Cellular phenotype is determined not only by genetic transcription but also by subsequent translation of mRNA into protein. Extracellular signals trigger intracellular pathways that distinctly activate translation. The 70/85-kDa S6 kinase (pp70S6k) is a central enzyme in the signal-dependent control of translation, but its regulation in endothelial cells is largely unknown. Here we show that fluid flow (in the absence of an exogenous mitogen) as well as humoral agonists activate endothelial pp70S6k. Rapamycin, an inhibitor of the mammalian target of rapamycin (mTOR), and wortmannin, a phosphatidylinositol 3-kinase inhibitor, blocked flow-induced pp70S6k activation; FK-506, a rapamycin analog with minimal mTOR inhibitory activity, and PD-98059, an inhibitor of the flow-sensitive mitogen-activated protein kinase pathway, had no effect. Synthesis of Bcl-3, a protein whose translation is controlled by an mTOR-dependent pathway, was induced by flow and inhibited by rapamycin and wortmannin. Transcriptional blockade did not abolish the flow-induced upregulation of Bcl-3. Fluid forces may therefore modify endothelial phenotype by specifically regulating translation of certain mRNA transcripts into protein.

  • signal transduction
  • hemorheology
  • rapamycin
  • phosphatidylinositol 3-kinase
  • phenotype

fluid shear stress is a biomechanical force particularly relevant to endothelial cells, because they are positioned at the interface between flowing blood and the vessel wall. Shear stress activates transcription of a wide array of genes that may alter endothelial phenotype (reviewed in Ref. 15). The transcriptional events triggered by shear have been studied extensively and are generally regulated by kinase pathways that signal to the nucleus via transcription factors (4, 36). Ultimately, cellular phenotype depends on not only the program of genes transcribed but also whether these mRNA species are translated into proteins. The signaling pathways that regulate protein translation have not been characterized in endothelial cells. It is possible that key translational control switches and checkpoints are also regulated by fluid shear stress, providing yet another mechanism by which hemodynamic factors might influence endothelial phenotype.

One important regulator of translation and cell cycle progression is the 70/85-kDa S6 kinase (11). Previous studies using lymphoid cells and fibroblasts showed that this kinase phosphorylates the S6 polypeptide of the 40S ribosomal subunit in response to mitogenic stimuli (6, 11,20, 30, 42). The two isoforms of the kinase migrate as 70- and 85-kDa proteins on SDS-PAGE. Collectively, these isoforms are termed pp70S6k. The 85-kDa isoform has an additional 23-amino acid nuclear localization signal and is thus directed to the nucleus (34).

The activation of pp70S6k involves phosphorylation of numerous Ser/Thr residues by multiple kinases that appear to be regulated by distinct mechanisms (reviewed in Refs. 11, 17, 30, and32). Blocking experiments using rapamycin, an inhibitor of the mammalian target of rapamycin (mTOR) kinase, and wortmannin, an inhibitor of phosphatidylinositol 3-kinase (PI3K), have established that these kinases regulate pp70S6k activity. However, the complete array of kinases (and their means of regulation) responsible for pp70S6k activation remains to be identified. The complexity of the mechanisms that control the enzyme and its known effects on cellular function establish pp70S6k as a key intracellular checkpoint in the transduction of mitogenic and other extracellular signals.

The biomechanical activation of pp70S6k in endothelial cells has not been described. Here, we demonstrate that pp70S6k in primary cultured human endothelial cells can be independently activated by fluid flow, in addition to humoral agonists, and that changes in its activity are correlated with altered expression of a protein, Bcl-3. We therefore conclude that fluid forces can modify endothelial phenotype by directly regulating specialized pathways that control translation of mRNA transcripts into protein.



