Non-invasive Electroarthrography Measures Load-Induced Cartilage Streaming Potentials via Electrodes Placed on Skin Surrounding an Articular Joint

Abstract

Objective

We aimed to demonstrate that electroarthrography (EAG) measures streaming potentials originating in the cartilage extracellular matrix during load bearing through electrodes adhered to skin surrounding an articular joint.

Design

Equine metacarpophalangeal joints were subjected to simulated physiological loads while (1) replacing synovial fluid with immersion buffers of different electrolyte concentrations and (2) directly degrading cartilage with trypsin.

Results

An inverse relationship between ionic strength and EAG coefficient was detected. Compared to native synovial fluid, EAG coefficients increased (P < 0.05) for 5 of 6 electrodes immersed in 0.1X phosphate-buffered saline (PBS) (0.014 M NaCl), decreased (P < 0.05) for 4 of 6 electrodes in 1X PBS (0.14 M NaCl), and decreased (P < 0.05) for all 6 electrodes in 10X PBS (1.4 M NaCl). This relationship corresponds to similar studies where streaming potentials were directly measured on cartilage. EAG coefficients, obtained after trypsin degradation, were reduced (P < 0.05) in 6 of 8, and 7 of 8 electrodes, during simulated standing and walking, respectively. Trypsin degradation was confirmed by direct cartilage assessments. Streaming potentials, measured by directly contacting cartilage, indicated lower cartilage stiffness (P < 10⁻⁵). Unconfined compression data revealed reduced Em, representing proteoglycan matrix stiffness (P = 0.005), no change in Ef, representing collagen network stiffness (P = 0.15), and no change in permeability (P = 0.24). Trypsin depleted proteoglycan as observed by both dimethylmethylene blue assay (P = 0.0005) and safranin-O stained histological sections.

Conclusion

These data show that non-invasive EAG detects streaming potentials produced by cartilage during joint compression and has potential to become a diagnostic tool capable of detecting early cartilage degeneration.

Keywords

electroarthrography streaming potential electromechanics biomechanics early osteoarthritis cartilage degeneration

Introduction

Osteoarthritis (OA) is characterized by progressive deterioration of articular cartilage eventually resulting in loss of joint function and pain.^1,2 No treatments presently exist to modify or reverse the disease course at later, symptomatic stages, and late-stage OA can only be remedied by joint replacement. Identifying earlier, asymptomatic, stages of the disease, where cartilage deterioration has recently commenced, may present opportunities for treatment aimed at slowing disease progression.^1,3 To accomplish this, sensitive diagnostics of cartilage quality are required, as current methods, such as radiography, are not capable of detecting subtle changes in cartilage that may occur during early OA.⁴ Imaging methods, including ultrasound and magnetic resonance imaging are promising candidates; however, they are limited by high costs and technical expertise requirements.^4,5 Electroarthrography (EAG) is a non-invasive method for measuring cartilage streaming potentials, which result when the hydrated cartilage extracellular matrix (ECM) responds to compressive loading. The cartilage ECM consists of large hydrophilic proteoglycan molecules which are immobilized by a highly organized, predominantly type II collagen network. When cartilage is in equilibrium, the fixed, negative charge density of the glycosaminoglycan chains is electrically balanced by mobile positive ions distributed within interstitial fluid. During cartilage compression, a pressure gradient develops that displaces mobile positive ions relative to proteoglycan, which remain immobilized by the fibrillar collagen network, thereby producing an electrical signal known as the streaming potential.^6-8

Cartilage streaming potentials have been measured by directly contacting cartilage with electrodes using a variety of benchtop apparatus,^6-10 as well as an indentation probe.^11-14 These studies have consistently demonstrated that streaming potentials detect changes in cartilage quality, including, for example, biomechanical changes resulting from enzymatic degradation^8,10 and acute impact injury.¹³ Cartilage streaming potentials correlated to histological assessments, including Mankin score,^15-17 collagen architecture,¹⁵ as well as the International Cartilage Repair Society macroscopic score.^12,16 Correlations to biochemical composition,^8,10,15 chondrocyte viability, and apoptosis¹⁷ have also been demonstrated by several groups. Streaming potentials, measured non-destructively by lightly compressing articular surfaces using an indentation probe, were capable of discerning subtle cartilage degeneration in human knee joints that was not visible to the eye.¹¹ Indeed, these authors demonstrated greater effect sizes for streaming potentials, indicating greater sensitivity to cartilage degenerative changes, compared to indentation tests, unconfined compression, biochemical assay, or histological appearance.¹¹ EAG is a non-invasive method for measuring cartilage streaming potentials that involves placing contact electrodes on skin external to an articulation while the joint undergoes compressive loading.^18-22 Préville et al.¹⁸ introduced a clinical EAG protocol whereby adhesive electrodes were attached to skin surrounding one knee and the subject applied compressive load to the instrumented leg by shifting their body weight to achieve a one-legged stance. EAG measurements were thus acquired from 20 control subjects, as well as both knees of twenty patients diagnosed with radiographic bilateral knee OA and who had undergone unilateral total knee arthroplasty (TKA). EAG distinguished among these 3 groups of knees and detected statistically significantly higher streaming potentials, indicating better quality cartilage, in control knees compared with osteoarthritic knees, while EAG measurements from TKA knees were statistically null.

