Hyaluronic acid–British anti-Lewisite as a safer chelation therapy for the treatment of arthroplasty-related metallosis

Significance

Cobalt-containing alloys are useful for orthopedic applications due to their high mechanical strength, corrosion resistance, hardness, low wear rates, and fatigue resistance. Unfortunately, these prostheses release significant levels of cobalt ions, a complication only discovered after widespread implantation. Current treatment necessitates costly revision surgeries. As an alternative, we have made major steps toward the development of a biocompatible, noninvasive, injectable chelation therapy that would delay or prevent the need for surgical revision.

Abstract

Cobalt-containing alloys are useful for orthopedic applications due to their low volumetric wear rates, corrosion resistance, high mechanical strength, hardness, and fatigue resistance. Unfortunately, these prosthetics release significant levels of cobalt ions, which was only discovered after their widespread implantation into patients requiring hip replacements. These cobalt ions can result in local toxic effects—including peri-implant toxicity, aseptic loosening, and pseudotumor—as well as systemic toxic effects—including neurological, cardiovascular, and endocrine disorders. Failing metal-on-metal (MoM) implants usually necessitate painful, risky, and costly revision surgeries. To treat metallosis arising from failing MoM implants, a synovial fluid-mimicking chelator was designed to remove these metal ions. Hyaluronic acid (HA), the major chemical component of synovial fluid, was functionalized with British anti-Lewisite (BAL) to create a chelator (BAL-HA). BAL-HA effectively binds cobalt and rescues in vitro cell vitality (up to 370% of cells exposed to IC₅₀ levels of cobalt) and enhances the rate of clearance of cobalt in vivo (t_1/2 from 48 h to 6 h). A metallosis model was also created to investigate our therapy. Results demonstrate that BAL-HA chelator system is biocompatible and capable of capturing significant amounts of cobalt ions from the hip joint within 30 min, with no risk of kidney failure. This chelation therapy has the potential to mitigate cobalt toxicity from failing MoM implants through noninvasive injections into the joint.

Orthopedic biomaterials have been successfully in use for decades and have significantly improved patients’ lives (1). When conservative treatments to relieve severe osteoarthritis or other joint degenerating injuries fail, joint repair or replacement (arthroplasty) is usually considered (2, 3). Globally, more than 2.9 million joint replacement procedures are conducted annually (4, 5); nearly 7.8 million hip and 1.1 million knee arthroplasties were reported to have been performed between 2012 and 2019 (6). Further reports indicated that about 270,000 total hip arthroplasties (THAs) are performed annually in the United States, with a projection of 600,000 by 2030 (7, 8). Internationally, hip implantations are expected to grow by a compound annual growth rate of 1.2% for the next 35 y (9). In 2017, the global hip replacement market was valued at $7 billion and is expected to increase to $9 billion in 2026 (10).

The most commonly used materials for joint replacement devices include metals, polymers, and ceramic composites (2, 11). Since they remain as permanent fixtures, ideally over the lifetime of the patient (12–14), implant longevity has become a critical design factor. Improved durability of orthopedic biomaterials has allowed younger and more active patients to consider THA. Patients expect implants to withstand high-impact activities and last decades (15). Considering this, significant focus has centered on improving the durability of joint prostheses to reduce implant wear and failure.

To meet the demand for improved durability, metallic alloys are desirable materials for THA due to their high corrosion resistance, high strength, hardness, (16, 17) and ability to withstand cyclic loading (16). Cobalt-based alloys, such as cobalt-chromium-molybdenum alloy (ASTM F799), (18) were widely implanted as metal-on-metal (MoM) implants for THA in the early 2000s, as these MoM implants were expected to have superior implant stability and less volumetric wear (19). Concerns emerged when reports indicated 1/8 revisions rate of all total hip replacements within 10 y of implant, in which 60% of them were attributed to wear/corrosion and metallosis (20). Metallosis is a condition that arises when particles released from the articulation of metal components accumulate to a toxic concentration. This is associated with pain and local inflammation in surrounding tissues and evidenced by dark discoloration surrounding the implant (21).

Metallosis is difficult to diagnose (22, 23) because concentrations of metal ions in serum or urine may not correlate with local metallic build-up associated with implant failure (24). In severe cases, systemic heavy metal poisoning results. While individuals with MoM implants have been shown to develop measurable levels of Co and Cr, it is unclear to what extent the local intra-articular concentration of these ions correlates to implant malfunction and joint failure (25). This debris can disseminate farther through tissues and organs leading to rare systemic side effects (26, 27), including cardiomyopathy, neurological changes (e.g., auditory or visual impairments), and thyroid dysfunction (28–30). Metallosis is also linked to extreme pain, localized aseptic fibrosis, necrosis, bone deterioration, and implant loosening (31, 32). Systemic distribution of toxic metal ions can cause carcinogenesis and death (22, 33, 34).

Current treatments for arthroplasty-related metallosis are limited to revision surgery, which is indicated for implant malfunction or failure (35), and the clinical outcome of a revision depends partly on the amount of bone and damaged tissues excised (35, 36). Revision is highly invasive, (37), more cost-intensive (38), and may fail earlier than the primary surgery (39). Furthermore, these procedures can pose a substantial risk and may not be feasible due to penitent comorbidities such as heart disease, kidney disease, liver disease, and prior or current cancer (40).

It has been speculated that a minimally invasive treatment administered after initially observed metallosis could improve a patient’s quality of life, and reduce the need for revision surgery (3). One such potential treatment is chelation therapy: the application of metal-binding molecules (chelators) to sequester and enhance the clearance of metal ions (41). This is a standard treatment for systemic metal poisoning. Although traditional chelation therapy can be effective in extreme metal intoxication, systemic administration of high dosages can be toxic to multiple organs, particularly the kidney (42). This usually prevents widespread indication of use because the benefits may not outweigh the risks (43). Additionally, high local concentrations of cobalt in the poorly vascularized joint space may not be fully cleared due to cobalt saturation of surrounding tissues (44).

1.1. Creation of BAL-HA Chelation Therapy

To increase the benefits-to-risks ratio of chelation therapy, a biomimetic chelator, called BAL-HA was created. A stable, covalent chemical bond was formed between a traditional small molecular chelator British anti-Lewisite (BAL) and hyaluronic acid (HA). HA (the “Bio”) is a polysaccharide and macromolecule naturally found in synovial fluid. The polysaccharide backbone of the chelator increases compatibility and bioavailability and speeds elimination. Macromolecules are known to affect compatibility and, in some cases, mitigate toxicity of compounds by serving as a carrier that hides or masks the toxic effects. Here, HA is already approved for injections into the intra-articular space to treat osteoarthritis. BAL (the “Binder”) is an approved chelator used for the treatment of heavy metal poisoning and the chemical bond to HA does not interfere with the ability to bind toxic metals. This BAL-HA chelation compound was developed for treatment of cobalt-induced metallosis in MoM hip implant patients, Fig. 1.

