Shear Stress Regulates Forward and Reverse Planar Cell Polarity of Vascular Endothelium In Vivo and In Vitro
http://www.100md.com
Shannon McCue, Dorota Dajnowiec, Feng Xu
参见附件。
the Department of Laboratory Medicine and Pathobiology (S.M., D.D., B.L.L.), University of Toronto, Canada
Toronto General Research Institute (S.M., D.D., F.X., M.Z., B.L.L.), University Health Network, Canada
Department of Anatomy and Cell Biology (M.R.J.), College of Medicine, University of Florida, Gainesville.
Abstract
Cultured vascular endothelium displays profound morphological adaptations to shear stress that include planar cell polarity (PCP) that is directed downstream. Endothelial cells in blood vessels are also polarized; however, the direction of polarity is vessel specific, and shear-independent mechanisms have been inferred. The regulation of endothelial PCP is therefore controversial. We report that the direction of PCP in blood vessels is age and vessel specific; nonetheless, it is caused by shear-related regulation of glycogen synthase kinase-3 (GSK-3), a profound regulator of endothelial microtubule stability. When GSK-3 is inhibited, PCP reverses direction. Endothelium is the only cell type studied to date that can reverse direction of polarity. Tight regulation of GSK-3, microtubule dynamics, and cell polarity was also required for the striking morphological responses of endothelium to shear stress (cell elongation and orientation with shear). Finally, the cytoskeletal polarity displayed in blood vessels is associated with polarized (shear-directed) cell mitoses that have important effects on endothelial repair. Vascular endothelium therefore displays a novel mode of mechanosensitive PCP that represents the first example of a single cell type that can reverse direction of polarity.
Key Words: shear stress endothelium microtubules planar cell polarity
Introduction
Vascular endothelial cells display profound morphological adaptations to shear stress that include cell elongation in the shear direction that is driven by a novel reorganization of actin in stress fibers,1 redistribution of focal adhesion complexes,1,2 and partial disassembly and reassembly of cell-cell junctional complexes.3 Imposition of shear stress on cultured endothelium also induces planar cell polarity (PCP) in the downstream direction.4 PCP occurs when cell organelles, cytoskeleton, and/or adhesion complexes exhibit unidirectional organization along an axis that lies in the plane of a cell monolayer. This process was manifest in sheared endothelium in vitro as redistribution of microtubules and the microtubule-organizing center (MTOC) (from which microtubules emanate) to the downstream side of the cell nuclei, a hallmark of PCP.5
PCP has been of great interest to cell biologists because it provides directionality to cell functions. PCP in vertebrate cells has been studied primarily during directional migration at wound margins where it is initiated by integrin activation and multiple downstream pathways that involve the Rho GTPase, Cdc42, adenomatous polyposis coli (APC), glycogen synthase kinase-3 (GSK-3), and mammalian Diaphanous-1 (mDia1).6,7 Polarization of the microtubule system is important because motor proteins (kinesins and dyneins) can deliver cargoes (eg, cell membrane and protein) along microtubules to and from the cell anterior and posterior and because microtubules regulate the reorganization of cell adhesion complexes.8 This process is enhanced by microtubule stabilization through posttranslational modifications to tubulin that are coinduced with morphological polarity.9 The significance of these processes to sheared cultures of endothelium is uncertain because polarity is retained at 24 hours,4 when the cells are no longer motile.1
Endothelial cells display PCP in intact blood vessels10 as well as in vitro; however, the origins of this polarity are controversial because microtubule polarity in vivo varies among blood vessels.10,11 Specifically, PCP is often directed upstream in arteries and downstream in veins, a finding that has led to the inference that it is independent of shear stress.
We have investigated the role of microtubules in shear-dependent motility and PCP of endothelium both in vitro and in blood vessels. In both cases, PCP is shear dependent, and it is tightly regulated by GSK-3. Indeed, manipulation of GSK-3 could induce polarity of these cells to reverse direction, a capacity not displayed by any other cell type studied to date; furthermore, regulation of GSK-3 had dramatic effects on microtubule stability. Primary effects of GSK-3 were partly attributable to its influence on stability of microtubules, and these effects were critical to endothelial adaptations to shear stress because the cells failed to elongate and align with shear when GSK-3 activity or microtubule dynamics were suppressed with exogenous inhibitors or mutant forms of the kinase. GSK-3–related regulation of microtubules is therefore critical to morphological adaptations of endothelium to shear stress.
Materials and Methods
An expanded Materials and Methods section is available in the online data supplement at http://circres.ahajournals.org.
Imposition of Physiological Levels of Shear Stress
Porcine aortic endothelial cells were grown to 2 days postconfluence on glass slides or cover slips and then subjected to shear stress of 15 dyne/cm2 in parallel plate flow chambers. When indicated, GSK-3 activity was inhibited by treatment with LiCl or SB415286. Nocodazole (100 mmol/L) or taxol (5 ng/mL) were used to disrupt or hyperstabilize microtubules, respectively.
Immunofluorescence
Immunofluorescence staining of cells was viewed by laser scanning confocal microscope. For studies of -tubulin localization, nuclei were counterstained with propidium iodide or Sytox green.
Analysis of Endothelial Cell Alignment With Shear Stress
Endothelial cell elongation in the shear direction was assessed by determining cell length:width ratios using a novel, simple, and accurate methodology that precludes inaccuracies inherent in "shape factor" determinations that rely on cell perimeter-area relations (see online data supplement).
Microtubule-Organizing Center Localization
MTOCs were grouped (see supplemental Figure I) as upstream of nucleus (scored 1), upstream of midpoint of nucleus (scored 2; includes cells scored 1), and downstream of the midpoint of the nucleus (scored 3), which included cells downstream of the nucleus (scored 4).
Polarization of Acetylated Microtubules
Cultures were immunostained against acetylated tubulin and -catenin, and nuclei were counterstained with propidium iodide. An observer, blinded to whether shear direction was left-to-right or right-to-left, assessed the position of the majority of stable microtubules in each cell relative to the nucleus. Decoded data were presented as a percentage of cells with stable microtubules predominantly upstream versus downstream of the nucleus.
Immunoblotting
Ten micrograms of protein from cell lysates were subjected to polyacrylamide gel electrophoresis, and blots were probed with antibodies as indicated. Labeled secondary antibodies were visualized using enhanced chemiluminescence.
Microinjections
In 3 experiments, nuclei of at least 30 endothelial cells in postconfluent monolayers were microinjected with constitutively active GSK-3 (GSK3S9A). Coinjection of rhodamine-labeled dextran was used to localize microinjected cells. After 3 hours, flow of culture medium was initiated and cells were then examined after 24 hours of shear stress.
In Vivo Assessment of Cell Polarity
Blood vessels from adult and 5-week-old rabbits were immunostained en face. Some rabbits were treated with 7 mmol/L lithium carbonate in drinking water to inhibit GSK-3 for 48 hours before euthanasia.12,13
To examine the effects of reversing flow/shear direction in vivo, rabbits were anesthetized and the midabdominal aorta was narrowed by 50% with 5-0 suture.14 These stenoses produce a downstream vortex that reverses flow near the endothelial surface for up to 10 vessel diameters downstream of the stenosis (Figure 6A). One week later, the rabbits were killed by barbiturate overdose and vessels segments in the vortex region were stained en face.
To test for polarized mitosis, 3- and 5-week-old rabbits were given intramuscular injections of bromodeoxyuridine (BrdUrd). En face BrdUrd immunostaining was performed after 24 hours. At this time, BrdUrd-labeled cells had completed mitosis, so that doublets of daughter cells were detected and the relative positions of the labeled nuclei indicated the direction of mitosis. Alternatively, cultured cells were pulse labeled with BrdUrd immediately before shearing.
These experiments were approved by the Animal Care Committee of the University Health Network and were conducted in accord with the standards of the Canadian Council on Animal Care.
Statistics
Sample sizes are presented in figure legends. Effects of drugs (nocodazole, taxol, SB415286) on cell length-width ratios were assessed using a Dunnett’s test. Differences in MTOC position after shear stress, and position relative to the nucleus of acetylated microtubules, were determined using Student’s t tests. P<0.05 was taken as statistically significant.
