Influence of extracellular matrix scaffolds on histological outcomes of regenerative endodontics in experimental animal models: a systematic review

The protocol for this review was registered in The International Prospective Register of Systematic Reviews (PROSPERO) as (CRD42023433026) and conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Protocol (PRISMA-P) 2020 statement [37].

The research question was formulated as: Can decellularized ECM-derived scaffolds influence the histological outcome of regenerative endodontic applications regarding the nature and pattern of tissues?

Data were extracted from eligible articles using pre-designed data extraction tables (Microsoft Word) by two independent reviewers (H.E. and R.E.). Any disagreements were solved by discussion among them. Only the data related to in vivo experiments of the included studies were extracted for analysis. Also, in studies assessing both periodontal tissues and supporting structures regeneration as well as dentin pulp regeneration, data were extracted only for the latter. Extracted data included:

List of abbreviations mentioned in the tables is included in supplementary file 6.

Assessment of the reporting quality for all studies included in this systematic review was performed using the Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines [54] (H.E. and R.E.). In order to evaluate the 21 items, a modified scoring system [55] was used as follows: all subitems were reported “Yes” = 2 points, not all subitems were reported “Unclear (Uc)” = 1 point, or all subitems were not reported “No” = 0 points. Afterwards, quality coefficient (QC) was generated by calculating the sum of all the points obtained for each study divided by 42 “the maximum possible points per study”. The quality of study was reported as “Excellent” when QC was 0.8–1, “Average” if QC was 0.5–0.8 and “Poor” if QC was < 0.5.

Assessment of bias for all studies included in this systematic review was performed using SYstematic Review Centre for Laboratory animal Experimentation (SYRCLE) risk of bias (RoB) tool [56] (H.E. and R.E.). A modified scoring system [57] was used according to the total number of “yes” answers to the assigned questions (Yes = 1 point) The degree of bias was calculated as follows: High risk: 0–4, Moderate risk: 5–7 and Low risk: 8–10.

The records from the different databases were combined (n = 384) then duplicates were removed (n = 96). Then additional records identified through other sources were added (n = 3). For the initial screening of the imported records (n = 291), the titles of the papers were identified and decided whether they were relevant to the topic or not (n = 184). Reviews were excluded (n = 57) and editorials/books were excluded (n = 3). For the final screening, the abstracts of all relevant articles (n = 47) were then carefully appraised to identify eligible studies. Full texts of the relevant articles were screened. During this final screening phase, studies were excluded as: in vitro studies (n = 7), ex-vivo studies (n = 1), and studies not including dECM (n = 22). Excluded studies by full text screening (n = 30) and reasons for exclusion are listed in supplementary file 2. Total number of studies included in this systematic review was 17 studies (n = 17). The flow chart for search process and study inclusion are presented in (Fig. 1) according to PRISMA 2020 guidelines [37].

Different animal models were used in the included studies. The majority used mice [38, 39, 41, 42, 45, 46, 48, 50, 51, 53] followed by rats [43, 44, 49, 52], pigs [39, 40] and beagle dogs [23, 47]. The main source of ECM used was swine [23, 39, 41, 47, 53], human [46, 48, 50] or bovine dental pulp [43‐45]. Others used cell-generated ECM [38, 42], swine tooth bud [40], human amniotic membrane [49, 52] and rats’ submandibular glands [51]. Regarding the study model, scaffolds were ectopically transplanted in six studies [38, 39, 43, 44, 46, 52], semi-orthotopically transplanted in twelve studies [39‐42, 45, 47‐53] and orthotopically transplanted in just one study [23]. In terms of the regenerative approach, two thirds of the included studies used cell-seeded scaffolds [39‐42, 45, 46, 48‐53] while the rest used cell-free scaffolds [23, 38, 43, 44, 47]. Six of the former studies also compared between the two approaches [40, 45, 48, 49, 51, 52]. As for the scaffold form, this was either tissue sheets placed within either treated dentin matrix TDM [39, 42, 47, 51, 53] /tooth slices [41, 45, 48], or freeze-dried sponges [43, 49] or sheets [23, 44, 51], powder [46], hydrogel microspheres [50], injectable hydrogel [52, 53] or whole decellularized tooth bud [40]. Only four studies used crosslinked scaffolds [43, 49, 50, 52]. Regarding terminal sterilization, scaffolds were either immersed in penicillin/streptomycin [39, 41, 46, 47, 51, 53], or sterilized using ethylene oxide (EtO] [23, 50] or peracetic acid (PAA) [45]. The rest of studies did not report the method of sterilization [38, 40, 42‐44, 48, 49, 52]. Interestingly, the concentration of ECM components was reported by only six studies [23, 43, 49, 50, 52, 53]^. Regarding the methods of in vivo outcome assessment, only histological evaluation was performed in two studies [44, 46], while histological and immunohistochemical analysis were done in the rest of the studies. When ECM scaffolds were compared to other non-ECM-derived scaffolds, regeneration of both pulp-like and dentin-like tissues were reported in the ECM groups in eight studies [23, 38‐42, 47, 50] while pulp-like tissue was solely reported in five studies [44, 46, 49, 52, 53]. Regarding the cell homing approach versus cell transplantation approach, it was noted that in cell-seeded scaffolds, pulp-like and dentin-like tissue regeneration was more likely to be detected compared to cell-free scaffolds [39‐42, 46, 48, 50, 51]. Conversely, quality of pulp-like tissues in cell-seeded group was reported, in one study, to be less optimal that in cell-free groups [52].

