Original Research

ajog.org

OBSTETRICS

Metformin as a prevention and treatment for preeclampsia: effects on soluble fms-like tyrosine kinase 1 and soluble endoglin secretion and endothelial dysfunction Fiona C. Brownfoot, MBBS; Roxanne Hastie, BBiomed Sc; Natalie J. Hannan, B Sci, PhD; Ping Cannon, B Sci; Laura Tuohey, BSci; Laura J. Parry, B Sci, PhD; Sevvandi Senadheera, B Sci PhD; Sebastian E. Illanes, MBBS, MSc; Tu’uhevaha J. Kaitu’u-Lino, PhD1; Stephen Tong, MBBS, PhD1

BACKGROUND: Preeclampsia is associated with placental ischemia/ hypoxia and secretion of soluble fms-like tyrosine kinase 1 and soluble endoglin into the maternal circulation. This causes widespread endothelial dysfunction that manifests clinically as hypertension and multisystem organ injury. Recently, small molecule inhibitors of hypoxic inducible factor 1a have been found to reduce soluble fms-like tyrosine kinase 1 and soluble endoglin secretion. However, their safety profile in pregnancy is unknown. Metformin is safe in pregnancy and is also reported to inhibit hypoxic inducible factor 1a by reducing mitochondrial electron transport chain activity. OBJECTIVE: The purposes of this study were to determine (1) the effects of metformin on placental soluble fms-like tyrosine kinase 1 and soluble endoglin secretion, (2) to investigate whether the effects of metformin on soluble fms-like tyrosine kinase 1 and soluble endoglin secretion are regulated through the mitochondrial electron transport chain, and (3) to examine its effects on endothelial dysfunction, maternal blood vessel vasodilation, and angiogenesis. STUDY DESIGN: We performed functional (in vitro and ex vivo) experiments using primary human tissues to examine the effects of metformin on soluble fms-like tyrosine kinase 1 and soluble endoglin secretion from placenta, endothelial cells, and placental villous explants. We used succinate, mitochondrial complex II substrate, to examine whether the effects of metformin on soluble fms-like tyrosine kinase 1 and soluble endoglin secretion were mediated through the mitochondria. We also isolated mitochondria from preterm preeclamptic placentas and gestationally matched control subjects and measured mitochondrial electron transport chain activity using kinetic spectrophotometric assays. Endothelial cells or whole maternal vessels were incubated with metformin to determine whether it rescued endothelial dysfunction induced by either tumor necrosis factor-a (to endothelial cells) or placenta villous explanteconditioned media (to whole vessels). Finally, we examined the effects of metformin on angiogenesis on maternal omental vessel explants.

Cite this article as: Brownfoot FC, Hastie R, Hannan NJ, et al. Metformin as a prevention and treatment for preeclampsia: effects on soluble fms-like tyrosine kinase 1 and soluble endoglin secretion and endothelial dysfunction. Am J Obstet Gynecol 2016;214:356.e1-15. 0002-9378/free Crown Copyright ª 2016 Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ajog.2015.12.019

P

RESULTS: Metformin reduced soluble fms-like tyrosine kinase 1 and

soluble endoglin secretion from primary endothelial cells, villous cytotrophoblast cells, and preterm preeclamptic placental villous explants. The reduction in soluble fms-like tyrosine kinase 1 and soluble endoglin secretion was rescued by coadministration of succinate, which suggests that the effects of metformin on soluble fms-like tyrosine kinase 1 and soluble endoglin were likely to be regulated at the level of the mitochondria. In addition, the mitochondrial electron transport chain inhibitors rotenone and antimycin reduced soluble fms-like tyrosine kinase 1 secretion, which further suggests that soluble fms-like tyrosine kinase 1 secretion is regulated through the mitochondria. Mitochondrial electron transport chain activity in preterm preeclamptic placentas was increased compared with gestation-matched control subjects. Metformin improved features of endothelial dysfunction relevant to preeclampsia. It reduced endothelial cell messenger RNA expression of vascular cell adhesion molecule 1 that was induced by tumor necrosis factorea (vascular cell adhesion molecule 1 is an inflammatory adhesion molecule up-regulated with endothelial dysfunction and is increased in preeclampsia). Placental conditioned media impaired bradykinin-induced vasodilation; this effect was reversed by metformin. Metformin also improved whole blood vessel angiogenesis impaired by fms-like tyrosine kinase 1. CONCLUSION: Metformin reduced soluble fms-like tyrosine kinase 1 and soluble endoglin secretion from primary human tissues, possibly by inhibiting the mitochondrial electron transport chain. The activity of the mitochondrial electron transport chain was increased in preterm preeclamptic placenta. Metformin reduced endothelial dysfunction, enhanced vasodilation in omental arteries, and induced angiogenesis. Metformin has potential to prevent or treat preeclampsia. Key words: electron transport chain, metformin, preeclampsia, soluble

endoglin, soluble fms-like tyrosine kinase 1

reeclampsia is a serious pregnancy complication that globally is responsible for >100 maternal and 400 perinatal deaths each day.1-7 An important step in the pathophysiologic condition may be placental ischemia/hypoxia,8-15 which leads to the release of soluble fmslike tyrosine kinase 1 (sFlt-1)16-18 and soluble endoglin (sENG)19 into the maternal circulation.17,20-22 These cause endothelial dysfunction that leads to multisystem

356.e1 American Journal of Obstetrics & Gynecology MARCH 2016

organ injury.23-35 There are no treatments to arrest disease progression; expectant management and delivery remain the only treatment options.1,25,36-39 A medication that is safe in pregnancy, that reduces placental sFlt-1 and sENG secretion, that rescues endothelial dysfunction, and that is angiogenic may be effective in the treatment or prevention of preeclampsia. There is interest in the use of drugs that inhibit hypoxic inducible factor 1a (HIF1a) to treat preeclampsia.9,40,41 HIF1a is up-regulated with ischemia/

ajog.org hypoxia42 and facilitates sFlt-1 secretion.43,44 Therefore, drugs that block HIF1a activity may decrease sFlt-1 secretion. Indeed, the HIF1a inhibitors YC-140 and ouabain41 have been shown to reduce sFlt-1 secretion from placental tissues. However, the safety profile of YC-1 and ouabain in pregnancy are not known, and they are still in clinical trials for use in pulmonary hyperplasia45,46 and cancer,47,48 respectively. This prompted us to explore repurposing a medication that is thought to be safe in pregnancy that inhibits HIF1a and lead us to investigate metformin.49 Metformin is an oral hypoglycemic agent that is used to treat gestational diabetes mellitus.49-51 Recently, metformin has been reported to reduce breast52,53 and prostate54 cancer metastasis and prolong survival. This discovery renewed interest in the further exploration of its mechanism of action; it recently was shown to inhibit HIF1a by blocking complex I of the mitochondrial electron transport chain.53,55,56 Therefore, we hypothesized that metformin may reduce sFlt-1 secretion in women with preeclampsia. Metformin has been reported to have vasoprotective properties; epidemiologic studies have shown that it reduces cardiovascular morbidity in patients with polycystic ovarian syndrome26 and diabetes mellitus.57,58 This has been attributed to its ability to reduce vascular cell adhesion molecule 1 (VCAM-1),59,60 which is a molecule that is expressed on the luminal surface of blood vessels in the presence of inflammation and is increased in preeclampsia.61 Metformin has also been reported to induce vasodilation of diabetic rat vessels.62

Objective This study had 3 objectives: (1) to assess the effects of metformin on sFlt-1 and sENG secretion from primary placental and endothelial cells/tissues and to investigate whether these effects are mediated through mitochondrial electron transport chain inhibition; (2) to assess whether mitochondrial electron transport chain activity positively regulates sFlt-1 secretion and if preterm preeclamptic placenta have increased

OBSTETRICS mitochondrial electron transport chain activity; and (3) to assess whether metformin can reduce endothelial dysfunction, induce vasodilation, and stimulate angiogenesis in human omental arteries.

Materials and Methods Patient population We performed functional experiments in which we administered metformin to human tissues and assessed its effects on sFlt-1 and sENG secretion, mitochondrial electron transport chain function, and endothelial dysfunction. To perform our experiments, we examined tissues from placenta and blood vessels. We collected several types of placental tissues. We isolated human umbilical vein endothelial cells (HUVECs)63 and primary villous cytotrophoblast cells64 from the placenta and umbilical cord that had been collected from patients at term. We also collected placental villous explants from patients with severe early onset proteinuric preeclampsia (delivered at
34 weeks gestation by cesarean delivery) as defined by the American College of Obstetricians and Gynecologists (ACOG) guidelines.65 Placental biopsy specimens were taken from 4 random placental sites as recommended by the Co-Lab consortium.66 There was hypertension and proteinuria (>300 mg of protein in a 24-hour urine collection) in all cases. The placental villous explants were prepared, as previously described.63,67 To assess mitochondrial electron transport chain activity, we obtained a placental biopsy specimen from 23 women with severe preterm proteinuric preeclampsia (defined by ACOG 2013 guidelines65) and from 25 gestationally matched normotensive control subjects who had a preterm delivery without evidence of significant chorioamnionitis or maternal comorbidities (Supplementary Table). All the women who were diagnosed with preeclampsia had proteinuria. All placental samples were collected from women who underwent a cesarean delivery. Mitochondrial electron transport chain activity studies were performed, as previously described.68,69 We also collected omental tissue from patients who had elective term cesarean

Original Research

delivery and dissected omental arteries, as previously described.70 These omental arteries were used to assess the effects of metformin on vasodilation (by pressure myography) and angiogenesis (omental artery explant assay), as described later. This study was approved by The Mercy Health Human Research Ethics Committee (Institutional review board number R11/34; approved on the November 12, 2014); all women gave written informed consent.