Results
Microtubule Dynamics Is Required for Shear-Induced Changes in Endothelial Cell Shape
Postconfluent endothelial cells elongated in the direction of shear stress within 24 hours,1 with mean cell length in the shear direction becoming 2.5 to 3 times cell width (Figure 1A, 1B, and 1E). Microtubules in these cells emanated from the MTOC and projected throughout the cytoplasm, except that the perijunctional region was frequently spared (Figure 1A and 1B). These filaments became highly oriented in the direction of shear stress. Microtubules are required for cell elongation because treatment with nocodazole prevents cell elongation.15 This effect was observed even with nocodazole doses that fail to completely disrupt microtubules (Figure 1C and 1E). Furthermore, cells did not elongate when dynamic instability of microtubules was suppressed using the microtubule-stabilizing drug taxol. With either drug, cell "length" was not significantly different from cell "width" after 24 hours of shear stress (Figure 1D and 1E). These findings indicate an obligate role for dynamic instability of microtubules in the shear stress-induced motility of endothelial cells.
Shear Stress-Induced Planar Cell Polarity Is Accompanied by Profound Stabilization of Microtubules
Recently, Tzima et al found that the microtubule system (MTOC) of endothelium subjected to shear stress displayed Cdc42-dependent polarization to the downstream side of the nucleus after 24 hours of shear stress,4 although these findings contradicted a previous report of transient and upstream polarization at 6 hours of shear.16 We failed to detect polarization of microtubules or MTOCs after 6 hours of comparable levels of shear stress (not shown); however, we confirmed a high degree of downstream polarization after 24 hours (Figure 2A, 2B, and 2D).
Microtubules can be stabilized by posttranslational modifications to tubulin.5 Acetylation of tubulin coincides with stabilization,5 and it can be detected with an acetylation-specific antibody.17 Immunoblots detected low levels of acetylated tubulin in postconfluent monolayers of endothelium that were not exposed to shear stress; however, the acetylated form was dramatically upregulated by 1 hour of shear stress, and further increases were seen by 24 hours (Figure 3A). When the same antibody was used for immunofluorescence studies, only a sparse population of perinuclear microtubules was labeled in static cultures (Figure 3B), whereas extensive labeling of much of the microtubule population was detected after 24 hours of shear (Figure 3C). Stabilized microtubules were found predominantly downstream of the cell nucleus in sheared cells (Figure 3D). These findings indicate major modifications to microtubules under the influence of shear stress. Importantly, the cells have completed shape change by 24 hours,1 so polarity and microtubule stability are not associated with directional motility at this time.
GSK-3 Regulates Stabilization and Polarization of Microtubules by Shear Stress
GSK-3 regulates planar cell polarity in migrating cells; furthermore, this kinase suppresses microtubule stability by phosphorylating multiple microtubule stabilizing proteins.18,19 We therefore assessed the role of GSK-3 in endothelial cell responses to shear stress. Western blots failed to detect the inactive (phosphorylated) form of GSK-3 (p-GSK-3) in static cultures (Figure 4A). In contrast, p-GSK-3 was readily detected in sheared cells over a time course that matched that of tubulin acetylation, except that p-GSK-3 waned at 24 hours, when tubulin acetylation was maximal (Figures 3A and 4A). Total GSK levels were insensitive to shear stress in these experiments (not shown).
To test whether GSK-3 regulates morphological responses to shear stress, we exposed cells to shear with and without treatment with the GSK-3 inhibitors LiCl and SB415286. Cell elongation in response to shear stress was totally blocked by GSK-3 inhibition (Figure 4B through 4D). Our findings that GSK-3 is suppressed by shear stress, whereas total inhibition using exogenous inhibitors of GSK-3 blocks shear-induced changes in cell morphology, indicate that tight control over this kinase is critical to endothelial responses to shear. Accordingly, cells failed to elongate after nuclear microinjection of vectors expressing constitutively active GSK3S9A (Figure 4E and 4F). Polarity of MTOCs also was compromised, ie, 57% of MTOCs were downstream of the nuclear midpoint of microinjected cells, compared with more than 90% of uninjected cells. As expected, we also observed a sparser distribution of acetylated microtubules in cells expressing GSK3S9A. Microinjection of a kinase dead construct (GSK3K85A) was without effect, consistent with reports that this mutation does not yield dominant-negative function (J.A. Woodgett, personal communication, 2005).
Given the importance of microtubule dynamics in control of cell shape and the capacity of GSK-3 to regulate microtubule stabilization in other cells, we examined the capacity of GSK-3 to influence microtubule stability in unsheared endothelium. Treatment with LiCl (or SB415286; supplemental Figure II) to inhibit GSK-3 caused the distribution of stable microtubules to convert from a sparse perinuclear expression (Figure 3B) to an elaborate network (Figure 5A). This stabilization resulted in extensive elongation of microtubules, so that the total microtubule population circled the cell periphery in a toroid pattern (Figure 5B).
We asked whether microtubule polarity, as indicated by MTOC position, was affected by inhibition of GSK-3. Based on studies of migrating cells, we anticipated randomization of MTOC position after shearing cells in the presence of an exogenous inhibitor.20 Intriguingly, polarity was reversed by this treatment (Figure 2C and 2D), ie, the MTOCs redistributed to the upstream side of the nucleus after initiation of shear. These findings suggest that shear stress induces competing, GSK-3–dependent and GSK-3–independent, modes of polarity control in endothelium.
Endothelial Cell Polarity is Age and Vessel Specific In Vivo
MTOCs in the thoracic aorta and vena cavae of rats and pigs localize to the heart side of the nucleus (upstream in arteries, downstream in veins).11,21 The same is true for carotid arteries and vena cavae of rabbits, although polarity declines with age in the latter vessel, and it is not evident in thoracic aortas in this species.11,22 These findings led to the inference that endothelial polarity in vivo is cardiocentric and independent of blood flow. Although we confirmed that MTOCs were downstream of nuclei in the vena cava of the adult rabbit, upstream in the adult carotid artery and unpolarized in the adult thoracic aorta (data not shown), we also found that endothelium of the abdominal aorta reversed polarity with age; MTOCs were downstream of nuclei in 5-week-old rabbits and upstream in abdominal aortas of mature rabbits (Figure 6A and 6B and 6C, solid bars). MTOCs were distributed predominantly downstream of the nucleus in the vena cavae of young (Figure 6D) and mature (not shown) rabbits. These findings indicate that polarity of endothelium in vivo is vessel and age specific, and it is not exclusively cardiocentric.
Endothelial Cell Mitoses Also Polarize in the Shear Direction
Polarization of cell morphology and the microtubule system may prime cells for polarized mitosis; therefore, we also tested whether spontaneous mitoses of endothelial cells in immature rabbits were consistently unidirectional. When S-phase cells were labeled in vivo with BrdUrd and endothelial surfaces of arteries were prepared for imaging 24 hours later, the relative positions of the 2 daughter cells was consistently along the vessel axis. Figure 7A illustrates this feature in an in vivo preparation of the carotid artery of a 3-week-old rabbit; however, identical results were obtained in aortas and carotid arteries of 5-week-old rabbits. This response was genuinely flow derived because nonaxial mitoses were observed in the blind-ended stumps of carotid arteries that had been ligated (Figure 7B); furthermore, mitosis were flow aligned in vitro as well as in vivo. Indeed, even though in vitro flow was initiated (immediately) after exposure to BrdUrd, cell division was aligned with shear (Figure 7C through 7E). This alignment included cells that, after 24 hours of shear stress, were in telophase (Figure 7C), cytokinesis (Figure 7D), or were postmitotic (Figure 7E).
Endothelial Cell Polarity Is Shear Dependent In Vivo
Our finding that, under different conditions, shear stress can induce either upstream or downstream polarization of MTOCs in cell cultures led us to examine its influence on MTOC polarity in intact blood vessels. We surgically produced 50% narrowing (stenosis) of the midabdominal aorta of both 5-week-old and mature rabbits. This procedure causes a vortex to form downstream of the stenosis, and we confirmed previously that blood flow near the vessel wall, and therefore wall shear stress, was directed toward the heart for many vessel diameters distal to the stenosis14 (Figure 8A). Flow reversal at the vortex site eliminated the downstream polarity of the MTOCs in immature abdominal aortas and the upstream polarity in mature vessels (Figure 8B and 8C). A slight reversal of these polarities was not statistically significant. These findings indicate that polarization of endothelial MTOCs in vivo is shear dependent.
