Transport of immunoglobulin G and its subclasses across the in vitro-perfused human placenta

Transport of immunoglobulin G and its subclasses across the in vitro-perfused human placenta Antoine Berne,

Malek,

PhD, Ruth

Sager, Anthony

Zakher,

BSc, and Henning

Schneider,

MD

Switzerland

OBJECTIVE: The transport of immunoglobulin G and its subclasses 1 to 4 was investigated in the in vitro-perfused isolated cotyledon of the human placenta. STUDY DESIGN: An in vitro system with separate perfusion of the villous capillary system (fetal compartment) and the corresponding intervillous space (maternal compartment) was set up in an isolated cotyledon of human term placenta. After a 2-hour control phase with both compartments perfused in a closed circuit with NCTC-135 tissue culture medium together with Earl’s balanced salt solution (2: l), media were exchanged in both circuits and for the experimental phase immunoglobulin G (Sandoglobulin) together with carbon 14-labeled bovine serum albumin (5-10 r&i) was added to the maternal compartment at a concentration of 6 gm/L. During the experimental phase, lasting between 2 and 5 hours, samples were taken from the maternal and fetal compartments every 30 minutes up to 2 hours and every 60 minutes thereafter. RESULTS: During the control phase immunoglobulin G appeared in the maternal perfusate and reached a plateau at 60 to 80 mg/L, whereas the concentration in the fetal perfusate did not exceed 20 mg/L. A similar pattern of release was observed for hemoglobin, suggesting a washout of remains of blood from the intervillous space and the villous vascular compartment. After addition of immunoglobulin G to the maternal circuit during the first 2 hours in three of four experiments, no change in immunoglobulin G concentration was seen in the fetal circuit, and only in the fourth and fifth hours did the fetal concentration increase to 0.6% of the maternal concentration. In contrast, carbon 14-labeled bovine serum albumin was already detectable in the fetal circuit after 1 hour, but the level remained constant at 0.1% of the maternal concentration. Total immunoglobulin G transfer was estimated at 0.5% of the amount added to the maternal circulation, which was five times higher than total transfer of bovine serum albumin, Transfer was shown for all four subclasses. At the end of the experiment the ratio of immunoglobulin G, to immunoglobulin G, in the fetal perfusate was significantly higher than in the maternal perfusate (3.8 vs 1.8) suggesting preferential transfer of immunoglobin G,. CONCLUSION: Transfer of all four immunoglobulin G subclasses of a commercially available immunoglobulin G preparation across the human placenta from the maternal to the fetal side was demonstrated by the dual in vitro perfusion system. There is a preferential transfer for immunoglobulin G,. (AM J OBSTET GYNECOL 1995;173:760-7.)

Key words:

In vitro perfusion,

placental

transfer,

immunoglobulin

Intrauterine or perinatally acquired infections are important causes of perinatal mortality and morbidity. Because of the immaturity of the fetal immune system, immunoprotection of the fetus and neonate is dependent on the transmission of maternal antibodies across the placenta to the fetus during the last weeks of pregnancy plus the uptake of antibodies with the colostrum during breast-feeding.

From the Department of Obstetrics and Gynecology, University of Berne. Supported by the Swiss National Foundation (grants No. 3226361.89 and 32-39676.93). Received for publication February 14, 1994; revised December 21, 1994; accepted Januaq 12, 1995. Reprint requests: Antoine Malek, PhD, Department of Obstetrics and Gynecoloa University of Berne, Schanrenec&trasse 1, CH-3012 Bemze, Switzerland. Copyright 0 1995 by Mosby-Year Book, Inc. 0002.9378/95 $5.00 f 0 6/l/63408 760

G subclasses, human

Immunoglobulin G (IgG) circulating in the mother may be the result of active immunization or it may have been administered to the mother to provide passive immunity. The issue of active immunization during pregnancy of mothers lacking immunity against specific infections, in particular when living in endemic areas or when at increased risk of exposure, has recently received considerable interest.’ The IgG subclass production may vary with the immunizing agent, with polysaccharide antigens producing predominantly IgG, and protein-based vaccines producing IgG,.‘, ’ Differences in placental transfer of different subtypes therefore could be of relevance in the context of immunization in pregnancy.’ Pregnant women with immunothrombocytopenia have been treated by intravenous infusion of immunoglobulin and, although this treatment has proved effective in raising maternal platelets, reports on the influ-

