Endothelial CD44 cleavage by ADAM17 impairs shear stress mechanotransduction

Abstract

In individuals with type 2 diabetes (T2D), blood flow-mediated increases in endothelial shear stress fail to elicit a robust vasodilatory response. This defective flow-mediated dilation (FMD) is associated with loss of the endothelial glycocalyx, a mechanosensitive extracellular structure that lines the luminal side of blood vessels. Hyaluronan (HA), a polysaccharide constituent of the glycocalyx, is anchored to the plasma membrane by its primary cell-surface receptor, CD44, which is implicated in mechanotransduction of shear stress. Critically, CD44 is also a substrate of a disintegrin and metalloproteinase 17 (ADAM17), a sheddase that is elevated in T2D. However, it is currently unknown whether elevated ADAM17 activity enhances CD44 cleavage from the endothelium and whether this contributes to impaired mechanotransduction and reduced FMD in T2D. Herein, we report elevated plasma HA and ADAM17 activity in a cohort of women and men with T2D and impaired FMD. Moreover, reduced endothelial CD44 is coupled with impaired FMD in arteries isolated from diabetic (db/db) mice. We also provide support for CD44 as a mechanotransducer of HA-associated shear stress mechanosensation and ADAM17-mediated cleavage of CD44 attenuating shear stress mechanotransduction. Finally, using an in-vitro assay and surface plasmon resonance, we show that active recombinant human ADAM17 (ADAM17-r) cleaves recombinant human CD44 and that intraluminal incubation of isolated arteries with ADAM17-r reduces FMD. Collectively, this work supports the role of ADAM17-mediated cleavage of CD44 in impairing endothelial shear stress mechanotransduction.

NEW & NOTEWORTHY Impaired flow-mediated dilation (FMD) is an indicator of endothelial dysfunction in type 2 diabetes (T2D). However, the exact cellular and molecular mechanisms contributing to impaired mechanotransduction of shear stress in T2D are incompletely understood. This work supports the novel concept that increased endothelial activity of a disintegrin and metalloproteinase 17 (ADAM17) causes cleavage of cell-surface CD44, leading to loss of hyaluronan (HA)-associated shear stress mechanotransduction.

INTRODUCTION

Type 2 diabetes (T2D) is associated with the development of endothelial dysfunction characterized by an impaired ability to detect and transduce blood flow-mediated shear stress. Loss of shear stress mechanotransduction has been attributed to the degradation of the endothelial glycocalyx, a heterogeneous and hair-like structure lining the apical side of endothelial cells and a putative mechanosensor of flow (1–4). Notably, the endothelial glycocalyx is diminished in T2D, and this is associated with impaired mechanotransduction and reduced flow-mediated dilation (FMD) (5). Moreover, the glycocalyx is also implicated in regulating the redox environment (6, 7). Tempol, a superoxide dismutase mimetic, rescues FMD following selective enzymatic removal of glycocalyx components like heparan sulfates and sialic acids, but not in the case of hyaluronan (HA) (1). This suggests that HA is likely an important mechanosensor of shear stress (8). Unlike other components of the glycocalyx, HA is an anionic, nonsulfated, and nonsialylated glycosaminoglycan synthesized at the plasma membrane (9, 10). It is secreted onto the cell surface, where it forms a dense, honeycomb-like web close to the plasma membrane (9–11). However, HA is not directly attached to the cell surface; rather, it is tethered to the membrane by its primary cell surface receptor, CD44 (12, 13). Moreover, CD44 has recently been reported to mediate mechanotransduction in response to shear stress (14, 15). Accordingly, we posit that HA serves as a shear stress mechanosensor (i.e., detects shear stress), whereas CD44 is the mechanotransducer (i.e., transmits the shear stress signal). In support of this, endothelial cells deficient in CD44 exhibit impaired shear stress mechanotransduction, whereas blockade of CD44’s HA-binding site reduces FMD (14, 15).

Importantly, CD44 is also a substrate of a disintegrin and metalloproteinase 17 (ADAM17), a proinflammatory membrane-bound enzyme that regulates diverse signaling pathways by cleaving transmembrane proteins from the cell surface (16). We have previously shown that ADAM17 expression and activity are increased in the vascular wall and plasma, respectively, of individuals with T2D (17, 18). Moreover, pathways regulating ADAM17 activity are also enhanced in T2D, including increased intracellular Ca²⁺ mobilization and protein kinase C (PKC) activation (16, 18). In this regard, we have provided evidence that circulating factors elevated in T2D (e.g., neuraminidase) increase Ca²⁺-dependent endothelial ADAM17 activation (18), whereas other investigators have demonstrated enhanced PKC expression in endothelial cells biopsied from individuals with T2D (19). Collectively, this prior work supports the notion that vascular ADAM17 activity is elevated in T2D.

However, it is currently unknown whether enhanced endothelial ADAM17 activity causes cleavage of CD44, limiting endothelial sensitivity to shear stress in T2D. Given that CD44 is a substrate for ADAM17 in nonendothelial cells, and ADAM17 activity is elevated in T2D, it is conceivable that elevated endothelial ADAM17 activity contributes to impaired FMD in T2D by cleaving CD44. Herein, we tested the hypothesis that elevated endothelial ADAM17 activity causes enhanced cleavage of endothelial CD44, impairing HA-associated mechanotransduction of shear stress.