In vitro experiments that assessed metformin and mitochondrial electron transport chain activity inhibitors and substrates on sFlt-1 and sENG production and endothelial dysfunction HUVECs were plated at 24,000 cells/cm2 between passages 2 and 4 and cultured at 37 C in 20% O2 and treated for 24 hours. Primary villous cytotrophoblast cells were plated at 24,000 cells/cm2, incubated overnight to ensure larger viable villous cytotrophoblasts had adhered, washed to remove apoptotic mononuclear syncytiotrophoblast fragments, then treated for 24 hours at 500,000 cells/cm2 and treated for 48 hours at 37 C in 8% O2 to assess sFlt-1 and sENG secretion, respectively. Placental villous explants of approximately 20-mg tissue per well (dried weight) were treated for 72 hours at 37 C in 8% O2. HUVECs, primary villous cytotrophoblast cells, and placental villous explants were treated with 0, 1, 2, and 5 mmol/L metformin (Sigma Chemical Company, St. Louis, MO); HUVECs and primary villous cytotrophoblast cells were treated with 0, 0.5, and 1 mmol/L metformin  25 mmol/L succinate (mitochondrial electron transport chain substrate; Sigma Chemical Company). Primary villous cytotrophoblast cells were also treated with mitochondrial electron transport chain inhibitors rotenone (Sigma Chemical Company) at 0, 0.625, 1.25, 2.5, and 5 mmol/L or antimycin (Sigma Chemical Company) at 0, 0.156, 0.31, 0.63, and 1.25 mmol/L. Endothelial dysfunction was induced in (1) HUVECs that used a constant dose of 10ng/mL tumor necrosis factor a

MARCH 2016 American Journal of Obstetrics & Gynecology

356.e2

Original Research

OBSTETRICS

(TNFa; Sigma Chemical Company), (2) whole omental arteries with the use of 25% conditioned placental villous explant media (this media is collected 24 hours after being cultured with placental villous explants that were obtained from normal pregnancies at term), and (3) omental artery explants that used 250 ng/mL sFlt-1 at 37 C at 20% O2. Metformin was administered simultaneously to HUVECs at 0, 1, 2, or 5 mmol/L for 24 hours, to whole omental arteries at 0 and 5 mmol/L for 3 hours, and omental artery explants at 1 mmol/L for 120 hours.

Measurement of sFlt-1, sENG, and VCAM-1, sFlt-1 e15a and i13 Conditioned media were collected, and RNA was extracted with the RNeasy mini kit (Qiagen, Valencia, CA) from functional experiments and HUVEC endothelial dysfunction assay. Enzyme-linked immunosorbent assay for sFlt-1 and sENG was performed with the DuoSet VEGF R1/Flt-1 kit (R&D Systems by Bioscience, Waterloo, Australia) and a DuoSet Human Endoglin CD/105 ELISA kit (R&D Systems), respectively. RNA was quantified with the Nanodrop ND 1000 spectrophotometer (NanoDrop Technologies Inc, Wilmington, DE). RNA (0.2 mg) was converted to complementary DNA with the use of Applied Biosystems high capacity cDNA reverse transcriptase kit (Life Technologies, Mulgrave, Australia). Sybr gene expression assay for sFlt-1 e15a and sFlt-1 i13 (Geneworks, South Australia, Australia) was performed,71 and a taqman gene expression assay was used for VCAM-1 (Life Technologies).

Assessment of the effect of metformin on whole omental artery vasodilation Treated whole omental arteries were mounted on a pressure myograph organ bath (Living Systems Instrumentation, Burlington, VT). Incremental doses of 0.01 nmol/L to 1mmol/L bradykinin (Auspep, West Melbourne, Australia) were infused, and vasodilation was assessed with video microscopy (Diamtrak Software, Adelaide, SA, Australia).72

Assessment of the effect of metformin on angiogenesis with the use of omental artery explants Omental artery explants (0.5-mm rings) were stained with calcein AM (Merck Millipore, Darmstadt, Germany) and imaged at 40 magnification with the EVOS FL microscope (Life Technologies); outgrowth was assessed with image J (http://imagej.nih.gov/ij/).73

Statistical analysis Technical triplicates were performed for each experiment, with a minimum of 3 biologic replicates for each in vitro study (samples from different patients were used for each biologic replicate). When 2 groups were analyzed, a t-test (parametric) or a Mann-Whitney test (nonparametric data) was used. When 3 groups were compared, a 1-way analysis of variance test (parametric) or a Kruskal-Wallis test (non-parametric) was used. Statistical analysis was done with GraphPad Prism 6 software (GraphPad Software, La Jolla, CA). All data are expressed as mean  SEM; probability values of <.05 were considered significant. Detailed methods are included in the supplementary “Methods” section.

Results Metformin reduces sFlt-1 secretion from primary endothelial cells and placental tissues We assessed the effects of metformin on sFlt-1 secretion from endothelial and placental tissues because they are its main tissue source. Administering metformin dose-dependently reduced sFlt-1 secretion from endothelial cells (HUVECs; Figure 1, A) and primary cells that were isolated from placenta (villous cytotrophoblast cells; Figure 1, B). At the highest doses, metformin reduced endothelial and placental cell secretion by 53% and 63%, respectively. Metformin also reduced sFlt1 secretion from placental villous explants that were obtained from 4 women who had been diagnosed with preterm preeclampsia (delivery required <34 weeks gestation; Figure 1, C). We investigated the effect of metformin on messenger RNA (mRNA) expression of different sFlt-1 variants in

356.e3 American Journal of Obstetrics & Gynecology MARCH 2016

ajog.org the cells, or placental villous explant tissues. sFlt-1 i13 is the most abundant sFlt-1 variant in endothelial cells.74 Metformin dose-dependently reduced sFlt-1 i13 mRNA expression in endothelial cells (Figure 1, D). sFlt-1 e15a is the predominant variant expressed in human placenta.74 Metformin reduced sFlt-1 e15a mRNA expression in primary villous cytotrophoblasts cells (Figure 1, E) and placental villous explants (Figure 1, F) that were obtained from women with preterm preeclampsia. Thus, we conclude that metformin reduces sFlt-1 isoform expression and sFlt1 secretion in endothelial and placental cells/tissues, including placental villous explants from patients diagnosed with preterm preeclampsia.

Metformin reduces sENG secretion from primary endothelial and placental tissues We next investigated the effects of metformin on sENG secretion from primary endothelial cells and placental cells/ tissues. Metformin dose-dependently reduced sENG secretion from HUVECs (Figure 2, A) and primary villous cytotrophoblast cells (Figure 2, B). Metformin induced a trend towards a reduction in sENG secretion from preterm preeclamptic placental villous explants at 3 doses, but none of these decreases were significant (Figure 2, C).

Metformin reduces sFlt-1 and sENG secretion by inhibiting the mitochondrial electron transport chain Given that metformin inhibits mitochondrial electron transport chain activity by blocking complex I,53,55,56 we examined whether the decrease in sFlt-1 secretion was mediated through mitochondrial electron transport chain inhibition. Succinate is a substrate for complex II of the mitochondria, downstream of the effect of metformin blockade. Thus, if metformin was reducing sFlt-1 secretion by directly blocking complex I, then succinate should restore electron flow and rescue this effect. Indeed, succinate rescued a reduction in sFlt-1 secretion that was induced by endothelial cells (Figure 3, A) and primary villous cytotrophoblasts

ajog.org

OBSTETRICS

Original Research

FIGURE 1

FIGURE 2

Effect of metformin on soluble fms-like tyrosine kinase 1 secretion and isoforms e15a and i13 expression in endothelial cells and placental tissue

Effect of metformin on soluble endoglin secretion from endothelial cells and placental tissue

Metformin (0, 1, 2, 5 mmol/L) dose-dependently reduced soluble fms-like tyrosine kinase 1 secretion from A, endothelial cells, B, villous cytotrophoblast cells, and C, preterm preeclamptic placental villous explants. Metformin reduced endothelial cell expression of D, sFlt-1 i13 isoform, E, villous cytotrophoblast cells, and F, preterm preeclamptic placental villous explant messenger RNA expression of sFlt-1 e15a. The single asterisk indicates P < .05; the double asterisks indicate P < .01; the triple asterisks indicate P < .0001; the quadruple asterisks indicate P < .00001. sFlt-1, soluble fms-like tyrosine kinase 1. Brownfoot et al. Metformin decreases sFlt-1 and sENG, improves endothelial function, and is angiogenic. Am J Obstet Gynecol 2016.

cells (Figure 3, B). Succinate also rescued a decrease in sENG secretion that was induced by metformin in primary HUVECs (Figure 3, C) and villous cytotrophoblast cells (Figure 3, D). These data raise the possibility that the effects of metformin on sFlt-1 and sENG secretion are mediated through its effects on the mitochondria.