The present study aims to demonstrate that EAG originates in articular cartilage by controlling the conditions under which streaming potentials are produced. We used cadaveric equine fetlock (metacarpophalangeal) joints to study EAG response to altering the salt concentration of the solution bathing cartilage surfaces, as well as to enzymatic degradation of cartilage. The equine fetlock joint, consisting of the distal metacarpus, the proximal first phalanx and sesamoid bones, is suitable because it is not surrounded by musculature, which could produce high-frequency electrical signals in live animals, nor fat, which has been shown to diminish cartilage streaming potentials in a finite element simulation.¹⁹ Theoretically, higher ionic strength reduces streaming potentials because the presence of additional mobile ions increases electrical shielding. Streaming potential amplitude, measured directly from cartilage during compression, has indeed been shown to decrease as ionic strength increases^6,7,23 and we hypothesize in our first set of experiments that EAG will demonstrate a similar inverse relationship. A second set of experiments examined EAG response to enzymatic cartilage degradation by intra-articular injection of trypsin, a proteinase that digests proteoglycan and has been used to create in vitro models of early OA.^9,24-26 Cartilage exposed to trypsin develops a gradient of proteoglycan depletion over time that progresses from the articular surface toward the bone-cartilage interface.^27,28 We hypothesize that non-invasive EAG would diminish in response to trypsin degradation similar to direct measurements of cartilage streaming potentials, biomechanical and biochemical properties, and histological appearance.

Methods

Procedures were performed with Institutional Animal Care Committee approval in accordance with the Canadian Council on Animal Care (University of Guelph, AUP#1412). Three distal forelimbs, comprising the third metacarpus to the hoof, were obtained from skeletally mature horses ranging from 2 to 7 years of age and weighing 561 ± 49 kg. Forelimbs were frozen until testing. EAG measurements were acquired on each fetlock while it underwent a set of loading sequences in the presence of native synovial fluid followed by three immersion buffers with different electrolyte concentrations.

EAG Measurements

Each forelimb was mounted in a servohydraulic mechanical tester (Instron 8800, Norwood, MA, USA) with joint axes aligned to those of a 6-degree-of-freedom load cell (SI-1500-240, ATI Industrial Automation, Apex, NC, USA). EAG was performed on each fetlock during simulated physiological loading using gold-plated disk electrodes attached to 6 sites around the articulation. Electrodes were placed at the anterior (dorsal) phalanx/metacarpus interface, on the medial (EAG1) and lateral (EAG2) sides of the common digital extensor tendon, as well as at the medial (EAG3) and lateral (EAG4) phalanx/metacarpus interface and the medial (EAG5) and lateral (EAG6) metacarpus/sesamoid interface ( Fig. 1 ). Two additional electrodes were placed on the diaphysis of the metacarpus approximately 8 cm away from the articulation and served as negative reference and ground electrodes, respectively. EAG signals were acquired with a wireless data acquisition system (Clevemed BioRadio 150, Great Lakes NeuroTechnologies, Independence, OH, USA) at an acquisition rate of 600 Hz. Skin impedance was minimized by clipping hair, applying an abrasive sponge and cleaning with 70% ethanol prior to affixing electrodes with conductive gel (Ten20, Weaver & Co, Aurora, CO, USA). Compressive loads representing approximately half the load experienced during standing, standing, and walking were achieved with displacements of 8 mm, 15 mm, and 25 mm, respectively.²⁹ Loading sequences consisted of 10 cycles where displacement was applied at 5 mm/s, held for 5 seconds, and unloaded at 5 mm/s. Three measurements of change in flexion angle were made for each loading sequence by using a goniometer to measure joint angles at the zero and peak displacement positions. Using custom-designed software (Biomomentum, Laval, Quebec, Canada), EAG coefficients (μV/kg) for the last 5 cycles of each loading sequence were calculated by fitting EAG signals obtained at each electrode to measured axial load resulting from displacement applied to the whole joint.

Figure 1.

Placement of electroarthrography (EAG) electrodes on a right fetlock joint. Views are (A) lateral, (B) anterior (dorsal), and (C) medial. Electrodes were placed dorsally on skin over the articulation between the metacarpus and phalanx on the medial (EAG1) and lateral (EAG2) sides of the common digital extensor tendon. Mediolateral electrodes were placed over the medial (EAG3) and lateral (EAG4) interfaces between the metacarpus and phalanx. Two electrodes were placed on medial (EAG5) and lateral (EAG6) interfaces between the metacarpus and sesamoid bones. The negative reference and ground electrodes were placed on the metacarpus where skin is directly over bone at a distance of approximately 8 cm proximal from EAG1.

Adjustment of Electrolyte Concentrations with Different Immersion Buffers

EAG measurements were repeated four times on each joint. First, with cartilage surfaces bathed in native synovial fluid, followed by 1 of 3 immersion buffers, namely 0.1X phosphate-buffered saline (PBS) (0.014 M NaCl), 1X PBS (0.14 M NaCl), or 10X PBS (1.4 M NaCl). PBS solutions were made from a 10X stock solution consisting of 1369.4 mM NaCl, 97.4 mM Na₂HPO₄, 21.8 mM KH₂PO₄, and 26.5 mM KCl (BP665, Thermo Fisher Scientific, Waltham, MA, USA). Immersion buffers were adjusted to pH 7.4.

Native synovial fluid was removed by fully flexing the fetlock and inserting an 18-gauge needle into the joint space between the sesamoid and metacarpus bones first on the medial, and then lateral, aspects. To maximize synovial fluid removal, the needle was oriented downward to allow gravity to assist and the joint was repeatedly flexed and extended. Quantities of synovial fluid retrieved were 2.5 mL, 4 mL, and 5.5 mL for each of the 3 fetlocks, respectively.

