A multi-modal strategy for enhancing titanium–bone interface strength and antimicrobial properties of PMMA bone cements

Abstract

Infection prevention and retention at the bone-device interface are critical for long-term stability of titanium-based biomedical implants. This study utilized a multifactorial approach to assess bond strength and antibacterial activity of polymethyl methacrylate (PMMA) bone cements modified with polymerizable antimicrobials (quaternary ammonium methacrylate, QAM) for sustained antibiofilm effects. Four commercial dental adhesives (Gen 4, Gen 5, Gen 6, Gen 7) were tested between porcine cortical bone and PMMA bone cements (Jet Acrylic or Biomet Bone Cement R). Shear bond strength (SBS), interface analysis (SEM, FTIR/ATR) and cytotoxicity (MTT) were evaluated and compared to controls without dental adhesive, which is the current standard of care. Top performing materials were subjected to stringent mechanical cycling in a bioreactor, followed by SBS assessment. Finally, bone cements were modified with QAMs and evaluated with or without dental adhesive for physical properties, antibiotic activity and SBS. Refobacin was used as an antibiotic-containing commercial control. Results showed more than a twenty-fold increase in SBS after 24 h and up to a fifteen-fold improvement after mechanical cycling with the Gen 6 adhesive, compared to commercial control cements. Refobacin and the QAM-modified cement significantly reduced S. aureus biofilm biomass, but the QAM cements were significantly better able to maintain antimicrobial activity after water storage. Even though the shear bond strength was reduced with either QAM or Refobacin, the presence of the adhesive mitigated this reduction in comparison with the Refobacin applied without the adhesive. Overall, the results suggest that the application of dental adhesives prior to cementation provides an avenue for improving implant retention and that incorporating QAMs may offer a means of sustained infection control.

1. Introduction

The main complications involving implanted biomedical devices are recurrent infection [1] and loss of retention at the body-device interface [2] leading to, for example, cancellous bone formation following hip replacement [3]. Improving the prognosis of such interventions is a critical need, since at least 7 million people in the US alone are living with a hip or knee replacement [4]. According to the 2022 American Joint Replacement Registry, around 2.5 million total hip arthroplasty (THA) and total knee arthroplasty (TKA) cases were submitted from 2012 to 2021, with around 4% being revision procedures for each joint classification. Overall, while the success rate for these procedures is high, given the total volume of surgeries each year there is still a significant population who could benefit from increased prosthetic retention and infection prevention. However, current strategies to improve prognosis are not without drawbacks. FDA-approved polymethyl methacrylate (PMMA) bone cements have been used for over 60 years and although they function as a fast primary fixation of implant to bone, they do not guarantee a mechanically and biologically stable interface [5]. Depending on the substrate, the ability of the cement to penetrate into the bonded surface varies widely, with lower micromechanical retention and presumably lower stress resistance for a flat surface compared with a micro-patterned counterpart [6]. In addition, the use of metallic plates and screws can increase the rate of infection [7–9], the risk of fracture due to induction of mechanical stress, damage to the surrounding tissue and inhibition of bone remodeling, all of which require follow up surgery for correction [10,11]. A few studies have proposed the use of cyanoacrylate-based bone adhesives to improve interaction with tissues and increase retention without requiring additional features such as screws but results in vitro are conflicting [12–14].

High-strength, crosslinkable adhesives have been developed for other biomedical applications, for example in Dentistry. These materials are ubiquitously used to bond composite-based restorative materials to mineralized tissues in the tooth, with very efficient hybridization of the mineralized, collagen-rich tissues in the dental substrate [15], leading to clinical longevity increasing many-fold over the non-bonded counterparts [16]. Given the structural similarity between enamel/dentin and cortical/trabecular bone [17,18], it could be envisioned that these materials and procedures could be translated to bone applications. Bonding procedures in dentistry may involve only enamel, or a combination of enamel and dentin depending on the severity of decay. Similarly, prosthetic implant placement may result in material contact with cortical bone, trabecular bone, or a mixture of both depending on if the nature of the surgery is orthopedic or trauma-induced [19,20]. Based on what is known from the dental literature regarding hybridization of the dentin substrate with the adhesive material [21], it could be expected that bonding to the denser cortical bone, which has parallel orientation of lamellar units, would benefit from the creation of micromechanical retention similar to dentin, and/or from chelation of the hydroxyapatite, as is the case with phosphonated adhesives (self-etching) [22]. There are currently several FDA-approved adhesive materials used for dentin bonding in dentistry, which improve device retention in the oral cavity [23,24]. These are usually classified in four different generations (Fig. 1, supplementary information), each requiring different steps for application. In general, dental adhesive systems require etching the dental surface with 37% phosphoric acid to increase the surface area and clean the adherent substrate. In addition, materials have at least one component with hydrophilic character to allow for penetration of the water-rich dentin tubules and to form a hybrid layer composed of demineralized collagen and polymerized adhesive. This layer has the function of providing mechanical interlocking for the more viscous filling materials. Specifically in the case of self-etch materials, in addition to the micromechanical retention/mechanical interlocking, adhesion is achieved by chelation of the mineral content by acidic functionalities in the monomers of the adhesive [25]. Even though there are advantages and disadvantages for any of these products, when applied under ideal conditions, they all achieve acceptable clinical outcomes in dentistry and have been optimized for not only performance, but consistency and ease of application [26].

Fig. 1. Schematic overview of the experimental design and study workflow.

The study was conducted in two sequential phases. Phase I focused on screening dental adhesive systems combined with commercial bone cements to identify the adhesive with the most favorable bonding performance. Phase II assessed the antimicrobial performance of PMMA-based bone cements modified with quaternary ammonium methacrylate (QAM), using the selected adhesive from Phase I.

In addition to improving retention, another area of need for implantable devices is minimizing the risk of recurring infection [27,28]. Osteomyelitis is associated with trauma, orthopedic surgery and periprosthetic implant infection, with Staphylococcus aureus responsible for the vast majority of acute cases [29]. Several materials exist in the market with incorporated antibiotics, such as tobramycin-impregnated cements, which have shown up to a 50% reduction in infection rates [30]. While these are advantageous, the antibiotic effect is limited by the leaching of the therapeutic agent from the material, which is diffusion-dependent [31]. This leads to a significant reduction in antimicrobial efficacy [32], which is a concern for long-term implantable devices especially when placed in less-than ideal conditions, such as battlefield injury repair and treatment of open fractures [33]. One solution to overcome these limitations is to covalently tether the antibiotic compounds to the surface of the material, such as has been proposed for methacrylate-based restorative materials containing co-polymerizable quaternary ammonium methacrylates (QAMs) [34]. These have been attempted for dental adhesives and cements, with significant reduction of bacterial colonization, without compromising mechanical properties and retention [35] and with no cytotoxicity concerns [36]. These properties have been shown to be dependent on the structure of the QAM material [37]. For example, it was found that the greater the alkyl chain length of the QAMs in the bonding agents, the greater the inhibition of Streptococcus mutans, while also maintaining fibroblast and odontoblast viability and consistent bond strength [38]. Similarly, being methacrylate-based in the majority of commercial examples, bone cements have the potential to be co-polymerized with methacrylate-containing antibiotics, such as QAMs, or methacrylated imidazolium [39]. The primary advantage of these non-leachable materials is their sustained antimicrobial effect.

Given similar substrate compositions, there is underexplored potential for dental adhesives to be used as adjuvant materials in both bone cementation and to improve implant retention. Therefore, the objective of this study was to test the shear bond strength of bone cement materials used to bond titanium implants to bone when used in combination with (a) four different types of dental adhesives and (b) with the addition of QAM to the cement formulation. Emphasis was placed on providing a mechanical challenge under physiological conditions while also evaluating toxicity and antibacterial efficacy to ensure safe implementation within a closed biological system. The hypotheses tested were: (1) The use of dental adhesives in combination with commercial bone cements will lead to an increase in bond strength and retention without causing cytotoxicity; (2) The incorporation of QAM compounds in the cement, in conjunction with the highest performing adhesive, will decrease infection and improve retention.

