Polyunsaturated Fatty Acids Suppress PIEZO Ion Channel Mechanotransduction in Articular Chondrocytes

Abstract

Osteoarthritis (OA) is characterized by articular cartilage degeneration, leading to pain and loss of joint function. Recent studies have demonstrated that omega-3 (ω3) polyunsaturated fatty acid (PUFA) supplementation can decrease injury-induced OA progression in mice fed a high-fat diet. Furthermore, PUFAs have been shown to influence the mechanical properties of chondrocyte membranes, suggesting that alterations in mechanosensitive ion channel signaling could contribute to the mechanism by which ω3 PUFAs decreased OA pathogenesis. Here, we hypothesized that PUFAs may alter mechanical signaling through PIEZO1 (activated by changes in membrane tension) and TRPV4 (activated by physiologic mechano-osmotic signals), as these mechanosensitive cation channels have been shown to influence OA progression. Our results demonstrated that PUFAs reduced chondrocyte sensitivity to single-cell mechanical compression and to pharmacologic agonists of PIEZO1 and TRPV4, with ω3 PUFAs having the most significant effects overall. We also found that supplementation with ω6 PUFA linoleic acid (LA) altered the biophysical properties of chondrocytes, as evidenced by increased intracellular lipid droplet formation and more rapid membrane rupture in response to hypo-osmotic shock, suggesting that LA increases chondrocyte membrane susceptibility to damage. Our findings underscore the differential impacts of specific PUFAs on chondrocyte signaling and membrane properties and provide important considerations in the development of nutritional interventions to prevent or treat OA.

Introduction

Osteoarthritis (OA) is a debilitating and painful disease characterized by progressive degeneration of the articular cartilage, bone remodeling, synovitis, and deleterious changes in other joint tissues (1). While the etiology of OA is multifactorial, obesity is a major risk factor for OA development, as this condition both alters mechanical loading on the joint (1) and promotes metabolic dysfunction (2). The mechanism(s) by which altered joint loading promotes OA development is complex and multifactorial as well, but appears to involve the mechanosensing of increased tissue-level loads by the chondrocytes and other cells of the joint, leading to a dysregulation of the normal balance of anabolic and catabolic factors needed for maintaining cartilage homeostasis and function (1, 3).

Recent studies suggest that dietary and metabolic factors, particularly the quantity and type of free fatty acids (FAs), also contribute to the relationship between obesity and OA (4). Fatty acids make up a large part of the Western diet (2), and are categorized as saturated, monounsaturated, or polyunsaturated fatty acids (PUFAs) (5). Based on double bond positioning, PUFAs are further categorized as ω3 PUFAs, such as docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA), or ω6 PUFAs, such as arachidonic acid (AA) and linoleic acid (LA) (5). Both the quantity of ω6 PUFAs and the ratio of ω6:ω3 PUFAs correlate positively with the risk of obesity (2). While the evolutionary dietary ratio of ω6:ω3 PUFAs was 1:1, the modern Western diet averages a ω6:ω3 ratio of 20:1 (2). As mammals lack significant endogenous capabilities to modify FAs, the FA content consumed from the diet is extremely important in determining the overall content of FAs in the body, which influences disease progression (6). For example, among adults with OA, higher circulating ω6:ω3 FA ratios were associated with increased knee pain and decreased functionality (7). In chondrocytes specifically, ω3 PUFAs decrease markers of cartilage damage (ADAMTSs and MMPs) while ω6 PUFAs reduce chondrogenesis and increase cartilage degeneration (8). Moreover, Wu, et al. demonstrated that mice fed diets enriched in ω3 PUFAs had decreased OA progression, synovitis, and osteophyte formation following knee injury, even after controlling for body weight (4). Additionally, mice that were genetically modified to express the FA desaturase fat-1, which transforms ω6 PUFAs into ω3 PUFAs, demonstrated less severe post-traumatic OA and synovitis compared to wild-type mice of the same body weight (9). More recently, it has been shown that delivering fat-1 to mice using AAV-based gene therapy protected against high-fat diet (HFD)-induced metabolic dysfunction and OA (10). These findings suggest that the balance between dietary ω6 and ω3 PUFAs plays a critical role in the relationship between obesity and OA, beyond just the effects of increased joint loading. A number of potential mechanisms may explain the influence of FA composition on OA. For example, it has been hypothesized that increased ratios of ω6:ω3 PUFAs can significantly shift the local and systemic inflammatory processes in the eicosanoid pathway from the synthesis of anti-inflammatory, pro-resolving mediators (e.g., PGE3, LTB5) to a more inflammatory phenotype (e.g., PGE2, LTB4) (reviewed in (11)). Fatty acid composition can modulate a number of intracellular signaling pathways through direct membrane receptors (5). By serving as ligands for intracellular signaling pathways, ω6 PUFAs decrease chondrogenesis, while ω3 PUFAs increase the production of chondrogenic markers (5), suggesting the two types of PUFAs activate profoundly different pathways in chondrocytes.

It has also been hypothesized that PUFAs can influence cell properties, and therefore, mechanotransduction, through alterations in mechanical properties of cell membranes (12). Fatty acids are spontaneously incorporated into cell membranes (5), which results in changes in membrane thickness (13) and fluidity (12). Additionally, after being taken up by chondrocytes, excess FAs will accumulate into intracellular lipid droplets (14). Though the effects of lipid droplets on chondrocytes is still an active area of research, studies on other types of cells have shown that lipid droplets can alter both signaling pathways (15) and membrane properties (16). By altering the mechanical properties of the cell membrane through incorporation into the membrane or formation of lipid droplets, FAs may alter transmembrane mechanosensitive ion channel signaling (17). Specifically, PUFAs have been shown to modify the activation properties of PIEZO1 channels (17) and Transient Receptor Potential Vanilloid 4 (TRPV4) cation channels (18), which are key channels involved in chondrocyte mechanotransduction (19–21).

