Microbiome and Inflammatory Biomarkers Associated with Palatal Wound Healing

Abstract

Aim:

The clinical outcomes of a variety of surgical procedures highly depend on tissue repair and show high variability among patients. There is a gap in the literature on how the host inflammatory response, the microbiome, and the interplay between them can influence oral mucosa healing. In this pilot study, we aimed to evaluate the microbiome and biomarkers profiles in patients who had desired versus undesired wound healing in the palatal mucosa.

Methods:

Seventeen patients underwent a free gingival graft (FGG) for socket preservation. Palatal wound closure (WC) and epithelization (EPT) were assessed clinically. Biofilm from the palatal wound was collected before the surgical procedure and 3, 7, 14, and 30 days postoperatively. The inflammatory exudate was sampled at days 3 and 7. At 14 days posttreatment, patients were classified into two groups based on EPT rates: (1) undesired healing (UH) and (2) desired healing (DH).

Results:

No difference was observed in alfa diversity over time or between groups. In beta diversity, both UH and DH showed microbiome changes on days 3–7 and 7, respectively, compared to the baseline (p=0.01), returning to its initial condition 30 days later. There was a trend toward a different microbiome profile between groups at day 7 (p=0.08). Bacterium composition in DH showed a balance between healthy species and oral pathogens over time whereas UH composition was characterized by microorganisms correlated with epithelium invasion/cytotoxicity; virulence factor upregulation; and oral diseases, such as periodontitis and aphthous stomatitis, until day 30. UH showed an increase in IL-6, MCP-1, and MIP-1α over time, and DH showed a decrease in TIMP-1, IL-1β, and MIP-1α. On days 3 and 7, MIP-1α and MMP-2 showed greater concentrations of DH in the intergroup assessment, and MCP-1 increased at day 7 in UH.

Conclusion:

Specific microbiome/inflammatory profiles are associated with DH and UH.

Graphical Abstract

The aim of this study was to evaluate the influence of the microbiome on the palatal healing process of open wounds. The results showed that some bacterial species may be related to a slower wound closure process.

INTRODUCTION

Mucogingival defects around teeth and implants, such as gingival recession, peri-implant soft tissue dehiscence, and lack of keratinized tissue require surgical treatment to reestablish function and esthetics. Autogenous grafts, either connective tissue grafts or free gingival grafts (FGGs), are considered the gold standard for soft tissue reconstruction¹. However, autogenous soft-tissue grafts have some drawbacks, including hemorrhagic risk, patient discomfort, and a limited amount of tissue that can be harvested^2,3. Even with the best approach, great variability in clinical outcomes and patient’s perception such as discomfort and number of analgesics are often observed^4–6. Therefore, factors other than the surgical procedure per se may influence the surgical therapy’s success and posttreatment morbidity, including host inflammatory response and the area’s microbial composition.

Wound healing impairment has been a frequent issue in public systems worldwide. In the US, expenditures between U$28.1 and U$96.8 billion have been reported to treat wound healing problems⁷. Factors that have been associated with delays in tissue repair include systemic conditions, such as diabetes mellitus⁸, age⁹, and obesity¹⁰. Moreover, lack of oxygenation, infection, and insufficient blood supply are local factors that significantly affect tissue repair^10,11. However, not only unfavorable local and systemic conditions impair healing, but also innate characteristics of the host inflammatory response¹² and the microbiome environment can play a role here¹³.

Dysregulated innate responses can trigger unbalanced tissue remodeling which can compromise wound closure (WC)¹⁴. Studies evaluating skin healing showed that impaired inflammation can generate scars¹². Conversely, studies on fetuses, among whom very little inflammation is present, show that healing occurs rapidly and scarlessly¹⁵, suggesting a major role of inflammation during repair. Therefore, inflammation can influence the entire healing process, and any misbalance (excessive/diminished) in this process, can lead to delayed or unfavorable wound healing¹⁶. However, most of the studies on wound healing have mainly been conducted to evaluate skin, and minimal evidence is available on human oral mucosa, which has a different immunoinflammatory process¹⁷, with intimal and constant contact with highly diverse microbiota.

