Hypoxia-independent up-regulation of placental HIF-1α gene expression contributes to the pathogenesis of preeclampsia

Abstract

Accumulation of hypoxia inducible factor-1α (HIF-1α) is commonly an acute and beneficial response to hypoxia, while chronically elevated HIF-1α is associated with multiple disease conditions including preeclampsia (PE), a serious hypertensive disease of pregnancy. However, the molecular basis underlying the persistent elevation of placental HIF-1α in PE and its role in the pathogenesis of PE are poorly understood. Here we report that Hif-1α mRNA and HIF-1± protein were elevated in the placentas of pregnant mice infused with angiotensin II type I receptor agonistic autoantibody (AT₁-AA), a pathogenic factor in PE. Knockdown of placental Hif-1α mRNA by specific siRNA significantly attenuated hallmark features of PE induced by AT₁-AA in pregnant mice including hypertension, proteinuria, kidney damage, impaired placental vasculature, and elevated maternal circulating soluble fms-like tyrosine kinase-1 (sFlt-1) levels. Next, we discovered that Hif-1α mRNA levels and HIF-1± protein levels were induced in an independent PE model with infusion of the inflammatory cytokine LIGHT (tumor necrosis factor superfamily member 14). SiRNA knockdown experiments also demonstrated that elevated HIF-1α contributed to LIGHT-induced PE features. Translational studies with human placentas showed that AT₁-AA or LIGHT is capable of inducing HIF-1α in a hypoxia-independent manner. Moreover, increased HIF-1α was found to be responsible for AT₁-AA or LIGHT-induced elevation of Flt-1 gene expression and production of sFlt-1 in human villous explants. Overall, we demonstrated that hypoxia-independent stimulation of HIF-1α gene expression in the placenta is a common pathogenic mechanism promoting disease progression. Our findings reveal new insight to PE and highlight novel therapeutic possibilities for the disease.

Introduction

Preeclampsia (PE) is a life threating hypertensive complication of pregnancy and is a leading causes of maternal and neonatal morbidity and mortality^{1, 2}. Despite intensive research efforts and several large clinical trials, current strategies for managing PE remain inadequate and are limited to symptomatic therapy or the termination of pregnancy. Thus, uncovering novel factors and signaling pathways that contribute to the pathogenesis of PE are needed for the establishment of mechanism-based preventative and therapeutic strategies to improve the prognosis of the disease.

Hypoxia-inducible factor-1 (HIF-1) is a key transcription factor that plays a central role in the cellular response to low oxygen tension under physiologic and pathologic conditions^{3, 4}. HIF-1α is a heterodimer consisting of two subunits, α and β. Although HIF-1β is constitutively expressed, HIF-1α levels are precisely regulated by post-translational modification depending on oxygen tension. HIF-1α is rapidly degraded under normoxic conditions, but quickly stabilized when oxygen availability is reduced. Thus, hypoxia-induced HIF-1α is usually transient and brief at the protein level⁴. A variety of studies have shown that women with PE are characterized by persistently elevated placental HIF-1α levels that promote enhanced transcription of genes encoding soluble fms-like tyrosine kinase-1 (sFlt-1), soluble endoglin (sEng) and endothelin-1 (ET-1), all known to contribute to PE^5–7. However, the molecular basis underlying prolonged elevated placental HIF-1α in PE and the pathological role of sustained elevated HIF-1α in PE are largely unknown.

Numerous recent studies have shown that HIF-1α levels can be regulated by means that are independent of hypoxia⁸. For example, angiotensin II and the inflammatory cytokines TNF and IL-6 induce HIF-1α gene expression in vascular smooth muscle cells, kidney cells and hepatocytes, respectively^9–11. Multiple studies have revealed that inflammatory cytokines and autoantibodies are elevated in PE patients and contribute to pathophysiology PE^12–14. For example, early studies showed that injection or infusion of pathogenic autoantibodies such as the angiotensin II type 1 receptor agonistic autoantibody (AT₁-AA) or the inflammatory cytokine LIGHT into pregnant mice results in features of PE including hypertension, proteinuria, placental abnormalities, and increased circulating soluble fms-like tyrosine kinase-1 (sFlt-1), soluble endoglin, and endothelin-1^14–16. Thus, we hypothesized that the pathogenic autoantibody, AT₁-AA, and the inflammatory cytokine, LIGHT, stimulate placental HIF-1α production and in this way contribute to features of PE. Here we conducted both mouse and human studies to assess these hypotheses.

