Human Prostaglandin Transporter Gene (hPGT) is Regulated by Fluid Mechanical Stimuli in Cultured Endothelial Cells and Expressed in Vascular Endothelium in Vivo
Background—biomechanical forces generated by blood flow within the cardiovascular system have been proposed as important modulators of regional endothelial phenotype and function. This process is thought to involve the regulation of vascular gene expression by physiological fluid mechanical stimuli such as fluid shear stresses.
Methods and Results—We demonstrate sustained upregulation of a recently identified gene encoding a human prostaglandin transporter (hPGT) in cultured human vascular endothelium exposed to a physiological fluid mechanical stimulus in vitro. This biomechanical induction is selective in that steady laminar shear stress is sufficient to upregulate the hPGT gene at the level of transcriptional activation, whereas a comparable level of turbulent shear stress (a nonphysiological stimulus) is not. Various biochemical stimuli, such as bacterial endotoxin and the inflammatory cytokines recombinant human interleukin 1β cytokines (rhIL-1β) and tumor necrosis factor-α (TNF-α), did not significantly induce hPGT. Using a specific antiserum to hPGT, we demonstrate endothelial expression within the arterial vasculature and the microcirculation of highly vascularized tissues such as the heart.
Conclusions—Our results identify hPGT as an inducible gene in vascular endothelium and suggest that biomechanical stimuli generated by blood flow in vivo may be important determinants of hPGT expression. Furthermore, this demonstration of regulated endothelial expression of hPGT implicates this molecule in the regional metabolism of prostanoids within the cardiovascular system.
Vascular endothelium is both the source and the target of action of a variety of biologically important mediators. Uniquely positioned between blood and tissues, it plays a multifunctional role as an integrator and transducer of both blood-borne and locally derived signals.1 In addition to the many well-characterized humoral stimuli capable of regulating endothelial function (eg, growth factors, cytokines), it is now clear that biomechanical stimuli, such as the fluid shear stresses and cyclic strains produced by the pulsatile flow of blood within the cardiovascular system, are important determinants of endothelial phenotype in vivo. In vitro, fluid shear stresses within the physiological range have been demonstrated to induce a variety of responses in cultured endothelium, ranging from the acute stimulation of nitric oxide and prostacyclin (PGI2) release to direct effects on endothelial gene expression resulting in more sustained alterations in endothelial phenotype.2 3 4
Eicosanoids are important mediators of both physiological and pathophysiological processes within the cardiovascular system.5 6 Recently, a cDNA encoding a molecule capable of mediating the transport of prostanoids in rat and humans has been identified.7 8 This molecule, human prostaglandin transporter (hPGT), is a member of the 12-membrane-spanning superfamily of transporters and has been demonstrated to mediate the transport of prostaglandins across cell membranes in vitro. In particular, hPGT expressed in HeLa cells is capable of mediating the uptake of prostaglandins E2, E1, F2α, D2, and thromboxane B2, but not the PGI2 analogue iloprost. Although the cell types capable of expressing hPGT in vivo have not been identified, based on this in vitro transport data and the relatively broad tissue distribution of the hPGT mRNA by Northern analysis,7 8 Kanai, Lu, and colleagues have postulated a role for hPGT in the transport or clearance of prostaglandins in diverse tissues.
Here, we report the isolation from an endothelial cell library of a cDNA encoding hPGT and demonstrate that this gene is selectively induced in a sustained fashion in response to physiological levels of a uniform, laminar fluid shear stress in vitro. Furthermore, hPGT protein is expressed in the endothelium of tissues such as the heart, lung, and kidney, as well as in the endothelial lining of the arterial vasculature in humans. These results thus identify hPGT as a regulated gene in vascular endothelium, and support a role for hPGT in regional endothelial prostanoid metabolism in vivo.
Endothelial Cell Culture
Human umbilical vein endothelial cells (HUVEC) were isolated from multiple normal-term umbilical cords, pooled, and cultured in medium 199 supplemented with endothelial cell growth supplement (50 μg/mL, Collaborative Research Inc), heparin (50 μg/mL, porcine intestinal, Sigma), antibiotics (penicillin-G 100 U/mL, streptomycin 100 μg/mL, Sigma) and 20% fetal bovine serum. Cells at passage level 2 or 3 were replicate-plated on 0.1% gelatin-coated standard Petri dishes or specially designed plates fabricated from the same tissue-culture plastic (Costar), and allowed to grow to confluent densities before experimental use.