The microbiome plays a key role in modulating the healing process either to resolution or chronicity¹⁸ via several pathways, including toll-like receptor (TLR) activation and inflammation¹⁹. Commensal bacteria in contact with the skin wound’s edges can trigger chemokines release, such as CXCL10, by S. epidermidis, leading to improved repair steps²⁰. Conversely, nonresident bacteria, such as S. aureus, can jeopardize epidermis healing, thereby blocking immunological response due to the release of several endotoxins²¹. An interplay is observed between host response and microbiome composition in various biological fields^19,22. Focusing on the oral cavity, Delima et al.²³ observed a microbiome profile shift in pre-wound areas due to inflammation, and this may modify the healing process. However, there is a lack of studies evaluating inflammatory and microbiome profiles and how they can influence the palatal healing parameters. Therefore, we aimed to characterize the inflammatory and microbiome profiles of palatal mucosa wounds and associate them with delayed versus normal tissue repair.

METHODS

The present pilot study is a sub-study of a parallel-arm, double-blinded, and controlled RCT (ClinicalTrials.gov: NCT05171400) approved by the human subjects’ ethics board of São Paulo State University (UNESP) (CAAE: 29349720.9.0000.0077). The study followed the CONSORT STATEMENT and the Declaration of Helsinki of 1975 (revised in 2013). All subjects agreed to participate by signing an informed consent form.

Population

Subjects referred to the Division of Periodontics, Sao Paulo State University- UNESP, Brazil, in need of tooth extraction followed by ridge preservation were recruited from July 2021 to August 2022. Patients who fulfilled the following criteria were included in the study: aged≥18 years, no systemic issues that could affect tissue healing (e.g., diabetes mellitus), plaque and gingival index<25%, single-tooth extraction with both adjacent teeth free of clinical attachment loss/periodontally healthy, and no morphological or pathological conditions in the palatal area. The exclusion criteria were a systemic condition that contraindicates oral surgery, use of medications that influence wound healing or surgical protocol, smokers, pregnancy or lactation, previous surgical procedure at the same intervention area, oral mucosa lesions, and prosthesis rehabilitation covering the palate.

Surgical Procedure

Patients were enrolled in a biofilm control program, which consisted of dental hygiene instruction, prophylaxis, and scaling and root planning when needed. Carious lesions were restored, and hopeless teeth that did not fit into the inclusion criteria were removed. Clinical parameters, photographs, and periapical radiographs were taken.

The same operator (MMVM) as previously described performed the surgical intervention^2,3. Dexamethasone was prescribed 1 hour before the surgery. A minimal traumatic extraction was executed using intrasulcular incision and periotome. The fresh socket was irrigated with saline solution. Thereafter, a 2-mm-thick FGG was removed from the palatal area (between 1^st pre-molar and 1^st molar) using an 8-mm circular punch connected to a refrigerated low-speed handpiece (Figure 1A and 1G). To ensure graft removal standardization, an individual acrylic stent was placed in the palatal area containing the circular punch mold. The FGG was manually trimmed to fit and seal the fresh socket entrance and stabilized by interrupted sutures (Silk 4.0, Ethicon Johnsons do Brasil^®- São José dos Campos, São Paulo, BRA). Oral and written postoperative recommendations were given to the patients, including the following prescriptions: chlorhexidine 0.12% antimicrobial rinse twice daily for 2 weeks starting immediately after surgery and sodium dipyrone (500 mg every 8 h) in pain episodes. Sutures were removed after 7 days.

Figure 1.

Representative clinical cases of each group: Desired Healing group (A-F); Undesired Healing group (G-L).

Clinical measurement

One calibrated examiner (ACFB) measured clinical parameters. The calibration exercise was carried out by evaluating WC in the palatal area in standard photographs of 10 patients from previous studies in a two-day interval²⁴. The measurements were assessed using intraclass correlation, and its coefficient (ICC) reached 85%. The clinical measurements included WC, using a scale in the palatal area as a reference. Standardized photographs were taken with a professional camera with mirror placement of ~45° with the lens². The remaining area was measured in square millimeters using software (Image J^®– NIH, Bethesda, Maryland, USA) at baseline (trans-operative) and 7, 14, 21, 30, and 90 days after surgery; Epithelialization (EPT): standardized photographs following the same parameter mentioned above were taken after a disclosure solution (Disclosing Agent, Dentsply, Charlotte, NC, USA) was applied in the wound on days 7, 14, 30, 21, and 90 postoperatively²⁵. The epithelialized area was calculated as the percentage of the original wound area using software (Image J^®– NIH, Bethesda, Maryland, USA) calibrated with a measuring scale in the palate (Supplementary Figure 1).

Patients were classified into two categories based on EPT percentage on day 14: desired healing (DH) (patients who reached 95–100% of the original wound EPT on day 14 or undesired healing (UH) (patients who presented less than 95% of the original wound EPT on day 14). The cut-off values to categorize patients were based on a previous study². Palatal spontaneous healing had an epithelization average rate of 86.2% at 14 days after graft harvesting while palatal healing associated with an adjunctive therapy had an EPT rate of 91%. Combining both groups and getting the data of half of the patients who healed faster, a 95% EPT rate was observed². Therefore, it was established a 95%-EPT as the cut-off 14 days after surgery as desired healing.

