Heme Oxygenase-1 Is Not Decreased in Preeclamptic Placenta and Does Not Negatively Regulate Placental Soluble fms-Like Tyrosine Kinase-1 or Soluble Endoglin SecretionNovelty and Significance
Elevated placental release of the antiangiogenic factors, soluble fms-like tyrosine kinase-1 (sFlt-1) and soluble endoglin (sENG), is central to the pathophysiology of preeclampsia. It is widely accepted that heme oxygenase-1 (HO-1) is decreased in preeclamptic placenta and negatively regulates sFlt-1 and sENG production. We set out to verify these contentions. There was no difference in HO-1 mRNA or protein levels in preterm preeclamptic placentas (n=17) compared with gestationally matched controls (n=27). In silico analysis of microarray studies did not identify decreased placental HO-1 expression in preeclamptic placenta. Silencing HO-1 in primary trophoblasts did not affect sFlt-1 protein secretion after 24 or 48 hours. Silencing nuclear factor (erythroid-derived 2)-like 2 (transcription factor that upregulates HO-1) in trophoblasts also did not affect sFlt-1 secretion. Administering tin protoporphyrin IX dichloride (HO-1 inhibitor) or cobalt protoporphyrin (HO-1 inducer) into placental explants did not affect sFlt-1 or sENG secretion. Silencing HO-1 in 2 types of primary endothelial cells (human umbilical vein endothelial and uterine microvascular endothelial cells) significantly increased sFlt-1 secretion but not sENG secretion. However, HO-1 silencing selectively increased mRNA expression of sFlt-1 i13 (generically expressed sFlt-1 variant) but not of sFlt-1 e15a (sFlt-1 variant mainly expressed in placenta). Furthermore, adding tin protoporphyrin IX dichloride decreased sFlt-1, whereas adding HO-1 inducers (cobalt protoporphyrin, dimethyl fumarate, and rosiglitazone) either had no effect or increased sFlt-1 or sENG secretion (these trends are opposite to what is expected). We conclude that HO-1 expression is not decreased in preeclamptic placenta and HO-1 does not negatively regulate placental sFlt-1 and sENG secretion in placental or endothelial cells.
Preeclampsia complicates 2% to 7% of pregnancies,1 and accounts for 60 000 maternal deaths per year globally and far greater rates of fetal and neonatal losses.2 It is a multisystem disorder characterized by maternal endothelial dysfunction, leading to hypertension and maternal end-organ injury (kidneys, liver, hematological and neurological systems, and brain, leading to eclamptic seizures). Soluble factors released from the placenta contribute significantly to endothelial dysfunction and disease pathogenesis.
Central to the pathogenesis of preeclampsia is placental release of the antiangiogenic factors, soluble fms-like tyrosine kinase-1 (sFlt-1) and soluble endoglin (sENG), into the maternal circulation. They are mainly responsible for the maternal endothelial dysfunction and injury seen in clinical disease. Evidence for their important role in preeclampsia is strong.3–6 They are, by far, the most studied molecules in preeclampsia.
Heme oxygenase-1 (HO-1) is an inducible cytoprotective antioxidant enzyme, which catalyzes free heme into carbon monoxide, free iron, and biliverdin.7 It is upregulated by nuclear factor (erythroid-derived 2)-like 2 (Nrf2), a master regulator of the antioxidant response.8–10 When activated, Nrf2 translocates to the nucleus and upregulates antioxidant genes (including HO-1).
It was proposed that decreased placental HO-1 expression in preeclampsia is a key pathogenic step.11 The group proposed that HO-1 directly inhibits sFlt-1 and sENG release. Thus, lower HO-1 expression seen in preeclampsia is thought to increase sFlt-1 and sENG secretion. Decreased HO-1 leading to increased sFlt-1 and sENG production is now often cited as part of the pathophysiology of preeclampsia.2,12–21 A clinical extension of this concept is to identify drugs that induce HO-1 (given that this will decrease placental sFlt-1 and sENG secretion) as a therapeutic strategy to treat preeclampsia.22 For example, the rationale for a clinical trial examining whether pravastatin can be used to treat preterm preeclampsia was that preclinical data showed that statins upregulate HO-1 and decrease sFlt-1 production.13
The evidence that HO-1 expression is decreased in preeclamptic placenta seems limited to 2 reports. The first showed reduced HO-1 protein levels in 4 preeclamptic placentas and 4 controls.23 A more recent study compared placental HO-1 mRNA expression in 16 cases of preterm preeclampsia (32–37 weeks of gestation; n=16) and 16 control placentas obtained from term pregnancies (all >37 weeks).24 It is therefore possible that an increase in HO-1 expression with advancing gestational age may explain their findings rather than decreased HO-1 expression in preeclamptic placenta.