MATERIALS AND METHODS

Ethics and Approvals

All human study procedures were conducted in accordance with the Declaration of Helsinki and approved by the University of Missouri Institutional Review Board (IRB, No. 2008181, 2012106, 2038203). Written informed consent was obtained from all participants before their participation in the study. All animal study procedures received prior approval by the University of Missouri Animal Care and Use Committee and were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Experimenters were blinded when possible.

Human Participants

Data from 72 individuals with T2D (31 females and 41 males, age = 56.4 ± 1.1 yr, body mass index = 35.0 ± 0.7) and 33 age-matched participants without T2D (20 females and 13 males, age = 52.6 ± 2.3 yr, body mass index = 23.3 ± 0.4) were included in retrospective analyses. Exclusion criteria have been described elsewhere (5, 20, 21). However, for these retrospective analyses, only individuals ≤65 yr of age were included. This was to prevent an age difference between the two groups. Femoral artery FMD was assessed after an overnight fast, as previously described (20) and according to published guidelines (22). The femoral artery was chosen because it is a muscular artery that is highly responsive to shear stress and susceptible to dysfunction and disease. It should be noted that some participants included in these retrospective analyses, which include new data documenting the time course FMD and peak FMD relative to the shear rate stimulus (Fig. 1, A–B), were also part of previous publications addressing different research questions and in which FMD data were expressed in a different format (5, 20, 23).

Figure 1.

Reduced flow-mediated dilation (FMD) is associated with increased ADAM17 activity, lower endothelial CD44 content, and elevated plasma hyaluronan levels in T2D. A: femoral artery FMD time-course (0–180 s) profile following cuff deflation in humans (non-T2D, n = 19, women/men = 12/7; T2D, n = 26, women/men = 8/18). B: femoral artery peak FMD expressed relative to shear rate AUC (× 10³ a.u.) until peak dilation from the same subjects as in A. C: human plasma ADAM17 activity (non-T2D, n = 20, women/men = 12/8; T2D, n = 19, women/men = 12/7). D: human plasma concentration of hyaluronan (non-T2D, n = 32, women/men = 16/16; T2D, n = 51, women/men = 28/24). E: cannulated and pressurized mesenteric arteries isolated from db/db and C57BL/6J [wild-type (WT)] male mice were preconstricted with phenylephrine and exposed to increasing intraluminal flow rates to induce FMD or vasodilatory responses to sodium nitroprusside (SNP) (n = 4 animals/group). F: positive fluorescent area of endothelial ADAM17 in aortic en face preparations from WT vs. db/db male mice (n = 5 animals/group), with representative images; scale bar = 50μm. G: CD44 expression in lung endothelial cells isolated from WT vs. db/db male mice (n = 9 or 10 animals/group), with representative blots. All images were enhanced equally for visualization purposes only, and all analyses were performed using the raw data. Data are presented as means ± SE. Women are represented by triangles and men by squares (C and D). Two-way ANOVA or mixed-effects analyses with repeated measures were used to assess FMD and SNP-induced vasodilation in human femoral and mouse mesenteric arteries (A and E). Mann–Whitney test was used to determine differences between plasma ADAM17 activity (C), and two-tailed unpaired Student’s t tests were used in all other analyses. *P ≤ 0.05 vs. respective controls.

Animal Studies

C57BL/6J (wild-type) and db/db (a mouse model of T2D) female and male mice on the same background were obtained from Jackson Laboratory. Mice with ad libitum access to water were fed standard chow (5053-PicoLab Rodent Diet 20, LabDiet) and kept under 12 h:12 h dark/light cycles. Mice (12–14 wk old) were euthanized via CO₂ inhalation in accordance with AMVA guidelines, followed by cervical dislocation and exsanguination. Arteries were excised for microscopy and functional testing, and lungs were excised for endothelial cell isolation. Muscular mesenteric resistance arteries were chosen for FMD experiments because they are well-established to respond to shear stress and because there are multiple arteries per animal, which facilitates within animal treatments. Mouse lung endothelial cells were studied because of the technical challenge of isolating endothelial cells from mouse conduit arteries and the availability of commercially available mouse lung endothelial isolation kits (Miltenyi Biotec., Bergisch Gladbach, Germany). In brief, CD45⁺ cells were removed from a lung cell suspension following dissociation using a gentleMACs Octo dissociator (Miltenyi Biotec.). CD31⁺ cells were then magnetically separated and lysed in RIPA buffer (Invitrogen, Carlsbad, CA), EDTA (5 mM), and protease/phosphatase inhibitors (Invitrogen). Sample sizes were selected based on previous similar experiments conducted in our laboratory. All current and future reagents, stains, and antibodies are detailed in the Supplemental Material.

