Cell-specific increases in metallothionein expression in the human placenta perfused with cadmium

Reproductive Toxicology,Vol. 8, No. 4, pp. 297-306, 1994 Copyright © 1994Elsevier ScienceLtd Printed in the USA. All rights reserved 0890-6238/94$6.00 + .00

Pergamon 0890.6238(94)E0015-N

• Original Contributions

CELL-SPECIFIC INCREASES IN METALLOTHIONEIN EXPRESSION IN THE H U M A N PLACENTA PERFUSED WITH CADMIUM JOHN G. BREEN,*'~ CAROL E I S E N M A N N , * t STUART HOROWITZ,:~ a n d

RICHARD K . MILLER*t§ Departments of *Obstetrics and Gynecology, tEnvironmental Medicine, ~:Pediatrics, §Environmental Health Sciences Center, University of Rochester Medical Center, Rochester, NY, USA Abstract - - Metallothionein (MT) is a cysteine-rich protein that may have both a nutritional and a protective role by binding either physiologic or toxic concentrations of metals in cells. The objective of this investigation was to measure the inducibility of MT mRNA and protein, and to determine their specific cellular localization following exposure to 20/zM cadmium (Cd) in the perfused human placenta for periods up to 8 h. MT mRNA was quantitated using slot blot hybridization with an MT llagenomic clone and MT transcripts were localized via in situ hybridization. MT protein was measured using a l°9Cd-binding assay and localized by immunocytochemistry using a monoclonal antibody to MT in fresh placental tissue from nonsmoking mothers and in tissue perfused for 4 or 8 h with CdClz (20 pM). Perfusions were performed on placentae from which fresh samples were also obtained. In fresh term placentae, MT mRNA was barely detectable using both slot blot and in situ hybridization. Slot blot hybridization for MT message demonstrated a dramatic increase of at least 70-fold above fresh after 8 h of perfusion with 20/zM CdCiz. In this time hmne, however, statistically significant increases in MT protein were not detected. Following perfusion for 4 and 8 h with 20/zM CdCIz, accumulation of MT transcripts was shown, in situ, to occur in stromal and endothelial cells with a small but detectable increase in trophoblast cells. These results were consistent with the localization of the MT protein after perfusion with Cd and also correlated with areas of stromal edema. Thus, while the trophoblast cells are in direct contact with Cd in maternal perfusate, the cells of the villous core and fetal endothelial cells are most responsive to Cd exposure with regard to MT expression. Lower levels of MT expression in trophoblast cells may leave this cell type more susceptible to Cd toxicity, while higher levels of MT expression in cells of the villous core and endothelial cells may provide these cells with more protection. Key Words: metallothionein; human placenta; cadmium; placental perfusion; in situ hybridization; immunocytochemistry;

slot blot hybridization; trophoblast cells.

INTRODUCTION

like Cd and other stressors. Parizek (3) first demonstrated in 1964 that the rat placenta is a site for Cd toxicity. Further investigations revealed that the fetus was not directly affected, but rather, the placenta was directly intoxicated (4-6). The rat placenta accumulates Cd at up to 50 times the concentration in the maternal circulation (6,7). During dual human term placental perfusion with CdCI2, the concentration of Cd in the maternal perfusate decreases due to placental accumulation and only a small proportion of the Cd crosses into the fetal circulation even after 6 h of perfusion (8). In the placental cytosolic fractions, Cd is bound to both a high and a low (<10,000 MW) molecular weight protein fraction (8).

Throughout gestation, the human placenta modulates a number of metabolic functions including transport (1,2). Cellular defense mechanisms and their relationship to placental toxicity, however, have been poorly understood, especially the role of metal binding proteins, for example, metallothionein, in the amelioration of the toxicity of metals Dr. Horowitz is currently at the CardioPulmonary Research Institute, Winthrop University Hospital, Mineola, New York. Address correspondence to Dr. R. K. Miller, Department of Obstetrics/Gynecology, University of Rochester Medical Center, 601 Elmwood Avenue, Rochester, NY 14642, USA. Received 25 January 1994; Revision received 21 March 1994; Accepted 24 March 1994. 297

