Evidence for the functional activity of hypoxia-inducible transcription factors overexpressed in preeclamptic placentae

Placenta (2004), 25, 763–769 doi:10.1016/j.placenta.2004.02.011

Evidence for the Functional Activity of Hypoxia-inducible Transcription Factors Overexpressed in Preeclamptic Placentae A. Rajakumara, H. M. Brandona, A. Daftarya, R. Nessb and K. P. Conrada,c,* a Department of Obstetrics, Gynecology and Reproductive Sciences, Magee Womens Research Institute, University of Pittsburgh School of Medicine, Pittsburgh, PA 15213, USA; b Department of Epidemiology, Magee Womens Research Institute, University of Pittsburgh School of Medicine, Pittsburgh, PA 15213, USA; c Department of Cell Biology and Physiology, Magee Womens Research Institute, University of Pittsburgh School of Medicine, Pittsburgh, PA 15213, USA Paper accepted 23 February 2004

Placentas from women with preeclampsia overexpress the hypoxia-inducible transcription factor proteins, HIF-1a and -2a (Rajakumar 2001, Biol Reprod 64; p499–506 and p1019–1020). As a first step in evaluating whether HIF-a overexpressed in preeclamptic placentae is capable of transactivation, we tested its ability to bind to the DNA hypoxia response element (HRE). Six pairs of normal and preeclamptic placentae obtained by cesarean section were investigated. Three biopsy sites per placenta were analyzed. We first confirmed HIF-1a protein overexpression in the preeclamptic placentae using Western analysis. The ratios of the arbitrary densitometry units for HIF-1a protein from the preeclamptic and normal placentae (PE/NP) in the three biopsy sites were: 1:9 G 0:3, 1:7 G 0:2 and 1:8 G 0:2, each p ! 0:05 vs 1.0. (A ratio of O1.0 indicates that HIF-1a protein expression in placentas of women with PE exceeds that in placentas of NP women.) Conventional methods for extracting nuclear proteins and subsequent analysis by electrophoretic mobility shift assay were not suited for the frozen, archived samples (data not shown). Therefore, we employed DNA affinity chromatography using a biotinylated oligonucleotide representing the HRE of the erythropoietin gene coupled to streptavidin-coated Dynabeads. The HRE-bound proteins were then characterized by Western blot analysis. The PE/NP ratios of HRE-bound HIF-1a in the three biopsy sites from the six pairs of normal and preeclamptic placentae were 1:7 G 0:2, 2:1 G 0:4 and 2:4 G 0:5, each p ! 0:05 vs 1.0. Having established DNA-binding potential at least in vitro, we subsequently analyzed three proteins that have been shown to be regulated by HIF-a as downstream, molecular markers of HIF-1a activity in vivo. VEGF receptor Flt-1 and Flk-1 play key roles in angiogenesis. Tyrosine hydroxylase is the ratelimiting enzyme in catecholamine synthesis. All three genes contain functional HRE in their promoter sequences. Total proteins were extracted from the same biopsy samples that were used for total and HRE-bound HIF-1a. Using specific antibodies we performed Western analysis and the levels of these three proteins were quantitated. The Flt-1 and tyrosine hydroxylase proteins were significantly higher, and Flk-1 significantly lower in placentae from preeclamptic compared to normal pregnancies. In summary, HIF-1a protein overexpressed in preeclamptic placentae is capable of binding to its DNA recognition sequence in vitro, and modulates gene expression in vivo. Placenta (2004), 25, 763–769 Ó 2004 Elsevier Ltd. All rights reserved.

INTRODUCTION Placental hypoxia is likely to play an important role in normal placental development and pathology. The hypoxia-inducible transcription factors, HIF-1a and -2a, are major transducers of hypoxia signaling in most tissues including the human placenta leading to the regulation of numerous genes [1–3]. Consistent with the relative hypoxic environment of the placental intervillous space measured during the first trimester of * Corresponding author address: Magee Womens Institute, 204 Craft Avenue, Pittsburg, PA 15213, USA. Tel.: C1-412-641-6019; fax: C1-412-641-1503. E-mail address: [email protected] (K.P. Conrad). 0143–4004/$–see front matter

pregnancy [4–6], both HIF-1a and -2a protein (but not mRNA) are increased in the syncytiotrophoblast, villous cytotrophoblast and fetoplacental vasculature. At the end of the first trimester when intervillous blood flow and placental oxygenation begin to increase [5,6], the HIF-alpha proteins are correspondingly downregulated [1]. Both HIF-1a and -2a proteins are significantly overexpressed in preeclamptic placentae where they localize primarily to the nuclei of syncytiotrophoblast and fetoplacental blood vessels suggesting transactivational activity [7,8]. Because villous explants from preeclamptic placentae fail to adequately downregulate HIF-1a and -2a upon oxygenation in vitro, this metabolic abnormality may contribute to their overexpression in vivo [9]. In women destined to develop preeclampsia, Ó 2004 Elsevier Ltd. All rights reserved.