In chondrocytes, PIEZO1 channels are activated by high levels of membrane strain (>45%) resulting from supra-physiological mechanical loads associated with cell and tissue injury (20). The subsequent Ca²⁺ influx causes increased expression of inflammatory mediators and promotes chondrocyte apoptosis and senescence (21). Looking further downstream, PIEZO1 channels have been linked to OA progression, as the mRNA and protein levels of PIEZO1 are significantly higher in OA joints than in healthy joints (8). Additionally, an increase in inflammatory mediator interleukin 1 (IL-1) leads to an elevation in PIEZO1 activation in response to mechanical loading, suggesting PIEZO1 channels both respond and contribute to OA changes (8). In contrast to PIEZO1 channels, TRPV4 channels are activated primarily by osmotic stress, which may occur secondary to mechanical compression of the cartilage matrix and the ensuing exudation of interstitial fluid from the tissue (19). Under physiologic loading conditions, TRPV4 channels regulate the anabolic response of chondrocytes (22). Interestingly, TRPV4 channels have been suggested to mediate cartilage changes related to age- and obesity-induced OA (23, 24), while PIEZO1 channels are likely to be involved in regulating injury-induced OA pathogenesis (1).

Here, we examined the hypothesis that different PUFAs can influences PIEZO1 and TRPV4 mechanosignaling and inflammatory mediator expression in chondrocytes. We hypothesized that ω3 PUFAs would reduce ion channel signaling and inflammatory gene expression, while ω6 PUFAs would sensitize these responses. We investigated the influence of FAs on chondrocyte mechanosignaling using atomic force or confocal microscopy for intracellular calcium ion (Ca²⁺) imaging, and the role of FAs in biochemical signaling using gene expression analysis. We found that ω3 and ω6 PUFAs differently modulate PIEZO1 and TRPV4 activation in response to varied stimuli, and that the two types of PUFAs have varied effects on the expression of different inflammatory markers. Our findings support the notion that PUFAs, particularly ω3 PUFAs, can reduce pathologic mechanical signaling and alter inflammatory and senescence marker production in cartilage.

Materials and Methods

Chondrocyte isolation and cell culture

Stifle joints from skeletally mature pigs (5–6 months old, mixed breeds of Landrace Yorkshire and Duroc) were obtained from a regional abattoir on the day of slaughter. Articular cartilage was extracted from the middle of the femoral condyle and cultured in medium consisting of high glucose DMEM (Gibco, Thermo Fisher Scientific, Waltham, MA), 10% fetal bovine serum (FBS, Atlas Biologicals, Fort Collins, CO), 1.5% HEPES (Corning, Corning, NY), 1% MEM non-essential amino acid (Corning), 1% Pen/Strep (Gibco, Thermo Fisher Scientific), and 0.5% L-proline (Sigma Aldrich, St. Louis, MO).

Within 48 h of harvest, cartilage samples were digested for 1 to 1.5 hours using pronase (Worthington Biochemical, Lakewood, NJ) in a wash medium at a 1320 PKU/ml medium. The wash medium contained high glucose DMEM, 1x gentamycin (Gibco, Thermo Fisher Scientific), 1x kanamycin (Goldbio, St. Louis, MO), and 1x fungizone (Corning). After digestion, the medium containing protease was discarded, and tissue samples were incubated for 3 to 3.5 hours in a medium containing 0.4% collagenase type II (Worthington Biochemical).

Next, cells were isolated, cleaned, and resuspended in a growth medium containing 1.5% HEPES and 1% MEM Non-essential amino acid (Corning). In preparation for AFM microscopy, cells were counted and then seeded onto 12mm round coverslips #1.5 (Electron Microscopy Sciences, Hatfield, PA) at densities ranging from 50,000 to 100,000 cells per coverslip. In preparation for confocal microscopy, cells were transferred to high-resolution glass-bottom 96-well plates (Cellvis, Sunnyvale, CA). These cells were cultured for three days in an incubator set at 5% CO₂ and 37 °C. Throughout all experiments, serum-free mediums were maintained at an osmolarity of 380 mOsm to emulate the physiological conditions observed in cartilage [32, 33].

Fatty acid solutions and treatments

Fatty acids (Nu-Check-Prep Inc, Elysian, MN) were dissolved in 100 μl of ethanol, slowly added to fatty-acid free bovine serum albumin (BSA) solution (0.33 grams BSA in 4.9 ml of PBS) resulting in 5mM as final concentration of FA stock. To assure proper reconstitution FA solutions were placed on a shaker for 1 hour at 37°C. Lastly, the FAs were aliquoted into small glass vials, from which oxygen was removed using nitrogen gas. The vials were then sealed and stored at −20 °C.

In preparation for cell treatment, the FAs were diluted with a serum-free medium (1% PenStrep, 1% NEAA, and 1% HEEPS in DMEM) at an osmolarity of 380 mOsm, resulting in a final concentration of 100 μM. The FA solution was then filtered and added to the cells. The cells were cultured for three days before analysis.

Inflammation and Senescence Treatments

To simulate an inflammatory environment, porcine IL-1α (R&D Biosystems) was added to FA media to achieve a final concentration of 1 ng/mL. To induce senescence, H₂O₂ was added to FA media to achieve a final concentration of 100 μM. Two days after chondrocytes were first exposed to FAs, media containing FAs with either IL-1 or H₂O₂ were added to the cells. The chondrocytes were treated with IL-1 or H₂O₂ for 24 hours before gene expression analysis using RT-qPCR.