Methods

For detailed descriptions, refer to the Methods section in the online-only Data Supplement.

Results

Increased placental HIF-1α contributes to the development of PE features in an autoantibody-injection model of PE in pregnant mice

In order to examine a potential role of elevated HIF-1α in PE, we took advantage of an experimental model of PE in mice induced by the injection of patient-derived-IgG (PE-IgG) known to contain the pathogenic autoantibodies, AT₁-AA¹⁷. We found that Hif-1α gene expression was induced significantly in the placentas of mice injected with PE-IgG compared to the pregnant mice injected with IgG from normotensive pregnant women (NT-IgG) (Figure 1A). In contrast, no significant difference was observed in the kidneys between PE-IgG and NT-IgG-injected pregnant mice (Figure S1). We also confirmed that PE-IgG induced the elevation of placental HIF-1α expression at the protein level by immunoblotting (Figure 1B). Additionally, immunohistochemical analysis revealed that the PE-IgG induced elevation of HIF-1α protein expression throughout the placenta (Figure 1C and Figure 1D). In addition, the PE-IgG-induced elevation of placental HIF-1α expression was almost completely inhibited when PE-IgG was coinjected with losartan, an angiotensin II type 1 receptor (AT₁R) blocker, or with the autoantibody-neutralizing 7 amino acid epitope peptide (Figure 1A, 1B, 1C, and 1D). These results indicate that the elevation of placental HIF-1α resulting from injection PE-IgG was due to the activation of AT₁Rs by AT₁-AA.

Figure 1. HIF-1α is increased in placentas of AT₁-AA-treated mice via AT₁R activation.

Pregnant mice were injected with purified IgG from normotensive pregnant women (NT-IgG) or PE patients (PE-IgG) on E13.5 and E14.5. Losartan (Los), an angiotensin II receptor antagonist, or autoantibody-neutralizing 7 amino acid epitope peptide (7aa) were in some cases co-injected with PE-IgG. Samples were collected on E18.5.

(A) Placental Hif-α mRNA expression was quantified by real-time RT-PCR. (NT-IgG; n=6 mice, PE-IgG; n=7, PE-IgG+Los; n=4, PE-IgG+7aa; n=4), (**P<0.01 vs NT-IgG, ^##P<0.01 vs PE-IgG)

(B) Increase in placental HIF-1α protein level in mice injected with PE-IgG was detected by immunoblotting.

(C) HIF-1α protein in mouse placentas detected by immunohistochemistry (IHC). Arrows indicate the cells with positive staining in the nucleus. Scale bar, upper; 500µm, lower; 50µm. (Sp; spongiotrophoblast zone Lb; labyrinth zone)

(D) The positive staining for HIF-1α was quantified. (n=6 fields per placenta under ×100 magnification; 4 mice per group), (**P<0.01 vs NT-IgG, ^#P<0.05, ^##P<0.01 vs PE-IgG)

Global HIF-1α-deficient mice die in midgestation from cardiac and vascular malformation¹⁸. This embryonic lethality makes it difficult to examine the in vivo role of HIF-1α. To determine the pathophysiologic significance for PE-IgG-induced placental HIF-1α expression we conducted siRNA-induced in vivo knockdown of Hif-1α mRNA. Briefly, siRNA-encapsulated nanoparticles were injected into pregnant mice on E13.5 and E14.5, together with PE-IgG to specifically knockdown Hif-1α mRNA levels. As shown in Figure 2A, placental Hif-1α mRNA levels were successfully down-regulated in Hif-1α siRNA-injected mice compared with control scrambled siRNA-injected mice. We also confirmed the reduction of HIF-1α protein expression levels in the placentas of Hif-1α siRNA-injected mice by immunoblotting (Figure S2). As a result of in vivo knockdown of Hif-1α mRNA, we found that PE-IgG-induced diagnostic features of PE, hypertension and proteinuria, were significantly reduced compared with control siRNA-injected mice (Figure 2B and Figure 2C). Histologic analysis of mouse kidneys revealed that PE-IgG-induced pathologic changes seen in the glomeruli of control siRNA-injected mice (i.e., swollen glomeruli with narrowed capillary and Bowman’s spaces) were attenuated in the kidneys of Hif-1α siRNA-injected mice (Figure 2D). Additionally, we found that placentas of Hif-1α siRNA-injected mice displayed significantly less tissue damage including placental calcifications, a hallmark of placental distress observed in placentas of PE patients, as compared with those of control siRNA-injected mice (Figure S3). Moreover, we examined placental vasculature using CD31 staining. As a result, PE-IgG-induced disorganized and impaired vasculature in the labyrinthine zone of control siRNA-injected mice (low density of CD31-positive vessels and narrowed capillary spaces) was attenuated in the placentas of Hif-1α siRNA-injected mice (Figure 2E).

