Spectroscopic and microscopic investigation of the effects of bacteria on dental implant surfaces

Abstract

The surface morphology and chemical composition of commercially pure titanium dental implants and healing abutments exposed in vitro or in vivo to oral bacteria were studied. Implants exposed in vitro to media containing S. mutans and healing abutment retrievals were used to evaluate the effect of early bacterial colonizers. Another dental implant retrieval that failed due to peri-implantitis and implants exposed to an in vitro culture of P. gingivalis were analyzed to assess the effect of late stage colonizing bacteria. Using optical microscopy and X-ray photoelectron spectroscopy, it was determined that both S. mutans and P. gingivalis adversely affected the implant surface, with the early colonizing S. mutans creating conditions leading to more severe corrosion features being observed, including pitting and surface discoloration. Overall, this study demonstrated the ability of bacteria in the oral environment to solely induce changes in the oxidation states of titanium and subsequent corrosion of the implant surface, even in the absence of mechanical loads.

---

1. Introduction

Dental implants have been reported to have long-term predictability and performance.¹ The success of these implants is directly related to the degree of integration of the surface first with soft tissues and then with bone.^2,3 Titanium has been the most commonly employed material in the design of dental implants due to a combination of properties but particularly due to its ability to spontaneously form passive oxide layers. This passive oxide layer (TiO₂) has been reported to facilitate osseointegration because its porous structure allows for incorporation of calcium ions and phosphate.^4,5 A series of surface modification techniques, chemistry and material combinations have been proposed to further improve the surface properties of titanium in order to enhance osseointegration^6,7 or mitigate bacterial infections.^8–10 However, implant failures still occur at a rate of 5–10%, and due to the growing popularity of these devices, the incidence of short-term and long-term clinical complications has increased in recent years.¹¹

Implant failures can be induced by a number of individual or synergistic events.^12–16 Typically, these failure mechanisms are classified as either early or late stage failures.^17,18 Early stage failure occurs when the establishment of osseointegration is interrupted mainly due to bacterial contamination, premature overloading, impaired healing and excessive surgical trauma.^17,18 Late stage failures involve breakdown of the established osseointegration, having major etiological factors associated with bacteria-induced marginal bone loss (peri-implantitis) and excessive occlusal stresses.¹⁸ Bacteria is clearly involved in the early and late stages of implant failure, inducing surface changes that can trigger inflammatory reactions in soft tissues.^17–19 According to Zhao et al., the performance of a dental implant surface is dependent on a “race-for-the-surface” between human gingival fibroblasts and bacterial biofilm adhesion.² Early failure can be initiated by surface adhesion of early bacterial colonizers such as Streptococci and Actinomyces species, which upon formation of a biofilm will prevent sealing of epithelial tissues on the implant surface.^20,21 Failure to achieve soft tissue integration on implant surfaces will enable the first colonizers to infiltrate and interrupt the early stages of osseointegration.²² Attachment of these early colonizers will create the necessary anaerobic conditions for pathogenic late colonizers such as Aggregatibacter actinomycetemcomitans, Porphyromonas gingivalis and Fusobacterium nucleatum.^20,21,23 These periodontal pathogens can ultimately lead to bone loss and are considered causative agents of peri-implantitis.²⁴ Inflammation due to bacterial biofilm is then assumed as the primary etiological factor for peri-implantitis.²⁵ However, other factors including, excessive mechanical loads,²⁶ corrosion,^4,27 excess cement²⁸ and previous periodontal disease history²⁹ have also been reported to play an important role in the inflammatory process.

Although, bacteria is a major cause for both early and late stage failures, there are only a few studies that discuss the ability of bacteria to create unfavorable conditions leading to implant surface damage.^30–32 We demonstrated in a recent in vitro study that immersion of dental implants in medium containing Streptococcus mutans for 60 days led to significant changes in surface morphology with evidence of corrosion and generation of metal ions in solution.³³ Wilson et al. reported histopathological findings in soft tissue biopsies of implants with peri-implantitis, demonstrating that the predominant foreign bodies found were titanium and dental cement surrounded by inflammatory cells.³⁴ Retrospective studies of the surface of dental implants retrieved from patients with peri-implantitis also illustrate the degree of surface damage that can be imparted by late colonizers. Morphological features such as pitting, surface etching, discoloration, cracking, scratches and fractures were observed on the surface of these implants.^35–37

Bacterial biofilms are assumed to change the electrochemical environment surrounding an implant, which can lead to disruption of the surface oxide layer and induce osseointegration loss.^34,38 This is hypothesized based on two mechanisms: (1) contact and adhesion of early-colonizing planktonic bacteria on surfaces that release lactic acid during glycolysis, thereby decreasing the pH of the oral environment. Acidic environments can result in changes in the oxidation states of titanium and can lead to metal ion dissolution;^30,39,40 and (2) biofilm formation on implant surfaces, leading to the development of localized crevice environments on the implant surface. Oxygen depletion within these crevices will induce a significant decrease in local pH and consequent metal surface attack.^41–43 Implant micromotion, resulting from occlusal loading, in the presence of a stagnant acidic environment can lead to permanent disruption of the oxide layer through a mechanism of fretting-crevice corrosion.⁴⁴ Besides leading to active metal ion and particle generation, this process can result in disruption of osseointegrated and sealed surfaces, exacerbating inflammatory reactions.⁴⁵

Hence, it has been suggested that metal ions and wear release can affect the biological response of cells that are vital to attain implant integration with soft and bone tissues.^37,46 The possibility of implant corrosion mediated by bacteria requires a better understanding of the role of early and late colonizers on the surface of titanium dental implants. The primary goal of this study was to evaluate whether bacteria found in the oral cavity are able to change the surface topography and surface composition of dental implants and abutments exposed to them.

2. Experimental section

2.1. Samples

A pool of six dental implants and three healing abutments were analyzed from in vivo retrievals and in vitro testing conditions, along with control specimens for each group. Available implant information is summarized in Table 1.