All studies were analysed using Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines. Study design was described by all studies [23, 38‐53]. However, none of the studies provided precalculated sample size [23, 38‐53]. Inclusion and exclusion criteria were not properly described in any of the studies [23, 38‐53]. Randomization was carried out in only three studies [45, 49, 52]. Blinding was only carried out in four studies [43‐45, 52]. Outcome measures and experimental animals’ details were adequately described by all studies [23, 38‐53]. Statistical analysis of in vivo outcomes was not performed in ten studies [23, 28, 38, 40, 42, 44, 46, 47, 52, 53] Experimental steps were unclear in all studies [23, 38‐53]. Results were adequately described with statistical evaluation of each group in only five studies [41, 43, 45, 48, 51]. Data on the animal model and in vivo outcomes were properly described in abstracts in only seven studies [23, 39, 40, 45‐47, 52, 53]. The introduction section provided adequate background on the topic in all studies [23, 38‐53], however, none of them explained the rationale of using a specific animal model and its relevance to human biology. In all studies [23, 38‐53], objectives were clearly described. In one study [44], the ethical statement was not provided. Housing details and animal care were reported in only three studies [23, 40, 41]. Adequate interpretation of results as well as study protocol registration were done in all studies [23, 38‐53]. In one study [38], the generalizability of outcomes and clinical implications of experimental results were not stated. Data access statement was mentioned in only six studies [23, 47, 48, 50, 51, 53]. While declaration of interest statement was mentioned in all studies except for two [38, 39]. Overall, quality coefficient ranged between 0.5–0.7 indicating an “Average” grade of all the reviewed studies. The assessment criteria and their results are listed in Table 3.

All studies were analysed using SYstematic Review Centre for Laboratory animal Experimentation (SYRCLE) risk of bias (RoB) tool [56]. Regarding selection bias, in all studies, allocation sequence was unclear [23, 38‐53]. However, baseline characteristics of animals were similar among different groups in all studies [23, 38‐53]. Only three studies reported concealment of sample allocation [40, 49, 52]. Investigators were blinded in only two studies [44, 45]. Outcome assessors were blinded in four studies [44, 45, 49, 52]. Random allocation of animals in study groups, random selection of animals for assessment, and addressing incomplete outcome data, were mentioned in none of the included studies [23, 38‐53]. Regarding reporting bias, there was no selective reporting of outcomes in any of the studies [23, 38‐53]. All of the reviewed studies showed high risk of bias where twelve studies scored 2/10 in RoB assessment [23, 38‐43, 46‐48, 50‐53] while the remaining three studies [44, 45, 49] scored 4/10. The overall risk of bias was high in all studies. The risk of bias assessment criteria and their results are listed in Table 4.

Qualitative analysis of the studies was performed; however, meta-analysis was not feasible due to wide variations amongst studies in methods for assessment, decellularization protocols, animal models used, and nature and source of decellularized ECM.