Placenta (2004), Vol. 25

764

overexpression of HIF-alpha proteins in the placenta likely contributes to the dysregulation of numerous genes that perturbs placental function leading to impairment of trophoblast invasion during early gestation and elaboration of various proteins deleterious to the endothelium during late gestation. To begin investigating the molecular and functional consequences of this fundamental molecular abnormality, we tested whether the overexpressed HIF-1a protein in preeclamptic placentae (i) has the capacity for DNA binding, a prerequisite for transactivational activity, and (ii) corresponds with expression of oxygen-regulated genes in vivo that are known to be regulated by HIF-a. Specifically, we analyzed the expression of the VEGF receptors Flt-1 and Flk-1 that are involved in angiogenesis [10], as well as tyrosine hydroxylase which is the rate-limiting enzyme in catecholamine synthesis [11]. These gene products have been shown to contain functional hypoxia response elements (HRE) in their promoter sequence [12–14].

METHODS Reagents Dynabeads M-280 streptavidin was purchased from Dynal (Dynal Inc., Lake Success, NY, USA). Oligonucleotides representing the hypoxia response element (sense 5#-GCCCTACGTGCTGTCTCA-3# and antisense 5#-TGAGACAGCACGTAGGGC-3# (concatemer, three repeats)) were synthesized and biotinylated at the DNA synthesis facility of University of Pittsburgh. Cell culture Pheochromocytoma (PC-12) cells were purchased from ATCC (American Type Culture Collection, Manassas, VA) and were grown in Kaighn’s modified Ham’s F12 medium with 2 mM L-glutamine, 1.5 g/l sodium bicarbonate, 15% horse serum, 2.5% fetal bovine serum and 100 U/ml each of penicillin and streptomycin. Placental collection and processing Placentae were obtained from women with normal pregnancies and women with preeclampsia undergoing elective cesarean section at term under approval of the Institutional Internal Review Board of Magee-Womens Hospital. The diagnosis of preeclampsia was made based on the Working Group Report on High Blood Pressure in Pregnancy [15]. Gestational blood pressure elevation was defined as systolic blood pressure O140 or diastolic pressure O90 mmHg. Furthermore, the subjects were normotensive during early pregnancy or postpartum with no history of chronic hypertension. The preeclamptic subjects also had proteinuria, O300 mg/24 h or O2C on dipstick. Villous biopsy samples were collected using a systematic and unbiased approach as described in our previous studies [16]. After quickly rinsing the harvested tissues in three changes of

ice-cold sterile saline and blotting on sterile gauze to remove excess fluid, they were flash frozen in liquid nitrogen. Placental villous explants were prepared as described by Benyo et al. [16] with modifications, in order to standardize the technique for HIF–HRE binding using DNA affinity chromatography. Several cotyledons from a normal term placenta were excised at random and rinsed extensively in sterile saline to remove blood. Decidua basalis and large vessels were removed from the villous placenta by blunt dissection. The villous tissue was then finely dissected into 5–10 mg pieces while in a bath of sterile saline. The pieces were extensively washed two or three more times before explant culture. Villous explant culture Fifty milligrams of villous tissue was placed into each well of a 24-well plate (Becton Dickinson, Franklin Lakes, NJ) containing 1.0 ml of Medium 199 (Mediatech, Cellgro Herndon, VA) supplemented with 10% Fetal Bovine Serum (FBS, Summit Technology, Ft. Collins, CO) and 100 U/ml penicillin and streptomycin and 50 mg/ml gentamycin. Explants were incubated at 37 (C for a 12 h period on an orbital shaker (60 rpm, Belly Dancer, Stovall Life Science Inc., Greensboro, NC) under standard tissue culture conditions of 5% CO2–95% room air (nonhypoxic condition, pO2 w 140 mmHg or 20.94% O2) in a cell culture incubator (Forma Scientific, Marietta, OH). After a medium change, the plates were again placed on the orbital shaker at 37 (C under either nonhypoxic or reduced O2 (‘‘hypoxia’’, 2% O2–5% CO2–93% nitrogen, pO2 w 14 mmHg) in the cell culture incubators. After 4 h of normoxic and hypoxic exposures, the villous tissue samples were collected, excess medium was removed by blotting with sterile gauze cloth and the samples were flash frozen in liquid nitrogen. DNA-binding studies Probe preparation. A double-stranded probe was made by boiling equimolar concentrations of the two biotinylated oligonucleotides comprising the hypoxia response element from the erythropoietin gene [17] for 10 min and allowing them to cool to 42 (C slowly. The biotinylated doublestranded probe was immobilized to streptavidin-coated Dynal magnaspheres following the manufacturer’s protocol. Briefly, 2 ml of Dynabeads were washed once in binding and washing (B&W) buffer (10 mM Tris, pH 7.5, 1 mM EDTA, and 2 M NaCl) and resuspended in 1 ml of the B&W buffer. The double-stranded HRE oligonucleotide (1.3 mg/1.0 ml water) was mixed with the 1 ml of washed Dynabeads, thus bringing the final NaCl concentration to 1 M. This mixture was incubated at room temperature for 30 min. The unbound oligonucleotides were removed using a magnetic stand and decanting the buffer. These washes were saved and unbound DNA was measured at A260 to calculate the DNA bound to the Dynabeads. The final Dynabead–HRE complex was resuspended in 1! gel shift buffer (10 mM Tris, pH 7.8, 50 mM

