Biomarkers in External Apical Root Resorption: An Evidence-based Scoping Review in Biofluids

Background External apical root resorption (EARR), an unwanted sequela of orthodontic treatment, is difficult to diagnose radiographically. Hence, the current scoping review was planned to generate critical evidence related to biomarkers in oral fluids, i.e. gingival crevicular fluid (GCF), saliva, and blood, of patients showing root resorption, compared to no-resorption or physiologic resorption. Methods A literature search was conducted in major databases along with a manual search of relevant articles in the library, and further search from references of the related articles in March 2021. The initial search was subjected to strict inclusion and exclusion criteria according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines. Results Following PRISMA guidelines, 20 studies were included in the final review. The studies included human clinical trials and cross-sectional and prospective studies with/without control groups with no date/language restriction. Various biomarkers identified in EARR included dentinal proteins, enzymes, cytokines, and salivary proteins. Severe resorption had higher dentin sialoprotein (DSP) and resorption protein concentrations as well as lower granulocyte-macrophage colony-stimulating factor (GM-CSF) as compared with mild resorption. Increased DSP and dentin phosphophoryn (DPP) expression was found in physiologic resorption. Compared to controls, resorbed teeth showed a higher receptor activator of nuclear factor kappa B ligand/osteoprotegerin (RANKL/OPG) ratio. In contrast, levels of anti-resorptive mediators (IL-1RA, IL-4) was significantly decreased. Differences in force levels (150 g and 100 g) showed no difference in resorption, but a significant rise in biomarkers (aspartate transaminase [AST] and alkaline phosphatase [ALP]) for 150 g force. Moderate to severe resorption in young patients showed a rise in specific salivary proteins, requiring further validation. Limitations of the studies were heterogeneity in study design, biomarker collection, sample selection, and confounding inflammatory conditions. Conclusions Various biomarkers in biofluids indicate active resorption, while resorption severity was associated with DSP and GM-CSF in GCF, and a few salivary proteins. However, a robust study design in the future is mandated.