Volume 173, Number Am J Obstet Gynecol

3, Part

1

ence on neonatal thrombocytopenia have been conflicting? The lack of an effect on neonatal platelet count has been interpreted as an absence of placental transfer of therapeutically administered IgG. Smith and Hammarstrom,* however, have shown that with continuous treatment of the mother high neonatal IgG levels were seen at delivery. For a rational use of active immunization in pregnancy or of intravenous treatment with immunoglobulins for platelet autoimmune or alloimmune disease with the objective of providing the fetus with IgG, a thorough understanding of the placental transport mechanism is required. We therefore applied the dual in vitro perfusion system of an isolated cotyledon of the human placenta to study whether there is placental transfer of a commercially available preparation of immunoglobulin and whether there are differences in transfer rates for the four IgG subclasses. The question of proteolytic breakdown during placental passage was also addressed.

Material and methods In vitro perfusion method. For this study placentas obtained from uncomplicated deliveries at term after normal pregnancies, as judged by an appropriately grown newborn with normal Apgar scores and cord blood gas values, were used. After either vaginal or cesarean delivery the placenta was immediately taken to the laboratory, and a cotyledon was prepared for dual in vitro perfusion as originally described by Schneider et al9 After cannulation of a pair of a chorionic artery and vein the perfusion of the corresponding villous capillary system was started (fetal compartment). The isolated lobule was fixed in a perfusion chamber that was surrounded by a jacket connected to a warm water circuit at 37” C. Three or four blunt metal cannulas were introduced into the intervillous space by penetration of the decidual plate and connected to a second circuit for perfusion of the maternal compartment. The per&sate was composed of tissue culture medium NCTC-135 diluted with Earl’s solution (2 : 1) with addition of glucose (2 gm/L), dextran 40 (10 gm/L), heparin (2500 IU/L), and clamoxyl (250 mg/L). In each circuit a volume of 130 ml of perfusate was recirculated and continuously equilibrated with 95% oxygen and 5% carbon dioxide. Flow rates of 12 and 4 to 6 mlimin were used in the maternal and fetal circuits, respectively. Perfusion experiments. All experiments included a 30-minute period (prephase) of open perfusion of both compartments, to flush the blood out of the intervillous space and the villous vascular compartment and to allow recovery of the placental tissue from the ischemic period after delivery before the initiation of the artificial perfusion. The diffusion rate of antipyrine and creatinine was measured to assure adequate perfusion of the intervillous space and a good overlap of the two com-

Malek et al.