Inhibition of the mitochondrial electron transport chain reduces sFlt-1 secretion from primary villous cytotrophoblasts cells To our knowledge, the concept that the mitochondria regulate sFlt-1 secretion is novel. To obtain further evidence that this is the case, we examined whether other mitochondrial electron transport

Metformin (0, 1, 2, and 5 mmol/L) reduced soluble endoglin secretion from A, endothelial cells and B, villous cytotrophoblast cells. Metformin did not change soluble endoglin secretion from C, preterm preeclamptic placental villous explants. The single asterisk indicates P < .05; the double asterisks indicate P < .01; the quadruple asterisks indicate P < .00001. sENG, soluble endoglin. Brownfoot et al. Metformin decreases sFlt-1 and sENG, improves endothelial function, and is angiogenic. Am J Obstet Gynecol 2016.

chain inhibitors reduce sFlt-1 secretion. Administering rotenone, another complex I inhibitor, to primary villous cytotrophoblast cells reduced sFlt-1 secretion by 65% (Figure 4, A). Antimycin, a complex III inhibitor, also reduced sflt-1 secretion by 75% (Figure 4, B). These doses of rotenone and antimycin did not induce cell death

MARCH 2016 American Journal of Obstetrics & Gynecology

356.e4

Original Research

ajog.org

OBSTETRICS

FIGURE 3

FIGURE 4

Effect of metformin and mitochondrial electron transport chain complex I activity on soluble fms-like tyrosine kinase 1 and soluble endoglin secretion from endothelial cells and villous cytotrophoblast cells

Mitochondrial electron transport chain inhibitors and sFlt-1 secretion

The effect of metformin (0.5 and 1 mmol/L) on soluble fms-like tyrosine kinase 1 secretion from A, endothelial cells and B, villous cytotrophoblast cells is likely mediated through mitochondrial electron transport chain complex 1 inhibition because the effect is rescued in the presence of the mitochondrial electron transport chain substrate, succinate (25 mmol/L). Furthermore, metformin also reduced soluble endoglin secretion likely through mitochondrial electron transport chain complex 1 inhibition from C, endothelial cells (0.5, 1 mmol/L) and D, villous cytotrophoblast cells (1 mmol/L) because succinate (25 mmol/L) rescued its effect on secretion. The single asterisk indicates P < .05; the double asterisks indicate P < .01; the triple asterisks indicate P < .0001; the quadruple asterisks indicate P < .00001.

A, Rotenone (0, 0.62, 1.25, 2.5, and 5 mmol/L), a complex 1 mitochondrial electron transport chain inhibitor, and B, antimycin (0, 0.16, 0.31, 0.63, and 1.25 mmol/L), a complex III mitochondrial electron transport chain inhibitor, both significantly reduce soluble fms-like tyrosine kinase 1 secretion from villous cytotrophoblast cells. The single asterisk indicates P < .05; the double asterisks indicate P < .01; the triple asterisks indicate P < .0001. sFlt-1, soluble fms-like tyrosine kinase 1. Brownfoot et al. Metformin decreases sFlt-1 and sENG, improves endothelial function, and is angiogenic. Am J Obstet Gynecol 2016.

sENG, soluble endoglin; sFlt-1, soluble fms-like tyrosine kinase 1. Brownfoot et al. Metformin decreases sFlt-1 and sENG, improves endothelial function, and is angiogenic. Am J Obstet Gynecol 2016.

(MTS assay and calcein stain; data not shown). These studies provide further evidence that the mitochondria is involved in the regulation of sFlt-1 secretion.

Mitochondrial electron transport chain activity is up-regulated in preterm preeclamptic placenta Given the mitochondria appears to positively regulate sFlt-1 and sENG secretion,

we hypothesized that preeclamptic placentas might have increased mitochondrial electron transport chain activity. We therefore compared mitochondrial electron transport chain activity in preterm preeclamptic placentas (n ¼ 23) and normotensive gestation matched preterm control placentas (n ¼ 25; Supplementary Table contains baseline characteristics). We observed an increase in mitochondrial electron transport chain activity in

356.e5 American Journal of Obstetrics & Gynecology MARCH 2016

the preeclamptic placentas for all 4 complexes, and this was significant for complex II (Figure 5). Therefore, mitochondrial electron transport chain activity may be increased in preterm preeclamptic placenta.

Metformin reduces VCAM-1 expression on endothelial cells Endothelial dysfunction is associated with increased VCAM-1 expression in the endothelium.61,75 VCAM-1 is an adhesion molecule that is expressed on the luminal surface of blood vessels

ajog.org

OBSTETRICS

Original Research

FIGURE 5

FIGURE 6

Mitochondrial electron transport chain activity in preterm preeclamptic placenta compared with gestationally matched controls

Effect of metformin on endothelial cell vascular cell adhesion molecule 1 expression

Inflammatory cytokine tumor necrosis factor a increased endothelial cell expression of vascular cell adhesion molecule 1 and was significantly reduced with increasing doses of metformin (0, 1, 2, and 5 mmol/L). The single asterisk indicates P < .05; the triple asterisks indicate P < .0001. TNFa, tumor necrosis factor a; VCAM 1, vascular cell adhesion molecule 1. Brownfoot et al. Metformin decreases sFlt-1 and sENG, improves endothelial function, and is angiogenic. Am J Obstet Gynecol 2016.

Activity of all (A-D) complexes in the mitochondrial electron transport chain was increased in preeclamptic placenta (n ¼ 23), compared with preterm placental controls (n ¼ 25), and was significant for complex II (B). The single asterisk indicates P < .05. CS, citrate synthase. Brownfoot et al. Metformin decreases sFlt-1 and sENG, improves endothelial function, and is angiogenic. Am J Obstet Gynecol 2016.

and can cause an inflammatory mesh and snare circulating blood cells. Preeclampsia is also associated with increased circulating TNFa,76 which is a proinflammatory cytokine that upregulates VCAM-1.61,75 We therefore performed an in vitro assay for which we added TNFa to HUVECs, which significantly up-regulated VCAM-1 expression in endothelial cells (Figure 6). The administration of metformin significantly reduced TNFa-induced VCAM-1 expression, which suggests that it may have effects to decrease endothelial dysfunction.

Metformin induces vasodilation in maternal vessels that are isolated from the omentum We next set up an assay to mimic the fact that placental factors that are released

into the maternal circulation have a role in inducing whole blood vessel dysfunction. We incubated human omental arteries in either conditioned placental culture media or normal culture media (not incubated with placental tissue). The vessels that were incubated in normal culture media dilated 100% in response to bradykinin (endogenous vasodilator), but vessels that were incubated in placental culture media exhibited a significant impairment in vasorelaxation (40% less; Figure 7, A). This suggests that placental factors seem to impair vasorelaxation in our assay. When metformin was coadministered to vessels that were cultured with the placental culture media, this impairment of vasodilation was reversed, and the bradykinin-mediated relaxation did not

differ from the normal culture media control (Figure 7).

Metformin enhances angiogenic sprouting from omental vessel explants Reduced angiogenesis is thought to contribute to placental hypoxia and contribute to the development of preeclampsia.23,77 We explored whether metformin could rescue the inhibition of angiogenesis that is caused by sFlt-1. We devised a human angiogenesis omental ring explant assay that was based on a mouse aorta explant model.73 We obtained omental biopsy specimens at cesarean delivery, isolated the omental vessels, dissected them into small explants, and cultured them with or without sFlt-1. We found a significant reduction in angiogenic sprouting from the vessels in the presence of sFlt-1 (Figure 8). However, metformin rescued sFlt-1-induced inhibition of angiogenic sprouting (Figure 8). A PowerPoint presentation that summarizes these data is included in supplementary materials.

MARCH 2016 American Journal of Obstetrics & Gynecology

356.e6

Original Research

ajog.org

OBSTETRICS

Comment Primary findings of the study Metformin reduces sFlt-1 and sENG secretion from primary human tissues, possibly by inhibiting the mitochondrial electron transport chain. Second, mitochondrial electron transport chain activity positivity regulates sFlt-1 secretion, and mitochondrial electron transport chain activity is increased in preterm preeclamptic placenta. By using assays to replicate the endothelial and vascular dysfunction that may be occurring in preeclampsia, we found that metformin reduces endothelial dysfunction, improves vasodilation, and is angiogenic. Collectively, our results suggest that metformin has potential to prevent or treat preeclampsia.

