- Email: [email protected]
The immunoglobulin G fraction from plasma containing antiphospholipid antibodies causes increased placental thromboxane production
Volume 167 Number 6
enzymes and thrombocytopenia in the third trimester of pregnancy: an unusual case report and a review of the literature. AM J OBSTET GYNECOL 1989; 161 :322-3. 8. Riely CA. Acute fatty liver of pregnancy. Sem Liver Dis 1987;7:47-53. 9. Sibai BM. The HELLP syndrome (hemolysis, elevated liver enzymes, and low platelets): much ado about nothing? AM J OBSTET GYNECOL 1990;162:311-6. 10. Barton .JR, Sibai BM. Care of the pregnancy complicated
Hepatic histopathology in HELLP syndrome
by HELLP syndrome. Obstet Gynecol Clin North Am 1991;18:165-79. II. Goodlin RC, Holdt D. Impending gestosis. Obstet Gynecol 1981;58:743-5. 12. Thiagarajah S, Bourgeois F.J, Harbert GM, Caudle MR. Thrombocytopenia in preeclampsia: associated abnormalities and management principles. AM .J OBSTET GYNECOL 1984;150:1-7.
The immunoglobulin G fraction from plasma containing antiphospholipid antibodies causes increased placental thromboxane production Alan M. Peaceman, MD, and Karen A. Rehnberg, BA
Chicago, Illinois OBJECTIVE: Our objective was to evaluate whether the immunoglobulin G fraction from plasma containing high levels of antiphospholipid antibodies alters the production of prostacyclin or thromboxane when incubated with normal human placental tissue. STUDY DESIGN: The immunoglobulin G fraction was prepared from the pooled plasma of five volunteers with normal obstetric histories and no anti phospholipid antibodies. The immunoglobulin G fraction was prepared similarly from a patient with the antiphospholipid antibody syndrome. Doses of these immunoglobulin G fractions ranging from 0.3 to 3.0 mg were incubated with placental explants obtained from eight normal pregnancies, and prostacyclin and thromboxane production was assessed over 48 hours. RESULTS: Placental prostacyclin production was unaltered by incubation with either immunoglobulin G fraction at any of the doses tested. Placental thromboxane production tripled by 32 hours with the addition of 0.6, 1.5, and 3.0 mg of the anti phospholipid antibody fraction (p < 0.05) compared with baseline production but was unaltered by the addition of the normal pooled plasma fraction at any dose. The increase in thromboxane production with anti phospholipid antibody immunoglobulin G appeared to be dose related. CONCLUSION: The immunoglobulin G fraction prepared from plasma containing anti phospholipid antibodies caused increased placental thromboxane production without altering prostacyclin production. (AM J OBSTET GYNECOL 1992;167:1543-7.)
Key words: Antiphospholipid antibodies, prostacyclin, thromboxane, recurrent pregnancy loss, placental explants Antiphospholipid antibodies are circulating immunoglobulins that have recently been associated with poor obstetric outcomes, including recurrent abortion, maFrom the Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, Northwestern University Medical School, Northwestern Memorial Hospital. Presented at the Thirty-ninth Annual Meeting of the Society for Gynecologic investigations, San Antonio, Texas, March 18-21, 1992. Reprint requests: Alan M. Peaceman, MD, Northwestern University Medical School, 333 E. Superior St., Suite 410, Chicago, lL 60611.
6/6/41723
ternal thrombosis, fetal death, and early, severe preeclampsia. I The clinical findings of thrombosis, thrombocytopenia, or recurrent pregnancy loss in association with the laboratory detection of moderate or high levels of anticardiolipin or lupus anticoagulant has been termed the antiphospholipid antibody syndrome." Data from a BALB/c mouse model support a causative role for these antibodies in pregnancy complications and justifY attempts to suppress antibody activity in affected patients.' All proposed treatments are empiric, however, because no controlled trials have been performed 1543
1544
Peace man and Rehnberg
to date to document their efficacy in preventing adverse pregnancy outcomes. The choice of appropriate treatments is hampered by the absence of a defined pathophysiology for antiphospholipid antibody-mediated pregnancy loss. Initial studies suggested that lupus anticoagulant, the primary antiphospholipid antibody implicated in recurrent pregnancy loss, affects vascular endothelial cells, resulting in decreased prostacyclin production! Subsequent work has been inconsistent,5-9 and the site and mechanism of action for antiphospholipid antibodies remain in doubt. An alternative site in some patients may be on the placenta, with altered prostanoid production at that level being involved in the resulting adverse pregnancy outcome. We evaluated the hypothesis that the immunoglobulin G (IgG) fraction for plasma containing high levels of antiphospholipid antibody alters the production of prostacyclin or thromboxane when incubated with normal human placental tissue.
