|Home | About | Journals | Submit | Contact Us | Français|
Pre-eclampsia affects approximately 5% of pregnancies and remains a leading cause of maternal and neonatal mortality and morbidity in the United States and the world1,2. The clinical hallmarks of this maternal disorder include hypertension, proteinuria, endothelial dysfunction and placental defects. Advanced-stage clinical symptoms include cerebral hemorrhage, renal failure and the HELLP (hemolysis, elevated liver enzymes and low platelets) syndrome. An effective treatment of pre-eclampsia is unavailable owing to the poor understanding of the pathogenesis of the disease. Numerous recent studies3–5 have shown that women with pre-eclampsia possess autoantibodies, termed AT1-AAs, that bind and activate the angiotensin II receptor type 1a (AT1 receptor). We show here that key features of pre-eclampsia, including hypertension, proteinuria, glomerular endotheliosis (a classical renal lesion of pre-eclampsia), placental abnormalities and small fetus size appeared in pregnant mice after injection with either total IgG or affinity-purified AT1-AAs from women with pre-eclampsia. These features were prevented by co-injection with losartan, an AT1 receptor antagonist, or by an antibody neutralizing seven–amino-acid epitope peptide. Thus, our studies indicate that pre-eclampsia may be a pregnancy-induced autoimmune disease in which key features of the disease result from autoantibody-induced angiotensin receptor activation. This hypothesis has obvious implications regarding pre-eclampsia screening, diagnosis and therapy.
The pathophysiology of pre-eclampsia remains largely unknown. A widely held view is that placental ischemia, stemming from shallow trophoblast invasion and improper spiral artery remodeling, is a crucial initiating event1,2,6. Numerous studies have focused on circulating factors secreted by the ischemic placenta that contribute to the maternal syndrome7–10. The oxidative stress and vascular damage resulting from placental ischemia are believed to underlie the enhanced maternal inflammatory response associated with pre-eclampsia11. Immune mechanisms and the renin-angiotensin system are also implicated in pre-eclampsia3–5,12,15. These two concepts were united in a previous report16, in which it was shown that sera from women with pre-eclampsia contain autoantibodies that react with AT1 angiotensin receptors in a stimulatory fashion. Subsequent to these findings, multiple other groups, including our own, showed that many features of pre-eclampsia could be explained by the ability of these autoantibodies to activate AT1 receptors on a variety of cells and provoke biological responses that are relevant to the pathophysiology of pre-eclampsia17–22. However, previous work has been restricted to the use of in vitro systems and has been unable to specifically address the relevance of AT1-AAs to the defining features of pre-eclampsia, hypertension and proteinuria.
To evaluate the in vivo pathophysiological consequences of AT1-AAs, we introduced IgG (approximately 800 µg) from either normotensive pregnant women or pregnant women with pre-eclampsia into pregnant mice on day 13 of gestation. We chose day 13 because this stage of mouse pregnancy is comparable to early onset pre-eclampsia in humans and is a time at which we can reliably determine whether a mouse is pregnant. We initially used western blot analysis to show that human IgG was readily detectable for at least 5d after injection (Fig. 1a). By ELISA we found that human IgG persisted in the circulation of injected mice until the time of killing on gestation day 18 and that the IgG concentrations in injected mice were similar for mice injected with IgG from normotensive pregnant women or women with pre-eclampsia (Fig. 1b). To determine whether the injected antibody retained biological activity, we killed pregnant mice 5 d after antibody injection, purified IgG from the maternal mouse sera and assayed the isolated IgG for AT1 receptor agonistic activity with a reporter cell line in which AT1 receptor activation results in increased expression of a 4× NFAT cis element–driven luciferase reporter. The results (Fig. 1c) show that IgG from women with pre-eclampsia retained the ability to activate AT1 receptors for at least 5 d after retro-orbital injection into pregnant mice. In contrast, IgG isolated from pregnant mice injected with IgG from normotensive pregnant women did not stimulate luciferase synthesis (Fig. 1c). These results show that it is possible to introduce physiologically relevant concentrations of human IgG into pregnant mice and that the injected antibody persists in a biologically active form for many days in the maternal circulation.
