Force-Induced Polarized Mitosis of Endothelial and Smooth Muscle Cells in Arterial Remodeling
Arteries display highly directional growth and remodeling that are specific to increases in the mechanical loads imposed on them by blood pressure, blood flow, and lengthwise tensile forces that are transmitted from the tissues to which they are attached. This study examined the effect of mechanical forces on the direction in which mitosis delivers daughter cells, as a mechanism for directional growth. Lateral forces were imposed on surface integrins of cultured endothelial cells by seeding the cells with arginine-glycine-aspartate peptide–coated magnetic microspheres and applying a magnetic field. Video images revealed that the mitotic axis of dividing cells became highly biased in the direction of applied force. Distribution of cortactin, which participates in polarized mitoses driven by other stimuli, was highly sensitive to mechanical loading and interfering with cortactin function arrested cell growth. Smooth muscle cell mitoses also proved to be sensitive to mechanical force: when lengthwise force imposed on rabbit carotid arteries was altered by excision of a vessel segment and reanastomosis of the cut ends, direction of mitosis was dramatically altered. These findings indicate that influences of mechanical force can modulate the manner in which mitosis of vascular cells contributes to reorganization of arterial wall tissue.
Blood vessels are constantly subjected to mechanical forces that greatly influence their development, function, and pathology. For example, chronic increases in shear stress selectively gives rise to growth of arterial diameter,1,2 whereas increased pressure and lengthwise stretch cause wall thickening3 and axial growth,4 respectively. Smooth muscle cells, the only cells resident in the media, must achieve much of this remodeling, but it is unclear how these cells selectively deliver extracellular matrix or daughter cells after mitosis in different directions in response to different mechanical forces. The endothelial monolayer also remodels in response to hemodynamic forces, with shear stress inducing elongation, cytoskeletal alignment, and planar cell polarity of these cells. Again, growth and remodeling are affected, because regrowth after endothelial denudation proceeds much more rapidly along the shear axis then perpendicular to it.5
We hypothesized that mitosis of vascular cells induced by mechanical forces is polarized so that delivery of daughter cells occurs in a direction that is specific for the force that is applied to the tissue. We initially focused on endothelial cells because of profound morphological responses to shear stress that they display. Most notably, shear stress induces planar cell polarity of endothelium, a cell-polarizing response in which the microtubule system becomes highly directionalized.6,7 In particular, the polarized positioning of the endothelial microtubule organizing center and associated centrosome may play a pivotal role in aligning the spindle axis during mitosis. In addition, endothelial cells elongate in the shear direction, and cell shape alone can influence direction of mitosis.8 Accordingly, we found that endothelial cells exposed to shear stress underwent mitosis that was consistently aligned with shear stress both in vivo and in vitro; furthermore, alignment of cells was lost when blood flow was obstructed in vivo.7
In the current study, we tested whether imposing lateral force on cell surface–associated integrins, which are thought to be important in force transduction, is sufficient to polarize mitosis and/or whether cell elongation in the shear direction is required. We also asked whether the observations that we made with endothelium apply to mechanically loaded smooth muscle cells.
Imposition of Mechanical Load on Endothelial Cells via Magnetic Microspheres
To explore the role of local, ligand-specific loading of cell surface structures on mitosis, ferromagnetic microspheres coated with integrin ligand were seeded onto cell surfaces, allowed to stabilize for 1 hour, and exposed to magnetic fields.9 Endothelial cells at 90% to 100% confluence were cultured in sterile tissue culture dishes (Falcon, VWR) in Medium199 plus 10% FBS (Invitrogen), 1% amphotericin B (Fungizone), and 1% penicillin/streptomycin. The cells were seeded with 2 to 5 μm iron (II and III) oxide microspheres (Sigma) that were coated with cyclic arginine-glycine-aspartate (RGD) peptides (Peptides International), which bind to endothelial α5β1 and αvβ3 integrins,10 at a seeding density of 2 to 4 microspheres per cell. Microspheres coated with RGD peptides mimic ligation to vitronectin or fibronectin but not collagen. Four 3×12×12-mm neodymium magnets (Jobmaster Magnets Canada Inc) were stacked and placed within 1.5 to 2.0 cm of the field to be observed so that they exerted a mean lateral force of 1.13×10−2 pN per bead on the microspheres (see below). In control experiments, the magnets were omitted. Twenty-four–hour phase contrast videos of the cultures were made using a Nikon TE300 inverted optical microscope using a ×10 objective, with images captured at 5-minute intervals. For all of the conditions, 3 independent experiments were performed, with an average of 25 to 55 mitoses per experiment undergoing analysis. Analysis consisted of measuring the angle between force direction and nuclear centroids (geometric centers) of daughter cells at the end of cytokinesis using Simple PCI software.