Regarding functional studies on molecular regulation, there has been one study providing experimental evidence in primary human placental and endothelial tissues to suggest that HO-1 negatively regulates sFlt-1 and sENG production.11 Given the wide acceptance of these findings, it is perhaps important that these observations are verified.
Therefore, we explored the hypothesis that HO-1 expression is decreased in preterm preeclampsia and that HO-1 negatively regulates sFlt-1 and sENG. Importantly, we used a large cohort of placentas collected at a single center and compared placental HO-1 expression (mRNA and protein) from cases of preterm preeclampsia against gestation-matched normotensive preterm controls. To examine whether HO-1 inhibits sFlt-1 and sENG secretion, we performed experiments in primary placental cytotrophoblasts, placental explants, and 2 types of primary human endothelial cells. Our data do not support the contention that there is decreased placental HO-1 expression in preeclampsia that then negatively regulates sFlt-1 or sENG.
A summary of the Methods are presented here. Detailed information is available in the online-only Data Supplement.
Ethical approval was obtained for this study from the Mercy Heath Human Research Ethics Committee (R11/34). For the observational studies, placentas were obtained from (1) pregnancies complicated by preterm preeclampsia with or without fetal growth restriction (delivered in <34 weeks of gestation for fetal or maternal indications) and (2) normotensive preterm pregnancies that delivered a fetus of normal birthweight (delivered in <34 weeks of gestation). Preeclampsia was defined according to the criteria published by the American College of Obstetricians and Gynecologist in 2013.25
Tissues were washed in sterile phosphate buffered saline, snap frozen, and stored at −80°C until analysis. Tissue for histology was fixed in 10% buffered formalin. For the functional studies, we collected umbilical cords and placentas from term pregnancies undergoing cesarean section to harvest primary trophoblasts, placental explants, and human umbilical vein endothelial cells (HUVECs). Primary uterine microvascular cells (UtMVECs) were purchased from Lonza Walkersville, Inc (Walkersville, MD).
Preparation of Tissues for Functional Assays
Primary cytotrophoblast isolation was performed by Percoll gradient centrifugation followed by CD9 negative selection as previously described.26 To prepare placental explants, small pieces of villous tissues were excised, washed, and then dissected into small explants of 10 to 15 mg in weight. After treatments, the wet weight was recorded before the explants were snap frozen for later analysis.
To prepare HUVECs, the cord vein was cannulated and infused with phosphate buffered saline to wash out fetal blood. Collagenase was infused into the cord, and the HUVECs were recovered by pelleting and resuspension. They were then plated. The UtMVECs were thawed and plated. Primary cytotrophoblasts and placental explants were cultured at 8% oxygen (physiological for placenta), whereas HUVECs and UtMVECs were cultured at 20% oxygen.
Summary of Doses Used for Functional Studies
For siRNA experiments (performed on HUVECs, UtMVECs, and primary trophoblast cells), 10 nmol/L of siRNAs or 10 nmol/L of a negative control was administered and left for 24 or 48 hours.
For the small molecule experiments, we treated tissues/cells in various experiments with the following doses: 10 μmol/L of cobalt protoporphyrin (an HO-1 inducer; Frontier Scientific, Logan, UT), 20 μmol/L of tin protoporphyrin IX dichloride (SnPP; an HO-1 inhibitor; Santa Cruz, TX), 50 μmol/L of dimethyl fumarate, or rosiglitazone at doses indicated in results.
In experiments where we used cytokine stimulation, we added 20 ng/mL of vascular endothelial growth factor (VEGF) or 10 ng/mL of tumor necrosis factor-α (TNF-α) after 24 hours of siRNA treatment.
Assays to Measure Various Outputs
Methods to perform quantitative polymerase chain reaction (PCR), digital PCR, immunohistochemistry, Western blot, and ELISAs are described in the online-only Data Supplement.