Endothelial Cell Polarity Is Regulated by GSK-3 In Vivo
The flow dependence of endothelial cell polarity both in vivo and in vitro, and the capacity of GSK-3 inhibition to reverse shear-induced polarity in vitro, prompted us to assess the effects of GSK-3 inhibition in vivo. When lithium (LiCO3) was administered to rabbits for 48 hours before examination of the position of the MTOCs in blood vessels, both the upstream polarity of abdominal aortic endothelium of mature animals and the downstream polarity of endothelium of the abdominal aorta and vena cava of immature rabbits (Figure 6B through 6D) were suppressed. Furthermore, immunofluorescence staining for acetylated tubulin was suppressed by obstructing flow by carotid ligation for one week and partially restored in rabbits treated with LiCO3 in drinking water (supplemental Figure III). These findings confirmed that GSK-3 is an important regulator of shear-induced endothelial polarity both in cell culture and in blood vessels.
Discussion
This study has shown that shear stress induces tight control over microtubule organization, stability, and function and that this control is absolutely required for the dramatic morphological responses of endothelium to shear stress. Accordingly, drugs that induce either hyperstabilization of microtubules (taxol) or microtubule disassembly (nocodazole), or expression of a mutant (constitutively active) form of an important regulator of microtubules (GSK-3), blocked cell elongation and orientation under the influence of shear.
We confirmed that MTOCs displayed downstream polarity when endothelial cells were subjected to shear stress,4 and we showed that shear induces marked stabilization of microtubules as evidenced by tubulin acetylation. Although the effects were profound, stabilization must be tightly regulated during cell polarization, given our findings that the microtubule stabilizing agent, taxol, blocked cell morphological responses to shear stress. These findings demonstrate that shear stress induces primary manifestations of PCP.
Because shear had major effects on microtubules that proved important to cell morphological responses, we examined regulation of microtubule modifications. When cells migrate at wound edges, integrins are activated at the cell anterior, as new cell-substrate interactions occur, and downstream activation of Cdc42 leads to the anterior assembly of a complex that includes Cdc42 and the scaffold protein, Par6, atypical protein kinase C- (aPKC), and GSK-3. An outcome of this assembly is inhibitory phosphorylation of GSK-3, which is pivotal to establishment of PCP.20 GSK-3 is multifunctional, but its roles in PCP appear related to its capacity to dissociate APC from microtubule + ends.20 How inhibition of GSK-3, and the resulting association of APC with the peripheral tips of microtubules, effects microtubule redistribution is unknown, although it may promote microtubule binding to proteins/structures at the cell anterior that prevents them by being swept backward by retrograde flow of actin.23 Alternatively, motor proteins (dyneins) that are captured at the cell anterior may ratchet microtubules anteriorly.24 Although the other major manifestation of PCP, microtubule stabilization, appears to be controlled independently of MTOC redistribution,24 inhibition of GSK-3 again is involved. APC promotes the association with microtubules of stabilizing proteins such as EB1,19 and we found that phosphorylation of GSK-3 had profound effects on microtubule stability in endothelial cells.
Based on the pivotal role of GSK-3 in PCP of migrating cells, we examined its role in shear-induced responses. Shear-induced phosphorylation of GSK-3 coincided with acetylation of tubulin in microtubules, and we infer a causal link. At late time points, GSK-3 phosphorylation waned, whereas tubulin acetylation was sustained; therefore, other factors may participate in maintenance of shear-induced microtubule stability. Nonetheless, a role for inhibition of GSK-3 was strongly supported by our finding that exogenous inhibitors greatly enhanced tubulin acetylation in unsheared cells and so strongly stabilized these structures that they formed extremely long filaments that encircled the cell periphery in a toroid pattern. Again, tight regulation, rather than simple suppression, of GSK-3 activity was required for cell morphological responses because the cells failed to elongate when the activity of the kinase was blocked using exogenous inhibitors. Regulation of intracellular gradients, as well as absolute activity, of GSK-3 is probably required.20
Surprisingly, exogenous inhibitors of GSK-3 activity did not suppress MTOC polarity, as occurs in migrating cells20; they reversed it so that cells polarized in the upstream direction. Recent work indicates that at least some epithelial cells display posterior PCP when migrating; however, this is the first example of a single cell type that can reverse direction of polarity. Endothelium may therefore provide an informative model for studying directional polarity control. We infer from this observation that endothelial cells can undergo GSK-3–independent, shear-induced reverse PCP that is unmasked by the inhibitors.
Reversal of shear-induced microtubule polarity by inhibition of GSK-3 was particularly interesting given that, in some species, endothelial MTOCs are upstream of nuclei in arteries and downstream in veins.10,11,21 This previous observation was interpreted as indicating flow independence of endothelial cell polarity and novel alternatives were proposed, including sheet-like migration of peripheral vascular endothelium toward the heart.25 We found, however, that cardiocentric polarity was not ubiquitous in vivo; furthermore, polarity in the abdominal aorta of rabbits reversed with age. This finding, plus our observation that shear can induce upstream or downstream polarity in vitro, depending on p-GSK-3 levels, prompted us to directly assess the influence of shear on endothelial polarity in vivo. Experimental reversal of blood flow in the vortex that forms downstream from stenoses in vivo eliminated endothelial polarity, regardless of whether polarity was directed upstream or downstream in the unmanipulated artery. A modest reversal of polarity was not statistically significant, which may reflect either the less pronounced polarity that is observed in vivo or the precise nature of shear stress and its temporal variations downstream from stenoses. Nonetheless, these findings demonstrate that endothelial cell polarity in blood vessels is shear dependent.
Endothelial cell polarity was GSK-3 dependent in blood vessels, as it was in cell culture. This hypothesis was testable because lithium administration, to inhibit GSK-3 activity, is frequently used to treat bipolar disorder in humans.26 Lithium given in drinking water to rabbits suppressed or reversed polarity in both arteries and veins. Failure to always observe reversal of downstream polarity, as was seen in cell cultures, may indicate that circulating lithium levels fluctuated substantially when the rabbits were given lithium-treated water ad libidum, so that GSK-3 may have transiently but repeatedly escaped inhibition.
How microtubules influence morphological adaptations to shear stress is unclear. While the cell is changing shape, polarized microtubules might deliver cell membrane, protein and other cargoes to the cell poles (downstream pole in vitro) as well as regulate disassembly of substrate adhesion complexes. However, this simple interpretation is not readily reconciled with our previous finding that the actin assembly that drives shear-induced cell elongation proceeds with equal frequency in the upstream and downstream directions.1 Furthermore, polarization and stabilization of microtubules persisted at 24 hours of shear, after shape change had reached completion.1 The latter finding suggests that microtubule functions in planar cell polarity go beyond contributing to motility. Additional functions are unknown, although we have argued that the highly oriented distribution of the cytoskeleton and adhesion complexes in cells elongated by shear represents a low entropy state and that the organization of these dynamic structures must undergo constant renovation to prevent drift toward random orientation. The persistent interactions of highly stabilized and oriented microtubules with adhesion complexes might contribute to continuous maintenance of these structures, as might the steady state actin assembly/turnover that we described previously.1
The high degree of PCP displayed by endothelial cell morphology and cytoskeleton led to the hypothesis that mitosis would exhibit directional bias as well. Accordingly, spontaneous mitoses in the arterial endothelium of young, rapidly growing rabbits ubiquitously displayed mitosis along the shear vector. Shear dependence was confirmed when polarized mitoses were observed in sheared endothelium in vitro and when disruption of blood flow by carotid ligation caused orientation of mitoses to randomize in vivo. These findings indicate that mechanical forces may be potent regulators of directional tissue growth and repair. Possibly, mitotic spindle formation during polarized mitosis is associated with the biased distribution of microtubules that is induced by shear stress; alternatively, there may be an interplay between forces generated by mitotic microtubules and associated motor proteins at the cell cortex that align mitosis27 and externally applied forces that are transmitted to these structures.