To transition from synovial fluid to 0.1X PBS, the fetlock was subjected to a minimum of 2 lavages with 0.1X PBS followed by addition of a fresh quantity of 0.1X PBS prior to repeating EAG measurements. Lavage involved accessing the joint space as described above and delivering 5 mL 0.1X PBS medially and 5 mL 0.1X PBS laterally. Distribution throughout the joint was encouraged by manually flexing and extending 30 times followed by sinusoidal loading at 0.1 Hz with amplitude of 3 mm for 20 cycles. The lavage solution was then retrieved and conductivity measured (described below). Lavages were repeated until conductivity stabilized. Then 10 mL of fresh 0.1X PBS was delivered to the joint and it was incubated at 4°C for 1 hour. Lavage solutions were retrieved in quantities ranging from 2.5 mL to 10.5 mL (median 8.5 mL), indicating some solution remained in the joint and surrounding tissues with each lavage. EAG measurements were completed with joint surfaces bathed in 0.1X PBS, then the lavage procedure repeated to transition from 0.1X PBS to 1X PBS and finally from 1X PBS to 10X PBS.

Conductivity Measurements

Conductivity was measured using an Accumet AP85 (Thermo Fisher Scientific, Waltham, MA, USA) calibrated with a 12.88 mS/cm standard. 10X PBS solutions were diluted 10:1 with deionized water prior to conductivity measurement. The conductivity probe was cleaned with a mild soap solution and rinsed with deionized water between measurements. The certificate of analysis for the lot of PBS used reported conductivity of a 1X solution as 16.6 mS/cm at 25°C.

Trypsin Degradation of Cartilage

Fresh distal forelimbs from 1 horse (18 years old, 460 kg) were collected to investigate EAG sensitivity to trypsin degradation of cartilage. EAG measurements (as above) were performed on the right fetlock. Cartilage degradation was subsequently induced by intra-articular injection (as above) of 0.5% w/v trypsin (T9201, Sigma Aldrich, St. Louis, MO, USA) in Tris buffer (BP154-1 Fisher, Waltham, MA, USA) and incubation at 37°C for 3 hours.²⁷ EAG measurements were repeated post-trypsin. Confirmation of trypsin degradation was sought by directly measuring cartilage streaming potentials, biomechanical properties, histological appearance, and biochemical composition. The untreated contralateral fetlock served as a control for these destructive analyses. Direct measurements of cartilage streaming potentials were made at 75 sites on the articular surfaces of the metacarpus and phalanx with the Arthro-BST (Biomomentum Inc.); an arthroscopic device that records streaming potentials during manual cartilage indentation through an array of 37 microelectrodes (5 microelectrodes/mm²). The device calculates a quantitative parameter (QP, arbitrary units) corresponding to the number of microelectrodes in contact with cartilage when the sum of potentials reaches 100 mV. Cartilage disks, 3 mm in diameter, were subsequently isolated from load bearing regions^29,30 for testing in unconfined compression geometry. Each disk was subjected to five stress relaxation ramps of 2% strain. The fibril-reinforced poroelastic model³¹ was used to obtain fibril modulus (Ef), matrix modulus (Em), and hydraulic permeability (k). Cartilage disks were then fixed in 10% neutral buffered formalin and embedded in paraffin. Five-micrometer sections were stained with safranin-O/fast green/iron hematoxylin. Biochemical composition was quantified in adjacent cartilage disks submitted to the dimethylmethylene blue (DMMB) and hydroxyproline (HPr) assays.³²

Statistical Analysis

A full factorial analysis of variance with replicates was used to explain variation in EAG coefficients due to immersion buffer (synovial fluid, 0.1X PBS, 1X PBS, 10X PBS) and displacement (8 mm, 15 mm, 25 mm). For each electrode, immersion buffer and displacement were treated as fixed effects with repeated measurement across 5 loading cycles. The experiment was independently replicated 3 times with fetlock treated as a random effect. Tukey’s post hoc honestly significant difference was applied to identify significant differences in average EAG coefficients, accounting for multiple comparisons. Paired t tests were used to compare EAG coefficients, QP, biomechanical parameters, and biochemical parameters. Statistical analyses were performed in Statistica v.8 and reported as least square mean (lower 95% confidence limit, upper 95% confidence limit). 95% confidence intervals = ±1.96 × standard error of the mean. Descriptive data are reported as mean ± standard deviation.

Results

Conductivity measurements were made on synovial fluid and immersion buffers retrieved after EAG measurements ( Table 1 ) and within a temperature range of 20.7°C ± 0.7°C (n = 12). Results from a t test indicated that compared to synovial fluid, 0.1X PBS had significantly lower (P = 0.021) and 10X PBS significantly higher (P < 1 × 10⁻⁴) conductivities. There was no statistical difference detected between conductivity measured in synovial fluid and 1X PBS.

Table 1.

Conductivity of Synovial Fluid and Immersion Buffers Retrieved After Electroarthrography Measurements Were Performed [Least Square Mean (Lower 95% Confidence Interval, Upper 95% Confidence Interval), n = 3].

	Conductivity (mS/cm)
Synovial fluid	6.14 (5.13, 7.15)
0.1X PBS	4.09 (3.67, 4.51)
1X PBS	6.68 (6.23, 7.13)
10X PBS	43.47 (39.12, 47.82)

Taken together, across all displacements and fetlocks, average EAG coefficients increased (P < 0.05) at 5 of 6 electrodes in the presence of 0.1X PBS and decreased (P < 0.05) at all electrode sites in the presence of 10X PBS compared with synovial fluid ( Fig. 2 ). When cartilage surfaces were bathed in 1X PBS, EAG coefficients at 4 of 6 electrodes were significantly different (P < 0.05) compared with synovial fluid. These patterns, where EAG coefficient increases in the presence of 0.1X PBS and then decreased proportionally when cartilage is immersed in 1X PBS and 10X PBS, were most pronounced at electrodes placed on the anterior aspect of the fetlock (EAG1 and EAG2).