Methods Whole blood was obtained by venipuncture from five nonpregnant female volunteers in tubes containing sodium citrate. All five had normal medical and obstetric histories. The samples were centrifuged at 3000g to remove blood cells and platelets, and the plasma obtained was then centrifuged again at 3000g before the supernatants were pooled. This normal pooled plasma was then treated wtih sufficient calcium chloride to achieve a final calcium concentration of 5 mmoVL, and the resulting clot was removed with a glass rod with gentle stirring. An equal volume of a saturated solution of ammonium sulfate (27.5%) was added, precipitating the IgG fraction from the normal pooled plasma. This simple extraction technique was chosen for its high recovery of IgG and its ability to remove unidentified plasma factors that interfere with the radioimmunoassays. The normal pooled plasma IgG fraction thus derived was resuspended in phosphate-buffered saline solution and dialyzed against 5 L of phosphate-buffered saline solution to remove traces of the ammonium sulfate. Whole blood was obtained in a similar fashion from a nonpregnant patient with the diagnosis of the antiphospholipid antibody syndrome. This patient had a history of venous thrombosis and recurrent pregnancy loss; she was found to have a high titer of IgG anticardiolipin by highly specific enzyme-linked immunosorbent assay and to be strongly positive for lupus anticoagulant as determined by the Russell viper venom test. The plasma was processed in the same manner as the normal pooled plasma, and ammonium sulfate precipitation of the IgG fraction was also performed. The composition of the normal pooled plasma and
December 1992 Am J Obstet Gynecol
antiphospholipid antibody IgG fractions was evaluated with gel electrophoresis, both revealing approximately 40% IgG with albumin as the only other significant protein in the sample. Spectrophotometric analysis revealed similar total protein concentrations in the normal pooled plasma and antiphospholipid antibody IgG fractions (14 mg/ml). The IgG fraction from the patient with the antiphospholipid antibody syndrome retained its anticoagulant activity in vitro, prolonging the dilute Russell viper venom test to 62.5 seconds (control 37.0 seconds) when mixed I: I with normal serum. The processed normal pooled plasma IgG fraction also prolonged this clotting assay but to a lesser degree (52.0 seconds; upper limit of normal in our laboratory is 45 seconds). Both the normal pooled plasma and the antiphospholipid antibody IgG were frozen and stored for later use. Fresh placentas were obtained from eight normal term pregnancies and processed under sterile conditions within IS minutes of delivery. For each placental experiment the basal plate was removed, and multiple placental explants were dissected from the placenta. These ex plants were washed repeatedly with Dulbecco's modified Eagle medium (Gibco, Grand Island, N.Y.), cut into small pieces, and excess buffer was removed with sterile gauze. Approximately 350 mg of placental tissue in three or four fragments was then placed into separate incubation wells, each containing Dulbecco's modified Eagle medium. To each well was added either normal pooled plasma or antiphospholipid antibody IgG. The same four volumes of n9rmal pooled plasma and antiphospholipid antibody IgG (20, 40, 100, and 200 IJ-l) were tested in separate wells with tissue from each of the eight placentas. These volumes corresponded to 0.3, 0.6, 1.5, and 3.0 mg of IgG and produced IgG concentrations in the wells approximating 0.4%, 0.8%, 2.0%, and 4.0% of the concentrations found in maternal plasma. In addition, a well containing Dulbecco's modified Eagle medium but no added IgG was included with each placenta as a control. The amount of Dulbecco's modified Eagle medium added to each well was adjusted to create final well volumes of 7 m!. The wells were incubated at 37° C in a Dubnoff shaker located in a sterile hood, and 200 IJ-l aliquots were removed at 0, 8, 20, 32, and 48 hours. Because the relative amount of volume removed with each sampling was small (approximately 3%), replacement was not performed, and the concentration of IgG remained essentially constant. These aliquots were assessed for placental prostacyclin and thromboxane production by the radioimmunoassay of their stable metabolites, 6-keto-prostaglandin F 1", and thromboxane B2 • 10 All test wells were run in duplicate with the results averaged for analysis. Data are expressed in picograms per milli-
APA and placental prostanoid production
Volume 167 I\"umber (i
0.3 ml IIG
1000
A.
900
iii ~
B.
900 800
800
I
0.6 ml IIG
1000
1545
700
700 800
0:
800
E
800
E ......
500
APA
Ii!
400
~
300
200
~boL
100 10 TIME
1000
1000
1.5 ml IIG
c.
800
110
10
(hours)
3.0 ml IIG
1100
D.
800
700
700 ~
'" ...... E
800
E
800
Ii!
400
~
40
30
TIME
IlOO
1......
