent (Invitrogen, Mulgrave, VIC, Australia). Images were captured using Deltavision Core technology (Applied Precision, Washington DC, USA).
3. Results 3.1. Transthyretin concentrations in maternal and fetal circuits TTR was readily detected in maternal and fetal circuits of the dually perfused human placental lobule (Fig. 1) There was significant perfusion to perfusion variation in TTR levels in maternal and fetal circuits, leading to moderately large standard errors but in both circuits mean TTR levels rose progressively with time in a linear fashion (Pearson correlation r ¼ 0.950, p ¼ 0.001 for maternal circuit levels and r ¼ 0.967, p ¼ 0.0004 for fetal circuit levels). TTR levels were much higher in the maternal than in the fetal circuits, reaching 3920 726.4 pmol/L by 180 min in the maternal circuit and 90 36.1 pmol/L in the fetal circuit (n ¼ 5). Mean maternal/fetal circuit TTR ratio was 99.4 25.6 (n ¼ 7).
Fig. 2. Three compartment model (Maternal circuit, placental lobule and fetal circuit) used to fit maternal and fetal circuit TTR concentrations. kS represents continuous synthesis of TTR by the placental lobule. The elimination rate constants k1 and k2 represent flow from the placental lobule to the maternal and fetal compartments respectively. The elimination rate constant k3 represents flow from the maternal compartment to the placental lobule. Estimated rate constants and their asymptotic standard deviations and coefficients of variation are displayed in the associated table.
3.2. Mathematical modelling A two compartment (maternal and fetal) and several three compartment models (maternal, placental and fetal) were evaluated. The best fit, as judged by pseudo r2 values, was achieved by the three compartment model in Fig. 2 (pseudo r2 ¼ 1.000). Incorporation of a return pathway from the fetal to the placental compartment reduced the pseudo r2. Estimated values for the rate constants kS, k1, k2 and k3 are shown in Fig. 2. 3.3. Placental uptake of Alexa-labelled transthyretin As previously reported [9] trophoblasts in placental explants readily took up Alexa-labelled TTR. In the present study we increased incubation times and studied placental vessels for evidence of appearance of intravascular Alexa-labelled TTR. In Fig. 3A a sheaf of villus capillaries from 14 weeks gestation placenta is marked green with anti ZO-1 antibody, an excellent marker of placental endothelial tight junctions [14]. Alexa-labelled TTR (red) can also been seen distributed linearly with collocation of TTR and the endothelial marker in several areas. In a 15 weeks placental explant (Fig. 3B) the endothelium of a large villus stem capillary is marked green by ZO-1. Intravascular TTR (red) is easily seen adjacent to endothelium. Labelled intravascular TTR could also be detected within villus capillaries in a term placenta (Fig. 3C). Thus early and late placentas labelled TTR is transferred within 1 h from the maternal trophoblast surface to the villus capillaries, part of the
Fig. 1. Transthyretin (TTR) concentrations (mean SE pm/L) in maternal (black circles) and fetal (open squares) perfusate over time. Secretion was linear r ¼ 0.950, p ¼ 0.001 for maternal circuit levels and r ¼ 0.967, p ¼ 0.0004 for fetal circuit levels.
fetal circulation. We carefully examined vascular and stromal distribution of TTR in explants exposed to Alexa-labelled TTR and found it in distal villous vessels. To exclude the possibility that the Alexa label seen within the villi was free dye rather than conjugated to TTR we examined uptake of free Alexa Fluor-594 into early placental villi. No Alexa Fluor-594 dye uptake was detected (Fig. 4). 4. Discussion This study shows for the first time that placental TTR is secreted into the maternal and fetal circulations of the dually perfused placental lobule. Secretion is predominantly into the maternal circulation with a ratio of maternal to fetal concentrations of about 100:1 and over the duration of the perfusions there is no indication of a trend to equilibrium of TTR concentrations between the two circuits. TTR could conceptually reach the fetal circuit by direct secretion of trophoblastic TTR into fetal capillaries, reuptake of TTR secreted into the maternal circuit and subsequent secretion into the fetal circuit or a combination of the two. Studies in polarised JEG-3 cells referred to above suggest that the trophoblast may indeed have the capacity for both apical and basal secretion of TTR. To explore this further we fitted the data from the perfusion studies to a mathematical model. Interestingly the best fitting model combined direct secretion into maternal and fetal circuits and reuptake of TTR from the maternal circuit. The rate constant for reuptake is low but this may relate to absence of thyroxine in the perfusate. As indicated earlier [9] thyroxine enhanced tetramerization of TTR increases cell uptake significantly and this could be tested in the perfusion model. The model also required the assumption of continuous TTR synthesis by the perfused lobules. There are some limitations to this modelling approach. Because of lobule to lobule variations in TTR secretion some standard errors are large. The overall fit of the model, as judged by pseudo r2, is however excellent. Studies of transfer of labelled TTR from maternal to fetal circulations in the perfused placental lobule would either pose an unacceptable radiation risk or be prohibitively expensive. We therefore further explored the important question of whether maternally secreted TTR could be taken up by trophoblast and transferred to the fetal circulation by more closely examining uptake of Alexa-labelled TTR by human placental explants. It is clear from our experiments (Fig. 4) that the Alexa-labelled TTR, not
R.H. Mortimer et al. / Placenta 33 (2012) 252e256
255
Fig. 4. Immunofluorescent studies of Alexa Fluor-594 labelled TTR (red) and free Alexa Fluor-594 uptake by villus sections from early (16 week) placenta. Panel A shows significant Alexa Fluor-594 labelled TTR (red) uptake in trophoblasts. Panel B shows no Alexa Fluor-594 signal (red) was detected. Nuclei are labelled blue with DAPI. To ensure identical image analysis for the comparison of TTR uptake, all slides were scanned under same setting of exposure conditions including UV light transmission and exposure time. Captured images were displayed with same intensity scale for channel of Alexa Fluor-594 labelled TTR.