These adaptations of endothelium to shear may have implications for important vascular pathologies. For example, endothelial denudation when arteries are traumatized by angioplasty or stent implantation contributes to restenosis; consequently, rapid repair of endothelium is important to arterial integrity. Shear-induced endothelial cell polarity directs cell elongation and proliferation preferentially in the long axis of the vessel, the effects of which are vividly demonstrated by examining outgrowth of endothelium from branches of denuded arteries, which proceeds much more rapidly along the vessel axis than circumferentially.28 If substantial lengths of vessel are denuded, a bias in cell motility and mitosis in the lengthwise direction will accelerate repair.
Furthermore, continual microtubule-dependent remodeling of the endothelial monolayer in response to changing blood flow conditions, and associated disruption of cell-cell junctions,3 may contribute to high endothelial permeability in regions of complex blood flow at arterial branch sites and thereby contribute to the predilection of these sites for atherogenesis.29 At these locations, magnitude and direction of shear stress changes are caused by exercise, circadian rhythms in cardiac output, or longer-term changes in cardiovascular function that are associated with exercise regimens, reproductive cycles, pregnancy, and disease states.30
Acknowledgments
This work was supported by operating grant 15044 from the Canadian Institutes for Health Research and a studentship (to D.D) from Canadian Institutes of Health Research and the Canadian Hypertension Society. We thank Donna Johnston for technical assistance and James Woodgett for providing mutant GSK-3 constructs.
Footnotes
Original received July 22, 2005; resubmission received February 1, 2006; accepted February 24, 2006.
References
Noria S, Xu F, McCue S, Jones M, Gotlieb AI, Langille BL. Assembly and reorientation of stress fibers drives morphologic changes to endothelial cells exposed to shear stress. Am J Pathol. 2004; 164: 1211–1223.
Girard PR, Nerem RM. Shear stress modulates endothelial cell morphology and F-actin organization through the regulation of focal adhesion-associated proteins. J Cell Physiol. 1995; 163: 179–193. [Order article via Infotrieve]
Noria S, Cowan DB, Gotlieb AI, Langille BL. Transient and steady-state effects of shear stress on endothelial cell adherens junctions. Circ Res. 1999; 85: 504–514.
Tzima E, Kiosses WB, Del Pozo MA, Schwartz MA. Localized cdc42 activation, detected using a novel assay, mediates microtubule organizing center positioning in endothelial cells in response to fluid shear stress. J Biol Chem. 2003; 278: 31020–31023.
Gundersen GG, Gomes ER, Wen Y. Cortical control of microtubule stability and polarization. Curr Opin Cell Biol. 2004; 16: 106–112. [Order article via Infotrieve]
Raftopoulou M, Hall A. Cell migration: Rho GTPases lead the way. Dev Biol. 2004; 265: 23–32. [Order article via Infotrieve]
Etienne-Manneville S, Hall A. Cell polarity: Par6, aPKC and cytoskeletal crosstalk. Curr Opin Cell Biol. 2003; 15: 67–72. [Order article via Infotrieve]
Kaverina I, Krylyshkina O, Small JV. Microtubule targeting of substrate contacts promotes their relaxation and dissociation. J Cell Biol. 1999; 146: 1033–1043.
Lin SX, Gundersen GG, Maxfield FR. Export from pericentriolar endocytic recycling compartment to cell surface depends on stable, detyrosinated (Glu) microtubules and kinesin. Mol Biol Cell. 2002; 13: 96–109. [Order article via Infotrieve]
Rogers KA, McKee NH, Kalnins VI. Preferential orientation of centrioles towards the heart in endothelial cells of major blood vessels is reestablished after reversal of a segment. Proc Nat Acad Sci U S A. 1985; 82: 3272–3276.
Rogers KA, Kalnins VI. Comparison of the cytoskeleton in aortic endothelial cells in situ and in vitro. Lab Invest. 1983; 49: 650–654. [Order article via Infotrieve]
Ghribi O, Herman MM, Savory J. Lithium inhibits A beta-induced stress in endoplasmic reticulum of rabbit hippocampus but does not prevent oxidative damage and tau phosphorylation. J Neurosci Res. 2003; 71: 853–862. [Order article via Infotrieve]
Stambolic V, Ruel L, Woodgett JR. Lithium inhibits glycogen synthase kinase-3 activity and mimics wingless signalling in intact cells. Curr Biol. 1996; 6: 1664–1668. [Order article via Infotrieve]
Langille BL, Reidy MA, Kline RL. Injury and repair of endothelium at sites of flow disturbances near abdominal aortic coarctations in rabbits. Arteriosclerosis. 1986; 6: 146–154.
Malek AM, Izumo S. Mechanism of endothelial cell shape change and cytoskeletal remodeling in response to fluid shear stress. J Cell Sci. 1996; 109: 713–726.
Galbraith CG, Skalak R, Chien S. Shear stress induces spatial reorganization of the endothelial cell cytoskeleton. Cell Motil Cytoskeleton. 1998; 40: 317–330. [Order article via Infotrieve]
Palazzo AF, Eng CH, Schlaepfer DD, Marcantonio EE, Gundersen GG. Localized stabilization of microtubules by integrin- and FAK-facilitated Rho signaling. Science. 2004; 303: 836–839.
Zumbrunn J, Kinoshita K, Hyman AA, Nthke IS. Binding of the adenomatous polyposis coli protein to microtubules increases microtubule stability and is regulated by GSK3 phosphorylation. Curr Biol. 2001; 11: 44–49. [Order article via Infotrieve]
Bu W, Su LK. Regulation of microtubule assembly by human EB1 family proteins. Oncogene. 2001; 20: 3185–3192. [Order article via Infotrieve]
Etienne-Manneville S, Hall A. Cdc42 regulates GSK-3 and adenomatous polyposis coli to control cell polarity. Nature. 2003; 421: 753–756. [Order article via Infotrieve]
Kiosses WB, McKee NH, Kalnins VI. The distribution of centrosomes in endothelial cells of the rat aorta and inferior vena cava. Artery. 1997; 22: 251–265. [Order article via Infotrieve]
Kiosses WB, Kalnins VI. Age-related changes in the position of centrosomes in endothelial cells of the rabbit aorta. Exp Gerontol. 1993; 28: 69–75. [Order article via Infotrieve]
Gomes ER, Jani S, Gundersen GG. Nuclear movement regulated by Cdc42, MRCK, myosin, and actin flow establishes MTOC polarization in migrating cells. Cell. 2005; 121: 451–463. [Order article via Infotrieve]
Palazzo AF, Joseph HL, Chen Y-J, Dujardin DL, Alberts AS, Pfister KK, Vallee RB, Gundersen GG. Cdc42, dynein, and dynactin regulate MTOC reorientation independent of Rho-regulated microtubule stabilization. Curr Biol. 2001; 11: 1536–1541. [Order article via Infotrieve]
Kiosses WB, McKee NH, Kalnins VI. Evidence for the migration of rat aortic endothelial cells toward the heart. Arterioscler Thromb Vasc Biol. 1997; 17: 2891–2896.
Detera-Wadleigh SD. Lithium-related genetics of bipolar disorder. Ann Med. 2001; 33: 272–285. [Order article via Infotrieve]
Dogterom M, Kerssemakers JW, Romet-Lemonne G, Janson ME. Force generation by dynamic microtubules. Curr Opin Cell Biol. 2005; 17: 67–74. [Order article via Infotrieve]
Schwartz SM, Gajdusek CM, Reidy MA, Selden SCI, Haudenschild CC. Maintenance of integrity in aortic endothelium. Fed Proc. 1980; 39: 2618–2625. [Order article via Infotrieve]
Langille BL, Ojha M. Blood flow dynamics, atherosclerosis and bypass graft failure. Trends Cardiovasc Med. 1997; 7: 111–118.