Rajakumar et al.: Functional Activity of Hypoxia-inducible Transcription Factors

KCl, 50 mM NaCl, 1 mM EDTA, and 5% glycerol); the probe concentration was 0.5 mg/ml. Protein extraction. Those biopsy samples from preeclamptic placentae that had been previously analyzed for HIF protein expression by Western analysis and shown to overexpress the proteins were chosen for the DNA-binding study. Approximately 100 mg wet weight of each placental biopsy sample was homogenized using a sonicator for 90 s at 4 (C (Ultrasonic Processor, Tekmar, Cincinnati, OH, USA). Total proteins (i.e., both cytoplasmic and nuclear together) were extracted from these biopsy samples using 10 volumes of an extraction buffer containing 20 mM Tris, pH 7.8, 1.5 mM MgCl2, 2 M NaCl, 20% glycerol, 5 mM DTT, 0.5 mM sodium vanadate, 0.5 mM phenylmethyl sulfonyl fluoride (PMSF) and protease inhibitors (Protease inhibitor cocktail # III, 1000!, Calbiochem, San Diego, CA). The extract was then placed on a rocker for 30 min at 4 (C followed by centrifugation at 14,000 g for 15 min. The supernatant was dialyzed against 500 volumes of dialysis buffer containing 20 mM Tris, pH 7.5, 0.1 M KCl, 0.2 mM EDTA, 20% glycerol, 0.5 mM sodium vanadate, 0.5 mM PMSF and protease inhibitors overnight with one buffer change. Protein estimation was carried out using Biorad protein assay solution and the extract was stored in aliquots at
70 (C. DNA-binding reactions. Five micrograms HRE probe coupled to the Dynabeads was incubated with 2 mg of dialyzed protein extracts from the placental biopsy samples. This DNAbinding cocktail also contained 20 mg calf thymus DNA as carrier in gel shift buffer (10 mM Tris, pH 7.5, 50 mM KCl, 50 mM NaCl, 1 mM MgCl2, 1 mM EDTA, 5% glycerol, 0.5 mM sodium vanadate, 0.5 mM PMSF and protease inhibitors). After 30 min incubation, the HRE–protein–Dynabead complex was captured using a magnetic stand. After removing the buffer and two more washes and applications of the magnet, the captured complex was resuspended in 1! Laemmli’s sample buffer (50 mM Tris HCl, pH 6.8, 2% SDS, 10% glycerol) containing 5 mM DTT, 0.5 mM PMSF and ml per ml of protease inhibitors cocktail and was subjected to Western analysis for detection of HRE-bound HIF-1a protein. Western blot analysis Total proteins from villous explants were extracted in SDS containing denaturing buffer, using our published procedures [1] with slight modifications. Tissues were homogenized by sonication (Ultrasonic Processor, Tekmar, Cincinnati, OH, microprobe, setting 70 for 30 s) in 4 volumes of 1! Laemmli buffer (50 mM Tris HCl, pH 6.8, 2% SDS, 10% glycerol) containing 5 mM DTT, 0.5 mM PMSF and ml per ml of protease inhibitor cocktail. The crude homogenate was centrifuged at 13,400! g at 4 (C. The supernatant was then stored in aliquots at
70 (C. Protein estimation was carried out using the Biorad reagent. All preparations were finally