Table 1 Summary of implants used in the study

Sample	Brand	Dimension	Morphology and surface treatment	History	Retrieval stage
a Not available.b No exposure to medium containing bacteria.c Used as control against implants or components analyzed as early or late stage failures.
Implant (IC)	Straumann LLC., (Andover, MA)	4.1 × 10 mm	Regular Neck (RN), Sand-blasted, Large-grit and Acid-etched (SLA)	NE^b	None^c
Implant 1 (I1)				Immersed in medium containing Streptococcus mutans (in vitro)	Early
Implant 2 (I2)				Immersed in medium containing Porphyromonas gingivalis (in vitro)	Late
Implant 3 (I3)
Implant 4 (I4)
Implant 5 (I5)
Implant 6 (I6)	NA^a	4.1 × 8 mm	Unknown	Retrieved implant (in vivo)
Healing Abutment Control (HAC)	Straumann LLC., (Andover, MA)	4.5 × 6 mm	Regular connection (RC), conical emergence profile, machined surface	NE^b	None^c
(HA1)				Retrieved healing abutments (in vivo)	Early
(HA2)
(HA3)

---

A retrieved implant removed from a patient presenting a case of peri-implantitis and three healing abutments (HA) were received from a private periodontal practice (Dallas, TX). Patients provided consent as according to the guidelines of the Helsinki Declaration to donate their retrieved implants and abutments for research. Upon retrieval and storage, there were no identifiers that could be linked to the patient who donated the implant/abutment. Therefore, information like time elapsed from implant placement to implant removal and surface treatment were not available.

2.2. Methodology

2.2.1 In vivo retrieval analysis of dental implant and abutments. The surface of three regular connection healing abutments for bone-level implants was analyzed to evaluate the effect of early colonizing bacteria during the healing period in the oral environment. A failed dental implant removed from a patient with peri-implantitis was investigated to understand the effect of late colonizing bacteria on the surface in the in vivo environment. This particular implant suffered a fracture and was broken into two parts as received. The bottom part was analyzed as it contained the majority of the previously osseointegrated region.

A visual inspection was initially performed to detect particular areas of interest and gross features and to verify the severity of corrosion present (discoloration, cracking and debris). Specimens were then analyzed with a digital optical microscope (OM) for identification of surface features and failure mechanisms. Areas of interest were further analyzed with X-ray photoelectron spectroscopy (XPS) to evaluate the elemental composition and the nature of chemical interactions on the surface. After surface characterization, all specimens were subjected to a cleaning protocol for complete removal of biological deposits that could have covered important surface features. They were first cleaned in an aqueous solution with soap powder, followed by cleaning in deionized water and finally 70% ethanol. For this protocol, specimens were fully immersed in each of the solutions and subjected to sonication for 1 hour. Microscopy and XPS were then repeated.

2.2.2 In vitro immersion testing of dental implants. The dental implants used for in vitro immersion testing were cleaned by ultrasonication while immersed in 70% ethanol for one hour and then in deionized (DI) water for another hour. Prior to immersion in medium containing bacteria, the implant was immersed in ethanol, flame-sterilized and allowed to cool. When contamination was observed during testing, the implant was retrieved from the immersion medium and autoclaved to destroy any microbes attached to it. The autoclaved implant was imaged with optical microscopy and cleaned again using ultrasonication and flame sterilization before being placed in immersion medium again. After immersion testing was completed, the implants were removed from the immersion media and autoclaved.
In vitro immersion testing of dental implant in early colonizing bacteria. An SLA dental implant was exposed to an in vitro bacterial monoculture of S. mutans (UA159), which is an early colonizer.^20,47 The implant was immersed in the medium containing bacteria (pH = 5) for a period of 60 days in order to simulate the initial healing period. S. mutans was struck from a −80 °C freezer stock. The struck agar plate was placed in a gas chamber containing microaerophilic CO₂ condition. After 48 hours of incubation at 37 °C, individual colonies from the agar plate were inoculated in a 5 mL nutrient medium consisting of tryptic-soy broth. Growth of S. mutans was confirmed after overnight incubation for 24 hours by optical density (OD₆₀₀) measurement at 600 nm wavelength. Bacteria were replenished with fresh nutrient broth once every two days throughout the period of immersion. The test tube containing the implant immersed in bacteria was placed in a shaker rotating at a constant speed of 110 RPM at 37 °C.
In vitro immersion testing of dental implants in late colonizing bacteria. Four SLA dental implants were exposed to an in vitro bacterial monoculture of P. gingivalis (ATCC33277), which is considered as a late biofilm colonizer and one of the important strains in triggering inflammatory reactions leading to peri-implantitis.³¹ The implants were exposed to this immersion medium (pH = 8) for a period of 30 days to evaluate the ability of obligate anaerobic pathogen to contaminate and to adhere to the surface of the implant. The immersion medium was not changed or disturbed throughout the time of immersion.

2.2.3 Optical microscopy. The surfaces of three regular connection healing abutments for bone-level implants and a dental implant immersed in S. mutans were analyzed to evaluate the effect of early colonizing bacteria during the healing period in the oral environment. A failed dental implant removed from a patient with peri-implantitis and four dental implants immersed in P. gingivalis were investigated to understand the effect of late colonizing bacteria on the surface in vivo. Surface morphological features of the implants and healing abutments were obtained with a digital optical microscope (Keyence VHX-2000, Itasca, IL). The microscope was equipped with low (5–50×) and high magnification (100–1000×) lenses. This microscope provided 3D mapping of the surface of the specimens, which depicted the surface contour in terms of a color coded depth/height profile. The mapping or depth/height profile of the surface provided valuable qualitative information about surface features such as pitting, delamination and deformation characteristic of damaged surfaces.

2.2.4 X-ray photoelectron spectroscopy. Areas of interest on each specimen were further analyzed with X-ray photoelectron spectroscopy (XPS) to evaluate the elemental composition and the nature of chemical interactions on the surface. Two sets of experiments were performed to evaluate the elemental composition of the specimens described in Table 1 (IC, I1–I6, HAC and HA1–HA3). Initially, the analyses were performed as the samples were received without any cleaning or treatment procedures. In a second step, in order to investigate surface depth profile, samples were sputtered for 2 minutes using an argon source (5 kV). The sputtering process removes superficial layers from the specimen surface and contaminants, permitting new structural rearrangements and revealing other oxidation states. For each sample (referred to “as received” and “sputtered”), a survey spectrum was acquired to first identify the total elemental makeup of the surface followed by a high resolution scan for each element of interest identified in the sample.