Hypertension is a defining feature of pre-eclampsia. To determine whether IgG from women with pre-eclampsia has the ability to contribute to gestational hypertension, we measured systolic blood pressure at daily intervals after retro-orbital injection of IgG from pregnant women into pregnant mice on day 13 of gestation. IgG from women with pre-eclampsia but not IgG from normotensive pregnant women induced a significant increase in blood pressure that was evident 4 d after antibody injection (Fig. 1d). The increased blood pressure resulting from injection of IgG from women with pre-eclampsia was prevented by co-injection with losartan, an AT1 receptor antagonist (Fig. 1d), suggesting that the antibody-induced increase in blood pressure required AT1 receptor activation.
Antibody-injected mice were also analyzed for proteinuria, another defining feature of pre-eclampsia, by determining the ratio of urinary albumin to creatinine on gestation day 18, when the dams were killed. The results (Fig. 1e) show that a significant increase in urinary albumin occurred after injection of IgG from women with pre-eclampsia in contrast to IgG from normotensive pregnant women. The autoantibody-induced proteinuria was prevented by co-injection with losartan (Fig. 1e), indicating that it resulted from autoantibody-induced AT1 receptor activation. Overall, the results (Fig. 1) indicate that IgG from women with pre-eclampsia, in contrast to IgG from normotensive pregnant women, induced hypertension and proteinuria in pregnant mice and that these antibody-induced features of pre-eclampsia required AT1 receptor activation. Our inability to detect increased blood pressure and urinary protein in the pregnant mice injected with IgG from the controls allows us to conclude that it is unlikely that immune complexes formed between mouse antibodies raised against the injected human IgG account for the features observed in the pregnant mice injected with human IgG from subjects with pre-eclampsia.
AT1-AAs interact with a specific seven–amino acid sequence on the second extracellular loop of the AT1 receptor16. Competition experiments have shown that this seven–amino acid epitope peptide blocks autoantibody-induced AT1 receptor activation in a variety of cell lines16–22. To determine whether the hypertension and proteinuria in antibody-injected pregnant mice was a result of AT1-AAs, we injected mice on gestational day 13 with IgG from women with pre-eclampsia in the presence or absence of the seven–amino acid epitope peptide and subsequently monitored blood pressure and urinary protein. The results (Fig. 1d,e) show that antibody-induced hypertension and proteinuria were prevented by co-injection with the seven–amino acid epitope peptide. These findings indicate that the IgG-induced hypertension and proteinuria resulted from autoantibodies that interacted with a seven–amino acid epitope on the second extracellular loop of the AT1 receptor.
The ability of the seven–amino acid epitope peptide to block AT1-AA–mediated AT1 receptor activation16–22 suggests a physical association between AT1-AAs and the second extracellular loop of the AT1 receptor. We tested this possibility directly with a GST-peptide fusion protein containing a 27–amino acid sequence corresponding to the second extracellular loop of the AT1 receptor (GST–27-aa). We initially tested the ability of AT1-AAs to bind the GST–27-aa fusion protein by incubating total IgG from women with pre-eclampsia with the GST–27-aa fusion protein coupled to AminoLink Plus Coupling gel. Column chromatography was performed and the original IgG preparation was separated into a flow-through fraction that did not bind the GST–27-aa fusion protein and a fraction that initially bound and was subsequently eluted from the GST–27-aa peptide fusion protein. Antibodies in the flow-through and eluted fractions were assayed for their ability to detect AT1 receptor protein by western blot analysis. A commercially available antibody to the AT1 receptor was used as a positive control (Fig. 2a). The results (Fig. 2a) show that the antibodies in the flow-through fraction were unable to detect a protein at 43 kDa, corresponding to the AT1 receptor. However, the fraction that initially bound the column and that was subsequently eluted (that is, the elution fraction, Fig. 2a) readily detected a protein of approximately 43 kDa that co-migrated with the protein detected by the commercially available anti-AT1 receptor antibody. To determine whether the affinity-purified AT1-AAs retained biological activity and to determine the effectiveness of the affinity purification procedure, we assayed the flow-through and elution fractions for AT1-AA biological activity using the 4× NFAT cis element–driven luciferase reporter cell line mentioned above. The bioassay revealed essentially no activity in the flow-through fraction (Fig. 2b), which is consistent with the western blotting results (Fig. 2a). However, the elution fraction was highly enriched in AT1-AA activity, approximately 40-fold over the original IgG preparation (Fig. 2b) with a recovery of greater than 30% (Fig. 2c). The ability to stimulate luciferase synthesis in this cell line was inhibited by losartan and by the seven–amino acid epitope peptide (Fig. 2d). Thus, the affinity-purified AT1-AAs are highly enriched and retain biological activity.