Mean force per bead (F) was determined by capturing video images of beads suspended in culture medium at 37°C as they were drawn toward permanent magnets positioned as for experiments. Bead radius (R) and velocity (v) were measured and combined with viscosity of the culture medium (μ=0.0069 P at 37°C) to determine force on the beads using Stokes Law (F=6 πμRv).
Analysis of Endothelial Cell Alignment With Force
Cell shape can influence direction of mitosis,8 and endothelial cells coalign with shear stress7; therefore, we tested whether the cells also aligned with the direction of force imposed on magnetic microspheres bound to cell surface integrins, as described previously.7 Briefly, a rectangular grid, with lines parallel and perpendicular to axis of the magnetic force, was placed over the final frame (24-hour time point) of each movie, and the number of cells per millimeter that were encountered by lines parallel to the magnetic force axis was determined. The inverse of this number (millimeters per cell) provided the mean cell length parallel to the direction of applied force (L//). Mean cell length in the direction perpendicular to the applied force (L⊥) was determined in a similar manner. The ratio of these dimensions (L///L⊥) provided a measure of elongation in the direction of force. This simple method for assessing directional elongation of cells in monolayers avoids limitations of “shape factors” that are based on cell area/perimeter relations, limitations that are related to the indeterminacy of cell perimeter because of its fractal nature.
Role of Cortactin in Polarized Mitosis of Endothelial Cells
Cortactin can participate in orientation of mitotic spindles; therefore, endothelial cells were transfected by electroporation (Amaxa) with 10 μg of wild-type cortactin (control) and 10 μg of nonphosphorylatable mutant cortactin11,12 at 10×105 cell density. RGD-coated microspheres were again used to impose force and were subjected to 24 hours of magnetic force (as described above).
Immunofluorescence Detection of Proteins
Endothelial cells were fixed with 4% paraformaldehyde in 0.1 mol/L of PBS (2 mmol/L of NaH2PO4, 8 mmol/L of Na2HPO4, 150 mmol/L of NaCl, 0.1 mmol/L of CaCl2, and 0.1 mmol/L of MgCl2 [pH 7.4]) for 20 minutes at room temperature, washed, and then permeabilized with 0.2% Triton X-100. Cells were then incubated with primary antibodies to cortactin (Chemicon; 1:100) and α-tubulin (Sigma; 1:50) for 1 hour, washed with PBS, and incubated with fluorescently conjugated secondary antibodies for 30 minutes. Secondary antibodies included Alexa 568 goat anti-rabbit (Molecular Probes/Invitrogen; 1:100) and fluorescein isothiocyanate–conjugated donkey anti-mouse (Jackson ImmunoResearch Laboratories Inc; 1:100). Confocal images of samples were examined using a Olympus FluoView 1000 (IX 81) inverted microscope (×60 oil immersion objective, numerical aperture=1.4; Olympus America Inc).
Elevation of Axial Strain In Vivo
Animal experimental protocols were approved by the animal care committee of the University Health Network and were conducted in accordance with the Guidelines of the Canadian Council on Animal Care. We also examined the influence of tensile forces on polarized mitosis of vascular smooth muscle in vivo. Adult male New Zealand White rabbits (body weight, 3.0 kg ±0.5 kg) and 3-week-old weanling rabbits (body weight, 1.1 kg ±0.2 kg) were anesthetized by IM injection of 0.8 mL/kg of xylazine (20 mg/mL) and 1.8 mL of ketamine hydrochloride (90 mg/mL), and anesthesia was maintained with a continuous intravenous infusion of the 1:9 xylazine/ketamine mixture (0.03 mL/min). A cervical midline incision was made, a segment of the left carotid artery was excised, and the cut ends of the vessel were reanastomosed using an 8–0 polypropylene suture (Ethicon). The length of the excised artery was adjusted such that lengthwise stretch of the artery was increased from resting levels of 60% to 100%.4 The incision was closed, and the animals were given 0.5 mL of Longisil (150 000 IU/mL of benzathine penicillin G, 150 000 IU/mL of procaine penicillin G, 0.9 mg of methyl paraben, and 0.1 mL of propyl paraben) IM and 0.2 mL of the analgesic Buprenex (0.3 mg/mL of buprenorphine HCl) SC. Arteries from unmanipulated animals served as controls.
