INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX REFERENCES

SINCE THE 1960S, there has been an increasing interest in the effects of various hemodynamic shear-stress exposures on endothelial-intimal permeability, structure (17, 21, 22), and susceptibility to atherogenesis (4, 23, 24). Atheroma are slowly developing, raised intimal lesions that are the leading cause of morbidity and death in modern societies. Early studies suggest that hemodynamic factors could influence the endothelial cell surface barrier in an adverse manner to initiate the chain of aberrant interstitial mass transport and cellular events of atherogenesis (2, 5, 23, 27, 30). Because one of the initiating events appears to be endothelial barrier failure (3, 29, 30, 42, 45), most subsequent work has focused on endothelial cell biology. To do this, special tissue culture systems were developed in which controlled shear stresses could be applied to cultured monolayers of endothelial cells. Endothelial "responses" to these controlled shear-stress exposures were based initially on observations of endothelial cell integrity, shape, population density, and apparent rheological changes. With the use of such cultured endothelial monolayers, it was found that an increase in shear stress was associated with increased endothelial pinocytotic activity and restructuring of the endothelial cell cytoskeleton. Elongation and realignment of the endothelial cells in the direction of the imposed shear stress (9, 13) was also observed, confirming earlier in vivo observations (17).

More recently, there have been improvements in methodologies for estimation of genetic expression, e.g., measurements of changes in the patterns of cellular proteins, mRNA, transcription factors, etc. Accordingly, it is possible to use such measurements to identify the underlying intracellular processes associated with endothelial shear-dependent observations. This important field of research is relatively new and growing rapidly (16). A few references of relevance to the present work are noted among the following: it has been shown that laminar shear-stress exposure tends to upregulate the various genes that are thought to maintain the integrity of the endothelial surface; examples are the upregulation of the genes for production of superoxide dismutase (15), nitric oxide (15), prostacyclin (18), integrin (6, 49), plasminogen (14), etc.

In addition to studies of steady laminar flow fields, turbulent and temporally as well as spatially, varying flow fields also have been studied in cell culture systems and reported to activate various other unique patterns of genes (6, 10, 11, 31, 37). Some endothelial genes appear to be activated by particular aspects of the flow fields. For example, endothelial phenotypic expression has been reported to vary with location in endothelial spatial shear-stress gradients (12, 41). Moreover, the expression of certain other genes appears to depend on various temporal gradients of shear-stress exposure such as those generated by impulse, pulsatile, and ramp flows (1, 19). Studies of such spatial and temporal shear-stress gradients are of considerable interest and importance because they can provide new insights into the cascades of stress-dependent signaling mechanisms in endothelial cells.

Thus previous studies have provided a large body of important new information from endothelial cell monolayers. It is important to place such observations in the context of the intimal-medial tissue system. To this end, there have been efforts by a few workers to study the effect of flow on arterial mass transport in the in vivo state (2, 22, 42) or in perfused vessels (38). In addition, perfused vessel segments (5, 32, 35) or surgical preparations (7, 40) have been used for studies of stress-induced phenotypic expressions of endothelial and smooth muscle cells in the intact arterial structure. The observations from these kinds of tissue studies have an added physiological significance because the cells are interacting in the native arterial structural environment. However, these kinds of approaches can only provide limited experimental control for studies of certain more complex questions. For example, such control is essential for studies of processes that initiate and orchestrate the coevolution of cellular interactions that lead to the development of atheroma and preatheromatous intimal structures.

Thus, to study these more complex tissue choreographics, an experimentally controlled, metabolically supported, ex vivo artery preparation was developed in an effort to provide the requisite control in an arterial architectural arena. This tissue-oriented methodology consists of an ex vivo metabolically supported swine artery preparation and specially designed perfusion and flow-configuring devices. With the use of this methodology, new observations are presented that explore coevolving, shear-stress-dependent, arterial responses as they appeared during 0- to 95-h periods of arterial shear-stress adaptation. For the present protocols, the following types of simulated hemodynamic shear-stress exposures were studied: 1) uniform laminar flow, 2) flow "separation-reattachment phenomena," 3) multidirectional diverging-converging flows, and 4) stagnant flows.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX REFERENCES