761

partments. A low perfusion pressure in the fetal circuit 5 20 mm Hg was another indicator of an intact cotyledon. The actual experiment started with the closure of the perfusion circuits and a control phase of 2 hours with recirculation on the maternal and fetal side. Any significant decrease in per&sate volume in the fetal circuit indicates a defect in the membrane, and whenever the volume loss exceeded 4 ml/hr the experiment was terminated. After the 2-hour control phase, the perfusate was exchanged against fresh media on both sides and IgG (6 to 9 gm/L) was added to the maternal circuit. In all experiments carbon 14-labeled bovine serum albumin (10,000 to 30,000 disintegrationslmin per milliliter) was added as a macromolecular marker to the maternal side. IgG (Sandoglobulin) was provided by the Central Laboratory of the Swiss Red Cross, Berne. Radioactivity counting. The fraction of proteinbound Y-labeled bovine serum albumin radioactivity was determined by precipitation with trichloroacetic acid. A sample volume of 200 ~1 was precipitated with 400 ~1 of trichloroacetic acid (1 mol/L). After centrimgation at 14008 for 10 minutes the supernatant and pellet were dissolved separately in 10 ml of liquid scintillant and the 14C radioactivity was counted with a B-spectrometer (LS-180 1, Beckman, Fullerton, Calif.) with automatic quench correction with appropriate standards. IgG enzyme-linked immunosorbent assay. The following solutions were prepared: (1) solution A, 0.1 mol/L sodium bicarbonate, pH 9.8; (2) solution B, 0.9% sodium chloride (wt/vol) and 0.05% Tween-20 (wtivol); (3) solution (3, phosphate-buffered saline solution containing 0.05% Tween-20 (vol/vol), pH 7.3; and (4) solution D, substrate buffer containing 1 mmol/L diethanolamine and magnesium chloride, pH 9.8, with 4-nitrophenyl phosphate disodium salt 1 mg/ml as substrate. One hundred microliters of solution C containing rabbit antihuman IgG (1 pg/ml, Dakopatts, Glostrup, Denmark) was added to each well in microtiter plates. After overnight incubation (15 to 17 hours) at room temperature, the contents were aspirated. After three washes with solution B the plates were incubated with 300 ~1 of bovine serum albumin (1% [wt/vol] in solution C) for 1 hour to saturate the surfaces preventing nonspecific binding during the following steps. One hundred microliters of diluted standards or samples in solution C was added to the wells in the first row of the microtiter plates, followed by a 1: 2 dilution sequence in each column. After 2 hours of incubation at 37” C the contents were aspirated and each well was washed three times with solution B (300 ~1). Rabbit antihuman IgG conjugated with phosphatase (Dakopatts) was diluted 1: 1000 with solution C and added to each well (100 JJJ). After incubation for 2 hours the contents were aspirated and the wells washed four times with solution B (300 ~1). To each well 100 l.~l of

762

Malek

et al.

September 1995 Am J Obstet Gym01

60

Time (min) Fig. 1. Endogenous IgG appearing in maternal and fetal circuits during 2-hour control phase. Values are mean t SD, n = 4. Maternal (m) (closed triangles) vs fetal (F) ( cI osed circles) concentration: asterisk, fi < 0.01; tuo asterisks, fi < 0.001.

substrate buffer (solution D) in a defined order was added. After 30 minutes of incubation at 37” C the reaction was stopped, in each well with the same order, with 20 k1 of sodium hydroxide (4 mol/L). After 15 minutes the absorbance was determined at 405 nm with a spectrophotometer (Titertek Multiskan Plus, Flow Laboratories, Helsinki). A concentration linearity for IgG standard was found between 23 and 93 rig/L. A sample of human pregnancy serum was used as a control. The intraassay and interassay variations in control samples were 6.07% and 8.41%, respectively. IgG subclasses 1 to 4 were determined with an enzyme immunoassay kit (Binding Site, Birmingham, United Kingdom) with antibodies evaluated by an International Union of Immunological Society World Health Organization collaborative study.” Briefly, samples were incubated with standard and control human serum on precoated plates with subclass-specific mouse monoclonal antibodies. After the wells were washed, a second incubation with sheep antihuman IgG coupled to horseradish peroxidase followed. The reaction product was developed with orthophenylenediamine in buffered hydrogen peroxidase. The reaction was stopped with sulfuric acid, and the absorbance was determined at 450 nm with a spectrophotometer (Titertek Multiskan Plus, Flow Laboratories). The coefficient of variation evaluated for IgG,., in control human serum for intraassay and interassay was between 5.9 and 11.6. Hemoglobin was determined after conversion into hemoglobin cyanide with a Merckotest kit (Merck, Germany). Western immunoblot analysis. Perfusate samples were electrophoresed by polyacrylamide gel electrophoresis under nondenaturing conditions, together

with biotinylated molecular weight markers on 8% to 25% sodium dodecyl polyacrylamide gels and transferred to polyvinylidene difluoride membranes (TransBlot, Bio-Rad, Hercules, Calif.) with semidry electrophoretic transfer (Phastsystem, Pharmacia, Uppsala). Immunodetection was performed with a primary polyclonal antibody to human IgG produced in rabbit (Dako) at a dilution of 1: 600. Primary antibody binding was detected with a biotinylated secondary antibody (Dako) at a dilution of 1:500, followed by streptavidin-biotin-peroxidase (Dako). Diaminobenzidine tetrahydrochloride was used for detection of peroxidase activity (Sigma, St. Louis). Statistical analyses were made with the Mann-Whitney two-sample test and the Wilcoxon test.