FIGURE 7

Effect of metformin on whole blood vessel dilation

sFlt-1 secretion is regulated through the mitochondrial electron transport chain; metformin possibly blocks this pathway This is the first report to show that sFlt-1 secretion is regulated through the mitochondrial electron transport chain. Using 2 inhibitors, we showed that blocking complex I or III significantly reduced sFlt-1 secretion. We have generated circumstantial evidence that suggests that metformin inhibits sFlt-1 and sENG secretion by inhibiting complex I of the mitochondrial electron transport chain. The administration of succinate (a substrate for complex II of the mitochondrial electron transport chain) rescued the metformin-induced reduction in sFlt-1 secretion but had no effect on sFlt-1 secretion when administered alone. However, a limitation of this experiment is that succinate is also thought to stabilize HIF1a directly.78 Furthermore, metformin is known to have other intracellular targets, such as activating AMP-activated protein kinase.79 Another consideration is whether metformin may be inhibiting syncytialization to reduce sFlt-1 secretion. We do not believe this is likely for the following reasons: (1) metformin had the same effect on sFlt-1 secretion from endothelial cells, which do not fuse; (2) we observed a reduction in sFlt-1 secretion from intact placental villous explants, and (3) there was a

Whole omental blood vessels that were cultured with placental villous explant media and were compared with those blood vessels that were cultured with normal media have impaired relaxation to the vasodilator bradykinin (1 mmol/L). Metformin (5 mmol/L) reversed the effect of placental villous explant media on vasodilation and improved relaxation to levels that were seen in those vessels cultured in normal media. The single asterisk indicates P < .05. pEC50, half maximal effective concentration. Brownfoot et al. Metformin decreases sFlt-1 and sENG, improves endothelial function, and is angiogenic. Am J Obstet Gynecol 2016.

reduction in sFlt-1 mRNA from primary trophoblasts that indicated a direct effect on the transcriptional machinery of the cell. We note that we have not demonstrated conclusively that metformin is decreasing sFlt-1 secretion via its effects on the mitochondria.

Mitochondrial electron transport chain activity is increased in preterm preeclamptic placenta We found an increase in activity in all electron transport chain complexes in preterm preeclamptic placentas, which was significant for the activity of complex II. There are 2 other reports that have examined mitochondrial electron transport chain activity. One report was concordant with our results, also reporting an increase in complex II

356.e7 American Journal of Obstetrics & Gynecology MARCH 2016

activity in preeclamptic placenta,80 while the other concluded mitochondrial electron transport chain activity was decreased.81 Both reports examined only 6 preeclamptic placentas that were obtained predominately from term pregnancies (10 of the 12 placentas were at term).80,81 Our study is the first to examine preterm preeclampsia specifically, and we investigated a considerably larger cohort of samples compared with those previous reports. 80,81 Given that our data suggests that preeclamptic placentas have increased electron transport chain activity, we speculate that enhanced mitochondrial electron transport chain activity in preeclampsia may have a role in the increased sFlt-1 and sENG secretion seen in this disease.82

ajog.org

OBSTETRICS

Original Research

FIGURE 8

Effect of metformin on angiogenesis

Omental whole vessel rings cultured in the presence of soluble fms-like tyrosine kinase 1 have reduced vessel outgrowth, which is restored with the addition of metformin (1 mmol/L). The white arrows point at vessel outgrowth. Representative micrographs are shown by the asterisk that indicates P < .05 sFlt-1, soluble fms-like tyrosine kinase 1. Brownfoot et al. Metformin decreases sFlt-1 and sENG, improves endothelial function, and is angiogenic. Am J Obstet Gynecol 2016.

The therapeutic implication of our findings that the mitochondria regulates sFlt-1 and sENG secretion is that screening other mitochondrial electron transport chain inhibitors may identify new therapeutic candidates for preeclampsia.

Metformin rescues endothelial dysfunction and impaired vasodilation specific to preeclampsia VCAM-1 is an inflammatory protein that is up-regulated in preeclampsia

and associated with endothelial dysfunction.61 Metformin previously has been shown to reduce endothelial cell inflammatory gene and protein expression83,84 and serum VCAM-1 concentrations in patients with diabetes mellitus59 and impaired glucose tolerance.60 We demonstrated that, in the presence of TNFa that is an inflammatory cytokine that is up-regulated in preeclampsia,76 metformin reduced VCAM-1 expression in endothelial cells. Metformin has been shown to vasodilate mouse and rat vessels that

are affected by insulin resistance85 and diabetes mellitus,62 respectively. We developed an assay to assess whether metformin vasodilates maternal vessels that are incubated in placentalconditioned media. Indeed, metformin appeared to enhance the vasodilatory response to bradykinin. We propose that this technique could be used more widely to assess the effect of small molecules on vasodilation in the context of pregnancy. Experiments to determine whether this effect was dependent on an intact endothelium

MARCH 2016 American Journal of Obstetrics & Gynecology

356.e8

Original Research

OBSTETRICS

were not performed and merit investigation in the future.

Metformin improves angiogenesis There is an imbalance towards an antiangiogenic state in preeclampsia.16,19,20,23,82 We devised an assay to examine the effects of metformin on angiogenesis in maternal vessels. There are a number of other angiogenesis assays that could have been performed. These include tube-forming assays that use endothelial cells alone or are cocultured with fibroblasts and mouse or rat aortic ring assays.73 We believe our assay more closely represents the dysfunction that is present in preeclampsia because it examines human vessels, examines angiogenesis in heterogeneous tissues that includes both endothelial and vascular smooth muscle; we induced dysfunction with sFlt-1, the antiangiogenic molecule that is of direct relevance to preeclampsia. We demonstrated that sFlt-1 reduced human omental vessel outgrowth and that metformin rescued this effect. These data suggest that metformin has angiogenic effects on maternal vessels.

Metformin has potential to prevent and treat preeclampsia Metformin is safe in pregnancy and currently is used to treat gestational diabetes mellitus. Interestingly, a randomized trial that compared metformin and insulin to treat gestational diabetes mellitus49 showed a nonsignificant reduction in the incidence of gestational hypertension (3.9% metformin arm, compared with 6.2% insulin treatment) and preeclampsia (5.5% metformin arm, 7% insulin treatment) among those treated with metformin. A randomized trial86 that assessed the effect of metformin on reducing perinatal morbidity in obese women did not identify a difference in preeclampsia risk. However, it is possible that under-dosing and poor compliance may be an explanation for the reason that there was no effect on the incidence of preeclampsia. Furthermore, a metaanalysis of 2 randomized control trials that assessed the effect of metformin on pregnancy outcome in

patients with polycystic ovarian syndrome also did not show a reduction in preeclampsia.87 It is important to note that incidence of hypertensive disorders of pregnancy and preeclampsia were not the primary outcome of these previous trials.49,86 In light of our work, we propose that a clinical trial that will examine whether metformin can prevent or treat preeclampsia is justified. Given our findings of reduced sFlt-1 and sENG placental secretion and improved endothelial dysfunction, we believe metformin may be able to reduce both placental and maternal vascular aspects of preeclampsia’

Conclusion We have performed preclinical studies using primary human tissues to show that mitochondrial electron transport chain activity is up-regulated in preterm preeclamptic placenta and that the mitochondrial electron transport chain regulate villous cytotrophoblast cells and endothelial cell sFlt-1 and sENG secretion. Metformin reduces sFlt-1 and sENG secretion by inhibiting complex 1 of the mitochondria. Furthermore, metformin reduces key features of endothelial dysfunction that are specific to preeclampsia and enhances angiogenesis. Metformin may be a novel preventative or be therapeutic for preeclampsia and has the potential to reduce the burden of this major pregnancy complication. n Acknowledgments We thank the research midwives, Gabrielle Pell, Genevieve Christophers, Rachel Murdoch, and Debra Jinks, and the patients at Mercy Hospital for Women for participating in this research.