Langille BL, Dajnowiec D, Gotlieb AI. The role of rheology in atherothrombotic coronary artery disease. In: Fuster V, Topol EJ, Nabel EG, eds. Atherothrombosis and Coronary Artery Disease. Philadelphia: Lippincott Williams and Wilkins; 2005.
the Department of Laboratory Medicine and Pathobiology (S.M., D.D., B.L.L.), University of Toronto, Canada
Toronto General Research Institute (S.M., D.D., F.X., M.Z., B.L.L.), University Health Network, Canada
Department of Anatomy and Cell Biology (M.R.J.), College of Medicine, University of Florida, Gainesville.
Abstract
Cultured vascular endothelium displays profound morphological adaptations to shear stress that include planar cell polarity (PCP) that is directed downstream. Endothelial cells in blood vessels are also polarized; however, the direction of polarity is vessel specific, and shear-independent mechanisms have been inferred. The regulation of endothelial PCP is therefore controversial. We report that the direction of PCP in blood vessels is age and vessel specific; nonetheless, it is caused by shear-related regulation of glycogen synthase kinase-3 (GSK-3), a profound regulator of endothelial microtubule stability. When GSK-3 is inhibited, PCP reverses direction. Endothelium is the only cell type studied to date that can reverse direction of polarity. Tight regulation of GSK-3, microtubule dynamics, and cell polarity was also required for the striking morphological responses of endothelium to shear stress (cell elongation and orientation with shear). Finally, the cytoskeletal polarity displayed in blood vessels is associated with polarized (shear-directed) cell mitoses that have important effects on endothelial repair. Vascular endothelium therefore displays a novel mode of mechanosensitive PCP that represents the first example of a single cell type that can reverse direction of polarity.
Key Words: shear stress endothelium microtubules planar cell polarity
Introduction
Vascular endothelial cells display profound morphological adaptations to shear stress that include cell elongation in the shear direction that is driven by a novel reorganization of actin in stress fibers,1 redistribution of focal adhesion complexes,1,2 and partial disassembly and reassembly of cell-cell junctional complexes.3 Imposition of shear stress on cultured endothelium also induces planar cell polarity (PCP) in the downstream direction.4 PCP occurs when cell organelles, cytoskeleton, and/or adhesion complexes exhibit unidirectional organization along an axis that lies in the plane of a cell monolayer. This process was manifest in sheared endothelium in vitro as redistribution of microtubules and the microtubule-organizing center (MTOC) (from which microtubules emanate) to the downstream side of the cell nuclei, a hallmark of PCP.5
PCP has been of great interest to cell biologists because it provides directionality to cell functions. PCP in vertebrate cells has been studied primarily during directional migration at wound margins where it is initiated by integrin activation and multiple downstream pathways that involve the Rho GTPase, Cdc42, adenomatous polyposis coli (APC), glycogen synthase kinase-3 (GSK-3), and mammalian Diaphanous-1 (mDia1).6,7 Polarization of the microtubule system is important because motor proteins (kinesins and dyneins) can deliver cargoes (eg, cell membrane and protein) along microtubules to and from the cell anterior and posterior and because microtubules regulate the reorganization of cell adhesion complexes.8 This process is enhanced by microtubule stabilization through posttranslational modifications to tubulin that are coinduced with morphological polarity.9 The significance of these processes to sheared cultures of endothelium is uncertain because polarity is retained at 24 hours,4 when the cells are no longer motile.1
Endothelial cells display PCP in intact blood vessels10 as well as in vitro; however, the origins of this polarity are controversial because microtubule polarity in vivo varies among blood vessels.10,11 Specifically, PCP is often directed upstream in arteries and downstream in veins, a finding that has led to the inference that it is independent of shear stress.
We have investigated the role of microtubules in shear-dependent motility and PCP of endothelium both in vitro and in blood vessels. In both cases, PCP is shear dependent, and it is tightly regulated by GSK-3. Indeed, manipulation of GSK-3 could induce polarity of these cells to reverse direction, a capacity not displayed by any other cell type studied to date; furthermore, regulation of GSK-3 had dramatic effects on microtubule stability. Primary effects of GSK-3 were partly attributable to its influence on stability of microtubules, and these effects were critical to endothelial adaptations to shear stress because the cells failed to elongate and align with shear when GSK-3 activity or microtubule dynamics were suppressed with exogenous inhibitors or mutant forms of the kinase. GSK-3–related regulation of microtubules is therefore critical to morphological adaptations of endothelium to shear stress.
Materials and Methods
An expanded Materials and Methods section is available in the online data supplement at http://circres.ahajournals.org.
Imposition of Physiological Levels of Shear Stress
Porcine aortic endothelial cells were grown to 2 days postconfluence on glass slides or cover slips and then subjected to shear stress of 15 dyne/cm2 in parallel plate flow chambers. When indicated, GSK-3 activity was inhibited by treatment with LiCl or SB415286. Nocodazole (100 mmol/L) or taxol (5 ng/mL) were used to disrupt or hyperstabilize microtubules, respectively.
Immunofluorescence
Immunofluorescence staining of cells was viewed by laser scanning confocal microscope. For studies of -tubulin localization, nuclei were counterstained with propidium iodide or Sytox green.
Analysis of Endothelial Cell Alignment With Shear Stress
Endothelial cell elongation in the shear direction was assessed by determining cell length:width ratios using a novel, simple, and accurate methodology that precludes inaccuracies inherent in "shape factor" determinations that rely on cell perimeter-area relations (see online data supplement).
Microtubule-Organizing Center Localization
MTOCs were grouped (see supplemental Figure I) as upstream of nucleus (scored 1), upstream of midpoint of nucleus (scored 2; includes cells scored 1), and downstream of the midpoint of the nucleus (scored 3), which included cells downstream of the nucleus (scored 4).
Polarization of Acetylated Microtubules
Cultures were immunostained against acetylated tubulin and -catenin, and nuclei were counterstained with propidium iodide. An observer, blinded to whether shear direction was left-to-right or right-to-left, assessed the position of the majority of stable microtubules in each cell relative to the nucleus. Decoded data were presented as a percentage of cells with stable microtubules predominantly upstream versus downstream of the nucleus.
Immunoblotting
Ten micrograms of protein from cell lysates were subjected to polyacrylamide gel electrophoresis, and blots were probed with antibodies as indicated. Labeled secondary antibodies were visualized using enhanced chemiluminescence.
Microinjections
In 3 experiments, nuclei of at least 30 endothelial cells in postconfluent monolayers were microinjected with constitutively active GSK-3 (GSK3S9A). Coinjection of rhodamine-labeled dextran was used to localize microinjected cells. After 3 hours, flow of culture medium was initiated and cells were then examined after 24 hours of shear stress.
In Vivo Assessment of Cell Polarity
Blood vessels from adult and 5-week-old rabbits were immunostained en face. Some rabbits were treated with 7 mmol/L lithium carbonate in drinking water to inhibit GSK-3 for 48 hours before euthanasia.12,13
To examine the effects of reversing flow/shear direction in vivo, rabbits were anesthetized and the midabdominal aorta was narrowed by 50% with 5-0 suture.14 These stenoses produce a downstream vortex that reverses flow near the endothelial surface for up to 10 vessel diameters downstream of the stenosis (Figure 6A). One week later, the rabbits were killed by barbiturate overdose and vessels segments in the vortex region were stained en face.
To test for polarized mitosis, 3- and 5-week-old rabbits were given intramuscular injections of bromodeoxyuridine (BrdUrd). En face BrdUrd immunostaining was performed after 24 hours. At this time, BrdUrd-labeled cells had completed mitosis, so that doublets of daughter cells were detected and the relative positions of the labeled nuclei indicated the direction of mitosis. Alternatively, cultured cells were pulse labeled with BrdUrd immediately before shearing.
These experiments were approved by the Animal Care Committee of the University Health Network and were conducted in accord with the standards of the Canadian Council on Animal Care.
Statistics
Sample sizes are presented in figure legends. Effects of drugs (nocodazole, taxol, SB415286) on cell length-width ratios were assessed using a Dunnett’s test. Differences in MTOC position after shear stress, and position relative to the nucleus of acetylated microtubules, were determined using Student’s t tests. P<0.05 was taken as statistically significant.