20
(hours)
E
'"
CD
APA
~
300 200
100
20 TIME
30
40
400 JOG
200
NPP
~_ _"'--4t===~_""'fCONTROL
10
APA
600
800
eo
(hours)
NPP CONTROL
100 0
0
10
20 TIME
30
40
eo
(hours)
Fig. 1. Effect of IgG fraction on thromboxane production by normal human placental tissue. "Ibromboxane production over 48 hours for each dose of IgG is assessed by measurement of stable metabolite thromboxane B2 (TXB2) and is expressed in picograms per milligram of wet tissue. Measurements obtained from tissue incubated with antiphospholipid antibody (APA) and normal pooled plasma (NPP) IgG are shown and compared with baseline placental tissue production with no added IgG (Control). Four doses of IgG tested were as follows: A, 0.3 mg; B, 0.6 mg; C, 1.5 mg; and D, 3.0 mg. Data represent mean ± SE (n = 8). Asterisk, p < 0.05 by repeated-measures testing, antiphospholipid antibody versus normal pooled plasma IgG and control.
gram of wet tissue, with mean values ± 1 SE displayed. Statistical differences in prostanoid production were analyzed with repeated-measures testing. This placental explant technique has been validated previously, demonstrating tissue viability, effectiveness of high oxygen tension, and penetration of stimulatory substances. II This project was approved by the Institutional Review Board of Northwestern University Medical School.
Results Prostacyclin production by placental tissue incubated with each dose of normal pooled plasma or antiphospholipid antibody IgG is shown in Fig. 1. For comparison control wells that contained no added IgG and indicate baseline placental production are also included in each graph. For all of the concentrations tested prostacyclin production by normal placental tissue was
not altered significantly by the addition of either antiphospholipid antibody or normal pooled plasma IgG. Placental thromboxane production was not altered by the addition of 0.3 mg of either normal pooled plasma or antiphospholipid antibody IgG and was similar to baseline placental thromboxane production (Fig. 2, A). Placental thromboxane production remained unchanged with the addition of increasing concentrations of normal pooled plasma IgG but was increased with the higher concentrations of added antiphospholipid antibody IgG (Fig. 2, B through D). For the 0.6, 1.2, and 3.0 mg doses, the differences between antiphospholipid antibody and control or normal pooled plasma IgG were statistically significant by 20 hours of incubation (p < 0.05). By 32 hours the amount of thromboxane produced by placental tissue incubated with the three higher doses of antiphospholipid anti-
1546 Peaceman and Rehnberg
December 1992 Am J Obslet Gynecol
A.
0.3 mg IgG
2200 2000
Ci E
1Il00
C.
1Il00
1\1
1400
a.
1200
S: .c (ij LL
C) Q.
E CD ~
tD
2000
aE
APA
C.
S:
IIPP
1\1
s:::
a.
(ij
1Il00
APA
1600 1400 1200 1000
1000
LL
IlOO
C)
IlOO
600
E CD
600
Q. ~
400
tD
200 10
20
·30
~
400
200 10
110
aE
C.
S:
1800 1Il00
APA NPP CONTROL
s:::
a.
(ij
LL
c.
1.5 mg IgG
2000
1\1
1200 1000
aE
D.
C.
1Il00 1600
APA
S: 1\1
s:::
CONTROL
a.
(ij LL
9CD
IlOO
Q.
tD
(hours)
2000
800
~
60
40
30
3.0 mg IgG
2200
C)
0;-
20
TIME
TIME 2200
B.
0.6 mg IgG
2200
NPP
C)
400
E CD
200
tD
~
110
nilE
(hours)
10
20
nilE
30
40
110
(hours)
Fig. 2. Effect of IgG fraction on prostaeyelin production by normal human placental tissue. Prostaeyelin production over 48 hours for each dose of IgG is assessed by measurement of the stable metabolite 6-keto prostaglandin Fin (PGF 1 alpha) and is expressed in picograms per milligram of wet tissue. Measurements obtained from tissue incubated with anti phospholipid antibody (APA) and normal pooled plasma (NPP) IgG are shown and compared with baseline placental tissue production with no added IgG (Control). Four doses of IgG tested were as follows: A, 0.3 mg/ B, 0.6 mg; C, 1.5 mg; and D, 3.0 mg. Data represent mean ± SE (n = 8).
body IgG was at least three times greater than baseline placental production and the amount produced by tissue incubated with normal pooled plasma IgG.
Comment The addition of the IgG fraction derived from plasma containing antiphospholipid antibodies produced a consistent increase in thromboxane production by normal placental tissue over baseline production, whereas IgG from normal nonpregnant patients did not alter thromboxane production. This increase in thromboxane production with anti phospholipid antibody IgG appeared to be dose related. Conversely, no alteration in prostacyclin production by the placenta was seen with the addition of either normal pooled plasma or antiphospholipid antibody IgG at any concentration tested.