Fig. 1.

Schematic showing injection and action of our chelator (BAL-HA) in the intra-articular space of a MoM hip implant. This paper reports an animal model for metallosis as well as a noninvasive treatment option.

1.2. Creation of Metallosis Animal Model

To investigate the preclinical efficacy of our chelation therapy, an animal model (rat) of metallosis was developed that is highly valuable in testing the safety and efficacy of BAL-HA. This model could also be applied to the evaluation of other chelation therapies. It allows for evaluation of the biodistribution and elimination of cobalt ions following injection into the hip joints. The hip joints and major organs (liver, kidney, spleen, and heart) were analyzed to determine whether BAL-HA prevented tissue deterioration from cobalt ions toxicity. Furthermore, the off-target side effects and kidney health were investigated to confirm safety.

The findings reported in this work demonstrate that noninvasive, local injections of BAL-HA can be a safe treatment of metallosis from MoM implants. This approach could lead to the reduction, prevention, and treatment of metallosis-related surgical revisions.

2. Results

2.1. Synthesis and Characterization of BAL-HA.

BAL-HA was synthesized using 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC) and N-hydroxysuccinimide (NHS) to activate the carboxylic acids on HA and create a covalent bond to BAL (Fig. 2A). ¹H-NMR characterization displayed the emergence of chemical shifts associated with BAL and confirmed successful covalent coupling of BAL to HA (SI Appendix, Fig. S1A). Covalent functionalization can be confirmed via heteronuclear single quantum coherence (HSQC) NMR. In BAL-HA, carbon A (blue) (J value = 0.05) appears at 47.52 ppm. This contrasts with the control mixture of HA + BAL, where the corresponding carbon A (green) on BAL appears at 28.48 ppm (SI Appendix, Fig. S1B). The downfield shift of 19 ppm in ¹³C NMR, as shown by HSQC, suggests that the carbon center is in a more deshielding environment and is indicative of covalent conjugation. Further, the peaks at 28.48 ppm corresponding to unbound BAL do not appear in the BAL-HA ¹³C spectrum, confirming the effectiveness of the purification.

Fig. 2.

(A) Synthesis of BAL-HA. The polysaccharide carboxylic acids of HA are activated and functionalized with the chelator BAL. (B) SEM images of BAL-HA dialyzed individually with Ni²⁺, Cr³⁺, Cr⁶⁺, and Co²⁺, showing metal binding. (C, i) EDX data quantifying the attachment of cobalt to BAL-HA. (ii) The sulfur and nitrogen on BAL-HA are used to determine quantitative functionalization of BAL onto HA. (iii) Graphic illustrating 10:1 n/n nucleation of cobalt onto BAL-HA. (D) Inhibitory concentration comparison between BAL, BAL (when attached to BAL-HA), and BAL-HA. BAL-HA significantly increases the cytocompatibility of BAL to mammalian cells. (E) Images of fluorescently labeled subcellular compartments of NIH-3T3 fibroblasts exposed to 10 µg mL^–1 of BAL, HA, or BAL-HA for 24 h. (F) Cytocompatibility of HA, BAL, and BAL-HA. i) Cellular enumeration assessed using Hoechst 33342 to quantify DNA/nuclei content. ii) Ethidium homodimer-1 positive cells represent late apoptotic and late necrotic cells before detachment and indicate cytotoxicity. iii) Large fields of view (2 × 3 concatenations of 10× images) of fluorescence from Hoechst 33342 (blue, proliferation), Calcein AM (green, metabolism), ethidium homodimer-1 (red, dying cells), and phase contrast (grayscale, cellular morphology).

2.2. Qualitative Chelation Capacity of BAL-HA.

It was observed that BAL, which is a transparent liquid, forms strongly colored chelating complexes with cobalt and nickel ions (SI Appendix, Fig. S2A). The previous literature has reported this colorimetric response results from the formation of cobalt– or nickel–BAL coordination complexes; thus, colorimetry can be used as an indicator of chelation (45). Considering this, dispersions of BAL-HA in water were spiked with dilute concentrations of Co²⁺, Ni²⁺, Cr³⁺, and Cr⁶⁺ (SI Appendix, Fig. S2B). Cobalt and nickel solutions displayed a significant color change upon contact with BAL-HA, a phenomenon not observed with the unmodified HA control.

To qualitatively evaluate the strength of the chelation, the solutions were dialyzed against water to remove excess metal from the samples. Even after several days of dialysis and freeze-drying, BAL-HA spiked with cobalt and nickel remained dark brown, suggesting the retention of bound Co²⁺ and Ni²⁺ (SI Appendix, Fig. S2).

2.3. Quantitative BAL-HA Cobalt Chelation.

Energy dispersive X-ray spectroscopy (EDX) was used to quantify the capacity of BAL-HA to bind cobalt. EDX was chosen for elemental analysis (Fig. 2 C, i) due to its depth of analysis (~2 µm) compared to surface-level techniques. Nitrogen is only found in the HA backbone, while sulfur is found only in the BAL functionalization, so by calculating the ratio of those two elements, the functionalization can be easily determined (Fig. 2 C, ii). By comparing the cobalt to sulfur ratio, the amount of cobalt per BAL-functionalized unit can be readily determined. After chelation, BAL-HA was found to contain 10.3 cobalt atoms per 2.1 sulfur atoms, which equates to 4.9 cobalt atoms per BAL on average. The greater than stoichiometric equivalents can be explained by nucleation (Fig. 2 C, iii).

2.4. BAL-HA Improves BAL Cytocompatibility.

BAL-HA was found to have markedly better cytocompatibility than the small-molecule chelator, BAL. For preliminary studies, RAW 264.7 macrophages and NIH-3T3 fibroblasts were utilized because these cell lines approximate the cell types first encountered upon implantation. To test compatibility, the half maximal inhibitory concentration (IC50) of BAL-HA was measured and found to be 1,999 and 1,028 mg/mL for macrophages and fibroblasts, respectively. This is in stark contrast to BAL, which has an IC50 of 10.78 mg/mL for macrophages and 2.32 mg/mL for fibroblasts. When comparing the amount of BAL in BAL-HA to BAL directly, that is, when the amount of BAL as a part of BAL-HA is compared to neat BAL, BAL-HA is still significantly safer for cells. In this comparison, the IC50 of BAL in BAL-HA was 538.8 mg/mL for macrophages and 277.1 mg/mL for fibroblasts. The smallest observed increase in viability of cells by conjugating BAL to HA was in macrophages, thus, conjugation to HA makes BAL at least 50× safer to cells than neat BAL (Figs. 1 and Fig. 2 D, E, and F).