Results During the P-hour control phase there was a release of endogenous IgG into both circuits, reaching a plateau that, on the maternal side, was three to four times higher than in the fetal circuit. At the end of the control phase the IgG concentration was 87.5 i. 31.8 mg/L in the maternal and 21.3 t 7.9 mg/L in the fetal circuit (Fig. 1). A similar rise was seen in both circuits for hemoglobin concentration. After 2 hours of recirculation the concentration in the maternal circulation was 504 f 106 mg/L, which again is about three times the mean level of 140 -+ 44 mg/L reached on the fetal side (Fig. 2). After media exchange and addition of exogenous IgG and ‘X-labeled bovine serum albumin to the maternal circuit, the release of hemoglobin into both circuits continued and a new plateau was reached after 1 hour, which was slightly lower than in the control phase.

Volume

173, Number

3, Part 1

Malek

Am J Obstet Gynecol

zi

et al.

763

r cui t 0 F-Circuit

0

60

120

Time (min) Fig. 2. Hemoglobin release into maternal and fetal circuits during 2-hour control phase. Values are mean * SD, 12 = 4. Maternal (m) (closed triangles) vs fetal (F) (closed circles) concentration: asterisk, p c 0.01; two asterisks,p < 0.001.

60

120

180

240

300

Time (min) Fig. 3. IgG concentrations in fetal circuit during test phase. After 2-hour control phase media were exchanged and IgG was added to maternal circuit at concentration of 6 to 9 gm/L (n = 4).

The IgG levels measured in the fetal compartment during the first 2 hours after exchange of media and addition of exogenous IgG to the maternal side resembled the curve observed during the control phase (Fig. 3). The fetal concentrations after the first 2 hours of the test phase were, however, significantly lower when compared with the values found at the end of the control phase, indicating a continuation of the washout of remaining blood from villous capillaries, with no significant transfer of IgG from the maternal to the fetal side. During the fourth and ‘fifth hours after the addi-

,tion of IgG a clear rise in fetal IgG concentration was seen, demonstrating transfer from the maternal to the fetal side (Fig. 3). In one of the four experiments IgG transfer apparently started earlier with a linear rise in the fetal IgG concentration from the beginning of the test phase. The mean level of IgG measured on the fetal side after 300 minutes was 58.7 rt 12.4 mg/L, which is significantly higher than the concentration of 26.3 -+ 21.1 mg/L at 120 minutes ($J < 0.007). The individual values of the final concentration in the fetal circulation were 0.6%, 0.4%, 0.3%, and 0.8% of the initial maternal

764

Malek

et al.

September

Am J Obstet

MAB

67K43K30K, 20.1 K-w 14.4K-

MC

1995

Gynecol

DEFM

;;e

-7.

%u*

-

-

-

-

Fig. 4. Detection of IgG by Western immunoblot in perfusate samples collected from maternal and fetal circuits during control and test phases of one perfusion experiment. M, Molecular weight markers. Maternal (A) and fetal (B) samples were collected at end of control phase. From test phase initial maternal sample (C), maternal and fetal samples after 120 minutes (0, E), and fetal (F) sample after 300 minutes were collected.

level.

The

maternal

IgG

concentration

dropped

con-

and after 5 hours of perfusion reached a level of 4.2 to 6.9 gm,‘L. Maternal samples and the corresponding fetal samples at 120 minutes and at the end of the test phase were analyzed to compare the distribution of the IgG subclasses by use of an enzyme-linked immunosorbent assay (Table I). At 120 minutes and at the end of the experiment the fraction of IgG, was significantly higher in fetal samples than in the initial maternal sample. In contrast, fetal IgG, after 2 and 5 hours of pe&sion, expressed as a fraction of total IgG, was significantly lower than in the maternal sample. The ratio of fetal IgG,/IgG, at 120 minutes and at the end of the experiment was 2.8 and 3.8, respectively, which is significantly higher than the ratio in the maternal sample (Table I). Western blot analysis was performed to exclude breakdown of IgG. To demonstrate the ability of the Western immunoblot technique to detect peptides as fragments of IgG, a tryptic digestion of the IgG preparation (Sandoglobulin) was performed under the protein digestion conditions previously described.” The mixture of tryptic peptides was then electrophoresed. Western blot of the digested IgG showed a wide range of peptides, confirming the ability of the antibody to detect IgG fragments. The experimental samples horn the maternal and fetal sides showed identical patterns, with only one band in the range of 160,000 and no small IgG fragments (Fig. 4). We therefore could not confirm the breakdown during transfer, as was previously seen with iodine 125-labeled 1gG.l’ In addition, a serum sample obtained from the umbilical vein after delivery at term was also analyzed. Only one band was identified as IgG, with a molecular weight of 150,000 d. During the experiment the concentration of 14Clabeled bovine serum albumin in the maternal circuit dropped to 71% to 82% of the initial value. In contrast to IgG, where evidence of transfer from the maternal to tinuously

