Osteoimmunology is the crosstalk between cells from the immune and skeletal systems, suggesting a role of pro-inflammatory cytokines in the stimulation of osteoclast activity. Endotoxin or bacterial challenges to inflammatory cells are directly relevant to dental implant pathologies involving bone resorption, such as osseointegration failure and peri-implantitis. While the endotoxin amount on implant devices is regulated by standards, it is unknown whether commercially available dental implants elicit different levels of adherent-endotoxin stimulated cytokines. The objective of this work is to develop a model system and evaluate endotoxin-induced expression of pro-inflammatory cytokine genes relevant to osteoclast activation on commercially available dental implants. Murine J774-A1 macrophages were cultured on Ti disks with different level of lipopolysaccharide (LPS) contamination to define the time-course of the inflammatory response to endotoxin, as evaluated by reverse transcription polymerase chain reaction analysis. The developed protocol was then used to measure adherent endotoxin on commercially available packaged and sterile dental implants in the “as-implanted” condition. Results show that tested dental implants induce variable expression of endotoxin-stimulated genes, sometimes above the level expected to promote bone resorption in vivo. Results are unaffected by the specific surface treatment; rather, they likely reflect care in cleaning and packaging protocols. In conclusion, expression of genes that enhance osteoclast activity through endotoxin stimulation of inflammatory cells is widely different on commercially available dental implants. A reappraisal of the clinical impact of adherent endotoxins on dental (and bone) implant devices is required in light of increasing knowledge on crosstalk between cells from the immune and skeletal systems.
Bone homeostasis and remodeling occur throughout life in organisms that possess a skeleton. Bone homeostasis is often regulated by immune responses, particularly when the immune system has been activated or becomes pathologic.1 The name osteoimmunology was coined to identify studies related to interactions between the bone, hematopoietic, and immune systems.2–4 Crosstalk between activated lymphocytes and bone cells occurs throughout life because all mammals are constantly challenged by a variety of infectious agents, producing some level of constant low-grade immune system activation.
A particular instance of immune system activation involves implant devices. In addition to immunological response to surgery (ensuing onset of the inflammatory cascade and wound-healing mechanisms), interrogation of the implanted device by pertinent immune cells occurs at the implant site.5 Cytokines and chemokines produced by inflammatory cells on contact with the implant surface contribute to the peri-implant biochemical environment and to the overall host response. This topic has been widely investigated both in terms of material surface chemistry6,7 and surface topography,8–11 supporting concepts such as healing enhancement through the “programmed” release of pro-healing cytokines by macrophages on properly engineered implant surfaces.12
Beyond the physico-chemical properties of surfaces, it has been shown that endotoxin—the “uninvited guest”13—significantly affects inflammatory cells' response to implant materials, and hence may confound any effect of the material itself. Greenfield and coworkers have published an interesting series of papers14,15 showing that endotoxin adherent to implant surfaces is largely responsible for inducing osteoclast differentiation through production of interleukin-1β (IL-1β), IL-6, and tumor necrosis factor-α (TNF-α) on inflammatory cell-material contact. Adherent endotoxin was found at significant levels on the commonly used preparation of commercially pure titanium particles, as well as on orthopedic titanium implant surfaces. Removal of adherent endotoxin14 almost completely inhibited the responses to titanium (Ti) particles by both murine marrow cells and human peripheral blood monocytes. In vivo experiments showed that endotoxin removal reduced particle-induced osteolysis by 50–70%, while addition of lipopolysaccharide (LPS) to the “endotoxin-free” particles restored their ability to induce cytokine production and osteoclast differentiation in vitro. Thus, adherent endotoxin —and not inflammatory response to particle and materials— could be at the basis of the widely investigated “aseptic loosening” of orthopedic prostheses.
In a paper published in this journal in 2001, Wataha and coworkers investigated the effect of LPS contamination on the attachment of osteoblast-like cells on titanium in vitro,16 building on the clinical evidence that failing implants with loss of alveolar bone are associated with gram-negative bacteria that carry LPS in the bacterial cell wall. Their study failed to detect evidence on the effect of LPS on the attachment of osteoblast cells to titanium surfaces in vitro, concluding that “further research is needed to define the clinical liabilities of LPS during implant placement and maintenance.” 16 While not fully developed at that time, today the role of crosstalk between cells of the immune and the musculoskeletal systems provides the missing link that can shed more light on this topic.
