Metabolism
of Very-Low-Density Lipoprotein Triglyceride Cells: The Role of Lipoprotein Lipase Bartolome
Bonet, John
D. Brunzell,
by Human
Placental
Allen M. Gown, and Robert l-4.Knopp
Several studies have shown lipoprotein lipase (LPL) activity in human placenta, but the quantitative significance and cellular specificity of LPL in this organ are unknown. The objective of this report is to investigate the metabolism of very-low-density lipoprotein triglycerides (VLDL-TG) by the placenta, the role of LPL in this process, and the types of cells involved. Placental cells were obtained by enzymatic digestion (collagenase, hyaluronidase, and DNA-ase) and separated on a 40% Percoll gradient. The trophoblasts were the predominant cell type (80% to 85% pure) isolated at d = 1.033 to 1.048 and macrophages were predominant at d = 1.077 to 1.100 (>95% pure), as characterized by eight immunocytochemical assays using cell protein-specific monoclonal antibodies. Macrophages represented 50% to 80% of cells isolated, and trophoblasts, 40% to 50%. LPL activity was assessed by VLDL-TG hydrolysis in primary 3- to 4-day tissue culture. In a representative experiment, LPL activity (nmol fatty acids (FA)/mg protein/24 h) was 101.3 f 5.3 in macrophages and 29.9 f 8.5 in the predominant trophoblast cell types, with approximately 20% of these amounts incorporated and reesterified. VLDL-TG hydrolysis and cell lipid uptake in both placental cell types was essentially abolished by a monoclonal anti-LPL antibody. When compared with a model of hepatocytes (Hep G2 cells), the hydrolysis of VLDL-TG was almost undetectable in these cells. In contrast, free fatty acids (FFA) uptake by Hep G2 cells was fourfold to sixfold greater than that by macrophages and trophoblasts, respectively. In conclusion, macrophages and trophoblasts are the two predominant placental cells isolated by enzymatic digestion. The majority of VLDL-TG hydrolysis activity in tissue culture of placental cells arises from macrophages and is LPL-mediated. Compared with a cell model of hepatocyte function, VLDL-TGFA predominate as a source of FA for placental cells, while FFA predominate for the hepatocyte. Placental macrophages may play a role in the uptake of triglyceride-rich lipoprotein FA for placental storage, oxidation, and transfer to the fetus. Copyright 0 1992 by W.B. Saunders Company
A
MONG THE NUTRIENTS required by the fetus are fatty acids (FA), either for delivery of energy or for meeting essential biological requirements. The FA are available to the fetus either by hydrolysis of dietary chylomicrons or endogenous circulating very-low-density lipoprotein triglycerides (VLDL-TG) or by direct transfer of circulating free fatty acids (FFA) across the placenta. Lipoprotein iipase (LPL) plays a central role in the hydrolysis of VLDL-TG and in the uptake of FA.Q LPL activity is present in animal and human placenta,3-h and studies indicate that circulating TG-rich lipoproteins serve as a source of FA to the fetus.‘,” Studies in guinea pigs indicate a major role of placental LPL in TG hydrolysis and FFA appearance in the fetus4 In this animal, LPL is located in the maternal side of the placenta,4 and its activity increases by the end of gestation when circulating levels of VLDL-TG are higher. To further characterize the function of placental LPL in the human, we have examined the metabolism of TG-rich lipoproteins by cultured human placental cells and the extent to which LPL is responsible. As human placenta is a multicellular organ, we also were interested in determining which types of cells exhibited LPL activity. The results
From the Northwest Lipid Research Clinic and the Departments of Medicine, Obstetrics and Gynecology, and Pathology. University of Washington School of Medicine, Seattle, WA. Supported b Grant No. CCH-R510l061 from the US-Spain Joint Committee for Scientific and Technological Cooperation, Clinical Research Core of the Clinical Nutrition Research Unit at the University of Washington (CNRU), Grant No. DK35816 from the National Institutes of Health Grant No. DK 02456, and a gift from the Ortho Pharmaceutical Company. Address reprint requests to Robert H. Knopp. MD, Northwesf Lipid Research Clinic, 326 Ninth Ave, Seattle, WA 98104. Copyright 0 1992 by W.B. Saunders Company 0026-0495192/4106-OOOS$O3.OOjO 596
of the investigation demonstrate hydrolysis of VLDL-TG almost entirely by LPL in placental cells, with the majority of activity, if not all, arising from macrophages. METHODS Isolation of Placental Cefls Placentas were obtained from women undergoing elective cesarean section with normal pregnancies of 38 to 42 weeks gestation. Placentas were placed on ice immediately after delivery and brought to the laboratory. The process to obtain cells was started no later than 30 minutes after delivery. After removal of associated membranes. small pieces of placenta (2 to 3 cn$) were obtained, avoiding areas of calcification, and rinsed with cold saline several times. Placental tissue was carefully teased from vascular stroma and was coarsely minced, rinsed with Medium 199 (Flow Lab, McLean, NJ), and transferred to a 250.cc flask containing the following enzymatic solution: 0.05% to 0.1% collagenase class II (Worthington Biochemical, Freehold. NJ), O.OS% hyaluronidase type III (Sigma Chemical, St Louis, MO), 4 UimL DNA-ase type I (Sigma), and 1% fetal calf serum (FCS) (Gibco, Grand Island. NY) in HAM’s F-IO culture medium with glutamine (Flow Lab). The mixture was incubated in a shaking bath at 37°C. with the flask lying horizontally to maximize contact between enzymes and placental tissue. for 60 minutes. The supernatant was filtered through three layers of sterile gauze into SO-cc tubes containing 3 mL FCS, and centrifuged at 1,000 x g for 7 minutes. The pellets were washed twice with Medium 199 (Flow Lab) and pooled. This cellular suspension containing 30 to 80 x 10’ cells was then layered on a tube containing 30 mL of a 40% Percoil gradienty (Pharmacia. Piscataway, NJ). The cell suspension placed on the gradient was centrifuged at 2,500 x g for 15 minutes. Simultaneously, another gradient tube loaded with density markel beads (Pharmacia) was centrifuged to determine the density at which cells were localized. Three bands of cells were isolated with this method at d = 1.033 to I.048 (middle band). d = I.077 to I. 100 (bottom band). and d > I.140 the last containing red blood cells. When individual bands were isolated and reloaded onto Percoll gradients, cells always returned to their original density and no new Metabolism,
Vol 41, NO 6 (June),
1992: pp 596.603
PLACENTAL
TRIGLYCERIDE
METABOLISM
bands were observed. Cells from the different bands were washed twice with Medium 199 and counted with a hemocytometer. The yield of cells was 15 to 40 x 106 in the two non-red blood cell bands. It was difficult to assess total placental tissue weight because of the blood and fibrous tissue, and to weigh the individual pieces would increase the risk of contamination. It was not possible to express yield per unit of placental weight. The viability of the cells in each placenta was greater than 95% as assessed by Trypan Blue exclusion. Placental Cell Culture and Characterization Placental cells from the two non-red cell bands were suspended in HAM’s F-10 culture medium containing 10% serum obtained from the blood of the same placenta donor, 100 pU/mL penicillin, and 100 ug/mL streptomycin, with a final concentration of 1 X lo6 cells/ml. One milliliter of this cell solution was incubated in each well of a 24-well culture plate (Falcon) (Becton-Dickinson Labware, Lincoln Park, NJ) with 95% air and 5% CO2 at 37°C. After 48 hours, the cells were attached to the plates, the medium was changed, and experimental protocols were followed as described below. To identify the cells from both bands, two approaches were followed: In the first, 10 )LL of the cell solution was placed onto 12-well. teflon-coated, microscope slides (formerly Meloy, now Erie Scientific, Portsmouth, NH). After 48 hours in the incubator in the same conditions as described above, the slides were fixed in cold methanol and overlaid with the following panel of monoclonal and polyclonal antibodies: 35PHl1, a monoclonal antibody to cytokeratin no. 8, typically found in glandular epithelium, but also present in trophoblasts i”,it; 34BE12, a monoclonal antibody to high-molecular weight cytokeratins, eg, found in ductal and squamous epithelium i”,ii; 43PE8, a monoclonal antibody to vimentin, the intermediate filament expressed by mesenchymal cells”; HHF35 and CGA7, two monoclonal antibodies specific for muscle actinsi2,13;HAM-56, a monoclonal antibody specific for a cytoplasmic marker of monocytes/macrophagesi4; monoclonal and polyclonal antibodies to factor VIII-related antigen (Dako, Santa Barbara, CA) and endothelial-specific proteini5; and polyclonal antibodies to human chorionic gonadotropin (hCG) and human placental lactogen (hPL), both obtained from Dako. Localization was via an indirect immunofluorescence method, using fluoresceinconjugated anti-mouse or anti-rabbit antibodies (Dako). In the second approach, cells from the middle or bottom band were kept in tissue culture plates for 3 to 4 days, at which time the cells were scraped from the dish, using a rubber policeman. The resulting cell suspension was centrifuged at 1,000 x g, and the pellet thereby obtained was carefully removed and immediately fixed in methacarn fixative (60% methanol, 30% chloroform, and 10% glacial acetic acid) and then fixed overnight. Cell pelletparaffin blocks were obtained by standard processing and embedding procedures. Serial sections of the cell blocks were then cut, deparaffinized, and rehydrated through graded alcohols. The same pane1 of antibodies was then applied, with antibody localization performed by a slight modification of the standard avidin-biotin immunoperoxidase procedure with nickel chloride enhancement of the 3,3’-diaminobenzidine chromogen,t6 yielding a black reaction product. To determine the cell composition of placental tissue, without the possible artifacts induced by the enzymatic digestion, blocks of placental cotyledons were obtained and processed as described above for cell pellets. Serial sections of the tissue blocks were submitted to the same panel of antibodies described above. Antibody localization was via an indirect immunofluorescent method using fluorescein- or rhodamine-conjugated anti-mouse or anti-rabbit antibodies (Dako).