Vessel harvest. All procedures used in this study were carried out in accordance with institutional guidelines regarding the care and use of experimental animals and were done under aseptic conditions. All animals were fully anesthetized with telezol (6 mg/kg body wt)-xylazine (2 mg/kg body wt) followed by 2.5 mg/kg pentobarbital. Thereafter, pentobarbital was given as necessary to maintain a surgical level of anesthesia. A left lateral thorocotomy was performed, and heparin (3 U/ml blood) was then given and circulated 4 min before vessel excision. (This was to inhibit activation of platelets.) The descending thoracoabdominal aorta was then excised from young adult, ~70-kg, mixed sex, Hampshire-Durak, normocholesterolemic swine (n = 17). Swine are generally accepted models for studies of human atherogenesis. After excision of the artery, the vessel was slowly relaxed (to minimize potentially damaging viscoelastic and myogenic stresses in the tissue) and immersed in a 37°C bath of the sterile nutrient media that was used for all of the present flow studies. This media consisted of 50% pooled porcine serum plus 50% modified MCDB-131 solution (39) containing amino acids, vitamins, growth factors, etc. (Sigma MCDB-131 No. M-8737 or GIBCO-BRL MCDB plus ampicillin and amphotericin B No. 98-5100EC). The solution was modified to provide Mg²⁺ and HCO concentrations closer to those of the vessel donor values and adjusted to pH 7.4.

Metabolically supported ex vivo vessel preparation. The rapidly excised vessel was opened longitudinally along its dorsal aspect through the intercostal orifices. It was then stretched slowly to its in vivo length and clamped as a flat sheet along each longitudinally cut edge, with the endothelial side facing upward, in a specially designed tissue-holding device (THD; Fig. 1). The THD was then adjusted to restore the former in vivo circumferential dimension. Next, the sheet of tissue was tightly captured between the two parts of a "well assembly" device, a top well assembly (TWA) on the endothelial surface and a matching bottom well assembly (BWA) on the adventitial surface. When applied to the arterial surface, the device isolated the arterial sheet of tissue into 18 consecutive matching endothelial and adventitial chambers, respectively, along the vessel (as noted by the numbers on the side of the TWA in Fig. 1). As shown to the left in Fig. 1, the BWA chambers were equipped with pressure and adventitial perfusion lines. In the present study, the BWA chambers were also equipped with screen-covered, longitudinally grooved, adventitial support devices. The purpose of these was 1) to prevent downward deformation of the pressurized vessel surfaces so that the endothelial surface in each well was maintained in a flat configuration and 2) to direct a flow of fresh nutrient media along each adventitial surface to facilitate nutrient and metabolite diffusion at this surface.

View larger version (88K): [in this window] [in a new window]   Fig. 1.   A longitudinally opened, physiologically stretched artery (light yellow) with endothelial surface facing up and secured in the tissue-holding device (THD). The tissue is captured between the top (TWA) and bottom well assembly (BWA) parts of a well assembly device. The tubing emanating from the left of the BWA provides control of the requisite transmural pressure (100 mmHg) and the perfusion of nutrient media to the adventitial surfaces of each well. The four cuboidal Lucite devices extending above the TWA (wells 2-5) are the well mounts to secure selected flow-configuring devices that extend down into each well. Each tubing that extends upward from these mounts couples to the corresponding flow pumps as diagramed in Fig. 2.

View larger version (31K): [in this window] [in a new window]   Fig. 2.   Diagram of the overall ex vivo vessel system's flow logic for endothelial surface shear-stress field exposures and reagent refreshment. Circles at the top of the figure are flow pumps; the attached straight lines with arrows represent the directions of pump flow [solid lines show steady operation flow (STEADY OP) and dashed lines show superimposed refreshment flows every 2 h]. The large hatched [organ support system (OSS)] wall below the pumps encloses the environmentally controlled milieu of the ex vivo OSS system. FA, flow applicator; R, return well; F, flow (diverging and converging); W, waste well (stagnant flow). Partitions separating each well are hatched, and clear regions are apertures for overflow from the FA, R, and F wells to the W well during refreshments. The associated curved lines with arrows in each well suggest the approximate paths of the flow-velocity fields. The arrows in the coregistered squares below each well suggest the corresponding types of endothelial surface (Endo Surf) shear-stress fields.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Temporal contours and values (arrows) of the experimentally controlled pulsatile flows (Q) to the wells. cpm, Cycles per minute, i.e., pulse rate.

Designs for the flow-configuring devices. As mentioned, the quartet format of the wells (Fig. 2) was chosen in the present study so that at least 4 types of flow fields could be replicated at 4 separate regions along the 18-well preparation to average out site variability (30). For reference purposes, the sketched curved arrows in Fig. 2 in the FA, R, and F well chambers represent the general directions of the STEADY OP velocity fields. The corresponding directions of the endothelial surface shear-stress fields are suggested by the second set of arrows sketched in each of the squares below each well in Fig. 2. Because the boundaries, entrance conditions, and fluid constitutive properties of the flow were known, the velocity fields and corresponding endothelial shear-stress fields were evaluated by computational fluid dynamics (CFD). Because the requisite boundary conditions for such computations are determined by the geometries of the flow applicators in the wells, these designs will be described next.