298

Reproductive Toxicology

Following perfusion with CdCI2, the human term placenta demonstrates dose-dependent increases in toxicity at similar peak maternal circulation levels of Cd as was noted in the rat (10 to 100 /zM) (9). Oxygen consumption and the production of glucose and lactate are not affected, but after 4 h of perfusion with 10/zM CdC%, there are decreases in the tissue concentration and rate of secretion of human chorionic gonadotropin (hCG) as well as a decrease in the transfer of Zn into the fetal compartment (9). Fetal volume loss is apparent during perfusions with 20 and 100 /zM CdCI2 for 5 to 8 h. Placental histologic changes include edema, vacuoles in Hofbauer cells, mitochondrial swelling in syncytiotrophoblast cells, and finally trophoblastic necrosis (9). Metallothionein (MT), originally reported in 1957 by Margoshes and Vallee (10), is a small molecular weight, metal-binding protein that is reported to protect cells by binding toxic concentrations of a variety of metals (Cd, Zn, Cu) (1 l). The presence of MT protein in the human placenta was first described in 1984 by Waalkes and associates (12). MT protein has been detected in syncytiotrophoblast cells and in cells of the villous core in human term placentae using polyclonal antibodies (13). Further studies in the perfused human placenta extended the perfusion studies of Wier and Miller (8) by reporting the presence of MT protein in human placentae perfused for periods of 5 h following exposure to Cd (14). MT in the human placenta may have important implications for placental cellular defense as well as for both placental and fetal nutrition. The objective of these investigations was to determine the inducibility and the specific cellular localization of MT mRNA and protein following exposure to Cd (20 /zM) in the perfused human placenta for periods up to8h.

MATERIALS A N D METHODS

Tissue Human placentae were collected immediately after delivery from nonsmoking mothers at term. Subjects with complications during pregnancy were excluded from the study (for example, hypertension, diabetes mellitus). Placental weight and appearance of the organ was noted. Fresh tissue was immediately obtained from the placenta and either flash frozen in liquid nitrogen for later MT mRNA and protein analysis or fixed overnight in 10% buffered formalin (Sigma, St. Louis, MO) and embed-

Volume 8, Number 4, 1994

ded in paraffin for histology and in situ hybridization. Frozen tissue was stored at - 7 0 °C until used.

Placental perfusion Dual perfusion of the isolated human placental lobule utilized the methods described by Miller et al. (15,16) as modified from the original methods of Panigel (17) and Schneider et al. (18). Lactate, glucose, and oxygen consumption as well as net fetal oxygen transfer were used to confirm the metabolic activity of the perfused tissue. Flow rates were 3.0 mL/min on the fetal side and 15.0 mL/min on the maternal side. Fetal capillary integrity was confirmed by measuring fetal reservoir volume during the perfusion. Fetal arterial pressure was measured using a Grafco standard aneroid sphygmomanometer. Fetal perfusate was composed of M199 tissue culture media (Gibco, Grand Island, NY) with added heparin (15 IU/mL), glucose (2.0 g/L), dextran 40 (fetal--30 g/L, maternal--7.5 g/L, [Sigma, St. Louis, MO]), sulfamethoxazole (80 mg/L), trimethoprim (16 mg/L) and gentamycin (50 mg/L), but without phenol red. The fetal perfusate was gassed with 95% N2/5% CO2, while the maternal circulation was gassed with 95% O J 5 % CO2. During the first 2-h control period, perfusate samples were collected from the fetal reservoir and from the maternal arterial port every half hour for analysis of pO2, pCO2, pH, lactate, glucose, and hCG. Blood gasses and pH were measured using an Instrumentation Laboratories Model 1302 pH/blood gas analyzer. Glucose and lactate were measured using the YSI Lactate/Glucose analyzer, hCG was measured using the Tandem-hCG Antibody Set (Hybritech). At 2-and 4-h intervals, perfusates in both circuits were exchanged for fresh perfusates. In Cd experiments, after the initial 2-h control period, the fresh perfusate contained 20/zM CdCl2. This concentration was chosen based upon earlier published work that demonstrated the acute placental toxicity of 20 /~M CdCI2 in the perfusion model (9). After perfusion for 4 h with Cd, the perfusate was exchanged for fresh perfusate without Cd and the perfusion continued for an additional 4 h. After the experimental period was completed, the perfused tissue was collected in the same manner as described above for fresh tissue.