The affinity-purified AT1-AAs were tested for the ability to induce symptoms of pre-eclampsia when injected into pregnant mice. Retro-orbital injection of ~20 µg of these autoantibodies into pregnant mice resulted in a robust stimulation of blood pressure that was apparent within 3 d (Fig. 2e), as was the development of significant proteinuria (Fig. 2f). As with total IgG from women with pre-eclampsia, the hypertensive and proteinuric effects of affinity-purified AT1-AAs injection were inhibited by losartan or the seven–amino acid epitope peptide (Fig. 2e,f). None of the mice injected with control IgG showed a significant increase in blood pressure or urinary protein. These results provide direct evidence that autoantibodies from women with pre-eclampsia that bind the second extracellular loop of the AT1 receptor induce hypertension and proteinuria when injected into pregnant mice.
Pre-eclampsia is also associated with impaired renal function and characteristic alterations in renal histology. To evaluate the potential role of AT1-AAs in renal pathophysiology, we injected mice on gestational day 13 with IgG from normotensive pregnant women or from women with pre-eclampsia in the presence or absence of losartan or the seven–amino acid epitope peptide. Five days after antibody injection, pregnant mice were killed and their kidneys were isolated, fixed, sectioned and analyzed by H&E staining, which revealed extensive renal damage in the samples from the mice injected with the IgG from women with pre-eclampsia (Fig. 3a,b). The majority of the glomeruli seen in these mice showed normal to smaller-sized glomeruli with endothelial swelling, causing narrowing and total obliteration or occlusion of the glomerular capillary spaces (a condition termed endotheliosis; Fig. 3b). No marked cellular proliferation or increase in mesangial matrix was observed in these mice. These changes were not evident in mice after injection of IgG from normotensive pregnant women and were diminished when preeclamptic IgG was co-injected with either losartan or 7-aa (Fig. 3a–c). The glomeruli showed a characteristic ‘bloodless’ consolidated appearance, in contrast to the open glomerular tufts of the pregnant mice injected with IgG from normotensive pregnant women (Fig. 3b). Examination of kidneys from these mice by electron microscopy confirmed the glomerular changes observed by light microscopy. Narrowing and complete obliteration of the capillary loop spaces by the swollen cytoplasm of the endothelial cells are seen in the affected glomeruli, together with loss of the fenestrations of the endothelial cells (Fig. 3d). Foot process effacement is seen focally in the podocytes (Fig. 3d). Occasional electron-dense materials are also observed in the subendothelial and mesangial areas (Fig. 3d). These series of histological changes in the kidney are pathognomonic for the renal disease that often occurs in women with pre-eclampsia24. In contrast, no segmental glomerulosclerosis, significant mesangial cell proliferation or vessel wall changes were seen on light microscopy or electron microscopy (Fig. 3c,d). No renal histopathological changes were observed in mice injected with IgG from normotensive pregnant women (Fig. 3a–d). In mice co-injected with losartan or the seven–amino acid epitope peptide and IgG from women with pre-eclampsia, the histological changes were partially reversed (Fig. 3a–d). Consistent with these results, we found similar renal damages in pregnant mice injected with affinity-purified AT1-AAs (data not shown). Thus, this autoantibody can induce renal histopathological changes characteristic of pre-eclampsia via AT1 receptor activation.