View larger version (26K): [in this window] [in a new window]   Fig. 4.   R well simulation of flow "separation" and "reattachment" phenomenon. Top: diagram representing the region of flow separation (S) in the entrance of an arterial daughter branch. Lines with arrows represent regional flow directions. S1, point of flow separation; S2, point of flow reattachment. Bottom: diagram of R well simulation of the above types of arterial flows.

View larger version (42K): [in this window] [in a new window]   Fig. 5.   Top: numbered, color-coded, regionally averaged shear-stress magnitudes between selected shear-stress isopleths for FA, R, and F wells. Middle: color keys and mean values (±SD) for the coded regions shown in top. Bottom: corresponding computed shear-stress vector fields. The vector magnitude calibrations appear as the small horizontal bars in the parentheses. RD, dark blue region of interest; RL, light blue region of interest.

F well configuration. The flow configuration of the F well was designed to examine the effect of ordered radiating shear-stress fields and gradients. To do this, a two-pipe flow-configuring device was designed, as shown for the F well in Fig 2. This configuration had proximal (left) influx and distal (right) efflux pipe orifices located 3 mm above the endothelium in the midline of the well, as shown in Fig. 2. Unlike the R well, the surface of the F well experienced two, nonuniform, shear-stress fields, as sketched in Fig 2 and quantified in RESULTS (Fig. 5). The input pipe on the left created a diverging laminar stress field emanating from a central stagnation point below the pipe orifice. The efflux flow into the right pipe created a similar but converging stress field.

W well configuration. The last well (W well) in the quartet was a "zero shear-stress exposure" well, in which the efflux pipe was located remotely at the fluid surface 29 mm above the endothelial surface. The periodic 2-h waste flow from the other wells into the W well occurred through the aforementioned rectangular aperture of the left partition and then to the "waste" efflux pipe on the right. Moreover, this only occurred during the 20-s refreshment of the media in the quartet every 2 h. Accordingly, the shear stress on the W surface was virtually zero, even during the brief (20 s) refreshment periods. (A removable rectangular silicone rubber dam was inserted in the rectangular aperture across the top of the right partition to isolate the two quartets.) The design of the W well flow configuration was to examine the tissue responses to various durations of flow cessation temporarily in vessel branches during severe reapportionment of flow distribution or after (temporary or permanent) surgical occlusion of a vessel.

Computations of endothelial surface shear-stress distributions. Methodology for direct measurements of endothelial surface shear stresses in the small flow geometries of the present study is currently unavailable. Moreover, analytic mathematical formulas for the complex flow geometries of the study also do not exist. Accordingly, such shear-stress estimates must rely on approximations provided indirectly from calculations of the relevant flow velocity and corresponding shear-stress fields using a computer. Performing such approximations is referred to as CFD. Software for such complex calculations is commercially available. Doing such computations is essentially a two-step process. In the first, the enclosed geometry of the flow environment is defined and filled with a "finite-element computational mesh" that divides the entire volume of the flow system into a three-dimensional array of contiguous mechanically interacting finite fluid elements. The preceding geometries of the FA, R, F, and W wells define the requisite enclosed flow environment for the mesh. This mesh and the sizes of the finite elements in the mesh are detailed subsequently in the APPENDIX.

Evans blue dye-albumin complex permeability measurements. At the end of each experimental stress exposure, all flow devices were removed from the wells, cultures for bacteria and fungi were taken, and the nutrient media were then replaced with 100% sterile serum containing a known concentration (c₀; in nmol/cm³) of Evans blue dye-albumin complex (EBD-alb). This EBD-alb-containing serum medium was stirred continuously throughout this final EBD-alb uptake period (~1 h) to avoid concentration polarization of EBD-alb at the endothelial surfaces.

(1)

Apparent EBD-alb permeability calculation. For the present study, the concentration- and time-normalized uptake measurements were used to represent an "apparent" transendothelial EBD-alb permeability coefficient (; in cm/s), which was calculated from the fields of M_E measurements using the approximate relationship (29)

(2)

In Eq. 2, M_E (in nmol/cm²) is the transendothelial uptake (Eq. 1) at a given point on the endothelial surface of a well, c₀ (in nmol/cm³) is the concentration of EBD-alb in the serum, and T (in s) is the duration of endothelial exposure to the stirred serum-EBD-alb solution. As will be shown in RESULTS, stress-dependent permeability data tend to vary in magnitude depending on the location in the particular "hemodynamic" shear-stress field. In an effort to quantify the relevant aspects of these various permeability fields, the  data were expressed as the local area mean value of the permeability () for each region of interest. The FA, F, and W wells tended to have uniformly mottled fields, whereas the R well had two EBD-alb uptake regions of interest: the region of flow separation and the region of reattachment (as discussed further in RESULTS).

