Therapeutic Potential of Vasoactive Intestinal Peptide and its Derivative Stearyl-Norleucine-VIP in Inflammation-Induced Osteolysis.

The common use of dental and orthopedic implants calls for special attention to the immune response leading to peri-prosthetic bone loss and implant failure. In addition to the well-established microbial etiology for oral implant failure, wear debris and in particular titanium (Ti) particles (TiP) in the implant vicinity are an important trigger of inflammation and activation of bone resorption around oral and orthopedic implants, presenting an unmet medical need. Here, we employed bacterial-derived lipopolysaccharides (LPS) to model infection and TiP to model aseptic inflammation and osteolysis. We assessed inflammation in vitro by measuring IL1ß, IL6 and TNFα mRNA expression in primary macrophages, osteoclastogenesis in RANKL-induced bone marrow derived pre-osteoclasts and osteolysis in vivo in a mouse calvarial model. We also assessed the trans-epithelial penetrability and safety of the tested compound in rats. Our results show that a lipophilic super-active derivative of vasoactive intestinal peptide (VIP), namely stearyl-norleucine-VIP (SNV) presented superior anti-inflammatory and anti-osteoclastogenic effects compared to VIP in vitro. In the bacterial infection model (LPS), SNV significantly reduced IL1ß expression, while VIP increased IL6 expression. In the aseptic models of osteolysis, SNV showed greater suppression of in vitro osteoclastogenesis than VIP, and significantly inhibited inflammation-induced osteolysis in vivo. We also observed that expression levels of the VIP receptor VPAC-2, but not that of VPAC-1, dramatically decreased during osteoclast differentiation. Importantly, SNV previously shown to have an increased stability compared to VIP, showed here significant trans-epithelial penetration and a clean toxicological profile, presenting a novel drug candidate that could be applied topically to counter both aseptic and infection-related bone destruction.

Mechanism and Prevention of Titanium Particle-Induced Inflammation and Osteolysis.

The worldwide number of dental implants and orthopedic prostheses is steadily increasing. Orthopedic implant loosening, in the absence of infection, is mostly attributable to the generation of wear debris. Dental peri-implantitis is characterized by a multifactorial etiology and is the main cause of implant failure. It consists of a peri-implant inflammatory lesion that often results in loss of supporting bone. Disease management includes cleaning the surrounding flora by hand instruments, ultrasonic tips, lasers, or chemical agents. We recently published a paper indicating that US scaling of titanium (Ti) implants releases particles that provoke an inflammatory response and osteolysis. Here we show that a strong inflammatory response occurs; however, very few of the titanium particles are phagocytosed by the macrophages. We then measured a dramatic Ti particle-induced stimulation of IL1ß, IL6, and TNFα secretion by these macrophages using multiplex immunoassay. The particle-induced expression profile, examined by FACS, also indicated an M1 macrophage polarization. To assess how the secreted cytokines contributed to the paracrine exacerbation of the inflammatory response and to osteoclastogenesis, we treated macrophage/preosteoclast cultures with neutralizing antibodies against IL1ß, IL6, or TNFα. We found that anti-TNFα antibodies attenuated the overall expression of both the inflammatory cytokines and osteoclastogenesis. On the other hand, anti-IL1ß antibodies affected osteoclastogenesis but not the paracrine expression of inflammatory cytokines, whereas anti-IL6 antibodies did the opposite. We then tested these neutralizing antibodies in vivo using our mouse calvarial model of Ti particle-induced osteolysis and microCT analysis. Here, all neutralizing antibodies, administered by intraperitoneal injection, completely abrogated the particle-induced osteolysis. This suggests that blockage of paracrine inflammatory stimulation and osteoclastogenesis are similarly effective in preventing bone resorption induced by Ti particles. Blocking both the inflammation and osteoclastogenesis by anti-TNFα antibodies, incorporated locally into a slow-release membrane, also significantly prevented osteolysis. The osteolytic inflammatory response, fueled by ultrasonic scaling of Ti implants, results from an inflammatory positive feedback loop and osteoclastogenic stimulation. Our findings suggest that blocking IL1ß, IL6, and/or TNFα systemically or locally around titanium implants is a promising therapeutic approach for the clinical management of peri-implant bone loss.

Scaling of titanium implants entrains inflammation-induced osteolysis.