INTRODUCTION
External apical root resorption (EARR) may occur in a variety of conditions like bacterial invasions, trauma, neoplasms, and systemic or pressure conditions produced by the application of orthodontic forces. 1 External apical root resorption is also known as orthodontically induced root resorption (OIRR), which is an undesirable but common sequela of orthodontic tooth movement. 2,3 External apical root resorption has a multifactorial etiology and is associated with several risk factors predisposing patients to various degrees of root resorption. 2,3 The reported EARR incidence is variable: 90% in histological studies, 73% in radiological studies after tooth movement, 6%-13% depending on the type of teeth, and 1%-5% or 1%-2% depending on resorption severity. 3,4 Nevertheless, any grade of EARR severity is known to limit the outcome of successful orthodontic treatment and also cause oral dysfunction on progression. 5 The deleterious effects of EARR on tooth movement mandate early detection of resorption. However, early detection is not possible with the currently available diagnostic modalities that rely on twoor three-dimensional radiographs. 3,6 Radiographs are associated with limitations such as radiation exposure, inability to outline the active resorption process, and limited view and standardization of the resorption process. 3,5,7 Hence, there is a great need for non-invasive techniques or determination of biomarkers to detect root resorption early in susceptible patients. 5 To define the biomarkers in root resorption, a thorough understanding is needed of its pathophysiology in relation to the surrounding bone and the periodontal ligament housing different types of cells, matrix, and biological messengers, 8,9 as explained in Figure 1. Although the biomarkers released in the paracrine environment in the gingival crevicular fluid (GCF) have been extensively studied in bone resorption during orthodontic tooth movement, 10,11 a comprehensive study of all body fluid biomarkers (GCF, saliva, and blood) around teeth undergoing resorption is lacking. Various mediums have been evaluated for biomarker collection, of which GCF has the advantages of ease of repeatability, collection, and detection of early resorption. 5 Also, saliva has greater accessibility and ease, but is comparatively less specific to the underlying periodontal condition. 12 Various biomarkers are indicative of active resorption, with evidence supporting the presence of dentinal proteins including dentin sialophosphoprotein (DSPP), dentin sialoprotein (DSP), and dentin phosphophoryn (DPP) in GCF and saliva. 5,7,13 Of these, DSPP shows a continuous expression in amelogenesis and dentinogenesis 14 and is considered a potent resorption marker. Other markers responsible for osteoclastogenesis or extracellular matrix degradation, including pro-inflammatory cytokines (interleukins [IL], tumor necrosis factor, etc.) or matrix metalloproteinases (MMPs), have also been associated with the degree of resorption. 9,15,16 Alkaline phosphatase (ALP), an enzyme associated with early deposition of minerals and tissue calcification, may contribute toward pulpal repair and healing after traumatic insults or injury and shows variable expression in root resorption. 15 Hence, there are multiple mediators having distinct associations in resorption, some in tissue destruction and others in tissue repair, which show variable expression at different stages of resorption. The success of clinical orthodontic treatment in turn The bone/root resorption process is similar at cellular levels for both osteoclasts and odontoclasts. 8 Cellular differentiation from mononucleated (progenitor osteoclasts) to multinucleated (mature osteoclasts) involves proresorption and anti-resorption. The M-CSF is responsible for proliferation and differentiation of progenitor cells; proinflammatory cytokines are directly responsible for differentiation and activation of pre-osteoclasts, or activation of RANKL, which in turn leads to pro-resorptive gene expression (TRAP, etc.). Anti-resorptive OPG (decoy receptor for RANKL) and IL-1RA (receptor antagonist to IL-1) govern the resorptive activity, with the RANKL/OPG ratio being the primary governing factor. IL-1, interleukin-1β; IL-1RA, interleukin-1 receptor antagonist; IL-6, interleukin 6; M-CSF, macrophage colonystimulating factor; OPG, osteoprotegerin; PDL, periodontal ligament; RANKL, receptor activator of nuclear factor kappa-Β ligand; TNF-α, tumor necrosis factor-α; TRAP, tartrate-resistant acid phosphatase.
is dependent on early detection of EARR and on preventing and limiting the extent of this unwanted condition. Tarallo et al. have provided some evidence related to the role of GCF biomarkers in root resorption, but failed to establish an all-inclusive understanding of the dynamics of root resorption markers to identify the most potent biomarker that might show significant association in multiple oral biofluids. 17 Another study by Allen et al. examined salivary protein in orthodontic tooth movement, but it did not specifically target root resorption. 18 Hence, this scoping review addresses the gap in the literature to generate critical evidence related to biomarkers in all oral fluids (gingival crevicular fluid [GCF], saliva, blood) of patients showing root resorption, compared to no resorption or physiologic resorption.

Protocol
A scoping review was designed according to the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) guidelines specific to scoping reviews (PRISMA-ScR). The inclusion and exclusion criteria were defined (Table 1), and the study was registered in the Open Science Framework (https://osf.io/nep9z/). No funding was received for the study.

Eligibility
The research topic for determining literature eligibility was developed based on the PICOS model, as follows: Population, patients showing EARR on radiographs; Intervention, orthodontic forces; Comparison, no resorption or physiological resorption; Outcomes, change in biomarkers in biofluids. There was no limitation of date or language placed on the literature search.
Based on the above, the research question asked: was the variation in levels of biomarkers in the oral fluids associated with root resorption in patients undergoing orthodontic treatment in comparison to no resorption or physiologic resorption?

Information Sources and Search
In March 2021, a thorough literature search was conducted in the major databases: PubMed, Web of Science, J-Gate, Directory of Open Access Journals, Scopus, and Embase, along with related searches, manual searches, and tracking of references from the manual searches. Both MeSH and free-text terms were used to search most of the databases: "biomarkers," "root resorption," and "orthodontics" with the BOOLEAN terminology "AND." Duplicate results were removed.

Study Selection
The identification, screening, eligibility, and inclusion of studies were performed as detailed in the PRISMA flowchart shown in Figure 2. The search strategy was applied independently by two reviewers (PK and AC) strictly based on the inclusion and exclusion criteria (Table 1).
Any discordance was addressed by two reviewers (DKB and DB) for a final consensus. Duplicates were removed, and articles were screened based on their titles and abstracts. Full texts were then retrieved, and an in-depth review was performed to identify the final studies selected for this review. No quality assessment was done as it is not mandatory for scoping reviews, and the aim of the current scoping review was to present a broad scope of biomarkers identified to date in EARR. Studies related to root resorption by other causes, including traumatic forces or endodontic resorption, were excluded from the final selection. The primary outcome included the variation in expression of different biomarkers  In vitro studies  Animal studies  Studies on biomarkers but not on resorption  Studies on resorption but not on biomarkers  Case reports and reviews/opinions in root resorption, which was further correlated with their mechanism in the cellular remodeling process.