the fetal circuit could only be shown after 2 hours (Fig. 3), trace amounts of trichloroacetic acid-precipitable 14C activity were detectable on the fetal side 60 to 90 minutes after addition to the maternal perfusate. Over 5 hours there was no hrther rise in concentration of trichloroacetic acid-precipitable 14C in the fetal circuit, and the individual values of the final fetal concentration expressed were 0.07%, 0.13%, 0.06%, and 0.09 of the initial maternal level with a mean of 0.09%, which is significantly less than the corresponding value for IgG ($J < 0.001).

Comment Definitive proof of the transport of therapeutically administered IgG across the placenta is lacking. The technique of dual in vitro perfusion of an isolated cotyledon of the human placenta has been used to investigate active transport mechanisms and passive difision of small molecules.‘3, I4 For hydrophylic molecules a close correlation between permeability and molecular weight was found over a range of 60 (urea) to 5000 (inulin) d.14 More recently, the same technique has been used to study the permeability for macromolecules such as dextran, horseradish peroxidase, heparin, erythropoietin, and albumin.‘S-‘7 The only previous in vitro perhsion study of placental IgG transfer used labeled IgG and showed considerable breakdown, with a release of peptides into the fetal circulation.‘” There is serious concern that the iodine-labeling process may lead to changes in the molecular structure of IgG or other proteins so that these molecules may not be suitable as test compounds for all experimental purposes. The commercial preparation of IgG used in these experiments fulfills the criteria for integrity and purity as defined for therapeutic use by the World Health Organization.” More than 90% is present as intact IgG with

no

detectable

proportion

of

smaller

fragments,

Volume 173, Number Am J Obstet Gynecol

Table

3, Part 1

I. Concentration

Malek

of IgG subclasses as percent Maternal 59.5 32.5 4.8 3.3

initial t f + f 1.8

et al.

765

of total IgG

level

1.8 3.2 1.5 1.1

and the distribution of the IgG subclasses compares with the distribution found in normal adult plasma. Because a composite of unlabeled IgG was used in these experiments and in view of the very slow transfer process, any interference of trace amounts of endogenous IgG that continue to be washed out from the placental compartments had to be taken into account. The similarity in the washout profiles of IgG and hemoglobin during a 2-hour control phase suggests that the IgG appearing in the maternal and fetal compartment is part of the blood remaining inside the vessels or the intervillous space. Any transfer had to be measured against the background of this washout. It is interesting that for both IgG and hemoglobin the amount washed out on the maternal side was three to four times higher than on the fetal side, although normally levels for IgG and hemoglobin are higher in fetal than in maternal blood. This could be explained either by a more complete rinsing of the villous vascular system compared with the intervillous space during the initial 30-minute rinsing period or by a larger compartment being perfused on the maternal side compared with the fetal side. Various earlier attempts to measure the relative proportion of total placental volume represented by the intervillous space and the villous vascular system, respectively, gave inconsistent results. More recently it could be shown that both spaces are approximately equal in volume and each makes up 9% to 12% of the total tissue vo1ume.‘9 However, even these measurements were performed on delivered placentas, and their significance for the in vivo situation remains unclear. There is general agreement that placental transport of endogenous IgG is a Fc receptor-mediated process that predominantly develops in the third trimester of pregnancy. All three known leukocyte Fc receptor classes are present in the major cellular constituents ofthe placental barrier, including trophoblast, Hofbauer, and endothelial cells.” Uptake of maternal IgG is mediated by Fc receptors located on the syncytiotrophoblast.‘~ ‘I With different techniques such as immunocytochemistry combined with immunofluorescence, colloidal gold-complexed protein A, peroxidase-conjugated human IgG or tritiated IgG, several authors have shown that an adsorptive endocytosis in coated vesicles is the primary process to protect IgG against lysosomal digestion during pla-

