Endothelial Glycocalyx-Mediated Nitric Oxide Production in Response to Selective AFM Pulling

Abstract

Nitric oxide (NO) is a regulatory molecule in the vascular system and its inhibition due to endothelial injury contributes to cardiovascular disease. The glycocalyx is a thin layer of glycolipids, glycoproteins, and proteoglycans on the surface of mammalian epithelial cells. Extracellular forces are transmitted through the glycocalyx to initiate intracellular signaling pathways. In endothelial cells (ECs), previous studies have shown the glycocalyx to be a significant mediator of NO production; degradation of the endothelial glycocalyx layer (EGL) drastically reduces EC production of NO in response to fluid shear stress. However, the specific EGL components involved in this process are not well established. Recent work using short-hairpin RNA approaches in vitro suggest that the proteoglycan glypican-1, not syndecan-1, is the dominant core protein mediating shear-induced NO production. We utilized atomic force microscopy (AFM) to apply force selectively to components of the EGL of confluent rat fat pad ECs (RFPECs), including proteoglycans and glycosaminoglycans, to observe how each component individually contributes to force-induced production of NO. 4,5-diaminofluorescein diacetate, a cell-permeable fluorescent molecule, was used to detect changes in intracellular NO production. Antibody-coated AFM probes exhibited strong surface binding to RFPEC monolayers, with 100-300 pN mean adhesion forces. AFM pulling on glypican-1 and heparan sulfate for 10 min caused significantly increased NO production, whereas pulling on syndecan-1, CD44, hyaluronic acid, and with control probes did not. We conclude that AFM pulling can be used to activate EGL-mediated NO production and that the heparan sulfate proteoglycan glypican-1 is a primary mechanosensor for shear-induced NO production.

Figure 1

AFM pulling and AFM probe functionalization. Cartoons are not drawn to scale. (A) Cell-surface core proteins (syndecan-1, glypican-1, and CD44) and their associated GAGs (HS and HA) were stimulated with corresponding specific antibodies. (B) An AFM scanner stage is implemented to control XYZ-positioning and the deflection of a triangular cantilever is used to monitor indentation depth and pulling-off forces. Glass gridded slides were used to track location of cells probed with AFM for later imaging. Unprobed regions were imaged ∼2 mm away from the probing site. After background correction, percent activation was calculated by comparing the MFI of the probed region to the MFI of the unprobed regions. (C) Silicon nitride cantilevers are functionalized with amine groups, PEG linker, and antibody (Antibodies). Control functionalization schemes include isotype control using normal IgGs (included in Antibodies scheme), PEG linker with varying degrees of active termini (Linker), and bare (None) probes. To see this figure in color, go online.

Figure 2

Adhesion forces for functionalized pyramidal AFM probes on RFPEC monolayers. Mean adhesion forces ± SE values are shown for proteoglycans and CD44 in purple, for GAGs in blue, and for control probes in gray; n ≥ 60 indentations, with 20 indentations/cell and three to seven cells probed in each group. The asterisk indicates a significant difference (p < 0.01), using the two-tailed t-test for antibody-functionalized probes versus their isotype controls, HA probes and 1 M EA linker-sham probes versus a bare probe (none). The right panel displays the pulling-off region of a representative anti-glypican-1 force-distance curve during tip retraction (for complete force-distance curves, see Fig. S1B). The separation curve can include the unbinding of several individual bonds, but adhesion force here measures the maximal detachment force between the cell surface and AFM probe. To see this figure in color, go online.

Figure 3

Probing RFPEC with glypican-1 and HS antibodies initiates NO production. Functionalized AFM probes were used to pull on the indicated structures for 10 min. Mean DAF-2 T MFIs from probed regions were normalized by those from unprobed regions on the same slide and converted to activation percentages (mean ± SE), with unstimulated regions indicated by 0%. ^∗p < 0.05 compared to paired unstimulated regions. From left to right, n = 28, 20, 20, 22, 21, 16, 13, 11, and 19. The dotted line shows mean activation (n = 12) ± SE for RFPECs exposed to 20 dynes/cm² shear stress for 10 min, which is significantly different from static regions at 0% activation. Representative DAF-2 T fluorescent images are shown at the left for probed and unprobed regions of the same slide. The scale bar represents 20 μm. To see this figure in color, go online.

Figure 4

NO induced by 10 min shear stress is not affected by HA removal in RFPECs. DAF-2 T MFIs normalized by static control MFIs are shown as the mean ± SE (left to right: n = 14, 15, 14, and 18). ^∗p < 0.05 compared to paired static control. 20 dynes/cm² of shear stress was applied to shear samples for 10 min. To see this figure in color, go online.