Rapamycin was obtained from Alexis (San Diego, CA), tissue culture-grade DMSO was obtained from American Type Culture Collection (ATCC, Rockville, MD), and PD-98059 came from Calbiochem (La Jolla, CA). Platelet-derived growth factors A and B (PDGF-AB), tumor necrosis factor-α (TNF-α), oncostatin M, and interleukin-1β (IL-1β) were purchased from R&D Systems (Minneapolis, MN). [γ-32P]ATP was obtained from Amersham (Arlington Heights, IL). Other chemicals and reagents were purchased from Sigma (St. Louis, MO). All culture media and supplements were from Whittaker Bioproducts (Walkersville, MD) unless otherwise specified. DMEM was purchased from Sigma. SmGM-2 human aortic smooth muscle cell medium came from Clonetics (Walkersville, MD). Heat-inactivated fetal bovine serum (FBS) was supplied by Hyclone (Logan, UT).

Cell culture.

Confluent primary cultures of human umbilical vein endothelial cells (HUVEC) were established in 0.2% gelatin-coated six-well dishes as previously described (48). Cultures were used for experiments 1 day after they reached confluence (∼3–4 days after initial plating). On the day the cultures became confluent, they were serum deprived by a change in medium to medium 199 (M199) containing 0.1% pooled human serum. In experiments in which rapamycin, wortmannin, PD-98059, or actinomycin D was used as an inhibitor, the stated concentrations of these reagents (or an equivalent volume of DMSO carrier) were added to the HUVEC cultures 30 min to 2 h before the application of fluid shear stress.

A549 lung carcinoma cells were obtained from ATCC and cultured in Ham's F-12 medium supplemented with 10% FBS and penicillin/streptomycin. Chinese hamster ovary (CHO) cells were maintained in α-MEM supplemented with 10% FBS, penicillin/streptomycin, and l-glutamine (200 mM). Swiss 3T3 fibroblast cells were also obtained from ATCC and maintained in DMEM supplemented with 10% FBS. Human aortic smooth muscle cells (HASMC) were purchased from Clonetics and cultured in SmGM-2 medium. All cultures were serum deprived when confluent and were used for experimentation on the next day.

Application of fluid flow.

Fluid flow was applied to the confluent cultures with the use of an orbital shaker in accordance with previous reports (23, 29, 43). Although this technique does not result in uniform application of laminar shear stress across the entire monolayer, the majority of the cells are exposed to near maximal shear stress (τmax), which can be calculated asτmax=aρη(2πf)3 wherea is the radius of orbital rotation (1.4 cm), ρ is the density of the culture medium (1.0 g/ml), η is the viscosity of the medium (assumed to be 0.0075 poise), and f is the frequency of rotation (rotations/s) (23). At 200 rpm, τmax = 11.5 dyn/cm2. Selected experiments were also repeated using a cone-plate viscometer, which does impose laminar shear stress on the entire monolayer (8, 24).

Western blotting for pp70S6k, extracellular signal-regulated kinase 1/2, Bcl-3, and E-selectin.

Polyclonal rabbit antibodies against pp70S6k, Bcl-3, extracellular signal-regulated kinase 1 (ERK1; also cross-reactive with ERK2), and goat anti-rabbit horseradish peroxidase-conjugated antibodies were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). A phospho-specific antibody (rabbit polyclonal) against pp70S6k (pThr389) was obtained from New England Biolabs (Beverly, MA).

After experimental treatment, HUVEC cultures were placed on ice and immediately scraped under ice-cold PBS (pH 7.4). Cells were pelleted with gentle centrifugation, resuspended in reducing sample buffer (0.06 M Tris ⋅ HCl, pH 6.8, 2% SDS, 10% glycerol, 0.025% bromphenol blue, and 5% 2-mercaptoethanol) and boiled. The extracted proteins (10–20 μg/sample) were separated by SDS-PAGE and transferred to a polyvinylidene difluoride membrane (Immobilon-P; Millipore, Bedford, MA) for Western blotting. Primary antibodies were detected with 0.2 μg/ml goat anti-rabbit horseradish-peroxidase conjugated antibody. Bound antibody complexes were then visualized by enhanced chemiluminescence (Amersham) and exposure to radiographic film (Eastman Kodak, Rochester, NY). Immunoblotting for E-selectin was performed under nonreducing conditions with sample buffer lacking 2-mercaptoethanol.