Fetal I.20 min concentration 66.9 t 1.9 24.2 + 0.7 3.5 k 0.5 5.0 f 0.7 2.8

Fetal final

concentration

71.8 f 2.4 18.9 f 2.3 4.0 t 0.8 5.3 + 0.5 3.8

cental passage.3’ p2-25In addition to coated vesicles, uncoated vesicles, which are possibly responsible for the degradation of IgG in the lysosomal system, have been described.26, ” Of particular interest is the lag time of approximately 2 hours before IgG originating from the maternal compartment appeared on the fetal side. Horseradish peroxidase-labeled IgG was used to trace the pathway by electron microscopy, and the time needed for transfer was estimated by studying the tissue after different durations of in vitro perfusion. ** It took about 2 hours of perfusion of the intervillous space with labeled IgG before horseradish peroxidase tracer appeared in the endothelial cells of the villous capillaries. These findings obtained by cytochemistry of labeled IgG would explain the lag time observed until a rise in fetal IgG concentration could be shown in our experiments. In a perfusion system similar to the one used in this study, Morgan et aI.” observed a lag period of 2 to 3 hours before the transfer of original human serum IgG was observed, and for commercial intravenous immunoglobulin (Sandoglobulin) even 3 to 4 hours were needed. It was speculated that Sandoglobulin may contain a factor inhibiting the transfer of IgG. This inhibitory effect was confirmed by a reduced transfer of a platelet-specific antibody from the maternal to the fetal side after addition of Sandoglobulin to the maternal circulation.“8 In a separate series of experiments a higher concentration, 25 to 28 gm/L, of IgG was used on the maternal side. Consistent with the experiments using the 6 to 9 g&L concentration, there was no apparent transfer with a higher IgG concentration during the first 2 hours. Because of an increasing leak of fluid from the fetal to the maternal side, these experiments had to be terminated after 2 to 3 hours. The leak could be a result of the colloidosmotic pressure gradient, but this observation has so far not been explored any further. “C-labeled bovine serum albumin served as a macromolecular marker for the passive diffusion in the perfused placental lobule, and trace amounts of trichloroacetic acid-precipitable 14C activity appeared in the fetal circulation. The difference in transport behavior between bovine serum albumin and IgG is remarkable. The lag period before transfer could be detected was not seen for trichloroacetic acid-precipitable l*C activity and, although the total transfer of IgG over the observed time period of 5 hours was small, it was 5

766

Malek

et al.