At surgery or 3 days later, the rabbits were given IM injections of the thymidine analog bromodeoxyuridine (BrdU; 30 mg/kg), which incorporates into DNA in the S phase of the cell cycle. BrdU is cleared from the circulation within 60 minutes13; however, tissues were fixed ≥24 hours after injections, after BrdU-labeled cells had completed mitosis, so that the relative positions of doublets of daughter cell nuclei, detected by BrdU immunohistochemistry, could be used to assess the direction of mitosis.7 Six treatment groups were studied: groups 1 to 3, nonsurgical controls fixed 1, 3, or 7 days after BrdU injection, respectively; group 4, BrdU injection at surgery, fixed after 1 day; and groups 5 and 6, BrdU injection at 3 days after surgery, euthanized 1 or 4 days later, respectively. Later time points allow for assessment of relative migration of daughter cells after mitosis.
Tissue Fixation and Immunohistochemistry
Rabbits were euthanized with IV injection of barbiturate (pentobarbital sodium, 100 mg/kg, Bimeda-MTC Animal Health Inc) overdose. A rapid bilateral thoracotomy was followed by retrograde cannulation of the descending thoracic aorta and perfusion–fixation of the carotid arteries with 3% paraformaldehyde in PBS for 20 minutes at a perfusion pressure of 100 mm Hg. Whole-mount preparations of the vessel segments were prepared for en face immunohistochemistry. Alternatively, histological en face sections were stained with anti-BrdU antibody (Abcam), and images captured with a Nikon TE300 inverted optical microscope were analyzed using Simple PCI software to determine the axis of replication of daughter cells. The distance between centroids (geometric centers) of daughter cell nuclei was also determined to test for migration of daughter cells over time.
Sample sizes for in vivo experiments are presented in the Results sections and figure legends. All of the in vitro data represent a minimum of 3 independent replicates of each experiment. ANOVA followed by Bonferroni-corrected t tests were used to determine whether force application caused preferential orientation of mitosis or significant migration after mitosis. Student’s t test was used to determine whether force caused significant elongation of endothelial cells. P<0.05 was taken to indicate statistical significance.
Lateral Loading of Integrins Induces Orientation of Endothelial Cell Mitosis
Seeding RGD-coated microspheres, with and without application of force, did not affect mitosis rates or cell viability and, as expected, cells displayed random orientation of mitoses when no force was applied (Figure 1). During application of lateral force to RGD-coated beads over 24 hours, the orientation of cell division became strongly biased along the force vector. Accordingly, >50% of mitoses were oriented within 30° of the force vector, and ≈85% of mitoses were within 60° of the force vector.
Cell shape can influence orientation of mitosis,8 and endothelial cells elongate under shear stress; therefore, we tested whether loading integrins with microspheres induces similar elongation. Cells displayed significant elongation in the force direction length:width ratios after 24 hours, albeit to a much lesser extent than with application of physiological shear stress for the same duration. L///L⊥ became slightly <1.5 (Figure 2), whereas shear stress caused cells to become >2.5-fold longer in the shear direction than perpendicular to shear.7 However, no elongation was detectable at 16 hours after application of force (L///L⊥=1.04), a time by which 78% of the mitoses analyzed in Figure 1 were completed. These findings indicate that force, without coincident shape change, is probably an adequate stimulus for directional mitosis.
Role of Cortactin Is Control of Mitosis
Immunofluorescence imaging revealed that phosphorylated cortactin is ubiquitously expressed in confluent endothelial cells, including in the nucleus, diffusively through the cytoplasm and in a punctate distribution at the cell periphery (Figure 3, top). Imposition of lateral force effected redistribution of cortactin. Nuclear localization was lost, as was diffuse staining in the cell body and punctate staining at the cell periphery. Instead, the antibody decorated a subset of microtubules (Figure 3, bottom).
