Cytokine expression profile in the bone‐anchored hearing system: 12‐week results from a prospective randomized, controlled study

Abstract Objective To study the effect of implanting the percutaneous bone‐anchored hearing system (BAHS) itself and inflammation of the peri‐abutment skin warrant clarification. In this study, we aimed to acquire further insight into the immune responses related to BAHS surgery and peri‐implant skin inflammation. Materials and Methods During surgery and 12 weeks post‐implantation, skin biopsies were obtained. If applicable, additional biopsies were taken during cases of inflammation. The mRNA expression of IL‐1β, IL‐6, IL‐8, TNFα, IL‐17, IL‐10, TGF‐ß, MIP‐1α, MMP‐9, TIMP‐1, COL1α1, VEGF‐A, FGF‐2 TLR‐2, and TLR‐4 was quantified using qRT‐PCR. Results Thirty‐five patients agreed to the surgery and 12‐week biopsy. Twenty‐two patients had mRNA of sufficient quality for analysis. Ten were fitted with a BAHS using the minimally invasive Ponto surgery technique. Twelve were fitted with a BAHS using the linear incision technique with soft‐tissue preservation. Five biopsies were obtained during episodes of inflammation. The post‐implantation mRNA expression of IL‐1β (P = .002), IL‐8 (P = .003), MMP9 (P = .005), TIMP‐1 (P = .002), and COL1α1 (P < .001) was significantly up‐regulated. IL‐6 (P = .009) and FGF‐2 (P = .004) mRNA expression was significantly down‐regulated after implantation. Within patients, no difference between post‐implantation mRNA expression (at 12 weeks) and when inflammation was observed. Between patients, the expression of IL‐1β (P = .015) and IL‐17 (P = .02) was higher during cases of inflammation compared with patients who had no inflammation at 12‐week follow‐up. Conclusions As part of a randomized, prospective, clinical trial, the present study reports the molecular profile of selected cytokines in the soft tissue around BAHS. Within the limit of this study, the results showed that 12 weeks after BAHS implantation the gene expression of some inflammatory cytokines (IL‐8 and IL‐1β) is still relatively high compared with the baseline, steady‐state, expression. The up‐regulation of anabolic (COL1α1) and tissue‐remodeling (MMP‐9 and TIMP1) genes indicates an ongoing remodeling process after 12 weeks of implantation. The results suggest that IL‐1β, IL‐17, and TNF‐α may be interesting markers associated with inflammation.