597
An established cell line from human liver-tumor biopsies, designated Hep G2, has been shown to have morphological and metabolic characteristics similar to those of liver parenchymal cellsi7-i9; this cell line was used as a negative control for VLDL-TG hydrolysis, since hepatocytes usually do not have LPL activity. Hep G2 cells were grown in the same medium as the macrophages and trophoblasts, but it contained 10% FCS instead of autologous serum. Labeling WDL-TG Blood from healthy human donors who had fasted for 12 hours was collected in EDTA, and plasma was separated from cells at 2,500 x g for 20 minutes. The different lipoproteins used were isolated by sequential ultracentrifugation.20 Triolein-3H in hexane (Amersham-Searle, Chicago, IL), 0.50 mCi, was dried under NZ dissolved in 200 FL ethanol, transferred to a high-density lipoprotein (HDL) fraction, and incubated in a shaking bath at 37°C for 4 hours. The VLDL solution was added in a 5:l ratio of HDL cholesterol to VLDL cholesterol. Approximately 5 mg partially purified lipid-transfer protein (LTP-I)*’ (generously supplied by Dr J Albers and J Tollefson) was added to the medium and incubated at 37°C overnight. This solution was centrifuged at 40,000 x g for 24 hours. The top fraction, containing VLDL and more than 95% of the Triolein3H, was collected and overlaid on saline and centrifuged again for 24 hours. With this method, we consistently obtained an incorporation of Triolein-‘H exceeding 80% of the amount of radioactivity originally incubated with HDL. Oleic Acid-Human
Serum Albumin Solution
Ten microcuries of oleic acid in hexane labeled on its carboxyl group with i4C (Amersham-Searle) was mixed with 1 mL of an 0.7 mmol/L oleic acid solution, dried under Nr, and redissolved in 1 mL physiological saline and 0.02N NaOH, and then incubated in a shaking bath at 37°C for 1 hour. Once it was completely dissolved, 1 mL of a solution containing 1.4 mmol/L human serum albumin (HSA) essentially free of FA (Sigma) was added for a FFA to HSA ratio of 4:l. This solution of oleic acid-HSA (FA-HSA) was used in the tissue culture FA uptake experiments. Eqerimental
Conditions in Cell Culture
Eighteen hours before starting the experiment, the tissue culture medium was changed with fresh medium containing 2 U heparin/cc to stabilize LPL. When monoclonal anti-LPL antibodies were used, they were added to the medium at the start of the experiment. The monoclonal antibody, 5D2, developed against bovine milk LPL, inhibits human LPL activity by 99% and has no effect on hepatic lipase activity.** Cells were incubated with VLDL at a concentration of approximately 200,000 cpmiwell (50 kmol triolein/well) and the medium was sampled at 0, 4, and 24 hours. FFA from the medium were extracted with the method described by Belfrage and Vaughan.23 Cell pellets were washed 5 times with cold saline and then submitted to a lipid extraction with 1 mL hexane-isopropyl alcohol 3:2 (vol:vol).24 Two hundred microliters of the lipid extract was transferred to a scintillation vial to determine the amount of radioactivity incorporated into the cells. The remaining solution was dried under Nz and redissolved in 100 PL chloroform. Thin-layer chromatography (TLC) was used to isolate the lipid components.25 The protein concentration of the cells was determined by a modification of the method of Lowry et al.*” Simultaneously, some wells were incubated with FA-HSA at approximately 150,000 cpm/well, and at 0,4, and 24 hours, lipid extraction, TLC, and protein determination were performed as described above. Results are expressed as nmol/mg cellular protein based on
BONET ET AL
598
the radioactivity of the extract and the specific activity of the VLDL-TGFA3H or FFA-14C added to the medium. Statistical testing was performed using Student’s paired t test.27 RESULTS
Identification of Placental Cells
Using the tissue-digestion and cell-isolation procedures previously described, 30 to 80 x lo6 cells were typically isolated from the placentas of more than 15 different, normal, pregnant subjects, with 15 to 30 x lo6 cells from the middle band and 20 to 40 x lo6 cells from the bottom band. Under the light microscope, the cells of the two bands were distinctly different. The middle band contained cells that were heterogenous in both size and shape. The bottom band typically consisted of a homogenous population of small, round cells and a very small number of large, round cells, never more than 1% to 2% of the total number. After 2 days in culture, the cells from the middle band formed a syncytium (Fig 1A) and the cells from the bottom band remained as individual cells (Fig 1B). Both populations of cells were attached to the plates. Examination of the cells by immunofluorescence or immunoperoxidase procedures, using the monoclonal- and polyclonal-antibody panel (Figs 1 and 2), revealed the
majority of the large cells in the middle band to be positive with both anticytokeratin antibodies, 35PEll (Fig 2A) and 34BH12, and antivimentin antibody 43pE8 and the antibody to hPL (Fig lA), all findings characteristic of trophoblasts. A subpopulation of smaller cells was also positive with the anticytokeratin antibody 35BH11, and the antivimentin antibody 43PE8, and some were also positive with the anti-hPL antibody, or the anti-macrophage antibody HAM-56, indicating that the small cells from the middle band were a heterogenous population mainly comprised of trophoblasts and macrophages (Fig 2B). No cells in the middle band were positive with antibodies to hCG. Cells from the bottom band were comprised of two morphologically distinct cell populations: a very small number of large cells was positive with both anticytokeratin antibodies, 35PHll (Fig 2C) and 34PE12, the antivimentin antibody 43PE8, and the antibody to the trophoblast marker-protein hPL, identifying these cells as trophoblasts. The small cells comprising the great majority of cells in the bottom band were negative with the anticytokeratin antibodies, and positive only with the antivimentin antibody 43PE8, and the antimacrophage antibody HAM-56 (Fig 2D). None of the small cells in the bottom band expressed the trophoblastic marker hPL or anticytokeratin antibody
Fig 1. lmmunocytochemistry of cells from the middle end bottom bends placed onto microscope slides, after 48 hours of incubation. (A) Cells from the middle band overlaid with polyclonal antibodies to hPL. (9) Cells from the bottom band overlaid witf~ the monoclonal antibody HAMQB, a macrophage-specific antibody.
PLACENTAL TRIGLYCERIDE METABOLISM
Fig 2. lmmunocytochemistry of pellets obtained from cells of the middle and bottom bands after 3 days kept in tissue culture. (A) and (6) correspond to the middle band. (A) Overlaid with the monoclonal antibody 35BHll (which stains trophoblasts). (6) Overlaid with the monoclonal antibody HAM-56 (which stains macrophages]. (A) Reveals that most of the cells contained in the middle band are mainly trophoblasts, although macrophages are also present in the middle band. (C) and (D) correspond to the bottom band. (Cl Overlaid with the monoclonal antibody 356Hll. (D) Overlaid with the monoclonal antibody HAM-56. (B) Cells contained in the bottom band are mostly macrophages with occasional large trophoblasts.
35PHll. No cells in either band reacted positively for endothelial or smooth muscle cell markers. Thus, the majority of cells in the middle band were trophoblast in nature and formed a syncytiotrophoblast within 24 to 48 hours of incubation, although macrophages were present in this band. On the other hand, the large majority of cells in the bottom band were macrophages, with only a few cells showing characteristics of trophoblasts. When slides of placental tissue were subjected to a panel of monoclonal and polyclonal antibodies specific for different cell markers, the two predominant cell types were again macrophages and trophoblasts (Fig 3). The trophoblasts were localized on the periphery of the placental villi (Fig 3A), and macrophages were the predominant parenchyma1 cell (Fig 3B). Blood vessels reacted positively to factor VIII-related antigen and to HHF35 and C6A7, showing the presence of endothelial and smooth muscle cells (results not shown). None of these cells were obtained after enzymatic digestion. To further characterize the differences between middleand bottom-band cells, progesterone secretion by cells from each band was determined in tissue culture (Table 1). The data from one representative experiment reveal the middle band to secrete the largest amount of progesterone, with very little being secreted by cells of the bottom band. This observation confirms the different nature of cells from
middle and bottom bands, and supports the view that the great majority of cells of the middle band are steroid hormone-secreting trophoblasts, while the bottom-band cells are not. LPL Activity in Human Placental Cells
To characterize the metabolism of VLDL-TG by placental cells and the role of LPL, VLDL-TG labeled with 3H-oleic acid were incubated with cells fror : the middle band (trophoblast) or bottom band (macrophages) and with Hep G2 cells, a model of hepatocyte function as a negative control for LPL activity. VLDL-TGFA were also incubated in the absence of cells to detect any intrinsic VLDL-TG hydrolysis during the time of incubation (0, 4, and 24 hours). The amount of FFA generated in the absence of cells was small, but it was subtracted from the amount of FFA generated in the incubation wells containing cells, yielding cell-mediated FFA generated by VLDL-TG hydrolysis. The VLDL-TGFA were released as FFA into the medium and were incorporated into the cells at 0, 4, and 24 hours (a representative experiment is shown in Table 2). Both the appearance of labeled FFA in the medium and the uptake of labeled FA by the cells were consistently higher in macrophages than in the predominant trophoblast cell types (Table 2). In the experiment shown, labeled FFA in
BONET ET AL
Fig 3. lmmunocytochemistry of placental tissue. (A) Histological section of placenta villi overlaid with monoclonal antibodies to 356Hll. a marker of trophoblasts. (B) Slides of placental villi overlaid with monoclonal antibody to HAM-56, a marker of macrophages. This figure shows that the two main populations of term placental cells are macrophages, localized in interstitial tissue and trophoblasts, localized in the periphery.