MT probe labeling MT IIa genomic probe (19) was generously provided by Dr. Edward Dudak in the PBR-322 plasmid and subcloned into PBS-SKII. Both sense and antisense probes were synthesized from the MT IIa

Metallothionein in the h u m a n placenta • J. G. BREEN ET AL.

genomic template in PBS-SKII vector using RNA polymerase T7 or T3 according to the suppliers instructions (Promega). These were used as digoxigenin-UTP labeled riboprobes for northern blot, slot blot, and in situ hybridization. MTIIa is a genomic clone and is expected to cross-hybridize to all MT isoform transcripts. The DNA probe for glyceraldehyde-6-phosphate dehydrogenase (GAPDH) was labeled with digoxigenin using the Genius DNA labeling kit (Boehringer-Mannheim, Indianapolis, IN).

R N A isolation and analysis Total cellular RNA was isolated from human placentae using a variation of the Chomcynski method (20). Water, in all solutions, was treated with diethyl pyrocarbonate (DEPC; Sigma) to decrease the likelihood of RNAse contamination. Frozen tissue was ground in a Waring blender and further homogenized in a 4-M guanidine thiocyanate/0.1 M /3-mercaptoethanol/25 mM sodium citrate, pH 7/0.5% n-laurylsarcosine denaturing solution using a Dounce homogenizer. The homogenate was treated with 0.1 vol sodium acetate, pH 4.0, 1 vol of water equilibrated phenol and 0.2 vol chloroform/isoamyl alcohol (24 : 1). RNA was precipitated from the upper aqueous layer with 1 vol of isopropanol at - 2 0 °C. After centrifugation, the RNA pellet was treated with the denaturing solution and precipitated with 1 vol of isopropanol. After centrifugation, the RNA pellet was washed with 70% ethanol, centrifuged again, and dissolved in DEPCtreated water. RNA concentration was determined by absorbance at 260 nM, and integrity was assessed by ethidium bromide staining of RNA on agarose gels. Northern blots were performed by size fractionating 20 /zg of total placental RNA on a 1% agarose gel after denaturation with glyoxal. The RNA was transferred overnight onto Zeta probe nylon membranes (BioRad) by capillary action using 5× SSC (20× SSC = 3 M NaCI/0.3 M sodium citrate pH 7) as the transfer buffer. The membrane was baked at 80 °C for 2 h under vacuum, and transfer was assessed by staining a membrane strip with methylene blue. Slot blot hybridization was performed using a BioRad slot blot apparatus. Total placental RNA was denatured in glyoxal and loaded in a 5x SSC buffer. After transfer, the membrane was baked at 80 °C for 2 h under vacuum. Membrane hybridization was performed using the 3.2 kb MT IIa genomic clone of human metallothionein (19), which cross-hybridizes to all isoforms of

299

MT mRNA. GAPDH (Clonetech) DNA insert was used to normalize for the amount of total RNA. The membranes were pre-hybridized for 3 h at 68 °C in l x SSPE (0.1 M sodium phosphate/1.5 M NaC1/ 0.02 M EDTA), 2x Denhardts, 10% dextran sulfate, and yeast tRNA (200/zg/mL). Fresh solution was used overnight to hybridize digoxigenin labeled GAPDH DNA probe or digoxigenin labeled MT IIa RNA probe. The membranes were washed with 2× SSC/0.5% SDS at room temperature for 30 min, followed by three more washes in 0.1 x SSC/0.1% SDS at 68 °C. Hybridization was visualized by autoradiography on Kodak X-OMAT film. Densitometry was performed using the LKB Ultroscan XL Enhanced laser densitometer and only data within the linear range of the film was used.

MT in situ hybridization In situ hybridization was performed according to the method of Angerer and coworkers (21), with modifications. Deparaffinized tissue was rehydrated through an ethanol series and finally water. The 5 /zM tissue sections were treated with proteinase K (1/zg/mL) for 30 min at 37 °C, washed and dipped in 0.1 M triethanolamine-HC1 buffer (pH 8.0) and then triethanolamine buffer with acetic anhydride (I mL in 400 mL). The tissue was dehydrated and incubated overnight (55 °C, moist oven) with the MT IIa RNA probe solution (50% formamide, 30 mM NaC1, 10 mM Tris, pH 8, 1 mM EDTA, pH 8, l x Denhardts reagent (Sigma), 10% dextran sulfate, 0.5 mg/mL yeast tRNA and MT IIa RNA probe (0.6 ng//xL)). The optimal amount of probe (0.6 ng//zL) was determined by performing in situ hybridizations with MT IIa RNA probe concentrations of 0.1, 0.3, 0.6, and 1.2 ng//zL. Both sense and antisense probes were used in all in situ hybridization experiments. After hybridization the slides were washed in 4x SSC for 30 min at room temperature and then treated with RNAse A (20/xg/mL) (Sigma)/RNAse TI (I U/mL) (Boehringer-Mannheim, Indianapolis, IN). Washes with 2 x SSC and a stringent wash at 68 °C in 0.1 x SSC for 30 min followed. The cells were then washed in 0.1 x SSC for 30 min at room temperature and dehydrated. Detection of probe hybridization was performed according to the suppliers' instructions (Boehringer-Mannheim). Rehydrated tissue was incubated in buffer 1 (100 mM Tris-HC1, 150 mM NaC1, pH 7.5) and then incubated in buffer 1 with 0.3% Triton X-100 (Sigma) and 2% blocking solution (10 g blocking reagent (Boehringer-Mannheim)/0.1 M sodium maleate, pH 7.5). The tissue