All functional studies were repeated at least 3× where samples from different patients were used for each replicate (ie, n=3 for all functional experiments). Each of these 3 biological replicates was done in technical triplicate. The mean of these technical replicates for each experimental condition was calculated, and the combined means from the 3 biological repeats were used for the final statistical analysis. The data were then tested for normality, and the relevant statistical analysis was performed on the biological replicates (n=3). If there were multiple comparisons, a post hoc analysis was also done using either the Tukey (parametric) or Dunn test (nonparametric).
Analysis of Published Array Data Sets to Examine HO-1 Expression
Data submitted to Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo) were obtained and analyzed using the GEOR analysis program. Data were analyzed after normalization and processing. Gene expression was compared using the BioConductor package limma in R. Differentially expressed genes were calculated using t test with the Benjamini–Hochberg false discovery rate. P<0.05 was considered significant.
HO-1 mRNA Expression and Protein Expression in Placenta From Preterm Preeclamptic Pregnancies Are Not Decreased Compared With Gestation-Matched Controls
We first examined whether placental HO-1 expression is reduced in preeclampsia.23,24 We measured placental HO-1 expression in cases of preterm preeclampsia and preterm controls (normotensive women delivering a neonate of normal birthweight). The samples were selected from a tissue bank of samples collected contemporaneously at a single-tertiary hospital using the same tissue collection protocols.
Table S1 summarizes the clinical details of the entire cohort whose samples were used in this study. Notably, gestational age was not different between the cohorts.
There was no difference in HO-1 mRNA expression, measured by conventional quantitative PCR, between preeclampsia (n=12) and preterm controls (n=21; Figure 1A). We verified this finding using digital PCR. Digital PCR measures absolute mRNA copy number and is considered more precise than conventional PCR.27 There was no difference in HO-1 mRNA copy number measured using digital PCR between the preeclampsia cohort (n=15) and preterm controls (n=18; Figure 1B).
To confirm that our preeclamptic placentas are representative of the disease, we measured mRNA expression of 2 sFlt-1 variants, sFlt-1 i13 and sFlt-1 e15a.28 As expected, placental mRNA expression of both sFlt-1 variants was increased >2.5-fold in the preeclamptic samples, compared with preterm controls (Figure 1C and 1D; P<0.01).
We next measured HO-1 protein expression in placenta by Western blot analysis and densitometric analysis. There was no difference in HO-1 protein levels between preeclampsia (n=17) and preterm control (n=24) cohorts (Figure 1E; Figure S1 in the online-only Data Supplement shows all blots). HO-1 was expressed in the syncytiotrophoblasts and villous vessels (Figure 1F).
The variation in numbers used for the different assays (Western blot, digital PCR, and conventional PCR) is accounted for by the fact that there were differences in the number of samples available at the time we set out to perform the various assays.
There were differences in mode of birth among the 2 cohorts (Table S1). To examine whether this may be biasing our results, we re-examined our digital PCR data in the preterm control cohort and did not find differences in HO-1 mRNA copy number among those who labored and those delivered by cesarean section (Figure S2A). Also, HO-1 expression was not different between preeclamptic and preterm controls in a subanalysis restricted only to cases delivered by cesarean section (Figure S2B). Therefore, mode of delivery is unlikely to be biasing our results. Within the preterm control group, there was also no significant difference in HO-1 expression when we divided the cohort according to either indication for delivery (Figure S2C) or whether there was histological evidence of infection in the placenta (Figure S2D).
We next examined HO-1 expression in microarray data sets sourced from publically available microarray repositories. We identified 7 microarray data sets comparing placental samples from preeclampsia and controls. There was no evidence of significantly decreased HO-1 expression with preeclampsia in any microarray cohort (Table S2). Furthermore, there was no obvious global trend toward decreased HO-1 expression with preeclampsia. In fact, 5 microarray data sets showed a nonsignificant >3-fold increase in HO-1 expression with preeclampsia and only 2 showed a nonsignificant decrease in expression.
Overall, we were unable to verify that HO-1 is decreased in preeclamptic placenta at either the mRNA or protein level.
Effects of Functionally Altering HO-1 Expression and Activity on sFlt-1 and sENG Secretion From Primary Trophoblasts and Placental Explants
It is thought that placental-derived sFlt-1 and sENG are responsible for maternal endothelial dysfunction that occurs in preeclampsia. Therefore, we examined whether HO-1 regulates placental production of sFlt-1 and sENG. We functionally manipulated HO-1 expression and activity in primary human placental tissues and measured sFlt-1 and sENG.
