Single Cell Morphological Metrics and Cytoskeletal Alignment Regulate VCAM-1 Protein Expression

Abstract

In the initial stages of atherosclerosis, vascular adhesion molecule-1 (VCAM-1) is a surface protein that mediates leukocyte adhesion to the endothelium’s luminal surface. VCAM-1 expression is upregulated on endothelial cells (ECs) under pro-inflammatory conditions and is known to be modulated by fluid shear stress (FSS). High, pulsatile FSS induces endothelial elongation and cytoskeletal alignment and downregulates pro-inflammatory induced VCAM-1 expression, which is associated with an athero-protective EC phenotype. In contrast, athero-prone ECs under low, oscillatory FSS fail to elongate and maintain a cobblestone morphology with random cytoskeletal alignment, while VCAM-1 expression is upregulated. Whether EC shape and cytoskeletal alignment play a role in the regulation of VCAM-1 protein expression independent of FSS has not been previously determined. The goal of this study was to determine the effect of EC morphology, specifically cell elongation and alignment, and cytoskeletal alignment on VCAM-1 protein expression using topographical micropatterning of an endothelial monolayer and single cell image analysis techniques. Elongated ECs with an aligned cytoskeleton significantly downregulated VCAM-1 protein expression in the absence of FSS compared to planar controls. In addition, linear correlations between morphological metrics and protein expression showed that actin alignment had a significantly stronger effect on VCAM-1 expression than cell elongation. Functionally, monocytic U937 cells statically adhered less on micropatterns compared to planar substrates, in a VCAM-1 dependent manner. Therefore, endothelial cellular elongation and alignment as well as cytoskeletal alignment regulate VCAM-1 protein expression and immunogenic functions to produce a less inflammatory phenotype in the absence of hemodynamic effects.

Introduction

Atherosclerosis, a chronic inflammatory disease, is the most frequent underlying cause of coronary, peripheral, and carotid artery diseases, which annually affect millions of people worldwide [1]. Developing as early as childhood, atherosclerosis is characterized by progressive lipid and monocyte accumulation within the arterial wall, resulting in the development of hardened plaques. During the initial phase of this progression, atherogenesis, inflammatory cells are recruited from circulating blood to the endothelium and subsequently migrate through the endothelium into the vessel wall. This migration process is predominantly mediated by surface adhesion molecules on endothelial cells (ECs), specifically the surface expression of E-selectin, intercellular adhesion molecule-1 (ICAM-1), and vascular cell adhesion molecule-1 (VCAM-1).

An essential component in atherogenesis is hemodynamic fluid shear stress (FSS) acting on the endothelial lined vessel wall. In response to FSS, ECs alter their morphology, phenotype, and cytoskeletal alignment via a variety of mechanoreceptors and mechanotransducers. Low, oscillatory FSS, which occurs within the vasculature at arterial bifurcations, branch points, and areas of high curvature, is known to upregulate EC expression of adhesion molecules and increased vascular permeability in pro-atherosclerotic regions [2]. Unsteady FSS in these regions fails to orient and elongate ECs, which maintain a cobblestone morphology and random cytoskeletal alignment [3,4]. In contrast, regions of high, pulsatile unidirectional FSS exhibit low adhesion molecule expression, alignment of cytoskeletal components, and elongated ECs in the direction of flow, resulting in an athero-protective phenotype. These spatiotemporal responses contribute to EC heterogeneity in vivo and influence regional susceptibility to atherogenesis.

EC responses to local flow regimes have been highly studied, and implicate the significant role of the cytoskeleton. Decentralization theory suggests that the cytoskeleton enables ECs to distinguish between flow patterns by the transmission of mechanical signals from the luminal surface via spatial distribution of stresses throughout the cell to intracellular sites mechanically connected to the cytoskeleton, which include focal adhesions, cell-to-cell junctions, and the nucleus [5]. This theory proposes that the dynamic cytoskeleton is a critical component in EC mechanotransduction, resulting in the spatial integration of biomechanical signals and suggests that cell shape and cytoskeletal structure may regulate EC function [6]. To investigate the cytoskeleton’s role in EC function, micropatterning methods have shown success in altering EC morphological metrics (whole cell elongation, alignment, and cytoskeletal organization) by a wide variety of techniques [7,8].