Results
Microtubule Dynamics Is Required for Shear-Induced Changes in Endothelial Cell Shape
Postconfluent endothelial cells elongated in the direction of shear stress within 24 hours,1 with mean cell length in the shear direction becoming 2.5 to 3 times cell width (Figure 1A, 1B, and 1E). Microtubules in these cells emanated from the MTOC and projected throughout the cytoplasm, except that the perijunctional region was frequently spared (Figure 1A and 1B). These filaments became highly oriented in the direction of shear stress. Microtubules are required for cell elongation because treatment with nocodazole prevents cell elongation.15 This effect was observed even with nocodazole doses that fail to completely disrupt microtubules (Figure 1C and 1E). Furthermore, cells did not elongate when dynamic instability of microtubules was suppressed using the microtubule-stabilizing drug taxol. With either drug, cell "length" was not significantly different from cell "width" after 24 hours of shear stress (Figure 1D and 1E). These findings indicate an obligate role for dynamic instability of microtubules in the shear stress-induced motility of endothelial cells.
Shear Stress-Induced Planar Cell Polarity Is Accompanied by Profound Stabilization of Microtubules
Recently, Tzima et al found that the microtubule system (MTOC) of endothelium subjected to shear stress displayed Cdc42-dependent polarization to the downstream side of the nucleus after 24 hours of shear stress,4 although these findings contradicted a previous report of transient and upstream polarization at 6 hours of shear.16 We failed to detect polarization of microtubules or MTOCs after 6 hours of comparable levels of shear stress (not shown); however, we confirmed a high degree of downstream polarization after 24 hours (Figure 2A, 2B, and 2D).
Microtubules can be stabilized by posttranslational modifications to tubulin.5 Acetylation of tubulin coincides with stabilization,5 and it can be detected with an acetylation-specific antibody.17 Immunoblots detected low levels of acetylated tubulin in postconfluent monolayers of endothelium that were not exposed to shear stress; however, the acetylated form was dramatically upregulated by 1 hour of shear stress, and further increases were seen by 24 hours (Figure 3A). When the same antibody was used for immunofluorescence studies, only a sparse population of perinuclear microtubules was labeled in static cultures (Figure 3B), whereas extensive labeling of much of the microtubule population was detected after 24 hours of shear (Figure 3C). Stabilized microtubules were found predominantly downstream of the cell nucleus in sheared cells (Figure 3D). These findings indicate major modifications to microtubules under the influence of shear stress. Importantly, the cells have completed shape change by 24 hours,1 so polarity and microtubule stability are not associated with directional motility at this time.
GSK-3 Regulates Stabilization and Polarization of Microtubules by Shear Stress
GSK-3 regulates planar cell polarity in migrating cells; furthermore, this kinase suppresses microtubule stability by phosphorylating multiple microtubule stabilizing proteins.18,19 We therefore assessed the role of GSK-3 in endothelial cell responses to shear stress. Western blots failed to detect the inactive (phosphorylated) form of GSK-3 (p-GSK-3) in static cultures (Figure 4A). In contrast, p-GSK-3 was readily detected in sheared cells over a time course that matched that of tubulin acetylation, except that p-GSK-3 waned at 24 hours, when tubulin acetylation was maximal (Figures 3A and 4A). Total GSK levels were insensitive to shear stress in these experiments (not shown).
To test whether GSK-3 regulates morphological responses to shear stress, we exposed cells to shear with and without treatment with the GSK-3 inhibitors LiCl and SB415286. Cell elongation in response to shear stress was totally blocked by GSK-3 inhibition (Figure 4B through 4D). Our findings that GSK-3 is suppressed by shear stress, whereas total inhibition using exogenous inhibitors of GSK-3 blocks shear-induced changes in cell morphology, indicate that tight control over this kinase is critical to endothelial responses to shear. Accordingly, cells failed to elongate after nuclear microinjection of vectors expressing constitutively active GSK3S9A (Figure 4E and 4F). Polarity of MTOCs also was compromised, ie, 57% of MTOCs were downstream of the nuclear midpoint of microinjected cells, compared with more than 90% of uninjected cells. As expected, we also observed a sparser distribution of acetylated microtubules in cells expressing GSK3S9A. Microinjection of a kinase dead construct (GSK3K85A) was without effect, consistent with reports that this mutation does not yield dominant-negative function (J.A. Woodgett, personal communication, 2005).
Given the importance of microtubule dynamics in control of cell shape and the capacity of GSK-3 to regulate microtubule stabilization in other cells, we examined the capacity of GSK-3 to influence microtubule stability in unsheared endothelium. Treatment with LiCl (or SB415286; supplemental Figure II) to inhibit GSK-3 caused the distribution of stable microtubules to convert from a sparse perinuclear expression (Figure 3B) to an elaborate network (Figure 5A). This stabilization resulted in extensive elongation of microtubules, so that the total microtubule population circled the cell periphery in a toroid pattern (Figure 5B).
We asked whether microtubule polarity, as indicated by MTOC position, was affected by inhibition of GSK-3. Based on studies of migrating cells, we anticipated randomization of MTOC position after shearing cells in the presence of an exogenous inhibitor.20 Intriguingly, polarity was reversed by this treatment (Figure 2C and 2D), ie, the MTOCs redistributed to the upstream side of the nucleus after initiation of shear. These findings suggest that shear stress induces competing, GSK-3–dependent and GSK-3–independent, modes of polarity control in endothelium.
Endothelial Cell Polarity is Age and Vessel Specific In Vivo
MTOCs in the thoracic aorta and vena cavae of rats and pigs localize to the heart side of the nucleus (upstream in arteries, downstream in veins).11,21 The same is true for carotid arteries and vena cavae of rabbits, although polarity declines with age in the latter vessel, and it is not evident in thoracic aortas in this species.11,22 These findings led to the inference that endothelial polarity in vivo is cardiocentric and independent of blood flow. Although we confirmed that MTOCs were downstream of nuclei in the vena cava of the adult rabbit, upstream in the adult carotid artery and unpolarized in the adult thoracic aorta (data not shown), we also found that endothelium of the abdominal aorta reversed polarity with age; MTOCs were downstream of nuclei in 5-week-old rabbits and upstream in abdominal aortas of mature rabbits (Figure 6A and 6B and 6C, solid bars). MTOCs were distributed predominantly downstream of the nucleus in the vena cavae of young (Figure 6D) and mature (not shown) rabbits. These findings indicate that polarity of endothelium in vivo is vessel and age specific, and it is not exclusively cardiocentric.
Endothelial Cell Mitoses Also Polarize in the Shear Direction
Polarization of cell morphology and the microtubule system may prime cells for polarized mitosis; therefore, we also tested whether spontaneous mitoses of endothelial cells in immature rabbits were consistently unidirectional. When S-phase cells were labeled in vivo with BrdUrd and endothelial surfaces of arteries were prepared for imaging 24 hours later, the relative positions of the 2 daughter cells was consistently along the vessel axis. Figure 7A illustrates this feature in an in vivo preparation of the carotid artery of a 3-week-old rabbit; however, identical results were obtained in aortas and carotid arteries of 5-week-old rabbits. This response was genuinely flow derived because nonaxial mitoses were observed in the blind-ended stumps of carotid arteries that had been ligated (Figure 7B); furthermore, mitosis were flow aligned in vitro as well as in vivo. Indeed, even though in vitro flow was initiated (immediately) after exposure to BrdUrd, cell division was aligned with shear (Figure 7C through 7E). This alignment included cells that, after 24 hours of shear stress, were in telophase (Figure 7C), cytokinesis (Figure 7D), or were postmitotic (Figure 7E).
Endothelial Cell Polarity Is Shear Dependent In Vivo
Our finding that, under different conditions, shear stress can induce either upstream or downstream polarization of MTOCs in cell cultures led us to examine its influence on MTOC polarity in intact blood vessels. We surgically produced 50% narrowing (stenosis) of the midabdominal aorta of both 5-week-old and mature rabbits. This procedure causes a vortex to form downstream of the stenosis, and we confirmed previously that blood flow near the vessel wall, and therefore wall shear stress, was directed toward the heart for many vessel diameters distal to the stenosis14 (Figure 8A). Flow reversal at the vortex site eliminated the downstream polarity of the MTOCs in immature abdominal aortas and the upstream polarity in mature vessels (Figure 8B and 8C). A slight reversal of these polarities was not statistically significant. These findings indicate that polarization of endothelial MTOCs in vivo is shear dependent.