the medium (nmol/mg protein) was 101.3 +- 5.3 in macrophages versus 29.9 ? 6.4 in the trophoblast-rich band after 24 hours. The amount of label incorporated into cell lipid extract was a fairly constant proportion of the amount of FFA in the medium in both cell populations, and was approximately 20% of the amount of labeled FFA in the medium at 24 hours (Table 2). Most of the radioactivity in the cellular lipid extract was in the form of triolein. with very little as oleic acid (Table 2). These effects were replicated in all experiments from five different placentas. In both populations of placental cells, monoclonal antibodies against LPL that were added to the medium before incubation markedly decreased the appearance of labeled FFA in the medium, in most cases to below the limit of detection (Table 2). The uptake of labeled FA by the cells was similarly abolished. These effects were replicated in four separate experiments. When monoclonal antibodies Table 1. Progesterone Secretion in Placental Cells (ng/mL) Middle Band
Bottom Band
29.2 f 1.05
3.13 -t 0.24
NOTE. Cells from concentration Progesterone
the
middle
of 0.5 to 1 x concentration
band
or bottom
lo6 cells/ml.
band
Average
tested after 48 hours.
were
of four
at a wells.
against apolipoprotein (apo) A-I were added in the same way as the LPL antibodies, to serve as a control, neither the release of labeled FFA nor the uptake of labeled FA by the cells were affected (data not presented). Thus, VLDL-TG hydrolysis by cells is mediated by LPL. Uptake of FA-HSA by Placental Cells A comparison of the uptake and metabolism of FFA by both types of placental cells was undertaken. FA-HSA was added to the medium for comparison to FFA appearing in the medium from VLDL-TG hydrolysis in the previous experiments for similar incubation times (0, 4, and 24 hours). The incorporation of label was similar in macrophages and trophoblasts in four such experiments (representative experiment, Table 3) but was fourfold to sixfold higher in Hep G2 cells. Most of the intracellular radioactivity in macrophages and trophoblasts was found in FFA at 4 hours, but at 24 hours, most of the radioactivity was in cellular TGFA. These results indicate an ongoing esterification process following FA uptake. In contrast, esterified TGFA was the predominant form of radiolabeled FA in the Hep G2 cells at both 4 and 24 hours, indicating a more rapid esterification rate than in macrophages or trophoblasts. In the Hep G2 cells, phospholipids also comprised
PLACENTAL
TRIGLYCERIDE
METABOLISM
601
Table 2. Metabolism of VLDL-TG by Placental-Cell Bottom Band (Largely Macrophages) and Middle Band (Largely Trophoblasts)
in Tissue
Culture BottomBand (Hours)
VLDL-TGFA (Appearingas:)
mAb
FFA in medium
0
(-)
2.22 (0.97)X
(+I
24
33.5
101.3
(4.40)
1.39
Middle Band(Hours)
4
(5.30) 0
0
Hep G2 Cells(Hours)
0
4
24
0
0.77
4.26
29.9
0.22
4
24
0.42
1.96
(0.3) 0
(3.50) 0
(6.37) 0
(0.12) -
(0.36) -
(0.67) -
(0.70) Cell total lipid
f-1
0
1.4
26.3
(+)
0
0
0.26
5.40
0.38
2.76
6.15
(0.65) 0
(0.18) -
(1.02) -
(1.60)
0
(0.70) 0
(2.55)
(0.35)
0.41
0
(0.14) Cell total TGFA
0
(-)
0
(+)
0
21.2
0.79
(6.70)
(0.40)
0.70
0
0
0.25
4.73
(0.09)
(0.80)
0
0
0
1.13
1.48
(0.27) -
(0.31) -
(0.06)
0
l-1
Cell FFA
0.29
0
(+)
0
2.70 (0.24)
(0.07)
0.12
0
0
0.15
0.54
(0.05)
(0.05)
0
0
0
0.56
0.50
(0.11) -
(0.16) -
(0.04) NOTE. Results are presented as means in nmol/mg
cellular protein and values in parentheses are 2 SE. Values represent 2 to 4 wells/time.
Experiments with different placentas show similar results. Abbreviations:
+/-
mAb, incubation of cells with or without monoclonal anti-LPL antibody 502 added to the medium. Monoclonal antibody was
not tested in Hep G2 cells.
connection, the abundance of macrophages in the term placenta is noteworthy, and our studies confirm previous observations32-34 that macrophages comprise most of the interstitial cells of the human placental villi and include about 50% of isolated placental parenchymal cells. The amount of LPL activity found in the trophoblast-rich population was always lower than that in macrophages, approximating a ratio of 1:4. These data raise the possibility that most of the LPL activity found in the trophoblast-rich population is secondary to the macrophages present, with trophoblast celk that consistently comprise 10% to 20% of the cells present in the middle band (Fig 2B). However, we cannot completely exclude the possibility that some LPL activity resides in the trophoblasts. Despite numerous attempts, we were unable to obtain a sample of trophoblasts completely free of macrophages. Even doing a double-percoll isolation, we found macrophages contaminating the middle band. The high tendency of trophoblasts to aggregate after enzymatic digestion may actually entrap macrophages. Alternatively, LPL could originate in macrophages and migrate to trophoblasts, similar to the way LPL is synthesized in parenchymal cells and migrates to the apical surface of endothelial cells.34.35
about 40% of the intracellular radioactivity, in contrast to 10% in macrophages and trophoblasts (Table 3). The greater avidity of Hep G2 cells for FFA and the fourfold to sixfold greater esterification to cellular TG compared with placental macrophages and trophoblasts (Table 3) contrasts with the greater metabolism of medium VLDLTGFA by macrophages and trophoblasts than by Hep G2 cells. DISCUSSION
Our experiments show that macrophages and trophoblasts are the two main parenchymal components of human placenta. Macrophages isolated from human placenta possess the majority of TG-hydrolytic activity compared with trophoblasts, and this activity is inhibited more than 99% when both cell types are incubated with a specific monoclonal anti-LPL antibody that does not inhibit hepatic 1ipase.22 The finding of high hydrolytic activity per milligram protein in the macrophages is consistent with the findings of other investigators, that human blood monocyte-derived macrophages and macrophages from atheromatous lesions have been shown to contain LPL activity.29-3* In this Table 3. Uptake of Human Albumin-Bound
FFA W by Placental-Cell Bottom Band (Largely Macrophages) and Middle Band (Largely Trophoblasts)
in Tissue Culture
Bottom Band (Hours)
Cell FFA (Uptake As)
4
Hep G2 Cells (Hours)
Middle Band(Hours) 24
6.14 5 0.97
4
0.62 + 0.2
24
4
24
4.45 f 2.24
11.8 ‘- 1.48
25.1 t 3.54
Total lipids
1.56 2 0.32
TGFA
0.16 + 0.001
3.84 -+ 0.31
0.18 2 0.03
2.01 k 0.75
6.48 r 0.77
14.8 + 0.76
FFA
1.01 k 0.22
0.26 2 0.09
0.21 2 0.12
0.117 + 0.01
0.45 k 0.04
1.01 -+ 0.12
Phospholipid
0.27 -c 0.03
0.66 2 0.13
0.14 2 0.01
0.53 ? 0.3
3.76 2 0.03
9.53 f 0.43
NOTE. Two nanomoles oleic acid 1% was added per well. Two to six wells were averaged/time are presented as the mean 2 SE.