Fig. 2. In vivo specific knockdown of Hif-1α mRNA prevents the development of PE features induced by AT₁-AA.

In vivo siRNA-mediated Hif-1α mRNA knockdown was performed.

(A) The levels of Hif-1α mRNA in mouse placentas were quantified by real-time RT-PCR. (n=6 mice per group), (**P<0.01 vs NT-IgG, ^##P<0.01 vs PE-IgG+control siRNA)

(B) Blood pressure was measured by tail-cuff plethysmography on a daily base. (n=6 mice per group), (*P<0.05, **P<0.01 vs NT-IgG, ^##P<0.01 vs PE-IgG+control siRNA at the same time points)

(C) Proteinuria was determined as urine albumin to creatinine ratio by ELISA. (n=6 mice per group), (*P<0.05 vs NT-IgG, ^##P<0.01 vs PE-IgG+control siRNA)

(D) Renal histology assessed by PAS staining. Pathologic changes in kidneys of PE-IgG-treated WT mice (swollen glomeruli with narrowed capillary and Bowman’s spaces) were suppressed by siRNA knockdown of Hif-1α mRNA. Scale bar, 100µm.

(E) The staining of CD31, an endothelial cell marker, by immunohistochemistry (IHC) / immunofluorescence (IF). Scale bar, 200µm.

(F) The levels of Flt-1 mRNA in mouse placentas were quantified by real-time RT-PCR. (n=6 mice per group), (**P<0.01 vs NT-IgG, ^#P<0.05 vs PE-IgG+control siRNA)

(G) The levels of circulating sFlt-1 in mouse plasma were determined by ELISA. (n=6 mice per group), (*P<0.05 vs NT-IgG, ^#P<0.05 vs PE-IgG+control siRNA)

The Flt-1 gene is a direct transcriptional target of HIF-1α¹⁹. A splice variant encodes soluble Flt-1 (sFlt-1), an antiangiogenic factor secreted by the placenta into the maternal circulation, that is believed to contribute to the development of systemic endothelial dysfunction, hypertension and multi-organ damage, including the kidneys in PE patients²⁰. As such, we also found that Hif-1α mRNA knockdown in vivo suppressed the PE-IgG-induced elevation of Flt-1 mRNA in the placenta, as well as circulating sFlt-1 levels (Figure 2F and Figure 2G). These results provide in vivo evidence that the induction of HIF-1α in the placenta contributes to the development of pathogenic autoantibody-induced features of PE and is also involved in the increased sFlt-1 production.

Elevated HIF-1α contributes to the development of LIGHT-induced PE features

Emerging evidence indicates that an increased inflammatory response is involved in PE^{12, 13}. Supporting this concept, a recent study showed that a member of the TNF superfamily, LIGHT, is elevated in the circulation and placentas of PE patients and that the injection of LIGHT into pregnant mice induces features of PE including the overproduction of sFlt-1¹⁵. The following experiments were conducted to determine whether elevated HIF-1α contributes to LIGHT-induced PE features in pregnant mice. We found that LIGHT injection into pregnant mice resulted in increased levels of Hif-1α mRNA in placentas (Figure 3A) but not in kidneys (Figure S1). Next, we found that LIGHT-induced placental Hif-1α mRNA levels were significantly reduced by neutralizing antibodies specific for LIGHT receptors: lymphotoxin β receptor (LTβR) and herpes virus entry mediator (HVEM) (Figure 3A). These results indicated that LIGHT signaling via its receptors induced placental Hif-1α gene expression.

Fig. 3. Specific knockdown of mRNA in vivo suppresses the features of PE induced by LIGHT.