Immune-complex kinase activity assay of pp70S6k.

This assay was adapted from a previously described method (37). After experimental treatment, HUVEC were scraped under ice-cold PBS and pelleted by gentle centrifugation. The cells were then resuspended in albumin-free kinase buffer (20 mM HEPES, 150 mM NaCl, 20 mM MgCl2, 25 mM β-glycerophosphate, 1 mM dithiothreitol, 2 mM sodium orthovanadate, 10 μg/ml aprotinin, 100 nM leupeptin, 1 mM phenylmethylsulfonyl fluoride, and 25 μM ATP). A quantitative protein assay was performed on each sample (DC Protein Assay; Bio-Rad, Hercules, CA). pp70S6k was then immunoprecipitated by the addition of equivalent amounts of total protein from each sample (usually 500–750 μg) to separate wells of ELISA-grade 96-well plates (Costar, Cambridge, MA) that had been preadsorbed with protein A (10 μg/ml) at 37°C for 2 h and then coated with polyclonal anti-pp70S6k antibody (1 μg/well dissolved in kinase buffer) for 2 h at 4°C. The plates were incubated for 90 min at 4°C.

A kinase reaction mixture was prepared by the addition of 2 μl of [γ-32P]ATP to a solution containing 100 ng/sample S6 substrate peptide (Santa Cruz Biotechnology, Santa Cruz, CA) diluted in albumin-containing kinase buffer. In vitro phosphorylation (37°C) was initiated by the addition of the reaction mixture (50 μl) to each sample. The reaction was stopped after 30 min with 15 μl of 75 mM phosphoric acid. The samples were then loaded into individual phosphocellulose units (Pierce, Rockford, IL), centrifuged, and washed to remove unincorporated [γ-32P]ATP. The filters were then analyzed in a scintillation spectrometer.


Endothelial pp70S6k responds to fluid flow and selected endothelial agonists.

pp70S6k is partially phosphorylated in quiescent cells (20), giving rise to two or three bands on Western blots. Activation of pp70S6k by sequential phosphorylation results in the appearance of additional, more slowly migrating bands on SDS-PAGE (31). Exposure of HUVEC to fluid flow imposed by an orbital shaker or laminar fluid shear stress in a cone-plate viscometer produced obvious mobility shifts in endothelial pp70S6k (Fig.1 A). In no case did the results from the viscometer differ substantially from those obtained with the orbital shaker.

Fig. 1.

Fluid shear stress and selected endothelial agonists produce an electrophoretic mobility shift in pp70S6k. A: serum-deprived human umbilical vein endothelial cell (HUVEC) monolayers were kept static or exposed to shear stress (11.5 dyn/cm2) for 30 min using either an orbital shaker or a cone-plate viscometer. Samples were then prepared for Western blotting to detect pp70S6k as described in methods. Fluid shear stress triggered attenuation and disappearance of a lower band and appearance of an upper band, indicative of phosphorylation-dependent activation (see text). Blots are representative of numerous experiments (>10). B: serum-deprived HUVEC monolayers were stimulated with indicated agonist for 30 min. Samples were then prepared for Western blotting as described in A. Bottom panel was exposed for a longer period of time to visualize the less-abundant p85 isoform. Samples were treated with PBS (control), oncostatin M (OSM; 10 ng/ml), lipopolysaccharide (LPS; 100 ng/ml); interleukin-1β (IL-1; 10 U/ml), tumor necrosis factor-α (TNF; 10 U/ml), DMSO (1:10,000), phorbol 12-myristate 13-acetate (PMA; 10 nM), thrombin (2 U/ml), histamine (10 mM). Blot is representative of 3 separate experiments.