Despite efficient HO-1 siRNA gene silencing in primary trophoblasts (Figure 2A and 2C), there was no change in sFlt-1 secretion at either 24 or 48 hours (Figure 2B and 2D). We have previously noted that isolated primary cytotrophoblasts do not reliably produce sENG.26 Therefore, we could not assess whether silencing HO-1 affects sENG secretion in this assay.
Around 80% of the placental sFlt-1 transcripts are sFlt-1 e15a (or sFlt-1 v14), a recently described primate and placental-specific sFlt-1.28 The remaining sFlt-1 transcripts are sFlt-1 i13, a variant generically expressed not only in many tissues, notably endothelium, but also in brain and liver.28 Interestingly, silencing HO-1 increased sFlt-1 i13 expression but not sFlt-1 e15a (Figure 2E and 2F). However, this did not translate to changes in protein secretion of total sFlt-1 (Figure 2B and 2D—note the sFlt-1 ELISA indiscriminately detects both variants).
Nrf2 is a transcription factor that directly upregulates genes involved in the antioxidant response, including HO-1. Efficient silencing of Nrf2 in primary trophoblasts (Figure S3A) decreased HO-1 expression as expected (Figure S3B) but did not affect sFlt-1 secretion (Figure S3C). Interestingly, silencing Nrf2 decreased expression of the sFlt-1 variants (Figure S3D and S3E), but this is opposite to that expected, given that HO-1 is proposed to negatively regulate sFlt-1.
We next used small molecules to modulate HO-1 activity in trophoblasts. It was previously reported that administering SnPP (HO-1 inhibitor) into placental explants increased sFlt-1 release >2-fold under normoxic (5% oxygen) conditions.11 We set out to replicate this experiment, administering SnPP into placental explants (except we defined normoxia as 8% oxygen). In contrast to previous findings, SnPP failed to significantly affect sFlt-1 (Figure 2G) or sENG secretion (Figure 2H). Cobalt protoporphyrin (HO-1 inducer) also did not affect sFlt-1 (Figure 2G) or sENG secretion (Figure 2H).
The previous study concluding that HO-1 negatively regulates sFlt-1 and sENG report experiments where cells were treated with cytokines (interferon-γ, TNF-α, and VEGF-E) coadministered with agents to functionally manipulate HO-1 expression or activity.11 Therefore, we not only repeated siRNA HO-1 silencing experiments in trophoblasts but also coadministered either VEGF-E and TNF-α. Under these conditions, efficient HO-1 silencing (Figure 3A) still did not affect sFlt-1 secretion (Figure 3B). There was a significant increase in mRNA expression of sFlt-1 i13 (Figure 3C) but not of the placental-specific sFlt-1 variant, sFlt-1 e15a (Figure 3D). Of note, this did not translate to differences in sFlt-1 secretion (Figure 3B).
Overall, functional experiments in primary placental cells (with or without cytokine pretreatment) did not support the contention that HO-1 negatively regulates sFlt-1 or sENG secretion from placenta.
Effect of Silencing HO-1 Expression and Activity on sFlt-1 and sENG Secretion in Primary Endothelial Cells
The previous report proposing that HO-1 negatively regulates sFlt-1 and sENG performed most experiments in HUVECs.11 Therefore, we repeated our functional experiments in primary endothelial cells.
In contrast to our findings in primary placental tissues, efficient HO-1 silencing in HUVECs (Figure 4A) indeed increased sFlt-1 secretion (Figure 4B). Interestingly, when we examined mRNA expression, we found that silencing HO-1 significantly upregulated sFlt-1 i13 >2-fold (the main sFlt-1 variant in endothelial cells; Figure 4C) but had no effect on sFlt-1 e15a expression (the dominant sFlt-1 splice variant in placenta but minimally expressed in other tissues28; Figure 4D). Thus, it is possible that the fact that HO-1 specifically regulates sFlt-1 i13 (generic sFlt-1 variant and main endothelial-derived sFlt-1)28 but not placental sFlt-1 variant (sFlt-1 e15a) may explain why silencing HO-1 affects sFlt-1 protein secretion from endothelial cells but not placenta. Silencing HO-1 in HUVECs did not affect sENG secretion (Figure 4E).
We next silenced HO-1 in a second endothelial cell type, primary human UtMVECs. These contain a mix of venous and arterial endothelial cells, whereas HUVECs are only of venous origin. Like HUVECs, efficient silencing of HO-1 in UtMVECs (Figure 4F) also caused a significant increase in sFlt-1 secretion (Figure 4G) but did not affect sENG secretion (Figure 4H).