Pregnant mice were injected with saline or LIGHT with or without neutralizing antibodies against lymphotoxin β receptor (anti-LTβR) or herpes virus entry mediator (anti-HVEM) on E13.5 and E14.5. Additionally, in vivo siRNA-mediated Hif-1α mRNA knock down was conducted. Samples were collected on E18.5.

(A) Placental Hif-1α mRNA levels were quantified using real-time RT-PCR. (n=4–7 mice per group), (**P<0.01 vs saline, ^#P<0.05 vs LIGHT, ^††P<0.01 vs LIGHT+control siRNA)

(B) Blood pressure was measured by tail-cuff plethysmography on a daily base. (n=5–6 mice per group), (*P<0.05, **P<0.01 vs saline, ^##P<0.01 vs LIGHT+control siRNA at the same time points)

(C) Proteinuria was determined as urine albumin to creatinine ratio by ELISA. (n=5–6 mice per group), (*P<0.05 vs saline, ^##P<0.01 vs LIGHT+control siRNA)

(D) Placental Flt-1 mRNA levels were quantified using real-time RT-PCR. (n=5–6 mice per group), (**P<0.01 vs saline, ^#P<0.05 vs LIGHT + control siRNA)

(E) The levels of circulating sFlt-1 in mouse plasma were determined by ELISA. (n=6–7 mice per group), (**P<0.01 vs saline, ^#P<0.05 vs LIGHT+control siRNA)

To assess whether elevated HIF-1α in the placenta plays a detrimental role in the LIGHT-induced PE development as it does in PE-IgG-injected mice, we conducted in vivo knock-down of Hif-1α mRNA by injecting Hif-1α siRNA-encapsulated nanoparticles, together with LIGHT, into pregnant mice. Hif-1α siRNA injection successfully reduced the levels of placental Hif-1α mRNA compared with those of control siRNA-injected mice (Figure 3A). Moreover, preeclamptic features induced by LIGHT injection (hypertension and proteinuria) were attenuated significantly in Hif-1α siRNA-injected pregnant mice compared with control siRNA-injected pregnant mice (Figure 3B and Figure 3C). Additionally we found that Hif-1α mRNA knockdown attenuated the PE-IgG-induced elevation of Flt-1 mRNA levels, as well as circulating sFlt-1 protein (Figure 3D and Figure 3E). These findings indicate that placental elevated HIF-1α contributes to the development of PE features in the LIGHT-induced experimental model of PE.

HIF-1α is elevated in placentas of PE patients and AT₁-AA or LIGHT directly induces HIF-1α expression in cultured human placental villous explants independent of hypoxia

To extend our mouse findings to humans, we determined that the expression of HIF-1α was elevated in placentas of PE patients at both mRNA and protein levels compared to those of normotensive pregnant women (Figure 4A, 4B, and 4C). To determine whether AT₁-AA or LIGHT can directly induce HIF-1α gene expression in the human placenta independent of hypoxia, we used primary human placental villous explants isolated from normotensive pregnant women. We cultured human villous explants treated with PE-IgG, NT-IgG or LIGHT under ambient oxygen levels. We found that PE-IgG significantly induced HIF-1α mRNA levels compared to the NT-IgG-treated human villous explants and the induction was significantly reduced by co-treatment with losartan to inhibit AT₁R activation or 7 amino acid epitope peptide to neutralize AT¹-AA (Figure 4D). Similarly, we found that treatment of cultured villous explants with LIGHT resulted in increased HIF-1α mRNA levels compared to the controls (Figure 4E). Additionally, we also confirmed that PE-IgG or LIGHT induced the elevation of HIF-1α protein (Figure 4F). Thus, these results indicate that AT₁-AA and LIGHT are capable of directly inducing HIF-1α gene expression in cultured human villous explants independent of hypoxia.

Fig. 4. AT₁-AA or LIGHT is capable of directly inducing HIF-1α gene expression in human villous explants independent of hypoxia.

(A) Placental HIF-1α mRNA levels, determined by real-time RT-PCR, were increased in preeclampsia patients (PE) compared with normotensive pregnant women (NT). (n=10 per group), (*P<0.05 vs NT)

(B) The elevated levels of HIF-1α protein in placentas of PE patients is detected by immunohistochemistry. Scale bar, 200µm.