Endothelial Cell Culture Experiments

Human umbilical vein endothelial cells (HUVECs) from pooled donors (Lonza, Morristown, NJ) were used in all cell culture experiments because these cells are well-established to respond to shear stress, are easy to genetically manipulate, and are relatively consistent batch-to-batch. HUVECs were cultured under standard conditions in complete VascuLife EnGS endothelial media (Lifeline Cell Tech., Frederick, MD). To assess the role of CD44 in mechanotransduction of shear stress, cells were transfected with an siRNA targeting CD44 (20 nM; Dharmacon Reagents, Cambridge, United Kingdom) using Lipofectamine RNAiMAX transfection reagent (Invitrogen) complexed in Opti-MEM medium (Gibco, Grand Island, NY). To test the effects of pharmacological activation of ADAM17, cells were stimulated with phorbol 12-myristate 13-acetate (PMA,1 μM; Sigma-Aldrich, St. Louis, MO), or vehicle control (0.05% DMSO), for 15 min or 24 h. To assess the effects of ADAM17 overactivity, cells were transduced with ADAM17 or CMV-null (Ad.hADAM17, and Ad-CMV-null, Vector biolabs, Devault, PA) viral particles at a multiplicity of infection of 100. After 24 h, the media were replaced with fresh, virus-free media to remove unbound virus, and cells were cultured for an additional 24 h prior to experimentation. To determine the effects of active recombinant human ADAM17 (ADAM17-r, fragment predicted 52 kDa; R&D Systems, Minneapolis, MN), cells were exposed to ADAM17-r (1 μg/mL) or heat-inactivated ADAM17-r (1 μg/mL, 100°C for 30 min).

In-Vitro Shear Stress Mechanotransduction

Different approaches were used to evaluate shear stress mechanotransduction in HUVECs. As a general assessment of endothelial cell mechanotransduction, endothelial nitric oxide synthase (eNOS) expression was quantified in cells cultured on Ibidi μ-Slide VI 0.4 Luer channels (Ibidi, Fitchburg, WI) after exposure to 24 h of shear stress (15 dyn/cm² vs. static conditions) and fixation in 4% paraformaldehyde for 15 min. Upregulation of eNOS in response to shear stress is a well-established phenomenon (24, 25). More closely mimicking FMD, acute responses to shear stress were also measured in a live-cell assay. Cells cultured in Ibidi μ-Slide I 0.6 Luer channels (Ibidi) were loaded with Fluo-4 AM dye (Abcam, Cambridge, United Kingdom), a cell-permeable intracellular Ca²⁺ indicator, before a static time-control period (−300 to 0 s), which was followed by exposure to shear (15 dyn/cm²; 0–300 s). In both approaches, cells were exposed to flow using a computer-controlled flow device using an air-pressure pump and two-way switching valve to generate shear stress of 15 dyn/cm² (Ibidi). Finally, to determine eNOS activation in response to shear stress, cells were cultured on microscopy slides and exposed to 15 dyn/cm² of shear stress, using a parallel-plate flow system (Streamer; Flexcell, Burlington, NC) for 2 h. Cells were then collected in RIPA buffer for protein determination of p-eNOS/eNOS.

Assessment of FMD in Isolated Arteries

To demonstrate the impact of T2D on mechanotransduction of shear stress in isolated arteries, FMD was assessed in 2nd order (diameters: 220–280 μm) mesenteric arteries isolated from C57BL/6J and db/db mice, as previously described (23). In brief, isolated arteries were cannulated and pressurized at 70 mmHg before preconstriction with phenylephrine (10⁻⁶ M) and exposure to incremental increases in intraluminal flow (from 0.0 to 3.3 mL/h, with 0.3 mL/h increments every 2 min). To determine the effects of HA degradation, CD44-associated HA mechanosensation, and elevated ADAM17 activity on endothelial mechanotransduction of shear stress, FMD was assessed in mesenteric arteries isolated from C57BL/6J mice. Specifically, arteries from the same animal were exposed intraluminally to the following conditions (treatment vs. vehicle control; all intraluminal incubations were for 1 h; Fig. 2F) before FMD in three independent experiments: 1) hyaluronidase (15 μg/mL; Sigma-Aldrich) or heat-denatured (15 μg/mL; 1 h at 100°C) hyaluronidase, 2) anti-CD44 (KM201) antibody (0.25 mg/mL; Southern Biotech, Birmingham, AL) or IgG isotype control (0.25 mg/mL; Southern Biotech.), 3) active ADAM17-r (1 μg/mL; R&D Systems) or heat-inactivated ADAM17-r (1 μg/mL; 1 h at 100°C). Arteries were then exposed to cumulative concentrations of sodium nitroprusside (SNP; 10⁻⁸–10⁻⁴ M) to evaluate endothelium-independent vasodilation. Only vessels that constricted >30% to 80 mM potassium chloride before experimentation were analyzed.

Figure 2.