References 1. Redman CW, Sargent IL. Latest advances in understanding preeclampsia. Science 2005;308:1592-4. 2. Sibai B, Dekker G, Kupferminc M. Preeclampsia. Lancet 2005;365:785-99. 3. Lisonkova S, Joseph KS. Incidence of preeclampsia: risk factors and outcomes associated with early- versus late-onset disease. Am J Obstet Gynecol 2013;209:544.e1-12. 4. Thornton C, Dahlen H, Korda A, Hennessy A. The incidence of preeclampsia and eclampsia and associated maternal mortality in Australia

356.e9 American Journal of Obstetrics & Gynecology MARCH 2016

ajog.org from population-linked datasets: 2000-2008. Am J Obstet Gynecol 2013;208:476.e1-5. 5. Barton JR, O’Brien JM, Bergauer NK, Jacques DL, Sibai BM. Mild gestational hypertension remote from term: progression and outcome. Am J Obstet Gynecol 2001;184: 979-83. 6. Valent AM, DeFranco EA, Allison A, et al. Expectant management of mild preeclampsia versus superimposed preeclampsia up to 37 weeks. Am J Obstet Gynecol 2015;212:515. e1-8. 7. Mol BW, Roberts CT, Thangaratinam S, Magee LA, de Groot CJ, Hofmeyr GJ. Preeclampsia. Lancet 2015. Epub ahead of print. 8. Lunell NO, Lewander R, Mamoun I, Nylund L, Sarby S, Thornstrom S. Uteroplacental blood flow in pregnancy induced hypertension. Scand J Clin Lab Invest Suppl 1984;169:28-35. 9. Karumanchi SA, Bdolah Y. Hypoxia and sFlt-1 in preeclampsia: the “chicken-and-egg” question. Endocrinology 2004;145:4835-7. 10. Brosens I, Pijnenborg R, Vercruysse L, Romero R. The “Great Obstetrical Syndromes” are associated with disorders of deep placentation. Am J Obstet Gynecol 2011;204:193-201. 11. Quinn MJ. Preeclampsia: 2 placental phenotypes, 1 etiology? Am J Obstet Gynecol 2014;211:313-4. 12. Fisher SJ. Why is placentation abnormal in preeclampsia? Am J Obstet Gynecol 2015;213: S115-22. 13. McCarthy FP, Kingdom JC, Kenny LC, Walsh SK. Animal models of preeclampsia; uses and limitations. Placenta 2011;32:413-9. 14. Nevo O, Soleymanlou N, Wu Y, et al. Increased expression of sFlt-1 in in vivo and in vitro models of human placental hypoxia is mediated by HIF-1. Am J Physiol Regul Integr Comp Physiol 2006;291:R1085-93. 15. Soleymanlou N, Jurisica I, Nevo O, et al. Molecular evidence of placental hypoxia in preeclampsia. J Clin Endocrinol Metab 2005;90: 4299-308. 16. Maynard S, Min JY, Merchan J, et al. Excess placental soluble fms-like tyrosine kinase 1 (sFlt-1) may contribute to endothelial dysfunction, hypertension, and proteinuria in preeclampsia. J Clin Invest 2003;111:649-58. 17. Maynard SE, Karumanchi SA. Angiogenic factors and preeclampsia. Semin Nephrol 2011;31:33-46. 18. Nagamatsu T, Fujii T, Kusumi M, et al. Cytotrophoblasts up-regulate soluble fmslike tyrosine kinase-1 expression under reduced oxygen: an implication for the placental vascular development and the pathophysiology of preeclampsia. Endocrinology 2004;145: 4838-45. 19. Venkatesha S, Toporsian M, Lam C, et al. Soluble endoglin contributes to the pathogenesis of preeclampsia. Nat Med 2006;12: 642-9. 20. Kim MY, Buyon JP, Guerra MM, et al. Angiogenic factor imbalance early in pregnancy predicts adverse outcomes in patients with lupus and antiphospholipid antibodies: results of

ajog.org the PROMISSE study. Am J Obstet Gynecol 2016;214:108.e1-14. 21. Lee MS, Cantonwine D, Little SE, et al. Angiogenic markers in pregnancies conceived through in vitro fertilization. Am J Obstet Gynecol 2015;213:212.e1-8. 22. Redman CW, Staff AC. Preeclampsia, biomarkers, syncytiotrophoblast stress, and placental capacity. Am J Obstet Gynecol 2015;213:S9.e1-4. 23. Powe CE, Levine RJ, Karumanchi SA. Preeclampsia, a disease of the maternal endothelium: the role of antiangiogenic factors and implications for later cardiovascular disease. Circulation 2011;123:2856-69. 24. Young BC, Levine RJ, Karumanchi SA. Pathogenesis of preeclampsia. Annu Rev Pathol 2010;5:173-92. 25. Chaiworapongsa T, Chaemsaithong P, Yeo L, Romero R. Pre-eclampsia: part 1: current understanding of its pathophysiology. Nat Rev Nephrol 2014;10:466-80. 26. Sacks GP, Studena K, Sargent K, Redman CW. Normal pregnancy and preeclampsia both produce inflammatory changes in peripheral blood leukocytes akin to those of sepsis. Am J Obstet Gynecol 1998;179:80-6. 27. Charnock-Jones DS. Placental hypoxia, endoplasmic reticulum stress and maternal endothelial sensitisation by sFLT1 in preeclampsia. J Reprod Immunol 2015. Epub ahead of print. 28. Roberts JM, Taylor RN, Musci TJ, Rodgers GM, Hubel CA, McLaughlin MK. Preeclampsia: an endothelial cell disorder. Am J Obstet Gynecol 1989;161:1200-4. 29. Wright D, Syngelaki A, Akolekar R, Poon LC, Nicolaides KH. Competing risks model in screening for preeclampsia by maternal characteristics and medical history. Am J Obstet Gynecol 2015;213:62.e1-10. 30. Cohen JM, Kramer MS, Platt RW, Basso O, Evans RW, Kahn SR. The association between maternal antioxidant levels in midpregnancy and preeclampsia. Am J Obstet Gynecol 2015;213: 695.e1-13. 31. Bahado-Singh RO, Syngelaki A, Akolekar R, et al. Validation of metabolomic models for prediction of early-onset preeclampsia. Am J Obstet Gynecol 2015;213:530.e1-10. 32. Sakowicz A, Hejduk P, Pietrucha T, et al. Finding NEMO in Preeclampsia. Am J Obstet Gynecol 2015. Epub ahead of print. 33. Ray JG, Booth GL, Alter DA, Vermeulen MJ. Prognosis after maternal placental events and revascularization: PAMPER study. Am J Obstet Gynecol 2016;214:106.e1-14. 34. Van Veen TR, Panerai RB, Haeri S, et al. Cerebral autoregulation in different hypertensive disorders of pregnancy. Am J Obstet Gynecol 2015;212:513.e1-7. 35. Major HD, Campbell RA, Silver RM, Branch DW, Weyrich AS. Synthesis of sFlt-1 by platelet-monocyte aggregates contributes to the pathogenesis of preeclampsia. Am J Obstet Gynecol 2014;210:547.e1-7.

OBSTETRICS 36. Chaiworapongsa T, Chaemsaithong P, Korzeniewski SJ, Yeo L, Romero R. Preeclampsia part 2: prediction, prevention and management. Nat Rev Nephrol 2014;10:531-40. 37. Sibai BM, Mercer BM, Schiff E, Friedman SA. Aggressive versus expectant management of severe preeclampsia at 28 to 32 weeks’ gestation: a randomized controlled trial. Am J Obstet Gynecol 1994;171:818-22. 38. Gibson KS, Waters TP, Bailit JL. A risk of waiting: the weekly incidence of hypertensive disorders and associated maternal and neonatal morbidity in low-risk term pregnancies. Am J Obstet Gynecol 2015. Epub ahead of print. 39. Mor O, Stavsky M, Yitshak-Sade M, et al. Early onset preeclampsia and cerebral palsy: a double hit model? Am J Obstet Gynecol 2016;214:105.e1-9. 40. Brownfoot FC, Tong S, Hannan NJ, et al. YC-1 reduces placental sFlt-1 and soluble endoglin production and decreases endothelial dysfunction: a possible therapeutic for preeclampsia. Mol Cell Endocrinol 2015;413: 202-8. 41. Rana S, Rajakumar A, Geahchan C, et al. Ouabain inhibits placental sFlt1 production by repressing HSP27-dependent HIF-1alpha pathway. FASEB J 2014;28:4323-34. 42. Rajakumar A, Conrad KP. Expression, ontogeny, and regulation of hypoxia-inducible transcription factors in the human placenta. Biol Reprod 2000;63:559-69. 43. Kanasaki K, Palmsten K, Sugimoto H, et al. Deficiency in catechol-O-methyltransferase and 2-methoxyoestradiol is associated with preeclampsia. Nature 2008;453:1117-21. 44. Tal R. The role of hypoxia and hypoxiainducible factor-1alpha in preeclampsia pathogenesis. Biol Reprod 2012;87:134. 45. Stasch JP, Evgenov OV. Soluble guanylate cyclase stimulators in pulmonary hypertension. Handb Exp Pharmacol 2013;218: 279-313. 46. Stasch JP, Pacher P, Evgenov OV. Soluble guanylate cyclase as an emerging therapeutic target in cardiopulmonary disease. Circulation 2011;123:2263-73. 47. Yeo EJ, Chun YS, Cho YS, et al. YC-1: a potential anticancer drug targeting hypoxiainducible factor 1. J Natl Cancer Inst 2003;95: 516-25. 48. Zhang H, Qian DZ, Tan YS, et al. Digoxin and other cardiac glycosides inhibit HIF-1alpha synthesis and block tumor growth. Proc Natl Acad Sci 2008;105:19579-86. 49. Rowan JA, Hague WM, Gao W, Battin MR, Moore MP, Mi GTI. Metformin versus insulin for the treatment of gestational diabetes. N Engl J Med 2008;358:2003-15. 50. Spaulonci CP, Bernardes LS, Trindade TC, Zugaib M, Francisco RP. Randomized trial of metformin vs insulin in the management of gestational diabetes. Am J Obstet Gynecol 2013;209:34.e1-7. 51. Dhulkotia JS, Ola B, Fraser R, Farrell T. Oral hypoglycemic agents vs insulin in management