Because regulation of pp70S6k in endothelial cells has not been characterized, we also determined whether it was responsive to other endothelial agonists (Fig. 1 B). Marked mobility shifts (loss of the most rapidly migrating band) were observed in response to oncostatin M and histamine, whereas intermediate shifts (appearance or enhancement of higher molecular weight bands without loss of the lowest band) occurred in response to IL-1β, TNF-α, and phorbol 12-myristate 13-acetate (PMA). Insulin also induced mobility shifts in pp70S6k in HUVEC (see Responsiveness of pp70S6k to fluid flow depends on cell type). Interestingly, there was little or no shift in pp70S6k mobility in cells treated with thrombin or lipopolysaccharide (LPS) compared with that in control, unstimulated cultures even though adhesion molecule expression was induced in HUVEC treated with these agonists in parallel control assays (data not shown). The p85 nuclear isoform underwent the same phosphorylation pattern as the p70 cytoplasmic isoform in response to agonists that induced mobility (Fig. 1). In all experiments, the two isoforms responded identically. Therefore, only data relating to the p70 isoform are shown in subsequent Figs. 2-6.

The pp70S6k response to fluid flow occurred rapidly, with mobility shifts and increased phosphospecific immunoreactivity clearly evident within 5 min and maximal change apparent at 15 min (Fig.2 A). The flow-induced mobility shift of pp70S6k was also force dependent, with higher molecular weight bands appearing only at ≥100 rpm (∼4.1 dyn/cm2) on the orbital shaker. Higher levels of shear (200 rpm) produced a more pronounced shift with complete loss of both lower bands (Fig. 2 B). Time-course and force-response studies using the cone-plate viscometer produced similar results (data not shown).

Fig. 2.

Induction of pp70S6k phosphorylation by flow is time and force dependent. A: HUVEC were exposed to 11.5 dyn/cm2 peak fluid shear stress (orbital shaker) for indicated times and then processed for Western blotting to detect pp70S6k. Top: blot obtained using a polyclonal antibody to pp70S6k (not phospho-specific). Bottom: blot prepared from same experiment using a phospho-specific anti-pp70S6k (pThr389) antibody. B: HUVEC were exposed to indicated rotations per minute (rpm) for 30 min (orbital shaker) and then processed as in A to detect pp70S6k. Bottom panel depicts immunoreactivity against a phospho-specific antibody raised against pThr389. Calculated peak shear stresses correlating with 0, 25, 50, 100, and 200 rpm are 0, 0.5, 1.5, 4.1, and 11.5 dyn/cm2, respectively. Results are representative of 3 separate experiments using different HUVEC isolates.

Responsiveness of pp70S6k to fluid flow depends on cell type.

To determine whether activation of pp70S6k by fluid forces is a generic response to biomechanical stimulation, we subjected a variety of other cell types to fluid flow using the orbital shaker (Fig. 3). A mobility shift in response to flow was observed in HUVEC (Fig. 3, top) and HASMC (Fig. 3,bottom) but not in CHO cells, A549 epithelioid cells, or Swiss 3T3 fibroblasts (Fig. 3, middle). The signaling pathway to pp70S6k in the cell types unresponsive to flow was competent, because mobility shifts occurred in response to a relevant mitogen (either insulin or PDGF-AB). Thus vascular wall cells may be uniquely capable of transducing fluid forces to pp70S6k.

Fig. 3.

Flow-dependent signaling to pp70S6k depends on cell type. Confluent cultures of indicated cell types were kept static or exposed to peak fluid shear stress (11.5 dyn/cm2, orbital shaker) or a relevant mitogen for 30 min. pp70S6k was then detected by Western blotting. Insulin (100 nM) was the mitogen for HUVEC, Chinese hamster ovary (CHO), and A549 cells, whereas platelet-derived growth factors A and/or B (PDGF-AB; 10 ng/ml) was used to stimulate 3T3 cells and human aortic smooth muscle cells (HASMC). Results are representative of at least 3 different experiments.

In a separate set of studies, no pp70S6k mobility shift was observed in HUVEC exposed to conditioned media from separate shear-stressed HUVEC cultures (data not shown). This finding indicates that shear stress acts directly on endothelial cells to activate pp70S6k, rather than through an autocrine loop involving a stable soluble mitogen.

Endothelial pp70S6k kinase activity is upregulated by fluid flow and inhibited by rapamycin and wortmannin.