300

Reproductive Toxicology

was then incubated with antidigoxigenin antibody (Boehringer-Mannheim) (1:500) in buffer 1 with 1% blocking reagent/0.3% Triton X-100, Following a wash in buffer 1 for 10 min, the slides were incubated overnight with NBT/X-phosphate (Boehringer-Mannheim) in buffer 2 (2% blocking reagent dissolved in buffer 1) overnight and washed in TE buffer (10 mM Tris-HCl/1 mM EDTA). Slides were dehydrated and coverslips were placed using Permount.

MT protein analysis MT protein was measured using a variation of the Cadmium-Chelex assay (22). Flash frozen samples of placental tissue were homogenized in a Waring blender in buffer A (3 volumes of 30 mM Tris-HC1, pH 7) and centrifuged at 100,000 g (60 min, 0 °C). Supernatant was then heated for 3 min in boiling water, and to this final fraction I0 /~L ~°9Cd (740 kBq/mL 1 mM; Amersham) was added and allowed to incubate for 30 min. Chelex-100 resin (I00/zL, 66% w/v; Bio-Rad) was then added, and the mixture was shaken for 15 min. The sample was centrifuged at 10,000 g for 10 min and an aliquot of this solution, in 2.5 mL buffer A, was then subjected to G-25 chromatography in a PD-10 column (Pharmacia). Fractions (0.5 mL) were collected and counted in a Micromedic 4/600 Plus Automatic gamma counter. Values for fractions corresponding to ~°gCd-MT were pooled.

Volume8, Number4, 1994 combined with the post hoc Tukey test. A P value of less than 0.05 was considered significant. RESULTS During the first 2 h and then for the following 8 h, samples of maternal and fetal perfusate were obtained for later analysis of hCG, glucose, and lactate. Immediate measurements of maternal and fetal pO2, maternal and fetal pH, and perfusate volume were also obtained. After perfusion with 20 /zM CdCI2 for 4 (n = 3) or 8 (n = 3) h, glucose, lactate, and oxygen consumption as well as net fetal oxygen transfer did not differ from control experiments without Cd. There were also no significant changes in fetal perfusate volume, which would indicate a

MT immunocytochemistry MT protein was localized in fresh and perfused human term placentae, using a monoclonal mouse antibody to MT (clone E9, Dako) and performed according to the Vectastain Elite ABC kit instructions (PK-4002, Vector Laboratories), Formalin fixed, paraffin-embedded placental tissue sections (5 tzM) were deparaffinized in xylene and rehydrated in an ethanol series. In order to block endogenous peroxidase activity, the slides were incubated in a solution of 3% hydrogen peroxide in methanol for 10 min. Slides were then incubated with normal serum blocking solution for 30 min and then MT primary antibody for another 30 min in a moist chamber. MT protein was visualized using peroxidase-conjugated avidin-biotin complex. Controls for nonspecific staining were performed by eliminating the primary antibody from the protocol.

Statistics Statistics were performed using the Minitab IBM statistical analysis software. Tests used to determine significance were the one-way ANOVA

MT

!

Fig. 1. Northern blot hybridization of total RNA from fresh and cadmium (20/xM CdC12)perfused placentae: 20 /zg of total RNA was size fractionated on a 1% agarose gel and transferred to nylon membrane. The RNA was probed with the digoxigenin-UTP labeled, antisense MT IIa RNA probe. Hybridization occurs at 0.7 kb. Con = control perfusion for 8 h with normal media, 4 h = perfused for 4 h with 20 txM CdCI2, 8 h = perfused for 4 h with 20 tzM CdC12 and an additional 4 h with normal media.