The unpredictable regeneration of tissues following conventional regenerative endodontic procedures remains a major concern [65]. Several histological studies using blood-derived scaffolds have reported evidence of ectopic tissue regeneration [16, 65, 66]. Bone-like, cementum-like tissues and absence of odontoblast-like layer are often observed within regenerated tissues [67, 68]. Regenerated tissues appear to be mineralized counterparts of a mixture of tissues such as dentin, bone, cementum and periodontal tissues which have been collectively termed dentin associated mineralized tissues (DAMT) owing to their deposition and close association to the original inner canal wall dentin [69].

Results of this review indicate that decellularized ECM-derived scaffolds are novel biomimetic materials that can lead to enhanced angiogenesis and regeneration of dentin-pulp-like tissues compared to non-ECM-derived scaffolds [23, 38‐51]. Notably, the included studies that employed decellularized ECM-derived scaffolds, used other naturally-derived scaffolds as controls rather than synthetic scaffolds [23, 38‐51]. It was noted that thirteen of reviewed studies used IHC analysis to detect the expression levels of odontogenic/angiogenic markers. This method of evaluation is valuable for distinguishing the nature and quality of newly-formed tissues. The regenerated tissues were characterized by high expression of DSP and DPP, increased collagen deposition and neovascularization as reported by Ravidnran et al. [38] and Huang et al. [42] who both used cell-generated dECM scaffolds. These results were also in accordance with the findings of Chen et al. [39], Zhang et al. [40], Hu et al. [41], Alqahtani et al. [23] and Zheng et al. [50] who all used porcine dental pulp ECM-derived scaffolds. On the contrary, bone/cementum-like hard tissue formation was reported in the study by Fu et al. [47] in the acellular ECM group compared to high DSPP and DMP-1 expression in the laminin-coated ECM group. Formation of bone-like tissue was also observed by Bakhtiar et al. [43] when crosslinked bovine pulp dECM scaffold was used. This also coincides with the findings of Kim et al. [48] who reported hard tissue formation not resembling dentin when human dental pulp dECM was the scaffold, although they reported expression of some odontogenic markers such as DSPP and DMP-1 in the newly formed tissue. Enhanced angiogenesis and detection of pulp-like tissue were reported by Bakhtiar et al. [44, 49, 52] when either bovine pulp dECM or amniotic membrane dECM were used as scaffolds.

Generally, in the reviewed studies, it was evident that cell-seeded dECM-derived scaffolds triggered more organized tissue regeneration and robust neovascularization compared to cell-free scaffolds [51] except for one study [52]. However, the fact that those studies have used ectopic or semi-orthotopic models, cannot be overlooked. As the placement of the scaffold out of its original microenvironment may have had a significant effect on host cell recruitment and their behaviour [70]. Indeed, in the study by Alqahtani et al. [23], evidence of dentin-pulp regeneration was reported when the cell-homing approach was used in an orthotopic model which indicates the importance of providing the appropriate “niche”.

In the reviewed studies, there were significant limitations that may have affected outcome assessment. None of the studies compared between dECM-derived scaffolds and blood derived-scaffolds currently used in REPs. There was methodological heterogeneity in the histological assessment and interpretation of results. Additionally, the duration of studies ranged from 2 to 12 weeks in most of the included studies which may not be sufficient time to assess scaffold remodelling and long-term efficacy of the scaffold on dentin-pulp regeneration. Although all included studies had both in vitro and in vivo phases, we only extracted data that was relevant to the in vivo experiments and histological outcomes. The quality and risk of bias assessments indicated that most of reviewed records are preliminary studies lacking sufficient sample size, randomization, blinding and proper statistical analyses. Other limitations of this systematic review include the small number of available studies.

Future studies should work towards optimizing decellularization and sterilization protocols with maximum preservation of ECM architecture and innate growth factors content. Moreover, evaluating the effect of adding natural or synthetic modifiers to the dECM-derived scaffolds should be also addressed, with the purpose of improving specific mechanical characteristics, such as stiffness, viscoelasticity and biodegradation. Additionally, preclinical and animal models should mimic the clinical protocol of REPs with sufficient long-term follow up. The preparation of freeze-dried hydrogel form that can be stored as an off-the-shelf product could also aid in the clinical translation of using dECM-derived scaffolds in REPs. Until this “optimal” scaffold is available, it is important not to lose sight of clinically-relevant outcomes, namely; patient-centered and clinician-centered, as these continue to be measurable and reproducible.