We next not only repeated the HO-1 silencing experiments in endothelial cells but also coadministered VEGF-E and TNF-α. We confirmed the findings reported previously11 that silencing HO-1 in endothelial cells under the stimulus of both cytokines increased sFlt-1 secretion (Figure S4A and S4B). Specifically, silencing HO-1 increased sFlt 1-i13 but not sFlt-1 e15a expression (Figure S4C and S4D). We obtained the same findings when similar experiments were performed in UtMVECs (Figure S4E and S4F). Silencing HO-1 in UtMVECs with VEGF-E and TNF-α stimulation did not affect sENG secretion (Figure S4G).
Effect of Small Molecules That Inhibit or Induce HO-1 on sFlt-1 and sENG Secretion
We next examined whether upregulating HO-1 with cobalt protoporphyrin (HO-1 inducer) affects sFlt-1 secretion in HUVECs and UtMVECs. Cobalt protoporphyrin potently induced HO-1 expression by ≈19-fold in both endothelial cell types (Figure S5A and S5B). Surprisingly, this marked HO-1 induction did not affect sFlt-1 or sENG secretion (Figure S5C–S5E). Administering SnPP to inhibit HO-1 activation in UtMVECs did not affect sFlt-1 secretion (Figure S5F).
We not only repeated these experiments where we added SnPP to HUVECs but also coadministered either VEGF-E and TNF-α. Unexpectedly, we observed a significant decrease in sFlt-1 secretion under all conditions and a significant reduction in sENG secretion under control conditions (Figure S6A and S6B). These findings are opposite to the contention that HO-1 negatively regulates the production of sFlt-1 and sENG.
Finally, we added 2 small molecule HO-1 activators, dimethyl fumarate (a drug developed to activate Nrf2 as a treatment for multiple sclerosis) to HUVECs and rosiglitazone (developed to treat type II diabetes mellitus) to primary trophoblasts. Both drugs upregulated HO-1 expression (Figure S7A and S7D). However, dimethyl fumarate did not affect sFlt-1 secretion (Figure S7B) from HUVECs but increased sENG release (Figure S7C; opposite to the contention that HO-1 negatively regulates sENG production). Rosiglitazone did not affect sFlt-1 secretion from primary cytotrophoblasts (Figure S7E).
Thus, although our HO-1 silencing experiments support the contention that HO-1 negatively regulates sFlt-1 in endothelial cells (Figure 4; Figure S4), none of our experiments examining molecules support this contention (Figures S5–S7). Some of the results using small molecules were contrary to this hypothesis.
None of our experiments in either placental tissues or endothelial cells support the premise that HO-1 negatively regulates sENG production.
The contention that HO-1 expression is reduced in preeclampsia and that this decrease is directly responsible for increased sFlt-1 and sENG secretion is widely accepted as a key step in the pathogenesis of preeclampsia.2,12–21 The premise of a randomized clinical trial of pravastatin to treat preterm preeclampsia is based on the fact that statins upregulate HO-1 and this in turn decreases sFlt-1 production.13 In that study, the primary outcome is sFlt-1 production and HO-1 expression is a prespecified outcome. However, there has been no independent validation confirming whether HO-1 indeed regulates the production of sFlt-1 and sENG.
The evidence that HO-1 expression is reduced in preeclamptic placenta is limited to 2 reports where one had a small sample size23 and the second did not account for gestational age.24 Furthermore, Barber et al29 did not find that HO-1 was differentially expressed in preeclampsia. Therefore, the contention that HO-1 is decreased in preeclampsia merited further validation.
We measured HO-1 expression in preterm preeclampsia in a sizeable cohort. Importantly, we used preterm controls, examined mRNA expression levels by conventional and digital PCR, and measured protein expression. We also analyzed data from microarray repositories. Using these approaches, we were unable to validate the contention that there is decreased placental HO-1 expression in preeclampsia.
We were also unable to confirm the fact that HO-1 plays a major role in regulating sFlt-1 and sENG in placenta.2,12–21 We think that we were comprehensive and methodical in arriving at this conclusion. We silenced HO-1 expression, examined 2 timepoints, and measured the sFlt-1 mRNA expression of 2 variants and protein secretion. We silenced Nrf2, the upstream transcription factor regulating HO-1 expression. We also treated placental explants with various HO-1 inhibitors and inducers. Furthermore, we also repeated experiments where we stimulated the cells with VEGF-E and TNF-α.