Shear Stress Apparatus
Confluent HUVEC monolayers grown on 17.8-cm-diameter maxiplates (≈107 cells per plate) were introduced into a cone-plate flow apparatus9 consisting of a stainless steel cone rotating over a stationary baseplate. The culture medium present between the cone and the plate was constantly replenished at a rate of 0.5 mL/min during experiments, and the entire apparatus was maintained in a humidified 5% CO2-95% air atmosphere. The equations and calculations for describing the shear stresses generated in this cone-plate apparatus have been reported in detail previously.9 10 For laminar shear stress (LSS) at 10 dyne/cm2, we used a 0.5° cone at a rotational velocity of 100 rpm; for 4 dyne/cm2 or 12 dyne/cm2, the rotational velocity of the cone was 40 or 120 rpm, respectively. As described by Sdougos et al,10 the parameter R̃ is a function of the local radius from the center of the cone, the angular velocity of the cone, the angle of the cone itself, and the fluid kinematic viscosity of the media; at values <1, R̃ predicts laminar flow conditions. Turbulent shear stresses (TSS) can be generated by manipulating these variables to achieve R̃ values of >4. At a 3.0° cone angle, a rotational velocity of 135 rpm, and a radius of ≥3.5 cm, the R̃ values are >5 and turbulence is predicted and observed experimentally. For the nuclear runoff analysis, 2 maxiplates (≈107 cells per plate) were simultaneously prepared as above. One served as a static control, and one was exposed to 6 hours of steady LSS at 10 dyne/cm2 as above. The transcriptional rates of the hPGT and GAPDH genes were assayed as described previously,11 and the PUC-19 plasmid that contains no endothelial genes served as a negative control.
Northern and Reverse-Transcription–Polymerase Chain Reaction Analysis
RNA was isolated by guanidinium/phenol extraction, resolved on 1.2% agarose/formaldehyde gels and transferred to nylon membranes by capillary blot for Northern analysis. Differential display and semi- quantitative reverse-transcription–polymerase chain reaction (RT-PCR) was performed as previously described.11 For the Northern blots, the probe used consisted of the hPGT cDNA (isolated from a human endothelial library) labeled with random hexamers. The primers used in the RT-PCR analysis of hPGT were 5′ AGCCTTGGGGATGCTGTTTG3′ and 5′CATGGCAAGGGGAGAGGTAC3′. Primers for human GAPDH were purchased from Stratagene.
Antibody Preparation, Western Blot, and Immunostaining
A rabbit polyclonal anti-hPGT antiserum was raised against a synthetic peptide consisting of amino acids 627 to 639 of hPGT (RVKKNKEYNVQKAA) and affinity-purified on the immobilized peptide (Research Genetics). For the Western blot analysis, the HUVEC were exposed to the appropriate stimuli and then lysed in 20 mmol/L Tris-HCl pH 7.0, 1% SDS, 0.5% β-mercaptoethanol, 30% glycerol, and 1 mmol/L PMSF and immediately boiled for 5 minutes. Proteins were resolved on 10% acrylamide gels under denaturing conditions, blotted to nitrocellulose, and probed with the hPGT antisera at a concentration of 1/1000. Human tissue samples were collected at the time of surgery or autopsy, according to established institutional protocols. Frozen sections (4 to 5 μm) were used for immunostaining, with anti-hPGT serum applied at a concentration of 1/800 to 1/1000. As a control for the specificity of this antiserum, excess purified hPGT peptide was preincubated with the antiserum overnight before its use for tissue staining. In addition, preimmune serum was routinely included as a negative control (data not shown). The monoclonal antibody to human PECAM-1 (CD-31) was obtained from DAKO Corp. Immunostaining was detected with a secondary antibody conjugated to HRP by use of AEC as substrate, and the slides were counterstained with hematoxylin. For the immunofluorescence, individual sections were stained simultaneously with the human anti-CD31 antibody, and the rabbit anti-hPGT antiserum and the sections were developed with a goat anti-mouse secondary antibody conjugated to Texas Red (NEN Life Science Products) and a goat anti-rabbit antibody conjugated to FITC (Jackson Immuno Research Inc). The sections were visualized with a Nikon Microphot-FXA fluorescence microscope and imaged with a Power Macintosh–based image analysis system (Oncor).
