Autoantibody-mediated angiotensin receptor activation contributes to preeclampsia through tumor necrosis factor-alpha signaling

Abstract

Preeclampsia is a prevalent life-threatening hypertensive disorder of pregnancy for which the pathophysiology remains largely undefined. Recently, a circulating maternal autoantibody, the angiotensin II type I (AT(1)) receptor agonistic autoantibody (AA), has emerged as a contributor to disease features. Increased circulating maternal tumor necrosis factor alpha (TNF-alpha) is also associated with the disease; however, it is unknown whether this factor directly contributes to preeclamptic symptoms. Here we report that this autoantibody increases the proinflammatory cytokine TNF-alpha in the circulation of AT(1)-AA-injected pregnant mice but not in nonpregnant mice. Coinjection of AT(1)-AA with a TNF-alpha neutralizing antibody reduced cytokine availability in AT(1)-AA-injected pregnant mice. Moreover, TNF-alpha blockade in AT(1)-AA-injected pregnant mice significantly attenuated the key features of preeclampsia. Autoantibody-induced hypertension was reduced from 131+/-4 to 110+/-4 mm Hg, and proteinuria was reduced from 212+/-25 to 155+/-23 microg of albumin per milligram of creatinine (both P<0.05). Injection of AT(1)-AA increased the serum levels of circulating soluble fms-like tyrosine kinase 1 and soluble endoglin (34.1+/-5.1, 2.4+/-0.3 ng/mL, respectively) and coinjection with the TNF-alpha blocker significantly reduced their levels (21.7+/-3.4 and 1.2+/-0.4 ng/mL, respectively). Renal damage and placental abnormalities were also decreased by TNF-alpha blockade. Lastly, the elevated circulating TNF-alpha in preeclamptic patients is significantly correlated with the AT(1)-AA bioactivity in our patient cohort. Similarly, the autoantibody, through AT(1) receptor-mediated TNF-alpha induction, contributed to increased soluble fms-like tyrosine kinase 1, soluble endoglin secretion, and increased apoptosis in cultured human villous explants. Overall, AT(1)-AA is a novel candidate that induces TNF-alpha, a cytokine that may play an important pathogenic role in preeclampsia.

Figure 1. TNF-α is increased in AT₁-AA-injected pregnant mice

Serum TNF-α was elevated in PE-IgG injected pregnant mice but not in NT-IgG injected pregnant mice. Co-injection of losartan or 7-aa resulted in decreased serum TNF-α levels in PE-IgG injected pregnant mice. Non-pregnant animals injected with similarly purified human IgG fractions (white bars) did not demonstrate increased cytokine levels. n=9 for each variable. *P<0.05 versus NT-IgG. **P<0.05 versus PE-IgG.

Figure 2. TNF-α blockade reduces AT₁-AA-induced preeclamptic-like features

The key features of PE, hypertension (A) and proteinuria (B) present in the PE-IgG-injected pregnant mice were reduced by co-injection with a TNF-α blocker. In addition, sFlt-1 (C) and sEng (D) were also reduced with the co-injection of the autoantibody and the TNF-α blocker. n=9 for each variable. *P<0.05 versus NT-IgG. **P<0.05 versus PE-IgG.

Figure 3. Autoantibody-induced placental damage can be prevented by TNF-α neutralization

Placentas assessed by H&E staining (Panel A, 40X) indicate that PE-IgG injected mice had damaged placentas: calcifications (thin arrow) and fibrotic areas (thick arrow). Their labyrinth zones appear heterogeneous and have abnormal pools of blood (inset box). Placental apoptosis was assessed by TUNEL staining (Panel B, 10X, scale bar=1 mm). PE-IgG injected mice had increased apoptosis in their labyrinth zones as compared to NT-IgG injected animals. Quantification of the TUNEL assay (C) indicates a reduction in the TUNEL-positive cells in mice co-injected with PE-IgG and a TNF-α blocker as compared to the PE-IgG injected animals. n=18 placentas per variable, from 9 different mice in each group. Green; TUNEL-positive cells. Blue; DAPI-positive nuclei. Mice injected with NT-IgG or the anti-TNF-α antibody alone had unremarkable placentas. *P<0.05 versus NT-IgG. **P<0.05 versus PE-IgG.

Figure 4. Serum TNF-α positively correlates to AT₁-AA bioactivity

A positive correlation between the level of AT₁-AA bioactivity and serum TNF-α level was identified in preeclamptic women. Spearman’s rank correlation was used to determine an r-value (r=0.85, n=20, P<0.001).

Figure 5. TNF-α blockade prevents AT₁ receptor-mediated damage in human placental villous explants

Culturing human villous explants with PE-IgG resulted in TNF-α secretion (A). Co-culturing the explants with PE-IgG and losartan (5μM) or 7-aa (1μM) reduced the cytokine level. Apoptosis was increased in explants incubated with AT₁-AA and was partially diminished by blocking TNF-α activity as demonstrated by a TUNEL assay (B). Green; TUNEL-positive cells. Blue; DAPI-positive nuclei. 10X. Quantification of TUNEL staining (C) indicates that co-incubation with PE-IgG and an anti-TNF-α agent (5μg/ml) reduces the amount of apoptosis. Secretion of sFlt-1 (D) and sEng (E) were reduced by co-incubation of the autoantibody with an anti-TNF-α antibody. Six different placentas were collected, and from each, n=4 for every variable, total n=24 per variable. *P<0.05 versus NT-IgG. **P<0.05 versus PE-IgG.