2.5. Metallosis Animal Model Development.

Cobalt is a major metallic cause of metallosis and occurs naturally in the diet in the form of vitamin B12. In order to reduce the interference of the cobalt contents from Vit B12 in regular rat diets, a custom cobalt-free diet was designed (Envigo TEKLAD). The initial cobalt ion evaluation was carried out via inductively coupled plasma mass spectrometer (ICP-MS) before injections. Rats fed this cobalt-free diet were found to have significantly lower cobalt levels after 7 d compared to those fed a regular rat diet (Fig. 3 A, i and ii) [P (*) < 0.05]. All rats in subsequent experiments were fed cobalt-free diets for 7 d before cobalt injection.

Fig. 3.

(A–C) Cobalt concentration evaluations. (A) The initial data on rats with cobalt-free diet in comparison to the regular diet. (A, i) Serum cobalt levels were significantly higher in rats fed on regular diet after 7 d P (*) < 0.05. (A, ii) Urine cobalt levels were significantly higher in rats fed on a regular diet after 7 d P (*) < 0.05. (B) Evaluation of cobalt concentration in the saline and cobalt groups. (B, i and ii) Serum cobalt concentrations at the 6th and 24th hours in the cobalt group were significantly higher compared to the saline group P (*) (**) < 0.05, 0.01, 0.001. (C) The cumulative urine cobalt excretion at the 6th hour for the cobalt group was significantly higher compared to the control, P < 0.05. (D–F) Data from repeated experiment on cobalt-treated rats in comparison with saline. (D and E) The serum and urine cobalt concentrations at the 3rd and 6th hours for the cobalt group were significantly higher compared to the saline group and to all other time points. The cobalt ions at both 3rd hour were significantly higher than the 6th hour P (**) (***) < 0.01, 0.001; (F) Evaluations of the internal organs (spleen, heart, kidney, and heart) on day 14 showed no significant cobalt detected in both groups; (G) Gross and Histological Evaluations. The gross macroscopic evaluation of the joints at postinjections day 14 depicts that the cobalt group had severe inflammation, discoloration, pseudotumor, and bone degeneration (blue arrows), while the saline group has none of these features of metallosis. The H&E and Safranin O-stained areas of tissue show inflammation and cartilage degradation (black circle). The IHC stained areas of colocalization show cobalt and inflammation (yellow) Scales: Gross image is 3 mm; micrograph is 20 μm; (H) Quantified Tissue Deterioration of the Pilot Study. The gross macroscopic evaluation, the H&E, Safranin O, and IHC clearly show that the tissue damage seen in the cobalt group was significantly higher compared to the saline group. P (***) < 0.001.

The ICP-MS results for both the saline control group and the cobalt group indicated that both had equivalent serum and urine cobalt levels at 7 d precobalt injection. However, the 6th and 24th hour postinjection evaluations showed significantly higher cobalt levels for the cobalt group compared to the saline group, and all other time points. The 6th-hour cobalt level was significantly higher than that of the 24th hour (Fig. 3 B, i and ii) P (*) (**) (***) < 0.05, 0.01, 0.001. The cumulative urine cobalt excretion in the cobalt group, starting from the 6th hour, was significantly higher compared to the baseline; however, none of the successive cobalt levels following the 6th hour were significant compared to the preceding times (Fig. 3C), P < 0.05.

Furthermore, this experiment was repeated to evaluate earlier time points in greater detail for the maximum cobalt ion diffusion from the joints as reflected in circulation/excretion. The ICP-MS data revealed that the cobalt group had higher cobalt concentration than the saline group. The serum and urine cobalt concentration at the 3rd and 6th hour postinjection were statistically significant (Fig. 3 D and E) P (**) (***) < 0.01, 0.001. The cobalt concentrations at the 3rd hour for both serum and urine were significantly higher than the 6th hour (Fig. 3 D and E) P (***) < 0.001. This may indicate that a rapid diffusion of cobalt to the circulation occurs at the hip joint.

The ICP-MS evaluations for the internal organs (spleen, heart, kidney, and heart) on day 14 indicated no significant difference between the cobalt levels detected in each organ in both treatment groups. This may indicate that most of the injected cobalt did not accumulate in these organs (Fig. 3F). Grossly, the macroscopic evaluation of the harvested hip joints at postinjection day 14 showed that the cobalt group had severe inflammation, tissue discoloration, pseudotumor, and evidence of subchondral bone degeneration, while the saline group had none of these metallosis-associated symptoms of high cobalt levels in the joint. The Hematoxylin and Eosin and Safranin O–stained areas of tissue show inflammation and cartilage degradation (black circle), while immunohistochemistry showed the colocalization of cobalt and inflammatory cytokine IL-1b (yellow) (Fig. 3G). The tissue/joint deterioration evaluation, quantified by ImageJ, shows that the tissue damage seen in the cobalt group was significantly higher compared to the uninjured saline control group (Fig. 3H).

The Hematoxylin and Eosin and Safranin O staining of the hip joint tissue demonstrates that test animals have severe joint inflammation, tissue necrosis, and damaged cartilage. The immunohistochemistry staining for the presence of cobalt, inflammatory cytokine IL-1b, or evidence of tissue damage also shows that the test samples had significant tissue inflammation. The tissue deterioration quantified by ImageJ reveals that the percentage of inflammation that occurred in the cobalt group was significantly higher compared to the saline group, while the preserved cartilage percentage shown by safranin O was significantly higher in the saline group (Fig. 3H) P (***) < 0.001.

Analysis of the internal organs (the heart, liver, kidneys, and spleen) found little or no fibrosis, necrosis, inflammation, or cysts, suggesting that there were no off-target effects from cobalt or BAL-HA. Most of the tissues scored zero for these evaluations, showing no detectable changes. However, the saline group had a nonsignificant 0.2 ± 0.5 score for liver inflammation; and the cobalt group had 0.2 ± 0.4 for liver steatosis, and 0.2 ± 0.5 for neutrophilic inflammation of the spleen, which are all nonsignificant changes SI Appendix, Fig. S4, Table 1: pilot study.

2.6. Metallosis Treatment for 7- and 14-d Evaluation.

After establishing the efficacy of the metallosis animal model, the BAL-HA chelation therapy was evaluated in vivo. The initial 7- and 14-d experiment revealed that the chelator compound can significantly reduce cobalt levels systemically and in the surrounding joint cavity. The concentration of cobalt in the cobalt group in comparison with the BAL-HA group indicated that the BAL-HA group had significantly higher serum cobalt and urine cobalt on day zero (Fig. 4 A, i and ii P (*), (**) < 0.05, 0.01.

Fig. 4.