Experiments were performed with a PHI 5000 Versa Probe II X-ray Photoelectron Spectrometer (Chanhassen, MN). Measurements were taken at an angle of 45° in relation to the sample surface. Areas scanned in each analysis had dimensions of approximately 200 μm². A monochromatic Al Ka source of 1486.6 eV was used. The survey spectra were acquired using 187.850 eV pass energy and 0.8 eV step size, while the high resolution scans were obtained using 23.5 eV pass energy and 0.200 eV step size. The pressure in the analysis chamber was maintained below 10⁻⁸ Torr. The data was analyzed and normalized using CasaXPS 2.3.16 Software.

XPS allowed for identification of chemical elements present in the samples by measuring the binding energy spectra of emitted X-ray photoelectrons. A typical XPS spectrum is a plot of X-ray intensity versus binding energy (BE). Before a more detailed discussion of XPS data, some general information is provided regarding the XPS profiles of oxygen (O1s) and titanium (Ti2p), which are the main elements investigated in this study. The XPS spectrum of oxygen consists of a singlet peak and its BE varies according to the chemical environment in which oxygen molecules are located. For example, it is already established that when bound to titanium in the oxide layer (TiO₂), the BE of oxygen is around 529.70 ± 0.10 eV.⁴⁸ However, this BE can be shifted to higher values (530.99 ± 0.15 eV) when oxygen is in its hydroxyl form (OH), which can occur when either linked to Ti or bound in water.⁴⁸ The titanium spectrum is characterized by doublet peaks of Ti2p, which are denominated Ti2p_3/2 and Ti2p_1/2. To avoid dubiety in the discussion of results, only the binding energy of Ti2p_3/2 in the Ti spectrum will be investigated and discussed throughout this paper. XPS can detect four distinct electronic states of Ti2p (Ti2p_3/2) with BE values of 453.74 ± 0.32 eV (Ti⁰, metallic state), 455.34 ± 0.39 eV (Ti²⁺, TiO), 457.13 ± 0.35 eV (Ti³⁺, Ti₂O₃) and 458.66 ± 0.22 eV (Ti⁴⁺, TiO₂).⁴⁸ These different values of BE for both oxygen and titanium are important for probing the chemical environment in which these elements are located as well as their oxidation states. Moreover, it can be a powerful strategy to evaluate corrosion processes on the surface of dental implants, which is mainly characterized by damage of the oxide layer (TiO₂) and changes in the oxidation states of Ti.

3. Results

3.1. Surface morphology

Fig. 1A and 1B shows the morphological characteristics of the freshly packaged, unused control implant (IC), which exhibited undamaged interfaces. Fig. 1C shows the 3D mapping of the rough surface of the implant at 1000× magnification. This color coded map provided a surface profile with height increasing from 0 μm (dark blue) to 54 μm (red). The important feature to note is the uniform distribution of colored pattern with respect to height which further corroborates the unaffected nature of the surface.

Fig. 1 Dental implant control (IC): (A) low magnification of the control implant; (B) undamaged and deformation-free smooth collar-rough junction of the control implant; (C) 3D mapping of the surface illustrating implant profile without evidence of deformation.

3.1.1 Effects of early stage colonizing bacteria. Fig. 2 shows optical microscopic images of implant I1, which was immersed in in vitro culture medium of S. mutans. After 60 days of immersion, the surface of the implant was observed to present significant changes in surface coloration, indicative of titanium oxidation states⁴⁹ (Fig. 2A). Higher magnification of the smooth collar-rough surface junction revealed purple and yellow discoloration (black arrows in Fig. 2B). The 3D map (Fig. 2C) presented significant variation in the height profile characterized by intermittent changes in the morphologies in the form of peaks (blue arrows) and valleys (black arrows). These peaks and valleys resembled a characteristic surface feature of an acidic electrochemical attack known as “pitting” seen in a previous study.³⁶ The depth of these pits ranged from 25–35 μm with a diameter of 15–20 μm.

Fig. 2 (A–C) Implant immersed in S. mutans (I1): (A) low magnification image of the implant with discoloration and rust; (B) purple discoloration (black arrows) on the surface of smooth collar-rough junction; (C) 3D mapping of the surface displaying variation in the height profile with peaks (blue arrows) and valleys (black arrows). (D–G) Healing abutments (HA): (D) non-implanted control (HAC); (E) 3D mapping of the surface displaying minimal differences in height profile of HAC; (F) discoloration and scratches (black arrows) on the surface of retrieved healing abutment (HA2); (G) 3D mapping of the surface displaying minimal differences in height profiles of retrieved healing abutment HA2.

Microscopic analysis of the surface of healing abutments did not show significant surface deformation or corrosion features when compared with the control abutment (Fig. 2D). Fig. 2F illustrates the features present on the surface of the retrieved healing abutment HA2. There were discolorations and scratches (black arrows) observed on the surface of HA2 (Fig. 2F). However, since the abutments were color coded, it was difficult to visualize discoloration with optical microscopy. Also, the height profile (Fig. 2G) of scratches observed in HA2 did not show any significant change in the surface morphology in comparison to the control (Fig. 2E).

3.1.2 Effects of late stage colonizing bacteria. Microscopic images of implants immersed in P. gingivalis culture for a month are shown in Fig. 3. In comparison to the control, there was no indication of deformation or oxidation on the surface of these implants with the exception of implant I5. Implant I5 (Fig. 3A) displayed mild purple and yellow discoloration on the surface. 3D imaging (Fig. 3B) did not reveal the presence of peaks and valleys for this group of implants.