With millions of new dental and orthopedic implants inserted annually, periprosthetic osteolysis becomes a major concern. In dentistry, peri-implantitis management includes cleaning using ultrasonic scaling. We examined whether ultrasonic scaling releases titanium particles and induces inflammation and osteolysis. Titanium discs with machined, sandblasted/acid-etched and sandblasted surfaces were subjected to ultrasonic scaling and we physically and chemically characterized the released particles. These particles induced a severe inflammatory response in macrophages and stimulated osteoclastogenesis. The number of released particles and their chemical composition and nanotopography had a significant effect on the inflammatory response. Sandblasted surfaces released the highest number of particles with the greatest nanoroughness properties. Particles from sandblasted/acid-etched discs induced a milder inflammatory response than those from sandblasted discs but a stronger inflammatory response than those from machined discs. Titanium particles were then embedded in fibrin membranes placed on mouse calvariae for 5 weeks. Using micro-CT, we observed that particles from sandblasted discs induced more osteolysis than those from sandblasted/acid-etched discs. In summary, ultrasonic scaling of titanium implants releases particles in a surface type-dependent manner and may aggravate peri-implantitis. Future studies should assess whether surface roughening affects the extent of released wear particles and aseptic loosening of orthopedic implants.

Zinc enhances the phototoxic effect of blue light against malodour-producing bacteria in an experimental oral biofilm.

Oral malodour is thought to be caused mainly by the production of volatile sulfide compounds (VSCs) by anaerobic Gram-negative oral bacteria. Previous studies have shown that these bacteria are susceptible to blue light (400-500 nm wavelength). In the present study, we tested the effect of blue light in the presence of zinc, erythrosine B or both on malodour production in an experimental oral biofilm. Biofilms were exposed to a plasma-arc light source for 30, 60 and 120 s (equal to energy fluxes of 41, 82 and 164 J cm(-2), respectively) with or without the addition of zinc acetate, erythrosine B or both. After the light exposure, biofilm samples were examined for malodour production (by an odour judge) and VSC production (with a Halimeter), and VSC-producing bacteria were quantified using a microscopy-based sulfide assay (MSA) and in situ confocal laser scanning microscopy (CLSM). Results showed that exposing experimental oral biofilm to both blue light and zinc reduced malodour production, which coincided with a reduction in VSC-producing bacteria in the biofilm. These results suggest that zinc enhances the phototoxicity of blue light against malodour-producing bacteria.

The adhesion of oral bacteria to modified titanium surfaces: role of plasma proteins and electrostatic forces.

OBJECTIVES: Modifications of titanium (Ti) implant surfaces have a significant effect on early biofilm formation and the outcome of implant procedures. The aim of this study was to examine the role of plasma proteins and electrostatic forces in the adhesion mechanism of oral bacteria to modified Ti surfaces. MATERIALS AND METHODS: Ti discs with three different types of surface modifications, machined, acid-etched, and acid-etched and blasted, were examined for adhesion of oral bacteria: Streptococcus mutans, Porphyromonas gingivalis, and Fusobacterium nucleatum. Following pretreatment of the Ti with ion rich solutions or coating by human serum albumin or fibronectin, bacterial adhesion was examined by scanning electron microscopy and assessed quantitatively by DNA analysis. Ti coating by proteins as well as bacterial adhesion and their interrelationships were further investigated through confocal scanning laser microscopy. RESULTS: Acid-etched and blasted Ti surfaces exhibited significantly higher amounts of bacteria adhesion than the other two surfaces. Calcium was found to serve as a bridging agent in the adhesion process of S. mutans and F. nucleatum to Ti surfaces. Although albumin coating of the Ti reduced the adhesion of S. mutans to all surfaces, it had no influence on the adhesion of P. gingivalis or F. nucleatum. Coating the Ti with fibronectin enhanced P. gingivalis and F. nucleatum adhesion. CONCLUSIONS: Bacterial adhesion to Ti surfaces is roughness-dependent, and the adhesion mechanism is influenced by ions and proteins of the initial coating derived from the blood.

Wettability versus electrostatic forces in fibronectin and albumin adsorption to titanium surfaces.