(A and B) The cobalt concentration evaluation on days 7 and 14. (A, i and ii). The serum and urine cobalt concentrations of the BAL-HA group on day zero were significantly higher compared to the cobalt group, but all other time points. However, the days 1, 3, and 7 serum cobalt concentration and day 1 urine cobalt concentration of the Cobalt group were significantly higher than the BAL-HA group, indicating more retention of cobalt ion in the system P (*) (**) < 0.05, 0.01. (B, i and ii). The internal organs (spleen, heart, kidney, and liver) on days 7 and 14 showed no significant cobalt levels in both groups, indicating no accumulation of cobalt in the internal organs. (C) The gross and histology evaluations of day 14 study. The gross macroscopic evaluation of the joints on day 14 demonstrates that the cobalt group had severe inflammation and tissue deterioration (black circle), while the BAL-HA group has less inflammation. The H&E and Safranin O evaluations depicted areas of more inflammation (yellow circle) and cartilage degradation (yellow arrows) in the cobalt group than the BAL-HA group. Note: The crack in the BAL-HA tissue segment was a technical issue during staining (black arrow). The IHC staining depicted retained cobalt (red) and colocalization of cobalt and inflammatory cytokine IL-1b (yellow) in the cobalt group. (Scales: Gross image is 3 mm; micrograph is 20 µm.) (D) The quantified tissue deterioration for day 14 study. The gross macroscopic evaluation, the H&E, Safranin O, and IHC analysis show that the tissue damage seen in the cobalt group was significantly higher than the BAL-HA group on day 14. P (*) < 0.05, (***) < 0.001.

However, the days 1, 3, and 7 serum and day 1 urine cobalt concentrations in the Cobalt group were significantly higher than those of the BAL-HA group. On day zero (a few hours postinjection), the serum cobalt biodistribution indicated that chelator-treated rats had significantly higher cobalt ions in their circulation, compared to rats not given the chelator. Additionally, the urine cobalt excretion on day zero indicated that chelator-treated rats have significantly higher cobalt ion excreted at a mean elimination of 5,220 µg/L compared to 1,781 µg/L for rats not given the chelator (Fig. 4 A, i and ii) P (**) (*) < 0.01, 0.05 BAL-HA likely conjugated the cobalt ions and made them available in the blood, so they can be excreted by the kidney. Then, on days 1, 3, and 7, both groups had significantly reduced cobalt levels in circulation compared to day zero, however, the cobalt group retained significant cobalt in their system compared to BAL-HA. This could be because much of the cobalt ions in the BAL-HA group have been eliminated more rapidly on day zero, evidenced by the cobalt group’s significantly higher excretion on day 1. The ICP-MS evaluations from day 10 up to day 14 indicated extremely low quantities of cobalt ions in both groups, reflecting that most of the circulating cobalt was eliminated within the early days of the study (Figs. 3 A, ii and 4 A, i). Furthermore, the ICP-MS evaluations for the internal organs (spleen, heart, kidney, and liver) on days 7 and 14 indicated no significant cobalt ions in both groups, indicating that most of the circulating cobalt ions did not internalize to the organs, suggesting no off-target effects of BAL-HA treatment. (Fig. 4 B, i and ii). Gross macroscopic evaluations of the harvested hip joints on day 7 (SI Appendix, Fig. 5 A) and day 14 (Fig. 4C) show that the cobalt group had inflammation and tissue discoloration, while the BAL-HA group had intact and protected joint features. The quantified tissue/joint deterioration data depicts that the tissue damage in the cobalt group was significantly higher compared to the BAL-HA-treated group.

The immunohistochemistry staining for cobalt and inflammatory cytokine (IL-1b) shows that the cobalt group had significantly higher tissue inflammation. The tissue deterioration quantified for day 7- and -14 joints shows that the percentage inflammation and deterioration evaluation by H&E and IHC that occurred in the cobalt group were significantly higher compared to the BAL-HA. In contrast, the percentage of cartilage retained by the joint was significantly higher in the BAL-HA group, depicting cytocompatibility and joint protection (SI Appendix, Fig. 5 B and Fig. 4D).

The small-molecule off-target effect scoring on the internal organs that evaluated the heart, liver, kidney, and spleen revealed that none of these organs had any complications from the injected substances. Most of the tissues scored zero, showing no detectable changes. However, the cobalt group had mild liver inflammation of 0.3 ± 0.4 and 0.6 ± 0.3 scores for days 7 and 14, while the BAL-HA group had mild steatosis of 0.5 ± 0.5 scores for days 14. Both cobalt and BAL-HA groups had mild neutrophilic inflammation of the spleen (0.3 ± 0.6, & 0.6 ± 0.4), and (0.4 ± 0.8, & 0.2 ± 0.4) for days 7 and 14 respectively, which are all nonsignificant changes (SI Appendix, Fig. S4, Table 1: 7- and 14-day study).

2.7. Metallosis Treatment for 48-h Evaluation.

Given that a greater percentage of the injected cobalt was excreted within the first day postinjection, more evaluations were conducted to capture the earliest time points and rapidity of the bio-bind system in chelating cobalt from the joint cavity. A biodistribution and excretion experiment with time points of 30 min, 6 h, 18 h, 24 h, and 48 h revealed that BAL-HA can significantly bind cobalt ions in the surrounding joint cavity within the first 30 min. The ICP-MS results indicated that the BAL-HA group (chelator-treated rats) had a significantly higher serum cobalt concentration within 30 min postinjection and a subsequent significantly higher urine cobalt output at 30 min and 6 h P (*) < 0.05. However, none of the 18-h or 48-h cobalt values were significant between the groups (Fig. 5 A, i and ii). This trend supports earlier claims that the chelator system conjugates cobalt ions rapidly to make them available for excretion. Thus, BAL-HA conjugated cobalt ions within 30 min and made them available in circulation to be filtered out by the kidney between 30 min and 6 h. Furthermore, the ICP-MS evaluations for the internal organs (spleen, heart, kidneys, and heart) within 48 h indicated no significant quantity of cobalt in both groups, reflecting that most of the circulating cobalt did not lodge or internalize to the internal organs and that there was no off-target effect of the chelator system (Fig. 5 A, iii).

Fig. 5.