Preparations for microstructural studies. At the end of the EBD-alb exposure period, the well surfaces were rinsed very briefly with buffered saline and immediately covered with fixative (formalin or glutaraldehyde) for structural studies. Some of the experiments were fixed in a bath of 10% phosphate-buffered formalin and then transferred to a phosphate buffer containing 6% sucrose. Other tissue preparations were fixed in 3% phosphate-buffered glutaraldehyde for 24 h and then transferred to the buffered sucrose solution. After these fixation periods, the tissues were prepared for light microscopy by acetone and alcohol dehydration followed by glycolmethacrylate resin (GMA) infiltration. The tissues were then multiply embedded in GMA as a single block for comparative light microscopic study of the stress-dependent structural changes along contiguous wells. Each multiply embedded GMA block, which contained the four types of stress field-exposed tissues from a given quartet, was serially sectioned at 2 µm to obtain longitudinal arterial histological sections from the central axial region of all wells. These were stained with methylene blue-basic fuchsin. In addition, using a modified resin etching technique (47), some of the formalin-fixed tissue sections were also immunostained for endothelial cell nitric oxide synthase (ecNOS; Serotec No. MCA1746).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX REFERENCES

The results of the computed endothelial shear stress fields are presented in Fig. 5 for the FA, R and F flow configurations as noted at the top of each column. (The W shear-stress field was zero and thus does not appear.) Referring to the bottom row in Fig. 5, note that the computed endothelial shear-stress fields are represented by the magnitudes and directions of the vector field plots for each well. The small, horizontal bar at the bottom in Fig. 5 is the scale factor for that vector field. (Note also that the endothelial stress vectors are equal in magnitude to, but 180° from, the corresponding fluid shear-stress vectors.) The circles in each field represent the location of the flow supply pipes for reference purposes. Those for the R and F wells represent the actual flow applicator portion of the unit 3 mm above and facing the endothelial surface directly. The circle for the FA well does not communicate directly to the endothelial surface, but feeds the left vertical flow manifold, which was designed to produce the uniform flow field on the FA endothelial surface. The stress vector fields corresponding to the R well and the F well are seen to be nonuniform. These nonuniformaties are related to the converging or diverging flow fields around the pipe orifices. The corresponding three-dimensional velocity vector fields were too complex to depict in two dimensions, but followed the general paths that were indicated in the flow logic diagrams in Fig. 2 or 6.

View larger version (37K): [in this window] [in a new window]   Fig. 6.   Permeability response to stress. Top: diagram of the flow logic through the wells drawn in register with corresponding photographs of stress-exposed well surfaces. Bottom: en face color photographs of the corresponding stress-induced transendothelial intimal Evans blue dye-albumin (EBD-alb; blue) uptake patterns. Durations of stress exposure are noted at the left of each row of photographs. The missing three W well tissues on the right (stagnent flow) were left unexposed to the EBD-alb so that they could be used as added I0 reflectance wells for Eq. 1 while still providing their zero-stress, structural response information.

More importantly, the regional magnitudes of the associated shear-stress fields on the endothelial surfaces are represented in Fig. 5, top, by the numbered (1-12), color-coded regions in the figure. Each color-coded region represents the corresponding area-averaged shear-stress vector magnitudes between selected shear-stress isopleths. The corresponding numbered keys to this color coding appear in Fig. 5, middle, as the mean shear stress (±SD) for each of the numbered colored regions in Fig. 5, top.

Referring to the magnitudes of the FA endothelial surface shear-stress field, it can be seen that the shear stress was uniform at ~16 dyn/cm² over most of the FA flow channel. More specifically, stress was sharply elevated to ~26 dyn/cm² along the first 2 mm of the channel (the "entrance effect"), ~16 dyn/cm² over the next 7 mm, and ~3 dyn/cm² over the last 1 mm of the flow channel. Thus properties of the shear-stress field over the 7-mm-long, uniform shear-stress region of the FA regions were shown to be laminar and uniform except very near the well walls (Fig. 5, see FA vector field). The properties of the R and F stress fields were much more nonuniform and interesting than those of the FA fields. However, it is more appropriate to consider these more complex stress patterns in the DISCUSSION, after presentation of the corresponding tissue stress-response data.