point. Results are in nmol/mg cellular protein and
BONET ET AL
602
The amount of radioactivity found in placental ceil lipid extracts as a consequence of LPL activity was always proportional (20% to 25%) to the amount of FA released to the medium from the lipoproteins. Most of the radioactivity found in the cell lipid extract was in the form of TG, and not as free oleic acid. In addition, when macrophages and trophoblasts were incubated with FA-HSA, cell lipid extracts also showed that more than 90% of the intracellular radioactivity was present as TG, underscoring the dominance of the esterification pathway. This observation is consistent with the results found by other authors showing a high esterification rate of FFA to TG in human placental cells.36-3sWhen placental cells were incubated with monoclonal anti-LPL antibody, they did not acquire any radioactivity, indicating that the uptake of VLDL-TGFA is in the form of FFA, which, after becoming intracellular, are reesterified to TG. The current study therefore demonstrates that the placental cells can take up FA from VLDL-TG, and this process is mediated by prior hydrolysis of VLDL-TG by LPL and not by direct uptake of intact TG. The pattern uptake of FFA versus VLDL-TG-derived FA by placental macrophages and trophoblasts was different from that of Hep G2 cells. When the two types of placental cells were incubated with FA-HSA, the uptake of FFA after 24 hours of incubation was 4% to 6% of the initial amount contained in the incubation medium, in contrast to greater than 30% uptake of FFA by Hep G2 cells (Table 3). In contrast, when the three cell types were incubated with labeled VLDL-TGFA, cellular TGFA content was 15-fold greater in macrophages than in Hep G2 cells, and threefold greater in trophoblast cells compared with Hep G2 cells. These differences indicate that VLDL-TG is a more important source of intracellular FA than are circulating FFA for placental cells, while FFA are relatively more important to the Hep G2 cell. This comparison of FFA versus TGFA uptake underscores the importance of placental LPL activity in placental function, and suggests that circulating TGFA may be a more important source of placental FA than are albumin-bound FFA. The relative importance of TGFA over FFA as a source of intracellular placental-cell FA is in agreement with the observation of Thomas et al4 in a pregnant guinea pig placental perfusion model. In those studies, the concentration of maternal VLDL-TG-derived FFA was twofold
higher in fetal placental perfusate than in maternal plasma. In contrast, when labeled FFA were given to the mother, the fetal placental-perfusate concentration of FFA was 80% of that in maternal plasma. Both our studies and the studies of Thomas et al4 support the hypothesis that TG-rich lipoprotein may be a more efficient means to deliver FA to the placenta and fetus than are FFA. This mechanism could favor the delivery of essential FA and energy acquired from the diet via chylomicrons to the fetus if these alimentary particles are metabolized by the placenta in the same manner as endogenous VLDL. In vivo support for this finding is suggested by our recent observation that maternal TG concentrations are a predictor of infant birth weight39 and that placental LPL activity is a predictor of fetal guinea pig weight.4 It is important to place these findings of placental VLDL-TG metabolism in the context of other forms of lipid access to the placenta. Specifically, cholesterol has ready access to the placenta via LDL and the LDL receptor and via HDLz and a non-receptor-mediated mechanism.q This subject has been recently reviewed.41 In conclusion, the results presented in this report show that placental macrophages are an abundant cell type in isolates of placental cells. Placental macrophages have the capability to hydrolyze VLDL-TG, and this process is essentially entirely LPL dependent. Trophoblast cells show less activity, which may be due to contaminating macrophages. Regardless of source, these results show that LPL activity in placental cells has an essential role in the uptake of TGFA of TG-rich lipoproteins, and possibly, in their transfer to the fetus. Compared with Hep G2 cells, placental macrophages and trophoblasts are more avid in their uptake of VLDL-TGFA than FFA, suggesting that TGFA are the predominant FA source for the placenta. If macrophages are the main source of LPL activity, it is noteworthy that they are not in direct contact with maternal blood. The possible transfer of LPL from macrophages to trophoblasts, cells that are in direct contact with blood, and therefore, with lipoproteins, is currently being investigated. ACKNOWLEDGMENT
The authors are grateful to Carol Dorsett for technical tance, to Alan Greeves for typing. and to John Tollefson J. Albers for a gift of human transfer protein.
assisand John
REFERENCES 1. Olivecrona T, Bengtsson-Olivecrona G: Lipoprotein lipase and hepatic lipase. Curr Opin Lipid 1:222-230, 1990 2. Eckel RH: Lipoprotein lipase: A multifunctional enzyme relevant to common metabolic diseases. N Engl J Med 320:10601068,1989 3. Knopp RH, Warth MR, Charles D, et al: Lipoprotein metabolism in pregnancy, fat transport to the fetus, and the effects of diabetes. Biol Neonate 50:297-317,1986 4. Thomas CR, Lowy C, St Hillaire RJ, et al: Studies on the placental hydrolysis and transfer of lipids to the fetal guinea pig, in Miller RK, Thiede HA (eds): Fetal Nutrition, Metabolism, and Immunology: The Role of the Placenta. New York, NY, Plenum, 1984, pp 135-146 5. Elphick MC, Hull D: Rabbit placental clearing-factor lipase
and transfer to the fetus of fatty acids derived from triglycerides injected into the mother. J Physio1273:475-487, 1977 6. Mallov S. Alousi AD: Lipoprotein lipase activity of rat and human placenta. Proc Sot Exp Biol Med 119:301-306.1965 7. Hummel L, Schwartze A, Schirrmeister W, et al: Maternal plasma triglycerides as a source of fetal fatty acids. Acta Biol Med Germ 35:1635-1641,1976 8. Elphick MC, Filshie GM, Hull D: The passage of fat emulsion across the human placenta. Br J Obstet Gynaecol85:610-618.1978 9. Zeitler P, Markoff E, Handwerger S: Characterization of the synthesis and release of human placental lactogen and human chorionic gonadotropin by an enriched population of dispersed placental cells. J Clin Endocrinol Metab 57:812-818,1983 10. Gown AM, Vogel AM: Monoclonal antibodies to intermedi-
PLACENTAL TRIGLYCERIDE METABOLISM
ate sized filament proteins: Unique and cross-reacting antibodies. J Cell Biol95:414-424,1982 11. Gown AM, Vogel AM: Monoclonal antibodies to intermediate filament proteins of human cells II. Distribution of filament proteins in normal human tissue. Am J Patho! 114:309-321,1984 12. Gown AM, Vogel AM, Gordon D, et al: Smooth musclespecific monoclonal antibodies: Specific recognition of smooth muscle actin isotypes. J Cell Biol 100:808-813, 1985 13. Tsukda T, Tippens D, Mar H, et al: HHF35, a muscle-actinspecific monoclonal antibody. I. Biochemical and immunocytochemical characterization. Am J Patho! 126:51-60,1987 14. Gown AM, Tsukada T, Ross R: Human atherosclerosis II. Immunocytochemical analysis of the cellular composition of human atherosclerotic lesions. Am J Patho! 125:191-207,1986 15. Jaffe EA: Endothelial cells and the biology of factor VIII. N Engl J Med 296:377-383,1977 16. Gown AM, Vogel AM: Monoclonal antibodies to intermediate filament proteins of human cells: II. Distribution of filament proteins in normal human tissue. Am J Patho! 114:309-321,1984 17. Knowles BB, Howe CC, Aden PP: Human hepatocellular carcinoma cell lines secrete the major plasma proteins and hepatitis surface antigen. Science 206:497-499, 1980 18. Zannis VI, Breslow JL, San Giacomo TR, et al: Characterization of the major apolipoproteins secreted by two human hepatoma cell lines. Biochemistry 20:7089-7096,198l 19. Erickson SK, Fielding PE: Parameters of cholesterol metabolism in the human hepatoma cell line, Hep G2. J Lipid Res 27:875-883, 1986 20. Have! RJ, Eder H, Bragdon J: The distribution and chemical composition of ultracentrifugally reported lipoproteins in human serum. J Clin Invest 34:1345-1353,1955 21. Albers JJ, Tollefson JH, Chen CH, et al: Isolation and characterization of human plasma lipid transfer proteins. Arteriosclerosis 4:49-58,1984 22. Babirak SP, Iverius PH, Fujimoto WY, et al: Detection and characterization of the heterozygote state for lipoprotein lipase deficiency. Arteriosclerosis 9:326-334,1989 23. Belfrage P, Vaughan M: Simple liquid-partition system for isolation of labelled oleic acid from mixtures with glycerides. J Lipid Res 10:341-344,1969 24. Hara A, Radin NS: Lipid extraction of tissues with a low toxicity solvent. Anal Biochem 90:420-426,1979 25. Skipski VP, Barclay M: Thin-layer chromatography of lipids. Methods Enzymol14:530-598,1969
603
26. Lowry OH, Rosebrough NJ, Farr AL, et al: Protein measurement with the Folin phenol reagent. J Biol Chem 193:265-275,195l 27. Snedecor GW, Cochran WG: Statistical Methods (ed 6). Ames, IA, Iowa State University Press, 1967, pp 59 28. Chait A, Iverius PH, Brunzell JD: Lipoprotein lipase secretion by human monocyte derived macrophages. J Clin Invest 69:490-493,1982 29, Stray N, Letnes H, Blomhoff JP: Synthesis and secretion of lipoprotein lipase by human monocyte derived macrophages. Stand J Gastroenterol20:67-72,1985 (suppl107) 30. O’Brien K, Gordon D, Deeb S, et al: Localization of lipoprotein lipase mRNA to foam cell-rich regions of human atherosclerotic plaques. Clin Res 39:335A, 1991 (abstr) 31. Wood GW: Mononuclear phagocytes in the human placenta. Placenta 1:113-123,198O 32. Flynn A, Finke JH, Hilfiker ML: Placental mononuclear macrophages as a source of interleukin-1. Science 218:475-477, 1982 33. Castelluci M, Kaufmann P: Hotbauer cells, in Benirschke K, Kaufmann P (eds): The Pathology of the Human Placenta. New York, NY, Springer-Verlag, 1990, pp 71-80 34. Nilsson-Ehle P, Garfinkel AS, Schotz MC: Lipolytic enzymes and plasma lipoprotein metabolism. Annu Rev Biochem 49:667-693, 1980 35. Blanchette-Mackie EJ, Masuno H, Dwyer NK, et al: Lipoprotein lipase in myocytes and capillary endothelium of heart: Immunocytochemical study. Am J Physiol256:E818-E828,1989 36. Szabo AJ, DeLellis R, Grimaldi RD: Triglyceride synthesis by the human placenta. Am J Obstet Gynecol115:257-266,1973 37. Coleman RA: Placental metabolism and transport of lipid. Fed Proc 45:2519-2523,1986 38. Coleman RA, Haynes EB: Microsomal and lysosomal enzymes of triacylglycerol metabolism in rat placenta. Biochem J 217:391-397,1984 39. Knopp RH, Magee MS, Larson MP, et al: Alternative screening tests and birth weight associations in pregnant women with abnormal glucose screening. Diabetes 37:llOA, 1988 (abstr) 40. Lasuncibn MA, Bonet B, Knopp RI-L Mechanism of the HDLr stimulation of progesterone secretion in cultured placental trophoblast. J Lipid Res 32:1073-1087,199l 41. Knopp RH, Magee MS, Bonet B, et al: Lipid metabolism in pregnancy and lactation and effects on fetal growth and development, in Cowett RM (ed): Principles of Perinatal-Neonatal Metabolism. New York, NY, Springer-Verlag, 1991, pp 177-203