CD44 is a mechanotransducer of hyaluronan-associated shear stress sensing and flow-mediated dilation (FMD). A: cannulated and pressurized mesenteric arteries isolated from the same C57BL/6J female or male mouse exposed intraluminally to either hyaluronidase (n = 9 paired vessels/group, females/males = 5/4; 15 μg/mL, 1 h) or heat-denatured hyaluronidase (n = 6 paired vessels/group, females/males = 3/3; heat-treated at 100°C for 1 h before intraluminal exposure) were preconstricted with phenylephrine and exposed to increasing intraluminal flow rates to induce FMD or to sodium nitroprusside (SNP). B: CD44 protein expression following siRNA-mediated knockdown (n = 12 replicates/condition), with representative blots. C: flow-induced changes in intracellular Ca²⁺ in human umbilical vein endothelial cells (HUVECs) deficient in CD44 (15 dyn/cm², 0–300 s; n = 3 or 4 replicates/condition), with representative fluorescent images of Ca²⁺ following siRNA-mediated knockdown; scale bar = 200 μm. D: endothelial nitric oxide synthase (eNOS) expression in HUVECs following siRNA-mediated knockdown of CD44 and exposure to 24 h of 15 dyn/cm² shear stress (n = 4 replicates/condition), with representative images; scale bar = 100 μm. E: FMD and SNP-induced dilation in mesenteric arteries isolated from the same C57BL/6J female or male mouse exposed intraluminally to either anti-CD44 (KM201) (n = 8 paired vessels/group, females/males = 4/4; 0.25 mg/mL, 1 h) or IgG control (n = 7 paired vessels/group, females/males = 3/4; 0.25 mg/mL, 1 h). F: timeline of the standard protocol used for ex vivo functional testing experiments in mesenteric arteries. All images were enhanced equally for visualization purposes only, and all analyses were performed using the raw data. Data are presented as means ± SE. Two-way ANOVA or mixed-effects analyses with repeated measures were used to assess FMD and SNP-induced vasodilation in mesenteric arteries and flow-induced changes in intracellular Ca²⁺ in HUVECs (A, C, and E). Two-way ANOVA with Bonferroni’s post hoc was used to assess eNOS expression in CD44-deficient cells exposed to shear stress (D). Two-tailed unpaired Student’s t tests were used in all other analyses. *P ≤ 0.05 vs. respective controls and main effect of condition; #P ≤ 0.05 vs. shear stress (D).

Cell-Free Cleavage Assay

To demonstrate the specificity of ADAM17 for cleaving CD44, we performed an in vitro, cell-free cleavage assay. Recombinant human CD44 (CD44-r, full-length theoretical 66 kDa; 6.25 ng/μL; Abnova, Taipei, Taiwan) was incubated with ADAM17-r (62.5 ng/μL) or heat-inactivated ADAM17-r in the presence or absence of TAPI-0 (50 μM). Mixtures were incubated at 37°C for 8 h before addition of Laemmli, preparation for Western blot, and determination of protein band migration (indicative of cleavage).

Surface Plasmon Resonance

To demonstrate that ADAM17 cleaves CD44 and reduces binding to HA, molecular binding was measured using surface plasmon resonance (SPR) (OpenSPR, Nicoya, Kitchener, Canada). To control orientation of recombinant proteins, high-capacity carboxyl sensors (Nicoya) were functionalized with protein-A (40 ng/μL; Nicoya), which then bound the fc region of recombinant human CD44-fc chimera (CD44-r-fc; 5 ng/μL; R&D systems). Protein-A (Nicoya) was coupled to the sensor using sodium acetate (pH = 5). Ethanolamine (1 M; pH 8.5, Nicoya) was used to block unreacted sites on the sensor. Degassed and prefiltered PBS-T was selected as the solvent for CD44-r-fc and hyaluronan (HA; 100 kDa; 10 μM; Lifecore biomedical, Chaska, MN) and used as the running buffer. Before its injection, CD44-r-fc was incubated with active (5 ng/μL; R&D systems) or inactive ADAM17-r (5 ng/μL; 30 min at 100°C) for 8 h, and component C (1.85 μL; Anaspec, Fremont, CA) was added to the reaction to promote the activity of the enzyme. CD44-r-fc (± exposure to active ADAM17-r) binding to protein-A was evaluated before assessment of HA binding to CD44-r-fc. Both experiments (CD44-r-fc ± exposure to active ADAM17-r) were performed on the same high-capacity carboxyl sensor (n = 4).

ADAM17 Activity and Plasma HA Measurements

ADAM17 activity was measured in human plasma from a subset of participants and HUVECs using a SensoLyte TACE activity kit (AnaSpec), as previously described (18). Concentration of soluble HA in human plasma from a subset of participants was measured using an ELISA kit (R&D Systems).