Pre-eclampsia is frequently associated with placental abnormalities and intrauterine growth retardation26. To evaluate the potential contribution of AT1-AAs to these features of pre-eclampsia, we injected pregnant mice (gestational day 13) with IgG from normotensive pregnant women or women with pre-eclampsia and examined placentas and fetuses on gestational day 18. The placentas were significantly smaller in mice injected with IgG from women with pre-eclampsia (0.083 ± 0.015 g) in comparison to those from mice injected with IgG from normotensive pregnant women (0.125 ± 0.013 g; Supplementary Table 1 online). The reduction in placenta size was largely prevented by co-injection with losartan or the seven–amino acid epitope peptide, indicating that the reduction placental size or fetal weight was mediated by AT1-AA–induced AT1 receptor activation (Supplementary Table 1). Placentas from the mice injected with preeclamptic IgG showed increased intraparenchymal calcifications compared to those from mice injected with IgG from normotensive pregnant women (Supplementary Fig. 1 online). Significant fetal growth retardation was also observed by gestation day 18 (Supplementary Table 1) in mice injected with IgG from women with pre-eclampsia (1.28 ± 0.15 g for normotensive versus 1.01 ± 0.17 g for pre-eclamptic). The reduction in fetal weight was largely prevented when pre-eclamptic IgG was co-injected with losartan or the seven–amino acid epitope peptide (Supplementary Table 1).
Soluble fms-related tyrosine kinase-1 (sFlt1) is a soluble form of the vascular endothelial growth factor (VEGF) receptor that is significantly elevated in the plasma of women with pre-eclampsia and is believed to contribute to disease pathology by interference with VEGF signaling8,9,24. Soluble endoglin (sEng) is a soluble form of endoglin, a cell surface co-receptor of transforming growth factor-β receptor. It is also secreted by the placenta at elevated levels in women with pre-eclampsia and is believed to contribute to disease by interference with transforming growth factor-β signaling25. To evaluate the potential contribution of AT1-AAs to increased production of sFlt1 and sEng in pre-eclampsia, we gave pregnant mice a single injection (at gestation day 13) or double injections (at gestation days 13 and 14) of IgG from normotensive pregnant women or women with pre-eclampsia and determined the serum abundance of sFlt1 and sEng of injected mice on gestation day 18. The results (Supplementary Fig. 2 online) show that sFlt1 concentrations were significantly higher in pregnant mice receiving either single or double injections of IgG from women with pre-eclampsia over those observed with pregnant mice injected with IgG from normotensive pregnant women. sEng was also induced by injection with IgG from women with pre-eclampsia, although in this case two injections were required to see a significant increase (Supplementary Fig. 2). Autoantibody-mediated induction of sFlt1 and sEng in pregnant mice was inhibited by co-injection with losartan or the seven–amino acid epitope peptide, indicating that the induction was mediated by AT1-AAs and required AT1 receptor activation. These results indicate that autoantibodies from women with pre-eclampsia mediate a dosage-dependent stimulation of sFlt1 and sEng production in pregnant mice via AT1 receptor activation.
Autoantibody-induced hypertensive and proteinuric effects in pregnant mice could be mediated through the action of excessive sFlt1 production by the placenta22. To assess the potential contribution of AT1-AAs independently of excess sFlt1, we introduced IgG from normotensive pregnant women or women with pre-eclampsia into nonpregnant mice and monitored sFlt1 abundance, blood pressure and proteinuria at various times after injection. The concentration of sFlt1 was very low in nonpregnant mice injected with IgG from normotensive pregnant women and was not increased significantly by injection with IgG from women with pre-eclampsia (Fig. 4a). In contrast, sFlt1 concentrations were much higher in pregnant mice and showed a significant additional increase in pregnant mice injected with IgG from women with pre-eclampsia over levels observed in mice injected with IgG from normotensive pregnant women (Fig. 4a). Thus, nonpregnant mice have very low amounts of sFlt1 in comparison to pregnant mice, and the low concentration is not significantly increased as a result of injection with IgG from women with pre-eclampsia.