of Very-Low-Density Lipoprotein Triglyceride Cells: The Role of Lipoprotein Lipase Bartolome
Bonet, John
D. Brunzell,
by Human
Placental
Allen M. Gown, and Robert l-4.Knopp
Several studies have shown lipoprotein lipase (LPL) activity in human placenta, but the quantitative significance and cellular specificity of LPL in this organ are unknown. The objective of this report is to investigate the metabolism of very-low-density lipoprotein triglycerides (VLDL-TG) by the placenta, the role of LPL in this process, and the types of cells involved. Placental cells were obtained by enzymatic digestion (collagenase, hyaluronidase, and DNA-ase) and separated on a 40% Percoll gradient. The trophoblasts were the predominant cell type (80% to 85% pure) isolated at d = 1.033 to 1.048 and macrophages were predominant at d = 1.077 to 1.100 (>95% pure), as characterized by eight immunocytochemical assays using cell protein-specific monoclonal antibodies. Macrophages represented 50% to 80% of cells isolated, and trophoblasts, 40% to 50%. LPL activity was assessed by VLDL-TG hydrolysis in primary 3- to 4-day tissue culture. In a representative experiment, LPL activity (nmol fatty acids (FA)/mg protein/24 h) was 101.3 f 5.3 in macrophages and 29.9 f 8.5 in the predominant trophoblast cell types, with approximately 20% of these amounts incorporated and reesterified. VLDL-TG hydrolysis and cell lipid uptake in both placental cell types was essentially abolished by a monoclonal anti-LPL antibody. When compared with a model of hepatocytes (Hep G2 cells), the hydrolysis of VLDL-TG was almost undetectable in these cells. In contrast, free fatty acids (FFA) uptake by Hep G2 cells was fourfold to sixfold greater than that by macrophages and trophoblasts, respectively. In conclusion, macrophages and trophoblasts are the two predominant placental cells isolated by enzymatic digestion. The majority of VLDL-TG hydrolysis activity in tissue culture of placental cells arises from macrophages and is LPL-mediated. Compared with a cell model of hepatocyte function, VLDL-TGFA predominate as a source of FA for placental cells, while FFA predominate for the hepatocyte. Placental macrophages may play a role in the uptake of triglyceride-rich lipoprotein FA for placental storage, oxidation, and transfer to the fetus. Copyright 0 1992 by W.B. Saunders Company
A
MONG THE NUTRIENTS required by the fetus are fatty acids (FA), either for delivery of energy or for meeting essential biological requirements. The FA are available to the fetus either by hydrolysis of dietary chylomicrons or endogenous circulating very-low-density lipoprotein triglycerides (VLDL-TG) or by direct transfer of circulating free fatty acids (FFA) across the placenta. Lipoprotein iipase (LPL) plays a central role in the hydrolysis of VLDL-TG and in the uptake of FA.Q LPL activity is present in animal and human placenta,3-h and studies indicate that circulating TG-rich lipoproteins serve as a source of FA to the fetus.‘,” Studies in guinea pigs indicate a major role of placental LPL in TG hydrolysis and FFA appearance in the fetus4 In this animal, LPL is located in the maternal side of the placenta,4 and its activity increases by the end of gestation when circulating levels of VLDL-TG are higher. To further characterize the function of placental LPL in the human, we have examined the metabolism of TG-rich lipoproteins by cultured human placental cells and the extent to which LPL is responsible. As human placenta is a multicellular organ, we also were interested in determining which types of cells exhibited LPL activity. The results
From the Northwest Lipid Research Clinic and the Departments of Medicine, Obstetrics and Gynecology, and Pathology. University of Washington School of Medicine, Seattle, WA. Supported b Grant No. CCH-R510l061 from the US-Spain Joint Committee for Scientific and Technological Cooperation, Clinical Research Core of the Clinical Nutrition Research Unit at the University of Washington (CNRU), Grant No. DK35816 from the National Institutes of Health Grant No. DK 02456, and a gift from the Ortho Pharmaceutical Company. Address reprint requests to Robert H. Knopp. MD, Northwesf Lipid Research Clinic, 326 Ninth Ave, Seattle, WA 98104. Copyright 0 1992 by W.B. Saunders Company 0026-0495192/4106-OOOS$O3.OOjO 596
of the investigation demonstrate hydrolysis of VLDL-TG almost entirely by LPL in placental cells, with the majority of activity, if not all, arising from macrophages. METHODS Isolation of Placental Cefls Placentas were obtained from women undergoing elective cesarean section with normal pregnancies of 38 to 42 weeks gestation. Placentas were placed on ice immediately after delivery and brought to the laboratory. The process to obtain cells was started no later than 30 minutes after delivery. After removal of associated membranes. small pieces of placenta (2 to 3 cn$) were obtained, avoiding areas of calcification, and rinsed with cold saline several times. Placental tissue was carefully teased from vascular stroma and was coarsely minced, rinsed with Medium 199 (Flow Lab, McLean, NJ), and transferred to a 250.cc flask containing the following enzymatic solution: 0.05% to 0.1% collagenase class II (Worthington Biochemical, Freehold. NJ), O.OS% hyaluronidase type III (Sigma Chemical, St Louis, MO), 4 UimL DNA-ase type I (Sigma), and 1% fetal calf serum (FCS) (Gibco, Grand Island. NY) in HAM’s F-IO culture medium with glutamine (Flow Lab). The mixture was incubated in a shaking bath at 37°C. with the flask lying horizontally to maximize contact between enzymes and placental tissue. for 60 minutes. The supernatant was filtered through three layers of sterile gauze into SO-cc tubes containing 3 mL FCS, and centrifuged at 1,000 x g for 7 minutes. The pellets were washed twice with Medium 199 (Flow Lab) and pooled. This cellular suspension containing 30 to 80 x 10’ cells was then layered on a tube containing 30 mL of a 40% Percoil gradienty (Pharmacia. Piscataway, NJ). The cell suspension placed on the gradient was centrifuged at 2,500 x g for 15 minutes. Simultaneously, another gradient tube loaded with density markel beads (Pharmacia) was centrifuged to determine the density at which cells were localized. Three bands of cells were isolated with this method at d = 1.033 to I.048 (middle band). d = I.077 to I. 100 (bottom band). and d > I.140 the last containing red blood cells. When individual bands were isolated and reloaded onto Percoll gradients, cells always returned to their original density and no new Metabolism,
Vol 41, NO 6 (June),
1992: pp 596.603
PLACENTAL
TRIGLYCERIDE
METABOLISM
bands were observed. Cells from the different bands were washed twice with Medium 199 and counted with a hemocytometer. The yield of cells was 15 to 40 x 106 in the two non-red blood cell bands. It was difficult to assess total placental tissue weight because of the blood and fibrous tissue, and to weigh the individual pieces would increase the risk of contamination. It was not possible to express yield per unit of placental weight. The viability of the cells in each placenta was greater than 95% as assessed by Trypan Blue exclusion. Placental Cell Culture and Characterization Placental cells from the two non-red cell bands were suspended in HAM’s F-10 culture medium containing 10% serum obtained from the blood of the same placenta donor, 100 pU/mL penicillin, and 100 ug/mL streptomycin, with a final concentration of 1 X lo6 cells/ml. One milliliter of this cell solution was incubated in each well of a 24-well culture plate (Falcon) (Becton-Dickinson Labware, Lincoln Park, NJ) with 95% air and 5% CO2 at 37°C. After 48 hours, the cells were attached to the plates, the medium was changed, and experimental protocols were followed as described below. To identify the cells from both bands, two approaches were followed: In the first, 10 )LL of the cell solution was placed onto 12-well. teflon-coated, microscope slides (formerly Meloy, now Erie Scientific, Portsmouth, NH). After 48 hours in the incubator in the same conditions as described above, the slides were fixed in cold methanol and overlaid with the following panel of monoclonal and polyclonal antibodies: 35PHl1, a monoclonal antibody to cytokeratin no. 8, typically found in glandular epithelium, but also present in trophoblasts i”,it; 34BE12, a monoclonal antibody to high-molecular weight cytokeratins, eg, found in ductal and squamous epithelium i”,ii; 43PE8, a monoclonal antibody to vimentin, the intermediate filament expressed by mesenchymal cells”; HHF35 and CGA7, two monoclonal antibodies specific for muscle actinsi2,13;HAM-56, a monoclonal antibody specific for a cytoplasmic marker of monocytes/macrophagesi4; monoclonal and polyclonal antibodies to factor VIII-related antigen (Dako, Santa Barbara, CA) and endothelial-specific proteini5; and polyclonal antibodies to human chorionic gonadotropin (hCG) and human placental lactogen (hPL), both obtained from Dako. Localization was via an indirect immunofluorescence method, using fluoresceinconjugated anti-mouse or anti-rabbit antibodies (Dako). In the second approach, cells from the middle or bottom band were kept in tissue culture plates for 3 to 4 days, at which time the cells were scraped from the dish, using a rubber policeman. The resulting cell suspension was centrifuged at 1,000 x g, and the pellet thereby obtained was carefully removed and immediately fixed in methacarn fixative (60% methanol, 30% chloroform, and 10% glacial acetic acid) and then fixed overnight. Cell pelletparaffin blocks were obtained by standard processing and embedding procedures. Serial sections of the cell blocks were then cut, deparaffinized, and rehydrated through graded alcohols. The same pane1 of antibodies was then applied, with antibody localization performed by a slight modification of the standard avidin-biotin immunoperoxidase procedure with nickel chloride enhancement of the 3,3’-diaminobenzidine chromogen,t6 yielding a black reaction product. To determine the cell composition of placental tissue, without the possible artifacts induced by the enzymatic digestion, blocks of placental cotyledons were obtained and processed as described above for cell pellets. Serial sections of the tissue blocks were submitted to the same panel of antibodies described above. Antibody localization was via an indirect immunofluorescent method using fluorescein- or rhodamine-conjugated anti-mouse or anti-rabbit antibodies (Dako).