Micropatterning techniques, including topographic and spatial growth-restriction, enable a distribution of EC elongation in the absence of FSS. Previous studies using these methods have shown evidence of population-based VCAM-1 gene regulation by the cytoskeleton separate of hemodynamic effects [9–11], including a reduction of firm leukocyte binding to micropattern-elongated ECs. However, the studies performed were in the absence of a monolayer and were analyzed at the population-level. Thus, the relationship between physiologically relevant cytoskeletal elongation and immunogenic phenotype for a confluent endothelium at a single cell resolution remains unknown. The goal of this study was to investigate cytoskeletal dependence of VCAM-1 protein expression independent of FSS using topographic micropatterning and high-throughput image analysis techniques on a single cell basis. We hypothesized that an elongated morphology with an aligned cytoskeleton would downregulate VCAM-1 protein expression and, therefore, monocyte adhesion in the absence of flow. We quantified topographically aligned endothelial morphology, cytoskeletal alignment, and VCAM-1 protein expression using single cell image analysis techniques. Further, we correlated morphological metrics to protein expression to delineate which factors are major regulators of VCAM-1 protein expression. Lastly, we quantified static leukocyte adhesion to topographically elongated and planar endothelial monolayers.

Materials and methods

Substrate fabrication

Topographical micropatterned culture polyurethane substrates were fabricated and cleaned using an established protocol, as previously described [12]. Briefly, PDMS submasters with non-patterned, planar regions and parallel ridge and groove areas with a pitch of 6μm (pitch=ridge+groove) and a groove depth of 1μm were fabricated using soft lithography. Culture substrates were then fabricated by dispersing polyurethane on top of the PDMS patterns. 18×18mm glass coverslips (ThermoFisher) were then placed on top of the polyurethane and cured in an ultraviolet crosslinker (ThermoFisher) with 2J of long-wave light. Substrates were demolded and ethanol sterilized by ethanol and UV prior to cell seeding.

Cell culture

HUVECs (ATCC) at passage 6 were cultured in complete endothelial growth medium (VascuLife VEGF kit, Lifeline) with 10% FBS (Hyclone) concentration. Culture substrates were rinsed in Dulbecco’s phosphate buffered saline (PBS), incubated in rat tail collagen-I (50μg/mL, Corning) for at least 60min, and seeded with HUVECs at a density of 4.2×10⁴ cells/cm². HUVECs were cultured for 24hrs and then stimulated with TNF-α (100U/mL, R&D Systems) for another 24hrs before subsequent experiments. Culture substrates U937 cells (ATCC) were cultured in RPMI 1640 media (Gibco) supplemented with 10% FBS, 1% penicillin-streptomycin, and 2mM L-glutamine in ultra-low attachment surface flasks (Corning) up to 2×10⁶ cells/mL. All cells were incubated at 37°C and 5% CO₂.

Immunostaining and confocal microscopy

Immediately following experiments, cells were washed in warm PBS with Ca²⁺ and Mg²⁺ (Gibco) and fixed in warm 3.7% paraformaldehyde (PFA) for 15min. Cells were then stored in PBS with Ca²⁺ and Mg²⁺ at 4°C until further processing. HUVECs were permeabilized with 0.1% Triton-X100 and incubated in Image-iT FX signal enhancer (Invitrogen). For all staining, incubation with primary and secondary antibodies, as well as further blocking steps, were performed in the dark at room temperature. For morphology studies, the following sequence was used: 10% normal goat serum (ThermoFisher, 60min), VCAM-1 (1.4C3, Invitrogen, 1:100, 60min), anti-mouse IgG1 Alexa-488 (Invitrogen, 1:100, 60min), VE-cadherin/Alexa Fluor^® 594 (Santa Cruz, 1:100, 60min), and DAPI (Invitrogen, 1:10000, 5min). For cytoskeletal studies, the following sequence was used: Phalloidin (Invitrogen, 1:200, 60min), 10% normal goat serum (60min), VCAM-1 (1.4C3, 1:100, 60min), anti-mouse IgG1 Alexa-488 (1:100, 60 min), and DAPI (1:10000, 5min). Samples were mounted with ProLong Gold Antifade (Invitrogen) under glass coverslips (18×18×1.5mm, ThermoFisher) on frosted microscope slides (ThermoFisher).