765

boiled for 5 min and briefly centrifuged. The samples were loaded in 25 ml volume containing 50 mg of total protein and 5% b-mercaptoethanol and 0.005% pyronin (Sigma Chemical Co., St. Louis, MO) and separated on a SDS containing 7.5% polyacrylamide gel (20!20 cm, Owl Separations Systems Inc., Portsmouth, NH) at 100 V for 4 h. The proteins were next transferred to PVDF membranes (Immobilon, Millipore, Bedford, MA) using a semidry transfer system (The PantherÔ Semidry Electroblotter, Owl Separations Systems Inc., Portsmouth, NH) at a constant current of 1.0 mA per cm2. Detection of proteins was carried out after blocking the membranes with a 5% solution of non-fat dry milk for 1 h and incubating with the primary antibody. Anti-HIF-1a monoclonal (Cat # H72320) and rabbit polyclonal anti-HIF-2a (Cat # NB 100-122) antibodies were obtained from Transduction Laboratories, Lexington, KY, and Novus Biologicals, Littleton, CO, respectively. The HIF-1a antibody was diluted 1:200 to a final concentration of 1.25 mg/ml in TBS buffer (Tris buffered saline) containing 0.05% Tween-20, and the HIF-2a antibody was diluted 1:1000 to 2.2 mg/ml. Rabbit anti-human Flt-1 and and mouse monoclonal anti-mouse Flk-1 antibodies (Santa Cruz Biotechnology, Santa Cruz, CA USA, cat # SC-316 and Cat # SC-6251, respectively) were diluted 1:200 to contain mg/ml, respectively. The mouse monoclonal anti-human tyrosine hydroxylase and b-actin antibodies (Sigma Chemical Company, St. Louis, MO, cat # T1299 and Cat # A 5441, respectively) were diluted 1:1000 yielding final concentrations of 4.2 and 2.2 mg/ml, respectively. All primary antibodies were incubated with the membranes for 1 h at room temperature. The membranes were subsequently washed in TBS-T buffer three times for 10 min each and then incubated with alkaline phosphatase conjugated secondary antibody (1:5000 dilution, Promega, Madison, WI) for 30 min. The membranes were washed three more times in TBS-T for 10 min each. They were further washed in buffer without Tween-20 for 10 min and allowed to equilibrate in alkaline phosphatase buffer (100 mM Tris, pH 9.5, 150 mM NaCl) for 5 min. Chemiluminescent detection was carried out using the CDP-Star substrate (Boeringher Manheim, Indianapolis, IN) diluted 1:200 in the alkaline phosphatase buffer for 5 min. Membranes were exposed for different times to Kodak Bio-max–AR film. Densitometry The developed films for Western analysis containing the desired bands were scanned using a Hewlett Packard laser scanner (Scanjet 5370C, Hewlett Packard, Palo Alta, CA) into a PICT or TIFF file in grey scale. Densitometry was carried out using an automated digitizing software UN-SCANITÔ Gel Version 4.3 (Silk Scientific Inc., Orem, UT). Statistical analysis [18] The data in Figures 1C and 3B, and Table 1 are expressed as mean G SEM. Group means were compared by unpaired t-test in Table 1. The one-sample sign test was used to

Placenta (2004), Vol. 25

766

A

Oxygen conc. %

21

2

130 Kd

Total HIF-1α

42 Kd

β-actin DNA Bound HIF-1α

130 Kd

B

SITE #

1 NP

2 PE

NP

HIF-1α Protein (PE/NP Ratio)

PE

NP PE

130 Kd

Total HIF-1α

43 Kd

β-actin DNA Bound HIF-1α

130 Kd

C

3

2.5 2

*

*

1.5 1 0.5 0 HIF-1α Total Protein

HIF-1α DNA Bound

Figure 1. HIF-1a protein overexpressed in placentas of preeclamptic women exhibit DNA binding in vitro. (A) Western blot analysis of HIF-1a (top panel) and b-actin (middle panel) in the total proteins extracted from villous explants prepared from a normal placenta and exposed to 21% and 2% oxygen for 4 h. For DNAbinding study (bottom panel), total proteins were extracted in the presence of 2 M sodium chloride from the same samples and subjected to DNA-affinity chromatography using a biotinylated hypoxia response element (HRE) of Epo gene as described in the Methods section. The bound HIF–HRE (protein-DNA) complex was also subjected to Western blot analysis for HIF-1a protein detection. (B) A representative Western blot showing HIF-1a (upper panel) and b-actin (middle panel) expression in three sites of a placenta from a normal pregnant woman and three sites from a placenta from a woman with preeclampsia. The HIF– HRE (protein-DNA) complex from these samples was also subjected for HIF-1a protein detection (bottom panel). (C) The overexpressed HIF-1a protein and the HRE-bound HIF-1a proteins expressed as a ratio of PE/NP. Both *p ! 0:0001 vs 1.0 by one sample t-test.