2. Materials and methods

2.1. Bone Source

Fresh porcine jaws from Sus scrofa domesticus calves (6 months-old) were sourced from Carlton Farms (Carlton, OR, USA) and kept frozen until use. They were selected as a proxy for human bone, consistent with previous studies [40,41]. Flat samples were obtained from the ramus of the lower jaw, which provided a cortical surface for bonding. Sections of approximately 10 mm³ were cut, and the fragments were then stored in 0.5% Chloramine T solution for decontamination before bonding procedures. Bone pieces were embedded into epoxy (EpoxiCure 2, Buehler, Lake Bluff, IL, USA) within cylindrical PVC molds (16 mm in diameter and 20 mm in height), positioned so the cortical flat surface could be exposed for bonding. The procedure is illustrated in Fig. 3A.

Fig. 3.

(A) Top view of sample preparation steps once embedded in epoxy from left to right, respectively. (B) Jig used to standardize layer thickness using two 150 g weights. (C) Sample submerged in Millipore water within bioreactor chamber. (D) Zoomed out view of entire bioreactor machine. (E) Jig used to measure shear bond strength with separate stamp geometries. (F) Schematic representation of shear bond strength test from ISO 10477:2020 [42,99].

2.2. Cement and adhesive materials

PMMA-based commercial cements were used in this study. Bone Cement R (Zimmer Biomet, Warsaw, IN, USA) was selected as a commonly used cement in orthopedic practice, and Jet Acrylic (Lang Dental, Wheeling, IL, USA), which is used to produce dentures, was used as proxy, both independently and in conjunction with experimental monomers. The compositions of these materials, presented as a powder/liquid system, are listed in Table 1S.

For the first phase of the study, four adhesives were tested: Adper Scotchbond Multi-purpose (Gen 4, 3 bottle, total-etch adhesive, 3 M ESPE, St. Paul, MN), Adper Single Bond Plus (Gen 5, 2 bottle, total-etch adhesive, 3 M, ESPE), Clearfil SE Bond (Gen 6, two bottle, self-etch adhesive, Kuraray, New York, NY, USA), and Scotchbond Universal Plus (Gen 7, 1 bottle, self-etch adhesive, 3 M, ESPE). These are representative of each generation of dental adhesive commercially available, with all compositions of materials used shown in Table 2S.

Once the adhesive with the highest bond strength was identified, it was used in conjunction with antibiotic-modified cements (second phase of the study). The following cements were tested: non-antibiotic groups consisted of 1) Jet Acrylic, using a 1:1 wt ratio of powder to liquid; 2) Jet + HEMA − Jet Acrylic modified with 2-hydroxyethyl methacrylate (HEMA), added to assess potential effects of a longer side chain monomethacrylate on mechanical properties; antibiotic-containing groups consisted of positively charged, quaternary-ammonium based materials; 3) Jet + CTAC Cetyltrimethylammonium chloride, added as a non-polymerizable/leachable antibiotic (Fig. 2-A); 4) Jet + QAM Dimethylaminohexadecane methacrylate bromide, added as a co-polymerizable antibiotic, (QAM/DMAHDM, Fig. 2-B). Monomer incorporation was standardized by substituting 10 wt% of the liquid component with the respective monomer prior to mixing with the powder; 5) Refobacin Bone Cement R was used as a commercial antibiotic-containing control.

Fig. 2.

Chemical structures of the quaternary-ammonium-based antibiotics incorporated into PMMA cements during the second phase of the study. Cetyltrimethylammonium chloride (CTAC; left) was used as a non-polymerizable, leachable quaternary ammonium compound, whereas dimethylaminohexadecane methacrylate bromide (QAM/DMAHDM; right) was used as a co-polymerizable, covalently bound antimicrobial monomer.

2.3. Bonding procedures

Bone surfaces were standardized by removing a thin slice from the superior surface with a diamond saw (Accutom, Struers, Cleveland, OH, USA). Adhesives were applied to the “as cut” bone surface following the manufacturer’s recommended instructions to a standardized application area (28.3 mm², delimited by a rubber stamp), photopolymerized using an LED light-curing unit (Elipar, 3 M-ESPE) for 60 s with 1200 mW/cm² being delivered 2 mm away from the surface. Detailed protocols for each adhesive application can be found in the supplementary information.

Next, the two-component cement was mixed in a fume hood. The working time of the cements was approximately 150 s for a “doughy” phase to be achieved (at 25°C, with relative humidity of 40–60%, following manufacturer’s instructions), at which point the material was formed into a small ball, applied to the masked area in the center of the bone square, either directly onto the bone or atop the polymerized adhesive layer. Finally, a titanium alloy cylinder, Ti Al6 V4 (ASTM B348 GR5 UNS R56400), 6 mm in diameter and 5 mm thick, (ODEME Dental Research, Luzerna, SC, Brazil) was ultrasonically-cleaned at 37°C in deionized water for 10 min, followed by vigorous air drying, then coated with Z-prime plus (Bisco, Inc., Schaumburg, IL, USA), which was brushed on the bond surface for 15 sec, followed by solvent evaporation with a gentle air stream for 15 s. The titanium specimens was gently pressed onto the “doughy” cement ball and held in place with a 300 g weight applied during hardening (15 min), using the apparatus shown in Fig. 2B (ODEME Dental Research, Luzerna, SC, Brazil). After the initial cure, the specimens were stored in an incubator in MilliQ water for 24 h at 37° C. After this period, the samples were either immediately tested for shear bond strength (as described below) or subjected to mechanical cycling in a bioreactor (Fig. 2C).

2.4. Shear bond strength testing (SBS)

Shear bond strength tests were conducted according to ISO 10477:2020 [42]. Fig. 2E and F show the test apparatus that was mounted to a universal testing machine equipped with a 5 kN load cell (Criterion, MTS, Eden Prairie, MN, USA). Specimens were tested until failure at a rate of 1.0 mm/min and a 10 Hz data collection rate, the resultant peak force in N was recorded, and shear bond strength (MPa) was calculated by dividing the force by the bonded area. The SBS testing was conducted in the order indicated below:

1. Gen 4 – Gen 7 adhesives with Jet Acrylic were tested. Jet Acrylic without adhesive was used as a control.

2. Adhesive materials with the highest and lowest SBS were selected for further testing with the commercial Bone Cement R, where Jet Acrylic and Bone Cement R without adhesive were used as controls. Testing in this phase also included mechanical cycling in a bioreactor (described below).

3. The overall highest performing adhesive was tested in combination with modified cement materials (Jet + HEMA) and (Jet + QAM), where Jet Acrylic and Refobacin Bone Cement R were used as a controls. Testing in this phase included mechanical cycling in a bioreactor and bacterial aging (described below).

2.5. Mechanical cycling in the bioreactor

The bioreactor consisted of hermetically sealed chambers, to which pneumatic pistons were attached. Specimens were placed on the base of the chamber, stabilized in place by steel rods that nest notches produced on the specimen sides (Fig. 2C). The pistons were positioned on the Ti puck surface, the chambers were filled with water, and the connected in-house air pressure was turned on (Fig. 2D). Mechanical cycling consisted of 5.0 bar (200 N) at 1.5 Hz for 120 min (higher pressure, simulating periods of active service) followed by 2.5 bar (150 N) at 1.0 Hz for 30 min (lower pressure, simulating periods of rest), where completion of both phases represents 1 cycle. Before testing, a calibration was performed using the three pistons of the bioreactor to evaluate the load (N) at varying pressure inputs ranging from 1 to 5 (bar) across all pistons independently as seen in Fig. S2. A total of 67 cycles were run on all samples (n = 6 per group), which took 7 days. After testing in the bioreactor, if the specimens were still intact, the samples were removed, patted dry, and immediately tested for shear bond strength. For samples with varying bone cement formulations (n = 6 per group), the bioreactor ran for 48 cycles (5 days), with all other settings remaining the same. Once the test concluded, samples were carefully removed from the chambers, and the surrounding areas were gently cleansed and dried with cotton swabs. The specimens were then immersed in isopropanol for 20 min. Following this, they were rinsed with sterile deionized water for another 20 min and air-dried before proceeding with biological aging (described in proceeding sections).

2.6. Failure analysis

After shear bond strength testing, interfacial surface failure analysis was conducted and categorized (Bone Interface, Mixed, or Ti Interface). Optical images were obtained using a stereomicroscope (10x), M50, Leica, Deerfield, IL, USA). Additionally, dried bone samples and epoxy replicas of fracture surfaces were sputtered with Au-Pd at a thickness of 10 nm using a Denton Vacuum Desk II at 50 milliTorr and SEM images were obtained using a Helios Nanolab G3 DualBeam FIB-SEM (FEI, ThermoFisher Scientific, Hillsboro, OR, USA). Failures were classified as adhesive at the bone interface, adhesive at the titanium interface, or mixed. The vast majority of the failures happened due to delamination at either interface (adhesive failure), so the mixed failures include cohesive failure in the cement or bone. There was no instance where exclusively cohesive failure of the cement or bone were observed.