Ca²⁺ imaging using confocal microscopy

Cells were stained for one hour using two dyes, Fura red-AM (Invitrogen, Thermo Fisher Scientific) and Fluo4-AM (Invitrogen, Thermo Fisher Scientific), both of which bind to intracellular Ca²⁺. After staining, cells were washed and imaged under a confocal microscope (LSM 880, Zeiss, Dublin, CA) at 37°C using 488 nm excitation laser. The cells were imaged every 3 seconds for 10 minutes in the Yoda1 and GlaxoSmithKline1016790A (GSK101) experiments and for 4 minutes for the hypo-osmotic challenge experiment. The variations in imaging times were due to differences in the durations of the responses between pharmacologic agonists and osmotic stress. The imaging process included one minute of initial visualization before a control solution (DMSO, Sigma Aldrich) was introduced, an additional two minutes of observation before a stimulus solution was added, and a final 1 minute or 7 minutes of imaging to record the cells’ reactions to the stimulus. The stimulus solution contained Yoda1 (resulting in a final concentration of 2.5 μM), GSK101 (resulting in a final concentration of 2.5 μM), or water (resulting in an osmotic stress of −110 mOsm). Iso-osmotic conditions were maintained throughout the entire process of labeling, washing, and imaging cells. The resulting video data was processed in ImageJ, and each cell’s response to Ca²⁺ was quantified.

Live-dead imaging using confocal microscopy

Cells were labeled for 45 minutes using the live-dead kit (Calcein (Alive) and EthD-1 (Dead), Invitrogen, Thermo Fisher Scientific). Next, the cells were imaged using confocal microscopy to evaluate cell death in response to PUFA treatment.

Imaging of intracellular lipid droplets using confocal microscopy

Cells were stained for 30 minutes with the BODIPY stain, which labels intracellular lipid droplets. After staining, confocal microscopy was used to obtain Z-stack images of cells.

Imaging of cell size during osmotic challenge

Chondrocytes were imaged via brightfield microscopy while an osmotic shock was applied. From the resulting videos, the initial and final diameters of the cells were measured and used to quantify membrane strength and cell sensitivity to osmotic shock.

Atomic force microscopy to measure cell Ca²⁺ response to deformation

For atomic force microscopy (AFM) analysis, primary chondrocytes were stained for 2 hours with intracellular Ca²⁺ binding dye Fura2-AM (Invitrogen, Thermo Fisher Scientific). The staining solution was composed of Phenol red-free DMEM (Gibco, Thermo Fisher Scientific) with 1.5% HEPES buffer (Corning), 1% Pen Strep (Gibco Thermo Fisher Scientific), 1% MEM non-essential amino acid (Corning), 1% Na Pyruvate (Corning), 1% GlutaMax (Gibco, Thermo Fisher Scientific), and 0.1% L-Proline (Sigma Aldrich. Tipless cantilevers (7 to 13 N/m) were used to perform cell compression (Nanoandmore, Watsonville, CA) and calibrated using the manufacturer-provided thermal method for measuring stiffness. An Atomic Force Microscope (AFM; MFP-3D Bio, Asylum Research, Santa Barbara, CA) was used to compress chondrocytes while light sources with 340 and 380 nm wavelengths captured ratiometric Ca²⁺ images. The loaded cells were imaged 140 times over the period of 80 seconds. Cells that moved during the imaging process or were damaged (as evidenced by a quick decline in intracellular fluorescence) were excluded from the study. All experiments were conducted at 37°C. After collection, the video data of the intracellular Ca²⁺ response during mechanical pressure was processed through ImageJ software (U.S. National Institutes of Health, Bethesda, MD, https://imagej.nih.gov/ij/).

Image analysis

All images were processed using ImageJ. For cell viability analysis, percent viability was calculated as 100 times the total pixel intensity of live cells divided by total pixel intensity of live cells plus dead cells.

For confocal imaging, images were separated into two distinct channels, corresponding to the fluo-4 staining and fura-red staining while for AFM into the 340 nm and 380 nm imaging wavelengths. Next, signal intensity was normalized by thresholding the videos according to staining intensity, and average pixel intensities were standardized to the baseline signal level. Each cell’s maximum normalized pixel brightness value was then recorded as the cell’s intracellular Ca²⁺ response (ΔF_max/F).

For AFM studies, a control assessment was conducted in which the cantilever was positioned over the cells without exerting any pressure. The magnitude of these cells’ intracellular Ca²⁺ responses was recorded and standardized before the mean (M) and standard deviation (SD) were calculated. Cells that demonstrated a peak in their Ca²⁺ response profile and a Ca²⁺ response magnitude greater than the average response of the control group plus three times the standard deviation (ΔF_max/F > M + 3*SD) profile were deemed responders. The formula for identifying responders was designed to maximize the detection of true responders while minimizing the impact of signal noise. Cells not meeting these criteria were categorized as non-responders.

For confocal analyses, videos were processed in ImageJ, and custom MATLAB software was utilized to compute the proportion of responding cells and the average magnitude of the intracellular Ca²⁺ response. First, the software identified each cell in the image and reported the total number of cells. Then, the software normalized each cell’s intracellular Ca²⁺ response following Yoda1 addition to the cell’s response after DMSO addition. The software then calculated the average cellular response to DMSO addition, and the standard deviation of this response. Next, the software applied a threshold to determine whether the cellular response (ΔF_max/F) of each cell exceeded the average response after DMSO addition (Mˊ) plus five times its standard deviation (SDˊ) (Yoda1: ΔF_max/F > Mˊ + 5*SDˊ, GSK101: ΔF_max/F > Mˊ + 10*SDˊ, Hypo-osmotic: ΔF_max/F > Mˊ + 5*SDˊ). As these formulas were developed to distinguish true Ca²⁺ responses from noise, they were adjusted for the varying magnitudes of chondrocytes’ Ca²⁺ responses to mechanical loading, pharmacologic agonism and hypo-osmotic stress. Cells meeting these criteria were termed responders, and the percentage of cells deemed to be responders was recorded for each video. The average cellular response (ΔF_max/F) was also determined for each video.