Samples were imaged with a Zeiss LSM 880 inverted confocal microscope system (20X, NA=0.8; 10X, NA=0.45). Three randomly located images per pattern per replicate were collected for quantitative image analysis. Post-processing for publication was done in FIJI (SciJava).

Whole cell morphology quantification

Images were loaded into FIJI and color channels automatically split. A randomized 4×4 grid overlaying the VE-cadherin or F-actin channel was used to trace cells closest to grid points (n=6–12 cells/image). For each single cell tracing, cell area (A_s), perimeter (P), major and minor axis lengths, and major axis angle of a fit ellipse were measured. Cell morphology was determined by calculating shape index (SI), where 0 is equal to a perfect line and 1 is equal to a circle: SI = (4πA_S)/P². Deviance angle of the cell relative to the pattern was determined as the absolute value of the major axis angle minus the angle of the pattern. The deviance of planar cells was calculated using the pattern angle of the same substrate. Cell deviance angles of 0° is reported as perfectly aligned to the pattern and 90° as perpendicular to the pattern.

Cytoskeletal alignment quantification

Cytoskeletal alignment was determined on a single cell basis as previously described [12]. Single cell tracings were background subtracted using a rolling radius=50px, and locally autothresholded by means of the Midgrey autothreshold tool with radius=25px. Fibers were selected within the tracing bounds using the analyze particles tool with size≥2μm² and an SI≤0.30. Major axis angle of a fit ellipse and major lengths were measured for each selected fiber within the tracing bounds. The average fiber angle deviance was normalized for each cell by summing the product of each fiber angle and length and then dividing by the sum of the fiber lengths.

VCAM-1 protein expression quantification

The cells selected by the randomized grid pattern for whole cell morphology (n=6–12 cells/image) were analyzed for VCAM-1 expression. The protein expression was determined by measuring mean intensities within the regions of interest as defined by the cell perimeter. Intensities were corrected by background subtraction and normalized by individual cell area.

Correlations between morphological metrics and VCAM-1 expression

To identify on a single cell basis which morphological factors affected normalized VCAM-1 protein expression, linear correlations were determined as previously described [12]. Briefly, using single cell data, protein expression correlations were analyzed by linear regression between SI, whole cell deviance, actin fiber deviance, and VCAM-1 expression. Linear regression models were confirmed to be suitable by examining residual plots and normal quantile-quantile plots of the fitted values. Pearson’s correlation coefficients (R²) were used to determine the strength of the linear regressions for each substrate type, and probability values were used to determine the significance for each correlation.

Leukocyte static adhesion assay

U937 cells were labeled with 5μM CellTracker^™ Green CMFDA Dye (Invitrogen) for 30min according to manufacturer’s instructions. HUVECs were incubated after TNF-α stimulation with or without VCAM-1 primary antibody (100ng/mL) in serum-free culture media for 1hr. Fluorescently-labeled U937 cells were then co-cultured with HUVECs at 2×10⁶ U937 cells/substrate and allowed to statically adhere for 1hr. Cells were gently washed 3X with PBS with Ca²⁺ and Mg²⁺ to remove non-adhered monocytes and fixed in warm 3.7% PFA for 15mins. Substrates were mounted and imaged following the protocol above. Adhered leukocytes were manually counted and data is presented as the average U937 cells adhered per image field.

Statistics

All studies were performed in at least duplicate with sample sizes of n=3. Statistical analyses were performed with R (ver. 3.6.2) [13] by use of the package infer (ver. 0.5.2) [14]. Data are presented as mean±SEM for all studies unless otherwise noted. Histogram binwidths were determined using the Freedman-Diaconis rule. Data were analyzed using one-way ANOVA against substrate type or incubation condition, and Tukey’s HSD post-hoc test was subsequently performed where applicable. Statistical significance was determined for probability values<0.05.

Results

Topographic micropatterning effects on EC morphology and alignment

The effect of topographical micropatterning on EC elongation was determined by calculating SI for each cell. After 48hrs, ECs were successfully elongated and aligned on micropatterns independent of FSS (Fig. 1A). Planar controls had a rounded, cobblestone morphology with a SI of 0.69±0.15. Comparatively, micropatterned ECs had a more elongated morphology with a significantly decreased average SI of 0.52±0.02 (Fig. 1B, p<0.001). Planar cells also showed a random alignment orientation with a mean whole cell angle deviance of 40.25±3.16°. Micropatterned ECs had an average angle deviance of 12.87±1.22° relative to the underlying pattern (Fig. 1C, p<0.001). Topographical micropatterns also significantly aligned the actin cytoskeleton compared to planar controls, with average actin fiber angle deviances of 14.79±1.22° and 48.68±1.84°, respectively (Fig. 1D, p<0.001).