Identification of hPGT as a Regulated Gene in Cultured Vascular Endothelium
A differential display analysis in cultured HUVEC exposed to both defined fluid mechanical and cytokine stimuli identified a transcript of ≈4.5 kb that reproducibly demonstrated a selective upregulation in response to sustained steady LSS in vitro (data not shown). Sequence analysis of a corresponding cDNA derived from a HUVEC library initially revealed a novel gene that was homologous to a gene encoding the rat matrin F/G.12 13 Subsequently, Kanai and colleagues7 reported a longer reading frame within the rat matrin F/G cDNA and demonstrated that this longer form of the protein encoded a member of the family of 12-membrane-spanning transporters that was capable of specifically transporting prostaglandins. This group then identified the human homolog of this molecule and named it hPGT (for human prostaglandin transporter).8 The sequence of our HUVEC-derived cDNA contains all of the predicted reading frame of hPGT and is identical to that reported by Lu et al.8
As demonstrated in Figure 1⇓ by semiquantitative RT-PCR, the transcript encoding hPGT exhibits a selective upregulation in response to steady LSS in vitro. By 6 hours, the level of hPGT message is significantly increased over static controls, whereas a comparable time-averaged magnitude of TSS, or 20 U/mL of the cytokine rhIL-1β, does not significantly alter the level of this mRNA (Figure 1A⇓). Other biochemical stimuli such as lipopolysaccharide (LPS), interferon-γ, and TNF-α have been examined, and none of these altered the level of steady-state hPGT mRNA significantly (data not shown). After 24 hours of shear or cytokine stimuli, the selective upregulation of the hPGT gene by the steady laminar shear stimulus is even more evident (Figure 1B⇓). Figure 1C⇓ is a Northern analysis of total RNA isolated from HUVEC cultured under static conditions, or exposed to 24 hours of LSS, and probed with the hPGT cDNA. A major transcript of ≈4.5 kb is significantly upregulated, while less abundant species of ≈7.5 and 2.5 kb are also seen. This pattern is consistent with the Northern results reported by Lu et al.8 Conditioned media from shear stress-treated HUVEC was unable to upregulate hPGT expression in static HUVEC cultures (data not shown).
hPGT Gene Is Sensitive to Applied Shear Stress and Upregulation Results in Increased Protein Levels
To explore the laminar shear regulation of the hPGT gene in more detail, a series of shear preconditioning experiments were performed. As demonstrated in Figure 2⇓, when HUVEC are cultured under static conditions they express a low but detectable level of hPGT mRNA, which is significantly upregulated by 24 hours of steady LSS at 12 dyne/cm2. However, if the cells are preconditioned for 18 hours of steady LSS at 12 dyne/cm2 and then the applied shear stress is reduced to 4 dyne/cm2, the level of hPGT mRNA decays back to static levels (compare lanes 2 and 3). Lanes 4 and 5 demonstrate that steady LSS at 4 dyne/cm2 is not a sufficient stimulus to significantly induce hPGT message levels above static levels, but if the cells are preconditioned under this lower level of laminar shear for 18 hours and then LSS is stepped up to 12 dyne/cm2, the hPGT message is promptly induced. These results, together with those of Figure 1⇑, suggest that the hPGT gene is responding to both the nature of the applied fluid mechanical stimulus (eg, laminar versus turbulent shear stress) as well as the magnitude of the effective stimulus. In particular, it appears that a threshold of LSS between 4 and 10 dyne/cm2 is both necessary and sufficient for the induction and maintenance of hPGT expression in these cultured endothelial cells. The hPGT gene demonstrates a similar LSS-induced upregulation in cultured endothelial cells from bovine aortae, whereas no significant upregulation was observed in Cos-7 cells, a nonendothelial cell type exposed to the same biomechanical stimulus (data not shown).
Figure 3⇓ is a nuclear runoff analysis, demonstrating that the upregulation of hPGT message in HUVEC in response to the LSS stimulus is occurring at least in part at the level of transcriptional activation. These results suggest that the promoter of the hPGT gene is directly responsive to the LSS stimulus in vitro.