Endothelial Cell Polarity Is Regulated by GSK-3 In Vivo
The flow dependence of endothelial cell polarity both in vivo and in vitro, and the capacity of GSK-3 inhibition to reverse shear-induced polarity in vitro, prompted us to assess the effects of GSK-3 inhibition in vivo. When lithium (LiCO3) was administered to rabbits for 48 hours before examination of the position of the MTOCs in blood vessels, both the upstream polarity of abdominal aortic endothelium of mature animals and the downstream polarity of endothelium of the abdominal aorta and vena cava of immature rabbits (Figure 6B through 6D) were suppressed. Furthermore, immunofluorescence staining for acetylated tubulin was suppressed by obstructing flow by carotid ligation for one week and partially restored in rabbits treated with LiCO3 in drinking water (supplemental Figure III). These findings confirmed that GSK-3 is an important regulator of shear-induced endothelial polarity both in cell culture and in blood vessels.
Discussion
This study has shown that shear stress induces tight control over microtubule organization, stability, and function and that this control is absolutely required for the dramatic morphological responses of endothelium to shear stress. Accordingly, drugs that induce either hyperstabilization of microtubules (taxol) or microtubule disassembly (nocodazole), or expression of a mutant (constitutively active) form of an important regulator of microtubules (GSK-3), blocked cell elongation and orientation under the influence of shear.
We confirmed that MTOCs displayed downstream polarity when endothelial cells were subjected to shear stress,4 and we showed that shear induces marked stabilization of microtubules as evidenced by tubulin acetylation. Although the effects were profound, stabilization must be tightly regulated during cell polarization, given our findings that the microtubule stabilizing agent, taxol, blocked cell morphological responses to shear stress. These findings demonstrate that shear stress induces primary manifestations of PCP.
Because shear had major effects on microtubules that proved important to cell morphological responses, we examined regulation of microtubule modifications. When cells migrate at wound edges, integrins are activated at the cell anterior, as new cell-substrate interactions occur, and downstream activation of Cdc42 leads to the anterior assembly of a complex that includes Cdc42 and the scaffold protein, Par6, atypical protein kinase C- (aPKC), and GSK-3. An outcome of this assembly is inhibitory phosphorylation of GSK-3, which is pivotal to establishment of PCP.20 GSK-3 is multifunctional, but its roles in PCP appear related to its capacity to dissociate APC from microtubule + ends.20 How inhibition of GSK-3, and the resulting association of APC with the peripheral tips of microtubules, effects microtubule redistribution is unknown, although it may promote microtubule binding to proteins/structures at the cell anterior that prevents them by being swept backward by retrograde flow of actin.23 Alternatively, motor proteins (dyneins) that are captured at the cell anterior may ratchet microtubules anteriorly.24 Although the other major manifestation of PCP, microtubule stabilization, appears to be controlled independently of MTOC redistribution,24 inhibition of GSK-3 again is involved. APC promotes the association with microtubules of stabilizing proteins such as EB1,19 and we found that phosphorylation of GSK-3 had profound effects on microtubule stability in endothelial cells.
Based on the pivotal role of GSK-3 in PCP of migrating cells, we examined its role in shear-induced responses. Shear-induced phosphorylation of GSK-3 coincided with acetylation of tubulin in microtubules, and we infer a causal link. At late time points, GSK-3 phosphorylation waned, whereas tubulin acetylation was sustained; therefore, other factors may participate in maintenance of shear-induced microtubule stability. Nonetheless, a role for inhibition of GSK-3 was strongly supported by our finding that exogenous inhibitors greatly enhanced tubulin acetylation in unsheared cells and so strongly stabilized these structures that they formed extremely long filaments that encircled the cell periphery in a toroid pattern. Again, tight regulation, rather than simple suppression, of GSK-3 activity was required for cell morphological responses because the cells failed to elongate when the activity of the kinase was blocked using exogenous inhibitors. Regulation of intracellular gradients, as well as absolute activity, of GSK-3 is probably required.20
Surprisingly, exogenous inhibitors of GSK-3 activity did not suppress MTOC polarity, as occurs in migrating cells20; they reversed it so that cells polarized in the upstream direction. Recent work indicates that at least some epithelial cells display posterior PCP when migrating; however, this is the first example of a single cell type that can reverse direction of polarity. Endothelium may therefore provide an informative model for studying directional polarity control. We infer from this observation that endothelial cells can undergo GSK-3–independent, shear-induced reverse PCP that is unmasked by the inhibitors.
Reversal of shear-induced microtubule polarity by inhibition of GSK-3 was particularly interesting given that, in some species, endothelial MTOCs are upstream of nuclei in arteries and downstream in veins.10,11,21 This previous observation was interpreted as indicating flow independence of endothelial cell polarity and novel alternatives were proposed, including sheet-like migration of peripheral vascular endothelium toward the heart.25 We found, however, that cardiocentric polarity was not ubiquitous in vivo; furthermore, polarity in the abdominal aorta of rabbits reversed with age. This finding, plus our observation that shear can induce upstream or downstream polarity in vitro, depending on p-GSK-3 levels, prompted us to directly assess the influence of shear on endothelial polarity in vivo. Experimental reversal of blood flow in the vortex that forms downstream from stenoses in vivo eliminated endothelial polarity, regardless of whether polarity was directed upstream or downstream in the unmanipulated artery. A modest reversal of polarity was not statistically significant, which may reflect either the less pronounced polarity that is observed in vivo or the precise nature of shear stress and its temporal variations downstream from stenoses. Nonetheless, these findings demonstrate that endothelial cell polarity in blood vessels is shear dependent.
Endothelial cell polarity was GSK-3 dependent in blood vessels, as it was in cell culture. This hypothesis was testable because lithium administration, to inhibit GSK-3 activity, is frequently used to treat bipolar disorder in humans.26 Lithium given in drinking water to rabbits suppressed or reversed polarity in both arteries and veins. Failure to always observe reversal of downstream polarity, as was seen in cell cultures, may indicate that circulating lithium levels fluctuated substantially when the rabbits were given lithium-treated water ad libidum, so that GSK-3 may have transiently but repeatedly escaped inhibition.
How microtubules influence morphological adaptations to shear stress is unclear. While the cell is changing shape, polarized microtubules might deliver cell membrane, protein and other cargoes to the cell poles (downstream pole in vitro) as well as regulate disassembly of substrate adhesion complexes. However, this simple interpretation is not readily reconciled with our previous finding that the actin assembly that drives shear-induced cell elongation proceeds with equal frequency in the upstream and downstream directions.1 Furthermore, polarization and stabilization of microtubules persisted at 24 hours of shear, after shape change had reached completion.1 The latter finding suggests that microtubule functions in planar cell polarity go beyond contributing to motility. Additional functions are unknown, although we have argued that the highly oriented distribution of the cytoskeleton and adhesion complexes in cells elongated by shear represents a low entropy state and that the organization of these dynamic structures must undergo constant renovation to prevent drift toward random orientation. The persistent interactions of highly stabilized and oriented microtubules with adhesion complexes might contribute to continuous maintenance of these structures, as might the steady state actin assembly/turnover that we described previously.1
The high degree of PCP displayed by endothelial cell morphology and cytoskeleton led to the hypothesis that mitosis would exhibit directional bias as well. Accordingly, spontaneous mitoses in the arterial endothelium of young, rapidly growing rabbits ubiquitously displayed mitosis along the shear vector. Shear dependence was confirmed when polarized mitoses were observed in sheared endothelium in vitro and when disruption of blood flow by carotid ligation caused orientation of mitoses to randomize in vivo. These findings indicate that mechanical forces may be potent regulators of directional tissue growth and repair. Possibly, mitotic spindle formation during polarized mitosis is associated with the biased distribution of microtubules that is induced by shear stress; alternatively, there may be an interplay between forces generated by mitotic microtubules and associated motor proteins at the cell cortex that align mitosis27 and externally applied forces that are transmitted to these structures.