597
An established cell line from human liver-tumor biopsies, designated Hep G2, has been shown to have morphological and metabolic characteristics similar to those of liver parenchymal cellsi7-i9; this cell line was used as a negative control for VLDL-TG hydrolysis, since hepatocytes usually do not have LPL activity. Hep G2 cells were grown in the same medium as the macrophages and trophoblasts, but it contained 10% FCS instead of autologous serum. Labeling WDL-TG Blood from healthy human donors who had fasted for 12 hours was collected in EDTA, and plasma was separated from cells at 2,500 x g for 20 minutes. The different lipoproteins used were isolated by sequential ultracentrifugation.20 Triolein-3H in hexane (Amersham-Searle, Chicago, IL), 0.50 mCi, was dried under NZ dissolved in 200 FL ethanol, transferred to a high-density lipoprotein (HDL) fraction, and incubated in a shaking bath at 37°C for 4 hours. The VLDL solution was added in a 5:l ratio of HDL cholesterol to VLDL cholesterol. Approximately 5 mg partially purified lipid-transfer protein (LTP-I)*’ (generously supplied by Dr J Albers and J Tollefson) was added to the medium and incubated at 37°C overnight. This solution was centrifuged at 40,000 x g for 24 hours. The top fraction, containing VLDL and more than 95% of the Triolein3H, was collected and overlaid on saline and centrifuged again for 24 hours. With this method, we consistently obtained an incorporation of Triolein-‘H exceeding 80% of the amount of radioactivity originally incubated with HDL. Oleic Acid-Human
Serum Albumin Solution
Ten microcuries of oleic acid in hexane labeled on its carboxyl group with i4C (Amersham-Searle) was mixed with 1 mL of an 0.7 mmol/L oleic acid solution, dried under Nr, and redissolved in 1 mL physiological saline and 0.02N NaOH, and then incubated in a shaking bath at 37°C for 1 hour. Once it was completely dissolved, 1 mL of a solution containing 1.4 mmol/L human serum albumin (HSA) essentially free of FA (Sigma) was added for a FFA to HSA ratio of 4:l. This solution of oleic acid-HSA (FA-HSA) was used in the tissue culture FA uptake experiments. Eqerimental
Conditions in Cell Culture
Eighteen hours before starting the experiment, the tissue culture medium was changed with fresh medium containing 2 U heparin/cc to stabilize LPL. When monoclonal anti-LPL antibodies were used, they were added to the medium at the start of the experiment. The monoclonal antibody, 5D2, developed against bovine milk LPL, inhibits human LPL activity by 99% and has no effect on hepatic lipase activity.** Cells were incubated with VLDL at a concentration of approximately 200,000 cpmiwell (50 kmol triolein/well) and the medium was sampled at 0, 4, and 24 hours. FFA from the medium were extracted with the method described by Belfrage and Vaughan.23 Cell pellets were washed 5 times with cold saline and then submitted to a lipid extraction with 1 mL hexane-isopropyl alcohol 3:2 (vol:vol).24 Two hundred microliters of the lipid extract was transferred to a scintillation vial to determine the amount of radioactivity incorporated into the cells. The remaining solution was dried under Nz and redissolved in 100 PL chloroform. Thin-layer chromatography (TLC) was used to isolate the lipid components.25 The protein concentration of the cells was determined by a modification of the method of Lowry et al.*” Simultaneously, some wells were incubated with FA-HSA at approximately 150,000 cpm/well, and at 0,4, and 24 hours, lipid extraction, TLC, and protein determination were performed as described above. Results are expressed as nmol/mg cellular protein based on
BONET ET AL
598
the radioactivity of the extract and the specific activity of the VLDL-TGFA3H or FFA-14C added to the medium. Statistical testing was performed using Student’s paired t test.27 RESULTS
Identification of Placental Cells
Using the tissue-digestion and cell-isolation procedures previously described, 30 to 80 x lo6 cells were typically isolated from the placentas of more than 15 different, normal, pregnant subjects, with 15 to 30 x lo6 cells from the middle band and 20 to 40 x lo6 cells from the bottom band. Under the light microscope, the cells of the two bands were distinctly different. The middle band contained cells that were heterogenous in both size and shape. The bottom band typically consisted of a homogenous population of small, round cells and a very small number of large, round cells, never more than 1% to 2% of the total number. After 2 days in culture, the cells from the middle band formed a syncytium (Fig 1A) and the cells from the bottom band remained as individual cells (Fig 1B). Both populations of cells were attached to the plates. Examination of the cells by immunofluorescence or immunoperoxidase procedures, using the monoclonal- and polyclonal-antibody panel (Figs 1 and 2), revealed the
majority of the large cells in the middle band to be positive with both anticytokeratin antibodies, 35PEll (Fig 2A) and 34BH12, and antivimentin antibody 43pE8 and the antibody to hPL (Fig lA), all findings characteristic of trophoblasts. A subpopulation of smaller cells was also positive with the anticytokeratin antibody 35BH11, and the antivimentin antibody 43PE8, and some were also positive with the anti-hPL antibody, or the anti-macrophage antibody HAM-56, indicating that the small cells from the middle band were a heterogenous population mainly comprised of trophoblasts and macrophages (Fig 2B). No cells in the middle band were positive with antibodies to hCG. Cells from the bottom band were comprised of two morphologically distinct cell populations: a very small number of large cells was positive with both anticytokeratin antibodies, 35PHll (Fig 2C) and 34PE12, the antivimentin antibody 43PE8, and the antibody to the trophoblast marker-protein hPL, identifying these cells as trophoblasts. The small cells comprising the great majority of cells in the bottom band were negative with the anticytokeratin antibodies, and positive only with the antivimentin antibody 43PE8, and the antimacrophage antibody HAM-56 (Fig 2D). None of the small cells in the bottom band expressed the trophoblastic marker hPL or anticytokeratin antibody
Fig 1. lmmunocytochemistry of cells from the middle end bottom bends placed onto microscope slides, after 48 hours of incubation. (A) Cells from the middle band overlaid with polyclonal antibodies to hPL. (9) Cells from the bottom band overlaid witf~ the monoclonal antibody HAMQB, a macrophage-specific antibody.
PLACENTAL TRIGLYCERIDE METABOLISM
Fig 2. lmmunocytochemistry of pellets obtained from cells of the middle and bottom bands after 3 days kept in tissue culture. (A) and (6) correspond to the middle band. (A) Overlaid with the monoclonal antibody 35BHll (which stains trophoblasts). (6) Overlaid with the monoclonal antibody HAM-56 (which stains macrophages]. (A) Reveals that most of the cells contained in the middle band are mainly trophoblasts, although macrophages are also present in the middle band. (C) and (D) correspond to the bottom band. (Cl Overlaid with the monoclonal antibody 356Hll. (D) Overlaid with the monoclonal antibody HAM-56. (B) Cells contained in the bottom band are mostly macrophages with occasional large trophoblasts.
35PHll. No cells in either band reacted positively for endothelial or smooth muscle cell markers. Thus, the majority of cells in the middle band were trophoblast in nature and formed a syncytiotrophoblast within 24 to 48 hours of incubation, although macrophages were present in this band. On the other hand, the large majority of cells in the bottom band were macrophages, with only a few cells showing characteristics of trophoblasts. When slides of placental tissue were subjected to a panel of monoclonal and polyclonal antibodies specific for different cell markers, the two predominant cell types were again macrophages and trophoblasts (Fig 3). The trophoblasts were localized on the periphery of the placental villi (Fig 3A), and macrophages were the predominant parenchyma1 cell (Fig 3B). Blood vessels reacted positively to factor VIII-related antigen and to HHF35 and C6A7, showing the presence of endothelial and smooth muscle cells (results not shown). None of these cells were obtained after enzymatic digestion. To further characterize the differences between middleand bottom-band cells, progesterone secretion by cells from each band was determined in tissue culture (Table 1). The data from one representative experiment reveal the middle band to secrete the largest amount of progesterone, with very little being secreted by cells of the bottom band. This observation confirms the different nature of cells from
middle and bottom bands, and supports the view that the great majority of cells of the middle band are steroid hormone-secreting trophoblasts, while the bottom-band cells are not. LPL Activity in Human Placental Cells
To characterize the metabolism of VLDL-TG by placental cells and the role of LPL, VLDL-TG labeled with 3H-oleic acid were incubated with cells fror : the middle band (trophoblast) or bottom band (macrophages) and with Hep G2 cells, a model of hepatocyte function as a negative control for LPL activity. VLDL-TGFA were also incubated in the absence of cells to detect any intrinsic VLDL-TG hydrolysis during the time of incubation (0, 4, and 24 hours). The amount of FFA generated in the absence of cells was small, but it was subtracted from the amount of FFA generated in the incubation wells containing cells, yielding cell-mediated FFA generated by VLDL-TG hydrolysis. The VLDL-TGFA were released as FFA into the medium and were incorporated into the cells at 0, 4, and 24 hours (a representative experiment is shown in Table 2). Both the appearance of labeled FFA in the medium and the uptake of labeled FA by the cells were consistently higher in macrophages than in the predominant trophoblast cell types (Table 2). In the experiment shown, labeled FFA in
BONET ET AL
Fig 3. lmmunocytochemistry of placental tissue. (A) Histological section of placenta villi overlaid with monoclonal antibodies to 356Hll. a marker of trophoblasts. (B) Slides of placental villi overlaid with monoclonal antibody to HAM-56, a marker of macrophages. This figure shows that the two main populations of term placental cells are macrophages, localized in interstitial tissue and trophoblasts, localized in the periphery.