Fig. 3 (A and B) Implants immersed in P. gingivalis (I2–I5): (A) yellow discoloration (black arrows) on the surface of I5; (B) 3D mapping of the surface of the implant I5. (C–E) Dental implant (I6) extracted from a patient due to peri-implantitis: (C) whole view of the fractured implant (arrows represent the fractured area); (D) purple-yellow discoloration (black arrows) on the surface; (E) 3D mapping at 1000× magnification of the surface.

Surface analysis of the retrieved dental implant I6 from a patient with peri-implantitis revealed severe corrosion with the characteristic purple-yellow discoloration observed in several regions of the rough implant surface (Fig. 3D). Fig. 3C shows the bottom rough part of the failed implant while Fig. 3E shows a 3D mapping of the surface with no indication of roughness variation (absence of peaks and valleys).

3.2. Elemental composition

Results obtained from the control implant (IC) as received showed the presence of titanium (Ti2p), oxygen (O1s) and carbon (C1s) on the surface. Ti2p and O1s are related with surface chemistry while carbon is related to environmental contamination.³⁶ Two different peaks were observed for O1s at 533.31 eV and 531.74 eV, which are related to the oxygen in the oxide layer and in its hydroxyl form (OH), respectively. On the Ti spectrum, two peaks were observed at 459.84 eV and 465.74 eV, which are characteristic of Ti in its native oxide layer Ti⁴⁺ (TiO₂) form.^48,50,51 After sputtering the surface, two different states were observed for Ti at 454.56 eV and 456.05 eV, which can be assigned to Ti²⁺ and Ti³⁺, both being associated with oxidized states.⁵⁰ This result shows that even after sputtering, the oxide layer could be detected on the surface of the control implant by XPS (Fig. 4a).

Fig. 4 Ti2p spectra of (a) the implant control (IC) before and after sputtering; (b) comparison of Ti2p spectra of IC (sputtered) and implant I1 (as received and sputtered). Note the absence of Ti⁰ oxidation state on the surface of IC which indicates a thick oxide layer structure and presence of the Ti⁰ peak after sputtering the surface of I1.

3.2.1 Effects of early stage colonizing bacteria. Specimen I1, which was immersed in vitro in medium containing S. mutans, had two different areas analyzed on its surface. The analysis of the first area as received showed high carbon and oxygen concentrations. Both elements can be associated with environmental contamination and bacterial product.⁴ After sputtering the first area, iron was detected, as well as other elements (Cr and Mn) in trace amounts. No Ti was detected in the area analyzed before and after sputtering, which indicates that a thick layer of contaminants was present on the implant. A second area on I1 was evaluated in order to further investigate the corrosion features observed by optical microscopy (Fig. 2). In contrast to the results obtained from the analysis of the first area on the sample as received, titanium as Ti⁴⁺ (TiO₂) was detected as well as carbon, oxygen and calcium. The oxygen content was mostly related with its hydroxyl form.⁴⁸ Corroborating with this result, Ti⁴⁺ was detected only in trace amounts (Fig. 4b). After sputtering, titanium was detected in both its Ti²⁺ (TiO) as well as in Ti⁰ (metallic) states.

The XPS results obtained for all healing abutments (HA) were, in general, similar with the spectrum obtained for the control abutment (HAC). Two different areas of each HA were scanned. The first area was the screw, which is in direct contact with the implant placed in maxillar/mandibular bone. The second area was the upper section, which is in contact with soft tissues. Each of these areas was analyzed as received and after sputtering. On the screw area, titanium was observed only in its Ti⁴⁺ state before and after sputtering in all samples, with the exception of specimen HA3 where this element was not detected (Fig. 5). The other elements detected were oxygen, carbon, nitrogen and calcium. The BEs found for O1s (as received) are associated both with hydroxyl groups and TiO₂. However, after sputtering, only O1s related to TiO₂ was observed. The elements detected in the HA upper body as received were oxygen, titanium, nitrogen, calcium, carbon and silicon. The BE related to titanium was associated with Ti⁴⁺ (TiO₂). The oxygen spectra are in agreement with this result, in which both oxygen in its hydroxyl and TiO₂ forms are observed. Moreover, the amount of titanium detected in HA3 was lower than those observed on other samples, which indicates a thicker layer of organic contaminants from the oral environment (Fig. 5a). After sputtering, the BE related to Ti⁴⁺, Ti³⁺, Ti²⁺ and Ti⁰ (metallic titanium) was observed in all samples (Fig. 5b), showing that the oxide layer in the upper section of the HAs was thinner.

Fig. 5 Ti2p spectra (upper area) of healing abutments HAC and HA1–3: (a) as received and (b) after sputtering. Note that the same elemental composition is observed for the control and retrieved healing abutments.

3.2.2 Effects of late stage colonizing bacteria. Implants (I2–I5) were immersed in medium containing P. gingivalis for 30 days. In the first set of analyses (as received) of these implants, titanium and oxygen were observed in Ti⁴⁺ (TiO₂) state. In addition, the XPS spectrum of I5 showed interesting features because Ti²⁺ (TiO) and Ti⁰ (metallic) oxidation states were also detected before sputtering (Fig. 6a). After sputtering, Ti³⁺ (Ti₂O₃), Ti²⁺ (TiO) and Ti⁰ (metallic) states were detected in I2–I5 (Fig. 6b).

Fig. 6 Ti2p spectra of (a) I5 before and after sputtering; (b) implant control (IC) and I2–I5 after sputtering; (c) IC and I6 after sputtering.

The elements observed for the retrieved implant (I6) as received, were oxygen, carbon and titanium. The O1s spectrum consisted of two peaks that have binding energies associated both with hydroxyl form and also with TiO₂. Titanium was observed in trace quantity and associated with its Ti⁴⁺ (TiO₂) form. This result was expected since the implant was maintained in the oral environment. After sputtering the surface, both Ti²⁺ and Ti⁰ were detected (Fig. 6c).