2.7. Fourier transform infrared attenuated total reflectance spectroscopy (FTIR-ATR)

After completing the restoration procedures as previously outlined, 0.5 mm thick bone slices were obtained (Accutom-5, Struers, Cleveland, OH, USA) for chemical mapping (n = 3). The adhesive layer was analyzed using FTIR-ATR with a germanium tip (RaptIR FTIR Microscope IS50, Thermo Scientific, Waltham, MA, USA). The specimens were laid flat on a reflective glass microscope slide, and spectra were collected with 64 scans, at a resolution of 16 cm⁻¹, across the spectral region of 4000–400 cm⁻¹. The distance was standardized to 100 μm, resulting in 29 spectra obtained per specimen. Additionally, FTIR-ATR analysis was performed on the monomers before polymerization and on the bone using a diamond ATR accessory, using the same settings.

2.8. Confocal microscopy analysis

The same specimens that were used for FTIR-ATR analysis were also used for confocal analysis (n = 2). The sections were mounted on glass bottom dishes (0.17 mm thick, World Precision Instruments, Sarasota, FL, USA) and imaged without additional fluorescent markers, utilizing the inherent fluorescence of the adhesive. Fluorescence analysis was performed using a Zeiss LSM 880 confocal microscope (Carl Zeiss, Oberkochen, Germany) with an excitation wavelength of 488 nm and emission wavelength of 522 nm. Single-plane images were acquired and processed using ZEN 2010 software (Carl Zeiss, Oberkochen, Germany). The thickness of the adhesive layer was measured using ImageJ software (National Institutes of Health, Bethesda, MD, USA), based on the image scale.

2.9. Degree of conversion

Disc-shaped specimens (0.8 mm in thickness × 10 mm in diameter) of each experimental PMMA-based cement were fabricated using a silicone mold placed between two glass slides (n = 6). The degree of conversion (DC) was assessed immediately and after 48 h using near-infrared (near-IR) spectroscopy (Nicolet 6700; Thermo Fisher Scientific, Waltham, MA, USA). The degree of conversion was calculated based on the ratio of monomer to polymer peak areas of the carbon double bond peak (6165 cm⁻¹), using the following equation [43]:

$$DC(\mathrm{\%})=\left(1-\frac{6165{cm}^{-1}\mathit{cured}}{6165{cm}^{-1}\mathit{uncured}}\right)\times 100$$

2.10. Hardness

Following the degree of conversion analysis, Knoop microhardness (KHN) was evaluated on the same specimens (n = 6) using a microhardness tester (HMV-G Series, Shimadzu Corporation, Kyoto, Japan). A load of 980.7 mN was applied for 10 s, with 1 mm between each indentation. For each specimen, five measurements were taken, and the average value was calculated for each condition.

2.11. Water sorption and solubility

Disc specimens from each experimental group of PMMA-based cements were evaluated for water sorption (WS) and solubility (SL) following ISO 4049 standards (n = 6). Briefly, the cement samples were analyzed for their degree of conversion from monomer to polymer after 48 h. The initial dry mass (M1) was measured prior to immersion in 5 mL of Millipore water for 1 week. After the immersion period, the specimens were gently blotted dry with absorbent paper and weighed to record their mass after storage (M2). Subsequently, the samples were placed in a desiccator containing silica gel and connected to the house vacuum. Sample weights were measured daily until the final mass did not change to the nearest 0.0001 g (M3). WS and SL were calculated in μg/mm³ using Eqs. (1) and (2), with V denoting the disc volume in mm³. Disc leachates were analyzed by ¹H NMR spectroscopy (Bruker AMX 400 MHz).

$$WS=\frac{m2-m3}{V}$$ (1)

$$SL=\frac{m1-m3}{V}$$ (2)

2.12. Biofilm properties and biological aging

2.12.1. Sample preparation and biofilm inoculation

Staphylococcus aureus (S. aureus, ATCC^® 12600^™) was obtained in lyophilized form and rehydrated following the manufacturer's instructions (Supplementary Information). Prior to bacterial exposure, the cement disks were immersed in isopropyl alcohol for 20 min, thoroughly rinsed with sterile Milli-Q water, and air-dried under sterile conditions. The prepared specimens were then transferred to sterile 24-well plates, with each well containing one sample completely submerged in 1 mL of BHI broth inoculated with an overnight culture of S. aureus. Plates were incubated aerobically for 24 h at 37°C in a 5% CO₂ atmosphere on an orbital shaker to promote uniform biofilm formation.

2.12.2. Post-bioreactor biological aging

S. aureus was cultured and incubated to induce biological aging, and bioreactor specimens were submerged in the bacterial medium under aerobic conditions with continuous shaking in an incubator at 37°C in a 5% CO₂ atmosphere for 72 h. After the incubation period, each sample was immersed in isopropanol for 20 min to deactivate the bacteria, followed by rinsing with sterile deionized water for an additional 20 min. The samples were then air-dried under sterile conditions. Subsequently, shear bond strength testing and failure mode analysis were performed.

2.12.3. Antimicrobial assays

To assess planktonic bacterial growth, cement discs were carefully removed from the wells (n = 6), and the optical density (OD₆₀₀) of the remaining supernatant was immediately measured using a microplate reader (SpectraMax iD3, Molecular Devices) in absorbance mode. Prior to reading, the plate was subjected to orbital shaking to ensure homogeneous mixing. Experimental controls included BHI media alone and BHI media inoculated with S. aureus in the absence of cement discs. Following planktonic assessment, the retrieved discs were gently washed three times with sterile Milli-Q water to remove non-adherent bacteria. The samples were then air-dried and stained with 0.1% (w/v) crystal violet solution for 15 min at room temperature to visualize surface-bound biofilm. After staining, the discs were rinsed, allowed to dry completely, and imaged using a Dino-Lite digital stereomicroscope to qualitatively evaluate biofilm morphology. To quantify biofilm biomass, the crystal violet-stained biofilm was solubilized by adding 30% (v/v) acetic acid to each well, followed by shaking for 20 min. The discs were then removed, and the optical density of the solubilized dye was measured at 562 nm using the same microplate reader.

After measurement of the optical density of planktonic bacterial cultures, the supernatant from each well (n = 6) was serially diluted in Todd-Hewitt (TH) broth. An initial 1:9 dilution was prepared by adding 100 μL of the supernatant to 900 μL of sterile TH medium, followed by a total of six serial dilutions for each sample. For colony-forming unit (CFU) enumeration, blood agar plates (n = 3) were used and sectioned into six equal sectors. Aliquots of 10 μL from each dilution were pipetted onto the surface of the corresponding agar sector. The droplets were allowed to air-dry with the plate lids open under aseptic conditions, after which the plates were closed and incubated at 37°C in a 5% CO₂ atmosphere for 24 h to allow colony development. Following incubation, bacterial colonies were visually inspected for isolation, and CFU counts were performed using the sixth dilution, which consistently yielded well-isolated and countable colonies.

2.13. MTT assay

For the cytotoxicity assay, normal human dermal fibroblasts (NHDF; #CC-2511, Lonza, Walkersville, MD, USA) and normal human osteoblasts (NHOst; #CC-2538, Lonza) were used. The NHDF cells were cultured in a fibroblast growth medium (FGM) consisting of fibroblast basal medium (CC-3131, Lonza) supplemented with FGM2 SingleQuots (#CC-4126, Lonza) comprising of 2% fetal bovine serum (FBS), 1% recombinant human insulin, 1% recombinant human fibroblast growth factor and 1% gentamicin sulfate/amphotericin-B. The NHOst cells were cultured in Osteoblast growth medium (OGM) consisting of Osteoblast basal medium (#CC-3208; Lonza) supplemented with OGM^™ SingleQuots (#CC-4193, Lonza) comprising of 10% FBS, 1% Ascorbic acid, 1% gentamicin sulfate / amphotericin-B. Cells were replenished with media every 48 h and maintained in an incubator at 37°C under humidified atmosphere, provided with 5% CO₂ until they reach 80% confluence. On the day of experiment, the cells were washed with 4 mL of Dulbecco’s phosphate buffered saline solution twice (Thermo fisher scientific, Waltham, MA, USA), followed by detaching the cells from the culture flask using 5 mL of TrypLE^™ Express Enzyme (Gibco, Thermo fisher scientific). Cells were counted using a Countess II Automated Cell Counter (Thermo fisher scientific) then 10,000 cells per well were seeded in a 96-well plate and incubated for 24 h to reach 90% confluency.