ImageJ was used to measure the initial and final diameter of each cell, calculated by measuring each cell’s horizontal width and vertical height and averaging the two lengths. The final diameter of the cell was measured at the video frame in which the cell reached its maximal surface area, which occurred directly before the cell membrane failed and the cell exploded. If a cell did not explode, the final diameter was still measured at the video frame in which the cell was largest. The values obtained for the initial and final diameter were used to calculate each cell’s change in diameter, surface area, and volume.

To analyze the size lipid droplets that accumulated within each cell after PUFA treatment, maximum intensity projection (MIP) images were obtained from Z-stack images using ZenBlue software. The MIP images were then imported into ImageJ, and a threshold was applied so only the lipid droplets were visible. Then, the total area occupied by the lipid droplets and the mean intensity of the lipid droplets were measured for each cell.

RNA Extraction and RT-qPCR

After three days of treatment with FAs, cells were harvested, and mRNA was isolated (Norgen, Total RNA Purification Plus kit, Thorold, ON, Canada). First, cells seeded and treated in a 6-well plate were briefly washed with PBS. Next, 300 μl of the lysis buffer was added to each well. After 10 minutes, the lysate was collected in tubes snaped frozen in liquid nitrogen. The mRNA was harvested and purified before its concentration was measured using a NanoDrop. cDNA was generated from the mRNA (VILO Superscript Mastermix, Life Technologies, Carlsbad, CA) followed by real-time qPCR utilizing SYBR Green Master mix (Thermo Fisher Scientific). The delta delta Ct (ΔΔCt) approach was employed to quantify the relative gene expression of the target genes, and the expression levels were standardized to ACTB mRNA levels. The primer sequences used are detailed below:

Statistical analysis

The mean values for each group are displayed on every graph along with the standard error of the mean (±SEM). The statistical significance (p<0.05) between different groups was assessed using the one-way analysis of variance (ANOVA) followed by a Tukey post-hoc test.

Results

PUFA treatment decreased chondrocyte intracellular Ca²⁺ response to mechanical compression or Yoda1

To investigate the effects of FAs on PIEZO1 channel sensitivity, we exposed chondrocytes to high magnitudes of mechanical compression, known to activate Piezo1 channels, using AFM and a tipless cantilever. After 3 days of exposure to 100 μM FA or BSA, chondrocyte viability ranged from 80.2% to 85.2%. There was no significant difference in chondrocyte viability between cells treated with 100 μM FA and those exposed to the BSA control (p=[0.472–0.805]) (Figure S1). Supplementation with DHA, EPA, or LA significantly reduced the chondrocytes’ intracellular Ca²⁺ response to mechanical compression at 500 nN (loading rate of 1 μm/sec) compared to the BSA control (Figure 1A–D; p < 0.0001). AA or PA supplementation did not significantly alter chondrocytes’ intracellular Ca²⁺ response in comparison to the BSA control. Moreover, while 39.5% of cells treated with AA and 47% of the cells treated with BSA showed an increase in intracellular Ca²⁺ levels in response to mechanical compression, only 23 to 27% of cells treated with LA, DHA, EPA, or PA responded to the same stimulus (Figure 1C).

Figure 1.

Effect of PUFAs on intracellular Ca²⁺ response of chondrocytes to mechanical compression. A) Representative signaling trend for each FA. B) Intracellular Ca²⁺ fluorescence intensity ΔF_max/F for each FA. C) Percentage of responding cells for each FA. D) Considering only the responding cells, intracellular Ca²⁺ fluorescence intensity ΔF_max/F for each FA. Data presented as mean ± SEM. One-way ANOVA with Tukey’s post-hoc test was used for group comparisons in B-D. Different letters indicate statistical significance (p<0.05). If the same letter is assigned to multiple groups, those groups are not significantly different from each other. Conversely, if two groups are denoted with completely different letters, the two groups are significantly different. Cell count =34–106 for B and cell count = 9–43 for D. List of FAs: BSA (bovine serum albumin, control); LA (linoleic acid, ω6, pink); AA (arachidonic acid, ω6, purple), DHA (docosahexaenoic acid, ω3, light blue), EPA (eicosapentaenoic, ω3, dark blue), PA (palmitic acid, saturated long chain, red).

Next, we utilized confocal microscopy to evaluate the effects of different FAs on PIEZO1 sensitivity in chondrocytes exposed to Yoda1. All analyzed FAs significantly decreased PIEZO1 channel sensitivity to 2.5 μM of Yoda1 when compared to the control group (Figure 2A–C; p < 0.0001). The magnitude of the intracellular Ca²⁺ response was reduced by 65 to 69% in the AA, LA, DHA, and EPA groups compared to the BSA control, while the percentage of the responsive cells was reduced 69 to 81% in the AA, LA, DHA, and EPA groups compared to the BSA control (Figure 2B–C, p < 0.0001). As in our AFM experiment, chondrocytes treated with EPA exhibited the lowest Ca²⁺ response and the lowest percentage of responding cells. Chondrocytes treated with PA displayed a significantly lower intracellular Ca²⁺ response (p = 0.0341) and lower percentage of responding cells (p < 0.0001) than those treated with BSA, but a significantly greater Ca²⁺ response (p < 0.0001) and higher percentage of responding cells (p < 0.0001) compared to chondrocytes treated with ω3 or ω6 PUFAs.

Figure 2.