These adaptations of endothelium to shear may have implications for important vascular pathologies. For example, endothelial denudation when arteries are traumatized by angioplasty or stent implantation contributes to restenosis; consequently, rapid repair of endothelium is important to arterial integrity. Shear-induced endothelial cell polarity directs cell elongation and proliferation preferentially in the long axis of the vessel, the effects of which are vividly demonstrated by examining outgrowth of endothelium from branches of denuded arteries, which proceeds much more rapidly along the vessel axis than circumferentially.28 If substantial lengths of vessel are denuded, a bias in cell motility and mitosis in the lengthwise direction will accelerate repair.
Furthermore, continual microtubule-dependent remodeling of the endothelial monolayer in response to changing blood flow conditions, and associated disruption of cell-cell junctions,3 may contribute to high endothelial permeability in regions of complex blood flow at arterial branch sites and thereby contribute to the predilection of these sites for atherogenesis.29 At these locations, magnitude and direction of shear stress changes are caused by exercise, circadian rhythms in cardiac output, or longer-term changes in cardiovascular function that are associated with exercise regimens, reproductive cycles, pregnancy, and disease states.30
Acknowledgments
This work was supported by operating grant 15044 from the Canadian Institutes for Health Research and a studentship (to D.D) from Canadian Institutes of Health Research and the Canadian Hypertension Society. We thank Donna Johnston for technical assistance and James Woodgett for providing mutant GSK-3 constructs.
Footnotes
Original received July 22, 2005; resubmission received February 1, 2006; accepted February 24, 2006.
References
Noria S, Xu F, McCue S, Jones M, Gotlieb AI, Langille BL. Assembly and reorientation of stress fibers drives morphologic changes to endothelial cells exposed to shear stress. Am J Pathol. 2004; 164: 1211–1223.
Girard PR, Nerem RM. Shear stress modulates endothelial cell morphology and F-actin organization through the regulation of focal adhesion-associated proteins. J Cell Physiol. 1995; 163: 179–193. [Order article via Infotrieve]
Noria S, Cowan DB, Gotlieb AI, Langille BL. Transient and steady-state effects of shear stress on endothelial cell adherens junctions. Circ Res. 1999; 85: 504–514.
Tzima E, Kiosses WB, Del Pozo MA, Schwartz MA. Localized cdc42 activation, detected using a novel assay, mediates microtubule organizing center positioning in endothelial cells in response to fluid shear stress. J Biol Chem. 2003; 278: 31020–31023.
Gundersen GG, Gomes ER, Wen Y. Cortical control of microtubule stability and polarization. Curr Opin Cell Biol. 2004; 16: 106–112. [Order article via Infotrieve]
Raftopoulou M, Hall A. Cell migration: Rho GTPases lead the way. Dev Biol. 2004; 265: 23–32. [Order article via Infotrieve]
Etienne-Manneville S, Hall A. Cell polarity: Par6, aPKC and cytoskeletal crosstalk. Curr Opin Cell Biol. 2003; 15: 67–72. [Order article via Infotrieve]
Kaverina I, Krylyshkina O, Small JV. Microtubule targeting of substrate contacts promotes their relaxation and dissociation. J Cell Biol. 1999; 146: 1033–1043.
Lin SX, Gundersen GG, Maxfield FR. Export from pericentriolar endocytic recycling compartment to cell surface depends on stable, detyrosinated (Glu) microtubules and kinesin. Mol Biol Cell. 2002; 13: 96–109. [Order article via Infotrieve]
Rogers KA, McKee NH, Kalnins VI. Preferential orientation of centrioles towards the heart in endothelial cells of major blood vessels is reestablished after reversal of a segment. Proc Nat Acad Sci U S A. 1985; 82: 3272–3276.
Rogers KA, Kalnins VI. Comparison of the cytoskeleton in aortic endothelial cells in situ and in vitro. Lab Invest. 1983; 49: 650–654. [Order article via Infotrieve]
Ghribi O, Herman MM, Savory J. Lithium inhibits A beta-induced stress in endoplasmic reticulum of rabbit hippocampus but does not prevent oxidative damage and tau phosphorylation. J Neurosci Res. 2003; 71: 853–862. [Order article via Infotrieve]
Stambolic V, Ruel L, Woodgett JR. Lithium inhibits glycogen synthase kinase-3 activity and mimics wingless signalling in intact cells. Curr Biol. 1996; 6: 1664–1668. [Order article via Infotrieve]
Langille BL, Reidy MA, Kline RL. Injury and repair of endothelium at sites of flow disturbances near abdominal aortic coarctations in rabbits. Arteriosclerosis. 1986; 6: 146–154.
Malek AM, Izumo S. Mechanism of endothelial cell shape change and cytoskeletal remodeling in response to fluid shear stress. J Cell Sci. 1996; 109: 713–726.
Galbraith CG, Skalak R, Chien S. Shear stress induces spatial reorganization of the endothelial cell cytoskeleton. Cell Motil Cytoskeleton. 1998; 40: 317–330. [Order article via Infotrieve]
Palazzo AF, Eng CH, Schlaepfer DD, Marcantonio EE, Gundersen GG. Localized stabilization of microtubules by integrin- and FAK-facilitated Rho signaling. Science. 2004; 303: 836–839.
Zumbrunn J, Kinoshita K, Hyman AA, Nthke IS. Binding of the adenomatous polyposis coli protein to microtubules increases microtubule stability and is regulated by GSK3 phosphorylation. Curr Biol. 2001; 11: 44–49. [Order article via Infotrieve]
Bu W, Su LK. Regulation of microtubule assembly by human EB1 family proteins. Oncogene. 2001; 20: 3185–3192. [Order article via Infotrieve]
Etienne-Manneville S, Hall A. Cdc42 regulates GSK-3 and adenomatous polyposis coli to control cell polarity. Nature. 2003; 421: 753–756. [Order article via Infotrieve]
Kiosses WB, McKee NH, Kalnins VI. The distribution of centrosomes in endothelial cells of the rat aorta and inferior vena cava. Artery. 1997; 22: 251–265. [Order article via Infotrieve]
Kiosses WB, Kalnins VI. Age-related changes in the position of centrosomes in endothelial cells of the rabbit aorta. Exp Gerontol. 1993; 28: 69–75. [Order article via Infotrieve]
Gomes ER, Jani S, Gundersen GG. Nuclear movement regulated by Cdc42, MRCK, myosin, and actin flow establishes MTOC polarization in migrating cells. Cell. 2005; 121: 451–463. [Order article via Infotrieve]
Palazzo AF, Joseph HL, Chen Y-J, Dujardin DL, Alberts AS, Pfister KK, Vallee RB, Gundersen GG. Cdc42, dynein, and dynactin regulate MTOC reorientation independent of Rho-regulated microtubule stabilization. Curr Biol. 2001; 11: 1536–1541. [Order article via Infotrieve]
Kiosses WB, McKee NH, Kalnins VI. Evidence for the migration of rat aortic endothelial cells toward the heart. Arterioscler Thromb Vasc Biol. 1997; 17: 2891–2896.
Detera-Wadleigh SD. Lithium-related genetics of bipolar disorder. Ann Med. 2001; 33: 272–285. [Order article via Infotrieve]
Dogterom M, Kerssemakers JW, Romet-Lemonne G, Janson ME. Force generation by dynamic microtubules. Curr Opin Cell Biol. 2005; 17: 67–74. [Order article via Infotrieve]
Schwartz SM, Gajdusek CM, Reidy MA, Selden SCI, Haudenschild CC. Maintenance of integrity in aortic endothelium. Fed Proc. 1980; 39: 2618–2625. [Order article via Infotrieve]
Langille BL, Ojha M. Blood flow dynamics, atherosclerosis and bypass graft failure. Trends Cardiovasc Med. 1997; 7: 111–118.
Langille BL, Dajnowiec D, Gotlieb AI. The role of rheology in atherothrombotic coronary artery disease. In: Fuster V, Topol EJ, Nabel EG, eds. Atherothrombosis and Coronary Artery Disease. Philadelphia: Lippincott Williams and Wilkins; 2005.
您现在查看是摘要介绍页，详见ORG附件。