OBJECTIVES: Although the enhancement of plasma protein adsorption to titanium ( Ti ) following wetting has been recognized, the relationship between wettability and electrostatic forces has remained unclear. Thus, we have carried out a series of studies to determine the role of wettability and electrostatic forces on protein adsorption. METHODS: Titanium disks with different surfaces were wetted with a range of solutions, two of which contained divalent positive ions ( Ca and Mg ). Unwetted disks served as a control. Subsequently, the wetted disks were subjected to three treatment regimes: (1) incubation in human serum albumin (HSA) or human serum fibronectin (HSF); (2) drying the wetted disks, followed by incubation in HSA or HSF; and (3) following protein adsorption, the Ca originating in the wetting solutions was removed by divalent positive ions chelator treatment (EGTA), and the remaining quantities were assessed. The quantity of the adsorbed proteins was determined by ELISA. RESULTS: It was found that in the case of HSA, adsorption was enhanced by the wettability, the presence of Ca and Mg in the wetting solution, and the existence of rough surfaces. For HSF, the wettability and rough surfaces enhanced adsorption. CONCLUSION: The results demonstrate that in addition to wettability, the composition of the wetting solution affects the protein adsorption. While wetting reduces the time for the HSA and HSF adsorption to reach saturation, the electrostatic forces enhance the amount of HSA adsorption. Thus, the protein adsorption capacity of titanium rough surfaces can be selectively manipulated by changing of the wetting solution.

Endosseous implant anchorage is critically dependent on mechanostructural determinants of peri-implant bone trabeculae.

Low bone mass is highly prevalent among patients receiving endosseous implants. In turn, the implantation prognosis in low-density skeletal sites is poor. However, little is known about the mechanostructural determinants of implant anchorage. Using metabolic manipulations that lead to low bone density and to its rescue, we show here that anchorage is critically dependent on the peri-implant bone (PIB). Titanium implants were inserted horizontally into the proximal tibial metaphysis of adult rats 6 weeks after orchiectomy (ORX) or sham ORX. Systemic intermittent administration of human parathyroid hormone (1-34) [iahPTH(1-34)] or vehicle commenced immediately thereafter for 6 weeks. The bone-implant apparatus was then subjected to image-guided failure assessment, which assesses biomechanical properties and microstructural deformation concomitantly. Anchorage failure occurred mainly in PIB trabeculae, 0.5 to 1.0 mm away from the implant. Mechanically, the anchorage performed poorly in ORX-induced low-density bone, attributable mainly to decreased trabecular number. iahPTH(1-34) rescued the PIB density and implant mechanical function by augmenting trabecular thickness (Tb.Th). However, implant biomechanical properties in low-density bone were relatively insensitive to implant surface treatment that affected only the osseointegration (%bone-implant contact). These results support a model wherein anchorage failure involves buckling of the weakest trabecular struts followed by sequential failure of the stronger trabeculae. Treatment with iahPTH(1-34) induced thicker struts, which were able to delay and even prevent failure of individual elements, thus implicating trabecular thickness as a prime target for enhancing implant anchorage by systemic bone anabolic therapy.

Degradation of collagen-guided tissue regeneration membranes by proteolytic enzymes of Porphyromonas gingivalis and its inhibition by antibacterial agents.

Previous studies have shown that whole cells of several periodontal pathogenic bacteria including Porphyromonas gingivalis may degrade the clinically used regeneration membranes Biomend Extend and Bio-Gide. Fractionation of P. gingivalis cells revealed that cell membrane-associated proteases are responsible for the in vitro degradation of the collagen membranes. In the present study, the specific role of extracellular vesicles and the purified Arg-gingipain enzyme of P. gingivalis in the degradation of three differently cross-linked collagen membranes (Ossix; Bio-Gide and Biomend Extend) was examined. In addition, the inhibitory effect of antibacterial agents and antibiotics used in local periodontal therapy on the enzymatic degradation was evaluated. The data presented show that while all tested collagen membranes, are prone to lysis by oral bacterial proteases, cross-linked membranes are more resistant to proteolysis. Furthermore, therapeutical concentrations of the antibacterial and antibiotic agents chlorhexidine, cetylpyridiniumchloride, minocycline and doxycycline were found to partially inhibit the enzymatic breakdown of the membranes, while metronidazole had no such effect. These results suggest that the presence of P. gingivalis cells, extracellular vesicles and enzymes in the vicinity of regeneration membranes in the periodontium, may change their physical structure and therefore alter their biological properties. Furthermore, the use of cross-linked collagen membranes and antibacterial agents may significantly inhibit this proteolytic process.