Subsequently, we monitored blood pressure and proteinuria in IgG-injected nonpregnant mice. The results (Fig. 4b) show that IgG from women with pre-eclampsia, in contrast to IgG from normotensive pregnant women, stimulated an increase in blood pressure. None of ten mice injected with control IgG showed a significant increase in blood pressure. The autoantibody-mediated increase in blood pressure was inhibited by co-injection with losartan or the seven–amino acid epitope peptide (Fig. 4b). In addition, there was no significant difference between the autoantibody-induced hypertension observed in pregnant mice (Fig. 1b and Fig. 2e) and that in nonpregnant mice (Fig. 4b), suggesting that AT1-AA–mediated hypertension does not require antibody-induced sFlt1 produced by the placenta.
Finally, we measured the urinary protein in antibody-injected nonpregnant mice. The results (Fig. 4c) show that neither IgG from women with pre-eclampsia nor IgG from normotensive pregnant women induced proteinuria in antibody-injected nonpregnant mice. Consistent with the lack of antibody-induced proteinuria was the absence of antibody-induced renal histopathology (Fig. 4d). Specifically, the glomeruli from nonpregnant mice injected with IgG from women with pre-eclampsia showed no evidence of hypercellularity or endotheliosis and appeared normal in both size and shape (Fig. 4d). This is in contrast to the renal histopathological changes observed in pregnant mice injected with the IgG from women with pre-eclampsia (Fig. 3). These findings indicate that whereas IgG from women with pre-eclampsia is capable of inducing hypertension in nonpregnant mice, it is unable to induce proteinuria and glomerular endotheliosis at this antibody concentration, suggesting that AT1-AA–induced sFlt1 production from the placenta contributes to the renal damage and proteinuria observed in antibody-injected pregnant mice.
It is a widely held view that the maternal features of pre-eclampsia are secondary to placental ischemia2,26 and vascular damage resulting in a local and systemic inflammatory response. From this perspective, researchers developed a rat model of pre-eclampsia on the basis of experimentally induced placental ischemia resulting from reduced uterine perfusion pressure (RUPP)7,27–30 and found that tumor necrosis factor-α levels, along with those of other inflammatory markers, were elevated31,32. Such experimentally manipulated pregnant rats developed hypertension, proteinuria and other features of pre-eclampsia, including elevated sFlt133,34 and the production of AT1-AAs32,34. Sera from RUPP-manipulated pregnant rats stimulated increased endothelin-1 synthesis by endothelial cells through AT1 receptor activation7, a finding consistent with the presence of AT1-AAs. RUPP-induced hypertension was markedly attenuated by antagonism of the AT1 receptor35 or the endothelin receptor type A32, suggesting that RUPP-induced hypertension relied on AT1-AA–induced endothelin-1 synthesis. Collectively, these results indicate that placental ischemia and the associated vascular damage and inflammatory response may serve as key stimuli for AT1-AA production during pregnancy and that AT1 receptor activation has a major role in the hypertension produced by placental ischemia34.
Angiotensin II is a well recognized hypertensive agent that mediates its physiological effects primarily through AT1 receptor activation. Because AT1-AAs also mediate their physiological effects through AT1 receptor activation, we expected these autoantibodies to stimulate an increase in blood pressure in the absence of elevated sFlt1. Thus, it came as no surprise that these autoantibodies could induce hypertension in nonpregnant mice. Because both AT1-AAs and sFlt1 can stimulate a rise in blood pressure, it is possible that each contributes to hypertension associated with pre-eclampsia.