Effect of PUFAs on intracellular Ca²⁺ response of chondrocytes to Yoda1, GSK101, and hypo-osmotic shock. A) Representative signaling trend for each FA for Yoda1. B) Intracellular Ca²⁺ fluorescence intensity ΔF_max/F for each FA for Yoda1. C) Percentage of responding cells for each FA for Yoda1. D) Representative signaling trend for each FA for GSK101. E) Intracellular Ca²⁺ fluorescence intensity ΔF_max/F for each FA for GSK101. F) Percentage of responding cells for each FA for GSK101. G) Representative signaling trend for each FA for hypo-osmotic shock. H) Intracellular Ca²⁺ fluorescence intensity ΔF_max/F for each FA for hypo-osmotic shock. I) Percentage of responding cells for each FA for hypo-osmotic shock. Data presented as mean ± SEM. One-way ANOVA with Tukey’s post-hoc test was used for group comparisons in B, C, E, F, H and I. Different letters indicate statistical significance (p<0.05). If the same letter is assigned to multiple groups, those groups are not significantly different from each other. Conversely, if two groups are denoted with completely different letters, the two groups are significantly different. Number of experiments = 8–18. List of FAs: BSA (bovine serum albumin, control); LA (linoleic acid, ω6, pink); AA (arachidonic acid, ω6, purple), DHA (docosahexaenoic acid, ω3, light blue), EPA (eicosapentaenoic, ω3, dark blue), PA (palmitic acid, saturated long chain, red).

PUFA treatment decreased chondrocyte intracellular Ca²⁺ response to GSK101 but did not affect cellular response to hypo-osmotic stress

After analyzing the effects of FA supplementation on PIEZO1 activity, we wondered whether FA supplementation would influence other chondrocyte mechanosensitive channels. As TRPV4 is a mechanosensitive cation channel highly expressed in chondrocytes (1), we examined the effect of FAs on TRPV4 activation using confocal microscopy. We again supplemented chondrocytes with PUFAs, PA, or BSA control and activated TRPV4 using 1nM of GSK101, a TRPV4-specific agonist. The magnitude of the intracellular Ca²⁺ response was reduced by 59 to 80% in the AA, LA, DHA, and EPA groups compared to the BSA control, while the percentage of the responsive cells was reduced 56 to 79% in the AA, LA, DHA, and EPA groups compared to the BSA control (Figure 2D–F; p < 0.05). The EPA group again demonstrated the lowest intracellular Ca²⁺ response and the lowest proportion of responding cells. Chondrocytes treated with PA did not exhibit significant differences in the Ca²⁺ response magnitude or percentage of responding cells compared to the BSA control group.

As TRPV4 is endogenously activated by hypo-osmotic stimulation (25), we evaluated the effect of PUFAs on TRPV4 activity in response to hypo-osmolarity. FA treatment did not significantly affect chondrocytes’ intracellular Ca²⁺ response to a hypo-osmotic challenge of −110 mOsm (Figure 2G–I). Cells treated with DHA trended towards exhibiting a lower intracellular Ca²⁺ response and percentage of responding cells, but these findings did not reach significance (p = 0.5752 for intracellular Ca²⁺ response, p = 0.3415 for percentage of responding cells).

PUFAs alter the amount of excess membrane area of chondrocytes

Our finding that FAs altered the activation of PIEZO1 suggested that FA supplementation may directly modulate cell membrane integrity. To test this, we subjected chondrocytes to a hypo-osmotic shock and recorded the change in cell diameter, surface area, and volume prior to membrane rupture. Only treatment with LA resulted in a reduced percentages of change in cell diameter, surface area, and volume at the time of membrane rupture compared to the BSA control (Figure 3A–C, p < 0.0001). This finding suggested to us that LA reduced excess membrane area (e.g., membrane ruffling), making cells more susceptible to rupture.

Figure 3.

Effect of PUFAs on excess membrane area. Percentage of increase in cell A) diameter, B) surface area, and C) volume before membrane rupture in the presence of PUFAs. List of FAs: BSA (bovine serum albumin, control); LA (linoleic acid, ω6, pink); AA (arachidonic acid, ω6, purple), DHA (docosahexaenoic acid, ω3, light blue), EPA (eicosapentaenoic, ω3, dark blue), PA (palmitic acid, saturated long chain, red). Different letters indicate statistical significance (p<0.05). If the same letter is assigned to multiple groups, those groups are not significantly different from each other. Conversely, if two groups are denoted with completely different letters, the two groups are significantly different.

We continued to investigate the effects of FAs on membrane integrity exposing treated chondrocytes to osmotic stress. By adding either water or a hyper-osmotic medium to the iso-osmotic microenvironment (380 mOsm) in which cells were cultured, we subjected cells to a final osmolarity of 180, 280, 480, or 580 mOsm. The resulting Ponder’s values (slope of line reflecting normalized volume (V/V₀) vs. inverse osmolarity) ranged from 0.81 to 0.98, with chondrocytes treated with BSA demonstrating the lowest Ponder’s value and those treated with DHA demonstrating the highest (Figure 4A–F). We also found positive correlations between the percent increases in chondrocyte diameter, surface area, and volume after a hypo-osmotic shock and Ponder’s value among the treatment groups (Figure S2).

Figure 4.

Effect of PUFAs on chondrocytes’ sensitivity to osmotic shock. Ponder’s plots for A) BSA, B) LA, C) AA, D) DHA, E) EPA, F) PA. List of FAs: BSA (bovine serum albumin, control); LA (linoleic acid, ω6, pink); AA (arachidonic acid, ω6, purple), DHA (docosahexaenoic acid, ω3, light blue), EPA (eicosapentaenoic, ω3, dark blue), PA (palmitic acid, saturated long chain, red).