Trabecular bone gradient in rat long bone metaphyses: mathematical modeling and application to morphometric measurements and correction of implant positioning.

UNLABELLED: The distribution of trabecular structures in mammalian long bone metaphyses has been insufficiently explored. We show in rats that the trabecular bone structural parameters display a decreasing gradient, toward the diaphysis, that can be defined mathematically. This gradient is applicable for optimizing the reference volume in metabolic studies and for retrospective correction of implant positioning. INTRODUCTION: The mammalian metaphyseal trabecular bone is unevenly distributed. Hence, defining a standard reference volume is critical for morphometric analyses in metaphyseal sites. MATERIALS AND METHODS: The distal femoral and proximal tibial metaphyses of adult orchietomized (ORX) or sham-ORX rats were scanned by microCT 6 wk postoperatively. Morphometric analysis based on 3D image data was performed in 450-microm-thick transversal segments defined consecutively from the primary spongiosa toward the diaphysis. The results were subjected to curve-fit analysis. A similar approach was used for proximal tibial metaphyseal sites carrying titanium implants inserted horizontally 6 wk post-ORX and examined 2-12 wk after implantation. RESULTS: The respective curve-fit analysis in both femur and tibia revealed decreasing linear/quadratic and logarithmic gradients for all morphometric parameters in the sham-ORX animals. The ORX animals showed similar gradients with roughly similar slopes but lower values. For the bone volume (BV/TV) and connectivity (Conn.D) densities, the magnitude of the ORX effect vastly increased toward the diaphysis. The trabecular number was unaffected in ORX femora and tibias. The trabecular thickness showed a constant decrease in the femur and was unchanged in the tibia. These findings are useful for the determination and reporting of reference volumes in morphometric studies. Implementing the curve-fit analysis for retrospective correction of implant positioning revealed differences in BV/TV, Tb.N, Conn.D, and percent implant surface in contact with bone (%OI) between the sham-ORX and ORX rats. These differences were otherwise undisclosed. In addition, a temporal increase in %OI was shown only for the corrected measurements. CONCLUSIONS: We show the feasibility of modeling trabecular bone structures using mathematical tools. Such modeling may be used as an experimental tool. Moreover, if proven applicable to human skeletal structures, it may be further developed for the diagnosis of metabolic bone diseases and evaluation of therapeutic measures.

Adsorption of human plasma proteins to modified titanium surfaces.

OBJECTIVES: The aim of this study was to examine the effect of modified titanium (Ti) surfaces on the initial events of plasma proteins adsorption. MATERIALS AND METHODS: 'Ti disks' with three types of surface modifications were compared: machined, acid-etched and acid-etched and blasted. Physical and chemical characterizations of the surfaces were performed via scanning electron microscopy (SEM), atomic force microscopy (AFM) used for analysis of surface topography, characterization of the titanium oxide (TiO2) layer was carried out by X-ray photoelectron spectroscopy (XPS) and characterization of surface energy by the determination of contact angles. Evaluation of plasma proteins' adsorption to the treated Ti surfaces was performed by mass spectrometry, confocal laser scanning microscopy and XPS. Quantitative proteins' assessment was carried out by enzyme-linked immunosorbent assay. RESULTS: SEM images revealed major differences in the topography of the examined surfaces. Acid-etched and blasted Ti surfaces were found to have higher roughness values and a thicker TiO2 layer as compared with acid-etched and machined surfaces. Moreover, acid-etched and blasted surfaces showed high surface area differentiation, pointing to a high increase in the three-dimensional (3D) surface area over the 2D surface area compared with the other surfaces. Adsorption of plasma proteins to the acid-etched and blasted Ti surfaces was both qualitatively and quantitatively more intense compared with the machined and acid-etched surfaces. This was shown for each examined protein, total proteins and by the removal degree of the protein coat. CONCLUSIONS: The preferential adsorption of plasma proteins to the acid-etched and blasted Ti surfaces may be explained by its topographical characteristics and by the increase of the 3D surface area of this modified surface.

Parathyroid hormone 1-34 enhances titanium implant anchorage in low-density trabecular bone: a correlative micro-computed tomographic and biomechanical analysis.