To confirm solubility in the cell culture media, roughly 0.5 g of the commercial adhesive or adhesive/primer combination, depending on the generation, was weighed and allowed 24 h for desiccation to evaporate > 15% of the solvent. Using the remaining amount of monomer, DMSO was added to reach a concentration 0f 0.3 g/ml. Due to formation of precipitates in the cell culture media, this concentration was further diluted by adding 10 μL of the stock solution (0.3 g/ml) into 1000 μL of DMSO, reaching a new concentration of 0.3 mg/ml. 2 μL of this stock solution was then dissolved into 1 ml of the cell culture media without precipitate formation, creating a concentration of 6 μg/ml. The dental adhesive samples (3 mg/mL stock solution in DMSO) were serially diluted using 10-fold dilutions to obtain four concentrations: 6, 0.6, 0.06, and 0.006 μg/mL, using cell culture medium (FGM for HNDF and OGM for NHOst). To decouple the effect of DMSO from the adhesives, DMSO volumes corresponding to the sample concentrations were prepared by serial dilution using cell culture media, resulting in four concentrations: 1.8, 0.18, 0.018, and 0.0018 μL/mL. Then, 100 μL of the sample containing cell culture media was added to each well seeded with the cells and incubated for 24 h (n = 6).

After 24 h of incubation the cells were treated with 40 μL of 0.5 mg/mL of diphenyltetrazolium bromide (MTT; Sigma-Aldrich, St. Louis, MO, USA) prepared in PBS (wt/v) and incubated for 4 h in dark conditions. Following incubation, all the contents of the well plate (spent media + MTT reagent) were removed carefully without disturbing the cells. Then, 100 μL of DMSO was added to each well to lyse the cells and dissolve the produced formazan crystals, followed by incubation for 1 h in the dark. Absorbance was measured at 560 nm using a spectrophotometer (GloMax Discover, Promega Instrument, Version 4.0.0, Madison, WI, USA) with background absorbance (media only) corrected for all samples, and the percentage of cell viability was calculated using the following equation (3):

$$\mathrm{\%}Cellmetabolicactivity=\left[\frac{\mathit{ODsamples}}{\mathit{ODcontrol}}\right]\times 100$$ (3)

where OD samples is the Optical density of the samples at 560 nm and OD control is the Optical density of the control wells (only cells no samples) at 560 nm.

2.14. Statistical analysis

Data on degree of conversion, hardness, water sorption, solubility, optical density, biomass, and shear bond strength were collected and subjected to descriptive analysis. The datasets were further evaluated for normality and homoscedasticity using the Levene test (Jamovi Program, Version 2.2.5). Hardness, water sorption, solubility, optical density, biomass data were analyzed using one-way ANOVA. Shear bond strength data were examined via two-way ANOVA, while degree of conversion data was assessed using two-way repeated measures analysis of variance (ANOVA; Jamovi Program; Version 2.2.5). Post hoc comparisons were performed using the Tukey test, with a global significance threshold of 95% applied to all analyses.

3. Results

3.1. First phase: Screening adhesive materials

The values for shear bond strength for all adhesive materials tested in combination with Jet Acrylic as the cement material after 24-hour storage in 37°C Millipore water are shown in Fig. 4A. The Gen 6 adhesive showed the highest value at 21.6 ± 8.7 MPa, statistically different from the Gen 5 adhesive and the No Adhesive group (p < 0.001). The failure analysis (Fig. 4B) showed 100% adhesive failure at the bone interface for both the Jet acrylic and the commercial bone cement, while all the other adhesive groups also presented failure at the titanium interface, or cohesive failure of the cement itself.

Fig. 4.

(A) Shear bond strength (MPa) as a function of varying adhesive systems using only Jet Acrylic as the cement material. Different uppercase letters indicate statistically significant differences between the tested adhesive systems (p ≤ 0.05). (B) Failure analysis (% out of 100) of the same tested samples (n = 6) to indicate the fracture interface using the different adhesive systems combined with Jet Acrylic.

Results for the shear bond strength using the Biomet Bone Cement and Jet Acrylic in combination with the Gen 5 and Gen 6 adhesives tested after 24 h in 37°C Millipore water are shown in Fig. 5. The 2-way ANOVA showed that only the adhesive was significant (adhesive = p < 0.0001, cement = p = 0.4212, interaction = p = 0.1511). Gen 6 had the highest values (21.6 ± 8.7 and 19.4 ± 8.7 MPa for Jet Acrylic and Biomet Bone Cement, respectively), followed by the Gen 5 adhesive (p < 0.0001). Both the Jet Acrylic and Biomet Bone Cement without adhesives showed the lowest bond strength values (0.8 ± 0.7 and, 0.9 ± 0.4 MPa, respectively) and again failed exclusively on the bone interface.

Fig. 5.

(A) Shear bond strength (MPa) as a function of varying adhesive systems using two different cement materials. Different uppercase letters indicate statistically significant differences between the tested adhesive systems (p ≤ 0.05). (B) Failure analysis (% out of 100) of the same tested samples (n = 6) to indicate the fracture surface using the different cement materials within the adhesive systems.

The effect of mechanical cycling on shear bond strength of the samples is shown in Fig. 6 for the Gen 5 and Gen 6 adhesives combined with the Biomet Bone Cement. The 2-way ANOVA showed that the adhesive and the cycling were both significant (adhesive: p < 0.001, cycling: p = 0.0403, interaction: p = 0.3181). The No Adhesive group showed the lowest values of shear bond strength with or without mechanical cycling (p = 0.0403). The shear bond strength dropped for both adhesive systems after mechanical cycling, but not statistically for Gen 6.

Fig. 6.

(A) Shear bond strength (MPa) as a function of varying adhesive systems using only the Biomet Bone Cement R as the cement material with and without mechanical cycling in the bioreactor. The non-cycled samples were stored in 37°C Millipore water for 24 h, where the cycled samples were stored for 1 week under the same conditions. Different uppercase letters indicate statistically significant differences between the tested adhesive systems across both storage conditions (p ≤ 0.05). (B) Failure analysis (% out of 100) of the same tested samples (n = 6) to indicate the fracture interface using the different adhesive systems under the two cycling conditions.

3.1.1. SEM imaging

The failed bond interface of the adhesive groups in combination with the Jet Acrylic on bone is shown in Fig. 7. The top row images were taken at 20x magnification, whereas the bottom row images were taken at 200x magnification. The samples where the adhesives were used have a distinctively greater roughness compared to no adhesive.

Fig. 7.

(A: Top and Bottom) Surface of bone after failure using bone cement only as adhesive system. (B: Top and Bottom) Surface of bone after failure using Gen 5 adhesive prior to bone cement application. (C: Top and Bottom) Surface of bone after failure using Gen 6 adhesive prior to bone cement application.

3.1.2. Confocal imaging

Each image in Fig. 8 depicts the adhesive layer (represented in pink) situated between the cement material (top) and the cortical bone (bottom). There is little evidence of penetration of the adhesive into the bone. The thickness of the adhesive layer varies between 19–36 μm for Gen 6 and 39–78 μm for Gen 5.

Fig. 8.

Confocal images showing Gen 5 and Gen 6 adhesive layers of resin cement at 20x magnification. RC: resin cement; A: adhesive layer; B: cortical bone.