In 1981 Carreras et al. 4 originally proposed that lupus anticoagulant, the primary antiphospholipid antibody implicated in recurrent pregnancy loss, affects vascular endothelial cells, resulting in decreased prostacyclin production. This finding, however, could not be consistently reproduced with plasma taken from all patients with the clinical anti phospholipid antibody syndrome. 5 • 6 Subsequent work with cultured endothelial cell models failed to find any consistent alteration in prostanoid production by cells exposed to the IgG fraction from plasma that contained lupus anticoagulant. 7·9 Consequently, the site and mechanism of action for antiphospholipid antibodies remain to be elucidated. The placenta is a logical site for the action of antiphospholipid antibodies. Normal pregnancy is thought to be associated with a relative balance between
Volume lG7 "umber {i
the opposing effects of prostacyclin and thromboxane produced by the placenta. Any effect on placental prostanoid production could result in paracrine alteration of the local environment at the maternal-placental interface. Decreased prostacyclin and increased thromboxane production have been associated with vasoconstriction, platelet aggregation, uterine activity, and decreased uteroplacental blood flow. 'O The prostacyclinto-thromboxane ratio is known to be altered significantly in preeclampsia, a syndrome that shares many pathologic features with the antiphospholipid antibody syndrome. ' °It is therefore plausible that antiphospholipid antibodies are active at the maternalplacental interface and alter prostanoid production at that level. The data presented here support three hypotheses relating to the pathophysiologic makeup of the antiphospholipid antibody syndrome. First, altered placental thromboxane production rather than a change in prostacyclin production is involved in anti phospholipid antibody-related pregnancy loss. This differs from the findings of Carreras et al" but is consistent with the association of increased thromboxane production with abruptio placentae, preeclampsia, and other clinical manifestations of the anti phospholipid antibody syndrome. Second, antiphospholipid antibodies are active at the placental surface. With direct immunofluorescent studies antiphospholipid antibody IgG has been shown to bind with decidual tissue of treated mice. 3 Alterations in placental production of vasoactive substances have been implicated as contributing to other pregnancy complications as well. 10 The involvement of the placenta in the pathogenesis of this syndrome is consistent with the finding that most women are primarily symptomatic when pregnant. Last, the adverse effects of antiphospholipid antibodies may be self-contained in the IgG fraction and capable of causing characteristic changes in otherwise normal tissue from unaffected individuals. This concept of an IgG transferable factor has been shown to cause recurrent pregnancy loss in mice" and likely pertains to humans as well. If a consistent effect on thromboxane production can be confirmed with IgG fractions obtained from other
APA and placental prostanoid production
1547
patients with the antiphospholipid antibody syndrome, it might allow development of a bioassay for antiphospholipid antibody activity. Currently, testing for antiphospholipid antibodies, especially anticardiolipin antibody, is not uniform because significant interlaboratory variation exists among testing sites. 12 This variation likely leads to overtreatment of patients with falsepositive results and possibly undertreatment of some with false-negative results. Development and validation of a bioassay would allow more accurate identification of affected patients. REFERENCES
I. Triplett DA. Antiphospholipid antibodies and recurrent pregnancy loss. Am J Reprod Immunol 1989;20:52-67. 2. Harris EN. Syndrome of the black swan. Br J Rheumatol 1988;26:324-6. 3. Branch DW, Dudley DJ, Mitchell MD, et al. Immunoglobulin G fractions from patients with antiphospholipid antibodies cause fetal death in BALB/c mice: a model for autoimmune fetal loss. AM J OSSTET GYNECOL 1990;163: 210-6. 4. Carreras LO, Machin SJ, Deman R, et al. Arterial thrombosis, intrauterine death and "lupus" anticoagulant: detection of immunoglobulin intefering with prostacyclin formation. Lancet 1981; I ;244-6. 5. Carreras LO, Vermylen JG. "Lupus" anticoagulant and thrombosis - possible role of inhibition of prostacyclin formation. Thromb Hemost 1982;48:38-40. 6. Ros JO, Tarres MV, Baucells MV, Maired JJ, Solano JT. Prednisone and maternal lupus anticoagulant. Lancet 1983;2:576. 7. Dudley DJ, Mitchell MD, Branch DW. Pathophysiology of anti phospholipid antibodies: absence of prostaglandinmediated effects on cultured endothelium. AM J OBSTET GYNECOL 1990; 162:953-9. 8. Cariou R, Tobelem G, Belluci S, et al. Effect of lupus anticoagulant on antithrombogenic properties of endothelial cells - inhibition of thrombomodulin-dependent protein C activation. Thromb Hemost 1988;60:54-8. 9. Walker TS, Triplett DA, Javed N, Musgrave K. Evaluation of lupus anticoagulant: antiphospholipid antibodies, endothelium associated immunoglobulin, endothelial prostacyclin secretion, and antigenic protein S levels. Thromb Res 1988;51:267-81. 10. Walsh SW. Preeclampsia: an imbalance in placental prostacyclin and thromboxane production. AM J Ossn:r GvNECOI. 1985;152;335-40. II. Walsh SW, Behr MJ, Allen NH. Placental prostacyclin production in normal and toxemic pregnancies. AM J Ossn:!" GYNECOL 1985;151:110-5. 12. Peaceman AM, Silver RK, MacGregor SN, Socol ML. Interlaboratory variation in antiphospholipid antibody testing. AM J OSSTET GYNECOL 1992;166: 1780-7.