(A) The cobalt ion evaluation at the first 48-h study. (A, i and ii). The initial 48-h experiment revealed that the BAL-HA group has a significantly higher serum cobalt concentration within 30 min postinjection and a subsequent significantly higher urine cobalt output at 30 min and 6 h P (*) < 0.05. However, none of the 18-h to 48-h cobalt values were significant between the groups. (A, iii). The ICP-MS evaluations for the internal organs (spleen, heart, kidneys, and heart) indicated no significant quantity of cobalt in both groups. (B) The cobalt evaluation at the second 48-h study. (B, i and ii). The data indicate that the BAL-HA group has significantly higher serum cobalt concentration within 1 h, and at 3 h postinjection, with corresponding higher urine cobalt output in 1 h and at 3 h post injection P (*) < 0.05. (B, iii). The cumulative excreted cobalt evaluation indicated that the excreted cobalt at 6 h was significantly higher compared to the preceding times and also the 12 to 48-h time frame P (*) < 0.05, 0.01. (C) The gross and histological 48-h evaluation. The gross macroscopic evaluation of the joints at 48 h posttreatment revealed that the cobalt group had severe inflammation (black circle), while the BAL-HA group has less inflammation. The H&E and Safranin O staining depicts areas of inflammation, tissue deterioration (yellow circle), and cartilage degradation (black circle) in the cobalt group, more so than the BAL-HA group (yellow arrows). The IHC staining depicts a broader colocalization of cobalt and inflammatory cytokine IL-1b (yellow) in the cobalt group. (Scales: Gross image is 3 mm; micrograph is 20 μm.) (D) The total protein–creatinine panel (urine)—48 h. The comparison of each group to the saline (control) group at 48 h showed no significant difference between them at any time point in (i) the total protein, (ii) the creatinine, and (iii) the protein creatinine ratio. (E) The total protein–creatinine panel (urine)—day 21. The comparison of each group to the saline (control) group on day 21 showed no significant difference between the protein quantitative, creatinine, and protein creatinine ratio in all the groups.

After establishing that a greater percentage of the injected cobalt was excreted within the first 6 h postinjection, earlier time points between 6 h and 2 d postinjection were analyzed to gain further insight into the early pharmacokinetics of the system. The second experiment focused on biodistribution and excretion for time points of 1 h, 3 h, 6 h, 12 h, 24 h, and 48 h. The ICP-MS results revealed that the BAL-HA group had significantly higher serum and urine cobalt concentrations within 1 and 3 h postinjection (Fig. 5 B, i and ii) P (*) < 0.05.

However, there was no significant difference in cobalt concentration in the serum or urine between the groups after 6 h. This trend indicates that BAL-HA rapidly conjugated much of the cobalt within the first hour and enhanced the excretion by the kidney between 1 and 3 h. Furthermore, the cumulative excreted cobalt for the BAL-HA group increased rapidly up to 6 h and then leveled off (Fig. 5 B, iii) P (*) < 0.05, 0.01. The gross macroscopic evaluation of the harvested hip joints at 2 d postinjection shows that the cobalt group had inflammation and tissue discoloration, while the BAL-HA group has more intact and protected joint features (Fig. 5C and SI Appendix, Fig. 6A)

There was significantly more tissue/joint deterioration damage in the cobalt group compared to the BAL-HA treated group. The percentage of tissue inflammation and deteriorations revealed by H&E in the cobalt group was significantly higher compared to the BAL-HA group. The immunohistochemistry staining for cobalt, inflammatory cytokine (IL-1b), and evidence of tissue damage through the colocalization of cobalt and inflammatory signals shows that the cobalt group had significant tissue inflammation. The percentage of retained cartilage, visualized with safranin O, was significantly higher in the BAL-HA group, depicting cytocompatibility and joint protection (SI Appendix, Fig. 6B).

The small-molecule off-target effect analysis on the internal organs revealed that most of the tissues scored zero, showing no detectable changes; however, the cobalt group had mild liver steatosis of 0.4 ± 0.2 scores, while the BAL-HA group had mild inflammation of 0.4 ± 4.5 score within 48 h. Furthermore, both groups had mild neutrophilic spleen inflammation (0.6 ± 0.9 for cobalt, and 0.4 ± 0.8 for BAL-HA), but these values were not statistically significant (SI Appendix, Fig. S4, Table 1: 48-hour study).

2.8. Assessment of Renal Toxicity.

To assess the possibility the renal toxicity, urinary protein—a marker of kidney damage—was quantified in all 4 groups using a spot urine protein-to-creatinine ratio. There were no statistically significant differences between the groups (Fig. 5D). This suggests no acute kidney complication or failure risks posed by our macromolecular chelator.

To determine whether BAL-HA chelation therapy compromised renal function, serum was collected at 10 d postinjection, and creatinine and metabolites were measured. All 4 groups had comparable levels (no statistical significance) of creatinine, blood urea nitrogen, bicarbonate, calcium, ionized calcium, chloride, potassium, phosphate, sodium, total proteins, albumin, and globulin, consistent with preserved renal function (SI Appendix, Fig. 7). Kidney function was also assessed in animals at 21 d postinjection. Again, there was (Fig. 5E and SI Appendix, Fig. S8) no statistically significant difference among any of the treatment groups. These results further suggest that BAL-HA is not acutely toxic to the kidney.

3. Discussions

The millions of metallic hip prostheses that have already been implanted are still largely functional. Recently, it has been reported that metallosis can occur in a nonmetallic hip prosthesis, due to a fracture of the acetabular liner (metallic component) leading to abnormal metal–metal contact (46). Further investigation is warranted into ways to eliminate local intra-articular particles that build up and lead to implant toxicity and failure. Particles released from CoCrMo-based hip implants are mostly cobalt (around 68%) Co (47), and this debris can consist of metal–protein complexes, free metallic ions, oxides of metal/ inorganic metal salts, and organic metal forms (e.g., hemosiderin) (47, 48). Reports indicated that the observed in vivo toxicity is a cumulative effect of these forms of debris generated by implant (49); however, the soluble cobalt (Co²⁺) ions are the most cytotoxic, with the ability to damage DNA (genotoxicity) (50).

3.1. Chelation Therapy.

Chelation therapy, proposed as a prophylactic intervention, to protect surrounding tissues from inflammatory responses to metal ions/debris, and thereby prevent revisions, has been under investigation (3, 46, 51, 52). Previously known chelators attempted for arthroplasty-related metal toxicity include ethylenediaminetetraacetic acid (I.V) (53), 2,3-dimercapto-1-propanesulfonic acid (54), dimercaprol (BAL) (55), and N-acetylcysteine (NAC) (41) administered orally, intramuscularly, or intravenously (55). Gilbert et al. (56) reported the use of BAL on a 52-y-old male with a malfunctioning MOP with clinical manifestations, who received one three-day cycle of therapy, and the results indicated that the initial systemic Co blood level of 1,085 mcg/L was decreased by 33% (57). BAL is itself toxic with a narrow range of therapeutic use, with side effects such as pain at the site of injection and tendency to concentrate arsenic in some organs (58). Another report described a 75-y-old male with a revised malfunctioning MoM hip after 60 mo, who received oral and intravenous NAC therapy. The result revealed that Co/Cr decreased by 51% and 40%, respectively (55, 56). Furthermore, D'Ambrosi et al. (41) also reported that N-acetylcysteine reduced systemic blood cobalt and chromium levels in patients with MoM hip arthroplasty (41). However, excessive amounts of localized metal ions and particles produced in the intra-articular spaces still proved difficult to chelate, limiting the success of these therapies on arthroplasty. Our technique depicted a localized chelation therapy that targets the intra-articular space with evidence of metal ion chelation ability, joint protection from metal toxicity, and surrounding tissue deterioration. BAL-HA has the potential to revolutionize the field of chelation therapy as a much safer but still effective alternative to BAL. The animal model for metallosis presented in this work effectively demonstrated BAL-HA’s ability to clear cobalt from the body and significantly lessen the effects of cobalt poisoning without causing the typical side effects seen in chelation therapy with BAL alone.