The corresponding stress associated evolution of permeability changes in the FA-R and F wells are shown in the photographs of Fig. 6, bottom. These EBD-alb uptake data appear as the blue patterns in each coregistered column of wells under the OSS flow logic diagram (Fig. 6, top). The various durations of shear-stress exposure are noted at the left of each row of wells. Referring to the final row of well tissues, it can be seen that a 1-h exposure to the FA, F, and R types of shear-stress fields was not accompanied by a significant change in the intimal EBD-alb uptake responses. However, referring to the next row (23 h), it can be seen that the uptake increased during the initial 23 h of exposure to each of the different stress fields. With the exception of the F wells, such patterns appear as uniformly mottled blue with little relationship to the type of imposed stress field. However, over the ensuing stress exposure durations, these patterns appear to change progressively with time to converge toward the virtually stable new set of EBD-alb uptake patterns shown for the 75- and 95-h stress exposures (Fig. 6).

In addition to these visual assessments, the magnitudes of these EBD-alb permeability patterns were estimated by the photometric scanning and analytic procedures described in METHODS. The individual area-averaged permeability values ( × 10⁶ cm/s) for each major region of interest in the wells appear in the four scatter plots of Fig. 7. In Fig. 7, A-D, is on the ordinate and the duration of the respective stress exposure (t; in h) is on the abscissa. Each data point represents the mean (n = 4) value of the data for the particular shear-stress field at its four locations along the vessel to minimize location dependence in the data (30). Referring to Fig. 7A, the FA (circles) and the F (triangles) permeability responses were similar despite the radically different directions and magnitudes of the stress vectors on these endothelial surfaces. A single smooth "trend" curve for these FA and F stress-dependent permeability responses was sketched through these data points for subsequent comparisons to corresponding trends for the other types of stress field exposures (Fig. 7, B-D).

View larger version (24K): [in this window] [in a new window]   Fig. 7.   Scatter plots of individual mean stress-dependent EBD-alb permeability responses (open symbols). Ordinate, permeability ( × 106 cm/s); abscissa, duration of stress exposure (t; in h). The smooth, sketched curves represent approximations of the corresponding temporal trends of the shear-stress-induced permeability response data. The types of stress field exposures were as follows: ordered laminar FA/F flows (A); flow reattachment region (B); flow separation region (C); and flow stagnant (but refreshed) region (D).

The FA/F temporal trend curve suggests that after the abrupt transition from the in vivo stress-adapted state to the new ex vivo shear-stressed state, the permeability rose, after a brief 3- to 4-h latent period, to a maximum at ~23 h. The curve then declined gradually over the ensuing hours (75-95 h) toward the former in vivo-adapted values. Unlike the permeability patterns in the FA/F regions of interest, which were reasonably uniformly distributed over the well surface, the R well permeability patterns tended to approach a bimodal distribution (see the R well column of the photographs in Fig. 6). This can be seen most clearly for the R wells for the 75- and 95-h exposures in Fig. 6. Accordingly, the permeability estimates for each mode of the R well regions of interest were analyzed somewhat arbitrarily as from predominantly uniform regions of "light blue" or "dark blue." This division was admittedly subjective and ignored the regions of transitional uptake values between the two "uniform" regions of the R wells. The permeability data for the "light blue" region of interest in the R well are presented in Fig. 7B, labeled as R_L. As noted, this predominantly lightly stained region is around the flow return pipe (simulated flow reattachment region). The temporal data for the "dark blue" region of interest in the R well appear in Fig. 7C, labeled as R_D, the simulated flow separation region in the imposed flow field.

The temporal evolution of permeability in the reattachment region (R_L) rose to a maximum at 23 h and then decreased briefly with a trend resembling that of the FA/F curve in Fig. 7A; however, this decline reversed around the 40-h point and thereafter rose slowly to higher values. Referring to the R_D trend curve in Fig. 7C, this curve appears to have a shape that is similar to that of the R_L curve but has a greatly increased magnitude (note ordinate scale change in Fig. 7, C and D). Moreover, comparing these R_D response data with those in Fig. 7D for the flow "stagnation" simulation (W well), the permeability responses appear to be indistinguishable. Thus only the FA and F shear-stress fields appear to be associated with a normal permeability adaptation. Not surprisingly, all of the above stress-dependent permeability response data will be shown to be associated with interesting coevolving intimal-medial structural changes including raised intimal lesions, as described below.

Gross visual changes. Visual examination of the endothelial surface textures of unstained and only slightly stained EBD-alb uptake regions in all wells consistently revealed a smooth glistening surface that was flat except for the normal, slightly elevated patterns of the underlying bands of intimal connective tissue. In contrast, most darkly stained (high permeability) areas of the R_D (and W) endothelial surfaces developed elevated, edematous-appearing intimal regions covered with slightly gray fields of very small, rounded elevations. These "dark blue" regions presented a pebbly, gelatinous appearing surface texture resembling fields of small blisters. An en face photograph of such a surface is shown in Fig. 8 for a 23-h R well with the R_D (flow separation) region on the left and the R_L (flow reattachment) region on the right. In some of the edematous R_D regions, a grayish, opalescent, mucoid-appearing surface layer tended to obscure the blistered texture. The corresponding microstructural characteristics of these lesions, as well as normal regions, are detailed below.