To confirm that the observed upregulation of hPGT message is in fact resulting in an increase in hPGT protein levels, a polyclonal antiserum was raised against a synthetic peptide derived from the predicted amino acid sequence of hPGT. As demonstrated in Figure 4⇓, this antisera recognizes a protein of ≈80 to 85 kd in HUVEC total-cell lysates. This species is markedly induced by the LSS stimulus, whereas the TSS and cytokine stimuli result in little if any significant upregulation, consistent with the data for hPGT message presented above (Figure 4A⇓). The selective upregulation of hPGT protein appears sustained, such that after 48 hours of steady LSS, the level of hPGT protein remains significantly elevated above static levels (Figure 4B⇓). The predicted molecular weight of the hPGT protein based on the cDNA sequence is ≈79 kDa; thus, these results are consistent with the expression of hPGT protein from the open reading frame present within the hPGT cDNA identified8 and suggest that the hPGT protein may undergo some posttranslational modification.
hPGT Is Expressed in Vascular Endothelium in Vivo
To explore the tissue distribution of hPGT expression in vivo, we used the cDNA as a probe in a Northern analysis of RNA derived from various normal human tissues. A predominant transcript of ≈4.5 kb was easily detected, and less abundant species of ≈2.5 and 7.5 kb also were evident (data not shown). Although detectable amounts of hPGT mRNA were found in several tissues, including heart, placenta, and kidney, relatively high levels of hPGT transcripts were present in the pancreas, prostate, small intestine, and lung. These findings are in agreement with previously published results8 and suggest that the pattern (ie, size and complexity) of transcripts induced in cultured human endothelial cells in the current study is similar to that observed in intact normal human tissues in vivo.
To directly address the question of whether hPGT was expressed in vascular endothelium in vivo, a polyclonal anti-peptide antisera was used to examine the expression of hPGT immunoreactive protein in human vessels and myocardial tissues obtained at autopsy. As assessed by immunohistochemistry, hPGT is expressed within the microvascular endothelium present in normal myocardial tissue (Figure 5A⇓) and is present within the endothelial lining of all of the undiseased human arterial specimens we have examined to date, such as the epicardial coronary artery (Figure 5B⇓) or internal mammary artery (Figure 5C⇓). The specificity of the immunostaining was confirmed by the ability of excess peptide antigen to specifically inhibit tissue staining (Figure 5D⇓). To confirm that the hPGT protein detected in the myocardial and arterial tissues used was in fact expressed within the endothelium itself, we performed double-label immunofluorescence staining with the anti-hPGT serum and a monoclonal antibody directed against CD31 (PECAM-1), a specific marker of vascular endothelium in these tissues. As shown in Figure 6⇓, the immunofluorescent pattern produced by these antibodies was virtually identical. Colocalization of CD31 and hPGT was observed in vessels of all sizes, such as intramyocardial arterioles (Figure 6A⇓ through 6C⇓), larger epicardial arteries (Figure 6D⇓ through 6F⇓), or microvessels within the myocardium (Figure 6G⇓ through 6I⇓). Immunoreactive hPGT in a vascular pattern suggesting endothelial expression was also observed in other human tissues examined, such as the lung and kidney (Figure 7⇓).
The observation that cultured HUVEC that express very little hPGT under static conditions can be induced to express this protein by physiological levels of shear stress in vitro, together with the observation that hPGT is expressed in vascular endothelium in vivo, suggest that signals derived from blood flow may play a role as a regulator of hPGT expression in endothelium. A prediction of this hypothesis is that umbilical vessels (the source of the HUVEC used for the in vitro studies) would express this molecule. As shown in Figure 8⇓ by 2-color immunofluorescence, hPGT expression is readily detected in the luminal endothelium of both human umbilical veins and arteries.