(C) The positive staining for HIF-1α was quantified. (n=6 fields per section under ×100 magnification; NT; n=4, PE; n=5), (*P<0.05 vs NT)

(D) Human villous explants were treated with 100µg/ml NT-IgG or PE-IgG for 24 hours in the presence or absence of 5uM Losartan (Los) or 1uM autoantibody-neutralizing 7 amino acid epitope peptide (7aa). HIF-1α mRNA levels were quantified using real-time RT-PCR. (n=3 independent experiments), (**P<0.01 vs NT-IgG, ^##P<0.01 vs PE-IgG)

(E) HIF-1α mRNA levels in cultured human villous explants treated with 100pg/ml LIGHT for 24 hours were quantified using real-time RT-PCR. (n=3 independent experiments), (**P<0.01 vs saline injected mice)

(F) HIF-1α protein levels following treatment with IgG or LIGHT as described in D or E was assessed by immunoblotting. (NS: non-specific bands)

AT₁-AA or LIGHT-induced HIF-1α promotes FLT-1 gene expression and subsequent sFlt-1 secretion in human villous explants independent of hypoxia

We examined whether PE-IgG or LIGHT-induced HIF-1α is capable of promoting Flt-1 gene expression and subsequent sFlt-1 production in human placentas independent of hypoxia. To test this possibility, we treated human villous explants under ambient oxygen levels with PE-IgG or LIGHT in the presence or absence of CAY10585, a specific HIF-1α inhibitor. First, we confirmed that the elevation of HIF-1α induced by PE-IgG or LIGHT was suppressed by the treatment of CAY10585 (Figure 5A). We found that Flt-1 mRNA levels and amount of secreted sFlt-1 were increased by the treatment of human villous explants with PE-IgG or LIGHT (Figure 5B and Figure 5C). In contrast, treatment with a HIF-1α inhibitor, CAY10585, significantly reduced PE-IgG or LIGHT-induced Flt-1 gene expression and sFlt-1 production (Figure 5B and Figure 5C). To further validate the role of HIF-1α for Flt-1 gene induction and subsequent sFlt-1 production induced by PE-IgG or LIGHT, we conducted HIF-1α mRNA knockdown in human villous explants. We confirmed that the elevation of HIF-1α induced by PE-IgG or LIGHT was successfully down-regulated by the treatment of HIF-1α siRNA compared with control scrambled siRNA-treated group (Figure 5D). The increase in Flt-1 mRNA levels and the amount of secreted sFlt-1 induced by PE-IgG or LIGHT were significantly suppressed by the knockdown of HIF-1α mRNA in human villous explants (Figure 5E and Figure 5F). These results indicate that PE-IgG or LIGHT directly induces FLT-1 gene expression in a HIF-1α-dependent manner in cultured human villous explants independent of hypoxia.

Fig. 5. HIF-1α is responsible for AT₁-AA or LIGHT-induced elevation of Flt-1 gene expressioin and increased production of sFlt-1 in human villous explants independent of hypoxia.

(A) HIF-1α protein levels in human villous explants detected by immunoblotting. Human villous explants were pretreated with or without 10µM CAY10585 for 15 min and then treated with 100µg/ml NT-IgG or PE-IgG or 100pg/ml LIGHT for 24 hours. Explants were also treated with 10µM dimethyloxaloylglycine (DMOG), a prolyl hydroxylase (PHD) inhibitor for 24 hours as a positive control.

(B) FLT-1 mRNA levels were quantified using real-time RT-PCR. (n=4 independent experiments), (**P<0.01 vs NT-IgG or PBS-treated group, ^#P<0.05, ^##P<0.01 vs PE-IgG or LIGHT-treated group)

(C) Secreted sFLT-1 protein levels in the culture media were determined by ELISA. (n=5 independent experiments), (**P<0.01 vs NT-IgG or PBS-treated group, ^#P<0.05 vs PE-IgG or LIGHT-treated group)

(D) Knockdown of HIF-1α in human villous explants. The explants were treated with nanoparticle encapsulated control (con) -or HIF-1α-siRNA for 24 hours and then were treated with NT- or PE-IgG (100µg /ml) or LIGHT (100 pg/mL) for 24 hours. HIF-1α protein was assessed by immunoblotting.