View larger version (80K): [in this window] [in a new window]   Fig. 8.   Gross appearance of the endothelial surface after a 23-h exposure to a R well shear-stress field. On the left, observe the edematous field of small blister-like elevated lesions in the darkly EBD-stained RD (simulated flow separation) region. On the right, note the virtual absence of these lesions in the lighter-stained RL region (simulated flow reattachment).

Histological changes. The evolution of histological changes associated with the evolution of the permeability responses in Fig. 7 are represented by the photomicrographs in Fig. 9. The top photomicrograph represents a "control" tissue section from a nonstress-exposed but otherwise methodologically identically prepared (including 1-h EBD-alb exposure) ex vivo tissue preparation. This "control" histology provides a reference for evaluating structural differences due to preparative and EBD-alb exposure procedures from the structural differences induced solely by the duration and type of shear-stress exposures. Examples of salient stress-induced changes are represented by the paired FA and R_D tissue sections shown for 23-, 75-, and 95-h shear-stress exposures, respectively.

View larger version (82K): [in this window] [in a new window]   Fig. 9.   Photomicrographs of 2-µm-thick, methylene blue-basic fuchsin-stained resin-embedded tissue sections demonstrating the evolution of the shear-stress-induced intimal-medial histological responses: control tissue (top), FA tissues (left), and matching RD tissues (right). Durations of stress exposure are noted in the center of each set of paired FA and RD photomicrographs. Staining characteristics for methylene blue-basic fuchsin are as follows: endothelial nuclei, deep purplish-blue; cytoplasm, red; smooth muscle cell nuclei, dark blue; elastic lamella, white; tissue matrix, irregular grayish lavender material in the interlamellar spaces; and collagen, light pink. Bottom: resin-embedded, formalin-fixed FA and RD sections from a 75-h study immunostained for endothelial cell nitric oxide synthase (red) and counterstained with hematoxylin (blue). Solid bar = 50 µm.

Immunohistology for ecNOS. In view of the intimal devastation, cellular migration, and apparent sequestration or engulfment activity (see below) associated with the "surface" cells in the R_D intimal lesions, the question of their phenotypic identity arose. In many ways, these "surface" cells acted more like macrophages or migrating "modified" smooth muscle cells than endothelial cells. In an effort to clarify their identity, a formalin-fixed, 75-h pair of FA and R_D well tissues were immunostained for ecNOS. These sections appear in the photomicrographs in Fig. 9, bottom. Referring to the FA tissue on the left, it can be seen that the endothelial surface cells are strongly ecNOS positive (stained red). Referring to the corresponding R_D tissue section on the right, these "endothelial" cells also appeared to be ecNOS positive, although somewhat less so. Unlike the FA tissue, the R_D tissue also developed discrete ecNOS positive extracellular "granules." These appear to be scattered throughout the matrix of the subjacent intimal-medial tissues but not in or on the smooth muscle cells of this layer. (Negative control slides indicated no nonspecific binding.) This suggests that the "normal" FA and "abnormal" R_D intimal surface cells are of endothelial origin and are not macrophages or smooth muscle cells.

Early lesion ultrastructure. The endothelial ultrastructural features of the initial 23-h R_D stress-dependent responses are also of considerable interest. Typical changes are represented in the scanning and transmission electron micrographs of Fig. 10. The scanning electron micrographs (Fig. 10A) show a variety of elevated (light gray) structures lying on the (dark gray) endothelial surface. Looking more closely, this darker surface contains scattered small openings (black) to the subendothelial tissues. There also appear to be at least three other surface structures of note. The first are elevated, blebbed spheres with radiating cytoplasmic tethers to or from the subjacent surface. The second are two large, relatively flat structures resembling migrating cellular lamellapodia, one of which seems to be engaging the upper region of a blebbed sphere. The third are irregular patches of particulate debris scattered on the endothelial surface.