Data Charting
The data charting of these articles was performed by two investigators (PK and AC) independently, and any discordance was addressed by a third researcher (DKB). The criteria for data charting were according to JBI (Joanna Brigg's Institute) based on author, reference, and primary outcomes or results relevant to the broad research question. 17

Data Extraction
The data extraction from each study related to participant and study characteristics, the biomarker(s)

Figure 2. PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analysis) of Evidence Related to Biomarkers in EARR Based on Predetermined Search Strategy.
studied, the medium and technique of biomarker(s) study, and the outcomes related to biomarker(s) expression. Full details are given in Table 2.

Participant characteristics
The majority of studies had 20 or fewer participants. Three studies mentioned participants or teeth in two experimental and one control group with 20 patients in each group. 7,13, 19 Mah and Prasad mentioned two resorption groups; one group was examined for orthodontic resorption severity at 1-3 mm, while the second group looked at physiologic resorption of primary resorbing molars. 13 Balducci et al. classified the two experimental groups as mild (2 mm) and severe resorption (>2 mm) groups, 7 and George and Evans defined mild resorption ( 2 mm and severe resorption as >2 mm in their groups. 19 A total of 9 studies examined the resorption severity grades measured in mm, or classified it as mild/moderate/severe, or as coronal/apical resorption. 5,7,9,12,13,19,20,24,29 Resorption with respect to the duration of orthodontic treatment was considered in 7 studies. 7,9,19,20,23,27,30 While the majority of studies had both male and female participants, two studies investigated only female participants. 9,15 Most of the studies collected biomarkers for the experimental or control teeth from the maxillary central and lateral incisors. However, controls varied, depending on the study: for example, external (in different subjects), 5,7,9,12,13,19,20,24,25,27,29,30 or internal (baseline values), 22 antagonistic teeth, 15 and contralateral teeth, 16,21,26 while one of the studies mentioned no control. 28 Only five studies considered physiological resorption of the primary resorbing molars. 5,13,16,21,29

Study characteristics
The majority of studies were cross-sectional, although six mentioned the collection of samples at more than one observation time. 5,24,[26][27][28]30 There were four split-mouth design studies, 15,21,22,26 two of which considered 100 g force retraction on one side of the mouth and 150 g force on the other side. 15,22 The amount of resorption was judged radiographically in most studies, with intraoral periapical radiograph specified in six, 7,15,20,22,24,28 panorex in three, 5,12,27 and micro-computed tomography in only one study. 26