the medium (nmol/mg protein) was 101.3 +- 5.3 in macrophages versus 29.9 ? 6.4 in the trophoblast-rich band after 24 hours. The amount of label incorporated into cell lipid extract was a fairly constant proportion of the amount of FFA in the medium in both cell populations, and was approximately 20% of the amount of labeled FFA in the medium at 24 hours (Table 2). Most of the radioactivity in the cellular lipid extract was in the form of triolein. with very little as oleic acid (Table 2). These effects were replicated in all experiments from five different placentas. In both populations of placental cells, monoclonal antibodies against LPL that were added to the medium before incubation markedly decreased the appearance of labeled FFA in the medium, in most cases to below the limit of detection (Table 2). The uptake of labeled FA by the cells was similarly abolished. These effects were replicated in four separate experiments. When monoclonal antibodies Table 1. Progesterone Secretion in Placental Cells (ng/mL) Middle Band
Bottom Band
29.2 f 1.05
3.13 -t 0.24
NOTE. Cells from concentration Progesterone
the
middle
of 0.5 to 1 x concentration
band
or bottom
lo6 cells/ml.
band
Average
tested after 48 hours.
were
of four
at a wells.
against apolipoprotein (apo) A-I were added in the same way as the LPL antibodies, to serve as a control, neither the release of labeled FFA nor the uptake of labeled FA by the cells were affected (data not presented). Thus, VLDL-TG hydrolysis by cells is mediated by LPL. Uptake of FA-HSA by Placental Cells A comparison of the uptake and metabolism of FFA by both types of placental cells was undertaken. FA-HSA was added to the medium for comparison to FFA appearing in the medium from VLDL-TG hydrolysis in the previous experiments for similar incubation times (0, 4, and 24 hours). The incorporation of label was similar in macrophages and trophoblasts in four such experiments (representative experiment, Table 3) but was fourfold to sixfold higher in Hep G2 cells. Most of the intracellular radioactivity in macrophages and trophoblasts was found in FFA at 4 hours, but at 24 hours, most of the radioactivity was in cellular TGFA. These results indicate an ongoing esterification process following FA uptake. In contrast, esterified TGFA was the predominant form of radiolabeled FA in the Hep G2 cells at both 4 and 24 hours, indicating a more rapid esterification rate than in macrophages or trophoblasts. In the Hep G2 cells, phospholipids also comprised
PLACENTAL
TRIGLYCERIDE
METABOLISM
601
Table 2. Metabolism of VLDL-TG by Placental-Cell Bottom Band (Largely Macrophages) and Middle Band (Largely Trophoblasts)
in Tissue
Culture BottomBand (Hours)
VLDL-TGFA (Appearingas:)
mAb
FFA in medium
0
(-)
2.22 (0.97)X
(+I
24
33.5
101.3
(4.40)
1.39
Middle Band(Hours)
4
(5.30) 0
0
Hep G2 Cells(Hours)
0
4
24
0
0.77
4.26
29.9
0.22
4
24
0.42
1.96
(0.3) 0
(3.50) 0
(6.37) 0
(0.12) -
(0.36) -
(0.67) -
(0.70) Cell total lipid
f-1
0
1.4
26.3
(+)
0
0
0.26
5.40
0.38
2.76
6.15
(0.65) 0
(0.18) -
(1.02) -
(1.60)
0
(0.70) 0
(2.55)
(0.35)
0.41
0
(0.14) Cell total TGFA
0
(-)
0
(+)
0
21.2
0.79
(6.70)
(0.40)
0.70
0
0
0.25
4.73
(0.09)
(0.80)
0
0
0
1.13
1.48
(0.27) -
(0.31) -
(0.06)
0
l-1
Cell FFA
0.29
0
(+)
0
2.70 (0.24)
(0.07)
0.12
0
0
0.15
0.54
(0.05)
(0.05)
0
0
0
0.56
0.50
(0.11) -
(0.16) -
(0.04) NOTE. Results are presented as means in nmol/mg
cellular protein and values in parentheses are 2 SE. Values represent 2 to 4 wells/time.
Experiments with different placentas show similar results. Abbreviations:
+/-
mAb, incubation of cells with or without monoclonal anti-LPL antibody 502 added to the medium. Monoclonal antibody was
not tested in Hep G2 cells.
connection, the abundance of macrophages in the term placenta is noteworthy, and our studies confirm previous observations32-34 that macrophages comprise most of the interstitial cells of the human placental villi and include about 50% of isolated placental parenchymal cells. The amount of LPL activity found in the trophoblast-rich population was always lower than that in macrophages, approximating a ratio of 1:4. These data raise the possibility that most of the LPL activity found in the trophoblast-rich population is secondary to the macrophages present, with trophoblast celk that consistently comprise 10% to 20% of the cells present in the middle band (Fig 2B). However, we cannot completely exclude the possibility that some LPL activity resides in the trophoblasts. Despite numerous attempts, we were unable to obtain a sample of trophoblasts completely free of macrophages. Even doing a double-percoll isolation, we found macrophages contaminating the middle band. The high tendency of trophoblasts to aggregate after enzymatic digestion may actually entrap macrophages. Alternatively, LPL could originate in macrophages and migrate to trophoblasts, similar to the way LPL is synthesized in parenchymal cells and migrates to the apical surface of endothelial cells.34.35
about 40% of the intracellular radioactivity, in contrast to 10% in macrophages and trophoblasts (Table 3). The greater avidity of Hep G2 cells for FFA and the fourfold to sixfold greater esterification to cellular TG compared with placental macrophages and trophoblasts (Table 3) contrasts with the greater metabolism of medium VLDLTGFA by macrophages and trophoblasts than by Hep G2 cells. DISCUSSION
Our experiments show that macrophages and trophoblasts are the two main parenchymal components of human placenta. Macrophages isolated from human placenta possess the majority of TG-hydrolytic activity compared with trophoblasts, and this activity is inhibited more than 99% when both cell types are incubated with a specific monoclonal anti-LPL antibody that does not inhibit hepatic 1ipase.22 The finding of high hydrolytic activity per milligram protein in the macrophages is consistent with the findings of other investigators, that human blood monocyte-derived macrophages and macrophages from atheromatous lesions have been shown to contain LPL activity.29-3* In this Table 3. Uptake of Human Albumin-Bound
FFA W by Placental-Cell Bottom Band (Largely Macrophages) and Middle Band (Largely Trophoblasts)
in Tissue Culture
Bottom Band (Hours)
Cell FFA (Uptake As)
4
Hep G2 Cells (Hours)
Middle Band(Hours) 24
6.14 5 0.97
4
0.62 + 0.2
24
4
24
4.45 f 2.24
11.8 ‘- 1.48
25.1 t 3.54
Total lipids
1.56 2 0.32
TGFA
0.16 + 0.001
3.84 -+ 0.31
0.18 2 0.03
2.01 k 0.75
6.48 r 0.77
14.8 + 0.76
FFA
1.01 k 0.22
0.26 2 0.09
0.21 2 0.12
0.117 + 0.01
0.45 k 0.04
1.01 -+ 0.12
Phospholipid
0.27 -c 0.03
0.66 2 0.13
0.14 2 0.01
0.53 ? 0.3
3.76 2 0.03
9.53 f 0.43
NOTE. Two nanomoles oleic acid 1% was added per well. Two to six wells were averaged/time are presented as the mean 2 SE.