Metallothionein in the human placenta • J. G. BREEN ET AL.

301

MT mRNA and Protein Expression 100

a C:

8 ¢1 [ ] MT mRNA • MT protein

= m U.

= 0

_=

,,o Con o 8 hours

4 hours

20 ItM Cd8 hours

Fig. 2. Metallothionein mRNA and protein levels in the human, term placenta perfused with 20/xM CdCI2. Average values (n = 3 per group) and standard deviations are given. Slot blot hybridization was performed using the digoxigeninUTP labeled, antisense MT IIa RNA probe and the GAPDH digoxigenin labeled DNA probe (to normalize for total amounts of RNA). The ratio of MT to GAPDH signal was then used as the value for MT transcript level. This value was divided by the fresh placental value to give the fold increase above fresh. Protein levels are the averages of three n°gCd binding experiments. *indicates a significant difference from both control and the 4-h time point (p < 0.05, ANOVA with post hoc Tukey test).

leak in the placental vasculature. As noted for Cd exposure in previous placental perfusions, the release of hCG was significantly decreased after 2 h of exposure to 20/zM Cd (data not shown). Northern blot analysis revealed specific hybridization to a 0.7 kb class message in human placental RNA corresponding to MT transcripts. There was a time-dependent increase in MT mRNA in human placental tissue exposed to 20/~M CdCI2 (Figure 1). Specific hybridization to placental RNA was not observed with the sense-strand MT probe (data not shown). To measure the relative abundance of MT mRNA in the human placenta after perfusion with 20/zM CdC12, slot blot hybridization was performed on total R N A isolated from fresh human placentae, placentae perfused with normal media for an 8-h experimental period, placentae perfused for a 4-h experimental period with 20/zM C d C 1 2 , and placentae perfused for 4 h with 20/.tM CdCI2 and then 4 h more with normal media. In all experiments, a 2-h equilibration period preceded Cd exposure. MT mRNA levels were increased in a small but consistent manner after perfusion with normal media, compared to levels in fresh placental tissue, while after 4 h ofperfusion with Cd there was at least a 15fold increase in MT mRNA. A larger increase in MT transcripts was seen after a total of 8 h of perfusion following the initial 4-hour exposure to Cd. At this

8-hour time point, MT transcript abundance increased at least 70-fold relative to fresh tissue (Figure 2). In contrast to MT mRNA, while there were apparent increases in MT protein in some placentae, a statistically significant increase was not identified for the group (ANOVA with post-hoc T U K E Y test) (Figure 2). To determine which cell types accumulate MT mRNA during perfusion with Cd, in situ hybridizations were performed on paraffin tissue sections from all experiments. MT transcripts were barely detected in fresh tissue (Figure 3). Control perfusions demonstrated a small but consistent increase in the amount of MT transcripts relative to fresh tissue (Figure 3) and consistent with slot blot data (Figure 2). After 4 h of perfusion with Cd (20/zM CdCI2), there were increases in the intensity of signal. The trophoblast cells of the perfused tissue demonstrated small increases in MT mRNA. Large increases in MT transcripts occurred in cells of the villous core and fetal endothelial cells. After 8 h of perfusion (4 h with Cd and an additional 4 h without Cd) the intensity of hybridization in the stromal and fetal endothelial cells was increased dramatically above the 4-h exposure. There was no discernible change in the intensity of hybridization signal in trophoblast cells after the initial 4 h of exposure to Cd (Figure 3). This was a consistent finding in all Cd

Z

0

C~

Metallothioneinin the human placenta • J. G. BREENET AL. experiments. Therefore, much of the increase in MT mRNA abundance observed using slot blot hybridization can be accounted for by small increases in a large population of trophoblast cells and large increases in less numerous endothelial and villous core cells. To determine that MT protein synthesis increases in placental cells that also increase MT mRNA, immunocytochemistry was performed on paraffin tissue sections from all experiments. Localization in fresh tissue revealed low levels of MT protein in cells of the villous core, fetal endothelial cells, and trophoblast cells (Figure 4). After perfusion for 8 h with 20/zM Cd, immunodetectable MT protein levels in stromal and endothelial cells increased, while levels in trophoblast cells remained low (Figure 4). This was a consistent finding in all placentae examined. The highest levels of MT protein were observed in focal areas where edema in the villous architecture was also noted (Figure 4). This association between edema and higher levels of MT expression was most evident in 8-h perfusions with 20/xM Cd. No specific staining in control slides without the primary antibody was observed (data not shown).