compare the PE/NP ratio versus 1.0 in Figures 1C and 3B. A p value of !0.05 was considered significant.

RESULTS Preeclamptic patients demonstrated significant hypertension, proteinuria, and hyperuricemia (Table 1). Both gestational age at delivery and birth weight were lower in the preeclamptic group. One of the preeclamptic subjects had HELLP (hemolysis, elevated liver function and low platelets) syndrome. Standardization of DNA affinity chromatography for assessment of bound HIF-1a–HRE (protein-DNA) complex

is portrayed in Figure 1A. Cultured villous explants prepared from a normal placenta were subjected to 21% or 2% oxygen and analyzed for both total HIF-1a protein and bound HIF1a–HRE (protein-DNA) complex. As expected, total HIF-1a protein expression was increased by hypoxia. In the same tissues, bound HIF-1a–HRE (protein-DNA) complex was also increased by hypoxia, again as expected, thereby validating this technical approach. This methodology was subsequently applied to three matched sites in each of six placental pairs of normal pregnant and preeclamptic placentae. Figure 1B illustrates a representative Western blot showing total HIF-1a, b-actin, and the bound HIF-1a–HRE (protein-DNA) complex in the three

Rajakumar et al.: Functional Activity of Hypoxia-inducible Transcription Factors

Table 1. Clinical characteristics

Variable

Normal pregnancy (n ¼ 6)

767

A Preeclampsia (n ¼ 6)a

Maternal age (years) Race

30 G 2 6 Caucasian

Gestational age at delivery (weeks) Parity Mode of delivery Birth weight (g) Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Proteinuria Uric acid (mg/dl)

39.7 G 0.2

30 G 3 5 Caucasian, 1 African American 35.8 G 1.2*

6/6 Nulliparous 6/6 Cesarian section 3366 G 327 114 G 6

4/6 Nulliparous 6/6 Cesarian section 2642 G 469* 161 G 7*

70 G 4

101 G 4*

4/6 0-traceb NAd

6/6c 7.3 G 0.6e

1

2

3

180 Kd

FLT-1

150 Kd

1

B

2

230 Kd

FLK-1

Means G SEM. aOne patient had HELLP syndrome. bUrinary protein was unavailable for two patients with normal pregnancy. cR3+ on dipstick in 4 patients, 780 mg/24 h in 1 patient, protein/creatinine ratio of 2.2 in 1 patient. dNA, not available. eSerum uric acid was available for 5/6 patients. *p ! 0:01 vs normal pregnancy by unpaired t-test.

biopsy sites from one pair of normal and preeclamptic placentae. The densitometry values for each of the three biopsy sites from the normal and preeclamptic placentae were measured, and then expressed as a ratio: normal pregnancy/ preeclampsia (NP/PE) for each site. The resulting three ratios per placental pair were then averaged. In Figure 1C, the composite average for all six placental pairs of normal pregnant and preeclamptic placentae is portrayed. Both total HIF-1a protein and the bound HIF-1a–HRE (protein-DNA) complex were significantly increased by approximately two times in the preeclamptic relative to normal placentae. The characterization of the Flt-1, Flk-1 and tyrosine hydroxylase antibodies for Western analysis are portrayed in Figure 2. Both the unglycosylated 150 and glycosylated 180 kDa forms of Flt-1 are identified in a placental villous explant from a normal pregnancy as shown in Figure 2A, lane 1. These bands were eliminated after preabsorbing the antibody with the blocking peptide antigen in lane 2. Lane 3 depicts the Western analysis using rabbit IgG in the same concentration as the Flt-1 primary antibody. Both the glycosylated intermediate 190 and glycosylated mature 230 kDa forms of Flk-1 are identified in a placental villous explant from a normal pregnancy as shown in Figure 2B, lane 1. These bands were eliminated after preabsorbing the antibody with the blocking peptide antigen in lane 2. Lane 3 depicts the Western analysis using mouse IgG1 in the same concentration as the Flk-1 primary antibody. The 60 KDa form of tyrosine hydroxylase is also identified in a placental villous explant from a normal pregnancy as shown in Figure 2C, lane 1. Positive controls for tyrosine hydroxylase expression include human brain and PC-12 cells as depicted in lanes 2 and 3, respectively.