Localized distribution of cortactin and its phosphorylation by the nonreceptor tyrosine kinase Src8 have been implicated in directional mitosis; therefore, we expressed nonphosphorylatable cortactin in endothelial cells to test whether polarization of mitosis was affected. Surprisingly, growth arrest resulted (data not shown), possibly because of loss of target sites for the capture of astral microtubules.8,14,15
Imposition of Mechanical Load In Vivo Induces Orientation of Smooth Muscle Cell Mitosis
En face sections of carotid arteries from control animals showed consistent circumferential orientation of all of the daughter cell doublets, which indicated that mitoses were highly polarized in these unmanipulated vessels (Figure 4A through 4C). This distribution was not altered when tissue was examined 3 or 7 days after BrdU infusion; however, a slight but significant increase in separation of the daughter cells at later times indicates some postmitotic migration of the cells (Figure 5).
After longitudinal strain was increased from 60% to 100%, <50% of doublets labeled with BrdU at surgery exhibited circumferential orientation of mitosis 24 hours later (Figure 4D). When cells were labeled at 3 days and examined an additional 24 hours or 96 hours later, no preferential relative orientation of daughter cells was observed (Figure 4E and 4F). The data indicate that normalization of tensile stretch in arteries involved increased delivery of daughter smooth muscle cells in the axial direction.
During development, arterial adaptations to mechanical forces tune growth and reorganization of the vascular system to the changing blood supply demands of developing peripheral tissues, to the changing blood pressures that are required to drive these flows, and to the changing lengthwise tensions that are imposed on arteries by growth of contiguous tissues.16 Remodeling induced by mechanical forces also adapts the adult circulation to chronic changes in cardiovascular function, including those associated with exercise training, reproductive cycles, and pregnancy, and it affects the progression of vascular pathologies, including hypertension, atherosclerosis, and restenosis.3,17
Perhaps the most striking feature of these remodeling responses is that they consistently tend to normalize the mechanical loads imposed on the tissue, but it has remained unclear how vascular cells appropriately deliver new tissue in the circumferential, radial, or axial direction to selectively normalize shear stress, circumferential tension, or lengthwise tension, respectively. We have shown that an interesting remodeling of elastic lamellae contributes to this direction-specific tissue reorganization,18 and we now report that the delivery of daughter cells during mitosis is sensitive to directional forces and that integrins or integrin-associated molecules are capable of sensing force and initiating polarized mitosis.
The polarization of endothelial cell mitoses that followed the application of lateral forces to α5β1 and αvβ3 integrins mimicked the effects of shear stress that we reported recently.7 Likewise, the capacity of this force application to recapitulate shear-induced endothelial cell elongation, albeit to a lesser extent than that seen with physiological levels of shear, suggests that integrin loading is an adequate stimulus for a significant repertoire of endothelial cell responses to shear. Other sites of shear sensing/transduction have been implicated, with intriguing recent findings particularly implicating the endothelial cell–cell adhesion molecule platelet-endothelial cell adhesion molecule 1 (PECAM-1), and it could be argued that simple transmission of tension via cytoskeletal structures to PECAM-1 at cell junctions was eliciting cell elongation and polarized mitoses. However, additional studies, to be published elsewhere, have shown that magnetic microsphere loading of integrins versus PECAM-1 (using microspheres coated with anti–PECAM-1 antibody) produced very different profiles of intracellular signaling, for example, for mitogen-activated protein kinases, phosphatidylinositol 3′-kinase, and glycogen synthase kinase-3β. Indeed, 24 hours of PECAM-1 loading consistently led to apoptosis of most cells. These findings argue that the integrin loading that we studied here was producing a ligand-specific effect.
An important issue was whether polarized mitosis was a direct effect of mechanical loading or whether it was secondary to force-induced change in cell shape. We found, however, that change in cell shape was a late event, with no elongation being detectable after 16 hours. Similarly, shear-induced shape change requires 12 to 24 hours for postconfluent endothelium. Because the great majority (≈80%) of mitoses that occur in 24 hours of loading are completed by 16 hours, it is very likely that the highly polarized mitoses of these cells can be initiated as a primary response to force.