Figure 1:

Topographical micropatterning elongates and aligns individual ECs within an endothelial layer. (A) Representative 20X images of immunofluorescent staining of VE-cadherin (top) and F-actin (bottom) for each substrate after 48hrs of culture. Images are oriented such that 6μm micropatterned lanes are horizontal. Scale bars indicate 20μm. (B) Quantification of shape index of ECs on micropatterned and planar substrates. ECs are significantly more elongated on a single cell basis on micropatterns compared to planar substrates. Histograms summarize the distribution shape of shape indices for each substrate with a binwidth of 0.01. (C) Quantification of whole cell alignment as the angle deviance relative to the pattern. Micropatterned ECs are significantly more aligned compared to planar ECs. (D) Quantification of actin fiber alignment as deviance relative to the pattern. Actin fibers are significantly more aligned on a single cell basis for micropatterned ECs compared to planar ECs. Histograms summarize the distribution shape of the actin deviance for each substrate with a binwidth of 4. *** indicates statistical significance (p<0.001) compared to planar substrates.

Topographical micropatterning downregulates VCAM-1 protein expression

Planar (Fig. 2A) and micropatterned (Fig. 2B) single cell protein expression of VCAM-1 was determined using immunofluorescent staining and image analysis techniques. Planar substrates had an upregulated VCAM-1 protein expression with an average normalized VCAM-1 intensity of 2.10±0.33ABUs. Conversely, micropatterning had a significantly decreased protein expression, with an average normalized VCAM-1 intensity of 1.20±0.14ABUs (Fig. 2C, p=0.012).

Figure 2:

VCAM-1 protein expression. Representative 20X images of immunofluorescent staining of planar (A) and micropatterned (B) ECs of VE-cadherin, VCAM-1, and DAPI. Images are oriented such that patterns are horizontal. Note that polyurethane micropatterns are autofluorescent in the blue and green channels. Scale bars indicate 20μm. (C) Normalized VCAM-1 intensity quantification and density distribution on a single cell basis. * indicates statistical significance (p<0.05) compared to planar substrates.

VCAM-1 protein expression is correlated to morphological metrics

Image analysis quantification of ECs co-stained for VE-cadherin and VCAM-1 showed that elongation and alignment have a relationship with VCAM-1 protein expression. There was a weak but significant relationship between cell elongation and VCAM-1 in micropatterned ECs (Fig. 3A, R²=0.294; p=0.02). Conversely, this relationship was insignificant in planar ECs (R²=0.161; p=0.22). Examination between whole cell alignment and VCAM-1 protein expression also showed a weak but significant relationship for micropatterned cells (Fig. 3B, R²=0.276; p=0.03) but an uncorrelated relationship in planar cells (R²=−0.111; p=0.40). Image analysis quantification of endothelial monolayers co-stained for F-actin and VCAM-1 showed a strong, significant relationship between cytoskeletal alignment for micropatterned (Fig. 3C, R²=0.834; p<0.001) but not planar ECs (R²=−0.092; p=0.40). Examination also revealed a strong, significant correlation between actin and whole cell deviance between both micropatterned (Fig. 3D, R²=0.836; p<0.001) and planar ECs (R²=0.859; p<0.001).

Figure 3:

Relationships between VCAM-1 protein expression and morphological metrics. There are weak relationships between single cell morphological metrics (SI and deviance), while there exists strong relationships between single cell actin fiber alignment and both VCAM-1 expression and whole cell deviance. (A) Relationship between VCAM-1 protein expression and whole cell elongation. (B) Relationship between VCAM-1 protein expression and whole cell alignment. (C) Relationship between VCAM-1 protein expression and actin fiber alignment. (D) Relationship between whole cell and actin fiber alignment. All panels: blue: planar, red: micropatterned. Pearson correlations indicated in red and blue as appropriate.