Immunofluorescent Microscopy

Immunofluorescent images of HUVECs were obtained from cells seeded on Ibidi 15-well 3D and I 0.6 Luer channel μ-Slides following 24 h of shear stress or static conditions (Ibidi). Cells were fixed in 4% PFA and permeabilized with 0.1% Triton X-100 for 15 min (Triton X-100 was excluded when assessing CD44). Cells were blocked for 1 h in 10% goat serum followed by overnight incubation at 4°C in primary antibodies: CD44 specific to the extracellular (N-terminal) domain (Cell-surface CD44; 1:100; Abcam), ADAM17 (1:100; LSbio, Saratoga, CA; 1:100; Abcam), and eNOS (1:100; BD transduction, Franklin Lakes, NJ). Following overnight incubation, cells were washed in PBS and incubated with anti-mouse Alexa Fluor 488 or 647 (1:300-1:500; Invitrogen) and anti-rabbit Alexa Fluor 488 or 647 (1:300–1:500; Invitrogen). Nuclei were identified using DAPI (1:1,000; Sigma-Aldrich). In live-cell Fluo-4 AM experiments, temporal changes in intracellular Ca²⁺ fluorescence intensity were quantified from images taken every 15–30 s for 5 min. Widefield epifluorescent images were acquired with a Leica DMi8 automated microscope, with THUNDER imaging technology (Leica Microsystems, Inc., Morrisville, NC) using ×40/0.8 NA air or ×20/0.4 NA air objective. Automatically determined regions were imaged in triplicate per sample. Fluorescent intensity was quantified using Imaris software (Bitplane, Inc., Concord, MA) and normalized by cell number. Data presented are an average of the triplicate images.

For immunofluorescent assessment of isolated mouse aortic en face preparations, aortic rings were fixed in 4% PFA before being opened longitudinally and permeabilized in 0.5% Triton X-100 for 1 h. En face preparations were then blocked for 1 h in 5% goat serum and incubated overnight in a primary antibody for ADAM17 (1:200; LSbio). Confocal images were acquired with a Leica Stellaris 8 using a ×25/0.95 NA water immersion objective. Images were analyzed using Imaris software.

Western Blot

Protein expression was assessed in HUVEC lysates and isolated mouse endothelial cells prepared in RIPA buffer, EDTA (5 mM), and protease/phosphatase inhibitors. Proteins within samples were separated in Criterion Tris-Glycine eXtended-PAGE precast-gels (Bio-Rad, Hercules, CA) and transferred onto polyvinylidene difluoride membranes. Specific proteins were probed using the following primary antibodies: CD44 (1:1,000; Cell Signaling Technology, Danvers, MA), ADAM17 (1:1,000; Abcam), phosphorylated eNOS Ser¹¹⁷⁷ (p-eNOS^Ser1177; 1:1,000; Abcam), and eNOS (1:1,000; Cell Signaling Technology). Secondary antibodies used were horseradish peroxidase (HRP)-conjugated goat anti-rabbit (1:5,000; Bio-Rad) or Starbright Blue 700 anti-rabbit (1:2,000; Bio-Rad) and Starbright Blue 520 anti-mouse (1:2,000; Bio-Rad). Blots were imaged using a ChemiDoc XRS+ for chemiluminescent detection or ChemiDoc MP (Bio-Rad) for fluorescent detection. Protein bands were quantified by densitometry using Image Lab Software (v.6.1, Bio-Rad). All specific protein bands were normalized to total protein determined by UV-activated stain-free gels as a loading control.

Statistical Analyses

GraphPad Prism (v.10, Prism Software, La Jolla, CA) was used for statistical analyses. Shapiro–Wilk test was used to determine normality. The robust regression and outlier removal (ROUT) test identified and removed outliers based on a false discovery rate with Q = 5%. Statistical analyses consisted of Student’s t tests, Mann–Whitney U tests, one-way ANOVA, two-way with and without repeated-measures ANOVA, or mixed-effects model followed by Bonferroni’s post hoc test, when appropriate, as indicated in the figure legends. Data are presented as means ± SE. Significance was accepted at P ≤ 0.05. Details of statistical analyses can be found in the Supplemental Materials.

RESULTS

T2D Is Associated with a Reduced Vasodilatory Response to Flow, Increased Plasma ADAM17 Activity, and Elevated Concentration of Plasma-Soluble HA in Humans

We show the time-course profile of femoral artery FMD during reactive hyperemia in a cohort of females and males with T2D, relative to age-matched controls without T2D (Fig. 1A). As shown, the magnitude of FMD over the 180-s period following cuff deflation is markedly blunted in T2D. Peak FMD, expressed relative to the magnitude of shear stimulus during reactive hyperemia, was also reduced in this cohort of individuals with T2D (Fig. 1B). The effect of T2D on FMD was independent of sex (T2D by sex interaction P > 0.05), thus data from both sexes were combined (Fig. 1, A and B). Furthermore, we report that plasma ADAM17 activity and concentration of soluble HA were higher in individuals with T2D relative to counterparts without T2D (Fig. 1, C and D). These observations support a potential link between elevated ADAM17 activity, loss of HA from the endothelial glycocalyx, and impaired mechanotransduction of shear stress in T2D.

db/db Mice Exhibit Reduced Endothelial Sensitivity to Shear Stress and Lower Endothelial CD44 Content

In alignment with our observations in humans, we provide evidence that for a given magnitude of shear stress, the vasodilatory response is lower in db/db mice compared with wild-type controls (Fig. 1E). Notably, this observation was associated with increased expression of endothelial ADAM17 in en face preparations of arteries isolated from db/db versus wild-type mice (Fig. 1F). Moreover, CD44 content was lower in endothelial cells isolated from db/db mice versus controls (Fig. 1G).