Figure 2.

Effect of electrolyte concentration on electroarthrography (EAG) coefficients. EAG measurements were made at 6 sites while cartilage surfaces were immersed in native synovial fluid, 0.1X phosphate-buffered saline (PBS) (0.014 M NaCl), 1X PBS (0.14 M NaCl), or 10X PBS (1.4 M NaCl). Significant differences (P < 0.05) accounting for multiple comparisons were detected between immersion buffers unless otherwise identified as not significant (n.s.). Anterior electrodes, EAG1 and EAG2, as well as EAG3 at the mediolateral interface between the metacarpus and phalanx, demonstrated significantly different EAG coefficients for each solution. EAG coefficients measured at EAG4 and EAG5 were significantly different among synovial fluid, 0.1X PBS (0.014 M NaCl) and 10XPBS (1.4 M NaCl), but not between synovial fluid and 1X PBS (0.14 M NaCl). EAG coefficients measured at EAG6 were significantly different among synovial fluid, 1X PBS (0.14 M NaCl) and 10X PBS (1.4 M NaCl) but not between synovial fluid and 0.1X PBS (0.014 M NaCl). Bars are least square means ± 95% confidence intervals (n = 3 displacements × 3 fetlocks × 5 repetitions = 45).

During EAG measurements, fetlocks were subjected to displacements of 8 mm, 15 mm, and 25 mm, to approximate, respectively, half the load experienced during standing, standing, and walking. Average axial loads (n = 12) were 946 ± 161 N, 1935 ± 273 N, and 3240 ± 530 N, for 8 mm, 15 mm and 25 mm of displacement, respectively, which are within range of reported peak vertical forces.^29,33 Change in flexion angle (n = 12), measured manually with a goniometer, were 6.1° ± 0.4°, 11.4° ± 1.1°, and 18.1° ± 1.9° for 8 mm, 15 mm, and 25 mm of displacement, respectively. Significant differences among EAG coefficients due to displacement were detected for EAG2 and EAG3 but not for EAG4 or EAG5 ( Fig. 3 ). EAG1 and EAG6 measured significantly different EAG coefficients at 8 mm compared with 15 mm of displacement but not between 15 mm and 25 mm.

Figure 3.

Effect of displacement on electroarthrography (EAG) coefficients. EAG measurements were made at 6 sites while joints were subjected to displacements of 8 mm, 15 mm, or 25 mm, corresponding to average loads of 946 ± 161 N, 1935 ± 273 N, and 3240 ± 530 N, and change in flexion angles of 6.1° ± 0.4°, 11.4° ± 1.1°, and 18.1° ± 1.9°, respectively. Significant differences (P < 0.05) accounting for multiple comparisons were detected between displacements unless otherwise indicated as not significant (n.s.). Significant differences (P < 0.05) were detected for EAG2 and EAG3 but not for EAG4 or EAG5. EAG1 and EAG6 measured significantly different EAG coefficients at 8 mm compared with 15 mm of displacement but not between 15 mm and 25 mm of displacement. Bars are least square means ± 95% confidence intervals (n = 4 immersion buffers × 3 fetlocks × 5 repetitions = 60).

EAG coefficients were significantly reduced (P < 0.05) following trypsin treatment in 6 of 8 electrodes during standing and 7 of 8 during walking ( Fig. 4 ). QP, calculated from streaming potentials measured directly with the Arthro-BST, were higher (P < 10⁻⁵, n = 75), indicating lower cartilage stiffness, in trypsin-degraded cartilage compared with the contralateral control ( Fig. 5 ). Trypsin treatment reduced Em, representing proteoglycan matrix stiffness (P = 0.005, n = 6), but did not significantly alter Ef, representing collagen network stiffness (P = 0.15, n = 6), k (P = 0.24, n = 6), nor cartilage thickness (P = 0.76, n = 6) ( Table 2 ). Biochemical assays detected lower proteoglycan content (P = 0.0005) and no change in collagen content (P = 0.67) in cartilage isolated from the trypsin-treated fetlock compared with the contralateral control ( Table 2 ). Histology revealed a dramatic reduction in safranin-O staining due to trypsin treatment ( Fig. 6 ).

Figure 4.

Effect of trypsin degradation on electroarthrography (EAG) coefficients. EAG measurements were made at 6 sites pre- and post-trypsin on a single fetlock during (A) simulated standing (15 mm displacement), corresponding to peak axial loads approximately 1800 N and (B) simulated walking (25 mm displacement), corresponding to peak axial loads approximately 2870 N. Bars are least square mean ± 95% confidence interval for 5 loading cycles. *Indicates statistically significant differences (P < 0.05).

Figure 5.

Direct measurements of cartilage streaming potentials. Quantitative parameter (QP) maps were calculated from streaming potentials obtained with the Arthro-BST device at 75 locations on the phalanx and metacarpus in control and trypsin-treated fetlocks. QP is inversely proportional to cartilage stiffness so that higher values are interpreted as degradation. Joint surfaces from the trypsin-treated fetlock have higher QP, inferring degraded cartilage, compared with the contralateral control. Values are least square mean (lower 95% confidence interval, upper 95% confidence interval) for each joint surface. Circles indicate locations where cartilage disks were subsequently isolated for biomechanical testing and histology.

Table 2.

Cartilage Thickness, Biomechanical Parameters, and Biochemical Composition Measured on Cartilage Disks Isolated from Weightbearing Areas of Trypsin-Treated and Contralateral Control Fetlocks.^a.