Proteinuria and renal lesions, especially glomerular endotheliosis, are characteristic features of pre-eclampsia and occur in rat models of pre-eclampsia induced by elevated sFlt1 concentrations24. Because sFlt1 is induced by AT1-AAs in pregnant mice, it is possible that proteinuria and glomerular endotheliosis in antibody-injected mice are due to AT1-AA–mediated induction of sFlt1. Consistent with this possibility is the fact that IgG from women with pre-eclampsia readily induced hypertension in nonpregnant mice under conditions in which antibody-induced proteinuria and renal histopathological changes were lacking. In addition, AT1-AA–mediated sFlt1 induction is absent in nonpregnant mice (Fig. 4c) because the major source of sFlt1 (that is, the placenta) is missing15,23. It is likely that autoantibody-induced renal damage and proteinuria in pregnant mice are mediated through the induction of sFlt1.
It is possible that the increased occurrence of pre-eclampsia among women who are pregnant for the first time36 may result from an increased risk of placental ischemia in first pregnancies, owing to the small size of the spiral arteries in the uterus of such women. After these women become pregnant for the first time, the spiral artery diameters increase four to five times as they become flaccid and nonmuscular in comparison to those before pregnancy37. After pregnancy, the uterine spiral arteries do not return to their original small size but instead remain significantly larger37. Permanent enlargement of the uterine spiral arteries allows for enhanced placental perfusion in subsequent pregnancies36,37 and in this way may contribute to a reduced risk of ischemia, the resulting inflammatory response and the production of AT1-AAs. In pre-eclampsia, the decline of the inflammatory state that occurs at childbirth may deprive autoreactive lymphocytes of key cytokines required for continued AT1-AA production. Because AT1-AAs are predominantly of the IgG3 subclass38, with a relatively short 7-d half-life16, their levels decline substantially during the first few days after childbirth.
The antibody injection model of pre-eclampsia described here provides strong experimental support for our working hypothesis that pre-eclampsia is an autoimmune disease in which angiotensin receptor–activating autoantibodies contribute to many features of the disease. The biological properties of these autoantibodies can be blocked by a seven–amino acid epitope peptide that corresponds to a specific epitope associated with the second extracellular loop of the AT1 receptor. This fact suggests a common immunological origin for these autoantibodies in different individuals and has profound therapeutic potential. Available evidence suggests that AT1-AAs can be detected as early in pregnancy as 18 weeks, making them one of the earliest markers to identify women at risk for pre-eclampsia39. The fact that AT1-AAs can be detected many weeks before the symptoms of pre-eclampsia has far reaching implications regarding presymptomatic screening, disease diagnosis and treatment. If maternal circulating AT1-AAs contribute to pre-eclampsia, as our adoptive transfer animal model suggests, the timely identification and removal or inhibition of these autoantibodies from women with pre-eclampsia may provide considerable therapeutic benefit. If AT1-AAs play a major part in the etiology and pathophysiology of pre-eclampsia, as we hypothesize, it may one day be possible to block autoantibody-mediated AT1 receptor activation and thereby forestall or prevent the onset of the symptoms of pre-eclampsia.
We purchased cell culture medium RPMI 1640, antimycotics, antibiotics and FBS from Invitrogen. Losartan was a generous gift from Merck.
Pregnant human subjects with pre-eclampsia who were admitted to Memorial Hermann Hospital were identified by the obstetrics faculty of the University of Texas Medical School at Houston. The subjects were diagnosed with severe pre-eclampsia on the basis of the definition set by the National High Blood Pressure Education Program Working Group report. These criteria included the presence of a high blood pressure of ≥160/110 mm Hg and the presence of protein in urine of ≥3.5 g in a 24-h period or a random urinary protein/creatinine ratio of >0.3. These women had no previous history of hypertension. Other criteria included the presence of persistent headache, visual disturbances, epigastric pain or the HELLP syndrome in women with a blood pressure of ≥140/90 mm Hg. Normotensive pregnant individuals were characterized by uncomplicated pregnancies with normal-term deliveries. The clinical characteristics of the study subjects are summarized in Supplementary Table 2 online. We centrifuged blood samples at 18,000g for 10 min and stored the serum samples at −80 °C. The research protocol, which included informed consent from the subjects, was approved by the University of Texas Health Science Center at Houston Committee for the Protection of Human Subjects.