Figs. 9 and 10 present the analysis of Gen 5 and Gen 6 adhesives using FTIR-ATR spectroscopy coupled with microscopy, with a specific focus on the bone-restoration interface. Fig. 9/10A display the FTIR spectra of the adhesive monomer (before polymerization), the bone, and the adhesive layer at the interface. Key peaks such as Amide I (1650 cm⁻¹), Amide II (1550 cm⁻¹), Amide III (1250 cm⁻¹), and phosphate (${\text{PO}}_4^{3-}$ at 1020 cm⁻¹ and ${\text{PO}}_4^{3-}$ between 1030 and 1100 cm⁻¹) are visible in both the bone and the adhesive interface spectra, but do not appear in the adhesive layer. Fig. 9/10B present microscopy images captured using an FTIR-ATR coupled microscope, providing real-time visualization of the bone-restoration interface without causing sample destruction. The image distinctly shows the bone, resin cement, and adhesive layer, along with the collected spectra. Fig. 9/10C show the sequential FTIR-ATR spectra collected from the bone-restoration interface, illustrating the transition from bone to adhesive to resin cement. The peaks demonstrated an increase in the intensity of Amide I (1650 cm⁻¹), Amide II (1550 cm⁻¹), and phosphate peaks (1020 cm⁻¹ and 1030–1100 cm⁻¹) as the analysis transitions from the resin cement to the bone. Fig. 9/10D display the sequential spectra showing peak intensities, with areas in red representing the most intense peaks. The Gen 6 adhesive exhibits a strong red signal in the phosphate region, indicating a higher intensity of interaction with phosphate groups compared to the Gen 5 adhesive.

Fig. 9. FTIR-ATR Analysis of Gen 5 Adhesive Sample.

(A) FTIR-ATR spectra of the adhesive monomer, adhesive layer, and bone; (B) Microscopic image of the bone-restoration interface at 15x magnification, each of the black squares represents one spectrum; (C) Sequential FTIR spectra from the bone-restoration interface, showing the transition from bone to adhesive to resin cement; (D) FTIR-ATR spectra derived from (C), plotted by peak intensity.

Fig. 10. FTIR-ATR Analysis of Gen 6 Adhesive Sample.

(A) FTIR-ATR spectra of the adhesive monomer, adhesive layer, and bone; (B) Microscopic image of the bone-restoration interface at 15x magnification, each of the black squares represents one spectrum; (C) Sequential FTIR spectra from the bone-restoration interface, showing the transition from bone to adhesive to resin cement; (D) FTIR-ATR spectra derived from (C), plotted by peak intensity.

3.1.3. Cell viability

Cell metabolic activity for normal human dermal fibroblasts and normal human osteoblasts is shown in Fig. 11. Representative images of each cell type are shown in Figs. S3–4. Up to 0.6 μg/mL, there was no statistical decrease of cell metabolic activity for any of the groups. Gen 4 and Gen 6 materials showed a statistically significant decrease in cell viability at 6 μg/mL.

Fig. 11.

The MTT assay of dental adhesive samples was performed against (A) normal human dermal fibroblast cells or (B) normal human osteoblast cells after 24 h of incubation at 37°C. Different uppercase letters denote statistically significant differences within the group, while different lowercase letters denote statistically significant differences between the same concentrations across the different tested materials.

3.2. Second phase: Assessing the antimicrobial effect of PMMA modified with QAM

The shear bond strength values using varying cement formulations before and after aging are summarized in Fig. 12. Prior to aging, no statistically significant differences were observed between the control group (8.65 ± 2.56 MPa) and the Jet + HEMA group (11.83 ± 7.44 MPa; p = 0.469). However, only the Jet + HEMA group exhibited significantly higher bond strength compared to the Jet + QAM group (2.95 ± 1.02 MPa; p = 0.011). A similar trend was observed following aging: the control (9.02 ± 3.15 MPa) and Jet + HEMA (9.02 ± 5.17 MPa) groups remained statistically comparable (p = 1.000). However, both groups showed significantly greater bond strength than the Jet + QAM group (3.02 ± 1.81 MPa; p = 0.031), by about 66%. Additionally, aging did not exert a statistically significant influence on the bond strength within the control (p = 0.830), Jet + HEMA (p = 0.464), and Jet + QAM (p = 0.939) groups. As a means of benchmarking with relevant commercial materials, Biomet (no antibiotic) was compared with Refobacin Bone Cement R (antibiotic-containing) after aging in the bioreactor. The presence of antibiotic led to a decrease of around 42% in bond strength (14.7 ± 5.6 and 8.5 ± 3.9 MPa for the antibiotic free and the antibiotic containing material, respectively – Fig. S5), which was not statistically significant (p = 0.0519).

Fig. 12.

(A) Shear Bond Strength (MPa) as a function of varying cement formulations using Jet Acrylic alone and in combination with 10 wt% HEMA or QAM monomers with an adhesive pre-treatment using the Gen 6 adhesive. Groups were tested after storage of seven days in Millipore water incubated at 37°C (no bioreactor) or of five days in Millipore water under mechanical cycling in the bioreactor, followed by three days of exposure to S. aureus incubated at 37°C and 5% CO₂ atmosphere (yes bioreactor). Different uppercase letters indicate statistically significant differences between the tested cement systems (p ≤ 0.05). (B) Failure analysis (% out of 100) of the same tested samples (n = 6) to indicate the fracture interface using the different formulations of cement.

3.2.1. Properties of cements and biofilm morphology

Before the discs of varying cement formulations were evaluated for optical density and biomass accumulation, multiple tests were conducted to ensure the desired antimicrobial effect did not come at the cost of reduced physical properties of the cement. The degree of conversion values of the cements after two time periods and Knoop hardness (KHN) values are presented in Table 1. After 30 min, Jet Acrylic (74.1 ± 6.6%) exhibited lower values compared to Jet + HEMA (92.7 ± 11.4%; p = 0.020). No significant differences were observed when comparing Jet + CTAC, Jet + QAM, Biomet and Refobacin. After 48 h, all groups containing Jet reached full conversion (p > 0.05), but the Biomet-based cements only reached approximately 93%. When analyzing the effect of evaluation time, all groups showed significantly higher conversion after 48 h (p < 0.05). In terms of KHN, the Jet + CTAC group (13.5 ± 2.5) exhibited lower values compared to the Jet Acrylic control (17.6 ± 0.5; p < 0.001), Jet + HEMA group (17.8 ± 0.4; p < 0.001), and Jet + QAM group (16.6 ± 1.4; p = 0.008). The Biomet group had similar hardness all groups except Jet + CTAC (p = 0.029), while Refobacin was similar in hardness to Biomet, Jet + QAM and Jet + CTAC.

Table 1.

Mean and standard deviation of degree of conversion (%) and Knoop hardness (KHN).

	Degree of conversion		Knoop hardness (KHN)
	30 Minutes	48 Hours
Jet Acrylic	74.1 ± 6.6Ba	100.0 ± 0.0Ab	17.6 ± 0.5A
Jet + HEMA	92.7 ± 11.4Aa	100.0 ± 0.0Ab	17.8 ± 0.4A
Jet + CTAC	86.9 ± 14.6ABa	100.0 ± 0.0Ab	13.5 ± 2.5C
Jet + QAM	79.7 ± 1.0ABa	100.0 ± 0.0Ab	16.6 ± 1.4AB
Biomet	84.0 ± 2.7ABa	92.9 ± 0.4Bb	15.9 ± 0.7AB
Refobacin	83.5 ± 3.5ABa	93.2 ± 1.3Bb	14.5 ± 0.7BC

Conversion was reported after 30 min and 48 h. Uppercase refers to differences within columns and lowercase refers to differences within rows with respect to degree of conversion. Different letters indicate statistically significant differences (p < 0.05).

Water sorption (WS) and water solubility (SL) values for the evaluated PMMA-based cements are presented in Fig. 13-A. WS and SL were expressed in μg/mm³. Statistically significant differences among materials were detected for both parameters, as indicated by distinct upper-case letters for WS and lower-case letters for SL (p < 0.05). The WS and SL increased by 71 and 250%, respectively, for the Jet + QAM compared with the Jet material, and by 29 and −15%, respectively, for Refobacin compared with the Biomet material. This means Refobacin had lower solubility than Biomet despite having greater water sorption.

Fig. 13.