Original Research

of gestational diabetes: a systematic review and metaanalysis. Am J Obstet Gynecol 2010;203: 457.e1-9. 52. Jiralerspong S, Palla SL, Giordano SH, et al. Metformin and pathologic complete responses to neoadjuvant chemotherapy in diabetic patients with breast cancer. J Clin Oncol 2009;27: 3297-302. 53. Garcia A, Tisman G. Metformin, B(12), and enhanced breast cancer response to chemotherapy. J Clin Oncol 2010;28:e19-20. 54. Ranasinghe WK, Sengupta S, Williams S, et al. The effects of nonspecific HIF1alpha inhibitors on development of castrate resistance and metastases in prostate cancer. Cancer Med 2014;3:245-51. 55. Wheaton WW, Weinberg SE, Hamanaka RB, et al. Metformin inhibits mitochondrial complex I of cancer cells to reduce tumorigenesis. eLife 2014;3:e02242. 56. Andrzejewski S, Gravel SP, Pollak M, St-Pierre J. Metformin directly acts on mitochondria to alter cellular bioenergetics. Cancer Metab 2014;2:12. 57. Mather KJ, Verma S, Anderson TJ. Improved endothelial function with metformin in type 2 diabetes mellitus. J Am Coll Cardiol 2001;37:1344-50. 58. Jensterle M, Sebestjen M, Janez A, et al. Improvement of endothelial function with metformin and rosiglitazone treatment in women with polycystic ovary syndrome. Eur J Endocrinol 2008;159:399-406. 59. Kruszelnicka O, Chyrchel B, Golay A, Surdacki A. Differential associations of circulating asymmetric dimethylarginine and cell adhesion molecules with metformin use in patients with type 2 diabetes mellitus and stable coronary artery disease. Amino Acids 2015;47: 1951-9. 60. Caballero AE, Delgado A, AguilarSalinas CA, et al. The differential effects of metformin on markers of endothelial activation and inflammation in subjects with impaired glucose tolerance: a placebo-controlled, randomized clinical trial. J Clin Endocrinol Metab 2004;89:3943-8. 61. Austgulen R, Lien E, Vince G, Redman CW. Increased maternal plasma levels of soluble adhesion molecules (ICAM-1, VCAM-1, Eselectin) in preeclampsia. Eur J Obstet Gynecol Reprod Biol 1997;71:53-8. 62. Sena CM, Matafome P, Louro T, Nunes E, Fernandes R, Seica RM. Metformin restores endothelial function in aorta of diabetic rats. Br J Pharmacol 2011;163:424-37. 63. Brownfoot FC, Hannan N, Onda K, Tong S, Kaitu’u-Lino T. Soluble endoglin production is upregulated by oxysterols but not quenched by pravastatin in primary placental and endothelial cells. Placenta 2014;35:724-31. 64. Kaitu’u-Lino TJ, Tong S, Beard S, et al. Characterization of protocols for primary trophoblast purification, optimized for functional investigation of sFlt-1 and soluble endoglin. Pregnancy Hypertens 2014;4:287-95.

MARCH 2016 American Journal of Obstetrics & Gynecology

356.e10

Original Research

OBSTETRICS

65. American College of Obstetricians and Gynecologists. Report of the American College of Obstetricians and Gynecologists’ task force on hypertension in pregnancy. Obstet Gynecol 2013;122:1122-31. 66. Burton GJ, Sebire NJ, Myatt L, et al. Optimising sample collection for placental research. Placenta 2014;35:9-22. 67. Brownfoot FC, Tong S, Hannan NJ, et al. Effects of pravastatin on human placenta, endothelium, and women with severe preeclampsia. Hypertension 2015;66:687-97; discussion 445. 68. Frazier AE, Thorburn DR. Biochemical analyses of the electron transport chain complexes by spectrophotometry. Methods Mol Biol 2012;837:49-62. 69. Hastie R, Lappas M. The effect of preexisting maternal obesity and diabetes on placental mitochondrial content and electron transport chain activity. Placenta 2014;35: 673-83. 70. Wareing M, Myers JE, O’Hara M, et al. Phosphodiesterase-5 inhibitors and omental and placental small artery function in normal pregnancy and pre-eclampsia. Eur J Obstet Gynecol Reprod Biol 2006;127:41-9. 71. Whitehead CL, Palmer KR, Nilsson U, et al. Placental expression of a novel primate-specific splice variant of sFlt-1 is upregulated in pregnancies complicated by severe early onset preeclampsia. BJOG 2011;118:1268-71. 72. Senadheera S, Bertrand PP, Grayson TH, et al. Enhanced contractility in pregnancy is associated with augmented TRPC3, L-type, and T-type voltage-dependent calcium channel function in rat uterine radial artery. Am J Physiol Regul Integr Comp Physiol 2013;305:R917-26. 73. Baker M, Robinson SD, Lechertier T, et al. Use of the mouse aortic ring assay to study angiogenesis. Nat Protoc 2012;7:89-104. 74. Jebbink J, Keijser R, Veenboer G, van der Post J, Ris-Stalpers C, Afink G. Expression of placental FLT1 transcript variants relates to both

gestational hypertensive disease and fetal growth. Hypertension 2011;58:70-6. 75. Chaiworapongsa T, Romero R, Yoshimatsu J, et al. Soluble adhesion molecule profile in normal pregnancy and pre-eclampsia. J Matern Fetal Neonatal Med 2002;12:19-27. 76. Borzychowski AM, Sargent IL, Redman CW. Inflammation and pre-eclampsia. Semin Fetal Neonat Med 2006;11:309-16. 77. Ahmad S, Ahmed A. Elevated placental soluble vascular endothelial growth factor receptor-1 inhibits angiogenesis in preeclampsia. Circ Res 2004;95:884-91. 78. Mills E, O’Neill LA. Succinate: a metabolic signal in inflammation. Trends Cell Biol 2014;24: 313-20. 79. Takiyama Y, Harumi T, Watanabe J, et al. Tubular injury in a rat model of type 2 diabetes is prevented by metformin: a possible role of HIF1alpha expression and oxygen metabolism. Diabetes 2011;60:981-92. 80. Mando C, De Palma C, Stampalija T, et al. Placental mitochondrial content and function in intrauterine growth restriction and preeclampsia. Am J Physiol Endocrinol Metab 2014;306: E404-13. 81. Muralimanoharan S, Maloyan A, Mele J, Guo C, Myatt LG, Myatt L. MIR-210 modulates mitochondrial respiration in placenta with preeclampsia. Placenta 2012;33:816-23. 82. Levine RJ, Maynard SE, Qian C, et al. Circulating angiogenic factors and the risk of preeclampsia. N Engl J Med 2004;350:672-83. 83. Hattori Y, Suzuki K, Hattori S, Kasai K. Metformin inhibits cytokine-induced nuclear factor kappaB activation via AMP-activated protein kinase activation in vascular endothelial cells. Hypertension 2006;47:1183-8. 84. Viollet B, Guigas B, Sanz Garcia N, Leclerc J, Foretz M, Andreelli F. Cellular and molecular mechanisms of metformin: an overview. Clin Sci (Lond) 2012;122:253-70. 85. Katakam PV, Ujhelyi MR, Hoenig M, Miller AW. Metformin improves vascular function

356.e11 American Journal of Obstetrics & Gynecology MARCH 2016

ajog.org in insulin-resistant rats. Hypertension 2000;35: 108-12. 86. Chiswick C, Reynolds RM, Denison F, et al. Effect of metformin on maternal and fetal outcomes in obese pregnant women (EMPOWaR): a randomised, double-blind, placebo-controlled trial. Lancet Diabetes Endocrinol 2015;3: 778-86. 87. Vanky E, de Zegher F, Diaz M, Ibanez L, Carlsen SM. On the potential of metformin to prevent preterm delivery in women with polycystic ovary syndrome: an epi-analysis. Acta Obstet Gynecol Scand 2012;91: 1460-4.