3.2. BAL-HA Functionalization Chemistry.

Covalent chemistry was used to link a natural polysaccharide with a traditional chelating agent to improve compatibility compared to the clinical chelating agent alone. The toxicity of traditional chelators is in part due to their small size, which provides them access to tissues where they are toxic. For example, BAL has been documented to redistribute heavy metals to the brain (59). Covalently linking the chelator to a large biocompatible molecule limits its area of travel while allowing the body to recognize the conjugate as the larger polysaccharide and metabolize it effectively. Dynamic light scattering and zeta potential measurements provided insight into the effects of BAL modification on polymer crosslinking and electrostatic charge, respectively (SI Appendix, Fig. S9). These experiments showed that BAL-HA has substantially larger particle sizes than the HA starting material. Zeta potential measurements reveal that BAL-HA has a lower electrostatic charge than unmodified HA due to exchange of the charged carboxylate for neutral BAL. Lowered zeta potential may also in part be due to a potential side reaction of EDC/NHS coupling that results in crosslinking, where a charged carboxylate and alcohol form a neutral ester. Furthermore, the biomimetic portion of the macromolecular chelator improves compatibility, while the functionality of the chelator remains uncompromised. To create this covalent conjugate, an accessible, versatile, and scalable coupling chemistry was applied. Carbodiimide coupling chemistry is used to activate carboxylic acid functional groups on either the polysaccharide or chelator, which can then be attached by nucleophilic functional groups on the chelator or polysaccharide, respectively. With control over type of polysaccharide, type of covalent linkage, and type of chelator, there is a range in properties that can be selected for each intended application. Preliminary experiments identified one leading blend for cobalt chelation, BAL-HA. The carboxylic acids of the polysaccharide HA are activated using EDC/NHS chemistry to conjugate BAL (Fig. 2A).

3.3. Synthesis and Characterization of BAL-HA.

BAL was chosen as the chelator since it is approved for clinical use by the US Food and Drug Administration (FDA) and can chelate cobalt ions. HA was chosen as the polysaccharide since is the principal component of synovial fluid and is also FDA approved and widely used clinically (60, 61). To confirm covalent attachment, an unconjugated mixture, HA + BAL was analyzed; this was prepared using typical reaction conditions, but without EDC/NHS, and followed by purification. The shift in carbon A in the ¹³C-NMR spectrum on BAL supports that EDC/NHS coupling effectively uses the thiol on BAL to covalently couple to HA (Fig. 2A). This is in accordance with a report showing that primary thiols are stronger nucleophiles than primary alcohols and secondary thiols due to a higher polarizability and steric hindrance, respectively (62). Therefore, it is most likely that the primary thiol serves as the main nucleophilic moiety for chemical modification.

3.4. Qualitative Chelation Capacity of BAL-HA.

When the covalent bond is created between BAL and HA to create the chelation system, lone pairs remain accessible on the BAL, which allows BAL-HA to chelate cobalt effectively, as evidenced by formation of a colorimetric complex (SI Appendix, Fig. S2). Further, HA contains primary alcohols on the polymeric backbone that could participate in chelation. With the demonstration of cobalt chelation, it is likely that the polymeric system retains the capacity to additionally chelate chromium and nickel (63–65), further improving the potential of our chelators to rescue a failing MoM implant patient from systemic metallosis.

3.5. Quantitative BAL-HA Cobalt Chelation.

Nucleation has been evidenced by the quantitative data as well as imaging. It has also been reported in the literature as a known phenomenon with heavy metals (66). Chitosan, a polysaccharide, has also been reported as a nucleating agent for heavy metals (67). Quantitative and qualitative data strongly suggest that nucleation is the mechanism by which BAL-HA is able to achieve such a high chelation capacity.

3.6. BAL-HA Improves BAL Cytocompatibility.

As BAL is known to have significant toxicity even at low concentrations, it was imperative to assess the cytocompatibility of BAL-HA. To accomplish this, the two mammalian lines NIH-3T3 Fibroblasts and RAW 264.7 Macrophages were treated with a range of concentrations of BAL and BAL-HA to obtain an IC50 (the concentration of the material which yields a 50% cell viability) for both materials. The concentrations were standardized for the amount of BAL, such that BAL and BAL-HA results were directly comparable. Fibroblasts are a primary component of connective tissue and play a role in wound healing (68). Macrophages are immune cells that play a role in the innate immune response to a foreign material (69). This makes them appropriately representative cell types to assess the safety of an intra-articular injection. Neat BAL was found to have significant toxicity, while BAL-HA was found to have high cytocompatibility, effectively increasing the cytocompatibility of BAL (Fig. 1F).

3.7. Animal Model Development.

For a preclinical evaluation of the chelator, an animal model was essential (70), Previously reported models for hip implants include intra-articular injection of particles to reproduce tissue response around a failed arthroplasty (70, 71), and more recently, a model for arthroplasty pain (72). Some of these models, such as the air pouch method, involved the direct introduction of surrogate particles into muscles and the knee joints (73) to evaluate the in vivo local tissue effects postinjection of substances (73). However, to date, no animal model has been developed to evaluate both local and systemic effects of Co ions from MoM hip arthroplasty. Well-controlled experimental conditions were performed with a clinically relevant, reproducible animal model to evaluate the hypotheses that BAL-HA can chelate cobalt in vivo and that it can reduce the negative local and systemic effects of cobalt poisoning. Various small animals were considered. Mice have limited utility due to their small hip joint that poses difficulty to penetration and the containment of injected solution without capsular rupture. The movement of rabbits may not reproduce the cyclic motion of the human hip joint. Thus, rats were chosen due to the relative ease of handling and sufficient joint space to contain the injected solutions. Studies have also demonstrated the similarity between the rat hip and human hip in capsular insertion and the distal femoral neck, head-neck angle, and profile of the lesser and greater trochanters (74). In addition, similar investigative goals have demonstrated that murine synovial joints have the capability to undergo inflammatory changes that are close to humans’ response to degradation of MoM implants (75–77).