The signaling pathways through which force regulates orientation of mitoses are unknown. The microtubule system of cultured endothelial cells that have been exposed to shear stress displays planar cell polarity, an integrin-dependent behavior in which the microtubule organizing center moves downstream of the cell nuclei.19,20 This bias in microtubule positioning before the onset of mitotic spindle formation may underlie directional mitosis; therefore, critical signals downstream from integrins during initiation of planar cell polarity (eg, Cdc42, atypical protein kinase Cζ, and glycogen synthase kinase-3β) merit investigation in the regulation of polarized mitoses.19,20
Alternatively, directional cell division has been well studied for some cases where mitoses segregate differentiated from undifferentiated (stem) daughter cells after parent stem cell division.21,22 Although caution is needed when extrapolating from these studies of asymmetrical cell division in which daughter cells differ in the cellular constituents and often in size23,24 to mitoses that simply deliver similar daughter cells in defined directions, it is likely that the regulation of the 2 modes of polarized mitosis overlap. Therefore, it will be of interest to assess the roles, in directional vascular cell mitosis, of key players in the regulation of asymmetrical cell division, including heterotrimeric G proteins and their regulators (RGS7, RGS14, the guanidine nucleotide exchange factor, RIC-8, GoLoco, and NuMA).25,26
Asymmetrical cell division may also provide clues concerning how mitotic spindles become oriented in a specific direction during mitosis. In these mitoses, cell cortical domains, often composed of actin and actin-associated proteins (cortactin and ezrin), capture astral microtubules and exert pulling forces that wrench the spindle into a defined location, possibly through the actions of the microtubule-based motor protein dynein.20,23–25,27 Accordingly, culturing single isolated cells on micropatterned substrates of different shapes induces cells to divide along an axis that reflects the geometry of the substrate, and cortactin distribution is biased in a manner that reflects the direction of mitosis.8
The tensile forces imposed on integrins that proved capable of orienting endothelial cell mitosis also caused dramatic cortactin redistribution but in a manner that was unanticipated. A punctate pattern at the cell periphery, plus a diffuse distribution throughout the central regions of the cell, including the nucleus, prevailed in unloaded cells. After imposing tensile stress on integrins for 24 hours, cortactin was excluded from the nucleus and became largely localized to a subset of microtubules. This localization may reflect different roles for cortactin in endothelial and/or force-loaded cells; alternatively, microtubule-based transport may deliver cortactin to patches that orient mitoses in the early phases of the cell cycle.
Tyrosine phosphorylation of cortactin by Src was important in polarized mitoses of cells grown on patterned substrates8; therefore, we attempted to assess functions of cortactin by expressing a nonphosphorylatable mutant form of the protein. Surprisingly, expressing this modified cortactin induced growth arrest, a finding that underscores the importance of the protein in the control of mitoses but that neutralized attempts to define how it may participate in polarization of the process.
Our observation that mechanical forces influence the direction of mitosis of smooth muscle cells in vivo is potentially of central importance to arterial remodeling. We do not know whether cell shape or pulsatile arterial distension underlies the predominantly circumferential orientation of cell division in unmanipulated arteries, but the effects of perturbations in lengthwise mechanical load were striking and could clearly contribute to the marked axial growth that follows increases in these forces, whether these increases are experimental4 or because of growth of tissues to which arteries are tethered. Undoubtedly, this mode of remodeling may also participate in developmental and adaptive reorganization of many mechanically loaded nonvascular tissues. In some instances, (force-induced) mitosis in one direction may drive tissue growth in the same direction or in an orthogonal direction, for example, through intercalation of daughter cells between pre-existing resident cells. Something like the latter may occur during enlargement of flow-loaded arteries, when increases in vessel circumference3 accompany axial division of endothelial cells.7
Ultimately, control of mitoses and distribution of daughter cells are central to the most important vascular pathologies, including atherosclerosis, restenosis, and transplant arteriosclerosis. The capacity of normal and abnormal blood pressures and flows to influence the delivery of daughter cells may, therefore, impact significantly on these pathologies and may provide novel targets for therapeutic interventions.