Fig. 3. Immunofluorescent studies of Alexa Fluor-594 labelled TTR (red) uptake by villus sections from early (14 weeks gestation, A and 15 weeks gestation, B) and term (C) placental explants marked with 40 ,6-diamidino-2-phenylindole (DAPI) to visualise nuclei (blue) and zona occludens (ZO-1) antibody to mark capillary endothelium tight junctions (green). White bars indicate 10 mm. In A, a sheaf of villus capillaries (green) is well delineated. TTR is also distributed linearly and TTR and ZO-1 are collocated in several areas. In B a villus stem capillary is shown in cross section with intravascular TTR adjacent to endothelium in B. In C a villus capillary in a term placenta demonstrates linear accumulation of TTR.
256
R.H. Mortimer et al. / Placenta 33 (2012) 252e256
free Alexa dye, is taken up, which may suggest that placental uptake of TTR is by a specific receptor/channel not by simple diffusion. As shown in Fig. 3 the fluorescently labelled TTR added to incubation medium is co-located with fetal capillaries in placental villi not only in near term placentas but also in early placental tissue (14e15 weeks). To exclude the unlikely possibility that TTR was diffusing into the cut ends of placental blood vessels we carefully examined vascular and stromal distribution of TTR in explants and could find TTR only in distal villous vessels. We speculate that a placental TTR secretion and uptake system may have significant physiological and pathological implications. Firstly high concentrations of TTR at the maternal bloodetrophoblast interface may deliver TH to the large array of membrane TH transporters for subsequent delivery to the fetus. Secondly a TTR shuttle system may provide a mechanism for materno-fetal translocation of TH and retinol/retinol binding protein, both of which bind to TTR. Thirdly transfer of TH from the maternal to fetal circulations is significantly modulated by placental deiodination [15] and the presence of a high affinity intracellular TH binding protein could reduce TH exposure to Type 3 deiodinase. Lastly many xenobiotics bind to TTR [16]. These can interfere with TH binding and could reduce placental TH transfer but may also be carried into the fetal circulation by TTR. Most of the TTR in the maternal circulation comes from maternal liver. The relative contributions of maternal liver derived TTR and trophoblast derived TTR in the maternal placental circulation are unknown but it is very probable that maternal TTR could also be involved in TTR mediated thyroid hormone transfer. In the first trimester the fetus is totally dependent on placental transfer of maternal thyroid hormone, particularly thyroxine which in rat and human fetal brain is deiodinated to the biologically more active triiodothyronine [17,18]. We have recently reported that the low oxygen levels that human trophoblasts are exposed to in the first trimester significantly increase both expression and uptake of TTR in isolated human trophoblasts [19] suggesting that TTR mediated thyroid hormone transfer may be important at this time.
References [1] Jacobsson B, Collins VP, Grimelius L, Pettersson T, Sandstedt B, Carlstrom A. Transthyretin immunoreactivity in human and porcine liver, choroid plexus, and pancreatic islets. Journal of Histochemistry and Cytochemistry 1989; 37(1):31e7.