C

1

60 K d

2

3

TH

Figure 2. Characterization of the (A) Flt-1, (B) Flk-1 and (C) tyrosine hydroxylase antibodies. (A) Lane 1: 30 mg of total proteins were prepared from a biopsy sample of normal placenta and separated by electrophoresis on a 7.5% gel for detection of Flt-1. Lane 2: for the peptide neutralization, 2 mg of antibody was mixed with 10 mg of corresponding blocking peptide antigen in 1.0 ml buffer and incubated overnight at 4 (C. Lane 3: rabbit IgG was used as an additional negative control. (B) Lane 1: 30 mg of total proteins were prepared from a biopsy sample of normal placenta and separated by electrophoresis on a 7.5% gel for detection of Flk-1. Lane 2: mouse IgG was used as a negative control. (C) Lane 1: 30 mg of total proteins were prepared from a biopsy sample of normal placenta and separated by electrophoresis on a 10% gel for detection of tyrosine hydroxylase. Lane 2: total protein extract from a human brain autopsy sample, and lane 3: total protein extract from PC-12 cells were used as positive controls for tyrosine hydroxylase expression.

A representative Western blot is shown in Figure 3A depicting the expression of several genes known to be regulated by HIF-a in three biopsy sites each from one normal and preeclamptic placenta. The densitometry values for each of the three biopsy sites from the normal and preeclamptic placentae were measured, and then expressed as a ratio: normal pregnancy/preeclampsia (NP/PE) for each

Placenta (2004), Vol. 25

768 SITE #

A

1

2

DISCUSSION

3

NP PE NP PE NP PE HIF-1α

130 Kd

HIF-2α

115 Kd

NS FLT-1

180 Kd

230 Kd

B

FLK-1

60 Kd

TH

43 Kd

β-actin

3

* *

PE/NP Ratio

2.5 2

* *

1.5 1

* 0.5 0 HIF-1α HIF-2α FLT-1

FLK-1

TH

β-actin

Figure 3. The expression of HIF-1a and downstream oxygen-regulated genes in placentas of women with normal pregnancy and preeclampsia. (A) Representative Western blots showing HIF-1a and -2a protein expression in three placental biopsy sites each from a normal pregnant and preeclamptic woman. The protein expression of the VEGF receptors, Flt-1 and Flk-1, tyrosine hydroxylase, and b-actin in the same biopsy sites are also shown. (B) Densitometry was performed for the various protein bands and their levels recorded and expressed as a ratio of preeclampsia/normal pregnancy (PE/NP) for the three matched sites from each placental pair. An average of the three matched sites per placental pair was then calculated. The grand mean G SEM for the six normal and preeclamptic placental pairs are presented in (B). A ratio of 1.0 indicates comparable protein levels between the two groups. *p ! 0:02 by one sample t-test.

site. The resulting three ratios per placental pair were then averaged. In Figure 3, the composite average for all six placental pairs of normal pregnant and preeclamptic placentae is portrayed. In addition to overexpression of both HIF-1a and -2a, the preeclamptic placentae expressed significantly more Flt-1 and tyrosine hydroxylase proteins and significantly less Flk-1 protein relative to the placentae from normal pregnancies. There was no difference in the expression of b-actin.