We isolated the IgG fraction from human or mouse sera as previously described20 with GammaBind G Sepharose (Amersham Biosciences).
We used a pGEX-4T-1 GST expression vector (Promega) expressing GST and an in-frame DNA insertion encoding the 27 amino acids corresponding to the second extracellular loop of human AT1 receptor (accession code: NM_009585.2). Specifically, we mixed equal amounts of two complementary oligonucleotides corresponding to the second extracellular loop of human AT1 receptor flanked by EcoRI restriction enzyme sites, boiled them for 5 min and then cooled to 23 °C. We cut the annealed double-stranded DNA with EcoRI and subcloned it into the EcoRI polycloning site of the pGEX-4T-1 vector. We confirmed the correct sequence by DNA sequencing. We performed the expression and purification of the GST–27-aa fusion protein following the protocol provided by the vendor (Amersham Bioscience). The purified GST-peptide fusion proteins were either immediately used or stored at −80 °C.
We performed affinity purification of AT1-AAs with the MicroLink Protein Coupling Kit (PIERCE). Briefly, we coupled the GST-peptide fusion protein (200 µg) purified as above to an AminoLink Plus Coupling Gel column (PIERCE) for 4 h or overnight at 4 °C according to the protocol provided by the vendor. We loaded IgG fractions (400 µl) that had been purified by GammaBind G Sepharose chromatography from sera of donors with pre-eclampsia or normotensive donors on the columns and incubated for 3 h. We collected the flow-through by centrifugation at 1,000g for 1 min. We eluted bound IgG by centrifugation at 1,000g for 1 min after incubation with 100 µl 100 mM glycine (pH 2.7) for 10 min. We neutralized the low pH of the eluted fraction by adding 5 µl of 1 M Tris, pH 9.0. We performed all steps at 4 °C. The elution fraction (affinity-purified AT1-AAs) was either used immediately or stored at −80 °C.
For the neutralization experiments, we used a synthetic peptide corresponding to a sequence on the second extracellular loop of the human AT1 receptor, AFHYESQ. We purchased this peptide from Baylor College of Medicine.
We used C57BL/6J pregnant or non-pregnant mice (18–22 g; Harlan) in our study. All animal studies were reviewed and approved by the Animal Welfare Committee, University of Texas Houston Health Science Center. Mice were anesthetized with sodium pentobarbital (50 mg kg−1 intraperitoneally). We concentrated total IgG (~800 µg) or affinity-purified AT1-AAs (~20 µg) from 200 µl donor sera by lyophilization and resuspended them with ~200 µl saline and then introduced them into either 13-d pregnant mice or nonpregnant mice by retro-orbital injection. Some mice also received losartan (0.24 mg) or a seven–amino acid peptide corresponding to an epitope on the second extracellular loop of the AT1 receptor (1.5 mg). We collected plasma on gestation days 13 through 18 for measurement of sFlt1 concentration. We measured mouse systolic blood pressure by tail-cuff (AD Instruments) at each of these time points17,40. We collected mouse urine at gestation day 18 at the time the dams were killed. We quantified urinary albumin by ELISA (Exocell) and measured urinary creatinine by a picric acid colorimetric assay kit (Exocell). We used the ratio of urinary albumin to urinary creatinine as an index of urinary protein24.
We analyzed human IgG in mouse circulation by western blotting as previously described20. Briefly, we ran mouse plasma on 10% SDS-PAGE gels and transferred the gels to a nitrocellulose membrane. We probed the membrane with horseradish peroxidase–conjugated mouse antibody to human IgG (1:5,000 dilution, Jackson ImmunoResearch Laboratories). We detected the signal with an ECL Kit (Amersham Bioscience). In addition, we evaluated affinity-purified AT1-AA binding to the AT1 receptor by western blotting analysis. Specifically, we ran cellular extracts from the 4× NFAT cis element–driven luciferase reporter cells (prepared as described below) on 10% SDS-PAGE gels and transferred the gels to a nitrocellulose membrane. We then incubated membranes with either commercially available antibody to the AT1 receptor (1:1,000 dilution, Santa Cruz), flow-through fraction antibody (1:100 dilution) or the elution fraction antibody (1:100 dilution) as the primary antibody. We used horseradish peroxidase–conjugated antibody to rabbit IgG to detect the commercial antibody to AT1 receptor or antibody to human IgG for the flow-through or elution fraction antibodies as the secondary antibody. We detected the signal with an ECL Kit (Amersham Bioscience).