Main text continues on page 12
biomarkers such as the dentinal proteins DMP1, 7 DPP/PP, 7 DSP, 7 DSPP, 29 cytokines IL-6, 9 RANKL, 19 and RANKL/OPG ratio 19 showed better GCF detection in root resorption than in controls. But IL-1RA had higher levels in the controls versus the resorption group. 29 Studies evaluating the difference in resorption severity showed higher levels of DSP, 7 lower levels of granulocyte-macrophage colonystimulating factor, 26 and higher resorption protein concentrations in severe versus mild resorption (0.89 µg/µL ±0.32 µg versus 0.77 µg/µL ±0.21 µg, respectively). 19 Additionally, in comparison to nonresorbing teeth, mild and severe resorption showed higher RANKL/OPG ratios. 19 Specific protein bands in saliva have also been identified in mild to moderate resorption. 20 In physiologic root resorption of primary molars, upregulation of 37 resorption proteins was seen, as well as downregulation of 59 resorption proteins and IL-1RA levels, compared to no resorption groups in permanent molars. 16,21 A few longitudinal studies evaluated the rise or fall of biomarkers in GCF with resorption at different observation times. Kereshanan et al. mentioned the rise of DSP levels in GCF after the start of fixed orthodontic treatment compared with before treatment initiation, 5 while Thalanany et al. showed a significant increase in DSPP after two months of intrusion. 25 Protein abundance was also evaluated in GCF by Mohd Nasri et al. 28 In comparison, Ahuja et al. evaluated multiple markers that peaked at different observation times: IL-1β at 1 and 7 days, IL-4 at 1 and 3 days, interferon gamma (IFN-γ) at day 3, tumor necrosis factor-α at 3 hours and at 28 days, and IL-7 and granulocyte-macrophage colonystimulating factor (GM-CSF) at 28 days. 26 Two studies evaluated the efficacy of one method over another in the detection of biomarkers. Sha et al. found that DSPP detection by spectrophotometric ELISA (limit, 5.0 pg/mL) was less sensitive than electrochemical detection (limit, 0.5 pg/mL). 2 Lombardo et al. showed a modified micro-bead approach to be better than standard ELISA for DSP detection in GCF. 23 However, for salivary detection, Yashin et al. showed a significant increase in IL-7, IL-10, IL-12p70, and IFN-γ, and a significant decrease in IL-4 in moderate to severe resorption compared to controls. 12 That same study also showed lower osteocalcin in the blood for resorption compared to no resorption. 12 Salivary proteins have been shown to vary in young and adult root resorption groups, with an increased expression of 244 proteins in the moderate-to-severe young resorption group and only 58 proteins in the adult group compared to controls. 24 Additionally, 187 metabolites were identified by Zhou et al. in the saliva of root resorption groups compared to their no resorption group. 27 Cut-off values of biomarkers in orthodontic root resorption were studied by Mandour et al. at less than 432.6 pg/mL for IL-1RA and greater than 7.33 pg/mL for DSPP, with greater reliability for DSPP than IL-1RA. 29 Additionally, Zain et al. proved that treatment duration was a contributing factor for resorption, with the absorption spectrum of DSPP rising in subjects within 3, 6, and 12 months of treatment. 30 Studies have also evaluated changes in biomarkers in resorption associated with two different force levels (100 g and 150 g). 15 Wahab et al. showed a statistically significant increase in TRAP levels from baseline to 3-5 weeks for 100 g force and in AST at 5 weeks for 150 g force, with the ALP group only showing a slight increase in both force levels. 22

DISCUSSION
The variation in multiple biomarkers in EARR based on the outcome measurements of severity, physiologic resorption, and orthodontic treatment versus controls, different time intervals, and methods of detection is presented in Table 3. Figure 3 presents a pictorial compilation of all biomarkers studied in this review.
Wide heterogeneity was noticed in the reviewed studies with regard to tooth selection for resorption, study settings, biomarker selection, collection, and evaluation. However, the majority of studies took measures to alleviate confounding bias in terms of inflammation caused by coexistent periodontal or gingival inflammation. Several studies ensured good oral hygiene and gingival and periodontal condition by measuring probing depth, bleeding on probing, and the gingival index, since inflammation may alter the biomarker levels in biofluids. 2,5,7,9,12,13,15,16,20,21,23,24,[26][27][28]30,31 Furthermore, to rule out confounding variables for biomarker levels, many of the studies excluded patients with smoking, pregnancy, previous orthodontic treatment or systemic illness, and craniofacial disorders. A few studies also mentioned discouraging the use of antibiotics and anti-inflammatories 7,9,13,15,20,22,23,27,28 or mouthwashes like chlorhexidine, 22 but this was not a standard practice across all the studies.

Table 3. Evidence-based Compilation of Biomarkers in External Apical Root Resorption.
Continued on next page.

Biomarkers of Root-resorption in Oral Biofluids
Various biomarkers in the GCF were identified by this review, including dentinal proteins (DPP, DSP), cytokines (IL-6, OPG, OPN), RANKL, and enzymes (ALP, AST). A few of these were identified in the 2019 systematic review by Tarallo et al., 17 who evaluated EARR biomarkers in GCF from seven studies after quality assessment. However, this scoping review identified additional biomarkers, including DSPP, DPP, DMP1, cytokines, and their receptor antagonists (IL-1, 2, 4, 5, 6, 7, 8, 10, 12, 13, TNF-α, OPG, OPN, RANKL, and IL-1RA), along with resorption proteins in both the GCF and saliva. A recent review by Mona et al. evaluated protein-protein interactions of EARR biomarkers in variable study designs of human and animal studies, including case-control studies, reviews, and physiologic resorption. 32 However, it has limited applicability in studying resorption in clinical orthodontic practice, unlike this scoping review. In light of these data, future research should include a bioinformatics analysis for the biomarkers identified by this scoping review, to ascertain the protein interactions responsible for clinical resorption overlapping other periodontal and pathological problems.
The majority of studies in the current review identified dentin-specific proteins in EARR, espe-