3.8. Hip Joint Injection.

The injection technique relied on both the bony and soft tissue anatomy, fashioned after the approach used for pediatric hip arthrograms in the management of hip dysplasia and performance of closed hip reduction with confirmation using a nontoxic contrast agent (78–80). Previous studies have demonstrated greater reproducibility in performing intra-articular injections by targeting the articular recesses instead of the articulating joint space itself (70, 73, 81). It was found that focusing the needle tip toward the inferior neck which lies just proximal to the joint capsule allows for better tactile feedback to confirm appropriate depth, compared to the often-occurring penetration of the cartilaginous surface when aiming for the joint, and increased resistance to flow causing subsequent capsular disruption or extra-articular injection after repositioning of the needle from the cartilage surface. An anterior approach was used in a similar fashion to commonly performed pediatric arthrography as it minimized the amount of soft tissue encountered compared to a posterior approach (82, 83), while allowing the rat to be placed supine without additional manipulation of the limb to adequately access the articular recess (SI Appendix, Fig. S10).

To accurately determine the position of the needle with respect to the articular recess, a perfect circle technique was utilized where the visualization of the needle tip as a single point confirmed the needle position in 2 dimensions with the depth being confirmed by the tactile feedback achieved by reaching the level of the femoral neck. Successful hip injections were represented as an oblong collection of radiopaque fluid with a well-defined border overlying the hip joint. It is recommended that this orientation is confirmed on orthogonal views and after manipulation of the hind limb. Rupture of the joint capsule or extra-articular fluid placement is demonstrated by more diffuse positioning of the fluid within the anterior compartment of the proximal thigh and additional fluid displacement after hind limb manipulation.

3.9. Establishment of Metallosis.

To establish a significant pathology resembling adverse local tissue reaction to implant degradation particles, previous reports have injected high doses of metallic particles (70, 73, 81, 84). Wang et al. had three injection concentrations of CoCr particles (81). The thigh-dose group of 500 ppb CoCr particles [1.25 mg/mL (500 μg/kg)] had the highest serum cobalt with significant adverse local and systemic effects (81). Following this experiment, two groups were chosen: a low dose of 100 μg/mL and a high dose of 500 μg/mL hip injection. However, owing to no observable adverse tissue reaction on the low-dose group, the model was developed with only the high dose containing 500,000 ppb cobalt. Previously, a comparison of wear particles injected in experimental models indicated that most animal studies had doses several thousand (5 × 10⁶ to 2.87 × 10⁶) folds greater than the actual detected no observed adverse effect level and lowest observed effect level from patients with MoM hip implants (73, 85, 86).

The pilot studies depicted that out of the total 500 μg/mL injected cobalt per experiment, the test group (cobalt) had about (335.7 μg/L and 706 μg/L) in circulation from the first and second experiments, respectively, while the control (saline) group had 118 μg/L and 0 μg/L. The excretion profile showed that the cobalt group had significantly higher urine cobalt levels, which was expected, while the saline group had no detectable cobalt in the urine. The biodistribution and elimination profile were in accordance with some earlier reported patterns of injected particles (70, 73, 81); however, elimination was found to be more rapid here. There was no significant detectable cobalt in the internal organs, with no significant off-target effects being observed. A similar result was reported by Lu et. al (87). From the injected concentration, it was observed that 92156.89 μg/L and 122619.8 μg/L were eliminated from the cobalt group, while 407,843.11 μg/L and 377,380.2 μg/L were retained in the hip joint and surrounding tissues. This was also supported by the histology results which depicted a severe tissue deterioration as seen in adverse tissue reactions to metal debris (88, 89). The local inflammatory response and cytokine mobilization to metal debris were confirmed by the colocalization of the antibodies to inflammatory cytokine IL-1b. These data correspond to the wear debris-induced local inflammatory response seen in retrieved implants (3, 89).

3.10. Treatment of Metallosis.

For the treatment studies, injected 1 mg/mL of BAL-HA was injected 5 min after 500 μg/ml of cobalt. The 7th- and 14th-day evaluations revealed that BAL-HA had significant cobalt chelation capacity, leading to a significant excretion within 24 h. This was reflected by the significant serum cobalt biodistribution in the chelator group, compared to the cobalt group. Hence, BAL-HA conjugated cobalt ions and made them available for excretion. On days 1, 3, and 7, both groups had significantly reduced cobalt levels in circulation compared to day zero; however, the cobalt group still had significant cobalt in their system compared to BAL-HA, which was attributed to the conjugation and enhanced cobalt ions elimination effect in BAL-HA group. This short-term evaluation supported the project hypothesis that the chelator would aid in the elimination of cobalt from the body; however, the speed at which this occurred was a concern. Just as reported in this animal model development, no significant cobalt was detected in the internal organs, which accords with earlier reports (47, 73, 87). From the injected concentration, it was found that 5,783.59 μg/L was eliminated by the BAL-HA group in 14 d, while 2,450.29 μg/L was eliminated by the cobalt group. This was also supported by the histology results that depicted cartilage destruction and inflammatory deterioration in the cobalt group, significantly more so than the BAL-HA group (31, 85). The local inflammatory response and cytokine mobilization to metal debris (49, 90) also confirmed that more inflammation occurred at the cobalt group, while the BAL-HA group had insignificant inflammation, which implicates good tissue protection.

An acute evaluation was pursued to ascertain the rapidity of cobalt ion conjugation and elimination by BAL-HA. It served also to measure the rapidity of local inflammatory cytokine mobilization to metal debris. Owing to the difficulty in obtaining urine at short time points corresponding with blood collections, two separate studies were conducted to capture important times. The initial 48-h study revealed that BAL-HA rapidly conjugated cobalt ions within 30 min, as by a higher serum cobalt concentration and a subsequent significantly higher urine cobalt output at 30 min and 6 h. After establishing that a greater percentage of the injected cobalt was excreted within 6 h postinjection, the second 48-h evaluation captured earlier times points, in which the chelator group had significantly higher serum cobalt at 1 h and 3 h, with corresponding significantly higher urine cobalt output at 1 h and 3 h. Combining data from both studies revealed a trend of BAL-HA rapidly conjugating cobalt, and enhanced excretion from 30 min to 6 h postinjection. The 6th-hour postinjection was the maximum time for the chelation as supported by cumulative excretion data. From the injected concentration in each 48-h experiment, it was found that 176,449.9 μg/L and 283,172.5 μg/L were eliminated from the BAL-HA groups, while the cobalt groups had 143,191.8 μg/L and 103,819 μg/L eliminated from the first and second studies, respectively. These data supported the efficacy of BAL-HA in enhancing the elimination of localized cobalt at the hip joint. The histology results also revealed the protective effect from cartilage destruction and inflammatory deterioration. The local inflammatory response also confirmed that more inflammation occurred in cobalt groups (91, 92), while BAL-HA showed tissue protection from cobalt toxicity.

3.11. Off-Target Effect and Renal Toxicity.

A high dose of metal ions and/or chelators can infiltrate internal organs and elicit serious pathological features (20, 47, 81, 93). Analysis of the liver, kidney, and spleen analysis revealed no significant changes in tissues from all groups. This is similar to published findings, which showed limited concentrations of metals in internal organs from high-dose metallic particles (73, 86). One possible explanation may be the rapidity of the cobalt ion excretion in all groups (73, 94). Moreover, the mobilization of cobalt with BAL-HA likely had a protective effect on tissues, though some cobalt ions were retained within the joint.