Throughout our osmotic stress studies, we observed lipid droplets in cells treated with FAs (Figure S3). We used confocal microscopy to measure the size and presence of lipid droplets. The addition of all tested FAs significantly increased lipid droplet area (measured by lipid droplet area pixel intensity), and the percentage of cell area occupied by lipid droplets compared to the BSA control (Figure 5A–D; p < 0.0001, Figure S3). Interestingly, cells treated with LA had the largest lipid droplets and greatest lipid droplet pixel intensity.

Figure 5.

Effect of PUFAs on lipid droplet formation. A) Area of lipid droplets B) Pixel intensity of lipid droplets C) Total intensity of lipid droplets which is area of lipid droplets multiplied by their pixel intensity. D) Percentage of the cell area that is occupied by the lipid droplets. Data presented as mean ± SEM. One-way ANOVA with Tukey’s post-hoc test was used for group comparisons in A-D. Different letters indicate statistical significance (p<0.05). If the same letter is assigned to multiple groups, those groups are not significantly different from each other. Conversely, if two groups are denoted with completely different letters, the two groups are significantly different. Number of cells = 62–111. List of FAs: BSA (bovine serum albumin, control); LA (linoleic acid, ω6, pink); AA (arachidonic acid, ω6, purple), DHA (docosahexaenoic acid, ω3, light blue), EPA (eicosapentaenoic, ω3, dark blue), PA (palmitic acid, saturated long chain, red).

Differential effects of PUFAs on mRNA expression of senescence and inflammatory markers

After determining that different FAs have distinct effects on mechanosensitive signaling and membrane integrity in chondrocytes, we analyzed how FAs influence the expression of several key genes involved in cartilage maintenance and OA pathogenesis. Here, we supplemented chondrocytes with FAs and measured mRNA expression of different mechanosensitive ion channel, chondrogenic, adipogenic, inflammatory, and senescence gene markers (Table S1). FA treatment did not change PIEZO1 and PIEZO2 mRNA levels compared to the BSA control (Figure S4A–B; A: p = 0.0723, B: p = 0.1423). However, FA treatment significantly modulated TRPV4 mRNA expression (Figure S4C, P < 0.0001). In particular, DHA (p < 0.0001) and EPA (p < 0.0001), AA (p = 0.0004), and PA (p = 0.0005) significantly increased TRPV4 mRNA expression compared to the BSA control, while LA did not (p=0.053).

Next, we investigated the ability of FAs to modify the expression of three positive chondrogenesis markers, aggrecan (ACAN), collagen type II (COL2A1), and SOX9, as well as one negative chondrogenesis marker, collagen type I (COL1A1) (Figure S4D–G; D: p = 0.0045, E: p < 0.0001, F: p = 0.0109, G: p < 0.0088). Except for PA, all of the FAs included in our study reduced the expression of COL2A1. DHA and PA significantly reduced the expression of COL1A1. Our next experiment revealed that FA supplementation had no significant effect on the expression of adipogenesis markers (Figure S4H–I; H: p = 0.2743, I: p = 0.2372).

We then investigated the utility of FA supplementation in modulating the expression of inflammatory biomarkers (Figure S5A–E). Treatment with DHA, EPA, or PA reduced chondrocytes’ IL6 expression by 34–63%, while treatment with AA or LA increased IL6 expression by 34–36% (Figure S5C, p <0.0001). FA treatment also significantly affected IL8 expression, with each PUFA increasing IL8 mRNA expression by 79–160% (Figure S5D, p = 0.0033). When chondrocytes were exposed to an inflammatory environment, none of the FA treatment groups expressed significantly different levels of inflammatory genes compared to the BSA control (Figure S6A–E, p = 0.3215–0.9999).

Finally, we evaluated the effects of PUFAs on mRNA expression levels of different cellular senescence markers, including P16, P21, P53, CCL8, MMP3, MMP12, CXCL2, and TIMP1 (Figure S7A–H). Chondrocytes treated with DHA expressed significantly lower mRNA levels of P53 (p = 0.0047 compared to BSA) and MMP3 (p = 0.0002 compared to BSA), while chondrocytes treated with EPA or PA expressed significantly lower levels of MMP3 only. In contrast, LA supplementation significantly increased mRNA expression of MMP3 (p < 0.0001 compared to BSA) and MMP12 (p = 0.0135 compared to BSA). After chondrocytes were treated with H₂O₂, which is known to induce cellular senescence, FA treatment had minimal impact on senescence marker expression (Figure S6F–K), with only EPA significantly increasing the expression of a single senescence marker, p16, compared to the control (p = 0.0491). In summary, FA supplementation significantly influenced the expression of inflammatory and senescence markers by chondrocytes under baseline conditions, but these effects were notably diminished when chondrocytes were exposed to IL-1 or H₂O₂.

Discussion

Our findings demonstrate that the presence and type of FAs may significantly affect the activity of the mechanosensory ion channels PIEZO1 and TRPV4 in articular chondrocytes and modulate the expression of a range of anabolic and catabolic biomarkers involved in cartilage physiology and OA. Specifically, we observed that both ω3 and ω6 PUFAs attenuated chondrocytes’ intracellular Ca²⁺ response to mechanical compression, PIEZO1 agonism, and TRPV4 agonism. EPA, DHA and LA were particularly effective in reducing Ca²⁺ signaling after mechanical compression, while EPA, DHA, and AA most effectively attenuated Ca²⁺ signaling in response to GSK101. All PUFAs equally decreased Ca²⁺signaling in response to Yoda1. Thus, only EPA and DHA were superior in diminishing chondrocytes’ Ca²⁺ response across all three stimuli, suggesting that ω3 PUFAs are most effective at reducing Ca²⁺signaling overall. Furthermore, supplementation with the ω6 PUFA LA reduced the cells resistance to membrane rupture and increased the intracellular content occupied by lipid droplets, implying an increased susceptibility to mechanical damage. Thus, our findings support the hypothesis that different types of FAs have distinct influences on chondrocyte physical properties and mechanotransduction.