	Normal Cartilage	Trypsin Treated
Thickness (mm)	0.65 (0.51, 0.79)	0.67 (0.63, 0.71)
Em (MPa)*	0.53 (0.33, 0.72)	0.13 (0.02, 0.23)
Ef (MPa)	4.21 (2.14, 6.27)	2.15 (1.34, 2.96)
k (mm³/N·s)	6.6 × 10⁻³ (2.1 × 10⁻³, 1.1 × 10⁻²)	3.3 × 10⁻² (0, 6.9 × 10⁻²)
Proteoglycan (µg CSC/mg cartilage)*	100.39 (80.12, 120.67)	33.92 (27.89, 39.96)
Collagen (mg HYP/mg cartilage)	0.36 (0.28, 0.45)	0.31 (0.12, 0.51)

Proteoglycan content determined by measuring chondroitin sulfate (CSC). Collagen content determined by measuring hydroxyproline (HYP). Values are least square mean (lower 95% confidence interval, upper 95% confidence interval), n = 6. Groups compared using paired t tests with statistical significance identified by *.

Figure 6.

Safranin-O/fast green/iron hematoxylin stained histological sections. Histology was performed on cartilage disks isolated from the contralateral control fetlock (A-F) and trypsin-treated fetlock (G-L). Cartilage disks were obtained from specific sites identified in Figure 5 , including the lateral aspect of the phalanx (A, G), medial aspect of the phalanx (B, H), lateral aspect of the metacarpus-phalanx articulation (C, I), lateral aspect of the metacarpus-sesamoid articulation (D, J), medial aspect of the metacarpus-phalanx articulation (E, K) and, medial aspect of the metacarpus-sesamoid articulation (F, L). Scale bars are 250 µm.

Discussion

The conditions under which cartilage streaming potentials are produced were controlled first, by changing the ionic strength of the immersion buffer, and second, by directly altering cartilage quality with trypsin. EAG responded in the same manner as direct measurements under similar conditions where increasing salt concentration or trypsin degradation reduces cartilage streaming potentials. These results demonstrate that EAG, measured through skin electrodes, is indeed measuring streaming potentials produced by cartilage during compression. These experiments were conducted on a total of 5 metacarpophalangeal joints, which is a small sample size and a limitation of the study.

EAG Responds Inversely to Ionic Strength of Immersion Buffer

EAG coefficients decreased as conductivity, reflecting ionic strength, of the immersion buffer increased ( Fig. 2 ). This finding agrees with those of other investigators^6,7,23 who have measured cartilage streaming potentials in similar conditions by directly contacting cartilage with electrodes.

Frank and Grodzinsky⁶ used silver–silver chloride electrodes to measure streaming potentials produced by mature bovine cartilage disks during sinusoidal compression in confined compression geometry. These measurements were performed in real time while the ionic strength of the bath was increased from 10⁻⁴ to 10 M NaCl. These authors found that streaming potentials decreased monotonically as bath ionic strength increased but not at a constant rate. The greatest change in streaming potentials occurred when ionic strengths were in the range of the fixed charge density of the bovine cartilage used, approximately 0.12 to 0.15 M. The present study used PBS solutions with total salt concentrations of approximately 0.015 M, 0.15 M, and 1.5 M for 0.1X PBS, 1X PBS, and 10X PBS, respectively. Similar to Frank and Grodzinsky,⁶ EAG coefficients ( Fig. 2 ) decreased monotonically with regard to ionic strength; however, these results differ from those of Frank and Grodzinsky in that the relative change in EAG coefficient between 0.1X PBS and 1X PBS is smaller compared with the change between 1X PBS and 10X PBS. EAG coefficients measured in anterior electrodes in 1X PBS were approximately 80% of those measured in 0.1X PBS and approximately 5 times larger than those measured in 10X PBS. These differences may be due to the dissimilar experimental setups, where, in the present work, immersion buffers were exchanged by intra-articular access to the joint space. Multiple lavages were employed to clear the joint space of the previous solution as fully as possible; however, complete elimination could not be achieved given the complexities of the joint capsule where small pockets and folds exist that may retain traces of the previous solution. These effects, where traces of the previous solution contaminate the present solution, are more likely to affect the lowest concentration of immersion buffer. Evidence of this was observed in the conductivity of 4.09 (3.67, 4.51) mS/cm ( Table 1 ) measured in 0.1X PBS retrieved from the joint compared with 1.70 mS/cm measured in the original 0.1X PBS solution. By contrast, Frank and Grodzinsky’s⁶ experiments were performed directly on isolated cartilage in a well-controlled environment where they were able to ascertain precisely when cartilage had come to equilibrium following a change in bath ionic strength.

Similar relationships between streaming potentials and ionic strength were also obtained from experiments where spatially resolved streaming potentials were measured. Quenneville et al.²³ measured streaming potentials with a microelectrode array directly compressing bovine cartilage disks immersed in salt solutions at various ionic strengths. Streaming potentials were approximately 4 times higher at 0.01 M NaCl and 5 times lower at 0.5 M NaCl compared to measurements obtained at physiological ionic strength, 0.15 M NaCl. Additionally, the relationship between spatially resolved streaming potentials, explored by using a linear array of microelectrodes in contact with isolated cartilage disks subjected to compression, demonstrated that streaming potentials produced by cartilage immersed in 0.15 M NaCl were 4 to 6 times lower than those produced by the same cartilage disks equilibrated in 0.01 M NaCl.⁷ Comparable relationships were observed in the present study, notably in EAG coefficients measured in anterior electrodes ( Fig. 2 ), where cartilage streaming potentials were obtained indirectly by electrodes placed on skin exterior to the articulation in response to compressive loading of the whole joint.