To demonstrate that flow-induced phosphorylation was associated with increased pp70S6k enzyme activity, we performed Western blotting in parallel with immune-complex kinase activity assays to verify increased phosphorylation of the S6 substrate following exposure to flow (Fig. 4). Compared with control samples, the samples from HUVEC exposed to shear stress for 30 min contained 56 ± 29% more S6 peptide kinase activity (P < 0.001, n = 9 experiments). These results indicate that the mobility shift observed on SDS-PAGE is a valid surrogate for pp70S6k activation.

Fig. 4.

Flow-induced pp70S6k phosphorylation correlates with increased enzyme activity for S6 peptide and is inhibited by rapamycin and wortmannin. HUVEC were exposed to static or shear (11.5 dyn/cm2, orbital shaker) conditions for 30 min after pretreatment with rapamycin (10 nM), wortmannin (10 nM), or an equivalent concentration (1:10,000) of DMSO carrier. Samples were then scraped under ice-cold PBS and split for parallel processing for Western blotting (top) or an immune-complex pp70S6kactivity assay (bottom) as described in methods. Results displayed are from the same experiment and are representative of 9 different experiments.

In other cell lines, parallel and distinct pathways sensitive to inhibition by rapamycin and inhibitors of PI3K including wortmannin regulate pp70S6k activity (10, 32, 46). We found that endothelial pp70S6k was regulated in the same fashion: both the mobility shift and increased enzyme activity in response to fluid flow were inhibited by rapamycin and wortmannin (Fig. 4). FK-506, a structural analog of rapamycin without inhibitory activity against mTOR in certain other cell types (13, 40), did not inhibit pp70S6k phosphorylation in response to flow (data not shown). These results suggest that signaling pathways dependent on mTOR and PI3K are both necessary for biomechanical activation of pp70S6k in endothelial cells.

Inhibition of the flow-activated mitogen-activated protein kinase pathway does not prevent pp70S6k activation.

Studies using other cell types have indicated that althoughras-dependent ERKs can phosphorylate pp70S6k in vitro (25), the mitogen-activated protein kinase (MAPK) cascade is not the physiological activator of pp70S6k in vivo (3, 13). Nevertheless, ERK1/2 are activated by fluid shear stress in endothelial cells (29, 44) and can mediate increased protein translation in stimulated vascular cells (39). We therefore tested whether shear-induced activation of pp70S6k occurred independently of the MAPK pathway by using a soluble MAPK-ERK kinase inhibitor, PD-98059, that blocks activation of ERK1/2 (18). PD-98059 inhibited the flow-induced phosphorylation-dependent mobility shift of ERK1/2 in endothelial cells but had minimal effect on the phosphorylation of pp70S6k (Fig.5, lane 3). Treatment with rapamycin yielded the typical inhibition of pp70S6kphosphorylation but had no effect on ERK phosphorylation (Fig. 5,lane 4). These results indicate that although fluid shear stress activates ERK1/2 in parallel with pp70S6k (Fig. 5,lane 2), these pathways are distinct in endothelial cells.

Fig. 5.

Fluid flow activates pp70S6k and extracellular signal-regulated kinase 1/2 (ERK1/2) pathways independently. Confluent HUVEC monolayers were exposed to 11.5 dyn/cm2 peak fluid shear stress (orbital shaker) for 5 (ERK1/2) or 30 min (pp70S6k) after pretreatment (30 min) with PD-98059 (10 μM), rapamycin (10 nM), or an equivalent concentration (1:10,000) of DMSO. Cells were then lysed and processed for Western blotting using either an anti-ERK1 or anti-pp70S6k antibody. Results are from the same HUVEC isolate and are representative of at least 3 experiments.

Endothelial Bcl-3 expression is upregulated by shear stress and inhibited by rapamycin.