4. Discussion

This study evaluated the morphology and composition of dental implants and healing abutment surfaces exposed to bacteria during in vivo and in vitro conditions. Two sets of experimental conditions were developed: (1) analysis of the effect of early colonizers using implants that were immersed in vitro in a culture of S. mutans simulating the initial healing period and analysis of the surface of retrieved (in vivo) regular connection HAs. In general, dental implants are surgically screwed inside the maxillar/mandibular bone either as bone-level or tissue-level implants. In addition to the rough surface for hard tissue integration, tissue-level implants have a smooth collar that interacts with soft tissues and facilitates proper tissue growth and geometry for prosthetic attachment. HAs serve as closure caps of implants during the initial healing period (3–6 months) when implants are not exposed to mechanical occlusal loading. The HAs used in this study secure bone-level implants and play a vital role in interacting and arranging soft tissues. They occupy the supra- and sub-gingival soft tissue levels and hence are highly exposed to early bacterial colonizers. Therefore, HAs can serve as a model to understand the interaction of these bacterial species with the surface of titanium. (2) The second set of experimental conditions involved analysis of the effects of late bacterial colonizers using implants that were immersed in vitro in a culture of P. gingivalis and analysis of the surface of a retrieved implant (in vivo) that was removed due to peri-implantitis. Evaluation of in vitro bacteria-immersed samples and retrieved HAs allowed for understanding of the effects of bacterial load in the absence of mechanical forces while analyzing the surface of the retrieved implant provided information about the synergistic effects of bacteria and mechanical loads during the later stages of implantation.

Topographical analysis of the control implant specimen surface showed no evidence of deformation, as expected, in any of the rough and smooth collar areas of the implant (Fig. 1). The Ti spectrum of the sample as received showed a characteristic peak for Ti⁴⁺, which represents the native oxide layer (TiO₂) that is spontaneously formed on the surface of titanium when in contact with air.^52,53 Even after sputtering, the oxide layer could be detected on the surface of the control implant by XPS, despite having different oxidation states (Ti²⁺ and Ti³⁺) as shown in Fig. 4. This indicates that the oxide layer on the surface of the control implant was thick and stable, which can be also inferred from the absence of the metallic bulk state Ti⁰ peak. Previous depth profile studies have shown that the oxide layer may comprise of all oxidation species (Ti⁴⁺, Ti³⁺, Ti²⁺). Ti⁴⁺ is increased at the surface relative to the sub-oxides layers, while the concentration of Ti³⁺ and Ti²⁺ is enhanced closer to the metal-oxide interface.^52,53 According to Hashimoto and Tanaka,⁵⁴ after sputtering, Ti and O atoms are not able to recombine in the same coordination and concentration of TiO₂, and as a result, new oxidation forms can be detected in the XPS spectrum. Thus, it is reasonable to expect that titanium surfaces covered by a stable and thick oxide layer will exhibit peaks for these oxidation states (Ti⁴⁺ → Ti²⁺) upon sputtering due to removal of surface layers. In contrast, a surface with damaged or depleted oxide layers will exhibit the Ti⁰ peak, indicating bulk exposure upon removal of surface layers.

Microscopic analysis of the surface of the implant immersed in bacterial culture of S. mutans (I1, Fig. 2) showed significant morphological changes in relation to the control (Fig. 1). I1 displayed features that suggested an electrochemical attack. Areas of the smooth collar-rough junction interfaces of the implant showed the characteristic purplish and yellowish discoloration associated with the oxidation states of Ti³⁺ and Ti²⁺, respectively.⁴⁹ Pitting attack was also observed, indicating a mechanism of metal dissolution induced by breakdown of the oxide layer and active corrosion. The direct contact of the implant with the bacterial medium provided a worst case scenario for the surface, simulating a situation of high bacteria concentration such as in the presence of a biofilm. In the first area of I1 analyzed, high levels of carbon and oxygen were detected which can be associated with environmental contamination and, in this particular case, attributed to bacterial products.⁴ After sputtering, iron and other elements in trace concentrations were detected with no peaks for titanium, which indicated a thick layer of contaminants. Iron is one of the essential micronutrients for S. mutans growth, which can be correlated with previous exposure of the implant surface to bacteria.⁵⁵ In the second area evaluated, Ti⁴⁺ (TiO₂) was detected among other elements. After sputtering, Ti²⁺ and Ti⁰ were detected, suggesting that the oxide layer was almost completely removed after sputtering. Compared to the control specimen (Fig. 4), it can be inferred that I1 had a thinner oxide layer. It is well documented that S. mutans produces an acidic environment due to the generation of organic acids during its sugar metabolism,⁵⁶ which can produce the effects observed in the morphological and spectroscopy results. These results are supported by a previous investigation showing that S. mutans affected the corrosion behavior of commercially pure titanium by reducing its open circuit potential (OCP) and corrosion potential (E_corr), thereby accelerating surface corrosion processes.³⁰

Microscopic analysis of the three retrieved HAs did not reveal any features that could indicate surface damage (Fig. 2). The color of the HAs surfaces (deep purple) indicated that the implants were subjected to an anodization process which increases the thickness of the oxide layer and corrosion resistance.⁵⁷ The XPS results obtained for all HAs were, in general, similar with the spectrum obtained for the control abutment (Fig. 5). On the screw area, titanium was observed in its Ti⁴⁺ state before and after sputtering in all samples with the exception of specimen HA3, which was covered in a thick layer of biological materials. This reveals that minimal corrosion occurred as the oxide layer of HAs was much thicker than that of the implants analyzed, considering the same sputtering treatment was performed on both. On the cap region, the BE was associated with Ti⁴⁺ (TiO₂). After sputtering, Ti⁴⁺, Ti³⁺, Ti²⁺ and Ti⁰ was observed in all samples, including the control, showing that the oxide layer in the upper section of HA was thinner than that of the screw. The distinct surface chemistries of these two regions can also be related to the different environments in which those areas were located. The screw is seated inside and in direct contact with the implant while the upper section is maintained in contact with soft tissue and is therefore exposed to higher loads of bacteria.