The use of endosseous titanium implants is the standard of care in dentistry and orthopaedic surgery. Nevertheless, implantation in low-density bone has a poor prognosis and experimental studies show delayed implant anchorage following gonadectomy-induced bone loss. Intermittently administered human parathyroid hormone 1-34 [iahPTH(1-34)] is the leading bone anabolic therapy. Hence, this study assessed whether iahPTH(1-34) enhances titanium implant integration in low-density bone. Threaded titanium implants, 0.9 mm in diameter, were inserted horizontally into the proximal tibial metaphysis of 5-month-old rats, 7 weeks postorchiectomy (ORX). Subcutaneous administration of iahPTH(1-34), at 5, 25 and 75 microg/kg/day commenced immediately thereafter and lasted for 8 weeks. Quantitative micro-computed tomography (muCT) at the implantation site was carried out at 15 microm resolution using high energy and long integration time to minimize artifacts resulting from the high implant radiopacity. Osseointegration (OI) was calculated as percent implant surface in contact with bone (%OI) quantified as the ratio of "bone"-to-total voxels in contact with the implant. Additionally, the trabecular bone volume density (BV/TV), trabecular thickness (Tb.Th), trabecular number (Tb.N) and connectivity density (Conn.D) were measured in the peri-implant bone. All microCT parameters were stimulated by iahPTH(1-34) dose-dependently; the percent maximal enhancement was %OI = 143, BV/TV = 257, Tb.Th = 150, Tb.N = 140 and Conn.D = 193. The maximal values of %OI, BV/TV and Tb.Th in iahPTH(1-34)-treated ORX rats exceeded significantly those measured in the implantation site of untreated sham-ORX controls. The same specimens were then subjected to pullout biomechanical testing. The biomechanical parameters were also enhanced by iahPTH(1-34) dose-dependently, exceeding the values recorded in the sham-ORX controls. The percent iahPTH(1-34)-induced maximal enhancement was: ultimate force = 315, stiffness = 270 and toughness = 395. Except for the BV/TV and Tb.Th, there was no significant difference between the effect of the 25 and 75 microg/kg/day doses. There was a highly significant correlation between the morphometric and biomechanical parameters suggesting the use of quantitative CT as predictive of the implant mechanical properties. These findings demonstrate that iahPTH(1-34) effectively stimulates implant anchorage in low-density trabecular bone and thus the feasibility of administering iahPTH(1-34) to improve the clinical prognosis in low-density trabecular bone sites.

Flapless approach for removal of bone graft fixing screws and placement of dental implants using computerized navigation: a technique and case report.

This article presents a technique for the removal of the screws used to fix a bone graft and for the placement of dental implants in a flapless approach that utilizes the tracking technology of a computerized navigation system. A 24-year-old female patient injured in a terrorist bombing suffered from tooth loss and a bone defect in the maxilla. The area was grafted with bone from the chin in preparation for the placement of dental implants. Four months following the grafting procedure, the fixing screws were removed and the dental implants were placed in a flapless approach by the application of a specialized computerized navigation system. This technique emphasizes the potential of computerized navigation approaches in the facilitation of minimally invasive oral surgery.

Biological and clinical rationale for early implant loading.

This article questions whether early implant loading always leads to pseudointegration, as postulated by Branemark's original protocol, or whether the waiting period of 3 to 6 months can be significantly shortened in specific clinical situations and refined surgical protocols. The rationale for immediate loading of dental implants is discussed, and the evidence that immediate loading leads to implant fibrous encapsulation is reviewed. This evidence is opposed by conflicting indications that early loading induces bone growth. Micro- and macromovement are defined, and their role, as well as the importance of implant surface in the early loading protocol, is discussed. Recently published landmark human clinical trials on early implant loading also are reviewed, along with clinical, surgical, and prosthetic tips to take into consideration when planning an early implant loading case.

Intermittently administered parathyroid hormone 1-34 reverses bone loss and structural impairment in orchiectomized adult rats.