The in vivo effect of adherent endotoxin on osseointegration of titanium implants has recently been discussed by Omar and coworkers.17 LPS was first adsorbed through incubation from aqueous solution on machined and anodized dental implants that were then implanted in pig femoral diaphyses. After 2 weeks, histological analysis showed large areas of inflammatory infiltrates with active bone resorption, both around the neck as well as the middle and lower parts of LPS-adsorbed implants, independent from the nature of the implant surface. After 6 weeks, LPS-incubated implants demonstrated comparable bone morphology and amount in contact with the implant surfaces as implants not incubated in LPS. The quoted paper nicely describes the interplay between classically activated macrophage and osteogenic cells, ultimately leading to a positive solution of the endotoxin challenge in the peri-implant area. However, it is clear that the detected short time of 2 weeks' bone resorption and ensuing lack of stability around LPS-adsorbed implants is highly relevant from a clinical point of view, especially in light of the increasing demand for immediate or early loading. Thus, the amount of adherent endotoxin could rightfully be considered a further clinically relevant variable of dental implant surfaces. Since the 1980s, many papers describe the chemical contamination of clinically available implant surfaces18–20 and the interplay between biological stimulation and surface parameters,21–24 but no published comparison exists on the level of adherent endotoxin. It is worthy here to mention that the endotoxin amount on implant devices is regulated by standards and routinely checked by producers; however, measurements are performed on aqueous extracts, not directly on the implant surface.25
The aim of the present paper is to evaluate the potential for endotoxin-stimulated pro-inflammatory response of several commercially available dental implants. First, we validated a detection method for adherent endotoxin by developing an in vitro simple model that measures endotoxin activity directly on implant surfaces. In particular, gene expression of pro-inflammatory cytokines on implant surfaces can be routinely measured by Real-Time PCR. We followed the time course of inflammatory transcripts in a model system, involving the continuous murine macrophage cell line J774A-1 and intentionally LPS-contaminated titanium disks. We then showed that the level of LPS contamination, independent from surface topography, controls the short time (4 h) IL expression by the selected cell line. Finally, we performed the same measurement on commercially available, sterile and sealed dental implants (ie, in the “as-implanted” condition) from different producers to evaluate if and how much they elicit pro-inflammatory activity.
Methods and Materials
For the method validation, tests were performed on 8 mm diameter grade 4 Ti disks and on 3.75 × 13 mm grade 4 Ti dental implants; all samples were produced by the same supplier, and all surfaces were simply turned when received in our laboratory. All subsequent steps were performed according to our ISO 9001:2008 and ISO 13485:2004 quality standards and protocols. The following samples were prepared for the following objectives:
Evaluation of the time course and the dose-dependence of the response. Tests were performed on endotoxin-free titanium disks and on LPS-contaminated Ti disks. The endotoxin-free sample (negative control, coded Ctrl in the rest of the paper) was obtained by subjecting Ti disks to a dedicated proprietary cleaning treatment, involving both solvent and plasma (glow discharge) cleaning cycles, using a Plasma Finish microwave reactor placed inside a ISO7 clean room. The absence of detectable endotoxin on the endotoxin-free control sample was confirmed by LAL tests performed by an external lab, where the sensitivity of the test was 0.125 EU/mL. LPS-incubated samples were obtained by overnight incubation of formerly endotoxin-free samples in 1 μg/mL, 5 μg/mL, and 10 μg/mL LPS (Sigma) in phosphate-buffered saline. After incubation, the disks were washed three times in MilliQ pyrogen-free water (MilliQ Synthesis A-10, EMD Millipore Corp, Billerica, Mass) and dried under a laminary flow hood. With reference to the LPS concentration of the incubation solution, these samples are coded LPS1, LPS5, and LPS10 hereafter.
Evaluation of the surface topography contribution to gene expression involved in macrophage response to adherent endotoxin. Titanium implants were subjected to the following treatments in our lab:
Sandblasting using titanium oxides, 250–400 mesh, coded Tiblasted (Sa 1.83 μm, as measured by StereoSEM in a 130 × 120 μm area).
Double acid etching treatment, coded DAE (Sa 0.91 μm, as measured by StereoSEM in a 130 × 120 μm area).
Tiblasted and DAE samples, together with machined implants (coded Mach, Sa 0.47 μm, measured as above), were tested with both prepared under the following protocol: testing occurred after a solvent cleaning cycle involving nitric acid passivation, neutralization, and DI water after the complete cleaning cycle discussed above.