enzymes and thrombocytopenia in the third trimester of pregnancy: an unusual case report and a review of the literature. AM J OBSTET GYNECOL 1989; 161 :322-3. 8. Riely CA. Acute fatty liver of pregnancy. Sem Liver Dis 1987;7:47-53. 9. Sibai BM. The HELLP syndrome (hemolysis, elevated liver enzymes, and low platelets): much ado about nothing? AM J OBSTET GYNECOL 1990;162:311-6. 10. Barton .JR, Sibai BM. Care of the pregnancy complicated
Hepatic histopathology in HELLP syndrome
by HELLP syndrome. Obstet Gynecol Clin North Am 1991;18:165-79. II. Goodlin RC, Holdt D. Impending gestosis. Obstet Gynecol 1981;58:743-5. 12. Thiagarajah S, Bourgeois F.J, Harbert GM, Caudle MR. Thrombocytopenia in preeclampsia: associated abnormalities and management principles. AM .J OBSTET GYNECOL 1984;150:1-7.
The immunoglobulin G fraction from plasma containing antiphospholipid antibodies causes increased placental thromboxane production Alan M. Peaceman, MD, and Karen A. Rehnberg, BA
Chicago, Illinois OBJECTIVE: Our objective was to evaluate whether the immunoglobulin G fraction from plasma containing high levels of antiphospholipid antibodies alters the production of prostacyclin or thromboxane when incubated with normal human placental tissue. STUDY DESIGN: The immunoglobulin G fraction was prepared from the pooled plasma of five volunteers with normal obstetric histories and no anti phospholipid antibodies. The immunoglobulin G fraction was prepared similarly from a patient with the antiphospholipid antibody syndrome. Doses of these immunoglobulin G fractions ranging from 0.3 to 3.0 mg were incubated with placental explants obtained from eight normal pregnancies, and prostacyclin and thromboxane production was assessed over 48 hours. RESULTS: Placental prostacyclin production was unaltered by incubation with either immunoglobulin G fraction at any of the doses tested. Placental thromboxane production tripled by 32 hours with the addition of 0.6, 1.5, and 3.0 mg of the anti phospholipid antibody fraction (p < 0.05) compared with baseline production but was unaltered by the addition of the normal pooled plasma fraction at any dose. The increase in thromboxane production with anti phospholipid antibody immunoglobulin G appeared to be dose related. CONCLUSION: The immunoglobulin G fraction prepared from plasma containing anti phospholipid antibodies caused increased placental thromboxane production without altering prostacyclin production. (AM J OBSTET GYNECOL 1992;167:1543-7.)
Key words: Antiphospholipid antibodies, prostacyclin, thromboxane, recurrent pregnancy loss, placental explants Antiphospholipid antibodies are circulating immunoglobulins that have recently been associated with poor obstetric outcomes, including recurrent abortion, maFrom the Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, Northwestern University Medical School, Northwestern Memorial Hospital. Presented at the Thirty-ninth Annual Meeting of the Society for Gynecologic investigations, San Antonio, Texas, March 18-21, 1992. Reprint requests: Alan M. Peaceman, MD, Northwestern University Medical School, 333 E. Superior St., Suite 410, Chicago, lL 60611.
6/6/41723
ternal thrombosis, fetal death, and early, severe preeclampsia. I The clinical findings of thrombosis, thrombocytopenia, or recurrent pregnancy loss in association with the laboratory detection of moderate or high levels of anticardiolipin or lupus anticoagulant has been termed the antiphospholipid antibody syndrome." Data from a BALB/c mouse model support a causative role for these antibodies in pregnancy complications and justifY attempts to suppress antibody activity in affected patients.' All proposed treatments are empiric, however, because no controlled trials have been performed 1543
1544
Peace man and Rehnberg
to date to document their efficacy in preventing adverse pregnancy outcomes. The choice of appropriate treatments is hampered by the absence of a defined pathophysiology for antiphospholipid antibody-mediated pregnancy loss. Initial studies suggested that lupus anticoagulant, the primary antiphospholipid antibody implicated in recurrent pregnancy loss, affects vascular endothelial cells, resulting in decreased prostacyclin production! Subsequent work has been inconsistent,5-9 and the site and mechanism of action for antiphospholipid antibodies remain in doubt. An alternative site in some patients may be on the placenta, with altered prostanoid production at that level being involved in the resulting adverse pregnancy outcome. We evaluated the hypothesis that the immunoglobulin G (IgG) fraction for plasma containing high levels of antiphospholipid antibody alters the production of prostacyclin or thromboxane when incubated with normal human placental tissue.