| IN TR ODUC TION
The percutaneous bone-anchored hearing system (BAHS) has become an established treatment option for patients suffering from various types of hearing impairment. The BAHS consists of a titanium fixture, together with a pre-mounted skin-penetrating abutment, to which a sound processor can be attached. 1 The fixture is implanted in the retro-auricular temporal bone and relies on osseointegration for anchorage. 2 A percutaneous prosthesis penetrates the skin, thereby coming into direct contact with the outer environment, and it may elicit a variety of periprosthetic tissue responses, such as inflammation and infection. 3 In fact, adverse skin reactions, such as skin overgrowth, soft-tissue reactions, and infection, represent common complications. 4 Apart from discomfort and morbidity, recurrent episodes of inflammation can limit the use of the sound processor and in some cases lead to the extrusion of the fixture or voluntary implant removal. To assess the soft tissue surrounding the abutment, Holgers index is commonly used in clinical practice and it is frequently described as an endpoint analysis in trials, although questions regarding its validity warrant further study. [5][6][7] Materials implanted in hard or soft tissue stimulate different cell types. For instance, inflammatory cells produce cytokines, such as interleukin-1 (IL-1), IL-6, IL-10, and tumor necrosis factor-a (TNF-a), which are involved in regulating the immune response and wound healing. It has been hypothesized that several factors affect the biological response to implants; they include the surgical trauma, the shape and chemical characteristics of the material and the host tissue itself. 8,9 In addition, shear stress concentration in the mobile skin interfacing the rigid abutment may lead to micro-trauma and cell damage, resulting in a prolonged inflammatory state in the peri-abutment skin. 10 Several studies that correlated crevicular fluid cytokines to inflammatory conditions around dental implants, such as peri-implantitis, a destructive inflammatory process affecting the bone and soft tissues around osseointegrated dental implants, have been published. 15 In the last review, IL-1ß and TNF-a were identified as major cytokines in relation to peri-implantitis. In contrast to dental implants, our knowledge of the cytokine expression profile in the soft tissue and peri-abutment fluid surrounding extra-oral percutaneous implants is very limited.
Recently, Lennerås et al described the molecular profile associated with percutaneous femoral prostheses and correlated it to different microbiological and clinical parameters, including a modified Holgers index of the skin surrounding the abutment. Among several detected correlations, the expression of TNF-a correlated to the presence of S. aureus species, whereas the expression of MMP-8 correlated to polymicrobial detection. Nevertheless, no correlation was observed between the modified Holgers index and any of the analyzed genes. 16 In the field of BAHS, one study by Grant et al correlated the production of periabutment fluid exudate and its cytokine content to Holgers index, as an indicator of inflammation. 17 The Holgers index scores correlated to the fluid volume, as well as with the total amounts of IL-1ß and IL-8.
However, in the latter retrospective studies, the times of retrieval and analysis were relatively random in relation to implantation surgery.
The aims of this prospective clinical study were: (i) to evaluate the molecular profile of factors related to soft-tissue healing and inflammation at baseline and after 12 weeks of implantation in BAHS patients treated with either a linear incision or minimally invasive Ponto surgery (MIPS) approaches; (ii) to correlate the gene expression to clinical manifestations of soft-tissue complications, as judged by Holgers index, as well as to local and systemic factors that may affect the early outcomes of BAHS . The gene expression levels of IL-1b, IL-6, IL-8, TNF-a, IL-17, IL-10, transforming growth factor-beta (TGF-ß), MIP-1a, tissue metabolism (MMP-9, TIMP-1, collagen type 1 (COL1a1), vascular endothelial growth factor (VEGF)-A, basic fibroblast growth factor (FGF-2), Tolllike receptor (TLR)-2, and TLR-4 were determined in peri-abutment soft-tissue biopsies by quantitative real-time PCR (qRT-PCR).

| Population
This study is part of a multicenter randomized, controlled trial (RCT).
The study protocol has previously been published. 6 Patients were recruited at the out-patient ENT department at MUMC1. The inclusion criteria were at least 18 years of age and found to be eligible for unilateral BAHS surgery. The exclusion criteria were: (I) a history of immunosuppressive disease, (II) use of systemic immunosuppressive medication, (III) bilateral BAHS placement, (IV) relevant dermatological disease (eg, psoriasis, severe eczema), (V) participation in other studies, and (VI) when no suitable site for a 4-mm wide implant was found during surgery. In addition, patients had to agree to voluntary biopsies during surgery and the 12-week follow-up. All patients provided written informed consent.

| Procedures
Baseline characteristics including gender, age, body mass index, smoking habits, medical history, and medication were obtained in the case report forms. Patients received a Ponto Wide implant with a mounted abutment (Oticon Medical AB, Askim, Sweden). Prior to incision, skin thickness was measured to determine the appropriate abutment length according to the surgical manual. 19,20 Surgical techniques included the linear incision technique with soft-tissue preservation 6,21 or the MIPS technique. 6,22 An elaborate description is provided in the protocol publication. 6 Briefly, in the control group, the linear incision technique with soft tissue preservation was performed.
Here, a linear retro-auricular incision was made and a central area of periosteum is removed. The implant site was prepared using a guide drill and countersink drill. The BAHS was placed with 40-50 Ncm insertion torque setting. The incision was then closed with dermal sutures. The abutment was recovered using a 5-mm skin punch. 6,19,21 In the test group, the MIPS technique was performed. Skin and subcutaneous tissue was removed with a 5-mm punch. Remaining soft tissue was removed with a raspatorium. The MIPS-cannula was inserted and filled with saline. The implant site was prepared using the specifically designed guide drill and widening drill. After removal of the cannula, the BAHS was placed with 40-50 Ncm insertion torque setting. An installation indicator was used to aid in the estimation of complete insertion. 6,20,23 During surgery, bone quality was assessed by the surgeon while drilling as very soft, soft, medium, hard or very hard. Post-surgery, a healing cap with gauze drenched in ointment (Terra-cortril, Pfizer Laboratories, New York, USA) was placed on the abutment. Subjects were followed for regular follow-up visits at 9-days, 3 weeks, 12 weeks, and 1 year post-surgery. At 9 days post-surgery, the healing cap was removed. If necessary, local ointment (Terra-cortril) was used to    24 To quantify the mRNA expression levels, a qRT-PCR analysis was performed with a LightCycler480 (Roche) using a three-step PCR program. Relative gene expression levels were derived using the LinRegPCR (version 2016.1) method and normalized to the geometric average of two reference genes, cyclophylin A (CyloA) and beta-2-microglobulin (b2M). Non-detectable samples were imputed as half the lowest observed threshold. Fold changes were calculated using the delta-delta CT method.