HAs are not subjected to the same mechanical loading conditions as observed in implant–abutment–crown systems.⁵⁸ Therefore, any damage that may be present on the surface of retrieved HAs must have been a result of chemical attack due to the presence of early colonizing bacteria. This situation would be similar to the conditions imposed in the in vitro implant immersion testing in S. mutans. Studies on HAs are particularly scarce, probably due to the low probability of usage of these devices and their short life span in vivo.⁵⁹ Comparing the surface composition and morphology of I1 and HAs, it was observed that the HA surfaces were significantly less damaged. This can be due to the fact that the implant immersed in vitro in a culture of S. mutans was exposed to a higher bacterial load, which was supplemented by nutrients that provided ideal conditions for bacterial growth. However, in vivo (as in the case of HAs), a lower potential for bacterial growth exists due to the natural immune system assisting in the host defenses against bacterial contamination.⁶⁰ Furthermore, it can be assumed that the anodization process contributed to forming thicker and stable oxide layers, as indicated by the intense Ti⁴⁺ peaks that may have assisted in producing surfaces more resistant to the acidic environment produced by early colonizing bacteria.

Morphological analysis of the implants immersed in P. gingivalis (I2–I5) culture did not reveal signs of surface oxidation or corrosion features such as pitting attack or changes in surface roughness on the smooth and rough implant interfaces with the exception of one sample (I5) (Fig. 3). I5 in particular displayed discoloration, which was milder in comparison to the attack observed for the implant immersed in a culture of S. mutans (I1, Fig. 2). The discoloration suggested the possibility of damage due to the release of lipopolysaccharides (LPS), which is one the main metabolic products of P. gingivalis. Previous studies have reported that LPS negatively affects the stability of the oxide layer and corrosion behavior of Ti alloys.^61,62 Color formation on titanium has been reported to be a result of shifts in surface potential in the presence of acidic environments.⁵⁷ Interestingly, Ti⁴⁺ as well as Ti²⁺ and Ti⁰ were detected on I5 as received (Fig. 6a). This suggests that the oxide layer of I5 was possibly thinner than that of I2–I4 which only displayed Ti⁴⁺ before sputtering, corroborating with the corrosion features observed. The presence of Ti²⁺ and Ti⁰ after sputtering samples I2–I5 demonstrates that, as observed for I1, the oxide layer was thinner in comparison to the control implant. The morphological and XPS results indicate that immersion in a culture of P. gingivalis was less damaging to the implant surface in comparison to the effects of immersion in S. mutans. The difference in immersion time for the two group of specimens is not considered as a factor for the differences observed since we have demonstrated in a previous study that implant damage occurred as early as 3 days when immersed in S. mutans.³³ Similarly, a previous study reported that implants immersed in bacterial broth containing a mixture of early and late colonizers exhibited signs of pitting corrosion and surface damage after one month of exposure to in vitro culture.⁶³

Surface analysis of the retrieved implant from a patient with peri-implantitis (I6) revealed features that indicated severe corrosion and mechanical failure. The surface features and characteristic titanium discoloration were similar to those features observed for implants I1 and I2–I5, which were immersed in cultures of S. mutans and P. gingivalis, respectively. The 3D surface profile of I6 indicated smoother surfaces, possibly an indication that the surface layers were removed as a result of delamination caused by corrosion and wear processes. The XPS results for this implant were similar to those obtained for the implants immersed in S. mutans and P. gingivalis, with peaks detected for Ti⁴⁺ as received and Ti²⁺ and Ti⁰ after sputtering, indicating a thinner oxide layer structure in comparison to the control specimen.

The results from the morphological and spectroscopy studies revealed that oxide layer damage and corrosion can be induced by early and late stage colonizers on the surface of dental implants, even in the absence of mechanical loading. This is due to the creation of unfavorable acidic environments, which result in surface potential shifts that facilitate metal ion dissolution. This was particularly pronounced when the implant was exposed to a culture of S. mutans. Results from the in vivo retrieved implant illustrated synergistic effects of bacteria and occlusal loads on the surface of the implant with discoloration and delamination features, which were corroborated by XPS results demonstrating thinner oxide layer structure. The oxidation state of titanium may interfere with tissue integration, and the results of this study clearly illustrate the changes in oxidation states induced by bacteria in the presence and absence of mechanical loading.

Loss of passivity and corrosion of dental implants is a concerning issue. Metallic ions and particles resulting from these processes can deposit in soft tissues and induce macrophage accumulation at the implant site, which causes destruction of bone tissues and loss of integration.⁴ Features observed in this study, such as surface pitting and delamination, indicate processes of metal ion and wear generation. According to Mouhyi et al., corrosion can therefore be considered much more than a triggering factor and instead is a phenomenon underlying osseointegration and peri-implantitis.⁴

In summary this study showed that bacteria can create unfavorable electrochemical conditions that can disrupt the stability of the surface oxide layer. One of the limitations of this study was the use of monoculture in the in vitro studies instead of poly-microbial culture, which better resembles clinical conditions experienced by dental implants in vivo. However, understanding of the effects of individual bacterial strains on the implant surface is also important to provide information on the surface features induced by different species and their respective metabolites. In this study, S. mutans created an acidic electrochemical environment due to lactic acid metabolite whereas LPS releasing P. gingivalis resulted in a basic immersion medium. It is important to note that deleterious effects on the surface oxide layer were witnessed in both scenarios. Another limitation of this study was the lack of including electrochemical techniques as a direct means to characterize and quantify corrosion. Although other factors can play a major role in inducing failure, it is important to understand the individual contribution of each of these events. Therefore, future studies will thoroughly evaluate the role of a series of early and late colonizing bacterial strains in changing the morphology and surface composition of different implant materials under conditions of immersion and mechanical loading.

5. Conclusion

This study investigated morphological and compositional changes of the titanium surface, focusing on bacteria as one of the main etiological factors. Microscopic observations revealed deformations on dental implants surfaces exposed to both early and late stage colonizers with the exception of healing abutments showing no signs of physical surface damage. XPS was shown to be a powerful tool to detect and distinguish changes in surface chemistry. Combining the morphological features observed using optical microscopy with the detection of Ti in lower oxidation states, we were able to identify areas where corrosion took place. Reduction in the oxide layer thickness was demonstrated in dental implants and healing abutments in comparison to their respective controls. However, this reduction in the relative oxide layer thickness was much more pronounced for a clinically failed implant retrieval and implant exposed in vitro to S. mutans than for the implants immersed in vitro in P. gingivalis culture. Similarly, the cap region of HA retrievals, which were in direct contact with the oral cavity, exhibited a relatively thinner oxide layer in comparison to the screw region, which connects to the interior cavity of the implant. As the surface oxide layer ensures biocompatibility and corrosion resistance, changes in its morphology and composition can affect integration with soft and hard tissues. These results clearly indicated that bacteria can negatively impact the surface of dental implants and healing abutments.