Methods Whole blood was obtained by venipuncture from five nonpregnant female volunteers in tubes containing sodium citrate. All five had normal medical and obstetric histories. The samples were centrifuged at 3000g to remove blood cells and platelets, and the plasma obtained was then centrifuged again at 3000g before the supernatants were pooled. This normal pooled plasma was then treated wtih sufficient calcium chloride to achieve a final calcium concentration of 5 mmoVL, and the resulting clot was removed with a glass rod with gentle stirring. An equal volume of a saturated solution of ammonium sulfate (27.5%) was added, precipitating the IgG fraction from the normal pooled plasma. This simple extraction technique was chosen for its high recovery of IgG and its ability to remove unidentified plasma factors that interfere with the radioimmunoassays. The normal pooled plasma IgG fraction thus derived was resuspended in phosphate-buffered saline solution and dialyzed against 5 L of phosphate-buffered saline solution to remove traces of the ammonium sulfate. Whole blood was obtained in a similar fashion from a nonpregnant patient with the diagnosis of the antiphospholipid antibody syndrome. This patient had a history of venous thrombosis and recurrent pregnancy loss; she was found to have a high titer of IgG anticardiolipin by highly specific enzyme-linked immunosorbent assay and to be strongly positive for lupus anticoagulant as determined by the Russell viper venom test. The plasma was processed in the same manner as the normal pooled plasma, and ammonium sulfate precipitation of the IgG fraction was also performed. The composition of the normal pooled plasma and
December 1992 Am J Obstet Gynecol
antiphospholipid antibody IgG fractions was evaluated with gel electrophoresis, both revealing approximately 40% IgG with albumin as the only other significant protein in the sample. Spectrophotometric analysis revealed similar total protein concentrations in the normal pooled plasma and antiphospholipid antibody IgG fractions (14 mg/ml). The IgG fraction from the patient with the antiphospholipid antibody syndrome retained its anticoagulant activity in vitro, prolonging the dilute Russell viper venom test to 62.5 seconds (control 37.0 seconds) when mixed I: I with normal serum. The processed normal pooled plasma IgG fraction also prolonged this clotting assay but to a lesser degree (52.0 seconds; upper limit of normal in our laboratory is 45 seconds). Both the normal pooled plasma and the antiphospholipid antibody IgG were frozen and stored for later use. Fresh placentas were obtained from eight normal term pregnancies and processed under sterile conditions within IS minutes of delivery. For each placental experiment the basal plate was removed, and multiple placental explants were dissected from the placenta. These ex plants were washed repeatedly with Dulbecco's modified Eagle medium (Gibco, Grand Island, N.Y.), cut into small pieces, and excess buffer was removed with sterile gauze. Approximately 350 mg of placental tissue in three or four fragments was then placed into separate incubation wells, each containing Dulbecco's modified Eagle medium. To each well was added either normal pooled plasma or antiphospholipid antibody IgG. The same four volumes of n9rmal pooled plasma and antiphospholipid antibody IgG (20, 40, 100, and 200 IJ-l) were tested in separate wells with tissue from each of the eight placentas. These volumes corresponded to 0.3, 0.6, 1.5, and 3.0 mg of IgG and produced IgG concentrations in the wells approximating 0.4%, 0.8%, 2.0%, and 4.0% of the concentrations found in maternal plasma. In addition, a well containing Dulbecco's modified Eagle medium but no added IgG was included with each placenta as a control. The amount of Dulbecco's modified Eagle medium added to each well was adjusted to create final well volumes of 7 m!. The wells were incubated at 37° C in a Dubnoff shaker located in a sterile hood, and 200 IJ-l aliquots were removed at 0, 8, 20, 32, and 48 hours. Because the relative amount of volume removed with each sampling was small (approximately 3%), replacement was not performed, and the concentration of IgG remained essentially constant. These aliquots were assessed for placental prostacyclin and thromboxane production by the radioimmunoassay of their stable metabolites, 6-keto-prostaglandin F 1", and thromboxane B2 • 10 All test wells were run in duplicate with the results averaged for analysis. Data are expressed in picograms per milli-
APA and placental prostanoid production
Volume 167 I\"umber (i
0.3 ml IIG
1000
A.
900
iii ~
B.
900 800
800
I
0.6 ml IIG
1000
1545
700
700 800
0:
800
E
800
E ......
500
APA
Ii!
400
~
300
200
~boL
100 10 TIME
1000
1000
1.5 ml IIG
c.
800
110
10
(hours)
3.0 ml IIG
1100
D.
800
700
700 ~
'" ...... E
800
E
800
Ii!
400
~
40
30
TIME
IlOO
1......