More details will be provided in the “Results” section.
Beside samples used for process validation, a second set of samples was tested for actual measurements through the developed method. This set was made by 22 commercially available dental implants from different worldwide producers. All samples were received sterile and sealed in their original package, all well before their relevant expiration date. All were made from commercially pure titanium (cpTi).
Gene expression measurement through reverse transcription polymerase chain reaction (RT-PCR) was performed to evaluate the amount of adherent endotoxin. Tests were performed through the evaluation of the expression by J774A-1 macrophages of a few key genes involved in the inflammatory response to endotoxin: interleukin 1 (IL-1), interleukin 6 (IL-6), tumor necrosis factor-α (TNFα), MCP-1, COX-2, and MCSF.
A suspension of 1.05 ± 0.13 × 105 J774A-1 cells, cultured in DMEM containing L-glutamine (Gibco, INVITROGEN S.R.L, San Giuliano Milanese, Italy), 20% fetal bovine serum (FBS Gibco, INVITROGEN), penicillin, and streptomycin, was introduced into sterile 12-well polystyrene culture plates (CELLSTAR, Greiner Bio-One GmbH, Frickenhausen, Germany) containing the samples. Gene expression analysis was carried out using real-time reverse transcription PCR (qRT-PCR). Total RNA was extracted after 1, 4, and 24 hours using the MagMax Total RNA Isolation Kit (Applied Biosystems, Grand Island, NY). The RNA was assessed by ensuring the A260/A280 absorbance ratio was between 1.6 and 2.0. The extracted RNA was subsequently reverse transcribed to give cDNA using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems).
Relative quantification of the genes was obtained using Taq Man probes specific for each tested gene, with GAPDH as the reference gene. The amplification reactions were carried out in a StepOne thermal cycler (Applied Biosystems) in accordance with the manufacturer's instructions. To obtain the gene expression graphs, data were normalized using the StepOne software (Applied Biosystems) in accordance with the ΔCt standard method. Measurements were performed in triplicate for samples involved in the validation step. Measurements on the 22 test samples were obviously performed on a single specimen for each tested implant, so amplification reactions were performed in triplicate aliquots of cDNA to verify intra-experiment accuracy.
As a first step, expression of IL-1, IL-6, TNF-α, MCP-1, COX-2, and MCSF by J774A-1 macrophages was measured on control and LPS-adsorbed Ti disks as a function of time. As expected, all quoted genes were significantly overexpressed on LPS-contaminated samples, reflecting the onset of defense mechanisms against endotoxin challenge by macrophages. Among tested genes, both IL-1 and IL-6 showed significant upregulation as a function of LPS concentration. Figures 1 and 2 show fold-expression of IL-1 and IL-6, respectively, by J774A-1 macrophages adhering to the Ti disks as a function of time and concentration of LPS in the adsorption solution. Data are presented as fold expression over the value obtained on the control sample at 1 hour. The figure suggests that peak expression occurs at 4 hours (at least among the three time points investigated), and that there is a clear dose-dependent response of gene expression. Interestingly, while the LPS-free control sample shows an increasing trend of gene expression, likely plateauing in the 4- to 24-hour timeframe, macrophages on LPS-contaminated disks yield a burst response that is turned off, or at least attenuated, in the same 4- to 24-hour time span. To provide an analytical method for LPS contamination detection on Ti surfaces, Figures 1 and 2 suggest that the measurement of IL-1 or IL-6 expression by J774A-1 macrophages at 4 hours is a suitable approach, since it provides both sensitivity and dose-dependence within the tested range. Thus, the experimental protocol for adherent-endotoxin measurement would include evaluation of IL expression at 4 hours on test samples, using the control sample as a reference. When this is performed on the present data—that is, when 4-hour values of LPS1, LPS5, and LPS10 samples are expressed as fold expression over the 4-hour control value—the bar graph shown in Figure 3 was obtained. This is the “master curve” of the present adherent-endotoxin detection method: it shows that expression of both IL genes is dependent on the amount of LPS in the adsorption solution (and, therefore, on the amount of surface-adsorbed LPS), and that fold expression spans a significant analytical range, from slightly more than 10 to more than 100 in the tested conditions.