These results demonstrate for the first time that the gene encoding hPGT is regulated by fluid mechanical stimuli in cultured endothelial cells in vitro and is expressed in a selective fashion in the vascular endothelial lining of various sized vessels in vivo. The hPGT gene is induced in cultured HUVEC by a threshold magnitude of steady LSS between 4 and 10 dyne/cm2, but not by a comparable magnitude of TSS (a nonphysiological fluid mechanical stimulus) or by any of the humoral (eg, endotoxin, cytokine) stimuli examined. This pattern of response of the hPGT gene in vitro, together with our demonstration that hPGT protein is readily detectable in the endothelial lining of vessels in a variety of human tissues, suggests that the regulation of hPGT expression by blood flow may be an important mechanism governing its expression in vivo. The transcriptional activation of the hPGT gene resulting in increased hPGT protein indicates that this is an inducible gene potentially capable of responding to physiological stimuli in vivo. In addition, our protein expression data confirm and extend the report of Kanai et al,7 and are consistent with this previous in vitro data that hPGT protein is expressed from the longer reading frame present within rat matrin F/G cDNA.
The functional roles of hPGT expression within the vasculature are unknown. It has been postulated that hPGT may play a variety of roles in prostaglandin metabolism in vivo, including mediating the efflux of newly synthesized prostanoids from intracellular pools, the transepithelial transport of prostanoids, or the clearance of circulating prostaglandins from blood or other extracellular fluids.7 8 These suggestions are based on 2 observations: (1) that hPGT message is detectable at some level in most tissues but seems to be enriched in certain tissues, such as the pancreas and small intestine, that are rich in epithelium; and (2) that hPGT transiently overexpressed in HeLa cells in vitro mediates the uptake of a variety of prostanoids but does not appear capable of mediating the efflux of prostaglandins when expressed in Xenopus oocytes.
Our results demonstrating a predominantly endothelial-selective pattern of expression in tissues such as the heart and vasculature point to a previously unsuspected role for hPGT in the trafficking of prostanoids within the cardiovascular system. It has long been recognized that certain prostaglandins, such as PGE2 and PGF2α, are rapidly cleared by a single passage through the pulmonary or other vascular beds in vivo.14 15 16 17 This clearance occurs in the absence of any detectable enzymatic activity in blood capable of metabolizing these molecules, indicating that clearance by vascular cells may be playing a primary role. The expression of hPGT in the endothelium of vessels in the heart (as well as the lung and kidney), as demonstrated here, presents a possible mechanism for this phenomenon. A role for endothelial-expressed hPGT in the clearance of blood-borne prostanoids would be consistent with the observation of saturable, energy-dependent transmembrane transport of prostaglandins18 and the more recent observations that transport inhibitors such as probenecid or furosemide, which are capable of binding to and inhibiting diverse members of the family of 12-membrane transporters, including hPGT expressed in vitro, are capable of altering prostanoid uptake in vivo.7 8 19 20 A recent report examining the direct vascular effects of furosemide in humans demonstrated a prostaglandin-dependent venodilation that the authors hypothesized was caused by a stimulation of local prostaglandin synthesis.21 Our data suggest an alternative hypothesis, namely that the systemic administration of furosemide results in decreased vascular clearance of vasodilatory prostanoids mediated by endothelial-expressed hPGT, resulting in venodilation. It will be interesting to test these hypotheses directly via targeted over- and under-expression strategies of hPGT in vivo.
Interestingly, hPGT is the second example of a gene encoding a member of the 12-membrane-spanning transporters that demonstrates selective responsiveness to fluid mechanical stimuli in cultured vascular endothelial cells. We have recently reported that the gene encoding one of the isoforms of the bumetanide sensitive Na-K-2Cl cotransporter in humans (BSC2) exhibits a selective upregulation in response to sustained LSS, but not TSS, in HUVEC.22 Although the BSC2 gene also is regulated by inflammatory cytokines, its response to defined fluid mechanical stimuli in vitro is strikingly similar to that of the hPGT gene reported here. In both cases, the genes encoding these cell surface transporters are transcriptionally regulated by the LSS stimulus and demonstrate a selective response to a threshold magnitude of steady LSS between 4 and 10 dyne/cm2. These results suggest that these genes are not responding to the acute onset of applied shear stress and that a sustained magnitude of shear stress within the physiological range is both necessary and sufficient for expression of these genes in vitro. hPGT and BSC2 may represent the first examples of a larger class of molecules (ie, the superfamily of cell-surface transporters) that, in addition to other well-characterized classes of endothelial genes, such as the leukocyte adhesion molecules,2 are important functional components of endothelial phenotype that are modulated by biomechanical stimuli in vivo. In addition, the ability of endothelium to respond distinctly to differing fluid mechanical stimuli (eg, LSS versus TSS), suggests that fluid mechanical forces, which by their very nature demonstrate regional variations in amplitude and pattern with the cardiovascular system, may play a critical role in determining regional endothelial structure and function.