Micropatterning reduces monocyte adhesion to ECs

Because VCAM-1 is a mediator of the firm adhesion of leukocytes to the endothelial monolayer during atherogenesis, the effect of micropatterning on static leukocyte adhesion was quantified (Fig. 4A). Planar substrates treated with TNF-α had an average of 49.43±3.40 U937 cells adhered per image field. Comparatively, U937 cell adhesion was significantly decreased to 18.96±1.74 cells adhered per image field for micropatterned substrates treated with TNF-α (p<0.001). Blocking by incubation with anti-VCAM-1 primary antibody almost completely inhibited U937 cell adhesion on both planar and micropatterned substrates, which had averages of 7.67±1.71 (p<0.001) and 5.44±0.59 (p=0.0018) cells per field, respectively (Fig. 4B).

Figure 4:

Monocyte adhesion under static conditions. (A) Fluorescently-labeled U937 cells adhered under static conditions to ECs on planar and micropatterned substrates. Representative 10X images are shown for each substrate under TNF-α stimulation with or without incubation with VCAM-1 antibody for blocking studies. Scale bars indicate 20μm. (B) Fluorescently-labeled U937 cells were quantified per image field. U937 adhesion was significantly reduced on micropatterned monolayers compared to planar. Blocking with VCAM-1 antibody also significantly reduced adhesion for each substrate. ** and *** indicate statistical significance (p<0.01, 0.001 respectively). No significant difference was observed between substrates under blocking conditions (p=0.13).

Discussion

To study the role of morphological metrics independent of FSS on VCAM-1 protein expression, topographical micropatterning was utilized to generate distributions of EC elongation and alignment as well as cytoskeletal alignment. We showed on a single cell basis that elongated ECs with an aligned cytoskeleton significantly downregulate VCAM-1 protein expression in comparison with cobblestone ECs with a random cytoskeletal alignment. Linear correlations between morphological metrics and VCAM-1 protein expression showed that cell elongation and alignment had a weak relationship to protein expression, but actin alignment had a strong relationship, indicating a critical role of the cytoskeleton in VCAM-1 protein regulation. Downregulation of VCAM-1 also resulted in decreased static leukocyte binding, indicating the physiological impact of an aligned cytoskeleton.

While EC morphology and phenotype under FSS has been widely studied [3,6,15–17], micropatterning, including biochemically-restricted and topographic techniques, has emerged as a successful tool to alter and control EC morphology and function independent of flow [8]. Micropatterning via biochemical restriction has been previously performed to investigate the effects of cell spreading, shape, and cell-to-cell contact and adhesion on EC function [7,18]. This biochemical technique is used to form adhesive islands or lanes (typically 10–100μm wide) with one or very few cells attached. This lack of a confluent endothelial layer leaves ECs a limited surface area to form intercellular junctions, which decreases intracellular tension [19], and ultimately leaves the entire side of the cells freely exposed to the growth-restricted area of a substrate. In contrast, the use of topographical micropatterning with subcellular features (<10μm) allows the formation of a confluent endothelium, allowing for the physiologically relevant formation of cell-to-cell junctions and interactions. Topographic substrates have been shown to orient ECs by localizing focal adhesion proteins along the edges of micropatterned ridges [20]. Overall, topographical micropatterning both aligns and elongates ECs to the direction of the pattern but allows cells to freely spread, enabling a distribution of morphology on a single substrate. This study found that topographic micropatterning produces a significantly more elongated and aligned EC morphology in addition to cytoskeletal alignment, which supports our previous work [12].

High, pulsatile unidirectional FSS induces an elongated EC morphology and aligns cells to the direction of flow in vivo. These regions generally have an athero-protective EC phenotype, which is characterized by the downregulation of immunogenic protein expression and decreased expression of VCAM-1 [21–23]. Previous micropatterning studies have investigated the effect of cell morphology on VCAM-1 gene expression [9–11]. It has been shown that TNF-α - stimulated ECs elongated by biochemical restriction downregulate VCAM-1 mRNA expression compared to cobblestone ECs [9,11]. In addition, these studies showed that the addition of FSS further attenuated mRNA expression. This suggests that cytoskeletal alignment synergistically acts with FSS to regulate VCAM-1 gene expression. However, these gene expression studies were performed at the population-level and in the absence of a confluent monolayer. To gain further insight into VCAM-1 regulation, we investigated if topographical micropatterning of a confluent endothelial monolayer reduces single cell VCAM-1 protein expression. In this study, we sought to determine the effects of the individual cell and cytoskeletal morphological metrics on VCAM-1 protein expression independently. We found that topographical micropatterning downregulated VCAM-1 protein expression in comparison to planar ECs. This suggests that an elongated morphology and aligned cytoskeleton hold a role in VCAM-1 protein regulation independent of FSS, which corresponds well with the previous studies of downregulated VCAM-1 gene expression.