DISCUSSION In the current study, the accumulation and colocalization of MT mRNA and protein, after exposure to CdCl2 (20/.LM) in the dually perfused, human term placenta, are reported. Placental perfusion has a number of advantages, including the ability to verify the metabolic integrity of the tissue during a potentially toxic exposure for extended periods of time (15,16). Also, by exposing the placenta to Cd in the maternal perfusate, one is able to model the route of exposure in vivo (8). Therefore, using the perfusion model, placental MT analysis is the product of exposure to Cd exclusively through the maternal circulation. After exposure to CdCI2 for only 4 h, the perfused tissue accumulated 15-fold higher levels of

303

MT transcripts than were found in fresh tissue. While this was a large accumulation of MT message, the difference was not significant compared to control perfusions. In these studies, there appears to be a subset of placentae that do not accumulate MT mRNA as rapidly as others. For this reason, a statistically significant increase in MT mRNA was not observed until the final time point of 8 h. After 8 h of perfusion with Cd, the placentae accumulated an average of 70-fold higher levels of MT transcripts than noted in fresh tissue. This increase was significantly higher than control (p < 0.05). Even though variability in MT induction persisted, all placentae were responsive at this time point, and the lowest level of accumulation was still 30-fold higher than the level observed in fresh tissue. In a previous study, only marginal increases in MT protein (< 2-fold) have been demonstrated in the term human placenta after perfusion with CdCl2 (20/.tM) for 5 h (14). Also, cultured term trophoblast cells accumulate MT protein, after 24 h of exposure to 20/zM Cd, only 3- to 4-fold higher than levels in control cells (12). In the current study, while MT protein levels were increased as much as 3-fold in some placentae, the overall levels of MT protein did not increase significantly even after 8 h of perfusion. This result is consistent with other studies that demonstrate MT protein accumulation lagging behind the accumulation of MT mRNA (23,24). The physiologic importance of these modest, general tissue increases of MT protein is unclear. Cell specific increases in MT protein, however, may protect certain cells from Cd toxicity. As expected, in situ hybridization assays in fresh tissue revealed consistently low levels of MT mRNA. This is concordant with the results of the slot blot hybridization analysis for MT mRNA. Control perfusions (without Cd) for 8 h resulted in a barely detectable increase in MT transcripts in scattered trophoblast, stromal, and endothelial cells. Otherwise, most cells had MT transcript levels similar to fresh placental tissue. These results indicate,

Fig. 3. In situ hybridization for metallothionein mRNA in the human term placenta following perfusion with cadmium. In situ hybridization was performed using the antisense MT IIa RNA probe labeled with digoxigenin-UTP. Detection of message was performed with an alkaline phosphatase-antibody to digoxigenin. The color reaction, ranging from light brown to dark purple, is indicative of increasing amounts of MT message, a) MT antisense hybridization to fresh human term placental tissue, b) Hematoxylin and eosin stained adjacent section from fresh human term placenta, c) MT antisense hybridization to human term placenta perfused with normal media for 10 h, without Cd. d) Hematoxylin and eosin stained adjacent section from human term placenta perfused with normal media for 10 h. e) MT antisense hybridization to human term placenta perfused for 8 h with 20/~M CdCI2. T = trophoblast cells, S = stromal cells of the villous core (Hofbauer cells), and E = fetal endothelial cells, f) Hematoxylin-stained adjacent section from human term placenta peffused with 20/xM CdCI2for 8 h. g) MT sense hybridization to human term placenta perfused for 8 h with 20 p.M CdC12. Magnification = 146x for all photos.

304

Reproductive Toxicology

Volume8, Number 4, 1994

a

Fig. 4. In situ localization of MT protein in the human term placenta following perfusion with cadmium. Immunocytochemistry was performed using a monoclonal mouse anti-MT antibody (clone E9, Dako) and visualized using a peroxidase substrate kit (DAB) (Vector laboratories). The color reaction, ranging from light to dark, indicates the level of MT protein. (a) MT protein in fresh tissue. T = trophoblast cells, S -- stromal cells of the villous core (Hoffbauer cells), and E = fetal endothelial cells. (b) MT protein in human term placenta perfused for 8 h with 20/.LM CdCl2. T -- trophoblast cells, S -- stromal cells of the villous core (Hoffbauer cells), and E = fetal endothelial cells. Magnification = 200x for all photos.