[2] Defoe DM, Martone RL, Caldwell RB, Chang CW, Herbert J. Transthyretin secretion by cultured rat retinal-pigment epithelium. Investigative Ophthalmology and Visual Science 1992;33(4):911. [3] Schussler GC. The thyroxine-binding proteins. Thyroid 2000;10(2):141e9. [4] Refetoff S, Robin NI, Fang VS. Paramaters of thyroid function in serum of 16 selected vertebrate species e a study of PBI, serum T4, free T4 and pattern of T4 and T3 binding to serum proteins. Endocrinology 1970;86(4):793e805. [5] Dickson PW, Aldred AR, Menting JG, Marley PD, Sawyer WH, Schreiber G. Thyroxine transport in choroid plexus. The Journal of Biological Chemistry 1987;262(29):13907e15. [6] Herbert J, Cavallaro T, Martone R. The distribution of retinol-binding protein and its messenger RNA in the rat eye. Investigative Ophthalmology and Visual Science 1991;32(2):302e9. [7] Gitlin D, Perricelli A. Synthesis of serum albumin, prealbumin, alphafetoprotein, alpha-1-antitrypsin and transferrin by the human yolk sac. Nature 1970;228(5275):995e7. [8] McKinnon B, Li H, Richard K, Mortimer R. Synthesis of thyroid hormone binding proteins transthyretin and albumin by human trophoblast. The Journal of Clinical Endocrinology and Metabolism 2005;90(12):6714e20. [9] Landers KA, McKinnon BD, Li HK, Subramaniam VN, Mortimer RH, Richard K. Carrier-mediated thyroid hormone transport into placenta by placental transthyretin. Journal of Clinical Endocrinology and Metabolism 2009;94(7): 2610e6. [10] Balakrishnan B, Henare K, Thorstensen EB, Ponnampalam AP, Mitchell MD. Transfer of bisphenol A across the human placenta. American Journal of Obstetrics and Gynecology 2010;202(4):393.e1e7. [11] Cannell GR, Kluck RM, Hamilton SE, Mortimer RH, Hooper WD, Dickinson RG. Markers of physical integrity and metabolic viability of the perfused human placental lobule. Clinical and Experimental Pharmacology and Physiology 1988;15(11):837e44. [12] Balakrishnan B, Thorstensen EB, Ponnampalam AP, Mitchell MD. Transplacental transfer and biotransformation of genistein in human placenta. Placenta 2010;31(6):506e11. [13] Ralston JR ML, Sampson PF, Uno FK. Fitting pharmacokinetic models with program AR: new instructions for PCs and mainframes. Los Angeles, California: BMDP Statistical Software, Inc.; 1988 [pp. 1e49]. [14] Lievano S, Alarcon L, Chavez-Munguia B, Gonzalez-Mariscal L. Endothelia of term human placentae display diminished expression of tight junction proteins during preeclampsia. Cell and Tissue Research 2006;324(3):433e48. [15] Mortimer RH, Galligan JP, Cannell GR, Addison RS, Roberts MS. Maternal to fetal thyroxine transmission in the human term placenta is limited by inner ring deiodination. Journal of Clinical Endocrinology and Metabolism 1996; 81(6):2247e9. [16] Yamauchi K, Ishihara A, Fukazawa H, Terao Y. Competitive interactions of chlorinated phenol compounds with 3,30 ,5-triiodothyronine binding to transthyretin: detection of possible thyroid-disrupting chemicals in environmental waste water. Toxicology and Applied Pharmacology 2003;187(2): 110e7. [17] Morreale de Escobar G, Calvo R, Obregon MJ, Escobar del Rey F. Homeostasis of brain T3 in rat fetuses and their mothers: effects of thyroid status and iodine deficiency. Acta Medica Austriaca 1992;19(Suppl. 1):110e6. [18] Kester MH, Martinez de Mena R, Obregon MJ, Marinkovic D, Howatson A, Visser TJ, et al. Iodothyronine levels in the human developing brain: major regulatory roles of iodothyronine deiodinases in different areas. The Journal of Clinical Endocrinology and Metabolism 2004;89(7):3117e28. [19] Patel J, Landers K, Mortimer RH, Richard K. Expression and uptake of the thyroxine binding protein transthyretin is regulated by oxygen in primary trophoblast placental cells. The Journal of Endocrinology; 2011.
Contents lists available at SciVerse ScienceDirect
Placenta journal homepage: www.elsevier.com/locate/placenta
Secretion and transfer of the thyroid hormone binding protein transthyretin by human placenta R.H. Mortimer a, b, c, *, K.A. Landers a, B. Balakrishnan e, H. Li a, M.D. Mitchell d, e, J. Patel a, b, K. Richard a, b a
Conjoint Endocrine Laboratory, Royal Brisbane and Women’s Hospital, Herston 4029, Brisbane, Australia School of Medicine, The University of Queensland, Herston 4006, Brisbane, Australia c Disciplines of Medicine, Obstetrics and Gynaecology, The University of Queensland, Herston 4006, Brisbane, Australia d University of Queensland Centre for Clinical Research, Building 71/918, Royal Brisbane & Women’s Hospital Campus, Herston QLD 4029, Australia e Liggins Institute, The University of Auckland, Private Bag 92019, Victoria Street West, Auckland 1142, New Zealand b
a r t i c l e i n f o
a b s t r a c t
Article history: Accepted 3 January 2012
Context: The thyroid hormone and retinol binding protein transthyretin (TTR) is synthesised by human trophoblasts. Polarised JEG-3 choriocarcinoma cells grown in bicameral chambers secrete TTR predominantly apically but also basally and these cells and human trophoblasts also take up TTR suggesting that there may be a placental TTR shuttle that participates in materno-fetal transfer of thyroid hormones and retinol. Objectives and methods: Our objective was to investigate TTR secretion into the maternal and fetal circuits of the ex vivo dually perfused placental lobule to confirm that placenta secretes TTR into the fetal circulation. We also investigated translocation of Alexa Fluor-594 labelled TTR from incubation medium into the fetal placental capillaries in early (14e15 weeks) and term placental villus explants. Results: The perfused placental lobule secretes TTR into the maternal and fetal circuits. Secretion in both circuits is linear with time and is predominantly into the maternal circuit (mean maternal/fetal ratio 99.4 25.6). The mean data fitted well to a three compartment mathematical model (maternal circuit, placenta and fetal circuit, constant secretion of TTR and return of maternal circuit TTR to the placental compartment). Explants from early (14e15 weeks) and late (38e40 weeks) placentas translocated fluorescently labelled TTR from medium to villus (fetal) capillaries. Conclusions: Our results confirm that human placenta secretes TTR into maternal and fetal circulations and supports the hypothesis that placental TTR secreted into the maternal placental circulation can be taken up by trophoblasts and translocated to the fetal circulation, forming a TTR shuttle system. This may have important implications for materno-fetal transfer of thyroid hormones, retinol/retinol binding protein and xenobiotics (such as polychlorinated biphenyls) all of which bind to TTR. Crown Copyright Ó 2012 Published by Elsevier Ltd. All rights reserved.