Earlier studies report deleterious effects of chelation therapy on kidney health (42, 43). To evaluate the effect of our chelator on renal toxicity, a comprehensive renal function analysis was performed. Specifically, a total protein–creatinine panel was run for urine samples, which analyzed the total proteins, creatinine, and protein–creatinine ratio (95, 96), and a small animal renal panel was run for the blood samples, which analyzed the total proteins, albumin, globulin, A/G ratio, creatinine, urea nitrogen, bicarbonate, calcium, ionized calcium, chloride, potassium, phosphate, sodium, and cholesterol (97, 98).

Due to limited blood availability from the small animal model, early time points to assess acute renal toxicity were only evaluated for the total protein–creatinine panel for urine samples. No significant difference between the groups was recorded, specifically BAL-HA and the controls (saline). These data suggested that no acute renal toxicity is posed by BAL-HA.

At 10- and 21-d time points, rats were killed, allowing collection of sufficient blood volume for further testing. The renal panel performed on blood samples showed no significant difference in all proteins and electrolytes between the groups. The values were within the normal ranges according to earlier records (98), and they suggested no risk of renal complications posed by our chelation system. Both the total protein–creatinine panel and the small animal renal panel were conducted at the conclusion of the 21-d study, and the results indicated that no significant difference between the groups. These results were within the normal ranges in accordance with earlier reports (96, 98, 99) and support the claim that BAL-HA is not toxic to the kidney.

3.12. Limitations of the Study.

The limitations of this study include particle source and physicochemical properties, type of degradation product, dosing regimen, and rapid clearance of injected substances.

To ideally test tissue interactions of MoM degradation debris in a model, the characteristics of the degraded particles or ions injected into the animals should be similar and comparable to the standard wear debris from MoM patients. In this study, cobalt hexahydrate was dissolved in PBS, and this does not perfectly match the chemical composition of soluble cobalt ions generated by degrading MoM implants. In addition, although cobalt ion is the most cytotoxic/genotoxic wear debris, compared to micro- and nanoparticles, in an ideal metallosis from MoM, the biological manifestations are the result of a combined effect of all degradation products from the metal implant. Hence, injecting only cobalt does not truly represent the actual metallosis observed in patients.

To create a pathology similar to adverse tissue reactions to metal debris, a very high dose was required, thus, a single high dose of cobalt was injected into the joint space. However, data from such a dose that exceeded the realistic range of metal ions from malfunctioning MoM implants could cast doubt on whether the results reflect a sheer volume of material overload. Thus, administration of a large bolus of soluble cobalt may not be perfectly representative of the chemical characteristics and gradual release of particles observed in MoM implant patients.

Furthermore, the rapid clearance of the injected cobalt was a concern. A great percentage of the injected cobalt was eliminated within 24 h. In an ideal adverse reaction to metal debris and associated metallosis, the biological manifestations are the result of a gradual buildup of metallic particles and ions. A gradual series of injections to achieve a toxic level was considered; however, such a design would cloud the interpretation of the actual chelation effect of the BAL-HA. Hence, injecting only a specified one-time cobalt dose, and the chelator, helped to more directly assess the efficacy of the chelation therapy. In a human trial, the therapy may be designed for multiple time point injections of BAL-HA to contain the continuous generation of debris and ions seen in metallosis in MoM patients.

3.13. Future Direction.

In the future, a slow-release animal model of metallosis comprising both cobalt ions and metal debris particles could be developed. With this, the various components of the degraded products in patients with MoM implant could be more accurately represented and assessed. Additionally, the gradual process of metallosis development could be reflected. A slow-release chelation system will also be considered, where the chelator will be coated with a biomaterial that would allow a gradual release for long-term chelation of the metal ions and particles.

Furthermore, in the pursuit of a lasting treatment to osteoarthritis that would prevent the need of arthroplasty at the end-stages, regenerative engineering, a field that is established from the application of biological, chemical, and physical engineering principles toward the repair, regeneration, and restoration of living tissues by the combination of cells, biological cues, and biomaterials, has been projected to proffer a solution (100–107). This field can also be defined as the merger of tissue engineering, stem cell science, advanced materials, and developmental biology (103, 108–115). Currently, the therapeutic prospects of regenerative engineered products for osteoarthritis have been on the rise globally (108, 111, 116–126). However, some of the protocols of these products are still short of FDA requirement for clinical application (127, 128).

3.14. Conclusion.

A biocompatible chelator blend BAL-HA was successfully developed and assessed in this work. An animal model of metallosis that reflected the pathologies seen in adverse reaction from arthroplasty wear and corrosion debris was also successfully developed and utilized. Multiple evaluations indicated that the chelator system was efficacious in conjugation and elimination of cobalt in vivo. The off-target effect and kidney failure evaluations confirmed that the chelation therapy is safe, with no risk of complication. Overall, the data indicated that BAL-HA has the potential to mitigate cobalt toxicity caused by metal ions from failing MoM implants through noninvasive injections into the joints.

4. Materials and Methods

Full methods in SI Appendix.

4.1. Synthesis of BAL-HA.

Using EDC/NHS coupling chemistry, BAL-functionalized HA was synthesized from HA sodium salt by dissolving 100 mg of HA in deionized, followed by the addition of 505.4 mg of EDC hydrochloride.

4.2. Cell Culture and Bioassays.

NIH-3T3 murine fibroblasts and RAW 264.7 murine macrophages were utilized for cellular vitality assay. Cobalt toxicity and cytocompatibility of BAL and BAL-HA were assessed by exposing cells to varying concentrations of these materials (SI Appendix, Fig. S11).

4.3. Animal Model Development and Treatment.

Ethics approval was granted by the University of Connecticut Health Institutional Review Board (Institutional Animal Care and Use Committee protocol number #101786-0222). Six-month-old male breeders (CRL- Crl: CD (Sprague Dawley), weighing 500 to 600 g were utilized. The sham (positive control) received 0.9% saline hip injection; the control (negative control) received 0.3 μg/g body weight (BW) ⁵⁹Co ions; while the experimental (test group) received 0.3 μg/g (BW) ⁵⁹Co ions, followed by BAL-HA at 0.05 μg/g (BW) 5 min afterward. Blood, urine, and internal organs (spleen, heart, liver, and kidney) were harvested at designated times for evaluation (SI Appendix, Figs. 4 B, Table 2, S10, and S12). Data were analyzed using multifactor ANOVA with grouping factors of treatment and time followed by Tukey’s honestly significant difference (HSD) test for making pairwise comparisons.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix. Data have been deposited in Dryad (129).