(E) FLT-1 mRNA levels of human villous explants were quantified using real-time RT-PCR. (n=4 independent experiments), (**P<0.01 vs NT-IgG or PBS+control siRNA, ^#P<0.05, ^##P<0.01 vs PE-IgG or LIGHT+ control siRNA group)

(F) Secreted sFLT-1 protein levels in the culture media were determined by ELISA. (n=4 independent experiments), (**P<0.01 vs NT-IgG or PBS+control siRNA, ^##P<0.01 vs PE-IgG or LIGHT+ control siRNA group)

(G) Working model of our study. AT₁-AA or LIGHT-induced HIF-1α production in placental trophoblasts is followed by HIF-1α-mediated induction of FLT-1 gene expression and subsequent sFLT-1 protein production independent of hypoxia. Persistent elevation of HIF-1α causes chronic overproduction of sFLT-1 and contributes to impaired placental vascular development, maternal endothelial dysfunction and disease progression.

Discussion

Here we report that Hif-1α mRNA and HIF-1α protein levels were elevated in the placentas of two independent animal models of PE, based on the injection with AT₁-AA or LIGHT. We also showed that specific siRNA knockdown of Hif-1α mRNA attenuated hallmark features of PE including hypertension, proteinuria, kidney damage, impaired placental vasculature, and maternal elevated circulating sFlt-1 in both PE mouse models. These results indicate that increased HIF-1α gene expression is a common pathogenic factor contributing to PE. Extending animal studies to humans, we confirmed that HIF-1α mRNA and HIF-1α protein levels were elevated in the placentas of PE patients. Using human villous explant cultures under non-hypoxic conditions, we showed that AT₁-AA or LIGHT-induced HIF-1α mRNA and HIF-1α protein levels resulting in elevated FLT-1 mRNA levels and increased sFlt-1 secretion in a hypoxia-independent manner. Overall, we provide both in vivo animal studies and in vitro human evidence showing the pathogenic role of elevated HIF-1α gene expression in PE and hypoxia-independent mechanisms underlying its elevation in the placentas.

Numerous early studies showed that HIF-1α can be induced by non-hypoxic stimuli in various cell types⁶. For example, studies in vascular smooth muscle cells showed that angiotensin II (Ang II) stimulates HIF-1α production by a protein kinase C-mediated transcriptional activation of the HIF-1α gene expression and by a reactive oxygen species-dependent mechanism leading to enhanced translation of HIF-1α mRNA^{10, 21}. In this case, Ang II induces HIF-1α protein leading to an increase in vascular endothelial growth factor gene expression. Other studies have shown that Ang II induces HIF-1α mRNA production in renal glomerular cells²². Women with PE harbor autoantibodies (AT₁-AA) that mimic the action of Ang II, and activate the major angiotensin receptor, AT₁R. These pathogenic autoantibodies may serve as hypoxia-independent factors leading to the increased production of HIF-1α observed in placentas of women with PE. Supporting our hypothesis, we have provided human in vitro studies showing that purified PE-IgG, which contains AT₁-AA, activates AT₁Rs resulting in the induction of HIF-1α gene expression under non-hypoxic conditions. Additionally, we demonstrated that HIF-1α expression is induced significantly in the placentas of mice injected with PE-IgG. We realize that the PE-IgG used in our experiment is a complex mixture of immunoglobulins that is likely to contain other autoantibodies. However, we have shown in prior publications that the same effects are seen when AT₁-AA is specifically purified by affinity chromatography¹⁷. Furthermore, the PE-IgG or affinity purified AT₁-AA-induced PE features are blocked by losartan or the 7aa epitope peptide, indicating that the effects are mediated by interaction with the specific epitope on the AT₁ receptor. Our results are also in good agreement with those Wenzel et al. who showed that pregnant rats infused with a rabbit antibody that activates AT₁Rs resulted in elevated HIF-1α in rat placentas²³. Our current study has provided human and mouse evidence that the activation of AT₁Rs by AT₁-AA induces HIF-1α gene expression in the placenta, and thereby contributing to the development of PE.