Keywords: Placenta Transthyretin Secretion Transport
1. Introduction Transthyretin (TTR), a 56 KD homotetrameric protein that binds thyroid hormone (TH) and retinol binding protein, is synthesised and secreted by liver, choroid plexus, pancreas and retina [1,2]. Hepatic TTR, thyroxine binding globulin and albumin serve as important binding proteins for circulating TH [3], so that only about 0.03% of circulating T4 and 0.3% of circulating T3 are free [4]. The choroid plexus is a rich source of TTR which is secreted into the
* Corresponding author. Office of Health and Medical Research, Level 13, Queensland Health Building, 147-163 Charlotte St, Brisbane Q4001, Australia. Tel.: þ61 7 3234 1373; fax: þ61 7 3234 0107. E-mail addresses: [email protected], [email protected] (R.H. Mortimer).
cerebro-spinal fluid (CSF), perhaps contributing to TH delivery to brain [5]. The retina also makes TTR and this may be involved in TH transport within the eye [6]. Although early studies reported the presence of TTR in human yolk sac [7] the synthesis and secretion of TTR by human trophoblast was not reported until 2005 [8]. It was subsequently shown that human trophoblast and the human choriocarcinoma cell line JEG-3 internalise radiolabelled TTR [9]. Internalisation is enhanced by the presence of unlabelled T4 which induces tetramerization of TTR. T4 cross linked to TTR is also internalised. Using polarised JEG3 cells grown in bicameral chambers we demonstrated secretion of TTR predominantly via apical but also by basal cell membranes. These findings suggested that TTR might participate in a shuttle system in which trophoblast TTR is secreted into the maternal placental circulation and internalised back into the trophoblast.
0143-4004/$ e see front matter Crown Copyright Ó 2012 Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.placenta.2012.01.006
R.H. Mortimer et al. / Placenta 33 (2012) 252e256
Such a system would ensure high concentrations of TTR at the maternal surface of chorionic villi and, as TTR has greater affinity for T4 than T3, could provide a mechanism for preferential transfer of maternal T4 into trophoblast and, possibly, then into the fetal circulation. To further explore the hypotheses that the human placenta secretes TTR into maternal and fetal circuits and that TTR taken up into trophoblast via the apical membrane can be translocated into the fetal placental circulation we examined secretion of TTR into maternal and fetal circulations using dually perfused isolated human placental lobules. In vitro uptake of Alexa-labelled TTR by human term placental explants was also examined. The results indicate that while TTR secretion by human placenta is predominantly into the maternal circuit significant amounts appear in the fetal circuit. Incubation of early (14e15 weeks) and late (36e38 weeks) placental explants with Alexa-labelled TTR demonstrate translocation of TTR from incubation medium to villous capillaries, part of the fetal circulation. 2. Methods
253
tetramethyl-benzidine liquid substrate (Sigma) and absorbance read at 370 nm with a VersaMaxÔ Microplate reader. 2.3. ELISA validation 2.3.1. Standard curves and sensitivity The working ranges for the calibration curves were 5.5e545.5 pmol/L (300e3000 pg/mL). The curves generated were linear with a correlation coefficient of r 0.975. The lowest TTR concentration measurable outside the range of two standard deviations was determined as 35.5 pmol/L (CV 9%). 2.3.2. Precision data Assay variation was determined by its intra- and inter-assay reproducibility. Intra-assay reproducibility was determined using TTR concentrations in a range of 18.12e181.2 pmol/L (1000e10,000 pg/mL). Intra-assay reproducibility (coefficients of variation) was 2.28%. A total of nine replicates were measured on the same microplate on the same day using the ELISA method. To calculate inter-assay reproducibility, the ELISA assay was run on five different days in nonuplicate using the TTR concentrations mentioned above. Inter-assay coefficients of variation were between 8 and 21%. 2.3.3. Accuracy TTR recovery was determined in duplicate and triplicate on two different days (n ¼ 5) with 7 different concentrations of TTR (5.5e545.5 pmol/L). The recovery of added TTR ranged from 59 to 140% with coefficients of variation from 8 to 22%.