Interestingly, in endothelial cells, we did find that silencing HO-1 increased sFlt-1 secretion. As such, we did replicate some of the findings in the previous report.11 By examining the mRNA expression of the 2 major sFlt-1 splice variants, we have uncovered a possible explanation that why silencing HO-1 affects sFlt-1 secretion in HUVECs but not placental tissues. In both placental and endothelial cells, silencing HO-1 selectively increased sFlt-1 i13 mRNA expression (generically expressed in many tissues and the main sFlt-1 variant in endothelial cells)28 but not sFlt-1 e15a. Thus, it is possible that HO-1 may be differentially regulating sFlt-1 i13 but not the placental sFlt-1 variant, sFlt-1 e15a. Given that sFlt-i13 is the dominant sFlt-1 in endothelial cells, this may account for the significant difference in sFlt-1 secretion from endothelial cells but not from placenta. Although this differential regulation is interesting, it still suggests that HO-1 is unlikely to have a role in regulating placentally derived sFlt-1 where sFlt-1 e15a is by far, the dominant variant.28
We would note the possibility that sFlt-1 i13 may be regulated by HO-1 is tempered by the fact that we treated endothelial cells with 3 different HO-1 inducers (with up to 19-fold increase in expression) and 1 inhibitor; none of these experiments suggested that HO-1 negatively regulated sFlt-1 secretion. There are indeed other HO-1 inducers, such as sofalcone,30 pravastatin,31 and simvastatin,11 that are associated with a decrease in sFlt-1. However, given that this is not a consistent finding with HO-1 inducers, it raises the possibility that the decrease in sFlt-1 secretion with statin treatment may not be mediated through HO-1.
We were unable to obtain any evidence that HO-1 negatively regulates sENG in either placenta or endothelial cells.
We would make some further observations by comparing our work with the work by Cudmore et al.11 They performed most experiments in HUVECs rather than placental tissues.11 The evidence generated in placental tissues was limited to adding SnPP to explants, and HO-1 silencing experiments were not reported. As noted, although they showed that simvastatin induced HO-1 and decreased sFlt-1 production, they perhaps did not specifically demonstrate that this was mediated through HO-1.
A possible reason why Cudmore et al11 observed an increase in sFlt-1 protein in supernatants of lung biopsies of HO-1 knockout mice is that endothelial-derived sFlt-1 (not placental HO-1) may be regulating sFlt-1. We cannot explain their interesting observation that circulating sENG was increased in the HO-1 knockout mice (circulating sFlt-1 levels were not reported). sFlt-1 e15a is only expressed in primates, not in mice. This means our contention that HO-1 does not regulate sFlt-1 e15a (but may regulate sFlt-1 i13) cannot be explored using conventional gene knockout mouse models.
We do note that upregulating the antioxidant HO-1 may itself be beneficial as a therapeutic for preeclampsia because this condition is strongly associated with oxidative stress.32 Furthermore, our work has specifically examined placental HO-1 expression in preeclampsia and its role in regulating sFlt-1 and sENG. We have not excluded the possibility that HO-1 still plays an important role in placental development and even in the pathogenesis of preeclampsia through other mechanisms. In mouse models where HO-1 is genetically deleted, there is inadequate remodeling of the spiral arteries, diminished numbers of uterine natural killer cells at the maternal fetal interface, and fetal growth restriction.33–35 Furthermore, mothers carrying an HO-1+/− heterozygote develop hypertension in the latter stages of their pregnancy.34 However, whether these intriguing findings are relevant to early human placentation is yet to be proven.
The clinical implication of our findings is that the strategy of identifying drugs that induce HO-1 as a means to decrease sFlt-1 and sENG in preeclampsia may not be sound. In light of our data, we suggest that drugs that induce HO-1 also need to be independently examined to determine whether they decrease sFlt-1 and sENG production.
It seems widely accepted that HO-1 is decreased in preeclamptic placenta and negatively regulates sFlt-1 and sENG production. However, we have measured HO-1 expression in a large cohort of placentas and we have been unable to verify the premise that its expression is reduced in preeclamptic placenta. Furthermore, we have undertaken a series of functional studies and have been unable to validate the contention that HO-1 negatively regulates placental sFlt-1 or sENG production. Hence, we conclude that HO-1 is unlikely to be a key molecule that regulates sFlt-1 and sENG production in the preeclamptic placenta.