The streaming potential is a dynamic phenomenon that develops during cartilage compression. Interstitial fluid flows through the pores of the charged ECM in response to compressive load thereby displacing mobile positive ions relative to the fixed, negatively charged proteoglycan. The resulting electrical signal is influenced by the ionic strength of the bathing solution because of changes to the Debye length, defined as the thickness of the electrical double layer. At physiological ionic strength (~0.15 M NaCl), high electrostatic repulsion forces are generated because the Debye length is 0.8 nm, which coincides with the distance between fixed negative charge groups on chondroitin sulphate and half the distance between fixed negative charge groups on keratin sulfate. Lower ionic strengths mean that net charge occurs in the fluid phase at a greater distance from fixed glycosaminoglycan due to an increase in the Debye length resulting in higher streaming potentials, while higher ionic strengths confine charge density closer to glycosaminoglycan, resulting in shorter Debye lengths and decreased streaming potentials.^6,7,34,35

EAG Responds to Trypsin-Induced Cartilage Degradation

EAG coefficients, measured in the same fetlock before and after intra-articular injection of trypsin, were reduced when compressed to axial loads that approximated both standing and walking ( Fig. 4 ). Cartilage degradation following trypsin treatment was confirmed by direct assessment of cartilage streaming potentials, measured by compressing cartilage with a specialized indentation probe ( Fig. 5 ), as well as parameters obtained from biomechanical testing, biochemical assay ( Table 2 ), and histological appearance ( Fig. 6 ). These findings are consistent with the mechanism by which trypsin degrades the cartilage ECM.

Trypsin degradation has been used in a variety of models to simulate proteoglycan loss observed in early OA,^{8,24,26,28,36} where proteoglycan depletion begins at the articular surface and proceeds toward the subchondral bone.²⁷ Trypsin is a serine protease and cleaves proteoglycan at multiple sites along its core protein to produce 2 fragments.³⁷ The first fragment contains the G1 domain, which is a protein-hyaluronic acid complex ranging from 70 kDa to over 100 kDa,^38,39 and is critical to the aggregation of proteoglycan into large macromolecules. The second protein fragment includes a lengthy region between the G2 and G3 domains containing both chondroitin sulfate and keratin sulphate, which bear negative fixed charge groups.^37,38,40 Trypsin targets collagen to a lesser extent,^41,42 with Harris et al.⁴¹ measuring depletions of approximately 5% for hydroxyproline, compared with 70% for proteoglycan, in cartilage incubated in trypsin for 24 hours.

The present study is the first to demonstrate that EAG, acquired through electrodes placed on skin surrounding an articulation, is sensitive to cartilage degradation. Joint compression resulting in loads approximating standing and walking ( Fig. 4 ) produced EAG coefficients that were approximately 15% and 30%, respectively, lower than those measured in the same fetlock prior to incubation with trypsin. Direct measurements of cartilage streaming potentials, made with the Arthro-BST probe, correspond to EAG measurements, and reveal a general softening of cartilage ( Fig. 5 ). These streaming potential data are in agreement with biomechanical and biochemical assessments, which reveal that trypsin disrupted the cartilage ECM resulting in proteoglycan loss and reduced load-bearing capabilities ( Table 2 ). Histological sections ( Fig. 6 ) reveal a rather dramatic loss of proteoglycan, which may be expected to produce a greater reduction in EAG coefficients than what was observed ( Fig. 4 ). This apparent discrepancy may be due to the fact that histological sections were produced from cartilage disks after direct measurements of cartilage streaming potentials with the Arthro-BST probe and biomechanical testing. We suspect that proteoglycan fragments were released from the ECM during the time the joint surfaces and subsequently the isolated cartilage disks were submerged in PBS for these measurements. Proteoglycan content of the PBS solutions were not assessed. Furthermore, while trypsin disrupted proteoglycan by cleaving the large molecules into fragments, we could not confirm whether the fragments were released from the ECM during EAG tests when the joint was intact. It is possible that some fragments remained resident in the ECM and were able to provide some resistance to joint compression.⁹

EAG Demonstrates Potential to Provide a Non-invasive, Spatially Resolved Indicator of Cartilage Quality

Electrodes were adhered to skin symmetrically around the fetlock at the joint line and we observed that the amplitudes of EAG coefficients varied in a location-specific manner. EAG coefficients at anterior (dorsal) electrodes (EAG1 and EAG2) consistently had higher amplitudes than other sites. These electrodes were in closer proximity to cartilage surfaces undergoing compression compared to electrodes placed mediolaterally (EAG3 and EAG4) or at the metacarpus/sesamoid interface (EAG5 and EAG6). This pattern corresponds to reported load distribution in the fetlock,²⁹ where cartilage in the dorsal margin of the first phalanx, located immediately behind electrodes EAG1 and EAG2, experiences higher pressures during joint compression compared with the palmar margin. These data are also supported by a validated finite element model developed to better understand the forward problem of EAG.¹⁹ The model, which imposed current density sources of varying configurations in the medial and lateral knee joint compartments, demonstrated that EAG signals obtained at the skin surface were altered when the locations of source current density were changed. Sources placed close to the joint periphery produced well-defined maxima while sources located more internally resulted in broader maxima. EAG appears to reflect joint specific loading and is capable to a certain degree of providing spatially resolved information about cartilage quality.

As cartilage contact area is understood to increase progressively from the dorsal toward the palmar margins as compressive joint loads increase,^29,30 EAG coefficients were expected to be relatively stable at different displacements because they are normalized to axial load resulting from whole joint compression. However, we detected statistical differences due to displacement ( Fig. 3 ), suggesting non-uniform changes to cartilage involvement over the joint surfaces may occur as a function of displacement. Another contributor to these differences may be that the EAG coefficient is calculated for a complete loading cycle, meaning that it integrates information about cartilage quality over a range of joint angles. This suggests that an opportunity may exist for developing more sophisticated methods of analyzing EAG signals.