point. Results are in nmol/mg cellular protein and
BONET ET AL
602
The amount of radioactivity found in placental ceil lipid extracts as a consequence of LPL activity was always proportional (20% to 25%) to the amount of FA released to the medium from the lipoproteins. Most of the radioactivity found in the cell lipid extract was in the form of TG, and not as free oleic acid. In addition, when macrophages and trophoblasts were incubated with FA-HSA, cell lipid extracts also showed that more than 90% of the intracellular radioactivity was present as TG, underscoring the dominance of the esterification pathway. This observation is consistent with the results found by other authors showing a high esterification rate of FFA to TG in human placental cells.36-3sWhen placental cells were incubated with monoclonal anti-LPL antibody, they did not acquire any radioactivity, indicating that the uptake of VLDL-TGFA is in the form of FFA, which, after becoming intracellular, are reesterified to TG. The current study therefore demonstrates that the placental cells can take up FA from VLDL-TG, and this process is mediated by prior hydrolysis of VLDL-TG by LPL and not by direct uptake of intact TG. The pattern uptake of FFA versus VLDL-TG-derived FA by placental macrophages and trophoblasts was different from that of Hep G2 cells. When the two types of placental cells were incubated with FA-HSA, the uptake of FFA after 24 hours of incubation was 4% to 6% of the initial amount contained in the incubation medium, in contrast to greater than 30% uptake of FFA by Hep G2 cells (Table 3). In contrast, when the three cell types were incubated with labeled VLDL-TGFA, cellular TGFA content was 15-fold greater in macrophages than in Hep G2 cells, and threefold greater in trophoblast cells compared with Hep G2 cells. These differences indicate that VLDL-TG is a more important source of intracellular FA than are circulating FFA for placental cells, while FFA are relatively more important to the Hep G2 cell. This comparison of FFA versus TGFA uptake underscores the importance of placental LPL activity in placental function, and suggests that circulating TGFA may be a more important source of placental FA than are albumin-bound FFA. The relative importance of TGFA over FFA as a source of intracellular placental-cell FA is in agreement with the observation of Thomas et al4 in a pregnant guinea pig placental perfusion model. In those studies, the concentration of maternal VLDL-TG-derived FFA was twofold
higher in fetal placental perfusate than in maternal plasma. In contrast, when labeled FFA were given to the mother, the fetal placental-perfusate concentration of FFA was 80% of that in maternal plasma. Both our studies and the studies of Thomas et al4 support the hypothesis that TG-rich lipoprotein may be a more efficient means to deliver FA to the placenta and fetus than are FFA. This mechanism could favor the delivery of essential FA and energy acquired from the diet via chylomicrons to the fetus if these alimentary particles are metabolized by the placenta in the same manner as endogenous VLDL. In vivo support for this finding is suggested by our recent observation that maternal TG concentrations are a predictor of infant birth weight39 and that placental LPL activity is a predictor of fetal guinea pig weight.4 It is important to place these findings of placental VLDL-TG metabolism in the context of other forms of lipid access to the placenta. Specifically, cholesterol has ready access to the placenta via LDL and the LDL receptor and via HDLz and a non-receptor-mediated mechanism.q This subject has been recently reviewed.41 In conclusion, the results presented in this report show that placental macrophages are an abundant cell type in isolates of placental cells. Placental macrophages have the capability to hydrolyze VLDL-TG, and this process is essentially entirely LPL dependent. Trophoblast cells show less activity, which may be due to contaminating macrophages. Regardless of source, these results show that LPL activity in placental cells has an essential role in the uptake of TGFA of TG-rich lipoproteins, and possibly, in their transfer to the fetus. Compared with Hep G2 cells, placental macrophages and trophoblasts are more avid in their uptake of VLDL-TGFA than FFA, suggesting that TGFA are the predominant FA source for the placenta. If macrophages are the main source of LPL activity, it is noteworthy that they are not in direct contact with maternal blood. The possible transfer of LPL from macrophages to trophoblasts, cells that are in direct contact with blood, and therefore, with lipoproteins, is currently being investigated. ACKNOWLEDGMENT
The authors are grateful to Carol Dorsett for technical tance, to Alan Greeves for typing. and to John Tollefson J. Albers for a gift of human transfer protein.
assisand John
REFERENCES 1. Olivecrona T, Bengtsson-Olivecrona G: Lipoprotein lipase and hepatic lipase. Curr Opin Lipid 1:222-230, 1990 2. Eckel RH: Lipoprotein lipase: A multifunctional enzyme relevant to common metabolic diseases. N Engl J Med 320:10601068,1989 3. Knopp RH, Warth MR, Charles D, et al: Lipoprotein metabolism in pregnancy, fat transport to the fetus, and the effects of diabetes. Biol Neonate 50:297-317,1986 4. Thomas CR, Lowy C, St Hillaire RJ, et al: Studies on the placental hydrolysis and transfer of lipids to the fetal guinea pig, in Miller RK, Thiede HA (eds): Fetal Nutrition, Metabolism, and Immunology: The Role of the Placenta. New York, NY, Plenum, 1984, pp 135-146 5. Elphick MC, Hull D: Rabbit placental clearing-factor lipase
and transfer to the fetus of fatty acids derived from triglycerides injected into the mother. J Physio1273:475-487, 1977 6. Mallov S. Alousi AD: Lipoprotein lipase activity of rat and human placenta. Proc Sot Exp Biol Med 119:301-306.1965 7. Hummel L, Schwartze A, Schirrmeister W, et al: Maternal plasma triglycerides as a source of fetal fatty acids. Acta Biol Med Germ 35:1635-1641,1976 8. Elphick MC, Filshie GM, Hull D: The passage of fat emulsion across the human placenta. Br J Obstet Gynaecol85:610-618.1978 9. Zeitler P, Markoff E, Handwerger S: Characterization of the synthesis and release of human placental lactogen and human chorionic gonadotropin by an enriched population of dispersed placental cells. J Clin Endocrinol Metab 57:812-818,1983 10. Gown AM, Vogel AM: Monoclonal antibodies to intermedi-
PLACENTAL TRIGLYCERIDE METABOLISM
ate sized filament proteins: Unique and cross-reacting antibodies. J Cell Biol95:414-424,1982 11. Gown AM, Vogel AM: Monoclonal antibodies to intermediate filament proteins of human cells II. Distribution of filament proteins in normal human tissue. Am J Patho! 114:309-321,1984 12. Gown AM, Vogel AM, Gordon D, et al: Smooth musclespecific monoclonal antibodies: Specific recognition of smooth muscle actin isotypes. J Cell Biol 100:808-813, 1985 13. Tsukda T, Tippens D, Mar H, et al: HHF35, a muscle-actinspecific monoclonal antibody. I. Biochemical and immunocytochemical characterization. Am J Patho! 126:51-60,1987 14. Gown AM, Tsukada T, Ross R: Human atherosclerosis II. Immunocytochemical analysis of the cellular composition of human atherosclerotic lesions. Am J Patho! 125:191-207,1986 15. Jaffe EA: Endothelial cells and the biology of factor VIII. N Engl J Med 296:377-383,1977 16. Gown AM, Vogel AM: Monoclonal antibodies to intermediate filament proteins of human cells: II. Distribution of filament proteins in normal human tissue. Am J Patho! 114:309-321,1984 17. Knowles BB, Howe CC, Aden PP: Human hepatocellular carcinoma cell lines secrete the major plasma proteins and hepatitis surface antigen. Science 206:497-499, 1980 18. Zannis VI, Breslow JL, San Giacomo TR, et al: Characterization of the major apolipoproteins secreted by two human hepatoma cell lines. Biochemistry 20:7089-7096,198l 19. Erickson SK, Fielding PE: Parameters of cholesterol metabolism in the human hepatoma cell line, Hep G2. J Lipid Res 27:875-883, 1986 20. Have! RJ, Eder H, Bragdon J: The distribution and chemical composition of ultracentrifugally reported lipoproteins in human serum. J Clin Invest 34:1345-1353,1955 21. Albers JJ, Tollefson JH, Chen CH, et al: Isolation and characterization of human plasma lipid transfer proteins. Arteriosclerosis 4:49-58,1984 22. Babirak SP, Iverius PH, Fujimoto WY, et al: Detection and characterization of the heterozygote state for lipoprotein lipase deficiency. Arteriosclerosis 9:326-334,1989 23. Belfrage P, Vaughan M: Simple liquid-partition system for isolation of labelled oleic acid from mixtures with glycerides. J Lipid Res 10:341-344,1969 24. Hara A, Radin NS: Lipid extraction of tissues with a low toxicity solvent. Anal Biochem 90:420-426,1979 25. Skipski VP, Barclay M: Thin-layer chromatography of lipids. Methods Enzymol14:530-598,1969
603
26. Lowry OH, Rosebrough NJ, Farr AL, et al: Protein measurement with the Folin phenol reagent. J Biol Chem 193:265-275,195l 27. Snedecor GW, Cochran WG: Statistical Methods (ed 6). Ames, IA, Iowa State University Press, 1967, pp 59 28. Chait A, Iverius PH, Brunzell JD: Lipoprotein lipase secretion by human monocyte derived macrophages. J Clin Invest 69:490-493,1982 29, Stray N, Letnes H, Blomhoff JP: Synthesis and secretion of lipoprotein lipase by human monocyte derived macrophages. Stand J Gastroenterol20:67-72,1985 (suppl107) 30. O’Brien K, Gordon D, Deeb S, et al: Localization of lipoprotein lipase mRNA to foam cell-rich regions of human atherosclerotic plaques. Clin Res 39:335A, 1991 (abstr) 31. Wood GW: Mononuclear phagocytes in the human placenta. Placenta 1:113-123,198O 32. Flynn A, Finke JH, Hilfiker ML: Placental mononuclear macrophages as a source of interleukin-1. Science 218:475-477, 1982 33. Castelluci M, Kaufmann P: Hotbauer cells, in Benirschke K, Kaufmann P (eds): The Pathology of the Human Placenta. New York, NY, Springer-Verlag, 1990, pp 71-80 34. Nilsson-Ehle P, Garfinkel AS, Schotz MC: Lipolytic enzymes and plasma lipoprotein metabolism. Annu Rev Biochem 49:667-693, 1980 35. Blanchette-Mackie EJ, Masuno H, Dwyer NK, et al: Lipoprotein lipase in myocytes and capillary endothelium of heart: Immunocytochemical study. Am J Physiol256:E818-E828,1989 36. Szabo AJ, DeLellis R, Grimaldi RD: Triglyceride synthesis by the human placenta. Am J Obstet Gynecol115:257-266,1973 37. Coleman RA: Placental metabolism and transport of lipid. Fed Proc 45:2519-2523,1986 38. Coleman RA, Haynes EB: Microsomal and lysosomal enzymes of triacylglycerol metabolism in rat placenta. Biochem J 217:391-397,1984 39. Knopp RH, Magee MS, Larson MP, et al: Alternative screening tests and birth weight associations in pregnant women with abnormal glucose screening. Diabetes 37:llOA, 1988 (abstr) 40. Lasuncibn MA, Bonet B, Knopp RI-L Mechanism of the HDLr stimulation of progesterone secretion in cultured placental trophoblast. J Lipid Res 32:1073-1087,199l 41. Knopp RH, Magee MS, Bonet B, et al: Lipid metabolism in pregnancy and lactation and effects on fetal growth and development, in Cowett RM (ed): Principles of Perinatal-Neonatal Metabolism. New York, NY, Springer-Verlag, 1991, pp 177-203