times higher than that of bovine serum albumin, which in view of the larger molecular size of IgG (150,000 vs 68,000 d), is substantial. Further indirect evidence for a specific transport mechanism comes from the differences in transfer observed between IgG subclasses. The significantly higher concentration of IgG and its subclasses found in the fetal circuit after 5 hours, compared with 2 hours, indicates that all IgG subclasses cross the placenta. However, because of differences in transfer rates, the ratio of IgG,/IgG, changed and was significantly higher in the fetal than in the maternal circuit and increased between 2 and 5 hours. The relative concentrations of IgG, and IgG, remained similar in both circuits. We have recently shown a preferential transfer of IgG, compared with IgG, when concentrations in paired maternal peripheral vein and umbilical vein and artery samples were measured.“’ This study has shown a substantial increase in the fetal level of IgG during the third trimester, and at term the level was significantly higher in fetal than in maternal blood. Although a considerable rise in transport capacity could be shown for all subclasses, the increase in fetal IgG concentration in the third trimester was primarily related to the rise in IgG,, which at term in fetal blood was even higher than the total maternal IgG concentration. The level of IgG, in the fetal blood remained below the maternal concentration. The fraction of IgG, and IgG, in both maternal and fetal circulations were similar. We did not find any substantial degradation of IgG into smaller molecular weight fragments and therefore cannot confirm the previous findings of Contractor et al., I2 who have shown that when human lZSI-labeled IgG was used, 70% of the labeled IgG transported from the maternal into the fetal circuit was fragmented to small peptides. Radiolabeling of protein molecules with ‘9 could potentially cause alterations with denaturation, which may lead to a loss of biologic activity, as was shown for I” I-labeled erythropoietin.“O It may also yield the protein molecule labile so that during the transcellular pathway across the placental barrier degradation could occur. Previously observed breakdown of IgG therefore seems to be an artifact resulting from 9labeling and is of no physiologic relevance. The slow transfer of IgG in the in vitro system, with only 0.5% of the total amount added to the maternal compartment being transferred to the fetal side over a period of 5 hours in a perfusion preparation representing approximately 5% of the total placenta, may not be too different from the in vivo situation. It has been suggested that it may take as long as 3 weeks for exogenously administered IgG to equilibrate between maternal and fetal blood.31 In view of a half-life of 3 to 4 weeks of exogenous IgG in the maternal system, a considerable amount of the dosage given to the mother

September 1995 Am J Obstet Gynecol

will eventually reach the fetus, and maternal administration of intravenous immunoglobulin may be effective for fetal therapy.’ In cases of maternal idiopathic thrombocytopenic purpura an interval of several weeks between administration of immunoglobulin to the mother until delivery must be allowed, to achieve a therapeutic effect in the fetus.* The same applies for active immunization of the mother to provide immunoprotection of the fetus or neonate. We thank Prof. A. Morel1 from the Central Laboratory of the Swiss Red Cross, Berne, for his support and critical comment during this study. REFERENCES 1. Faix RG. Maternal immunization to prevent fetal and neonatal infection. Clin Obstet Gynecol 1991;34:277-87. 2. Anderson P, Insel RA. Prospects for overcoming maturational and genetic barriers to the human antibody response to the capsular polysaccharide of Hemophilus injuenzze type B. Vaccine 1988;6:188-91. 3. Eichhorn MS, Granoff DM, Hahm MH, Quinn A, Shackelford PG. Concentrations of antibodies in paired maternal and infants sera: relationship to IgG subclasses. J Pediatr 1987;111:783-8. 4. Schur PH, Alpert E, Alper C. Gamma G subgroups in human fetal cord and maternal sera. Clin Immunol Immunopathol 1973;2:62-6. 5. Davies SV, Murray JA, Gee H, Giles HM. Transplacental effect of high-dose immunoglobulin in idiopathic thrombocytopenia (ITP) [Letter]. Lancet 1986;1:1098. 6. Pappas C. Placental transfusion of immunoglobulin in immunothrombocytopenic purpura [Letter]. Lancet 1986; 1:389. 7. Sacher RA, King JC. Intravenous gamma-globulin in pregnancy: a review. Obstet Gynecol Surv 1988;44:25-34. 8. Smith CIE, Hammarstrom SL. Intravenous immunoplobulin in pregnancy. Obstet Gynecol 1985;66(suppl):39-40. 9. Schneider H, Paniael M, Dancis 1. Transfer across the perfused human p&en& of antipsine, sodium, and leutine. AM J OBSTET GYNECOL 1972;114:822-8. 10. Jefferis R, Reimer CB, Skvaril F, et al. Evaluation of monoclonal antibodies having specificity for human IgG subclasses: results of an IUISWHO collaborative study. Immunol Lett 1985;10:223-52. 11. Malek A, Hy M, Honegger A, Rose K, Brenner-Holzach 0. Fructose-1,6-bisphosphate aldolase from Drosophila melunogas&: primary structure prediction and comparison with vertebrate aldolase. Arch Biochem Biophys 1988; 266:10-31. 12. Contractor SF, Eaton BM, Stannard PJ. Uptake and fate of exogenous immunoglobulin G in the perfused human placenta. J Reprod Immunol 1983;5:265-73. 13. Schneider H, Mohlen K-H, Dancis J. Transfer of amino acids across the in vitro perfused human placenta. Pediatr Res 1979;13:236-40. 14. Schneider H, Sodha RJ, Progler M, Young MPA. Permeability of the human placenta for hydrophilic substances studied in the isolated dually in vitro perfused lobe. Contrib Gynecol Obstet 1985;13:98-103. 15 Edwards D, Tones ClP, Sibley CP, Nelson DM. Paracellular permeability pathways in the human placenta: a quantitative and morphological study of maternal-fetal transfer of horseradish peroxidase. Placenta 1993;14:63-73. 16. Bajoria R, Contractor SF. Transfer of heparin across the human perfused placental lobule. Pharmacol 2I Pharm 1992;44:952-9. 1 17 Malek A, Sager R, Eckhardt K-U, Bauer C, Schneider H. Lack of transport of erythropoietin across the human