Author and article information From the Translational Obstetrics Group, Department of Obstetrics and Gynaecology, University of Melbourne, Mercy Hospital for Women (Drs Brownfoot, Hannan, Kaitu’u-Lino, and Tong and Ms Hastie, Cannon, and Tuohey) Heidelberg, Victoria, Australia; Biosciences, University of Melbourne, (Drs Parry and Senadheera), Parkville, Victoria, Australia; and Department of Obstetrics and Gynecology, Laboratory of Reproductive Biology, Faculty of Medicine, Universidad de Los Andes, Santiago, Chile (Dr Illanes). 1 These authors have contributed equally to this article. Received Nov. 30, 2015; revised Dec. 15, 2015; accepted Dec. 15, 2015. Supported by The National Health and Medical Research Council of Australia (NHMRC; #1048707, #1046484, #1101871) and an Arthur Wilson RANZCOG scholarship; by an Australian Postgraduate Award and an AVANT scholarship (F.B); by a CR Roper Research Fellowship (N.J.R.); the NHMRC provided salary support (#1050765 [S.T.]; #1062418 [T.K.L.]; #628549 [S.S.]). The funders had no role in study design, data collection, analysis, decision to publish or the preparation of the manuscript. The authors report no conflict of interest. Corresponding author: Stephen Tong, MBBS, PhD. [email protected]

ajog.org Supplementary: Methods Cell culture media Placental tissues, endothelial cells, and omental vessels were cultured in a variety of media. Human umbilical vein endothelial cells (HUVECs) were cultured in M199 media (Life Technologies, Mulgrave, Australia) that contained 10% fetal calf serum, 1% antibioticantimicotic (Life Technologies), 1% endothelial cell growth factor (Sigma Chemical Company, St. Louis, MO), and 1% heparin; primary trophoblasts and placental explants were cultured in DMEM high Glutamax (Life Technologies) that contained 10% fetal calf serum (Sigma Chemical Company) and 1% antibiotic-antimicotic (Life Technologies).

Isolation of HUVECs and trophoblasts from human placenta Placentas were obtained from women who had undergone elective cesarean delivery at term. HUVECs were isolated by obtaining the umbilical cord, infusing it with 10 mL (1 mg/mL) of collagenase (Worthington, Lakewood, NJ), and flushing the cells through. as previously described.1 Primary trophoblasts were isolated by scraping 50 g of cotyledon tissue from the placenta, subjecting it to enzyme digestion buffer, and separating cells with the use of a discontinuous Percoll gradient then subjecting them to negative selection with CD9 to remove contaminating cells, as previously described.2

Cell viability assays (MTS assay and calcein stain) Cell viability assays were performed for all HUVEC and primary trophoblast functional experiments with the use of either CellTiter 96-Aquesous One solution (Promega, Madison, WI) or calcein stain (Merck Millipore, Darmstadt, Germany) and were quantitated with a Fluostar omega fluorescent plate reader (BMG Labtech, Victoria, Australia).

Whole omental vessel collection, pressure myography, and omental vessel explant outgrowth Omental vessel tissue collection

Omental tissue was obtained from women with normal pregnancies

OBSTETRICS who had undergone elective cesarean delivery at term. The omental biopsy specimens were transferred to cold Krebs’ physiologic solution (PSS) that contained 112 mmol/L NaCl, 25 mmol/ L NaHCO3, 4.7 mmol/L KCl, 1.2 mmol/ L MgSO4.7H2O, 0.7 mmol/L KH2PO4, 10 mmol/L HEPES, 11.6 mmol/L Dglucose, and 2.5 mmol/L CaCl2.2H2O; pH 7.4; Sigma Chemical Company) and washed 3 times to remove the anesthetic. The average inner lumen diameters of arteries that were used were 280-320 mm at 60 mm Hg in zero calcium. All experiments were performed at <20 hours after surgery. Whole vessel pressure myography

Whole omental arteries were cleaned carefully of connective and adipose tissue. Arteries were cultured for 3 hours at 4 C in treatment groups (control group, M199 media [Sigma Chemical Company] or treatment groups, M199 with 25% conditioned media that had been exposed to placental explants for 24 hours and M199 with 25% conditioned media with 5 mmol/L metformin [Sigma Chemical Company]) for incubation studies. All arteries were mounted in a pressure myograph organ bath (Living Systems Instrumentation, Burlington, VT) and continuously superfused with PSS (37 C) at a rate of 4 mL/ min. In the absence of intraluminal flow, arteries were pressurized gradually to 60 mm Hg, warmed to 37 C, and checked for pressure leaks over a 50-minute equilibration period. The outer diameters were measured through video microscopy (Diamtrak Software, Adelaide, SA, Australia). Before the start of each experiment, the viability of the smooth muscle was tested with potassium physiologic saline solution (KPSS, 100 mmol/L) or thromboxane agonist 9,11-dideoxy-9a,11a-methanoepoxy prostaglandin F2a (U46619, 1 mmol/L; Sigma Chemical Company) washed with PSS to regain basal diameter. Arteries were preconstricted with thromboxane agonist U46619; endothelial function was tested by applying bradykinin (10 mmol/L; Auspep, West Melbourne, Australia), and washed to

Original Research

regain basal diameter. Arteries were then submaximally preconstricted (approximately 70% of maximal response to KPSS or U46619) with U46619 (Sigma Chemical Company); vascular reactivity to drugs were assessed in the following manner. In incubation studies, the effect of treatments on vascular reactivity was assessed with bradykinin (0.01-1000 nmol/L). At the end of the experiment, smooth muscle integrity was tested in response to the endotheliumindependent, nitric oxide donor, sodium nitroprusside (10 mmol/L). The organ bath was then washed out with PSS, and Ca2þ-free Krebs solution was added for 20 minutes to determine the maximum diameter. Ca2þ-free Krebs solution contained no added CaCl2.2H2O but contained EGTA (2 mmol/L). Human omental vessel explants

To assess changes in angiogenesis potential, a human omental vessel outgrowth assay was conceived and performed. Omental samples were collected from patients who had had an elective cesarean delivery at term. An arteriole was dissected carefully out of the fat, and the vessel was flushed with a 26G needle with OptiMEM (Life Technologies). The vessel was cut into 0.5-mm rings and serum-starved in OptiMEM at 20% O2, 5% CO2 at 37 C overnight. The ring was embedded in a collagen matrix (1 mg/mL in DMEM, pH adjusted so slightly basic using NaOH) in a 96-well plate with 1 omental ring per well. Rings were treated in media that contained OptiMEM with 2.5% fetal calf serum (Sigma Chemical Company) and 1% antibiotic antimycotic (Life Technologies)  250 ng/ mL soluble fms-like tyrosine kinase 1 (sFlt-1) and 0 or 1 mmol/L metformin (Sigma Chemical Company). For each patient sample (n ¼ 4), we used 5 rings (technical replicates) per treatment group. Treatments were changed every 48 hours, and the experiment was continued for 120 hours. Calcein AM (Merck Millipore, Darmstadt, Germany) was added to the wells for 45 minutes at 37 C, and images were obtained at the same magnification (40) with the EVOS FL microscope (Life

MARCH 2016 American Journal of Obstetrics & Gynecology

356.e12

Original Research

OBSTETRICS

Technologies). Vessel outgrowth at the completion of the experiment was determined by a calculation of the area of growth with the computer program Image J (http://imagej.nih.gov/ij/).

Assessment of mitochondrial electron transport chain activity in preterm preeclamptic placenta and preterm gestationally matched placenta Placental collection for mitochondrial electron transport chain activity assay

Placental tissue was obtained from 23 women with severe preterm preeclampsia that was defined with the American College of Obstetricians and Gynecologists 2013 guidelines3 and from 25 gestationally matched preterm control subjects who delivered either because of spontaneous preterm labor without evidence of infection (confirmed on histopathologic evaluation), hypertensive disease, or maternal comorbidities. Baseline patient clinical characteristics are detailed in the Supplementary Table. Placental tissue was processed within 30 minutes of delivery. Placental tissue, excluding fetal membranes, was washed in sterile phosphate buffered saline, frozen within 15 minutes, and stored at e80 C. Isolation of mitochondria from preeclamptic and preterm placental samples

Intact mitochondria were isolated, as previously described.4 Briefly, to isolate intact mitochondria, the placental tissue was homogenized in ice-cold Zheng buffer (210 mmol/L mannitol, 70 mmol/L sucrose, 5 mmol/L HEPES, and 1 mmol/L EGTA, pH 7.2) and centrifuged at 600g for 10 minutes at 4 C; the supernatant was transferred to an icecold microcentrifuge tube and freeze thawed 3 times in a methanol bath (1-minute freeze followed by 4-minute thaw at room temperature).4 Electron transport chain activity