Figure 3. Evidence-based Compilation of Biomarkers in External Apical Root Resorption (EARR).
ALP, alkaline phosphatase; AST, aspartate transaminase; DMP1, dentin matrix protein 1; DSP, dentin sialoprotein; DPP, dentin phosphophoryn; DSP, dentin sialoprotein; DSPP, dentin sialophosphoprotein; GM-CSF, granulocytemacrophage colony-stimulating factor; IFN-γ, interferon gamma; IL, interleukin; IL-1, interleukin-1β; OPG, osteoprotegerin; OPN, osteopontin; PP, dentin phosphophoryn (alternate abbreviation in the literature; RANKL, receptor activator of nuclear factor kappa-Β ligand; TNF-α, tumor necrosis factor-α; TRAP, tartrate-resistant acid phosphatase. cially DSPP, DSP, DPP/PP, and DMP1. Of these, the DSP and DPP proteins are the most abundant noncollagenous proteolytic cleavage products of DSPP found in dentin (5%-8% and 50%, respectively). This review also identified DSP as a potent resorption marker, 5,7,16,23 both in orthodontic and physiologic resorption. 2,5 It is more dentin-specific than DPP and is found in odontoblasts and the extracellular matrix of pre-dentin, dentin, and dental pulp, but is not prevalent in bone, cartilage, ameloblasts, or other oral tissue components. 5 However, the presence of DSP and DPP in control subjects with no resorption 5,13 also indicates the release of dentinal matrix proteins in the GCF from pulpal cells during root mineralization in young permanent teeth with patent apices. These dentinal matrix proteins may not be exclusively present in dentin, since both are products of a larger precursor protein, DSPP, which is also present in osteoblast cells. 5 Osteopontin is another glycosylated protein of the dentin matrix and bone, produced by odontoblasts along with other bone precursors such as cementum and macrophages. The current review shows the presence of degraded fragments (54 kDa and 66 kDa) of OPN in the GCF of mild and severe resorption. 19 This occurs as a result of the enzymatic activity of cysteine proteases, causing degradation of bone and the dentin extracellular matrix, which is also seen in periodontal disease. 33 In addition, this review found that different cytokines, including proresorptive IL-6, show higher GCF levels in severe compared to no resorption, 9 which is supported by rat studies showing an association of IL-6 with induction and further progress of mechanically induced root resorption. 34 Furthermore, IL-6 has an established role in osteoclastogenesis and bone remodeling associated with orthodontic force application by inducing RANKL and osteoclasts formation. 35 Additionally, osteoclastogenesis is governed by the RANKL/OPG ratio, 10 as seen in the current review, where this ratio was significantly higher in severe resorption than in controls. 19 Other clastogenic mediators (TNF- and IL-7) also augment resorption in GCF, 26 with previous literature supporting their role in bone resorption in orthodontic tooth movement. 10 The orthodontic force levels, 150 g force versus 100 g force, seem to have no effect on tooth resorption. Nevertheless, 150 g force application causes a significant increase in ALP on the mesial side within one week compared to 100 g force. 15 Alkaline phosphatase (ALP) is known to support osteoblastic activity. 11 A similar rise in ALP was seen in previous studies at 1 to 3 weeks 36 and at 2 weeks after orthodontic force application. 37 The TRAP and AST enzymes also vary with the level of force. The TRAP levels showed a significant rise from baseline with 100 g force but not with 150 g force. 22 The AST on the other hand showed a significant rise with 150 g force within 5 weeks, but not with 100 g force. 22 Previous literature also supports a rise in TRAP proportionally with the orthodontic force magnitude, 31 and higher AST levels at compression versus tension sites, thus favoring the resorptive activity. 38 This review found that salivary metabolome was associated with specific clusters of metabolites in EARR using partial least squares discriminant analysis, which may be further explored for diagnosis of resorption. 27 These clusters include purine and arachidonic acid metabolites, known for chemotaxis of inflammatory cells as well as periodontal damage propagation/resorption. 39 This further produces reactive oxygen species causing a shortage of local oxygen concentrations, and triggering the RANKL pathway. 39 Thus, these metabolites may indicate resorption as well as periodontal damage, further confirming the need to ascertain periodontal health when performing such biomarker studies or examining the reciprocal effect of periodontal inflammation on these biomarkers and on resorption.
Best practices for biomarkers isolation and detection have also been highlighted by this review. While several of the reviewed studies primarily mentioned conventional ELISA, two comparative studies established the increased sensitivity of electrochemical over spectrophotometric ELISA, 2 as well as micro-beads over conventional ELISA. 23 These conventional and microbead assays offer several advantages: they are sensitive, non-invasive, include no radiation exposure, provide stage-wise monitoring and at-risk assessment, and can be used to diagnose and predict the clinical course of therapy. 13 This review also found a newer non-invasive approach for non-targeted metabolomics using highresolution nuclear magnetic resonance spectroscopy. 27 This method can identify newer mediators or varied human disease pathways in the EARR domain, offering significant benefits by providing multi-component information simultaneously.
Hence, the current review answers our primary research question by examining the variation in levels of all biomarkers in EARR which can be isolated in the oral fluids. The resorption markers have been studied in orthodontic treatment as well as in comparison with physiologic resorption. In addition, this review also highlights the best methods for biomarker isolation. It also mentions the study design drawbacks for consideration in future evaluations and proposes further bioinformatic analysis of identified cellular markers.