We acknowledge clinical research midwives Gabrielle Pell, Debra Jinks, Rachel Murdoch, and Genevieve Christophers, the obstetrics and midwifery staff, and patients at the Mercy Hospital for Women.
Sources of Funding
This study was funded by grants (no. 1061977 and 1048707) and received salary support (no. 1050765, 1062418, and 628927) from the National Health and Medical Research Council of Australia and by the University of Melbourne Research CR Roper Fellowship (to N.J. Hannan).
The online-only Data Supplement is available with this article at http://hyper.ahajournals.org/lookup/suppl/doi:10.1161/HYPERTENSIONAHA.115.05847/-/DC1.
- Received May 14, 2015.
- Revision received June 1, 2015.
- Accepted August 6, 2015.
- © 2015 American Heart Association, Inc.
- Maynard SE,
- Min JY,
- Merchan J,
- Lim KH,
- Li J,
- Mondal S,
- Libermann TA,
- Morgan JP,
- Sellke FW,
- Stillman IE,
- Epstein FH,
- Sukhatme VP,
- Karumanchi SA.
- Koskenkorva-Frank TS,
- Weiss G,
- Koppenol WH,
- Burckhardt S.
- Cudmore M,
- Ahmad S,
- Al-Ani B,
- Fujisawa T,
- Coxall H,
- Chudasama K,
- Devey LR,
- Wigmore SJ,
- Abbas A,
- Hewett PW,
- Ahmed A.
- Robillard PY,
- Dekker G,
- Chaouat G.
- Powe CE,
- Levine RJ,
- Karumanchi SA.
- Wang A,
- Rana S,
- Karumanchi SA.
- Zhou CC,
- Ahmad S,
- Mi T,
- Abbasi S,
- Xia L,
- Day MC,
- Ramin SM,
- Ahmed A,
- Kellems RE,
- Xia Y.
- Ramma W,
- Ahmed A.
- 25.↵American College of Obstetricians and Gynecologists; Task force on Hypertension in Pregnancy. Hypertension in Pregnancy. 2013. http://www.acog.org/Resources-And-Publications/Task-Force-and-Work-Group-Reports/Hypertension-in-Pregnancy. Accessed September 1, 2014.
- Kaitu’u-Lino TJ,
- Tong S,
- Beard S,
- Hastie R,
- Tuohey L,
- Brownfoot F,
- Onda K,
- Hannan NJ.
- Whale AS,
- Huggett JF,
- Cowen S,
- Speirs V,
- Shaw J,
- Ellison S,
- Foy CA,
- Scott DJ.
- Jebbink J,
- Keijser R,
- Veenboer G,
- van der Post J,
- Ris-Stalpers C,
- Afink G.
- Barber A,
- Robson SC,
- Myatt L,
- Bulmer JN,
- Lyall F.
- Onda K,
- Tong S,
- Nakahara A,
- Kondo M,
- Monchusho H,
- Hirano T,
- Kaitu’u-Lino T,
- Beard S,
- Binder N,
- Tuohey L,
- Brownfoot F,
- Hannan NJ.
- Brownfoot F,
- Hannan N,
- Binder N,
- Walker S,
- Cannon P,
- Hastie R,
- Onda K,
- Kaitu’u-Lino T.
- Linzke N,
- Schumacher A,
- Woidacki K,
- Croy BA,
- Zenclussen AC.
Novelty and Significance
What Is New?
We were unable to confirm the widely held view that heme oxygenase-1 (HO-1) expression is decreased in preeclamptic placenta.
We have generated evidence contrary to the contention that HO-1 negatively regulates placental secretion of soluble fms-like tyrosine kinase-1 or soluble endoglin.
What Is Relevant?
In preeclampsia, elevated placental release of the antiangiogenic factors soluble fms-like tyrosine kinase-1 and soluble endoglin causes maternal endothelial dysfunction and end-organ injury.
It is widely accepted that HO-1 negatively regulates placental soluble fms-like tyrosine kinase-1 and soluble endoglin secretion, and there is decreased HO-1 expression in preeclamptic placenta.
HO-1 expression is not decreased in preeclamptic placenta and does not negatively regulate placental soluble fms-like tyrosine kinase-1 and soluble endoglin secretion.