We determined the concentrations of sFlt1 and sEng in mouse plasma with commercial kits (R&D Systems). We evaluated the concentration of human IgG in mouse plasma by collecting mouse plasma at various time points as indicated in Figure 1b, diluting the plasma 300-fold and quantifying human IgG abundance with a commercial ELISA kit (PIERCE).
We harvested placentas and kidneys of mice and fixed them in 4% formaldehyde overnight at 23 °C. Tissues were infiltrated and embedded in paraffin. We cut 4-µm serial sections and stained them with H&E by standard techniques24,40.
After killing the mice, we isolated kidneys24. We fixed tissue samples (1-mm3 in size) in 3% glutaraldehyde overnight. We then rinsed the samples in buffer, exposed them to 1% osmium tetroxide, dehydrated them and embedded them in araldite-epon mixture. We prepared semi-thin tissue sections (0.6-µm) and stained them with uranyl acetate and lead citrate and then examined them with a JEOL 1210 transmission electron microscope (JEOL Corporation).
We measured the biological activity of affinity-purified IgG (AT1-AAs) and IgG isolated from donor sera or mouse plasma (100 µl) with GammaBind G Sepharose as previously described21. Briefly, we plated Chinese hamster ovary cells (1 × 105 cells) containing stably integrated copies of a AT1 receptor minigene and a 4× NFAT-driven luciferase minigene on 24-well plates and grew them overnight. The next day, we changed the cells to serum-free medium and treated them with IgG (1:10 dilution) or affinity-purified AT1-AAs (1 µg) for 24 h. We measured luciferase activity in cell lysates with the luciferase assay kit (Promega)21.
All values are expressed as the mean ± s.e.m. We analyzed data for statistical significance with GraphPad Prism software (Graphpad). A value of P < 0.05 was considered significant.
We thank Merck for the gift of losartan. This work was supported by US National Institutes of Health grants HL076558 (to Y.X.) and HD34130, March of Dimes grant 6-FY06-323 and the Texas Higher Education Coordinating Board (to R.E.K.).
Note: Supplementary information is available on the Nature Medicine website.
AUTHOR CONTRIBUTIONSC.C.Z. performed total IgG purification; blood pressure, urinary protein, sFlt1 and sEng measurements in antibody-injected pregnant mice; fetal kidney and placenta isolation and cell culture bioassays for AT1-AA biological activity; and contributed to writing, editing and generation of figures. Y.Z. carried out antibody injections, affinity purification and characterization of AT1-AAs, western blot experiments and ELISA for human IgG; and contributed to writing and generation of figures. R.A.I. procured human subjects’ approvals, maintained the database of de-identified human subject information, measured blood pressure and urinary protein in antibody-injected nonpregnant mice, performed kidney and placental histology and contributed to writing and generation of figures. H.Z. provided expertise in kidney and placental histology and contributed to writing and editing of histological studies in both placenta and kidney. T.M. injected mice with antibody and helped with blood pressure and other measurements. E.J.P. provided expertise regarding the electron microscopic analysis of kidney. M.J.H. provided expertise in placental histology. S.M.R. provided expertise in pre-eclampsia, collected human subject samples amd analyzed human subject data. R.E.K. provided expertise in pre-eclampsia and the renin angiotensin system and contributed to the writing and editing of the manuscript. Y.X. was the principal investigator of the study, oversaw the overall design of experiments and interpretation of all results, organized the manuscript including text and figures, and did the final editing of the manuscript.