Volume 173, Number Am J Obstet Gynecol

18.

19.

20.

21. 22,

23.

24.

Malek et al.

3, Part 1

placenta as studied by an in vitro perfusion system. Pflugers Archiv 1994;427:157-61. Immunodeficiency: report of a WHO scientific group. Geneva: World Health Organization, 1978; WHO technical report series 630:3-80. Barker G, Cunliffe N, Bardsley WG, D’Souza SW, Donnai P, Boyd RDH. Fetal and maternal blood volumes in shed human placentae: discrepant results comparing morphometry to haemoglobin content. Placenta 1988;9:28996. Sedmak DD, Davis DH, Singh U, van de Winkel JGJ. Expression of IgG Fc receptor antigen in placenta and on endothelial cells on humans. Am J Path01 1991;138:17581. Johansen PM, Brown PJ. Fey receptors in the human placenta. Placenta 1981;2:355-70. Leach L, Eaton BM, Firth JA, Contractor SF. Uptake and intracellular routing of peroxidase-conjugated immunoglobulin-G by the perfused human placenta. Cell Tissue Res 1990;261:383-8. Griffiths GD, Kershaw D, Booth AG. Rabbit peroxidaseantiperoxidase complex (PAP) as a model for the uptake of immunoglobulin G by the human placenta. Histochem J 1985;17:867-81. King BF. Absorption of peroxidase-conjugated immunoglobulin G by human placenta: an in vitro study. Placenta 1982;3:395-406.

25.

26.

27.

28.

29.

30.

31.

Leach L, Eaton BM, Firth JA, Contractor SF. Immunogold localization of endogenous immunoglobulin-G in ultrathin frozen sections of the human placenta. Cell Tissue Res 1989;257:603-7. Lin CT. Immunoelectron microscopic localization of immunoglobulin G in human placenta. J Histochem Cytochem 1980;28:339-46. Gruenberg J, Howell KE. Membrane traffic in endocytosis: insights from cell-free assays. Annu Rev Cell Biol 1989;5: 453-81. Morgan CL, Cannel1 CR, Addison RS, Minchinton RM. The effect of intravenous immunoglobulin on placental transfer of a platelet-specific antibody: Anti-PIflu. Transfus Med 1991;1:209-16. Malek A, Sager R, Schneider H. Maternal-fetal transport of immunoglobulin G and its subclasses during the third trimester of human pregnancy. Am J Reprod Immunol 1994;32:8-14. Coldwasser E. Erythropoietin and red cell differentiation. In: Cunningham D, Goldwasser E, Watson J, Fox DF, eds. Control of cellular division and development. New York: Alan R. Liss, 1981:487-94. Newland AC. Intravenous immunoglobulin therapy in chronic idiopathic thrombocytopenic purpura. In: Morel1 A, Nydegger UE, eds. Clinical use of intravenous gammaglobulins. London: Academic Press, 1986:203-16.

l-SOO-55-MOSBY number links you to the full text of articles published in over 25,000 journals, including all Mosby journals. MOSBY Document Express@, a rapid response information retrieval service, provides quick turnaround, 24-hour availability, and speedy delivery methods. For inquiries and pricing information, call our toll-free 24-hour order line: l-SOO-55MOSBY, outside the United States: 415-259-5046; fax: 415-259-5019; E-mail: [email protected]. This

767