This work provides evidence for the functional activity of hypoxia-inducible transcription factors in placentae from women with preeclampsia. The major findings are (1) HIF1a protein overexpressed in preeclamptic placentae is capable of binding to the DNA hypoxia response element in vitro, and (2) target genes known to be regulated by the HIF pathway are altered in preeclamptic placentae in vivo. We previously demonstrated overexpression of HIF-1a and -2a proteins, but not mRNA in the placentae of women with preeclampsia relative to normal term placentae [7,8]. This conclusion was based on a systematic, and unbiased placental sampling approach, in which eight biopsy sites were investigated for each placenta. Furthermore, the expression of HIF-1a and -2a proteins was not significantly different between normal term and preterm placentas without preeclampsia [7,8]. Interestingly, overexpression of both HIF-1a and -2a persisted in cultured villous explants from preeclamptic placentae even after 24 h in culture under standard tissue culture conditions—20.94% oxygen or PO2 w 130 torr [9]. Because HIF-1a and -2a are usually rapidly degraded in tissues under oxygenated conditions via the ubiquitin– proteasomal pathway [19], these results suggest an abnormality in this degradative pathway for HIF-1a and -2a (and possibly for other proteins) in preeclamptic placentae, a hypothesis that we are currently testing. The overexpression of HIF-1a and -2a in preeclamptic placentae was primarily confined to the cytoplasm and nucleus of the syncytiotrophoblast and fetoplacental blood vessels [7,8]. The nuclear localization suggests but does not prove functional activity. To begin investigating functional activity, we first tested whether the HIF-1a overexpressed in preeclamptic placentae was capable of binding to the DNA hypoxia response element—a prerequisite for subsequent transactivational activity. The results depicted in Figure 1 provide evidence for the DNA-binding potential of HIF-1a in preeclamptic placentae, insofar as both the total and DNA-bound HIF-1a protein, the latter determined by DNA affinity chromatography in vitro, were increased to the same extent by approximately two times relative to normal term placentae. Having established DNA-binding potential at least in vitro, we next investigated three proteins which are known to be regulated by the HIF pathway as molecular signatures of HIF activity in vivo. The VEGF receptors Flt-1 and Flk-1 play key roles in angiogenesis [10], and tyrosine hydroxylase is the ratelimiting enzyme in the formation of catecholamines [11]. Flt-1 and tyrosine hydoxylase have been generally shown to be increased by hypoxia [13,14], whereas the direction of change for Flk-1 varies perhaps as a function of the tissue source [12]. These proteins were analyzed by Western analysis in the same biopsy sites used to investigate total and DNA-bound HIF-1a and/or -2a and we carefully validated the primary antibodies used for their detection (Figure 2). All three proteins were significantly altered in the preeclamptic relative to normal term

Rajakumar et al.: Functional Activity of Hypoxia-inducible Transcription Factors

placentae. Both Flt-1 and tyrosine hydroxylase were increased, whereas Flk-1 was reduced (Figure 3). These results are consistent with increased transcriptional activation by the HIF pathway in vivo. We focused on Flt-1, Flk-1 and tyrosine hydroxylase in preeclamptic placentae primarily as molecular indicators of HIF activity. However, the overexpression of Flt-1 protein in preeclamptic placentae is interesting in its own right. This finding complements recent evidence for increased expression of both Flt-1 and soluble Flt-1 mRNA in the preeclamptic placenta [20]. Possibly, the increase in Flt-1 and sFlt-1, the latter arising from alternative splicing, are instigated by HIF1a overexpression in the preeclamptic placenta. Of additional interest is the expression of tyrosine hydroxylase protein by the normal human placentae, and the

769

overexpression of this enzyme in placentae from preeclamptic women. These intriguing data both confirm and extend work of Manyonda et al. [21]. These investigators reported increased maternal venous concentrations of norepinephrine in preeclampsia, as well as increased tyrosine hydroxylase activity and mRNA expression in preeclamptic placentae, thereby suggesting a placental contribution to the elevated circulating levels of norepinephrine in the disease. Another source may be the increased sympathetic activity reported during preeclampsia [22]. Although the majority of studies show an increase in circulating catecholamines in preeclampsia, there are several exceptions. Nevertheless, the mixed a1/b adrenergic antagonist, labetolol, is effective in reducing the blood pressure during preeclampsia [23], which is consistent with elevation of circulating catecholamines and/or sympathetic activity.

ACKNOWLEDGEMENTS We are grateful to the Prenatal Exposures Preeclampsia Prevention (PEPP) staff for their assistance in placental collection. This work was supported by PPG PO1 HD30367 and NIH RO1 HL56410. Portions were presented in abstract form [J Soc Gynecol Invest 2003, 10, 304A].