Enzymatic Degradation of HA, CD44 Knockdown, or Inhibition of the HA Binding Site on CD44 All Impaired Mechanotransduction of Shear Stress

We demonstrated that mesenteric arteries isolated from wild-type mice and exposed intraluminally to hyaluronidase exhibit reduced FMD, without affecting endothelium-independent vasodilation (Fig. 2A). In support of the notion that CD44 plays a role in mechanotransduction of shear, we first showed that acute increases in intracellular Ca²⁺ following a rapid increase in shear stress are attenuated in live cells deficient in CD44 (Fig. 2, B and C). We also demonstrated that, following exposure to 24 h of flow, eNOS upregulation was reduced in CD44-deficient cells (Fig. 2D). Moreover, blockade of the HA-binding site on CD44 reduced FMD in isolated arteries (Fig. 2E). Taken together, these data support the assertion that CD44 transduces HA-associated shear stress mechanotransduction.

Elevated Endothelial ADAM17 Activity Increased Cleavage of CD44 in Association with Impaired Shear Stress Mechanotransduction

Here we report that pharmacological activation of ADAM17 caused enhanced cleavage of CD44, as determined by supernatant/lysate CD44 content, and was prevented by coincubation with TAPI-0, an inhibitor of ADAM17 (Fig. 3A). This phenomenon was associated with reduced eNOS activation (p-eNOS/eNOS ratio) in cultured cells exposed to 2 h of shear stress (15 dyn/cm²) (Fig. 3B). To specifically interrogate the role of ADAM17, we overexpressed ADAM17 in HUVECs (ADAM17-OE). ADAM17-OE led to an increase in its activity, measured using a cleavable fluorogenic probe specific to ADAM17, and cleavage of CD44 (supernatant/lysate) (Fig. 3C). We also recapitulated our earlier findings in CD44 knockdown cells by measuring an attenuated shear stress-induced increase in intracellular Ca²⁺ and eNOS expression in ADAM17-OE cells (Fig. 3, D–F). Moreover, this loss of mechanotransduction was associated with lower levels of cell-surface CD44 (Fig. 3E). These data support the premise that increased ADAM17 activity is associated with CD44 cleavage and attenuated shear stress mechanotransduction.

Figure 3.

Active ADAM17 cleaves endothelial cell-surface CD44 impairing mechanotransduction of shear stress. A: ADAM17 activity (n = 6–8/condition); and cleavage of CD44 in the absence or presence of TAPI-0 (50 μM) in human umbilical vein endothelial cells (HUVECs) stimulated with phorbol 12-myristate 13-acetate (PMA) (1 μM; 24 h; n = 7 or 8 replicates/condition), with representative blots. B: endothelial nitric oxide synthase (eNOS) activation following exposure to shear stress (15 dyn/cm²; 2 h), in HUVECs stimulated with PMA (1 μM; 24 h; n = 12 replicates/condition), with representative blots. C: ADAM17 expression (n = 12 replicates/condition), activity (n = 12 replicates/condition), and cleavage of CD44 (n = 7–11 replicates/condition) in ADAM17-OE HUVECs, with representative blots. D: flow-induced changes in intracellular Ca²⁺ in ADAM17-OE HUVECs following exposure to shear stress (15 dyn/cm², 0–300 s; n = 3 or 4 replicates/condition), with representative images; scale bar = 200 μm. E and F: fluorescent intensity of ADAM17(left) or cell-surface CD44 (E, right) and total eNOS (F), in HUVECs and following exposure to shear stress (15 dyn/cm²; 24 h), with representative images (E and F, n = 4 replicates/condition). G: cleavage of human recombinant CD44 (approximately 70 kDa, 75 ng/well) by human recombinant active ADAM17 (ADAM17-r, approximately 64 kDa; 750 ng/well; 8 h) in a cell-free assay in the absence or presence of TAPI-0 (50 μM), with representative blots (n = 6 replicates/condition). H: fluorescent intensity of cell-surface CD44 in HUVECs exposed to either ADAM17-r (1 μg/mL, 1 h) or heat-inactivated (100°C for 30 min) ADAM17-r control with representative images (n = 6–8 replicates/condition); scale bar = 100 μm. I: schematic representation of the surface plasmon resonance (SPR) system and experimental workflow. J: representative sensorgrams of resonance units (RUs) showing CD44-r-fc binding to protein A–functionalized high-capacity carboxyl sensors following incubation with either active or heat-inactivated ADAM17-r (5 ng/μL, 8 h; n = 4/condition) and subsequent binding of hyaluronan (HA) to CD44-r-fc (n = 4/condition). K: FMD and sodium nitroprusside (SNP)-induced dilation in isolated mesenteric arteries exposed intraluminally to either ADAM17-r (n = 6, females/males = 3/3; 1 μg/mL, 1 h) or heat-inactivated ADAM17-r (n = 7, females/males = 4/3). All images were enhanced equally for visualization purposes only, and all analyses were performed using the raw data. Data are presented as means ± SE. Two-way ANOVA with repeated measures was used to assess flow-induced changes in Ca²⁺ in HUVECs and FMD and SNP-induced vasodilation (D and K). Two-way ANOVA with Bonferroni’s post hoc analyses were used to assess eNOS activation in PMA-stimulated cells and in ADAM17-OE cells exposed to shear stress (B, E, and F). One-way ANOVA was used to assess differences in cleavage of CD44 (A and G). Mann–Whitney test was used to determine differences between cell-surface CD44 (H). Two-tailed unpaired Student’s t tests were used in all other analyses. *P ≤ 0.05 vs. respective controls and main effect of condition; #P ≤ 0.05 vs. PMA (A), shear stress (B, E, and F), ADAM17-r (G); &P ≤ 0.05 vs. ADAM17-OE (E).