A further required step for the qualification of the present approach as a test method is the demonstration of selectivity. In particular, while the aforementioned data were obtained on machined disks, it has been reported that surface topography affects macrophages response.10–12 Hence, it would be impossible to meaningfully compare macrophage gene expression obtained on surfaces with different topographies. This would be a serious limitation of the method given the immense variety of existing approaches to control surface roughness of dental implants.26 To check this point, we measured 4-hour IL expression by J774A-1 macrophages on Ti implants showing different topographies: Mach, Tiblasted, and DAE, as defined in the “Methods and Materials” section. Measurements were performed after surface treatment and simple solvent cleaning, and after surface treatment followed by the complete endotoxin-removal cycle previously described. Obtained results are shown in Figure 4. Interestingly, the solvent-cleaned Mach is more proinflammatory than any other sample. Both sandblasting and acid etching—which destroy the pristine, environment-exposed surface—show some effectiveness in decreasing the endotoxin response to Ti surfaces. In addition, Figure 4 shows that both IL-1 and IL-6 expression is significantly dampened by the full endotoxin-removal cycle as compared to simple solvent cleaning. Most importantly, it shows that macrophages on the tested surfaces express the same level of IL transcripts after endotoxin removal, irrespective of the significant variation in surface roughness, as encoded by the Sa value and by the specific topography. These data show that the method provides the required selectivity: in the short time (4 h), IL expression by macrophages is not a function of physical parameters, such as surface roughness, but is solely controlled by the defense response to endotoxin.
Having shown that the test method provides sensitivity, dose-response, and selectivity, we moved to its actual application to clinically relevant samples. In particular, macrophages were cultured on 22 titanium dental implants from different worldwide producers. All samples were sterile and sealed in their package, in their “as sold” condition. Samples encompassed most of the presently adopted approaches to surface roughening: some were sandblasted, some acid etched, and some subjected to electrochemical treatment.27 No sample had a machined surface.
Obtained results are summarized in Figure 5, showing 4-hour fold expression of IL-6 by J774A-1 cells over that measured on a control sample. As a reference, horizontal lines in the figure show the fold expression obtained on purposely LPS-contaminated LPS1, LPS5, and LPS10 samples, as reported in Figure 3. In considering these data, it is important to remember that the reference—that is, the control sample—was not the same one for validation of the method (samples LPS1, LPS5, and LPS10) or for the testing of the actual implants. Thus, the direct comparison of the data rests on the underlying assumption that the cleaning cycle adopted to prepare the different control samples yields the same level of adherent endotoxin and, hence, the same macrophage response. While this assumption should be taken after due consideration, the data of Figure 5 indicate that clinically available dental implants show wide variation of adherent endotoxin, evoking a significantly different device-induced macrophage activation. Seven out of the 22 tested samples show endotoxin-induced IL-6 expression higher than that promoted by a titanium surface incubated overnight in a 1 μg/mL LPS solution: among these seven samples, three show endotoxin-induced IL-6 expression higher than that promoted by a titanium surface incubated overnight in a 5 μg/mL LPS solution, and one of them higher than that promoted by a titanium surface incubated overnight in a 10 μg/mL LPS solution. A few implant surfaces show the lack of any response to adherent endotoxin, suggesting an almost perfect control of surface contamination in the production and packaging steps. Considerations stemming from these data are reported in the following section.
The growing field of osteoimmunology underlines the role of inflammatory stimuli in triggering bone-loss pathologies. Crosstalk between cells from the immune and skeletal systems is of particular relevance in clinical dentistry that features remodeling bone tissue in close proximity to the thriving oral bacterial population.28,29 The present work focused on activation of inflammation by adherent endotoxin on implant surfaces, in addition to the response to bacteria. The classical bacterial endotoxin is LPS, the primary outer cell wall component of Gram-negative bacteria. However, Gram-positive bacteria also produce molecules, such as lipoteichoic acid (LTA) and peptidoglycan, with very similar biological effects.30 Thus, many possible sources of bacterial endotoxins (LPS, LTA, peptidoglycan, etc) exist that might lead to high levels of adherent endotoxin on implant devices.
The adopted approach exploits the direct response of a model continuous cell line, robust and reproducible to endotoxin stimuli. By using a set of intentionally LPS-contaminated samples, we showed that expression of key inflammatory genes is directly related to the amount of adherent endotoxin, and that the 4-hour response time point provides the required sensitivity, dose-response dependence, and selectivity. More sophisticated and clinically relevant (eg, human monocytes) cell lines are not required for the present scope,31 since the role of well-behaved J774A-1 murine macrophages in the present approach is to act as a sensitive “probe” that provides the required answer.