Increasing evidence suggests that neighboring ECs are highly heterogeneous and do not respond as a uniform population to control vascular function [24]. By utilizing image analysis techniques at the single cell level, cell-to-cell variation within a population can provide further context into biological pathways. We showed that micropatterning significantly downregulates VCAM-1 protein expression in TNF-α -stimulated ECs. However, we saw a high variability of expression on a single cell basis. This led to us to investigate whether cell and cytoskeletal morphological metrics were correlated with VCAM-1 protein expression using single cell image analysis techniques. While whole cell elongation and alignment had a weak, but significant, relationship with VCAM-1 expression, we found that actin alignment was strongly correlated for micropatterned ECs. This suggests that actin alignment plays a greater role than cell elongation in the regulation of VCAM-1. We found no correlation between planar cells, possibly because there were few with cytoskeletal alignment between 0–15°. Our findings support the theory of decentralization, which suggests that increased intracellular tension induced by an aligned cytoskeletal structure and increased intercellular junction points is critical to regulating various mechanotransduction pathways.

VCAM-1 is critical in facilitating leukocyte recruitment and adhesion to the vessel wall and is the first adhesion molecule expressed before atherosclerotic plaques are formed and is a known indicator of vascular inflammation [25]. Because we saw a downregulation in VCAM-1 protein expression in our studies, we investigated the effect of micropatterning on the static adhesion of U937 cells, which are leukocytes that express the integrin Very Late Antigen-4 (VLA-4), the VCAM-1 binding ligand. TNF-α-stimulated ECs were incubated with or without a monoclonal VCAM-1 antibody to determine the specificity of the monocyte-EC interaction. Micropatterned substrates had a significantly decreased amount of U937 cells adhered compared to planar controls. This data is consistent with protein expression data where VCAM-1 was downregulated for micropatterned ECs compared to planar ECs. In addition, the blocking of VCAM-1 led to a near inhibition of U937 cells adhesion, implicating VCAM-1 as the main contributor to firm leukocyte adhesion on these planar and micropatterned ECs.

While this study investigated the flow-independent effects of EC morphological metrics and cytoskeletal alignment on VCAM-1 expression, several limitations should be addressed in future studies. Even though the absence of flow is itself an injury model, the additive impact of FSS on topographic endothelial monolayers would be informative to study the combined effects. Future studies should quantify SI, alignment, and immunogenicity of topographically micropatterned ECs under laminar and disturbed flow patterns in order to address whether the cytoskeletal-dependent regulation of VCAM-1 acts synergistically or independent of hemodynamic effects within a monolayer. Additionally, the role of microtubules should be investigated to determine the individual roles of cytoskeletal regulation of VCAM-1 expression.

This study demonstrated for the first time that on a single cell basis topographical micropatterning alters EC protein expression involved in atherogenesis. By performing single cell image analyses, we were able to correlate the relationships between morphological metrics and VCAM-1 protein expression independent of flow. In addition, this study demonstrated that micropatterning induced downregulation of VCAM-1 was sufficient to significantly decrease leukocyte binding. Taken together, these results demonstrate that endothelial VCAM-1 expression can be regulated by cytoskeletal alignment with topographical micropatterning of a complete endothelial layer. This work further delineates the role that the cytoskeleton plays in EC mechanotransduction of critical inflammatory markers in the initiation of cardiovascular diseases.

Highlights.

• Topographies induce endothelial monolayer to elongate with cytoskeletal alignment

• Cellular elongation and alignment are independent of fluid flow

• Single cell image analysis describe endothelial cell heterogeneity within monolayer

• Single cell analyses correlate morphological metrics to VCAM-1 protein expression

• Topography aligned cytoskeletons reduces cell immunogenicity