The rapamycin-sensitive signaling pathway targeting pp70S6kcontrols translation of a distinct subset of mRNA transcripts into protein (6). We determined whether activation of pp70S6k by fluid shear stress resulted in detectable changes in endothelial protein expression by using the expression of Bcl-3 as a marker (26). The mRNA transcript for Bcl-3 has features in common with other transcripts handled by the rapamycin-sensitive translational pathway (6, 26, 47). Furthermore, we found that translation of Bcl-3 is blocked by rapamycin and by inhibitors of PI3K in another cell system (47).

First, we verified an earlier report that resting HUVEC contain Bcl-3 mRNA (28) (not shown). We detected Bcl-3 protein in HUVEC as multiple bands migrating at ∼56 kDa, which represent different states of phosphorylation because phosphatase treatment results in fewer, more mobile bands on SDS-PAGE (unpublished observations; Refs. 9 and 47). We then found that immunodetectable Bcl-3 is rapidly increased (within 15 min) in endothelial cells exposed to fluid flow (Fig.6 A). The time course of Bcl-3 accumulation in HUVEC in response to flow closely parallels the time course of pp70S6k activation (Fig. 2 A).

Fig. 6.

Fluid flow induces endothelial expression of Bcl-3 protein by a rapamycin-sensitive pathway. A: Bcl-3 expression is rapidly upregulated in response to fluid flow. HUVEC were exposed to 11.5 dyn/cm2 peak shear stress (orbital shaker) for indicated times and then processed for Western blotting performed to detect Bcl-3. Results are representative of 2 separate experiments. B:flow-induced Bcl-3 expression is inhibited by rapamycin. HUVEC were exposed to 11.5 dyn/cm2 peak shear stress (orbital shaker) for 1 h after 30-min pretreatment with rapamycin (10 nM) or an equivalent concentration (1:10,000) of DMSO. Samples were then collected and Western blots prepared. C: Bcl-3 upregulation in response to flow occurs despite transcriptional blockade by actinomycin D (Act D). HUVEC were exposed to 11.5 dyn/cm2 shear stress (orbital shaker) for 1 h after pretreatment for 2 h with actinomycin D (10 μg/ml) or an equivalent concentration (1:10,000) of DMSO. Western blots for Bcl-3 were then prepared (left). To ensure that transcription had been blocked by actinomycin D, HUVEC from the same culture were stimulated with TNF-α (100 U/ml) for 4 h in presence of actinomycin D (10 μg/ml) or DMSO (1:10,000). Western blots were then prepared for E-selectin (right).

Rapamycin inhibited upregulation of Bcl-3 by fluid flow (Fig.6 B) as did wortmannin (not shown). Thus the shear-induced pattern of expression of Bcl-3 protein is correlated with the activation and inhibition of pp70S6k and is consistent with regulation by PI3K- and mTOR-dependent pathways.

The rapid upregulation of Bcl-3 protein within 15 min of exposure to fluid shear stress suggests a direct induction of translation. We explored this further by inhibiting transcription in HUVEC with actinomycin D before the application of flow. Although baseline expression of Bcl-3 protein was attenuated by actinomycin D, transcriptional inhibition failed to eliminate the flow-induced upregulation of Bcl-3 (Fig. 6 C, left). In contrast, TNF-α-induced expression of E-selectin (an event that is entirely dependent on transcription; Ref. 14) was completely abolished (Fig.6 C, right). These data indicate that increased synthesis of Bcl-3 can occur in the absence of transcription. Fluid flow also resulted in increased incorporation of [35S]Met into several bands migrating at 55–60 kDa in autoradiographs prepared from HUVEC lysates immunoprecipitated with the anti-Bcl-3 antibody (not shown). Taken together, these findings indicate that the flow-induced upregulation of Bcl-3 (Fig. 6 A) occurs, at least in part, via increased rates of translation mediated by biomechanical activation of mTOR- and PI3K-dependent pathways (Fig. 6 B) and correlates closely with the time course of pp70S6k activation (Fig. 2 A).