Immunological-biomarker assessment

On days 7 and 14, the inflammatory exudate was collected from the palatal wound, placing the paper point at the wound’s edges. The area was dried and isolated with cotton rolls. Suction was used during the entire sample collection procedure to avoid saliva contamination. Sterile paper points were used and kept in position for 40 seconds, and any blood contamination was discarded. Samples were stored in sterilized microtubes containing buffer solution at −80°C. Before multiplex analysis, the samples were re-concentrated using a spin filter 3K/PK100, focusing on better detection. The biomarkers were assessed using multiplex assays (MILLIPLEX^® Multiplex Assays kits - San Luis, Missouri, USA) via Luminex (xMAP)²⁶. The following biomarkers were evaluated: cytokines (IL-1β, IL-6, IL-13, TNFα), chemokines (MIP-1α, MCP-1), proteinases (MMP-2, MMP-9, TIMP-1, TIMP-2), and growth factors (EGF, FGF-2/FGF Basic, PDGF-BB, VEGF). A Bradford assay (Quick Start^™ Bradford Protein Assay, Hercules, California, USA) was carried out to quantify each sample’s total protein content and used for biomarker concentration adjustment (pg/mg)²⁶.

Microbiome assessment

Microbiological samples were collected from the wound using a sterilized swab before surgery (epithelium area) and on days 3, 7, 14, and 30 postoperatively. The area was dried and isolated with cotton rolls. Suction was used during the entire sample collection procedure to avoid saliva contamination. Samples were stored in sterilized microtubes containing 300 μl phosphate buffer saline 0.05% Tween 2 (PBS) at −80°C. DNA extraction was performed using a specific kit according to the manufacturer’s instructions (The MasterPure Complete DNA&RNA Purification Kit - Biosearch Tech^™, Hoddesdon, UK). For sample enrichment, 16S rRNA genes PCR were executed using universal primers as 27F(5’-ACGGYTACCTTACGACTT–3’) and 1492R(5’-AGAGTTTGATCMTGGCTCAG-3’). Two regions of the 16sRNA were selected to be analyzed (V3–V4) using the following primers, which were trimmed to barcodes for sample identification: 341F(5’-CCTACGGGNGGCWGCAG-3’) and 785R(5’-GACTACHVGGGTATCTAATCC-3’)²⁷ (Integrated DNA Technologies, Inc. – Coralville, Iowa, USA). The samples were sequenced on the 2X300 kits Miseq instrument (Illumina).

Statistical Analysis

Demographic and clinical analysis results are presented as mean±standard deviation. After the Shapiro–Wilk test to analyze data normalization distribution, age and biological sex were analyzed using the T-test and Fischer Exact Test, respectively. WC and EPT were evaluated using a two-way ANOVA test with the Tukey test for multiple comparisons (α=5%).

Biomarker concentrations are displayed as mean±standard deviation and were evaluated using the paired t-test/Mann-Whitney and t-test/Wilcoxon test based on the data’s Gaussian distribution. Microorganisms selected after the microbiome composition analysis (Fold-Change>1) were combined with biomarker concentration at the same time point to evaluate immune and microbiological display during tissue healing by principal component analysis (PCA).

Microbiome analysis was conducted using QIIME2, PhytoToAST, and R software. Alfa diversity was implemented using the Shannon estimator index. Intra- and intergroup differences in these parameters were assessed using the Wilcoxon test (per pair) and Mann-Whitney test, respectively. The beta diversity matrices of bacterial communities based on weighted Unifrac distance were represented by principal coordinate analysis and assessed using the permutational multivariate analysis of variance (PERMANOVA) test. The differences in samples dispersion between groups were tested using PERMDISP. Intergroup microbiome composition was evaluated using analysis of composition of microbiomes with bias correction (ANCOM-BC) focusing on fold change in bacterial abundance adjusted for multiple testing (p<0.05). A linear discriminant analysis-effect size (LEfSe) was conducted to determine relevant bacteria at each time point between groups. Only scores (log10) over 2 and p-values≤0.05 were considered in this analysis. Raw sequencing is available in the Sequence Read Archive (#14636272).