All mitochondrial respiratory complex activity was performed by spectrophotometric assays, as previously

described.4 The Fluostar Omega Fluorescent Plate Reader (BMG Labtech, Victoria, Australia) was used to measure absorbance. First, the activity of the mitochondrial matrix enzyme citrate synthase was measured by the rate of free sulfhydryl group production. This was determined by the administration of thiol reagent 5,5-dithio-bis-(2-nitrobenzoic acid) to placenta, which reacts with free sulfhydryl to produce thio nitrobenzoate acid. Placental samples were sonicated on ice for 1 minute to disrupt the mitochondrial membrane and 1 mmol/L 5,5dithio-bis-(2-nitrobenzoic acid) and 12.5 mmol/L acetyl CoA were added to 10 mL of placental sample or blank, and the absorbance was read every 15 seconds for 2 minutes at 412 nmol/L. To begin the reaction, 5 mmol/L of oxaloacetic acid was then added, and absorbance was read every 15 seconds at 412 nmol/L for 4 minutes. The activity of complex I was measured by the determination of the transfer of electrons from b-nicotinamide adenine dinucleotide to coenzyme Q. This was achieved by the activation of complex 1 by adding 5mmol/L b-nicotinamide adenine dinucleotide, 50 mmol/L potassium cyanide, 0.5 mmol/L antimycin A, 10% bovine serum albumin, and 5.0 mmol/L oxidized coenzyme Q to 10 mL of placental sample or blank and the inhibiting of the complex by the addition of rotenone 0.125 mmol/L and reading the absorbance every 15 seconds for 3 minutes at 340 nmol/L. Complex II activity was next assessed by the measurement of the reduction of coenzyme Q on oxidation of succinate to fumarate. Baseline activity was assessed initially by the addition of 50 mmol/L potassium cyanide, 0.5 mmol/L antimycin A, 100 mmol/L succinate, and 0.125 mmol/L rotenone to 10 mL of placental sample or blank; the absorbance was read every 15 seconds for 2 minutes at 280 nmol/L; then 5 mmol/L of coenzyme Q was added, and the rate of reduction was assessed by the reading of absorbance every 15 seconds for 3 minutes at 280 nmol/L. The activity of complex III was assessed next by the measurement of the reduced form of decyl benzoquinone,

356.e13 American Journal of Obstetrics & Gynecology MARCH 2016

ajog.org which is a short chain of analogue of endogenous ubiquinone that donates electrons to cytochrome c. Baseline activity was assessed by the addition of 50 mmol/L potassium cyanide, 0.125 mmol/L rotenone, 10% bovine serum albumin, 25 mmol/L n-dodecyl-b-DMaltoside, 10 mmol/L reduced duroquinone, and 1 mmol/L oxidized cytochrome c to 10 mL of placental sample or blank; absorbance read every 12 seconds for 4 minutes at 550 nmol/L; 25 mmol/L of L-ascorbate was added to complete the reaction, and the absorbance was read every 15 seconds for 5 minutes at 550 nmol/L. The activity of complex IV, the terminal enzyme in the electron transport chain, was measured by the assessment of the loss of reduced cytochrome c as it is oxidized. Baseline activity was determined by the addition of reduced cytochrome c to 10 mL of placental sample or blank; the absorbance was read every 15 seconds for 5 minutes at 550 nmol/L; then 50 mmol/L of potassium ferricyanide was added to stop the oxidation of cytochrome c, and the absorbance was read every 15 seconds for 2 minutes at 550 nmol/L. Electron transport chain activity was calculated as first-rate constants (first rate constant per minute per milligram) that were determined by the subtraction of the final absorbance reading after the addition of inhibiting substance from the absorbance before and the data plotted as a slope against time. The slope is denoted the first-order rate constant. Activity was normalized to milligrams of tissue or citrate synthase (maker of mitochondrial content). Enzyme-linked immunosorbent assay (ELISA) analysis

Concentrations of sFlt-1 and soluble endoglin were measured in conditioned cell/tissue culture media with the use of the DuoSet VEGF R1/Flt-1 kit (R&D Systems by Bioscience, Waterloo, Australia) and a DuoSet Human Endoglin CD/105 ELISA kit (R&D Systems) according to the manufacturer’s instructions. The coefficient of variation for our ELISA sFlt-1 was 3.4% and for sENG was 2.4%.

ajog.org Real timeepolymerase chain reaction (PCR)

RNA was extracted from placental explants and HUVECs using an RNeasy mini kit (Qiagen, Valencia, CA) and quantified using the Nanodrop ND 1000 spectrophotometer (NanoDrop technologies Inc, Wilmington, DE). 0.2 mg of RNA was converted to complementary DNA with the use of Applied Biosystems high capacity complementary DNA reverse transcriptase kit (Life Technologies), as per manufacturer guidelines. The coefficient of variation for the PCR for vascular cell adhesion molecule 1 was 5.6%; the coefficient of variation for the PCR sFlt-1-i13 and e15a was 3.24%. A taqman gene expression assay was performed for VCAM-1 (Life Technologies). Real timeePCR was performed on the CFX 384 (Bio-Rad, Hercules, CA) with the use of FAM-labeled Taqman universal PCR mastermix (Life Technologies) with the following run conditions: 50 C for 2 minutes; 95 C for 10 minutes; 95 C for 15 seconds, and 60 C for 1 minute (40 cycles). Sybr gene expression assay for sFlt-1 e15a and sFlt1 i13 was used. Primers were designed as previously described (Geneworks, South Australia, Australia).5 RT-PCR was performed with the following run conditions: 95 C for 20 minutes; 95 C for 0.01 minutes, 60 C for 20 minutes, 95 C for 1 minute (39 cycles), melt curve

OBSTETRICS 65 C to 95 C at 0.05 C increments at 0.05 seconds. All data were normalized to GAPDH as an internal control and calibrated against the average Ct of the control samples. Results were expressed as fold change from control.

Statistical analysis Technical triplicates were performed for each experiment, with a minimum of 3 biologic replicates (different patient samples) for each in vitro study. Data were tested for normal distribution and analyzed with the use of the appropriate parametric or nonparametric test. The statistical software we used was GraphPad Prism 6 (GraphPad Software, La Jolla, CA). When 3 groups were compared, a 1-way analysis of variance (for parametric data) or Kruskal-Wallis test (for nonparametric data) was used. Post-hoc analysis was carried out with either the Tukey (parametric) or Dunn’s test (nonparametric). When 2 groups were analyzed, either an unpaired t-test (parametric) or a Mann-Whitney test (nonparametric) was used. Concentration response curves from omental arteries were fitted to a sigmoidal curve with nonlinear regression to calculate the sensitivity of each agonist (pEC50) or maximum relaxation (Emax). All data were expressed as mean  SEM; probability values of <.05 were considered significant.

Original Research

References 1. Brownfoot FC, Hannan N, Onda K, Tong S, Kaitu’u-Lino T. Soluble endoglin production is upregulated by oxysterols but not quenched by pravastatin in primary placental and endothelial cells. Placenta 2014;35:724-31. 2. Kaitu’u-Lino TJ, Tong S, Beard S, et al. Characterization of protocols for primary trophoblast purification, optimized for functional investigation of sFlt-1 and soluble endoglin. Pregnancy Hypertens 2014;4:287-95. 3. American College of Obstetricians and Gynecologists. Report of the American College of Obstetricians and Gynecologists’ task force on hypertension in pregnancy. Obstet Gynecol 2013;122:1122-31. 4. Frazier AE, Thorburn DR. Biochemical analyses of the electron transport chain complexes by spectrophotometry. Methods Mol Biol 2012;837:49-62. 5. Whitehead CL, Palmer KR, Nilsson U, et al. Placental expression of a novel primate-specific splice variant of sFlt-1 is upregulated in pregnancies complicated by severe early onset preeclampsia. BJOG 2011;118:1268-71.

MARCH 2016 American Journal of Obstetrics & Gynecology

356.e14

Original Research

ajog.org

OBSTETRICS

SUPPLEMENTARY TABLE

Baseline characteristics of placenta collected that compares mitochondrial electron transport chain activity in placentas that were complicated by preterm preeclampsia vs controls (normotensive, preterm delivery) Characteristic

Preeclampsia (n ¼ 23)

Preterm (n ¼ 25)

P value

Maternal age, y

30.6  5.9

31.1  6.6

.78

Gestation, wka

29.6  2.3

29.4  2.4

.75

28.3  5.5

27.6  7.9

.58

169.9  15.4

119.8  14.7

< .0001

103.6  9.9

70.4  14.1

< .0001

1190  448.5

a

2a

Body mass index, kg/m

Systolic blood pressure, mm Hg

a

Diastolic blood pressure, mm Hg

a

1393  475.7

.14

Nulliparity, n (%)

15 (65)

9 (36)

.08

Intrauterine growth restriction (birthweight, <10%), n (%)

11 (48)

10 (40)

.77

a

Birthweight, g

Data are shown is mean  standard deviation. Brownfoot et al. Metformin decreases sFlt-1 and sENG, improves endothelial function, and is angiogenic. Am J Obstet Gynecol 2016.

a

356.e15 American Journal of Obstetrics & Gynecology MARCH 2016