Conclusion

This study demonstrates that externally measured EAG is indeed detecting streaming potentials originating in cartilage during joint loading. The non-invasive EAG method has the potential to be developed into a sensitive diagnostic of cartilage health that may contribute to the early detection and treatment of OA and other degenerative joint diseases.

Footnotes

Authors’ Note

This study was performed at the Comparative Orthopaedic Research Laboratory, Department of Clinical Studies, University of Guelph.

Author Contributions

AC, MG, EQ, KG, PS, MDB, and MBH participated in study conception and design. AC and MG acquired and analyzed data. AC and SBB carried out statistical analyses. All authors participated in data interpretation. AC drafted the manuscript and it was critically revised by all authors. All authors approved the final version of the manuscript. AC takes responsibility for the integrity of the work as a whole.

Acknowledgments and Funding

We acknowledge technical assistance provided by Michele Beaudoin-Kimble, Felipe Garcia, Mohamed Hoba, and Sotcheadt Sim. Salary support for AC provided by the Natural Sciences and Engineering Research Council of Canada.

Declaration of Conflicting Interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: MG and EQ are the owners of Biomomentum Inc.

Ethical Approval

Procedures were performed with Institutional Animal Care Committee approval in accordance with the Canadian Council on Animal Care (University of Guelph, AUP#1412).

Animal Welfare

The present study followed international, national, and/or institutional guidelines for humane animal treatment and complied with relevant legislation.

ORCID iDs

Adele Changoor

Shelley B. Bull

References

Ryd

Brittberg

Eriksson

Jurvelin

Lindahl

Marlovits

, et al. Pre-osteoarthritis: definition and diagnosis of an elusive clinical entity. Cartilage. 2015;6(3):156-65.

Glyn-Jones

Palmer

AJR

Agricola

Price

Vincent

Weinans

, et al. Osteoarthritis. Lancet. 2015;386(9991):376-87.

Lotz

Kraus

VB.

New developments in osteoarthritis. Posttraumatic osteoarthritis: pathogenesis and pharmacological treatment options. Arthritis Res Ther. 2010;12(3):211.

Novakofski

Pownder

Koff

Williams

Potter

Fortier

LA.

High-resolution methods for diagnosing cartilage damage in vivo. Cartilage. 2016;7(1):39-51.

Viren

Saarakkala

Kaleva

Nieminen

Jurvelin

Töyräs

Minimally invasive ultrasound method for intra-articular diagnostics of cartilage degeneration. Ultrasound Med Biol. 2009;35(9):1546-54.

Frank

Grodzinsky

AJ.

Cartilage electromechanics—I. Electrokinetic transduction and the effects of electrolyte pH and ionic strength. J Biomech. 1987;20(6):615-27.

Garon

Legare

Guardo

Savard

Buschmann

MD.

Streaming potentials maps are spatially resolved indicators of amplitude, frequency and ionic strength dependent responses of articular cartilage to load. J Biomech. 2002;35(2_suppl):207-16.

Bonassar

Jeffries

Paguio

Grodzinsky

AJ.

Cartilage degradation and associated changes in biochemical and electromechanical properties. Acta Orthop Scand Suppl. 1995;266:38-44.

Frank

Grodzinsky

Koob

Eyre

DR.

Streaming potentials: a sensitive index of enzymatic degradation in articular cartilage. J Orthop Res. 1987;5(4):497-508.

10.

Legare

Garon

Guardo

Savard

Poole

Buschmann

MD.

Detection and analysis of cartilage degeneration by spatially resolved streaming potentials. J Orthop Res. 2002;20(4):819-26.

11.

Sim

Chevrier

Garon

Quenneville

Lavigne

Yaroshinsky

, et al. Electromechanical probe and automated indentation maps are sensitive techniques in assessing early degenerated human articular cartilage. J Orthop Res. 2017;35(4):858-67.

12.

Becher

Ricklefs

Willbold

Hurschler

Abedian

Electromechanical assessment of human knee articular cartilage with compression-induced streaming potentials. Cartilage. 2016;7(1):62-9.

13.

Changoor

Coutu

Garon

Quenneville

Hurtig

Buschmann

MD.

Streaming potential-based arthroscopic device is sensitive to cartilage changes immediately post-impact in an equine cartilage injury model. J Biomech Eng. 2011;133(6):061005.

14.

Changoor

Fereydoonzad

Yaroshinsky

Buschmann

MD.

Effects of refrigeration and freezing on the electromechanical and biomechanical properties of articular cartilage. J Biomech Eng. 2010;132(6):064502.

15.

Sim

Chevrier

Garon

Quenneville

Yaroshinsky

Hoemann

, et al. Non-destructive electromechanical assessment (Arthro-BST) of human articular cartilage correlates with histological scores and biomechanical properties. Osteoarthritis Cartilage. 2014;22(11):1926-35.

16.

Abedian

Willbold

Becher

Hurschler

In vitro electro-mechanical characterization of human knee articular cartilage of different degeneration levels: a comparison with ICRS and Mankin scores. J Biomech. 2013;46(7):1328-34.

17.

Mickevicius

Pockevicius

Kucinskas

Gudas

Maciulaitis

Noreikaite

, et al. Impact of storage conditions on electromechanical, histological and histochemical properties of osteochondral allografts. BMC Musculoskelet Disord. 2015;16:314.

18.