In this report we show that fluid flow, a biomechanical stimulus particularly relevant to endothelial cells, rapidly activates pp70S6k, leading to increased phosphorylation of the S6 polypeptide component of the 40S ribosomal subunit. Activation of pp70S6k by flow is blocked not only by rapamycin but also by wortmannin, suggesting that PI3K is also subject to biomechanical control. Accumulation of Bcl-3, a protein regulated by a rapamycin-sensitive translational control pathway (47), occurs rapidly after exposure to flow. This response is abolished by rapamycin, an inhibitor of mTOR, but persists despite transcriptional blockade. Because Bcl-3 transcripts are constitutively present, these findings suggest that fluid forces can directly regulate translation of new protein from available mRNA. Although fluid shear stress is known to influence nuclear signaling and transcriptional events in endothelium, these are the first data demonstrating that fluid flow regulates a translational checkpoint (pp70S6k) in human endothelial cells.

Fluid flow activates pp70S6k.

We showed that flow activates pp70S6k (Figs. 1-3) and that this may be a response unique to cells of the vessel wall (Fig.3). The transient nature of pp70S6k activation by flow (Fig. 2) is similar to that observed for shear activation of ERK1/2 (44). Activation of these pathways in a controlled, transient manner allows a discrete signal to be communicated from the extracellular environment to the interior of the cell and restores the capacity of the cell to respond to additional stimuli at a later time.

The mechanism by which pp70S6k is activated has proven to be quite complex (11, 17, 30, 32). Multiple, distinct hierarchical phosphorylation steps involving more than five separate Ser/Thr residues must occur for full and complete activation. The activation sequence appears to culminate in phosphorylation of Thr-229 in the catalytic domain by 3-phosphoinositide-dependent kinase 1 (PDK1) (2,33). However, PDK1 is constitutively active and, despite its name, appears to not require the presence of 3-phosphoinositides to phosphorylate Thr-229 (2). [PDK1 does require the presence of 3-phosphoinositides to activate its other known substrate, protein kinase B (2).] Access to Thr-229 on pp70S6k by PDK1 appears to require phosphorylation of other residues including Thr-389 [by mTOR (7)] to relieve steric hindrance (17, 30). It is likely that the extracellular signals that regulate pp70S6kactivity control the phosphorylation state of these other Ser/Thr residues, which produce the structural conformation that obscures Thr-229 from PDK1. These signals appear to be mediated by both protein kinase C (PKC)- and PI3K-dependent pathways (11, 17, 30).

PKC has been implicated in pp70S6k activation, and we have observed pp70S6k mobility shifts in endothelial cells treated with PMA (Fig. 1). The ras-dependent MAPK cascade is an obvious candidate PKC-dependent pathway (38) that might mediate signals to pp70S6k. ERK1/2 can phosphorylate pp70S6k in vitro (25) and are also activated by fluid shear stress (29, 44). However, phosphorylation of pp70S6k by ERK1/2 does not result in physiological activation (25). Our results also indicate that the signaling pathway to pp70S6k is distinct from the MAPK pathway, because we observed phosphorylation of pp70S6k by fluid flow even when the MAPK pathway was blocked by PD-98059 (Fig. 5).

Although PKC can signal to pp70S6k, PI3K-dependent signaling appears to be dominant (11, 12). Whereas PDK1 can phosphorylate Thr-229 in the absence of 3-phosphoinositides, other regulated phosphorylation steps appear to depend on PI3K activation (17, 30). Our finding that pp70S6k phosphorylation induced by flow is inhibited by wortmannin (Fig. 4) suggests that PI3K is also activated by shear stress. A recent report that shear-dependent activation of akt (protein kinase B) is blocked by PI3K inhibitors is also consistent with this interpretation (16).

Fluid flow activates a rapamycin-sensitive translational control pathway.

We wanted to demonstrate that the activation of pp70S6k by flow is associated with changes in protein expression by endothelial cells. Because of the central role of pp70S6k in the rapamycin-sensitive translational control pathway (6, 30), we examined whether flow could induce translation of a marker protein, Bcl-3, and whether this response could be inhibited by rapamycin. Although the biological function of Bcl-3 is obscure, its expression was studied because its mRNA possesses characteristics that predict translational control by a rapamycin-sensitive pathway and because we recently found that it is synthesized via such a pathway in stimulated human platelets (27, 47). We observed a similar pattern of regulated Bcl-3 expression in HUVEC exposed to fluid flow. Translation of Bcl-3 was increased with a time course that closely paralleled activation of pp70S6k. Furthermore, rapamycin inhibited flow-induced Bcl-3 synthesis (Fig. 6).