2.1. Placental perfusion Seven term placentas were obtained with informed consent from healthy nonsmoking women undergoing elective caesarean section in Auckland City Hospital (Auckland, New Zealand). This use was approved by the regional ethics committee. Ex vivo dual perfusion (maternal and fetal circuits) was done by a well-validated method [10e12]. In brief, perfusions were established within 30 min of delivery by catheterisation of a fetal (chorionic) vessel and perfusate circulated through the fetal side using a digitally controlled pump. The perfused cotyledon was then excised from the placenta and mounted fetal side up in the perfusion chamber. The maternal circuit was then accessed by cannulas inserted through the maternal placental surface. Initially perfusates in both compartments were recirculated for 1 h then replaced by fresh perfusate. The perfusate consisted of phenol red-free medium-199 (M-3769; SigmaeAldrich, St Louis, MO) cell culture media supplemented with 25 g/L polyvinylpyrrolidone PVP-40 (SigmaeAldrich), 1 g/L bovine serum albuminefraction V (SigmaeAldrich), 2 g/L glucose (SigmaeAldrich), 20,000 IU/L heparin (Multiparin, CP Pharmaceuticals Ltd, Wexham, UK), and 48 mg/L gentamicin reagent solution (GIBCO; Invitrogen, Auckland, New Zealand). Maternal perfusate was gassed with 95% O2 and 5% CO2 while fetal perfusate was gassed with (5% N2 and 5% CO2). After equilibration, perfusate was replaced with fresh perfusate containing antipyrine (SigmaeAldrich, St Louis, MO) at 40 mg/mL and fluorescein isothiocyanate (FITC)-dextran, (SigmaeAldrich, St Louis, MO) 12.5 mg/mL. The perfusions were then carried on for up to 3 h in recirculating mode. 7.5 mL of perfusate was collected at each time interval from both circuits. Antipyrine was measured by HPLC and FITCdextran was measured spectrofluorometrically [10]. Perfusions were continued for up to 180 min. Viability and metabolic activity of perfused cotyledons was assessed by measuring glucose utilization, lactate production (Hitachi 902 autoanalyzer; Hitachi High Technologies Corp, Tokyo, Japan), and human chorionic gonadotropin (hCG) secretion (enzyme-linked immunosorbent assay). Stringent quality control measures were observed, namely a fetal flow rate of 4 mL/min; a maternal flow rate of 10 mL/min; a fetal pressure 30e60 mmHg; a pH 7.2e7.4; and a feto-maternal fluid shift of less than 3 mL/h. The mean (SD) feto/ maternal ratio of antipyrine rose to 0.64 0.17 by 180 min, indicating satisfactory perfusion. No FITC-dextran was detected in the fetal circuit and this together with absence of significant volume change in the fetal circuit excludes significant leakage from the maternal to fetal circuits. 2.2. Transthyretin ELISA assay TTR was measured in perfusate with an ‘in-house’ enzyme-linked immunosorbent assay (ELISA). Anti-rabbit polyclonal TTR antibody (“capture antibody”) was obtained from Dako (Dako Australia, Campbellfield, VIC) and anti-sheep polyclonal TTR antibody conjugated with horseradish peroxidase (HRP) from Abcam (Cambridge, England). Maxisorp plates (Nunc, In Vitro Technologies, Sydney, NSW) were coated with capture antibody, 1 mg/well and incubated at 4 C overnight. The following day plates were washed (0.05% Tween 20 in PBS, pH 7.4), coated with blocking buffer (3% bovine serum albumin in PBS/0.05% Tween 20/0.5% Azide, pH 7.4) for 1 h and again washed. TTR (Sigma) standards ranging from 5.5 to 545.5 pmol/ L (300e3000 pg/mL) and perfusate samples were added to appropriate wells and incubated overnight at 4 C on a rocking platform. Plates were washed and incubated with HRP conjugated anti-TTR antibody (0.12 ug/well) for 1 h at room temperature. Following washing, color was developed over 30 min with 3,30 ,5,50
2.4. Statistics and mathematical modelling Pearson correlation coefficients were calculated by standard methods. p values of <0.05 were regarded as significant. Data are expressed as mean SE unless indicated otherwise. Mean TTR concentration data fitted best to a three compartment model (placenta and maternal and fetal circuits) by solving differential equations (dPconc/ dt ¼ change in placental compartment with time, dMconc/dt ¼ change in maternal compartment with time and dFconc/dt ¼ changes in fetal compartment with time), using the BMDP program AR [13]. TTR concentration data were fitted to the model (placental lobule and maternal and fetal circuits) by solving differential equations dPconc/dt, dMconc/dt and dFconc/dt. The program finds the estimated parametric values in non-linear functions which minimize the weighted sum of squares of residuals between the observed values and the estimated non-linear functions. In the equations below kS represents synthesis of TTR in the lobule during the time of perfusion. The elimination rate constants k1 and k2 represent transfer from the placental to the maternal and fetal compartments respectively and the elimination rate constant k3 represents transfer from the maternal to the placental compartment. Mconc and Fconc refer to initial concentrations of drug in the maternal and fetal circuits respectively and Pconc represents the (unknown) concentration within the placental lobule. Instantaneous net flux is the product of elimination rate constants and compartmental concentration of TTR. dPconc =dt ¼ kS ðk1 þ k2Þ*Pconc þ k3*Mconc