LIMITATIONS
Although the reviewed studies met all inclusion and exclusion criteria, there was an extensive heterogeneity of biomarkers, including a wide range of cytokines, dentinal proteins, receptors, and colonystimulating factors, as well as resorptive proteins and metabolites. The study designs were also varied, mostly cross-sectional using single observation samples, although a few studies evaluated resorption longitudinally with variation in mediator levels at different time points. None of the reviewed studies performed randomization to examine the effects of variable orthodontic forces or treatments on resorption. The sample size was generally small and unequal between the experimental and control groups in the majority of studies. Other confounders were unequal male-to-female ratio, no standardization of study prerequisites related to inflammatory conditions or history of smoking, and antibiotics or antiinflammatories, all of which may have a bearing on biomarker levels.

CONCLUSIONS
The conclusions of this scoping review may be summarized as follows:  Several biological markers have been identified in external apical root resorption in various oral body fluids (GCF, saliva, and blood). These include dentinal proteins, cytokines, enzymes, and protein metabolites.
 Higher severity of resorption showed increased DSP, 7 DPP, 7 and RANKL/OPG ratio 19 and higher resorption protein concentration 19 compared to mild resorption, although the evidence is scanty.
 Cut-off values of biomarkers for root resorption were mentioned with IL-1RA (<432.6 pg/mL) and DSPP (>7.33 pg/mL), but this evidence requires further validation. 29  Detection of DSPP by electrochemical ELISA (limit, 0.5 pg/mL) is more sensitive than spectrophotometric ELISA (limit, 5.0 pg/mL). 2 Furthermore, DSP detection in the GCF by modified micro-bead approach proved better than standard ELISA. 23 Several points for further investigation are suggested based on the findings of the current review:  Next steps include identifying the most sensitive and specific biomarkers (dentinal proteins/inflammatory cytokines/metabolites) in the GCF or saliva for early-stage EARR detection, and evaluating them repeatedly during the progress of treatment. A biosensor point-of-care screening device based on the most potent biomarker to detect root resorption is also suggested. 40  Cut-off levels for biomarkers need to be established, and a non-invasive clinical test developed for early diagnosis of iatrogenic resorption.
 Study designs should be standardized to generate unbiased high-quality evidence.
 Bioinformatic analysis is needed to identify the protein interactions, which may also overlap with other oral inflammatory conditions including external cervical resorption in chronic periodontitis. 41