View larger version (85K): [in this window] [in a new window]   Fig. 10.   A: scanning electron micrograph of an endothelial surface region exposed to a 23-h shear-stress field of simulated flow separation phenomenon (RD). Vertical white bar represents 10 µm. B: transmission electron micrograph of an endothelial surface region also exposed to a 23-h RD flow field. The vertical dimension of B represents 5 µm. Note the phagocytosis of the apoptotic-appearing cell on the left by the neighboring (endothelial?) cell on the right.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX REFERENCES

The methodology detailed in this report represents an effort to develop an experimentally accessible arterial intimal-medial tissue system in which controlled, simultaneous observations of induced pathophysiological processes can be studied. In its present state, the methodological simulation of an in vivo artery has obvious imperfections, as implied in the detailed descriptions in METHODS. Those imperfections most relevant to interpretation of the present observations are as follows: 1) cell-free serum and modified MCDB-131 nutrient media instead of blood, 2) new planar instead of cylindrical mural strain fields, 3) metabolic dependency of the outer medial lamellar layers on diffusion from the adjacent adventitial flow of nutrient media instead of from the vasovasorum, 4) potential sequella of the initial tissue harvest and preparative procedures, and 5) shear-stress estimation in the FA system is very sensitive to the value of . Further discussion of these "imperfections" exceeds the scope of this report and must be the subject of continuing studies.

Despite its imperfections, this methodology provided previously unavailable flexibility for experimental design and control of many experimental variables in arterial tissue preparations from swine (and ultimately from human donor arteries). The present study provided a well-controlled set of new data regarding temporal sequences of intimal-medial functional and structural changes associated with arterial adaptation to defined flow fields of interest. It was shown that when a previously in vivo-adapted artery is challenged by a new mechanochemical environment, it reacts first with an increased permeability. This probably represents a physiological dismantling of the endothelial barrier in preparation for adaptation to its new laminar or disturbed flow environment.

After the initial 23-h exposure to the new experimentally controlled environment, the subsequent sequences of permeability and structural changes appeared to become more shear-stress field dependent. Two distinctly different temporal paths of adaptive dynamics were observed. One path was shown to be a constructive adaptive sequence with a return to normal intimal barrier function. This was associated with the locally ordered laminar shear-stress field exposures of the FA and F wells. In contrast, the other path was shown to be a progressively destructive sequence of increasing intimal permeability, edema, loss of matrix, structural degradation, and raised intimal lesion formation. This path was associated with the disordered, low shear-stress fields of the R_D and W regions of the artery. These data are of interest regarding in vivo atherogenesis.

Atherosclerosis normally occurs in vivo at particular localized sites of "physiologically" increased permeability in the chronically stress-adapted arterial tree (23, 30, 45). In the present work, the stress-induced permeability patterns of the 75- and 95-h stress exposures appear to be approaching the end stages of such stress adaptation. Therefore, these temporally averaged, ex vivo response data should be reasonable approximations to that of an in vivo, physiologically stress-adapted vessel, were it exposed to the same type of chronic laminar or low shear-stress field patterns used in the present study. Accordingly, it is of interest to compare the shapes of these 75- to 95-h simulated "physiologically adapted" EBD-alb permeability patterns to the patterns and values of the corresponding "chronic" shear-stress exposures shown in Fig. 5. This comparison allows one to examine the relationships of regional stress patterns to the corresponding regional physiologically increased permeability in the 75- to 95-h "stress-adapted" tissues shown in Fig. 6. Referring to Fig. 5, top and middle, as well as Figs. 6 and 9 together, simple inspection indicates that increased EBD-alb permeability and raised intimal lesions correlate negatively with the intensity of the shear-stress exposure. Referring to Fig. 5, these comparisons suggest that stresses between 0.38 (regions 5-8) and 0 dyn/cm² (W well) were unable to sustain survival of the endothelial barrier and associated intimal structure. Thus the integrity of the endothelial surfaces and subjacent tissues depended on maintenance of organized, temporally averaged shear-stress fields having magnitudes of at least 1.43 dyn/cm², i.e., region 12 of the F well. Referring to the R_L well stress field, it appears that shear stresses as low as 0.15 dyn/cm² appear to maintain some initial but temporary protection for the endothelial barrier. However, by 75-95 h of exposure, the barrier in all regions of the R well appeared to be failing, with raised lesions continuing to develop, particularly in R_D shear-stress region 8 (0.03 dyn/cm²).

As mentioned in RESULTS, comparisons of the FA and R_D medial structural responses in Fig. 9 indicated that the subjacent medial tissues also play an active and significant role in the arterial adaptive remodeling processes. This role appears to be one of cellular, connective tissue and matrix hyperplasia to strengthen or preserve the intimal-medial junctional tissue barrier. The similarity of this medial hyperplastic remodeling in both the FA and R_D tissues suggests that the driving forces for these medial structural responses were not related directly to the detailed types of stress field exposures. Instead, this remodeling appeared to be driven by other related forces such as increased interstitial chemical and hydration gradients, as well as the generation of various metabolites and stress-induced paracrine substances associated with the antecedent permeability response trends. In any case, the sequence and nature of the medial adaptive processes for FA and R_D environments appeared to be indistinguishable. Such adaptive processes are of interest and importance in understanding the pathophysiological mechanisms of atheroma formation.