In summary, we have demonstrated that the gene encoding hPGT is expressed in human vascular endothelium in vivo, and manifests a selective upregulation in response to a physiologically relevant fluid mechanical stimulus in vitro. This is the first report of cell-type specific expression of hPGT in vivo and identifies the hPGT gene as a regulated species. Although the function of hPGT in endothelium is unknown, these results suggest that hPGT may be involved in the trafficking of prostanoids within the cardiovascular system and raise the testable hypothesis that alterations in the expression or functional status of hPGT may be involved in certain cardiovascular disease states.
This work was supported in part by grants from the National Heart, Lung, and Blood Institute (R37-HL51150 and P50-HL56985, to Dr Gimbrone) and by a sponsored research agreement between the Brigham and Women’s Hospital and Millennium Pharmaceuticals, Inc, in collaboration with Eli Lilly. Dr Topper is the recipient of a Howard Hughes Medical Institute Fellowship for Physicians. The authors gratefully acknowledge the assistance of Jeanne M. Kiely, Maria Dichiara, Tobi Nagel, William Atkinson, and Kay Case.
- Received June 22, 1998.
- Accepted July 27, 1998.
- Copyright © 1998 by American Heart Association
Davies PF. Flow-mediated endothelial mechanotransduction. Physiol Rev. 1995;75:519–560.
Resnick N, Gimbrone MA Jr. Hemodynamic forces are complex regulators of endothelial gene expression. FASEB J. 1995;9:874–882.
Kanai N, Lu R, Satriano JA, Bao Y, Wolkoff AW, Schuster VL. Identification and characterization of a prostaglandin transporter. Science. 1995;268:866–869.
Topper JN, Cai J, Falb D, Gimbrone MA Jr. Identification of vascular endothelial genes differentially responsive to fluid mechanical stimuli: cyclooxygenase-2, manganese superoxide dismutase, and endothelial cell nitric oxide synthase are selectively up-regulated by steady laminar shear stress. Proc Natl Acad Sci U S A. 1996;93:10417–10422.
Hakes DJ, Berezney R. Molecular cloning of matrin F/G: a DNA binding protein of the nuclear matrix that contains putative zinc finger motifs. Proc Natl Acad Sci U S A. 1991;88:6186–6190.
Nakayasu H, Berezney R. Nuclear matrins: identification of the major nuclear matrix proteins. Proc Natl Acad Sci U S A. 1991;88:10312–10316.
Hagen AA, Gerber JN, Sweeley CC, White RP, Robertson JT. Levels and disappearance of prostaglandin F2α in cerebral spinal fluid: a clinical and experimental study. Stroke. 1977;8:672–675.
Bito LZ, Salvador EV. Effects of anti-inflammatory agents and some other drugs on prostaglandin biotransport. J Pharmacol Exp Ther. 1976;198:481–488.
Pickkers P, Dormans TP, Russel FG, Hughes AD, Thien T, Schaper N, Smits P. Direct vascular effects of furosemide in humans. Circulation. 1997;96:1847–1852.
Topper JN, Wasserman SM, Anderson KR, Cai J, Falb D, Gimbrone MA Jr. Expression of the bumetanide-sensitive Na-K-Cl cotransporter BSC2 is differentially regulated by fluid mechanical and inflammatory cytokine stimuli in vascular endothelium. J Clin Invest. 1997;99:2941–2949.The gene encoding a human prostaglandin transporter (hPGT) is induced in cultured human vascular endothelium exposed to a physiological fluid mechanical stimulus in vitro. This biomechanical induction is selective in that other stimuli failed to induce hPGT. hPGT protein is expressed in human vascular endothelium in vivo. These results identify hPGT as an inducible gene in vascular endothelium and suggest that biomechanical stimuli generated by blood flow in vivo may be important determinants of hPGT expression. This demonstration of regulated endothelial expression of hPGT implicates this molecule in the regional metabolism of prostanoids within the cardiovascular system.