Acknowledgements

The authors would like to acknowledge Alessandro Bizzarri, Damiano Fortuna and Emiliano di Cicco for the support with some of the logistics of the research, specifically with dental implants cultured with P. gingivalis. The authors thank the University of Texas at Dallas for providing startup funds for this research (D. Rodrigues) and the fellowship from Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) (I. M. G.).

References

1. C. N. Elias, in Implant Dentistry – A Rapidly Evolving Practice, ed. Prof. I. Turkyilmaz, Rijeka, 1st edn, 2011, pp. 319–365.
2. B. Zhao, H. C. van der Mei, G. Subbiahdoss, J. de Vries, M. Rustema-Abbing, R. Kuijer, H. J. Busscher and Y. Ren, Dent. Mater., 2014, 30, 716–727.
3. E. Eisenbarth, J. Meyle, W. Nachtigall and J. Breme, Biomaterials, 1996, 17, 1399–1403.
4. J. Mouhyi, D. M. Dohan Ehrenfest and T. Albrektsson, Clin. Implant Dent. Relat. Res., 2012, 14, 170–183.
5. Y.-R. V. Shih, Y. Hwang, A. Phadke, H. Kang, N. S. Hwang, E. J. Caro, S. Nguyen, M. Siu, E. a. Theodorakis, N. C. Gianneschi, K. S. Vecchio, S. Chien, O. K. Lee and S. Varghese, Proc. Natl. Acad. Sci. U. S. A., 2014, 111, 990–995.
6. M. Herrero-Climent, P. Lázaro, J. Vicente Rios, S. Lluch, M. Marqués, J. Guillem-Martí and F. J. Gil, J. Mater. Sci.: Mater. Med., 2013, 24, 2047–2055.
7. S. Anil, P. S. Anand, H. Alghamdi, and J. A. Jansen, in Implant Dentistry – A Rapidly Evolving Practice, ed. I. Turkyilmaz, Rjeka, 1st edn, 2011, pp. 83–108.
8. H. Tsuchiya, T. Shirai, H. Nishida, H. Murakami, T. Kabata, N. Yamamoto, K. Watanabe and J. Nakase, J. Orthop. Sci., 2012, 17, 595–604.
9. S. Svensson, F. Suska, L. Emanuelsson, A. Palmquist, B. Norlindh, M. Trobos, H. Bäckros, L. Persson, G. Rydja, M. Ohrlander, B. Lyvén, J. Lausmaa and P. Thomsen, Nanomedicine, 2013, 9, 1048–1056.
10. M. Kazemzadeh-Narbat, S. Noordin, B. a. Masri, D. S. Garbuz, C. P. Duncan, R. E. W. Hancock and R. Wang, J. Biomed. Mater. Res., Part B, 2012, 100, 1344–1352.
11. A. Mellado-Valero, P. Buitrago-Vera, M.-F. Sola-Ruiz and J.-C. Ferrer-Garcia, Medicina Oral Patologia Oral y Cirugia Bucal, 2013, 18, e869–e876.
12. M. G. Manda, P. P. Psyllaki, D. N. Tsipas and P. T. Koidis, J. Biomed. Mater. Res., Part B, 2009, 89, 264–273.
13. S. M. Heckmann, J. J. Linke, F. Graef, C. Foitzik, M. G. Wichmann and H.-P. Weber, J. Dent. Res., 2006, 85, 711–716.
14. J. C. M. Souza, M. Henriques, R. Oliveira, W. Teughels, J.-P. Celis and L. A. Rocha, Biofouling, 2010, 26, 471–478.
15. T. P. Chaturvedi, Clin., Cosmet. Invest. Dent., 2013, 5, 57–61.
16. J. M. Salcetti, J. D. Moriarty, L. F. Cooper, F. W. Smith, J. G. Collins, S. S. Socransky and O. Steven, Int. J. Oral Maxillofac. Implants, 1997, 12, 32–42.
17. Y. Manor, S. Oubaid, O. Mardinger, G. Chaushu and J. Nissan, J. Oral Maxillofac. Surg., 2009, 67, 2649–2652.
18. M. C. L. G. Santos, Braz. J. Oral Sci., 2002, 1, 103–111.
19. M. Esposito, P. Thomsen, L. E. Ericson, L. Sennerby and U. Lekholm, Clin. Implant Dent. Relat. Res., 2000, 2, 18–32.
20. A. J. van Winkelhoff, R. J. Goené, C. Benschop and T. Folmer, Clin. Oral Implants Res., 2000, 11, 511–520.
21. I. S. Yeo, H. Y. Kim, K. S. Lim and J. S. Han, Int. J. Artif. Organs, 2012, 35, 762–772.
22. W. L. Chai, I. M. Brook, a. Palmquist, R. van Noort and K. Moharamzadeh, J. R. Soc., Interface, 2012, 9, 3528–3538.
23. K. Subramani and D. Wismeijer, Int. J. Oral Maxillofac. Implants, 2012, 27, 1043–1054.
24. M. Albertini, L. López-Cerero, M. G. O'Sullivan, C. F. Chereguini, S. Ballesta, V. Ríos, M. Herrero-Climent and P. Bullón, Clin. Oral Implants Res., 2014, 1–5.
25. A. Mombelli, N. Müller and N. Cionca, Clin. Oral Implants Res., 2012, 23(6), 67–76.
26. J.-H. Fu, Y.-T. Hsu and H.-L. Wang, Eur. J. Oral Implantol., 2012,(5), S91–S103.
27. D. G. Olmedo, D. R. Tasat, G. Duffó, M. B. Guglielmotti and R. L. Cabrini, Acta Odontol Latinoam., 2009, 22, 3–9.
28. T. G. Wilson, J. Periodontol., 2009, 80, 1388–1392.