(1)
dMconc =dt ¼ k1*Pconc k3*Mconc
(2)
dFconc =dt ¼ k2*Pconc
(3)
Quality of fit of the two compartment model was measured by the pseudo r2 (1.0 e ratio of the weighted residual sum of squares to (N-1) times the weighted variance). 2.5. Alexa Fluor labelling of TTR and placental uptake studies Term placentas (38e40 weeks gestation) were obtained from healthy nonsmoking women undergoing repeat elective caesarean section in the Royal Brisbane and Women’s Hospital. Early pregnancy placentas were obtained from terminations at 14e16 weeks gestation at a private clinic. The use of placentas was approved by the Royal Brisbane and Women’s Hospital and Queensland Institute of Medical Research Ethics Committees and fully informed consent was obtained from each donating woman. Placental villi were dissected into about 8 mm long branches from the end of the terminal villus. After three washes in cold saline solution, the placenta villi were incubated in serum free Dulbecco’s Modified Eagle Medium (DMEM, Sigma, Castle Hill, NSW, Australia) in a humidified atmosphere of 95% O2 and 5% CO2 for 1 h. Fluorescent labelling of TTR (Alexa-TTR) was performed with an Alexa Fluor-594 Labelling Kit according to manufacturer’s instructions (Invitrogen, Mulgrave, VIC, Australia). For TTR uptake, placenta villi were incubated in serum free DMEM uptake medium containing 1 mM of Alexa-TTR plus 10 mM T4 or Alexa 594 dye only for 1 h. Tissue was counterstained with 40 ,6-diamidino-2-phenylindole (DAPI) to visualise nuclei and zona occludens (ZO-1) monoclonal antibody conjugated to Alexa Fluor 488, 1:200, (Invitrogen, Mulgrave, VIC, Australia) to mark the fetal capillary
254
R.H. Mortimer et al. / Placenta 33 (2012) 252e256
endothelium. Villi were mounted on slides with ProLong Gold antifade reagent (Invitrogen, Mulgrave, VIC, Australia). Images were captured using Deltavision Core technology (Applied Precision, Washington DC, USA).
3. Results 3.1. Transthyretin concentrations in maternal and fetal circuits TTR was readily detected in maternal and fetal circuits of the dually perfused human placental lobule (Fig. 1) There was significant perfusion to perfusion variation in TTR levels in maternal and fetal circuits, leading to moderately large standard errors but in both circuits mean TTR levels rose progressively with time in a linear fashion (Pearson correlation r ¼ 0.950, p ¼ 0.001 for maternal circuit levels and r ¼ 0.967, p ¼ 0.0004 for fetal circuit levels). TTR levels were much higher in the maternal than in the fetal circuits, reaching 3920 726.4 pmol/L by 180 min in the maternal circuit and 90 36.1 pmol/L in the fetal circuit (n ¼ 5). Mean maternal/fetal circuit TTR ratio was 99.4 25.6 (n ¼ 7).
Fig. 2. Three compartment model (Maternal circuit, placental lobule and fetal circuit) used to fit maternal and fetal circuit TTR concentrations. kS represents continuous synthesis of TTR by the placental lobule. The elimination rate constants k1 and k2 represent flow from the placental lobule to the maternal and fetal compartments respectively. The elimination rate constant k3 represents flow from the maternal compartment to the placental lobule. Estimated rate constants and their asymptotic standard deviations and coefficients of variation are displayed in the associated table.
3.2. Mathematical modelling A two compartment (maternal and fetal) and several three compartment models (maternal, placental and fetal) were evaluated. The best fit, as judged by pseudo r2 values, was achieved by the three compartment model in Fig. 2 (pseudo r2 ¼ 1.000). Incorporation of a return pathway from the fetal to the placental compartment reduced the pseudo r2. Estimated values for the rate constants kS, k1, k2 and k3 are shown in Fig. 2. 3.3. Placental uptake of Alexa-labelled transthyretin As previously reported [9] trophoblasts in placental explants readily took up Alexa-labelled TTR. In the present study we increased incubation times and studied placental vessels for evidence of appearance of intravascular Alexa-labelled TTR. In Fig. 3A a sheaf of villus capillaries from 14 weeks gestation placenta is marked green with anti ZO-1 antibody, an excellent marker of placental endothelial tight junctions [14]. Alexa-labelled TTR (red) can also been seen distributed linearly with collocation of TTR and the endothelial marker in several areas. In a 15 weeks placental explant (Fig. 3B) the endothelium of a large villus stem capillary is marked green by ZO-1. Intravascular TTR (red) is easily seen adjacent to endothelium. Labelled intravascular TTR could also be detected within villus capillaries in a term placenta (Fig. 3C). Thus early and late placentas labelled TTR is transferred within 1 h from the maternal trophoblast surface to the villus capillaries, part of the
Fig. 1. Transthyretin (TTR) concentrations (mean SE pm/L) in maternal (black circles) and fetal (open squares) perfusate over time. Secretion was linear r ¼ 0.950, p ¼ 0.001 for maternal circuit levels and r ¼ 0.967, p ¼ 0.0004 for fetal circuit levels.