| Statistical analysis
Statistical analyses were performed using R version 3.3.2 (R Foundation for Statistical Computing, Austria). Statistical significance was established at P .05. Gene mRNA expression is presented for all genes.
Due to the small sample size and non-normality of the data, nonparametric tests were executed. Differences between baseline and the 12-week follow-up were compared using Wilcoxon's signed rank test.
Differences between surgical techniques were compared at baseline   Table 1. Soft-tissue outcomes at follow-up are presented in Table 2.

| Post-implantation expression
The expression of mRNA for the selected cytokines at baseline and the 12-week follow-up is shown in Figure 1 and Supporting Information Figure S1. The results for mRNA gene expression at 12 weeks postimplantation compared with baseline and differences between surgical techniques are presented in Table 3. The variation in mRNA expression at baseline for the different cytokines was high.
The analysis revealed that, 12 weeks post-implantation, a significantly higher expression of the inflammatory markers, IL-1b and IL-8, was observed compared with baseline for both surgical techniques, whereas the IL-6 mRNA expression was significantly down-regulated.
Tissue metabolism markers, MMP9, TIMP-1, and COL1a1, were significantly up-regulated 12 weeks post-implantation. The vascular endothelial growth factor, VEGF, displayed a trend toward increased mRNA expression at 12 weeks. Post-implantation, the mRNA expression of basic fibroblast growth factor, FGF-2, was significantly decreased.
Compared with baseline, no significant differences were observed for the bacterial marker, TLR-2.

| Surgical technique
The comparative analysis between the two different surgical techniques revealed no significant difference in gene expression in any of the genes at 12 weeks (Table 3). In spite of this, trends toward a higher expression of IL-8 and IL-1b were found in the linear incision group compared with the MIPS group ( Figure 2; Table 3). No significant differences in gene expression were observed between the groups at baseline (Supporting Information Table S3).   up-regulated compared with patients without an episode of inflammation during follow-up (Table 4, between patients; Inflammation; Supporting Information Figure S2). For the inflammation samples at the episode of inflammation, TLR-2, which is related to bacterial recognition, and the inflammatory mediator, TNF-a, displayed a trend toward higher mRNA expression compared with samples from patients without inflammation. In contrast, in patients experiencing inflammation, no differences in mRNA expression were observed at the time of inflammation compared with the non-inflamed state at 12 weeks (Table 4, Within patients, Inflammation). Similarly, there were no differences in gene expression at 12 weeks post-implantation between the 15 patients who did not experience an episode of inflammation during follow-up compared with the seven patients who did (Table 4, Between patients; Post-implantation).

| Clinical parameters
Multiple exploratory correlations were performed between gene expression and the clinical parameters. Significant correlations with a correlation coefficient of 0.6 or higher are presented in Table 5.

| Summary
To our knowledge, this is the first prospective study to investigate the molecular profile related to inflammation after BAHS implantation. Skin

| BAHS implantation
Our results indicate that a continuous state of increased immune activation is present within the soft tissue surrounding the abutment, despite a lack of macroscopic signs of inflammation. This is in line with previous microscopic observations. 14

| Surgical technique
MIPS was designed on the assumption that reduced surgical trauma would lead to improved outcomes. 31

| Clinical parameters
Smoking, BMI and diabetes have all been postulated as factors for softtissue reactions or implant loss in BAHS and dental studies. [34][35][36] Recently, Sayardoust et al showed that smoking is associated with different gene profiles post-surgery for dental implants. 37 Here, we observed that, with intact skin, no gene expression differences were observed. However, when the skin integrity is impaired by the BAHS, smoking status increases the expression of genes related to inflammation and tissue remodeling indicating a possible biological relationship for this risk factor. Strong to moderate correlations were found for IL-8 and MMP-9 expression. BMI and diabetes might affect the gene expression as well. However, only moderate to weak correlations were found here (Supporting Information Table S5). We found no correlation between pain scores and gene expression for BAHS.