Preville

Lavigne

Buschmann

Hardin

Han

Djerroud

, et al. Electroarthrography: a novel method to assess articular cartilage and diagnose osteoarthritis by non-invasive measurement of load-induced electrical potentials at the surface of the knee. Osteoarthritis Cartilage. 2013;21(11):1731-7.

19.

Han

Buschmann

Savard

The forward problem of electroarthrography: modeling load-induced electrical potentials at the surface of the knee. IEEE Trans Biomed Eng. 2014;61(7):2020-7.

20.

Zhu

Buschmann

Savard

Mechanical loading of knee articular cartilage induced by muscle contraction can be assessed by measuring electrical potentials at the surface of the knee. J Biomech. 2016;49(3):338-43.

21.

Zhu

Garon

Quenneville

Buschmann

Savard

Decrease of the electrical potentials measured on the surface of the knee and produced by cartilage compression during successive loading cycles. J Biomech. 2016;49(14):3587-91.

22.

Zhu

Garon

Quenneville

Buschmann

Savard

Electrical potentials measured on the surface of the knee reflect the changes of the contact force in the knee joint produced by postural sway. Gait Posture. 2017;52:159-64.

23.

Quenneville

Garon

Légaré

Buschmann

MD.

Load and streaming potential responses of articular cartilage as a function of compression speed during indentation (Abstract). Paper presented at: Transactions of the 49th Annual Meeting of the Orthopeadic Research Society; February 2003; New Orleans.

24.

Wang

Zheng

Qin

Huang

Lam

Leung

, et al. Real-time ultrasonic assessment of progressive proteoglycan depletion in articular cartilage. Ultrasound Med Biol. 2008;34(7):1085-92.

25.

Korhonen

Laasanen

Toyras

Lappalainen

Helminen

Jurvelin

JS.

Fibril reinforced poroelastic model predicts specifically mechanical behavior of normal, proteoglycan depleted and collagen degraded articular cartilage. J Biomech. 2003;36(9):1373-9.

26.

DiSilvestro

Suh

JK.

Biphasic poroviscoelastic characteristics of proteoglycan-depleted articular cartilage: simulation of degeneration. Ann Biomed Eng. 2002;30(6):792-800.

27.

Moody

Brown

Bowden

Crawford

McElwain

Oloyede

AO.

In vitro degradation of articular cartilage: does trypsin treatment produce consistent results?

J Anat. 2006;209(2_suppl):259-67.

28.

Qin

Zheng

Leung

Mak

Choy

Chan

Ultrasound detection of trypsin-treated articular cartilage: its association with cartilaginous proteoglycans assessed by histological and biochemical methods. J Bone Miner Metab. 2002;20(5):281-7.

29.

Brama

Karssenberg

Barneveld

van Weeren

PR.

Contact areas and pressure distribution on the proximal articular surface of the proximal phalanx under sagittal plane loading. Equine Vet J. 2001;33(1):26-32.

30.

Easton

Kawcak

CE.

Evaluation of increased subchondral bone density in areas of contact in the metacarpophalangeal joint during joint loading in horses. Am J Vet Res. 2007;68(8):816-21.

31.

Soulhat

Buschmann

Shirazi-Adl

A fibril-network-reinforced biphasic model of cartilage in unconfined compression. J Biomech Eng. 1999;121(3):340-7.

32.

Hoemann

. Molecular and biochemical assays of cartilage components. In: De Ceuninck

Sabatini

Pastoureau

, editors. Cartilage and osteoarthritis: structure and in vivo analysis. Vol 2. Totowa, NJ: Humana Press; 2004. p. 127-56.

33.

Bobbert

Alvarez

CBG

van Weeren

Roepstorff

Weishaupt

MA.

Validation of vertical ground reaction forces on individual limbs calculated from kinematics of horse locomotion. J Exp Biol. 2007;210(Pt 11):1885-96.

34.

Buschmann

Grodzinsky

AJ.

A molecular model of proteoglycan-associated electrostatic forces in cartilage mechanics. J Biomech Eng. 1995;117(2_suppl):179-92.

35.

Basser

Grodzinsky

AJ.

The Donnan model derived from microstructure. Biophys Chem. 1993;46(1):57-68.

36.

Fan

Xiaofei

Pengling

Xiran

Yan

Yubo

, et al. Correlations between X-ray attenuation and GAG content of different cartilage layers based on contrast agent enhanced micro-CT. Conf Proc IEEE Eng Med Biol Soc. 2015;2015:6366-9.

37.

Heinegard

Hascall

. Aggregation of cartilage proteoglycans. 3. Characteristics of the proteins isolated from trypsin digests of aggregates. J Biol Chem. 1974;249(13):4250-6.

38.

Neame

PJ.

Extracellular matrix of cartilage: proteoglycans. In: Woessner

Howell

, editors. Joint cartilage degeneration: basic and clinical aspects. New York: Marcel Dekker, Inc; 1993. p. 109-38.

39.

Bonnet

Dunham

Hardingham

TE.

Structure and interactions of cartilage proteoglycan binding region and link protein. Biochem J. 1985;228(1):77-85.

40.

Fosang

Neame

Hardingham

Murphy

Hamilton

JA.

Cleavage of cartilage proteoglycan between G1 and G2 domains by stromelysins. J Biol Chem. 1991;266(24):15579-82.

41.

Harris

Jr Parker

Radin

Krane

SM.

Effects of proteolytic enzymes on structural and mechanical properties of cartilage. Arthritis Rheum. 1972;15(5):497-503.

42.

Laasanen

Töyräs

Hirvonen

Saarakkala

Korhonen

Nieminen

, et al. Novel mechano-acoustic technique and instrument for diagnosis of cartilage degeneration. Physiol Meas. 2002;23(3):491-503.