(A) Water sorption (WS) and water solubility (SL) expressed in μg/mm³ for Biomet, Refobacin, Jet, and Jet + QAM cements (n = 6). Because WS and SL are not directly comparable, different upper-case letters indicate statistically significant differences among groups for WS, whereas different lower-case letters indicate statistically significant differences among groups for SL. (B) Quantification of planktonic S. aureus by optical density (OD) measured at 600 nm (n = 6). (C) Biofilm biomass quantified after staining with 0.1% crystal violet and solubilization with 30% acetic acid; representative images of the stained cement disks are shown above each corresponding bar (n = 6). (D) Viable S. aureus counts expressed as colony-forming units per milliliter (CFU/mL), with representative images of 5% blood agar plates shown above the respective bars (n = 6 wells, in triplicate agar plates). For panels (B-D), different upper-case letters denote statistically significant differences within the same cement group before and after the water sorption/solubility assay (pre or post WS/SL), while different lower-case letters denote statistically significant differences among cement groups within the same evaluation time point (pre or post WS/SL). Data are presented as mean ± standard deviation. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

The antibacterial response of the materials was assessed before and after the water sorption/solubility (WS/SL) assay. Planktonic - Staphylococcus aureus growth, quantified by optical density at 600 nm (OD₆₀₀), is shown in Fig. 13-B. For the pre-WS/SL condition, data were normally distributed (Shapiro-Wilk, p = 0.291), although homogeneity of variances was not met (Levene’s test, p < 0.001). One-way ANOVA revealed a significant effect of material on planktonic bacterial growth (p < 0.001), with post-hoc Tukey analysis identifying statistically significant differences among groups (p < 0.05). After WS/SL, normality (p = 0.593) and homogeneity of variances (p = 0.037) were assessed, and one-way ANOVA again showed a significant effect of material (p < 0.001), with multiple pairwise differences detected by Tukey’s test.

Biofilm biomass quantification, assessed by crystal violet staining and shown in Fig. 13-C, followed similar trends. In the pre-WS/SL condition, the data met normality assumptions (Shapiro-Wilk, p = 0.199), while variances were heterogeneous (Levene’s test, p = 0.002). One-way ANOVA demonstrated a significant difference among cement groups (p < 0.001), and post-hoc analysis identified statistically significant pairwise comparisons (p < 0.05). After WS/SL, biomass data remained normally distributed (p = 0.551) with homogeneous variances (p = 0.598). One-way ANOVA indicated a significant effect of material (p = 0.005), with selected group differences detected by Tukey’s post-hoc test.

Viable bacterial counts, expressed as colony-forming units per milliliter (CFU/mL), are presented in Fig. 13-D. For all microbiological assays (OD₆₀₀, biomass, and CFU/mL), different upper-case letters indicate statistically significant differences within the same material before and after WS/SL, while different lower-case letters indicate statistically significant differences among materials within the same evaluation time point (p < 0.05).

4. Discussion

Arthroplasty procedures are ubiquitous, spanning from correction of birth defects to replacement of aging or injured joints, with total hip (THA) and knee (TKA) replacement being by far the most common procedures, with the most reliable results. Even with high success rates, the burden of revision surgery, with associated health complications, is still a major concern. The top three indications supporting revision include aseptic loosening, periprosthetic joint infection, and instability, which is expected to grow in the coming years due to increased primary procedures and population growth [44]. These can often be much more expensive in addition to having lower overall success. Therefore, this study sought to address the loss of retention and recurring infections by leveraging existing technologies that are currently applied to other mineralized tissues in the body. We tested an array of commercial, FDA-approved dental adhesives (detailed description in the supplementary information) as a preliminary step to bone cement application to increase retention and improve the strength of bone-implanted biomedical devices. We also tested experimental formulations of the bone cement itself that incorporated known antimicrobial monomers to reduce bacterial colonization at the bone-implant interface.

We first set out to identify the adhesive system that would provide the highest increase in bond strength when used in conjunction with one commercial PMMA-based cement. Each type of adhesive was applied according to manufacturer’s instructions directly prior to application of the freshly mixed bone cement. All adhesives improved the bond strength by at least 20-fold, and therefore the first hypothesis was accepted. For all adhesive types, the lower viscosity product applied ahead of the highly viscous PMMA-based bone cement led to a better seal of the substrate for several reasons. First, the acid etching, either by pre-treatment with mild phosphoric acid or even milder acidic moieties in the monomers leads to additional micromechanical retention and chelation (for Gen 6 and Gen 7 adhesives) of the mineral content in bone. This means that, as expected based on the similar composition, bone tissue behaves similarly to dentin in terms of bonding. The self-etching, 2-bottle Gen 6 adhesive yielded the highest shear bond strength for all the conditions tested. The value was statistically similar to all the other adhesive groups except the total-etch, one bottle Gen 5 adhesive, which yielded the lowest bond strength. This was somewhat expected based on the behavior described in the literature on dentin applications. The fact that the primer and bond component are combined in a single bottle makes the total-etch, Gen 5 adhesive more hydrophilic, creating susceptibility to water degradation and incompatibility with self-cure materials due to the acidic nature of the monomers. [45] Conversely, the separate bond step in the Gen 6 produces a hydrophobic layer for the cement to bond to. In terms of the failure mode, the groups where no adhesive was applied failed within the bone interface, which means the cement had virtually no adhesion to the bone. In general, the self-etch materials, with the milder, acidic monomers in the primer composition (Gen 6 and Gen 7), had a majority of mixed/titanium interface failures, while for the total etch adhesives (Gen 4 and Gen 5), the failure mode was more composition dependent. The Gen 4 system had primarily bone debonding, which is similar to the no adhesive group, albeit with a much greater bond strength value. One way to explain this observation is based on the viscosity of the different steps: the basic composition of the primer step for this product contains 2-hydroxyethyl methacrylate (HEMA) dissolved in some organic solvent, with very low viscosity. While technically this allows for better penetration through the micromechanical retentions created with acid etching, for the bone substrate it may have led to uneven coating. It is important to note that the bone was etched for 15 s only, following manufacturer’s recommendations, but further preliminary analysis of the bone substrate with increased etching times demonstrates greater roughness for longer acid etching times, as measured with a roughness tester as well as observed with SEM imaging (data not shown). In future studies, varying specific surface treatments could help further optimize the bonding, such as variable etching duration and identifying mechanisms that affect the hybrid bonding layer in different ways depending on the substrate [46,47].

Interestingly, further evidence that supports the viscosity hypothesis discussed above is the fact that the Gen 5 adhesive, which has a combination of 2-hydroxyethyl methacrylate (HEMA) and higher viscosity monomers, presented significantly more failures at the titanium surface. The lower viscosity HEMA likely resulted in more efficient penetration, enhancing the micromechanical retentions compared to the Gen 4 group. The Gen 6 and 7 materials use milder acid, as mentioned, so the micromechanical retention is expected to be less, which explains the greater incidence of mixed failures. However, the self-etch adhesives (Gen 6 and 7) also chelate the substrate, as mentioned previously, helping explain not only the more diverse failure modes, but also the overall higher bond strength (though only statistically different for Gen 5 and Gen 6). In summary, although dentin and cortical bone differ in nanostructure, they are both collagen-rich and follow similar denaturation mechanisms when exposed to either thermal or chemical treatment, [48] therefore would be affected similarly during the etching of the substrate. Even though the differences in performance of the different generation adhesives depends both on the overall type (total- or self-etch; separate or combined bond step) and on the brand-specific composition, the fact that the different types of adhesives are generally ranking similarly to dentin in terms of bond performance is further evidence that the bone indeed is behaving similarly to dentin [14,49,50].

Given the results observed thus far, the best and the worst performing groups were then used to validate the Jet acrylic (denture material) results against the commercially available surgical bone cement (Biomet). The lack of statistical difference of bond strength results between the Jet acrylic and the Biomet bone cement when used alone, or with the Gen 5 and 6 adhesives confirms that the Jet was a good proxy for the clinically relevant bone cement. This was expected based on their essentially identical compositions. For the purposes of this study, Jet acrylic was used as a more economical alternative to screen all the different adhesives in terms of shear bond strength. Regardless of the cement, the groups without any adhesive showed a bond strength 20 times lower and experienced 100% failures at the bone interface. In contrast, the use of either Gen 5 or Gen 6 adhesives resulted in significantly higher bond strength and introduced different failure modes, with less than half of the failures occurring solely at the bone interface. Additionally, it is noteworthy that on specimens without adhesives, the insufficient bond strength caused the cement to fail to adhere to the bone. Confocal imaging revealed variations in adhesive layer thickness, with the Gen 6 adhesive showing a thinner layer compared to Gen 5. In this case, this difference is explained by the presence of inorganic fillers in the Gen 5 adhesive (Adper Single Bond), which slightly increases the viscosity of the adhesive layer. It is important to point out that it was not the objective of this study to systematically vary the composition of adhesives, but to compare different materials that are commercially available, to ultimately facilitate translation to clinical applications.