We observed that PUFA supplementation reduced chondrocyte intracellular Ca²⁺ response to high-strain mechanical compression applied at a level specific for activating PIEZO1 channels. In previous studies, it has been shown that activation of PIEZO1 is dependent on the magnitude of effective membrane strains, which depends on a number of factors including applied cellular strain magnitude and rate, extracellular osmolarity, and the amount of excess membrane area. One mechanism through which PUFAs may decrease PIEZO1 mechano-activation is by the force from the lipid principle, which states that changes in the composition and properties of the lipid bilayer result in adjustments to membrane forces, which in turn cause membrane-embedded proteins to undergo conformational changes (26). Several studies have supported this principle by demonstrating differences in membrane properties following the incorporation of FAs into the lipid membrane. Specifically, both DHA and EPA increase membrane elasticity (27) and membrane fluidity (28), suggesting that ω3 PUFA incorporation into the cell membrane results in decreased membrane deformation and strain in response to mechanical loading. As membrane strain resulting from chondrocyte deformation causes PIEZO1 activation (29), the changes in membrane properties resulting from PUFA incorporation likely influenced PIEZO1 mechano-activation. We also found that PUFA supplementation did not affect PIEZO1 mRNA expression. This finding further supports the force from the lipid hypothesis (26), as it suggests that the decreases in PIEZO1 activity were due in large part to changes in membrane forces and not to changes in mRNA expression.

While all of the PUFAs decreased chondrocytes’ response to high-strain mechanical compression, none of them changed chondrocytes’ Ca²⁺ response to hypo-osmotic stress. While changes in osmolarity are associated with alterations in the excess membrane area and effective membrane tension, it has been shown that the TRPV4 channel is gated by extracellular osmolarity and not by membrane tension per se, unlike PIEZO1 (25, 29). Furthermore, following a hypo-osmotic stress, the ω6 PUFA LA altered the physical properties of chondrocytes, as evidenced by a reduction in the peak cell diameter, surface area, and volume at the time of membrane rupture, indicating a reduction in normal membrane ruffling and excess membrane area at iso-osmotic conditions. LA also increased the percent of the cell area occupied by lipid droplets, which could explain why it decreased membrane integrity. This hypothesis is supported by previous studies on adipocytes, which show that lipid droplet accumulation alters the physical properties of these cells as evidenced by increased cell stiffness (16). Thus, our finding that LA increases lipid droplet accumulation and decreases membrane area is consistent with previous studies reporting of LA-induced chondrocyte injury and decreased chondrocyte function (30). This mechanism could also contribute to the relationship between increased ω6 PUFA dietary intake, OA pathogenesis and progression (8). Of note was the finding that treatment with different FAs had a significant influence on the biophysical properties of chondrocytes, as evidenced by their Ponder’s values (intracellular osmotically active fraction) (31) and its relationship to increased FA accumulation intracellularly. This finding raised the possibility that the addition of FAs may have a significant effect on the fraction of osmotically active intracellular components within chondrocytes, and therefore, how chondrocytes physically respond to varying extracellular osmolarity, as occurs with cartilage loading in situ (32). The addition of FAs revealed a correlation between the percent increases in chondrocyte diameter, surface area, and volume after a hypo-osmotic shock and the Ponder’s values, suggesting that chondrocytes’ ability to modulate cell volume in response to increasing osmotic pressures after FA treatment is related to increased intracellular osmotic activity. Taken together, these findings provide additional support for the hypothesis that FAs can modulate the biophysical and biological responses of chondrocytes to mechanical signals.

While the Ca²⁺ current that followed the hypo-osmotic shock was transient, returning to baseline in about 100 seconds, the current induced by GSK101 stimulation was sustained for the duration of the experiment, lasting over 200 seconds. TRPV4 is activated by relative changes in osmolarity, not by absolute osmolality (25). Thus, it is only transiently activated by a hypo-osmotic shock before quickly returning to its baseline signaling. In contrast, pharmacologic stimulation of TRPV4 creates a Ca²⁺ current of longer duration. In 2009, Phan, et al. demonstrated that the Ca²⁺ current stimulated by a hypo-osmotic shock returned to baseline within 100 seconds, but the Ca²⁺ current in response to TRPV4 activator 4α-phorbol 12,13-didecanoate lasted at least 300 seconds (25). Thus, one can conclude that pharmacologic agonists have a longer duration of effect on TRPV4 signaling compared to hypo-osmotic shocks.

In addition to examining chondrocytes intracellular Ca²⁺ responses to mechanical and osmotic stress, we also directly tested their response to PIEZO1 and TRPV4 activation by utilizing specific pharmacologic agonists of these channels (Yoda1 and GSK101, respectively). Fatty acid supplementation significantly reduced chondrocyte intracellular Ca²⁺ response to either Yoda1 or GSK101. As PIEZO1 channels are implicated in injury-induced OA (29) and TRPV4 channels have been linked to age-induced OA (1), ω3 PUFA supplementation could be useful for treating or preventing OA of both etiologies (1). Moreover, our finding that FA supplementation reduced chondrocyte intracellular Ca²⁺ response to Yoda1 and GSK101 in the absence of mechanical stimulation implies that FAs modulate intracellular Ca²⁺ signaling by affecting pathways involved in PIEZO1 and TRPV4 directly, although these effects may still arise from alterations in channel activity through interactions with chondrocyte membrane structure and properties that are altered by FAs.