RESULTS

Seventeen patients were assigned to two groups after EPT assessment on day 14, undesired (UH, n=9) and desired healing (DH, n=8). There was no difference between groups regarding sex (UH: 6 females, DH: 5 females; p=1) or age (UH: 40.44±8.70 yo, DH: 43.125 ± 9.32 yo; p=0.549).

Clinical Outcome

The intragroup assessment was similar between groups with complete WC and EPT after 21 days (p>0.05; Figure 1). No difference was observed in wound size at baseline between the groups, which characterizes homogeneity in surgical wounds (p=0.84). However, better WC was reported in DH on days 7 (p=0.001) and 14 (p=0.002) than in UH. Not only were better rates observed in WC, but greater EPT percentages were detected in patients assigned to DH on day 14 (p=0.019; Supplementary Table 1).

Immunological Outcome

A reduction in IL-1β, MIP-1α, and TIMP-1 concentration from days 3 to 7 after surgery was observed in the DH group. IL-6, MCP-1, and MIP-1α increased in the UH group over time. Regarding intergroup assessment, MIP-1α showed higher concentration after 3 days postoperatively and MMP-2 on day 7 for DH whereas MCP-1 concentration was higher on day 7 in the UH group (Table 1).

Table 1.

Inflammatory biomarker concentrations after surgical intervention (pg/mg)

Groups	Biomarkers	3 days	7 days
DH	IL-1β	0.42±0.30 Aa	0.23±0.13 Ba
UH		0.27±0.21 Aa	0.27±0.30 Aa
DH	IL-6	0.41±0.20 Aa	1.08±1.14 Aa
UH		0.62±0.57 Aa	0.96±0.65 Ba
DH	IL-13	0.02±0.05 Aa	0.02±0.04 Aa
UH		0.005±0.003 Aa	0.005±0.005 Aa
DH	TNF-α	0.015±0.002 Aa	0.018±0.02 Aa
UH		0.005±0.003 Aa	0.01±0.004 Aa
DH	MIP-1α	2.24±1.88 Aa	1.22±0.87 Ba
UH		0.80±0.40 Ab	1.60±1.10 Ba
DH	MCP-1	0.09±0.04 Aa	0.08±0.05 Aa
UH		0.12±0.14 Aa	0.29±0.22 Bb
DH		13.60±4.33 Aa	6.24±0.80 Ba
UH	TIMP-1	13.44±8.92 Aa	7.61±5.76 Aa
DH	TIMP-2	10.29±4.66 Aa	19.68±19.21 Aa
UH		10.04±6.56 Aa	24.00±18.30 Aa
DH	MMP-2	9.85±8.84 Aa	9.25±6.30 Aa
UH		5.09±1.80 Aa	4.88±3.86 Ab
DH	MMP-9	1047,66±532.06 Aa	1818,20±1785,88Aa
UH		846.04±387.22 Aa	947.56±542.10 Aa
DH	EGF	0.09±0.03 Aa	0.07±0.06 Aa
UH		0.09±0.08 Aa	0.07±0.05 Aa
DH	FGF-2	0.16±0.23 Aa	0.63±1.36Aa
UH		0.10±0.07 Aa	0.21±0.16Aa
DH	PDGF-BB	0.07±0.14 Aa	0.23±0.24 Aa
UH		0.03±0.014Aa	0.45±0.96 Aa
DH	VEGF-α	0.02±0.003 Aa	0.021±0.03 Aa
UH		0.006±0.06 Aa	0.05±0.03Aa

Abbreviation: DH, desired healing group. UH, undesired healing group.

Different uppercase letters indicate statistically significant intra-group difference – Paired t-test/Mann-Whitney test, p<0.05 Different lowercase letters indicate statistically significant inter-group difference at the same time point – T-test/Wilcoxon test, p<0.05

Microbiome outcome

Alpha diversity analysis showed no significant difference in the within-sample diversity assessed using the Shannon index (p>0.05; Figure 2). In contrast, a difference in the beta diversity analysis (between samples) was observed after intragroup analyses (Figure 3). UH and DH showed a shift in microbiome on days 3–7 and 7 postoperatively, respectively, compared to baseline (p=0.01) and later time points (p≤0.02; Figure 3A). On day 30, this microbiome tends to return to its initial diversity. A tendency of distinct beta diversity was only observed on day 7 at intergroup assessment (p=0.08; Figure 3B).

Figure 2.

Alpha diversity represented by Shannon Diversity Index.

Figure 3.