fetal circulation. We carefully examined vascular and stromal distribution of TTR in explants exposed to Alexa-labelled TTR and found it in distal villous vessels. To exclude the possibility that the Alexa label seen within the villi was free dye rather than conjugated to TTR we examined uptake of free Alexa Fluor-594 into early placental villi. No Alexa Fluor-594 dye uptake was detected (Fig. 4). 4. Discussion This study shows for the first time that placental TTR is secreted into the maternal and fetal circulations of the dually perfused placental lobule. Secretion is predominantly into the maternal circulation with a ratio of maternal to fetal concentrations of about 100:1 and over the duration of the perfusions there is no indication of a trend to equilibrium of TTR concentrations between the two circuits. TTR could conceptually reach the fetal circuit by direct secretion of trophoblastic TTR into fetal capillaries, reuptake of TTR secreted into the maternal circuit and subsequent secretion into the fetal circuit or a combination of the two. Studies in polarised JEG-3 cells referred to above suggest that the trophoblast may indeed have the capacity for both apical and basal secretion of TTR. To explore this further we fitted the data from the perfusion studies to a mathematical model. Interestingly the best fitting model combined direct secretion into maternal and fetal circuits and reuptake of TTR from the maternal circuit. The rate constant for reuptake is low but this may relate to absence of thyroxine in the perfusate. As indicated earlier [9] thyroxine enhanced tetramerization of TTR increases cell uptake significantly and this could be tested in the perfusion model. The model also required the assumption of continuous TTR synthesis by the perfused lobules. There are some limitations to this modelling approach. Because of lobule to lobule variations in TTR secretion some standard errors are large. The overall fit of the model, as judged by pseudo r2, is however excellent. Studies of transfer of labelled TTR from maternal to fetal circulations in the perfused placental lobule would either pose an unacceptable radiation risk or be prohibitively expensive. We therefore further explored the important question of whether maternally secreted TTR could be taken up by trophoblast and transferred to the fetal circulation by more closely examining uptake of Alexa-labelled TTR by human placental explants. It is clear from our experiments (Fig. 4) that the Alexa-labelled TTR, not
R.H. Mortimer et al. / Placenta 33 (2012) 252e256
255
Fig. 4. Immunofluorescent studies of Alexa Fluor-594 labelled TTR (red) and free Alexa Fluor-594 uptake by villus sections from early (16 week) placenta. Panel A shows significant Alexa Fluor-594 labelled TTR (red) uptake in trophoblasts. Panel B shows no Alexa Fluor-594 signal (red) was detected. Nuclei are labelled blue with DAPI. To ensure identical image analysis for the comparison of TTR uptake, all slides were scanned under same setting of exposure conditions including UV light transmission and exposure time. Captured images were displayed with same intensity scale for channel of Alexa Fluor-594 labelled TTR.
Fig. 3. Immunofluorescent studies of Alexa Fluor-594 labelled TTR (red) uptake by villus sections from early (14 weeks gestation, A and 15 weeks gestation, B) and term (C) placental explants marked with 40 ,6-diamidino-2-phenylindole (DAPI) to visualise nuclei (blue) and zona occludens (ZO-1) antibody to mark capillary endothelium tight junctions (green). White bars indicate 10 mm. In A, a sheaf of villus capillaries (green) is well delineated. TTR is also distributed linearly and TTR and ZO-1 are collocated in several areas. In B a villus stem capillary is shown in cross section with intravascular TTR adjacent to endothelium in B. In C a villus capillary in a term placenta demonstrates linear accumulation of TTR.
256
R.H. Mortimer et al. / Placenta 33 (2012) 252e256
free Alexa dye, is taken up, which may suggest that placental uptake of TTR is by a specific receptor/channel not by simple diffusion. As shown in Fig. 3 the fluorescently labelled TTR added to incubation medium is co-located with fetal capillaries in placental villi not only in near term placentas but also in early placental tissue (14e15 weeks). To exclude the unlikely possibility that TTR was diffusing into the cut ends of placental blood vessels we carefully examined vascular and stromal distribution of TTR in explants and could find TTR only in distal villous vessels. We speculate that a placental TTR secretion and uptake system may have significant physiological and pathological implications. Firstly high concentrations of TTR at the maternal bloodetrophoblast interface may deliver TH to the large array of membrane TH transporters for subsequent delivery to the fetus. Secondly a TTR shuttle system may provide a mechanism for materno-fetal translocation of TH and retinol/retinol binding protein, both of which bind to TTR. Thirdly transfer of TH from the maternal to fetal circulations is significantly modulated by placental deiodination [15] and the presence of a high affinity intracellular TH binding protein could reduce TH exposure to Type 3 deiodinase. Lastly many xenobiotics bind to TTR [16]. These can interfere with TH binding and could reduce placental TH transfer but may also be carried into the fetal circulation by TTR. Most of the TTR in the maternal circulation comes from maternal liver. The relative contributions of maternal liver derived TTR and trophoblast derived TTR in the maternal placental circulation are unknown but it is very probable that maternal TTR could also be involved in TTR mediated thyroid hormone transfer. In the first trimester the fetus is totally dependent on placental transfer of maternal thyroid hormone, particularly thyroxine which in rat and human fetal brain is deiodinated to the biologically more active triiodothyronine [17,18]. We have recently reported that the low oxygen levels that human trophoblasts are exposed to in the first trimester significantly increase both expression and uptake of TTR in isolated human trophoblasts [19] suggesting that TTR mediated thyroid hormone transfer may be important at this time.