29. G. a. Van Der Weijden, K. M. Van Bemmel and S. Renvert, J. Clin. Periodontol., 2005, 32, 506–511.
30. J.-C. Chang, Y. Oshida, R. L. Gregory, C. J. Andres, T. M. Barco and D. T. Brown, Biomed. Mater. Eng., 2003, 13, 281–295.
31. A. Pier-Francesco, R. J. Adams, M. G. J. Waters and D. W. Williams, Clin. Oral. Implan. Res., 2006, 17, 633–637.
32. V. A. R. Barão, C. J. Yoon, M. T. Mathew, J. Chia-Chun Yuan, C. D. Wu and C. Sukotjo, J. Periodontol., 2014, 85, 1–14.
33. S. Sridhar, T. G. Wilson, K. L. Palmer, P. Valderrama, M. T. Mathew, S. Prasad, M. Jacobs, I. M. Gindri and D. C. Rodrigues, Clin. Implant Dent. Relat. Res., 2015, 17, e562–e575.
34. T. G. Wilson Jr, P. Valderrama, M. Burbano, J. Blansett, R. Levine, H. Kessler and D. C. Rodrigues, J. Dent. Res., 2015, 86, 9–15.
35. J. R. Goldberg and J. L. Gilbert, J. Biomed. Mater. Res., Part B, 2003, 64, 78–93.
36. D. C. Rodrigues, P. Valderrama, T. Wilson, K. Palmer, A. Thomas, S. Sridhar, A. Adapalli, M. Burbano and C. Wadhwani, Materials, 2013, 6, 5258–5274.
37. D. C. Rodrigues, R. M. Urban, J. J. Jacobs and J. L. Gilbert, J. Biomed. Mater. Res., Part B, 2009, 88, 206–219.
38. M. Y. H. Chin, A. Sandham, J. de Vries, H. C. van der Mei and H. J. Busscher, Biomaterials, 2007, 28, 2032–2040.
39. G. Mabilleau, S. Bourdon, M. L. Joly-Guillou, R. Filmon, M. F. Baslé and D. Chappard, Acta Biomater., 2006, 2, 121–129.
40. M. Koike and H. Fujii, Biomaterials, 2001, 22, 2931–2936.
41. M. T. Mathew, S. Abbey, N. J. Hallab, D. J. Hall, C. Sukotjo and M. A. Wimmer, J. Biomed. Mater. Res., Part B, 2012, 100, 1662–1671.
42. V. a. Barão, M. T. Mathew, W. G. Assunção, J. C. Yuan, M. a. Wimmer and C. Sukotjo, J. Dent. Res., 2011, 90, 613–618.
43. A. B. Novaes, S. L. S. de Souza, R. R. M. de Barros, K. K. Y. Pereira, G. Iezzi and A. Piattelli, Braz. Dent. J., 2010, 21, 471–481.
44. L. Gremillard, S. Meille, J. Chevalier, J. Zhao, V. Fridrici, P. Kapsa, J. Geringer and J. Uribe, in Degradation of Implant Materials, ed. N. Eliaz, Springer US, New York, 1st edn, 2012, pp. 195–252.
45. T. P. Chaturvedi, Indian J. Dent. Res., 2009, 20, 91–98.
46. A. Siddiqi, A. G. T. Payne, R. K. De Silva and W. J. Duncan, Clin. Oral Implants Res., 2011, 22, 673–680.
47. K. M. B. Sarah Hiyari, Arch. Ophthalmol., 1950, 44, 146–154.
48. M. C. Biesinger, L. W. M. Lau, A. R. Gerson and R. S. C. Smart, Appl. Surf. Sci., 2010, 257, 887–898.
49. S. Matsuda and A. Kato, Appl. Catal., 1983, 8, 149–165.
50. J. T. Mayer, U. Diebold, T. E. Madey and E. Garfunkel, J. Electron Spectrosc. Relat. Phenom., 1995, 73, 1–11.
51. T. Godfroid, R. Gouttebaron, J. Dauchot, P. Leclère, R. Lazzaroni and M. Hecq, Thin Solid Films, 2003, 437, 57–62.
52. A. F. Carley, P. R. Chalker, J. C. Riviere and M. W. Roberts, J. Chem. Soc., Faraday Trans. 1, 1987, 83, 351.
53. G. Lu, S. L. Bernasek and J. Schwartz, Surf. Sci., 2000, 458, 80–90.
54. S. Hashimoto and A. Tanaka, Surf. Interface Anal., 2002, 34, 262–265.
55. G. Spatafora, M. Moore, S. Landgren, E. Stonehouse and S. Michalek, Microbiology, 2001, 147, 1599–1610.
56. W. F. Liljemark and C. Bloomquist, Crit. Rev. Oral Biol. Med., 1996, 7, 180–198.
57. A. Karambakhsh, A. Afshar, S. Ghahramani and P. Malekinejad, J. Mater. Eng. Perform., 2011, 20, 1690–1696.
58. B. Yilmaz, E. McGlumphy and I. Turkyilmaz, J. Oral Implantol., 2009, 35, 201–203.
59. S. D'Ercole, A. Piattelli, G. Marzo, A. Scarano and D. Tripodi, J. Res. Dent. Sci., 2013, 8, 109–114.
60. L. Evrard, Implant Dent. – A Rapidly Evol. Pract., 2011, pp. 531–544.
61. F. Yu, O. Addison, S. J. Baker and A. J. Davenport, Int. J. Oral Sci., 2015, 7, 179–186.
62. L. P. Faverani, W. G. Assunção, P. S. P. De Carvalho, J. C. C. Yuan, C. Sukotjo, M. T. Mathew and V. A. Barao, PLoS One, 2014, 9, e93377.
63. F. J. Gil, A. Rodriguez, E. Espinar, J. M. Llamas, E. Padullés and A. Juárez, Int. J. Oral Maxillofac. Implants, 2012, 27, 64–68.

---