Our results are consistent with previous studies showing that various FAs have distinct effects on PIEZO1 activity (17). Our data demonstrates that supplementation with individual FA results in unique changes in PIEZO1 and TRPV4 activity, as well as in the expression of a range of chondrocyte- and OA-associated biomarker gene expression. For example, we found that AA differentially affects PIEZO1 activity based on stimulation mode (mechanical vs pharmacologic). Additionally, our finding that DHA reduces PIEZO1 activation agrees with the report from Romero, et al. (17), who reported that EPA and AA increase PIEZO1 activation (17), although we found that both of these PUFAs decrease PIEZO1 activation. We hypothesize that differences in cell types and PIEZO1 activation modes between our work here and Romero et al. may explain these differential findings.

In addition to examining the effects of PUFA supplementation on mechanosensitive Ca²⁺ ion channel activity and gene expression, we also investigated how PUFAs modulate the expression of OA-associated inflammatory markers in chondrocytes. One of the markers we tested, IL-6, is a cytokine that promotes the cartilage degradation, inflammation, and pain associated with OA (33). While ω3 PUFAs (EPA and DHA) reduced IL-6 expression under baseline conditions, ω6 PUFAs (AA and LA) increased IL-6 expression. However, we did not measure the absolute levels of IL-6 expression which limits our ability to determine the precise impact that PUFAs have on modulating inflammation in the joint. Despite this limitation, the finding that ω3 PUFAs were able to suppress IL-6 expression under baseline conditions is consistent with a range of previous in vitro and in vivo studies of chondrocytes and OA pathology. Specifically, ω3 fatty acids have been shown to suppress expression of markers of cartilage damage (e.g., ADAMTSs and MMP3) in addition to decreasing synovial GAG loss and oxidative stress while ω6 fatty acids decrease chondrogenesis and increase cartilage degeneration (5). However, both ω3 and ω6 PUFAs increased IL-8 expression, a cytokine involved in chemotaxis in the joint (34). Moreover, none of the FAs significantly affected inflammatory marker expression when chondrocytes were exposed to IL-1. These results suggest that the overall effects of ω3 and ω6 PUFAs on the homeostatic balance of anabolic and catabolic activities involved in chondrogenesis and inflammation may require longer term or even in vivo studies and remain to be determined.

The mechanism behind these findings likely involves FA metabolites and their downstream effects. ω6 PUFAs are metabolized into arachidonic acid, which is converted into potent eicosanoids (35). Eicosanoids are a class of pro-inflammatory mediators, which includes leukotrienes and prostaglandins. Both leukotrienes and prostaglandins promote the production of inflammatory mediators and collagenase in cartilage, which contribute to OA progression (36). While all ω6 PUFAs are converted to potent pro-inflammatory eicosanoids such as PGE2 and LTB4 (35), ω3 PUFAs EPA and DHA have distinct metabolic pathways (37). EPA is metabolized into weak eicosanoids and resolvins, while DHA is primarily converted into resolvins, protectins, and maresins (37), which have mixed pro- and anti-inflammatory effects (38). These studies provide a potential explanation for why PUFAs demonstrate mixed pro- and anti-inflammatory properties. Additionally, previous research has demonstrated that inflammatory environments can disrupt metabolic pathways in chondrocytes (39), which may account for the altered responses of chondrocytes to FAs after treatment with IL-1.

Just as PUFAs had varied effects on expression of inflammatory mediators, they were also varied in their ability to reduce chondrocyte senescence marker expression, a hallmark of aging, OA, and joint injury (40, 41). DHA, EPA, and PA each decreased mRNA levels of one or more senescence markers and did not increase mRNA levels of any senescence markers. In contrast, AA and LA each increased the expression of two senescence markers. However, inducing senescence in chondrocytes led to different gene expression patterns, with EPA being the only FA to increase expression of a single senescence marker. Additionally, it is important to note that only relative levels of senescence markers were measured, which may not accurately represent true levels of cell senescence or apoptosis.

In summary, supplementing chondrocytes with both ω3 and ω6 PUFAs effectively decreased PIEZO1- and TRPV4-mediated Ca²⁺ signaling, with ω3 fatty acids EPA and DHA emerging as the most potent modulators of Ca²⁺ signaling overall. However, the effects of PUFA supplementation on inflammatory marker and senescence marker expression were less clear. These findings provide new insights into the role of PUFAs as potential OA therapeutics. Directions for future studies include investigating the effects of PUFA supplementation on PIEZO1 and TRPV4 activation in animal models, further clarifying the distinct mechanisms by which Yoda1 and mechanical stress cause PIEZO1 activation and elucidating the relative effects of FAs on mechanical properties of the chondrocyte membrane and on biochemical signaling pathways.

Abbreviations

AA

arachidonic acid

AAV

adeno-associated virus

ADAMTS

a disintegrin and metalloproteinase with thrombospondin motifs

AFM

atomic force microscopy

ANOVA

analysis of variance

BSA

bovine serum albumin

Ca²⁺

calcium ion

DHA

docosahexaenoic acid

DMEM

Dulbecco’s modified Eagle medium

DMSO

dimethyl sulfoxide

EPA

eicosapentaenoic acid

FA

fatty acid

FBS

fetal bovine serum

HEPES

4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

IL-1

interleukin 1

LA

linoleic acid

LTB4

leukotriene B4

LTB5

leukotriene B5

MEM

modified Eagle medium

MIP

maximum intensity projection

MMP

matrix metalloproteinase

OA

osteoarthritis

PA

palmitic acid

Pen/strep

penicillin and streptomycin

PGE2

prostaglandin E2

PGE3

prostaglandin E3

PUFA

polyunsaturated fatty acid

SEM

standard error of the mean

TRPV4

transient receptor potential vanilloid 4

ω3

omega 3 (fatty acid)

ω6

omega 6 (fatty acid)

Data Availability:

All data needed to evaluate the conclusions of the paper are presented in the article and the Supplementary Information. Additional data related to this paper are available from the corresponding author upon request.