References [1] Jacobsson B, Collins VP, Grimelius L, Pettersson T, Sandstedt B, Carlstrom A. Transthyretin immunoreactivity in human and porcine liver, choroid plexus, and pancreatic islets. Journal of Histochemistry and Cytochemistry 1989; 37(1):31e7.
[2] Defoe DM, Martone RL, Caldwell RB, Chang CW, Herbert J. Transthyretin secretion by cultured rat retinal-pigment epithelium. Investigative Ophthalmology and Visual Science 1992;33(4):911. [3] Schussler GC. The thyroxine-binding proteins. Thyroid 2000;10(2):141e9. [4] Refetoff S, Robin NI, Fang VS. Paramaters of thyroid function in serum of 16 selected vertebrate species e a study of PBI, serum T4, free T4 and pattern of T4 and T3 binding to serum proteins. Endocrinology 1970;86(4):793e805. [5] Dickson PW, Aldred AR, Menting JG, Marley PD, Sawyer WH, Schreiber G. Thyroxine transport in choroid plexus. The Journal of Biological Chemistry 1987;262(29):13907e15. [6] Herbert J, Cavallaro T, Martone R. The distribution of retinol-binding protein and its messenger RNA in the rat eye. Investigative Ophthalmology and Visual Science 1991;32(2):302e9. [7] Gitlin D, Perricelli A. Synthesis of serum albumin, prealbumin, alphafetoprotein, alpha-1-antitrypsin and transferrin by the human yolk sac. Nature 1970;228(5275):995e7. [8] McKinnon B, Li H, Richard K, Mortimer R. Synthesis of thyroid hormone binding proteins transthyretin and albumin by human trophoblast. The Journal of Clinical Endocrinology and Metabolism 2005;90(12):6714e20. [9] Landers KA, McKinnon BD, Li HK, Subramaniam VN, Mortimer RH, Richard K. Carrier-mediated thyroid hormone transport into placenta by placental transthyretin. Journal of Clinical Endocrinology and Metabolism 2009;94(7): 2610e6. [10] Balakrishnan B, Henare K, Thorstensen EB, Ponnampalam AP, Mitchell MD. Transfer of bisphenol A across the human placenta. American Journal of Obstetrics and Gynecology 2010;202(4):393.e1e7. [11] Cannell GR, Kluck RM, Hamilton SE, Mortimer RH, Hooper WD, Dickinson RG. Markers of physical integrity and metabolic viability of the perfused human placental lobule. Clinical and Experimental Pharmacology and Physiology 1988;15(11):837e44. [12] Balakrishnan B, Thorstensen EB, Ponnampalam AP, Mitchell MD. Transplacental transfer and biotransformation of genistein in human placenta. Placenta 2010;31(6):506e11. [13] Ralston JR ML, Sampson PF, Uno FK. Fitting pharmacokinetic models with program AR: new instructions for PCs and mainframes. Los Angeles, California: BMDP Statistical Software, Inc.; 1988 [pp. 1e49]. [14] Lievano S, Alarcon L, Chavez-Munguia B, Gonzalez-Mariscal L. Endothelia of term human placentae display diminished expression of tight junction proteins during preeclampsia. Cell and Tissue Research 2006;324(3):433e48. [15] Mortimer RH, Galligan JP, Cannell GR, Addison RS, Roberts MS. Maternal to fetal thyroxine transmission in the human term placenta is limited by inner ring deiodination. Journal of Clinical Endocrinology and Metabolism 1996; 81(6):2247e9. [16] Yamauchi K, Ishihara A, Fukazawa H, Terao Y. Competitive interactions of chlorinated phenol compounds with 3,30 ,5-triiodothyronine binding to transthyretin: detection of possible thyroid-disrupting chemicals in environmental waste water. Toxicology and Applied Pharmacology 2003;187(2): 110e7. [17] Morreale de Escobar G, Calvo R, Obregon MJ, Escobar del Rey F. Homeostasis of brain T3 in rat fetuses and their mothers: effects of thyroid status and iodine deficiency. Acta Medica Austriaca 1992;19(Suppl. 1):110e6. [18] Kester MH, Martinez de Mena R, Obregon MJ, Marinkovic D, Howatson A, Visser TJ, et al. Iodothyronine levels in the human developing brain: major regulatory roles of iodothyronine deiodinases in different areas. The Journal of Clinical Endocrinology and Metabolism 2004;89(7):3117e28. [19] Patel J, Landers K, Mortimer RH, Richard K. Expression and uptake of the thyroxine binding protein transthyretin is regulated by oxygen in primary trophoblast placental cells. The Journal of Endocrinology; 2011.