in agreement with the slot blot data, that control perfusions of up to 8 h have very little effect on the abundance o f M T transcript. This observation is important when one considers the numerous conditions that have been shown to induce MT expression in a variety of systems. In placentae perfused for 4 h with CdC12 (20 /zM), there were increases in the intensity of MT

hybridization in stromal and endothelial cells. The level of M T message in trophoblast cells was only marginally increased above levels observed following control perfusions (without Cd). During 8-h Cd experiments, the amount of MT transcripts in stromal and endothelial cells dramatically increased. We had e x p e c t e d the levels of MT m R N A in all cell types to increase after 8 h, especially in trophoblast

Metallothionein in the human placenta • J. G. BREEN ET AL.

cells, which are in direct contact with Cd in the maternal perfusate. However, hybridization in trophoblast cells was unchanged between 4 and 8 h of perfusion with Cd. This observation was further supported by reducing the antisense MT probe concentration to 0.1 ng//zL where hybridization signal was absent in trophoblast cells, but intense signal remained in cells of the villous core and endothelial cells. Note that, after the initial 4-h exposure, fresh perfusate without Cd was added and, therefore, the tissue was no longer exposed to Cd in the maternal perfusate. It has been shown previously in this laboratory that Cd is accumulated in the placenta to levels 50 times that of the maternal perfusate (8,9). After 8 h of perfusion with Cd, placental Cd levels in the present study, in agreement with the previous reports, were 7.8 --- 0.6/.tg/g (unpublished data-Carol Eisenmann). Therefore, the increased amounts of MT mRNA observed after four additional hours of perfusion were, most likely, the direct or indirect result of Cd already concentrated in the tissue. MT protein was co-localized with MT mRNA, suggesting the sites of MT synthesis. After perfusion with 20/zM Cd for 8 h, immunodetectable MT protein is correlated with the histologic appearance of stromal edema (Fig. 4c). Wier and associates (9) have demonstrated sporadic stromal edema in placentae perfused with I0/xM Cd for 8 h, while 20/~M Cd produced more extensive edema during 8-h perfusions. Whether the increases in MT transcripts and protein in stromal and endothelial cells in these affected areas were the result of local concentrations of Cd or the result of an indirect effect of the Cd-induced edema has not been established. Many xenobiotics and disease states, including cycloheximide, maternal diabetes mellitus, syphilis, toxoplasmosis, and cytomegalovirus, are able to produce many of the pathologic effects observed following placental Cd intoxication (subsyncytiotrophoblastic vacuoles, vacuolization of Hofbauer cells, and villus stromal edema) (1,9,25). Villus stromal edema has been associated with fetal morbidity and mortality (26). Thus, the induction of MT mRNA and protein in endothelial cells and stromal cells may afford protection to these cell populations during toxic insult from a number of stressors. Consequently, in 8-hour perfusions with 100 ~M Cd when some stromal edema is evident and endothelial cell morphology appears normal, trophoblast cells are necrotic (9). In fact, in both the rodent and human placenta, the most dramatic effects of Cd toxicity are observed in the trophoblast cells (9,27). Therefore, the lower level of MT induction in

305

trophoblast cells may make these cells more sensitive to toxicants. - - We would like to thank the clinical staff in Obstetrics and Gynecology at the Strong Memorial and Highland Hospitals for their assistance in obtaining human placentae. We would also like to thank Lynn Jessee for technical assistance with placental perfusions and Richard H. Watkins for in situ hybridizations. This research was supported in part by National Institutes of Health grants ES02744 and ES01247 and an NIEHS Training Grant (ES07026- for J. Breen and C. Eisenmann).