Intimal pathophysiological implications. Were some of the hemodynamically dependent lesions that have been described to occur in a blood-filled artery, then the exposed intimal collagen obviously would be thrombogenic. Moreover, with or without thrombus deposition, the detailed patterns of transmural mass transport of water and macromolecules would be greatly distorted from that in the normal permeability case. In the normal case, the endothelial surface is essential for the continued maintenance of an optimized cellular, structural, and molecular environment in the arterial interstitium. Among other things, this normal interstitial state is maintained by the structure-dependent diffusive and convective paths of mass transport across the pressurized endothelial cell layer and its subjacent, gelled matrix system. The normal unpressurized endothelial cell surface and basement membrane act as a diffusive barrier that, when pressurized, also acts as a macromolecular sieving barrier (28, 29). In this case, the macromolecular flux becomes "sieved" or retarded with respect to the transmural water flux at the plasma-endothelial interface (27, 29). Accordingly, the chemically active concentration of plasma macromolecules at the normal endothelial (or at the basement membrane) interface would become progressively higher than that in the plasma were it not continuously diffused or swept away by the adjacent blood flow (27-29). Moreover, the associated transendothelial pressure drop related to this sieving is essential for maintenance of an optimal hydration of the compact (gelled) state of the subjacent endothelial-intimal matrix. With failure of the endothelial surface barrier, such as occurred in the R_D and W tissues, the transendothelial pressure gradient is lost, allowing the intimal gel phase to swell and ultimately dissolve. The previous, normally protein-poor, intimal plasma filtrate no longer can be maintained. Instead, this dilute filtrate is replaced by virtually pure plasma molecules mixed with soluble matrix molecules from the now unconstrained dissolution of the intimal gel-phase matrix (gel  sol) (8, 34, 44). The overall transmural pressure gradient will now drive a convective flux of this unsieved mixture of macromolecules further into the deeper tissue layers. When this flux encounters a new macromolecular barrier, e.g., the aforementioned compact subjacent intimal-medial layer (Fig. 9), a new sieving site will be created. Because this site is sequestered from the blood flow, the resulting accumulation of rejected solutes can rise to abnormally high (chemically active) interstitial concentrations (27-29). These regions finally reach new but abnormally high steady-state transmural concentration distributions. This new steady-state mass transport occurs at the point when diffusive fluxes (mostly retrograde to the lumen) from the sieving site become equal to the rates of convective solute rejection at the site (26, 27). Unlike the associated accumulation of soluble matrix molecules, the associated interstitial accumulations of plasma-borne atherogens at the site, e.g., oxidized free and bound LDL (46), only can be removed by inflammatory phagocytic transport and degradation (thus initiating an inflammatory atheroma?). The R_D intimal-medial structural responses shown in Fig. 9, right, provide the necessary architectural elements for such pathological mass transport scenarios.

Relationship to collateral fields of research. Most of the present set of observations are unique but have relevance to a number of previous important observations. First, assuming that the present gelatinous lesions are related to subsequent atheromatous changes, this finding would be in agreement with earlier correlative data, suggesting that human atheroma appear to develop at presumed sites of low or disturbed shear-stress fields (33). Second, the gross and microscopic architectures of the induced intimal lesions of this report are very similar to the human "gelatinous lesions" described in the older histopathological literature (36, 50) as potential precursors of atheromas. Third, it has been shown that human gelatinous lesions occurred most frequently in regions near or on the edges of mature, raised atheroma (36). In the case of elongated raised human atheromas, it was noted that gelatinous lesions tended to localize frequently at the upstream and particularly at the downstream end of the raised atheroma. These are sites where locally disturbed, low shear-stress fields, similar to the present R_D fields, might be expected. Fourth, in this regard, it has been shown that such sites also have a significantly higher incidence of apoptotic endothelial cells than other regions (48). Taken together with the observations in Ref. 36, one could infer colocalization of gelatinous lesions with apoptotic activity as also might be inferred from the electron micrographs in Fig. 10 of the present study. Fifth, this inference would be consistent with recent tissue culture studies in which sustained exposure to laminar shear stress was shown to prevent chemically induced activation of cellular apoptotic enzyme systems (15). In contrast, disturbed flow has been shown to induce apoptosis in the endothelial cell monolayer (20). Finally, partial ligations of the left carotid artery in immature (3 wk old) rabbits have been shown to be associated with greater apoptotic degradation of the arterial endothelial and smooth muscle cells in ligated than in nonligated control arteries (7).