This study demonstrates that physical forces imposed on vascular tissue have a dramatic effect on cell division and the directional delivery of daughter cells during arterial remodeling. This capacity of force transduction to orient mitosis has the potential to contribute greatly to the preferential wall thickening that accompanies hypertension, the growth of arterial diameter that follows elevation of blood flow rate, and the lengthwise growth of arteries that occurs when adjacent tissues grow or hypertrophy. As such, this phenomenon is central to the coordination of arterial growth with developmental changes in arterial function, the structural adaptation of the mature circulation to long-term alterations in cardiovascular function, and pathologic remodeling of vascular tissues.
We thank Donna Johnston for technical assistance.
Sources of Funding
The study was supported by grant T4910 from the Heart and Stroke Foundation of Ontario to B.L.L. and a Canadian Institutes for Health Research operating grant and Premier’s Research Excellence Award to A.K. D.D. was supported by a studentship from the Canadian Institutes for Health Research and Canadian Hypertension Society. P.J.B.S. was supported by a studentship from the Canadian Institutes for Health Research. T.C.V.R. was supported by the Charles Hollenberg Summer Studentship Program. J.T.K.L. was supported by a Heart and Stroke Foundation of Ontario summer studentship.
- Received February 23, 2007.
- Revision received March 16, 2007.
- Accepted April 13, 2007.
Langille BL, O’Donnell F. Reductions in arterial diameter produced by chronic decreases in blood flow are endothelium-dependent. Science. 1986; 231: 405–407.
Langille BL. Blood flow-induced remodeling of the artery wall. In: Bevan JA, Kaley G, Rubanyi G, eds. Flow-Dependent Regulation of Vascular Function. Oxford, UK: Oxford University Press; 1995.
Jackson ZS, Gotlieb AI, Langille BL. Wall tissue remodeling regulates longitudinal tension in arteries. Circ Res. 2002; 90: 918–925.
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.
McCue S, Dajnowiec D, Jackson MR, Langille BL. Shear stress regulates forward and reverse planar cell polarity of vascular endothelium in vivo and in vitro. Circ Res. 2006; 98: 939–946.
Goldschmidt ME, McLeod KJ, Taylor WR. Integrin-mediated mechanotransduction in vascular smooth muscle cells–frequency and force response characteristics. Circ Res. 2001; 88: 674–680.
Luscinskas FW, Lawler J. Integrins as dynamic regulators of vascular function. FASEB J. 1994; 8: 929–938.
El Sayegh TY, Arora PD, Laschinger CA, Lee W, Morrison C, Overall CM, Kapus A, McCulloch CA. Cortactin associates with N-cadherin adhesions and mediates intercellular adhesion strengthening in fibroblasts. J Cell Sci. 2004; 117: 5117–5131.
El Sayegh TY, Arora PD, Fan LZ, Laschinger CA, Greer PA, McCulloch CA, Kapus A. Phosphorylation of N-cadherin-associated cortactin by Fer kinase regulates N-cadherin adhesion strength. Mol Biol Cell. 2005; 16: 5514–5527.
Kriss JP, Revesz L. Distribution and fate of bromodeoxyuridine and bromodeoxycytidine in mouse and rat. Cancer Res. 1962; 22: 1257–1264.
Grill SW, Howard J, Schaffer E, Stelzer EHK, Hyman A. The distribution of active force generators controls mitotic spindle position. Science. 2003; 301: 518–521.
Hwang E, Kusch J, Barral Y, Huffaker TC. Spindle orientation in Saccharomyces cerevisiae depends on the transport of microtubule ends along polarized actin cables. J Cell Biol. 2003; 161: 483–488.
Miyachi M, Tanaka H, Yamamoto K, Yoshioka A, Takahashi K, Onodera S. Effects of one-legged endurance training on femoral arterial and venous size in healthy humans. J Appl Physiol. 2001; 90: 2439–2444.
Wong LCY, Langille BL. Developmental remodeling of the internal elastic lamina of rabbit arteries. Effect of blood flow. Circ Res. 1996; 78: 799–805.
Yamashita YM, Jones DL, Fuller MT. Orientation of asymmetric stem cell division by the APC tumor suppressor and centrosome. Science. 2003; 301: 1547–1550.