The developed test method was used to evaluate the amount of adherent endotoxin on commercially available implant surfaces. Obtained data show a vast variation of adherent endotoxin. A few of the tested samples are virtually endotoxin-free, while the response is similar to that obtained after overnight incubation in endotoxin solution in some instances. These differences likely reflect the quality of procedures adopted in the production and packaging steps and the care adopted to remove and prevent endotoxin contamination.
From a basic point of view, present data show that the amount of adherent endotoxin is an additional variable that affects cell response on dental implant surfaces. While the latter are often discussed in terms of chemico-physical variables, such as surface chemistry and topography,32,33 present data show that cell response (at least in the case of the tested cell line) is overwhelmingly dominated by “biological cleanliness,” or the amount of adherent endotoxin. This is obviously true for tests involving inflammatory cells, such as the J774A-1 macrophages adopted in the present work; however, implications are far ranging since most cells respond to endotoxin stimuli, including fibroblasts, endothelial, and osteoblasts. Thus, every study on cell response to implant surfaces (and material in general) should include the evaluation of adherent endotoxin among surface properties that require proper characterization and definition.14 This is particularly true for samples prepared in common lab environments that often lack facilities available to producers of medical devices.
A further consideration involves the clinical implications of present findings. In a recent paper, Omar and coworkers showed that activated human monocytes communicate pro-osteogenic signals to human Mesenchimal Cells (hMSCs).17 The signals involve regulation of autologous BMP-2 in the hMSCs, where the contribution by LPS stimulation of monocytes overwhelmed the effect of the surface properties. The quoted paper uses an in vivo pig model to show that intentionally contaminated titanium implants (1h LPS adsorption from a 10 μg/mL solution) resulted in excessive resorption/remodeling activity at the early 2 weeks of implantation. However, after 6 weeks, LPS-induced resorption areas were replaced with a higher percentage of bone contact, suggesting a possible major role for osteoclastic feedback on the process of bone formation at the interface. Thus, early upregulation of both bone resorption and bone formation genes could be associated with significant increase in implant stability.
This quoted data indicate the outcome of long-term exposure of bone tissue to LPS-contaminated implants, underlining the complicated and fascinating mechanisms that control healing and new tissue generation within our bodies once again. In clinical practice, especially in light of immediate or early loading, the detection of significant (see the impressive Figure 11 in Omar et al17 ) bone resorption and lack of bone tissue around LPS-contaminated implants underline the relevance of our present findings. A few of the clinically available implants we tested show endotoxin response in the range obtained on LPS-contaminated samples, using a similar concentration to that adopted in the quoted reference, as well as the ability to induce the described profound in vivo effects. Also note that our protocol involves overnight LPS adsorption, while the quoted paper involved adsorption lasting just 1 hour, which likely resulted in a lower adsorbed amount that remained deeply effective in vivo, as compared to our LPS-contaminated reference samples. Thus, it is possible that early loading of the most contaminated samples we tested could find no properly regenerated bone tissue; more generally, data of Omar et al. show that the amount of adherent endotoxin modulates time and amount of peri-implant bone regeneration. This also confirms our previous suggestion at the in vivo level: the amount of adherent endotoxin is an additional variable required for proper characterization of dental implant surfaces. Since its contribution can overwhelm that of “classical” chemico-physical parameter, evaluation of the adherent endotoxin amount is highly required for a correct understanding of peri-implant bone regeneration.
The last comment involves the observation that, among tested implants, a few of them are virtually free from adherent endotoxin, even when probed by sensitive and surface-specific tests as in the present study. This result underlines that methods and techniques exist to produce and market actual “implant devices,” as opposed to “titanium fixtures,” that are designed and produced with a proper understanding and control of cell interfacial biology highly relevant for the intended clinical use.
In conclusion, the present data show that short time (4 h) expression of proinflammatory genes—in particular, IL-1 and IL-6—by J774A-1 macrophages is directly and selectively related to the amount of adherent endotoxin, and it is largely independent from surface topography. The application of this approach to several clinically available dental implants shows significant heterogeneity among tested products, some of them virtually free from adherent endotoxin, with a few showing IL upregulation similar to that detected after overnight incubation in LPS solutions. Given the ascertained in vivo effect on peri-implant bone regeneration provided by LPS contamination, evaluation of adherent endotoxin should be reappraised and ranked among relevant surface properties required for proper understanding of interfacial tissue response to dental implants.