The effects of the adhesive application on bone were further assessed using FTIR-ATR analyses and microscopy imaging. Initially, FTIR spectra were acquired for the adhesive monomer, the hybrid layer, and the bone tissue, allowing for the identification of unique absorbance peaks for each component (Fig. 9/10A). An image at 15x magnification was then captured using an FTIR microscope, which not only provided a detailed view of the adhesive-bone interface but also enabled the precise selection of points for spectral analysis (Fig. 9/10B). These data revealed the spectral changes that occurred after the adhesive polymerized and demonstrated the chemical alterations during the bonding process between the adhesive and the bone. The emergence of peaks in the adhesive layer's spectrum, compared to the adhesive monomer, corresponding to phosphate (1020 cm⁻1 and 1030–1100 cm⁻1) and amide groups (around 1650 cm⁻1, 1550 cm⁻1, and 1250 cm⁻1), indicated these interactions. This was supported by the sequential spectra (Fig. 9/10C), which showed Amide peaks corresponding to collagen and phosphate peaks reflecting the bone's mineral composition, primarily hydroxyapatite. The presence of these peaks in the adhesive layer, but not in the resin, confirms that the adhesive is not merely present but actively interacting with the bone’s mineral content. Both the pre-treatment by H₃PO₄ with the Gen 5 and the acidic monomers in the primer of the Gen 6 led to additional micro-mechanical retention and/or chelation of the mineral content in bone, consistent with prior observations in dentin [51]. Additionally, as illustrated in Fig. 10D, the Gen 6 adhesive showed a more pronounced phosphate peak, indicating a stronger chemical interaction than the Gen 5 adhesive. This is expected, given that the Gen 6 adhesive contains 10-MDP, a phosphonated compound known for its strong affinity for hydroxyapatite [52,53] which is crucial for effective bonding [52,53]. Hydroxyapatite, a calcium phosphate mineral integral to bone structure, plays a significant role in bone rigidity and strength and is present in both cortical and trabecular bone [54]. 10-MDP's ability to form a chemical bond with hydroxyapatite is attributed to its phosphate ester group, which can dissociate to form P-OH groups that react with calcium ions in hydroxyapatite. This interaction leads to the formation of a stable MDP-Ca salt, which is key to achieving durable bonding in both enamel and dentin, and as demonstrated here, potentially also bone, where hydroxyapatite is a major component of the mineralized matrix [52,53]. This finding is further supported by the shear bond strength results, since the Gen 6 adhesive showed statistically significant improvements compared to Gen 5. This suggests that the thickness is not a determinant of bond strength, but rather the chemical interaction between the adhesive and the bone.

Importantly, the clinically relevant bone cement material with and without adhesive was also tested under simulated physiological load using a bioreactor [55,56]. Physical activity and various forces exerted on joints contributes to the design and limits of arthroplasty materials [57], therefore evaluation when subject to mechanical cycling is important to evaluate the potential clinical feasibility of using adhesive systems for bonding to the bone. Among the many methods available to evaluate the performance and mechanical properties of joint replacement materials [58–62], this study assessed shear bond strength after simulated use involving compressive loading under physiologic conditions [63,64]. The loads selected here ranged from 155 to 176 N, applied to a surface contact area of 7.1 mm², resulting in a stress range of 22–25 MPa, which matches the parameters commonly used in the arthroplasty literature. [65] The cycles used here also intended to simulate periods of activity alternated with periods of rest. Bond strength values post-cycling decreased for all materials compared to the non-cycled specimens as expected [23,66], though were not statistically significant. However, the 10–20-fold increase in bond strength obtained with the use of adhesives was maintained even after cycling. In terms of failure mode, the trend continued for exclusively bone interface failure without the presence of an adhesive. Interestingly, the Gen 6 group had 100% of the failures at the Ti interface, which was different than tested without cycling, which showed mixed and bone interface failures for the same adhesive. This indicates that the self-etch adhesive is better at preserving the interface integrity which also agrees with dentin literature [67,68] and is demonstrated by our interface analyses and FTIR results. One additional application in which this could be useful is the fixation of metal plates in vertebrae stabilization. The stabilization of vertebrae following injury, degeneration, or surgical intervention represents one of the most technically demanding challenges in spinal surgery. Conventional approaches have relied heavily on screw-based fixation systems, including plates with screws, pedicle screw-rod constructs, and polymethylmethacrylate (PMMA)-augmented pin or screw constructs [69–71]. However, a growing body of research has begun to explore alternatives that either reduce reliance on screws or eliminate them entirely, driven by concerns over screw loosening, implant failure, neurovascular injury, and imaging interference [72,73]. This falls outside the scope of this manuscript, but it is another area where these results could find utility. Future studies could use adhesives based on acrylamides, which are shown to resist very low pH or enzyme degradation [74], which is relevant in cases where there is inflammation present, such as in the immediate post-operatory condition, or in infections.

Even though dental adhesives are FDA-approved for clinical use, cytotoxicity concerns for this proposed off-label use still need to be investigated. In this study, the cytotoxicity of these materials against fibroblast and osteoblast cell lines were evaluated, as relevant cell types this material may encounter in this application. The dilutions selected here are estimated to be above the potential leachates for these materials, [75] simulating a worst-case scenario of tissues being exposed to unreacted monomer. Gen 5 and Gen 7 did not show any cytotoxicity, regardless of the concentration. Gen 4 and Gen 6 adhesives only had statistically higher cytotoxicity at the highest concentration tested (6 μg/mL), which is much higher than previously reported concentrations of leachates in dental adhesives, and unlikely to be found in clinical applications. [75] Based on these results, the Gen 6 adhesive was selected for the next set of experiments, where the addition of quaternary-ammonium methacrylates (QAM) to the PMMA-based cement was tested to reduce bacterial colonization, in this case S. aureus. Again, in the presence of contamination, antibiotic therapy is likely needed, which serves as the justification for these studies. PMMA-based bone cements infused with antibiotics are common, but the inherent antibiotic burst-release effect limits long-term efficacy, which can lead to reintervention, and potential for antibiotic resistance [76,77]. In this study, we purposefully added co-polymerizable antibiotics to the cement, which was tested against the following controls: a cement containing no QAM, a cement containing a QAM analogue without methacrylate functionalization (CTAC) and Refobacin, a commercial antibiotic-containing cement [12,14,61,78]. One additional control based on 2-hydroxyethyl methacrylate (HEMA) was incorporated to assess whether the addition of a monomer with longer side chains (as compared to the methyl group in methyl methacrylate) would disrupt physical and mechanical properties [79]. In summary, all physical properties for the modified materials (HEMA, CTAC or QAM) were either similar to or better than the unmodified controls (Jet Acrylic and Refobacin), with the following exceptions: Jet + CTAC had statistically lower KHN hardness value and the water sorption and solubility were greater for the CTAC and QAM materials compared to the unmodified control, which had similar sorption/solubility to the HEMA-modified material. This can be explained by the increased hydrophobicity imparted by the long alkyl chain present in the quaternary ammonium materials (QAM and CTAC) [80,81]. Other studies have shown similar trends. For example, Liang et al. studied QAM-modified urethane dimethacrylates and found these polymers had higher water sorption and solubility compared to BisGMA-based materials, despite having higher degree of conversion [82]. However, a reduction in mechanical properties was observed, likely due to the addition of the long alkyl chain QAM on the urethane dimethacrylate backbone. Nevertheless, the increase in water sorption was attributed to weaker intermolecular interactions between polymer chains and increased charge concentrations within the materials. As far as the antimicrobial effect, as expected, the quaternary ammonium-containing materials (Jet + CTAC and Jet + QAM), as well as Refobacin prevented S. aureus biofilm formation on the surface of cement discs, with significantly lower values for biofilm biomass. Previous research has found that these monomers are compatible with dimethacrylate-based adhesives through copolymerization to create antibacterial polymers and provide a surface charge density that can inhibit biofilm formation [83].