ADAM17 Cleaves Cell-Surface CD44, Reducing Shear Stress Sensitivity

Using a cell-free approach, we demonstrated that active ADAM17-r cleaves CD44-r, in a metalloproteinase-dependent manner (Fig. 3G). Furthermore, we showed that HUVECs exposed to active ADAM17-r (1 μg/mL; 1 h) exhibit lower cell-surface CD44, as determined using a monoclonal antibody specific to the extracellular N-terminal of CD44 (Fig. 3H). Using SPR, we strengthened the notion that CD44 cleavage contributes to loss of HA-associated shear stress mechanotransduction by demonstrating that HA binding to CD44-r following exposure to active ADAM17-r is reduced (Fig. 3, I and J). Notably, we also report that intraluminal incubation of isolated arteries with ADAM17-r impaired FMD, without affecting SNP-induced vasodilation (Fig. 3K).

DISCUSSION

A diminished ability of blood vessels to dilate in response to increased blood flow-mediated shear stress (e.g., during an FMD assay) is a well-established marker of endothelial dysfunction and cardiovascular disease risk in T2D (23, 26). Moreover, endothelial dysfunction coincides with and precedes arterial stiffening and atherosclerosis development, considered hallmark vasculopathies of T2D (27–29). Thus, uncovering the molecular and cellular mechanisms underlying impaired shear stress mechanotransduction may be a catalyst for identifying novel therapeutic targets or strategies to ameliorate cardiovascular disease in T2D and beyond. Herein, we report that ADAM17, a metalloproteinase with elevated sheddase activity in T2D, cleaves CD44 from the endothelial cell surface, reducing shear stress mechanotransduction associated with HA (Fig. 4).

Figure 4.

Schematic illustrating the primary findings. CD44 anchors extracellular hyaluronan to the endothelial plasma membrane. Hyaluronan serves as a shear stress mechanosensor (i.e., detects shear stress), whereas CD44 is the mechanotransducer (i.e., transmits the shear stress signal). In the healthy state (left), CD44 transduces the shear stress signal sensed by hyaluronan, leading to a normal flow-mediated dilation (FMD) response. In type 2 diabetes (right), endothelial ADAM17 overactivity causes CD44 cleavage and the consequent loss of hyaluronan-associated mechanotransduction of shear stress, leading to an impaired FMD response.

Specifically, we provide evidence that plasma HA concentration and ADAM17 activity are elevated in individuals with T2D who exhibit reduced FMD. These findings support and extend our previous reports that ADAM17 expression is enhanced in isolated human T2D arteries. Notably, our overexpression data support the notion that increased ADAM17 expression alone can precipitate elevated ADAM17 activity in HUVECs. Moreover, data from others report elevated plasma HA in prediabetes, type-1 diabetes, and pulmonary hypertension, all of which are characterized by endothelial dysfunction (17, 30–32). Given that HA is a primary constituent of the mechanosensitive glycocalyx, it is reasonable to suspect that elevated plasma HA concentration is evidence of glycocalyx erosion (4). Although hyaluronidases are reported to be elevated in T2D, it is also rational to posit that cleavage of CD44, and subsequent release of HA, contributes to this pool of soluble HA (33). This is because CD44 anchors extracellular HA to the cell membrane (12, 13). However, it is important to note that vascular CD44 has been reported to be upregulated in conditions such as atherosclerosis (34). This may still be due to increased cleavage of the extracellular portion of CD44 because this can in turn stimulate γ-secretase-dependent cleavage of the CD44 intracellular domain (ICD) (i.e., sequential proteolytic cleavage) (35). The ICD of CD44 translocates to the nucleus, where it can signal an increase in transcription and expression of CD44 (35). Thus, total CD44 content may not always be indicative of cleavage. Taken together, our work lends support to another potential mechanism by which HA and the endothelial glycocalyx may be degraded in T2D.