Acknowledgments

REFERENCES 1. Benirschke K, Kaufmann P. Pathology of the human placenta. 2nd ed. New York: Springer-Verlag; 1990. 2. Slikker W, Miller RK. Placental metabolism and transfer-role in developmental toxicology, In Kimmel CA, BuekeSam J, eds. Developmental toxicology. 2nd ed. New York: Raven Press; 1994:245-86. 3. Parizek J. Vascular changes at sites of estrogen biosynthesis produced by parenteral injection of cadmium salts: the destruction of the placenta by cadmium salts. J Reprod Fertil. 1964;7:263-4. 4. Levin AA, Miller RK. Fetal toxicity of cadmium in the rat: maternal vs fetal injections. Teratology. 1980;22:1-5. 5. Levin A, Miller RK. Fetal toxicity of cadmium in the rat: Decreased utero-placental blood flow. Toxicol Appl Pharmacol. 1981;58:297-305. 6. Levin AA, Miller RK, di Sant'Agnese PA. Heavy metal alterations of placental function: A mechanism for the induction of fetal toxicity in cadmium. In: Clarkson TW, Nordberg GF, Sager PR, eds. Reproductive and developmental toxicity of metals. New York: Plenum Press; 1983:633-54. 7. Sonawane BR, Nordberg M, Nordberg GF, Lucier GW. Placental transfer of cadmium in rats: Influence of dose and gestational age. Environ Health Perspect. 1975;12:97-102. 8. Wier PJ, Miller RK. The pharmacokinetics of cadmium in the dually perfused human placenta. Trophoblast Res. 1987;2:357-66. 9. Wier PJ, Miller RK, Maulik D, di Sant'Agnese PA. Toxicity of cadmium in the perfused human placenta. Toxicol Appl Pharmacol. 1990;105:156-71. 10. Margoshes M, Vallee BL. A cadmium protein from equine kidney cortex. J Am Chem Soc. 1957;79:4813-4. 11. Andrews GK. Regulation of metallothionein gene expression. Progr Food Nutr Sci. 1990;14:193-258. 12. Waalkes MP, Poisner AM, Klaassen CD, Wood GW. Metallothionein like proteins in human placenta and fetal membranes. Toxicol Appl Pharmacol. 1984;74:179-84. 13. Goyer RA, Haust MD, Cherian MG. Cellular localization of metallothionein in the human placenta. Placenta. 1992;13: 349-55. 14. Boadi WY, Yannal S, Urbach J, Brandes JM, Sumner KH. Transfer and accumulation of cadmium and the level of metallothionein in perfused human placentae. Arch Toxicol. 1991 ;65:318-23. 15. Miller RK, Wier PJ, Maulik D, di Sant'Agnese PA. Human placenta in vitro: characterization during 12 hr of dual perfusion. Contrib Gynecol Obstet. 1985;13:77-84. 16. Miller RK, Wier PJ, Perez D'Gregorio R, Eisenmann C, di Sant'Agnese PA, Shah Y, Neth Jessee L. Human dual placental perfusions--criteria for toxicity evaluations. Meth Toxicol. 1993;3B:246-59. 17. Panigel M. Placental peffusion experiments. Am J Obstet Gynecol. 1962;84:1664-72. 18. Schneider H, Panigel M, Dancis J. Transfer across the per-

306

19. 20. 21.

22. 23.

Reproductive Toxicology fused human placenta of antipyrine, sodium and leucine. Am J Obstet Gynecol. 1972;114:822-8. Karin M, Richards RI. Human metallothionein genes--primary structure of the metallothionein-II gene and a related processed gene. Nature. 1982;299:797-802. Chomcynski P, Sacchi N. Single-step method of RNA isolation by acid guanidiinium thiocyanate-phenol-chloroform extraction. Anal Biochem. 1986;162:156-9. Angerer LM, Stoler MH, Angerer RC. In situ hybridization with RNA probes: an annotated recipe. In Valentino KL, ed. In situ hybridization: applications to neurobiology. New York: Oxford University Press; 1987:42-70. Bartsch R, Klein D, Summer KH. The Cd-Chelex assay: a new sensitive method to determine metallothionein containing zinc and cadmium. Toxicology. 1990;64:177-80. Bauman JW, Cherukury M, McKim JM, Liu Y, Klaassen

Volume 8, Number 4, 1994

24. 25.

26. 27.

CD. Induction of hepatic metallothionein by paraquat. Toxicol Appl Pharmacol. 1992;117:233-41. Luce MC, Schyberg JP, Bunn CL. Metallothionein expres~ sion and stress responses in aging human diploid fibroblasts. Exp Gerontol. 1993;28:17-38. Miller RK, Wier PJ, Shah Y, Perez-D'Gregorio R. Human placental protein production in vitro: 12 hours of perfusion. In: VSP; Utrecht, The Netherlands: Placental and endometrial proteins. Tomoda Y, Mizutani S, Narita O, Klopper A, eds. 1988:453-60. Naeye RL, Maisels J, Lorenz RP, Botti JJ. The clinical significance of placental villous edema. Pediatrics. 1983;71: 588-94. di Sant'Agnese PA, DeMesy Jensen K, Levin A, Miller RK. Placental toxicity of cadmium in the rat: an ultrastructural study. Placenta. 1983;4:149-64.