The cements were further tested in a physiologically relevant environment, similar to what was described for the adhesive screening conducted in the first phase of this study. Shear bond strength was tested with and without mechanical and bacterial challenges, in the bioreactor and via incubation in S. aureus cultures, respectively. Given the non-deleterious effects of QAMs on the properties of PMMA-based cements, and the observed antimicrobial properties of the modified cements (Fig. 13), it was somewhat surprising to see a greater than 50% drop in shear bond strength in Jet + QAM, irrespective of aging condition. This drop was not observed with the Jet + HEMA group. Likewise, Refobacin also showed approximately 44% lower bond strength than Biomet, the antibiotic-free commercial control (Fig. S5), tested in the bioreactor. One possible explanation for this relates to the water sorption and solubility increase seen for Jet + QAM and Refobacin (Fig. 13A). The non-aged specimens were exposed to water for 24 h, while the aged specimens were exposed to water for 5 additional days, followed by 3 days in S. aureus culture. Therefore, it is possible that the increased water sorption during this period caused a plasticizing effect, weakening the bonded interface and lowering the shear bond strength. A previous study observed a similar phenomenon with QAMs added to BisGMA/TEGDMA/HEMA dental resins, where 3-day water storage resulted in nearly a 50% reduction in Young’s modulus and 80% reduction in ultimate stress in QAM-modified materials compared to the control [84]. Additionally, a reduction in flexural modulus has been reported for QAM-modified self-cure resins and pointed to both increased water sorption and dissolution issues of the QAM in the resin as an explanation for the reduced mechanical properties [85]. In the case of Refobacin, these results can be explained by at least two factors: 1. this cement presented the lowest degree of conversion of all materials tested, explaining the lower mechanical properties and higher WS/SL. Though not assessed for this particular material in this study, this also brings concerns in terms of biocompatibility; 2. the presence of gentamicin, a complex aminoglycoside antibiotic, with a strong polycationic character at physiological pH, very likely increased the polarity of the material, explaining the greater WS/SL compared with Biomet. One additional explanation relates to the pH of self-etching adhesives, and its potential deleterious effects on the redox polymerization of self-cure materials [86]. In fact, dental manufacturers promote specific adhesives to be paired with specific resin cements. It has been shown that some self-curing resin composites are incompatible with certain bonding systems due to factors like pH, miscibility, adhesive viscosity and initiator concentration [87,88]. With the variety of adhesive application and curing methods, paired with the vast range of manufacturers and material choices, system optimization is a complex issue beyond the scope of this paper. However, it can be speculated that shear bond strength of the QAM-modified cement could be improved with different adhesive monomers.

Planktonic growth, biofilm formation, and CFU analyses provide complementary insight into how PMMA-based cements influence bacterial behavior both at the material surface and in the surrounding environment. While surface-adhered biofilm is most directly implicated in periprosthetic infection, planktonic bacterial viability reflects the capacity of materials to release antimicrobial cues or generate bactericidal interfaces that affect bacteria beyond direct contact [89]. In the present study, these outcomes were strongly material-dependent and further modulated by water sorption and solubility (WS/SL), highlighting the importance of assessing antimicrobial performance under aging-relevant conditions. Comparing the commercial bone cements (Biomet and Refobacin) with their Jet-based analogues, the optical density data of planktonic S. aureus clearly demonstrated that incorporation of quaternary ammonium methacrylate (QAM) into PMMA cement provided the most pronounced and stable antibacterial effect (Fig. 13B). Jet + QAM consistently exhibited the lowest planktonic bacterial growth both before and after WS/SL, suggesting that its antibacterial activity was likely due to material-bacteria interaction. To confirm this hypothesis, the water in which the discs had been stored was lyophilized and analyzed with 1H NMR (Fig. S6, supplementary information). The liquid was only collected after 7 days, which means any trace of QAM in the solution represents the cumulative release over that period, likely not representative of the amount still left in the polymer disc itself. Albeit at a low concentration, there were traces of QAM in the 7 day cumulative leachate. Therefore, one additional experiment was conducted: discs (n = 6) of the control material (JET) and the Jet modified by adding QAM were produced with the same dimensions used in the WS/SL experiment. These were stored in deionized water for a total of 14 days, replacing the water at 7, 9 and 14 days. The storage water was collected and lyophilized to isolate potential leachates, which were re-suspended in d-chloroform for ¹H NMR analysis. The discs were then subjected to biofilm assays as described, and the results are shown in Figs. S7 and S8. In summary, the antibiofilm effect is maintained at 14 weeks, even in the absence of QAMs in the leachates. This observation is consistent with the established contact kill mechanism of polymerizable quaternary ammonium compounds, which exert bactericidal effects through electrostatic interactions between the positively charged surface and the negatively charged bacterial membrane, leading to membrane disruption and cell lysis [90,91]. In contrast, the gentamicin-containing cement (Refobacin) demonstrated a reduction in planktonic bacterial growth prior to WS/SL but showed diminished antibacterial effectiveness after prolonged water exposure. This behavior is consistent with the well-documented burst-release profile of antibiotic-loaded PMMA cements, in which an initial high concentration of antibiotic is rapidly eluted, followed by a sharp decline below inhibitory level [92,93]. The post-WS/SL increase in planktonic bacterial growth for Refobacin suggests that water conditioning likely exhausted the readily available gentamicin reservoir, leaving the cement surface unable to suppress bacterial proliferation.

The biofilm biomass results further corroborate the planktonic findings while emphasizing surface-mediated effects (Fig. 13C). Both Biomet and Jet Acrylic consistently supported high levels of S. aureus biofilm formation before and after WS/SL, underscoring the inherent susceptibility of unmodified PMMA surfaces to bacterial colonization [5 94,95]. Refobacin and Jet + QAM, however, showed significantly reduced biofilm accumulation, with Jet + QAM demonstrating the most robust inhibition prior to aging. After WS/SL, although overall biomass decreased across all groups, likely due to surface conditioning and changes in surface chemistry, the relative ranking among materials remained similar. Notably, Jet + QAM maintained reduced biofilm formation comparable to Refobacin, despite the absence of antibiotic release. This suggests that QAM-modified PMMA exerts its antibiofilm effect primarily through surface charge-mediated interactions rather than diffusible antimicrobial activity. Importantly, CFU quantification provided a stringent and biologically meaningful endpoint that integrates both surface attachment and bacterial viability. The complete absence of recoverable S. aureus from Jet + QAM specimens in both pre- and post-WS/SL conditions confirms that this material not only limits biofilm biomass but effectively prevents bacterial survival and/or adhesion altogether (Fig. 13D). This finding strongly supports the hypothesis that copolymerized QAMs create a bactericidal surface rather than merely inhibiting growth [96]. In contrast, the re-emergence of viable bacteria on Refobacin after WS/SL highlights the vulnerability of antibiotic-based strategies to environmental conditioning and further underscores the advantage of non-leaching, contact-active antimicrobial approaches [97]. Taken together, the planktonic, biofilm and CFU data indicate that covalently integrated QAM incorporation provides a stable and persistent antimicrobial effect that is maintained even after prolonged water exposure. From a clinical perspective, this is particularly relevant given the aqueous and protein-rich environment surrounding orthopedic implants, where materials are continuously exposed to fluids that can accelerate leaching, plasticization, and surface conditioning [98]. The ability of Jet + QAM to suppress both planktonic growth and surface colonization after WS/SL suggests that such materials may better withstand early post-operative conditions and reduce the risk of bacterial persistence at the bone-implant interface. While Jet + QAM demonstrated reduced shear bond strength, likely associated with increased water sorption and plasticization effects, the pairing with the adhesive meant those bonds strength results were still about 10-fold higher than the cement not containing antibiotics placed without the adhesive. Future work should therefore focus on optimizing QAM structure, concentration, and polymer compatibility to preserve antibacterial activity while mitigating adverse effects on interfacial strength and mechanical performance. Strategies such as incorporating shorter alkyl chain QAMs, multifunctional crosslinkers, or hybrid antimicrobial systems may offer promising pathways to achieve this balance.

5. Conclusions

The results of the present study demonstrate that the use of dental adhesives prior to bone cement application significantly improves bonding interactions in a relatively controlled environment, such as seen in planned arthroplasty surgery, and may be a suitable strategy to improve outcomes. Cytotoxicity results demonstrate that this approach is safe for at least two cell lines relevant to this application, and that the improvement in bond strength is highly significant, even after mechanical cycling. Furthermore, the use of adhesives is able to counterbalance the reduction in bond strength observed with QAM-modified bone cements compared to the current standard of care (no adhesive application). Given that the addition of QAM led to significant S. aureus biofilm inhibition without compromising the degree of conversion or hardness of the cement, this approach leveraging FDA-approved dental adhesives seems like a promising strategy to resolve key health and safety concerns, which is likely to facilitate translation.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.matdes.2026.115970.

Data availability

Data will be made available on request.