REFERENCES [1] Rajakumar A, Conrad KP. Expression, ontogeny, and regulation of hypoxia-inducible transcription factors in the human placenta. Biol Reprod 2000;63:559–69. [2] Semenza GL. HIF-1 and mechanisms of hypoxia sensing. Curr Opin Cell Biol 2001;13:167–71. [3] Wenger RH, Gassmann M. Oxygen(es) and the hypoxia-inducible factor-1. Biol Chem 1997;378:609–16. [4] Jauniaux E, Hempstock J, Greenwold N, Burton GJ. Trophoblastic oxidative stress in relation to temporal and regional differences in maternal placental blood flow in normal and abnormal early pregnancies. Am J Pathol 2003;162:115–25. [5] Jauniaux E, Watson AL, Hempstock J, Bao YP, Skepper JN, Burton GJ. Onset of maternal arterial blood flow and placental oxidative stress. A possible factor in human early pregnancy failure. Am J Pathol 2000;157:2111–22. [6] Rodesch F, Simon P, Donner C, Jauniaux E. Oxygen measurements in endometrial and trophoblastic tissues during early pregnancy. Obstet Gynecol 1992;80:283–5. [7] Rajakumar A, Whitelock KA, Weissfeld LA, Daftary AR, Markovic N, Conrad KP. Selective overexpression of the hypoxia-inducible transcription factor, HIF-2alpha, in placentas from women with preeclampsia. Biol Reprod 2001;64:499–506. [8] Rajakumar A, Whitelock KA, Weissfeld LA, Daftary AR, Markovic N, Conrad KP. Overexpression of the hypoxia-inducible transcription factors in placentas from women with preeclampsia. Biol Reprod 2001;64:1019–20 [Erratum]. [9] Rajakumar A, Doty K, Daftary A, Harger G, Conrad KP. Impaired oxygen-dependent reduction of HIF-1alpha and -2alpha proteins in preeclamptic placentae. Placenta 2003;24:199–208. [10] Ferrara N, Gerber HP, LeCouter J. The biology of VEGF and its receptors. Nat Med 2003;9:669–76. [11] Dluhy R, Lawrence JE, Williams GH. Endocrine hypertension. In: Reed P, Larsen HMK, Shlomo Melmed, Polonsky Kenneth S, editors. Williams text book of endocrinology. Philadelphia, Pennsylvania, USA: Saunders, An Imprint of Elsevier Science; 2003. p. 552–85. [12] Elvert G, Kappel A, Heidenreich R, Englmeier U, Lanz S, Acker T, et al. Cooperative interaction of hypoxia-inducible factor-2alpha (HIF-2alpha)

[13]

[14]

[15] [16]

[17]

[18] [19]

[20]

[21]

[22]

[23]

and Ets-1 in the transcriptional activation of vascular endothelial growth factor receptor-2 (Flk-1). J Biol Chem 2003;278:7520–30. Gerber HP, Condorelli F, Park J, Ferrara N. Differential transcriptional regulation of the two vascular endothelial growth factor receptor genes. Flt-1, but not Flk-1/KDR, is up-regulated by hypoxia. J Biol Chem 1997;272:23659–67. Schnell PO, Ignacak ML, Bauer AL, Striet JB, Paulding WR, CzyzykKrzeska MF. Regulation of tyrosine hydroxylase promoter activity by the von Hippel–Lindau tumor suppressor protein and hypoxia-inducible transcription factors. J Neurochem 2003;85:483–91. NIH. Working Group Report on High Blood Pressure in Pregnancy. 2000; 1–24. Benyo DF, Miles TM, Conrad KP. Hypoxia stimulates cytokine production by villous explants from the human placenta. J Clin Endocrinol Metab 1997;82:1582–8. Wang GL, Semenza GL. Characterization of hypoxia-inducible factor 1 and regulation of DNA binding activity by hypoxia. J Biol Chem 1993;268:21513–8. Zar J. Biostatistical analysis. Englewood Cliffs, NJ: Prentice Hall; 1984. Lando D, Peet DJ, Whelan DA, Gorman JJ, Whitelaw ML. Asparagine hydroxylation of the HIF transactivation domain a hypoxic switch. Science 2002;295:858–61. Maynard SE, Min JY, Merchan J, Lim KH, Li J, Mondal S, et al. Excess placental soluble fms-like tyrosine kinase 1 (sFlt1) may contribute to endothelial dysfunction, hypertension, and proteinuria in preeclampsia. J Clin Invest 2003;111:649–58. Manyonda IT, Slater DM, Fenske C, Hole D, Choy MY, Wilson C. A role for noradrenaline in pre-eclampsia: towards a unifying hypothesis for the pathophysiology. Br J Obstet Gynaecol 1998;105: 641–8. Schobel HP, Langenfeld M, Gatzka C, Schmieder RE. Treatment and post-treatment effects of alpha- versus beta-receptor blockers on left ventricular structure and function in essential hypertension. Am Heart J 1996;132:1004–9. Umans J, Lindheimer MD. Antihypertensive treatment. In: Lindheimer MD, Roberts JM, Cunningham FG, editors. Chesley’s hypertensive disorders in pregnancy. Stanford, CT: Appleton & Lange; 1999. p. 581–604.