Rapamycin forms a complex with an intermediate, FK-506-binding protein-12, which then binds to mTOR and inhibits its activity (5). mTOR is an upstream mediator of pp70S6k and has recently been shown to phosphorylate Thr-389 and to also control the phosphorylation state of several other pp70S6k residues (32). Rapamycin inhibits proliferation in mesenchymal (13) and lymphoid (1, 19) cell lines and selectively inhibits translation of mRNAs that share structural features with Bcl-3 (6, 21, 41).

It is now known that mTOR lies upstream of at least two different translational regulators: pp70S6k and 4E-binding protein-1 (4E-BP1) (6), which regulates the activity of eukaryotic initiation factor 4E (22). Theoretically, the inhibitory effects of rapamycin on Bcl-3 expression could be due to inhibition of 4E-BP1. However, rapamycin remains an informative reagent for the study of pp70S6k because the branch point of the rapamycin-sensitive pathway leading to 4E-BP1 occurs immediately upstream from pp70S6k (45). Experiments in which blocking antibodies directed against pp70S6k are microinjected into cells mimic the effects of rapamycin treatment (34). Furthermore, we are unaware of any stimulus that selectively activates pp70S6k over 4E-BP1 or vice versa. Although we have not shown unequivocally that Bcl-3 expression requires activation of pp70S6k (as opposed to 4E-BP1), we believe that these events are mechanistically linked for several reasons. First, pp70S6k is a key component of the rapamycin-sensitive translational control pathway (6). Second, both pp70S6k and Bcl-3 are induced by a biomechanical stimulus. Third, rapamycin inhibits both pp70S6k activation and the flow-induced accumulation of Bcl-3.

In summary, most studies exploring the mechanism(s) by which shear stress alters endothelial phenotype have focused on gene transcription and its regulatory processes (35). Ultimately, cellular phenotype is dependent on not only the transcription of the mRNA species for these genes but also the synthesis of the protein products encoded by these transcripts. We show for the first time that fluid flow can specifically regulate the activity of pp70S6k, a key enzyme in specialized translational control, in human endothelial cells. Thus fluid flow may influence endothelial gene expression and vascular phenotype by initiating signals to posttranscriptional checkpoints as well as to transcriptional pathways. Regulation at this additional level provides the cell with a means of governing the translation of constitutively present mRNA species and also allows the cell to respond to an external stimulus more rapidly than if transcription of new mRNA were required. Thus coordinate activation of distinct pathways targeting transcription and translation by fluid forces likely adds precision and versatility to the mechanisms that determine endothelial phenotype.


We gratefully acknowledge the valuable technical assistance of Donelle Benson.


  • Address for reprint requests and other correspondence: L. W. Kraiss, Program in Human Molecular Biology and Genetics, Eccles Institute of Human Genetics, Bldg. 533, Univ. of Utah, 15 North 2030 East, Salt Lake City, UT 84112 (E-mail: larry.kraiss{at}hmbg.utah.edu).

  • This work was supported in part by American Heart Association-Utah Affiliate Grant 9606235S to L. W. Kraiss, who is the recipient of a Wylie Scholarship in Academic Vascular Surgery from the Pacific Vascular Research Foundation and a member of the Fellowship-to-Faculty Transition Program at the University of Utah, which is supported by a grant from the Howard Hughes Medical Institute. This work was also supported by the Nora Eccles Treadwell Foundation and National Heart, Lung, and Blood Institute Grant HL-44525.

  • These data were presented in preliminary form at the 70th Scientific Sessions of the American Heart Association in Orlando, FL, in October 1997 (Circulation 96: I-49, 1997).

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.


View Abstract