Male osteoporosis is emerging as a central theme in bone research. As in females, hypogonadism appears as a principal risk factor in men that leads to bone loss and increased fracture incidence. Intermittently administered parathyroid hormone (PTH) reverses bone loss in sex hormone-deprived women and female animals and increases bone mass in elderly men and normal male animals. This study was carried out to assess whether the PTH anabolic activity is also effective in adult castrated males and to gain insight into the underlying tissue processes. Bilateral orchiectomy (ORX) or sham-ORX was performed in 13-week old rats. Five weeks later, the ORX rats were treated intermittently with human PTH(1-34), 80 microg/kg/day or vehicle for 6 weeks. Femora were evaluated by quantitative micro-computed tomography followed by dynamic histomorphometry. The trabecular bone volume density showed 40% and 56% ORX-induced loss in the distal metaphysis at 6 weeks and 12 weeks post-ORX, respectively. PTH(1-34) induced supraphysiologic recovery of this bone loss (155% recovery) consequent to a vast increase in trabecular thickness (174% over sham-ORX controls) and a partial reversal (62%) of the decrease in trabecular number. As compared with the results in 12-week, orchiectomized vehicle-administered rats, the PTH(1-34) treatment induced a significant decrease in osteoclast number (20%) and twofold increase in bone formation rate. While ORX did not affect the femoral diaphysis, PTH(1-34) induced marked cortical thickening via the stimulation of endosteal mineral appositional rate (154% over ORX rats). These data portray PTH(1-34) as a highly potent bone anabolic agent in adult ORX rats, mainly by increasing both the trabecular and cortical thicknesses through its effect on osteoblasts and osteoclasts. The adult ORX rat is useful for investigating the processes involved in bone anabolic activity in castrated osteoporotic males and for the development of bone anabolic agents for treating this condition.

Retrospective clinical review of dental implants placed in a university training program.

Many dental schools offer implant clinical training at the pre- and postdoctoral levels, but little has been published on the clinical outcomes of implants placed in those programs. A post-entry chart review was conducted of all Branemark and Taper-Lock implants placed in a university clinic by faculty-student teams. Case information was gathered on data reporting forms and entered into a computer spreadsheet program. Survival rates were calculated as percentages. Cumulative implant survival was 96% for all 303 implants placed. Eight of the 12 implants that failed were placed by 2 operators, only 1 of whom was in the early stages of implant training. Branemark implant survival was 94.9% (n = 198) at 36 months; failures occurred between stage 2 and 3 months (n = 8) in mandibles and after 12 months (n = 2) in maxillae. Taper-Lock implant survival was 98.1% (n = 105) at 24 months; no mandibular failures occurred, but maxillary failures occurred before stage 2 (n = 1) and after 12 months of loading (n = 1). Taper-Lock implants exhibited a slightly higher (3.2%) cumulative survival rate at 24 months compared to Branemark implants. Differences in the numbers of implants placed (105 vs 198) and follow-up times (24 months vs 36 months) may have skewed the comparative results of Taper-Lock and Branemark implants, respectively, in this study. Implant survival for both systems was similar at 24 months of follow-up, and clinician experience did not appear to be an influencing variable on implant survival.

Enzymatic degradation of collagen-guided tissue regeneration membranes by periodontal bacteria.

Bacterial infection in the vicinity of guided tissue regeneration barrier membranes was shown to have a negative effect on the clinical outcomes of this increasingly used technique. Several oral and specifically periodontal bacteria were shown to adhere to such membranes in vivo and in vitro with a higher affinity to membranes constructed from collagen. The present study examined the role of periodontal bacteria and their enzymes in the degradation of commercially used collagen membranes. Degradation of two collagen membranes [Biomend (Calcitek, Colla-Tec Inc., Plainsboro, NJ) and Bio-Gide (Geistlich Biomaterials, Wolhousen, Switzerland)] labeled by fluorescein isothiocyanate was examined by measuring soluble fluorescence. Porphyromonas gingivalis, Treponema denticola and Actinobacillus actinomycetemcomitans and their enzymes were evaluated. Collagenase from Clostridium hystolyticum was used as a positive control. While whole cells of P. gingivalis were able to degrade both types of membranes, T. denticola could degrade Bio-Gide membranes only and A. actinomycetemcomitans whole cells could degrade none of the membranes. Fractionation of P. gingivalis cells revealed that cell membrane associated proteases were responsible for the degradation of the two collagen membranes. In T. denticola, the purified major phenylalanine protease was found to be responsible for the degradation of Bio-Gide membranes. These results suggest that proteolytic bacterial enzymes may take part in the degradation of collagen barrier membranes used for guided tissue regeneration.