A growing body of studies indicate that PE is characterized by increased circulating levels of proinflammatory cytokines such as TNF-α, IL-1β, IL-6, IL-17 and LIGHT^{12, 13, 15}. A pathologic role for these cytokines is supported by experimental evidence showing that infusion of these cytokines into pregnant rodents produces features of PE, including the production of AT₁-AA^{15, 16}. These results suggest that the elevation of inflammatory cytokines is an early event in the development of PE, and functions upstream of AT₁-AA production. Additional evidence shows that inflammatory cytokines induce HIF-1α under non-hypoxic conditions⁶. For example, TNF-α and IL-1β activate HIF-1α gene expression in human hepatoma cells²⁴. TNF-α increases HIF-1α mRNA levels as well as its target genes GLUT-1 and GLUT-3 in HEK293 cells⁹. IL-6 induces both HIF-1α mRNA and HIF-1α protein levels, and the expression of the HIF-1α target gene, erythropoietin, in hepatocytes¹¹. IL-1β induces HIF-1α-mediated VEGF secretion in trophoblast cells²⁵. Thus, a pathogenic role for HIF-1α in response to elevated inflammatory cytokines associated with PE may be an important contributor to disease pathogenesis. However, a role for HIF-1α in increased inflammatory cytokine-induced features in PE was not recognized prior to the results of our experiments reported here with the inflammatory cytokine, LIGHT. As a member of the TNF superfamily, LIGHT is known to be elevated and contributes to PE features in pregnant mice¹⁵. As with AT₁-AA, we have shown here that infusion of LIGHT stimulates production of HIF-1α in placentas of pregnant mice and in human villous explants independent of hypoxia. The increased levels of HIF-1α stimulate the transcriptional activation of the Flt-1 gene, thereby providing for excessive sFlt-1 production leading to the features of PE. Our data represent the first in vivo animal evidence showing the pathologic role of HIF-1α in inflammatory cytokine-induced PE pathophysiology.

Hypoxia is known to induce HIF-1α transiently, largely as a result of protein stabilization²⁶. However, how HIF-1α remains persistently elevated in the placenta in the setting of PE was previously unknown and the role of HIF-1α in PE remained unclear. Here we provide both in vivo animal and in vitro human evidence that AT₁-AA and LIGHT are two hypoxia-independent factors that stimulate Hif-1α gene expression resulting in subsequent HIF-1α-mediated activation of Flt-1 gene expression. Our studies support a working model (Figure 5G) in which AT₁-AA or LIGHT-induced HIF-1α expression in placental trophoblasts is followed by HIF-1α-mediated induction of Flt-1 gene expression and subsequent sFlt-1 production independent of hypoxia. Considerable evidence now indicates that elevated production of sFlt-1 plays a pathogenic role in PE²⁰. Thus, interfering with persistently elevated HIF-1α is likely to reduce the overproduction of sFlt-1 and slow the progression of the disease (Figure 5G).

Perspectives

In conclusion, our current studies have added significant new insight to the pathogenesis of PE by identifying the detrimental role of chronic elevated placental HIF-1α initially triggered by hypoxia-independent factors, a pathogenic autoantibody and an inflammatory cytokine. Chronically elevated placental HIF-1α promotes excessive sFlt-1 production and disease progression. Supporting this working model, we demonstrated that reducing elevated HIF-1α by siRNA-induced mRNA knockdown successfully halted HIF-1α-induced sFlt-1 production and prevented disease development in two independent mouse models of PE in vivo. Altogether, these findings suggest that HIF-1α suppression may serve as a target for pharmacological intervention for PE.

Novelty and significance.

What is new?

• HIF-1α gene expression and protein levels were induced in the placentas of two independent animal models of PE infused with AT₁-AA or LIGHT.

• In vivo knockdown of HIF-1α gene expression using siRNA attenuated hallmark features of PE including hypertension, proteinuria, kidney damage, impaired placental vasculature, and maternal elevated circulating sFlt-1 in both PE mouse models.

• Using human villous explant culture, we found that AT₁-AA or LIGHT directly induced HIF-1α gene expression and upregulated HIF-1α was responsible for AT₁-AA or LIGHT-induced elevation of Flt-1 gene expression and the production of sFlt-1 independent of hypoxia.

What is relevant?

Our current studies have provided new insight to the pathogenesis of PE by identifying the detrimental role of chronically elevated placental HIF-1α initially triggered by hypoxia-independent factors. Additionally, our discoveries indicate therapeutic possibilities targeting HIF-1α.

Summary

We have provided both mouse and human evidence that increased HIF-1α in the placenta plays a general pathologic role in the pathogenesis of PE induced by a pathogenic autoantibody or an inflammatory cytokine. Our findings highlight novel therapeutic possibilities for PE.