20
(hours)
E
'"
CD
APA
~
300 200
100
20 TIME
30
40
400 JOG
200
NPP
~_ _"'--4t===~_""'fCONTROL
10
APA
600
800
eo
(hours)
NPP CONTROL
100 0
0
10
20 TIME
30
40
eo
(hours)
Fig. 1. Effect of IgG fraction on thromboxane production by normal human placental tissue. "Ibromboxane production over 48 hours for each dose of IgG is assessed by measurement of stable metabolite thromboxane B2 (TXB2) and is expressed in picograms per milligram of wet tissue. Measurements obtained from tissue incubated with antiphospholipid antibody (APA) and normal pooled plasma (NPP) IgG are shown and compared with baseline placental tissue production with no added IgG (Control). Four doses of IgG tested were as follows: A, 0.3 mg; B, 0.6 mg; C, 1.5 mg; and D, 3.0 mg. Data represent mean ± SE (n = 8). Asterisk, p < 0.05 by repeated-measures testing, antiphospholipid antibody versus normal pooled plasma IgG and control.
gram of wet tissue, with mean values ± 1 SE displayed. Statistical differences in prostanoid production were analyzed with repeated-measures testing. This placental explant technique has been validated previously, demonstrating tissue viability, effectiveness of high oxygen tension, and penetration of stimulatory substances. II This project was approved by the Institutional Review Board of Northwestern University Medical School.
Results Prostacyclin production by placental tissue incubated with each dose of normal pooled plasma or antiphospholipid antibody IgG is shown in Fig. 1. For comparison control wells that contained no added IgG and indicate baseline placental production are also included in each graph. For all of the concentrations tested prostacyclin production by normal placental tissue was
not altered significantly by the addition of either antiphospholipid antibody or normal pooled plasma IgG. Placental thromboxane production was not altered by the addition of 0.3 mg of either normal pooled plasma or antiphospholipid antibody IgG and was similar to baseline placental thromboxane production (Fig. 2, A). Placental thromboxane production remained unchanged with the addition of increasing concentrations of normal pooled plasma IgG but was increased with the higher concentrations of added antiphospholipid antibody IgG (Fig. 2, B through D). For the 0.6, 1.2, and 3.0 mg doses, the differences between antiphospholipid antibody and control or normal pooled plasma IgG were statistically significant by 20 hours of incubation (p < 0.05). By 32 hours the amount of thromboxane produced by placental tissue incubated with the three higher doses of antiphospholipid anti-
1546 Peaceman and Rehnberg
December 1992 Am J Obslet Gynecol
A.
0.3 mg IgG
2200 2000
Ci E
1Il00
C.
1Il00
1\1
1400
a.
1200
S: .c (ij LL
C) Q.
E CD ~
tD
2000
aE
APA
C.
S:
IIPP
1\1
s:::
a.
(ij
1Il00
APA
1600 1400 1200 1000
1000
LL
IlOO
C)
IlOO
600
E CD
600
Q. ~
400
tD
200 10
20
·30
~
400
200 10
110
aE
C.
S:
1800 1Il00
APA NPP CONTROL
s:::
a.
(ij
LL
c.
1.5 mg IgG
2000
1\1
1200 1000
aE
D.
C.
1Il00 1600
APA
S: 1\1
s:::
CONTROL
a.
(ij LL
9CD
IlOO
Q.
tD
(hours)
2000
800
~
60
40
30
3.0 mg IgG
2200
C)
0;-
20
TIME
TIME 2200
B.
0.6 mg IgG
2200
NPP
C)
400
E CD
200
tD
~
110
nilE
(hours)
10
20
nilE
30
40
110
(hours)
Fig. 2. Effect of IgG fraction on prostaeyelin production by normal human placental tissue. Prostaeyelin production over 48 hours for each dose of IgG is assessed by measurement of the stable metabolite 6-keto prostaglandin Fin (PGF 1 alpha) and is expressed in picograms per milligram of wet tissue. Measurements obtained from tissue incubated with anti phospholipid antibody (APA) and normal pooled plasma (NPP) IgG are shown and compared with baseline placental tissue production with no added IgG (Control). Four doses of IgG tested were as follows: A, 0.3 mg/ B, 0.6 mg; C, 1.5 mg; and D, 3.0 mg. Data represent mean ± SE (n = 8).
body IgG was at least three times greater than baseline placental production and the amount produced by tissue incubated with normal pooled plasma IgG.
Comment The addition of the IgG fraction derived from plasma containing antiphospholipid antibodies produced a consistent increase in thromboxane production by normal placental tissue over baseline production, whereas IgG from normal nonpregnant patients did not alter thromboxane production. This increase in thromboxane production with anti phospholipid antibody IgG appeared to be dose related. Conversely, no alteration in prostacyclin production by the placenta was seen with the addition of either normal pooled plasma or antiphospholipid antibody IgG at any concentration tested.