| Strengths and limitations
This is the first prospective study to evaluate cytokine expression within the soft tissue around the abutment in BAHS. The influence of both installing the device and inflammation was investigated within   The resulting sample size is limited, thereby restricting us to create models that include several factors or performing more elaborate factor analyses. Although, cytokine expression will vary from 12 weeks and onward we assumed that for the overall group expression is relatively stable in order to compare 12-week expression for subjects without inflammation to the cases of inflammation. In the comparative analysis, we would have preferred to execute all the qRT-PCR reactions in triplicate. However, the cDNA yield was too low to do this and several patients were excluded due to cDNA yields that were too low. We investigated mRNA expression, which may differ from protein expression. However, to limit this possibility, we used protein coding mRNA transcripts. The location of sampling was close to the BAHS, but it may have differed from patient to patient, due to manipulation and movements. At follow-up, several patients mentioned that they believed they had experienced more complaints such as crust formation or episodes of inflammation after the 12-week biopsy. By obtaining a biopsy in the area close to the abutment, a possible source of infection may have been created. In future studies, the use of the periostrip paper as a less invasive alternative for collecting the peri-abutment fluid should be considered. 17 A non-invasive method might increase willingness to provide samples, thereby increasing the sample size and enabling the collection of samples at several time points. In dental implants, bone quality is usually assessed using the classification introduced by Lekholm and Zarb. 38 This qualification includes a radiographic assessment combined with an tactile assessment performed during drilling.
Unfortunately, no such qualification exists for BAHS. Also, radiological assessment of the future implant location is rare. Therefore, in this study we used an explorative bone quality assessment scale based on tactile feedback only.

| Perspective
Implant-associated inflammation and infections represent a challenge in BAHS and other medical devices. 29 New submerged active and passive bone conduction devices that result in intact skin have been introduced in the last few years. 39,40 Even though they seem like good alternatives, they may provide too low amplification for some patients.
Moreover, these solutions require more invasive surgery, are probably more expensive and have relatively large imaging artifacts of at least 9-10 cm for MRI investigations. 41 Another approach would be to adjust the surface properties of the percutaneous abutment.
Abutments coated with hydroxyapatite that are believed to integrate with the skin have been introduced. 42 An alternative approach using extra-smooth abutment surfaces is currently being investigated (NCT02304692). Antibiotic-releasing, steroid-releasing and silvercoated abutments might be possible as well. It is hypothesized that the introduction of skin preservation techniques will help the immune system surrounding the abutment to remain intact, thereby improving immune responses. In addition, the use of punch-only techniques could reduce skin movements. Immune modulation, abutment properties, surgical damage and reduced skin movements may all contribute to the occurrence of inflammation. The combination of bacterial data, cytokine expression, skin movements, and long-term follow-up data may provide additional insights into implantation and inflammation. 6 Here, we found that implantation itself may already result in a state of continued inflammation.

| CON CL USI ONS
As part of a randomized, prospective, clinical trial, the present study reports the molecular profile of selected cytokines in the soft tissue around BAHS. Within the limit of this study, the results showed that 12 weeks after BAHS implantation the gene expression of some inflammatory cytokines (IL-8 and IL-1b) is still relatively high compared with the baseline, steady-state, expression. The up-regulation of anabolic (COL1a1) and tissue-remodeling (MMP-9 and TIMP1) genes indicates an ongoing remodeling process after 12 weeks of implantation. The results suggest that IL-1b, IL-17, and TNF-a may be interesting markers associated with inflammation.

SUPPORTING INFORMATION
Additional Supporting Information may be found online in the supporting information tab for this article.