As extracellular HA is implicated in mechanotransduction of shear stress and is fixed to the plasma membrane by surface receptors, we determined whether CD44, its primary receptor, transduces HA-associated mechanosensation. First, we corroborate previous reports that hyaluronidase impairs FMD in isolated arteries (1, 8). Then we demonstrate that genetic knockdown of CD44 reduces mechanotransduction of shear stress in HUVECs. Although we observed only a partial loss of mechanotransduction, this finding supports previous reports of ablation of shear stress mechanotransduction in CD44-deficient endothelial cells (14). The partial loss is likely explained by the presence of multiple, parallel, and redundant mechanosensory pathways such as those mediated by heparan-sulfated glypican-1 or piezo channels (36). To demonstrate that HA is reliant on CD44 to transduce shear stress, we show that functional blockade of the HA binding site on CD44 reduces shear stress sensitivity in isolated arteries exposed to increasing intraluminal flow. This functional blockade of the interaction between HA and CD44 has been previously reported to inhibit HA-dependent signaling and reduce FMD in mice, in vivo (15, 37). In sum, these findings fortify the concept that CD44 is a mechanotransducer of shear stress associated with HA.

Given that CD44 has been demonstrated to be cleaved by ADAM17 in nonendothelial cells (16), and ADAM17 is elevated in T2D (17, 18), we hypothesized that increased ADAM17 activity causes enhanced cleavage of endothelial CD44 and impairs mechanotransduction of shear stress. In support of our hypothesis, we report that increased endothelial ADAM17 activity augmented the cleavage of CD44 and reduced cell-surface CD44 in HUVECs. Notably, this was associated with an attenuated upregulation of eNOS, and blunted intracellular Ca²⁺, in response to shear stress. These are indicative of impaired shear stress mechanotransduction, a phenotype recapturing what was observed in HUVECs silenced for CD44. Unexpectedly, shear stress augmented the expression of ADAM17 and CD44. ADAM17 maturation and trafficking is regulated by inactive rhomboid 1 and 2 (iRhom1/2), and iRhom1 has been shown to increase in response to shear stress (38). This iRhom1-associated induction of ADAM17 may facilitate endothelial realignment to shear stress (whereas iRhom2 is more potently induced in inflammatory conditions) (38). Moreover, shear stress has been shown to enhance glycocalyx thickness (39), although the exact components that are responsive to shear are unknown. Reinforcing the premise that ADAM17 cleaves cell-surface CD44, we also show that active ADAM17-r cleaves CD44-r in a cell-free assay, lowers HA binding to CD44-r as determined by SPR, and reduces cell-surface CD44 in HUVECs. Congruently, intraluminal exposure of isolated arteries to active ADAM17-r impairs FMD. Although we did not test whether genetic deletion or inhibition of ADAM17 prevents cleavage of CD44 and reduced FMD in a mouse model of T2D, we have recently reported that endothelial-specific knockdown of ADAM17 mitigates loss of FMD in Western diet-fed obese mice (40). Together, our findings support the basis for ADAM17 cleaving CD44 and impairing mechanotransduction of shear stress.

In aggregate, this work supports the idea that endothelial ADAM17 activity cleaves CD44, impairing shear stress mechanotransduction and implicating elevated ADAM17 activity in T2D-associated endothelial dysfunction. As shear stress mechanotransduction is a universal and critical function of endothelial cells, we use vessels and cells from multiple vascular beds (i.e., elastic/muscular conduit and resistance arteries, as well as venous and microvascular endothelial cells). Thus, this mechanism may be conserved across diverse endothelia. However, further research is required to clarify these pathways, decipher the exact mechanisms by which endothelial cell ADAM17 becomes overactive in T2D, the molecular signals that contribute to targeting of CD44, and whether targeting endothelial ADAM17 activity represents a feasible therapeutic strategy to improve cardiovascular outcomes in T2D.

DATA AVAILABILITY

The data underlying this article can be shared upon reasonable request to the corresponding authors.

GRANTS

This work is supported, in part, by the National Institutes of Health Grants R01HL151384 (to L.A.M.-L. and J.P.), R01HL137769 (to J.P.), R21DK116081 (to C.M.-A.), the Veterans Affairs Merit Grant 1I01CX002399 (to C.M.-A. and J.P.); grants from the American Heart Association 23PRE1020897 (to G.P.), 25DIVSUP1463861 (to F.I.R.-P.), and 24EIA1248820 (to J.P.); and the São Paulo Research Foundation FAPESP, 2024/09267-4 (to L.F.-S.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

G.P.: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Validation, Visualization, Writing – original draft, Writing – review & editing; M.J.C.: Data curation, Formal analysis, Investigation, Methodology, Writing – review & editing; J.H.B.-O.: Data curation, Formal analysis, Investigation, Methodology, Writing – review & editing; M.A.A.: Data curation, Formal analysis, Investigation, Methodology, Writing – review & editing; N.I.: Data curation, Investigation, Methodology, Writing – review & editing; O.M.L.: Data curation, Formal analysis, Investigation, Methodology, Writing – review & editing; F.I.R.-P.: Data curation, Formal analysis, Investigation, Methodology, Software, Supervision, Visualization, Writing – review & editing; L.F.-S.: Data curation, Formal analysis, Investigation, Methodology, Writing – review & editing; C.M.-A.: Funding acquisition, Investigation, Resources, Writing – review & editing; L.A.M.-L.: Conceptualization, Funding acquisition, Investigation, Methodology, Resources, Supervision, Writing – review & editing; J.P.: Conceptualization, Funding acquisition, Investigation, Methodology, Resources, Supervision, Writing – original draft, Writing – review & editing.