In 1981 Carreras et al. 4 originally proposed that lupus anticoagulant, the primary antiphospholipid antibody implicated in recurrent pregnancy loss, affects vascular endothelial cells, resulting in decreased prostacyclin production. This finding, however, could not be consistently reproduced with plasma taken from all patients with the clinical anti phospholipid antibody syndrome. 5 • 6 Subsequent work with cultured endothelial cell models failed to find any consistent alteration in prostanoid production by cells exposed to the IgG fraction from plasma that contained lupus anticoagulant. 7·9 Consequently, the site and mechanism of action for antiphospholipid antibodies remain to be elucidated. The placenta is a logical site for the action of antiphospholipid antibodies. Normal pregnancy is thought to be associated with a relative balance between
Volume lG7 "umber {i
the opposing effects of prostacyclin and thromboxane produced by the placenta. Any effect on placental prostanoid production could result in paracrine alteration of the local environment at the maternal-placental interface. Decreased prostacyclin and increased thromboxane production have been associated with vasoconstriction, platelet aggregation, uterine activity, and decreased uteroplacental blood flow. 'O The prostacyclinto-thromboxane ratio is known to be altered significantly in preeclampsia, a syndrome that shares many pathologic features with the antiphospholipid antibody syndrome. ' °It is therefore plausible that antiphospholipid antibodies are active at the maternalplacental interface and alter prostanoid production at that level. The data presented here support three hypotheses relating to the pathophysiologic makeup of the antiphospholipid antibody syndrome. First, altered placental thromboxane production rather than a change in prostacyclin production is involved in anti phospholipid antibody-related pregnancy loss. This differs from the findings of Carreras et al" but is consistent with the association of increased thromboxane production with abruptio placentae, preeclampsia, and other clinical manifestations of the anti phospholipid antibody syndrome. Second, antiphospholipid antibodies are active at the placental surface. With direct immunofluorescent studies antiphospholipid antibody IgG has been shown to bind with decidual tissue of treated mice. 3 Alterations in placental production of vasoactive substances have been implicated as contributing to other pregnancy complications as well. 10 The involvement of the placenta in the pathogenesis of this syndrome is consistent with the finding that most women are primarily symptomatic when pregnant. Last, the adverse effects of antiphospholipid antibodies may be self-contained in the IgG fraction and capable of causing characteristic changes in otherwise normal tissue from unaffected individuals. This concept of an IgG transferable factor has been shown to cause recurrent pregnancy loss in mice" and likely pertains to humans as well. If a consistent effect on thromboxane production can be confirmed with IgG fractions obtained from other
APA and placental prostanoid production
1547
patients with the antiphospholipid antibody syndrome, it might allow development of a bioassay for antiphospholipid antibody activity. Currently, testing for antiphospholipid antibodies, especially anticardiolipin antibody, is not uniform because significant interlaboratory variation exists among testing sites. 12 This variation likely leads to overtreatment of patients with falsepositive results and possibly undertreatment of some with false-negative results. Development and validation of a bioassay would allow more accurate identification of affected patients. REFERENCES
I. Triplett DA. Antiphospholipid antibodies and recurrent pregnancy loss. Am J Reprod Immunol 1989;20:52-67. 2. Harris EN. Syndrome of the black swan. Br J Rheumatol 1988;26:324-6. 3. Branch DW, Dudley DJ, Mitchell MD, et al. Immunoglobulin G fractions from patients with antiphospholipid antibodies cause fetal death in BALB/c mice: a model for autoimmune fetal loss. AM J OSSTET GYNECOL 1990;163: 210-6. 4. Carreras LO, Machin SJ, Deman R, et al. Arterial thrombosis, intrauterine death and "lupus" anticoagulant: detection of immunoglobulin intefering with prostacyclin formation. Lancet 1981; I ;244-6. 5. Carreras LO, Vermylen JG. "Lupus" anticoagulant and thrombosis - possible role of inhibition of prostacyclin formation. Thromb Hemost 1982;48:38-40. 6. Ros JO, Tarres MV, Baucells MV, Maired JJ, Solano JT. Prednisone and maternal lupus anticoagulant. Lancet 1983;2:576. 7. Dudley DJ, Mitchell MD, Branch DW. Pathophysiology of anti phospholipid antibodies: absence of prostaglandinmediated effects on cultured endothelium. AM J OBSTET GYNECOL 1990; 162:953-9. 8. Cariou R, Tobelem G, Belluci S, et al. Effect of lupus anticoagulant on antithrombogenic properties of endothelial cells - inhibition of thrombomodulin-dependent protein C activation. Thromb Hemost 1988;60:54-8. 9. Walker TS, Triplett DA, Javed N, Musgrave K. Evaluation of lupus anticoagulant: antiphospholipid antibodies, endothelium associated immunoglobulin, endothelial prostacyclin secretion, and antigenic protein S levels. Thromb Res 1988;51:267-81. 10. Walsh SW. Preeclampsia: an imbalance in placental prostacyclin and thromboxane production. AM J Ossn:r GvNECOI. 1985;152;335-40. II. Walsh SW, Behr MJ, Allen NH. Placental prostacyclin production in normal and toxemic pregnancies. AM J Ossn:!" GYNECOL 1985;151:110-5. 12. Peaceman AM, Silver RK, MacGregor SN, Socol ML. Interlaboratory variation in antiphospholipid antibody testing. AM J OSSTET GYNECOL 1992;166: 1780-7.