Beta Diversity metric; PCoA (3-axis visualization) using weighted Unifrac distance (A); Statistically significant difference among undesired 3 days and 7 days post.op. (red dots) versus undesired baseline, 14, and, 30 post.op. (pale pink dots) (p=0.01). Statistically significant difference among desired 7 days post.op. (dark blue dots) versus desired baseline, 3, 14, and, 30 post.op (light blue dots) (p=0.01). PCoA (2-axis visualization) using weighted Unifrac distance (B). PERMANOVA; p<0.05. Samples dispersion’s analysis was tested by PERMDISP test.

Taxonomic identification evidenced different microbiome compositions between groups during follow-up (Figure 4). At baseline, a greater abundance of well-known periodontopathogens (e.g., Tannerella forsythia, Treponema denticola, Selenomonas noxia) was seen in UH than in DH. Moreover, some other species, such as Rothia aeria and Catonella sp._HMT_451, were more abundant in UH. In contrast, DH assessment at the same time point has shown more symbiotic commensal-rich community (e.g.,Gemella haemolysans, Dialister spp., and Granulicatella elegans).

Figure 4.

Species with differential abundance between groups by ANCOM-BC test. The bar graphic represents the fold change for each time-point comparison between groups. Only species differentially abundant in intergroup comparison were included in this graphic representation. P-values were adjusted for multiple testing (p<0.05; Holm-Bonferroni method).

On days 3, 7, and 14, an increase in the number of different species abundance between groups was seen. On day 3, microorganisms associated with periodontitis, such as Porphyromonas gingivalis, Treponema sp., Treponema socranskii, Selenomonas sp., and Fretibacterium fastidusum, were enriched in UH whereas Prevotella sp. and some oral inherited microbiomes (e.g., Gemella sanguinis and Haemophilus sp_HMT_036) were reduced. On day 7, an increase in Prevotella sp along with Selenomonas sp., Fretibacterium fastidusum, and Porphyromonas sp. was observed in the UH microbiome profile whereas a slight Treponema sp increase was observed in the DH group. Treponema sp, Veillonella sp, Treponema denticola, Treponema socranskii, Tannerella forsythia, Fretibacterium fastidusum, and Acinetobacter johnsonii are examples of microorganisms that characterize the UH microbiome profile on day 14. Conversely, DH presented a commensal-like community, mainly comprising Campylobacter concisus, Streptococcus parasanguinis clade 411, parvula, and Streptococcus anginosus. Interestingly, also with a higher wound closure at 14 days, the UH microbiome profile was still characterized by a greater abundance of Treponema socranskii and Prevotella sp along with Tannerella forsythia, Fusobacterium nucleatum_subsp._vicentii, Fusobacterium sp._HMT_203, and Acinetobacter johnsonii, as seen at days 3 and 7 of follow-up, indicating some stable microbiome in this group during the healing process.

After LEfSe analysis, Rothia aeria (ef=2.21; p=0.012) was detected as a relevant species in the UH group at baseline whereas Prevotella jejuni (ef=2.49; p=0.003), Haemophius sp_HMT_036 (ef=2.841; p=0.027), Campylobacter concisus (ef=2.22; p=0.003), and Prevotella nanceiensis (ef=2.12; p=0.029) had specific significant effects in DH (Supplementary Figure 2). No specific bacteria were detected as significant in LEfSe analysis on day 3. On the 7th day, Abiotrophia defective (ef=2.24; p=0.024) was identified as a discriminative species in the UH group whereas Alloprevotella rava (ef=2.71; p=0.123), Catonella morbi (ef=2.80; p=0.044), Ralstonia pickettii (ef=2.29; p=0.036), and Lachnoanaerobaculum orale (ef=2.00; p=0.016) were associated with DH (Supplementary Figure 1). Vellonella parvula (ef=3.02; p=0.009) and Streptococcus parasanguinis clade 411 (ef=2.81; p=0.009) were reported as significant microorganisms in DH on day 14 (Supplementary Figure 3). Lastly, Moraxella osloensis (ef=3.68; p=0.48) was featured as a relevant species associated with DH on day 30 (Supplementary Figure 2).

Immunological and microbiome biomarkers’ interaction

PCA of immunological outcomes and microbiome composition showed inflammatory biomarkers such as MMP-2, MIP-1α, FGF-2, IL-13, PDGF-BB, and TNF-α vectors along with Prevotella sp._HMT_305, Megasphaera micronuciformis, and Lachnospiraceae_[G-2]_bacterium_HMT_096 in the PCA’s first dimension on day 3, mostly correlated with DH. Porphyromonas gingivalis, Treponema sp HMT 257, as well as TIMP-1, IL-1β and IL-6 were associated with UH (Figure 5A).

Figure 5.

Inflammatory biomarkers and microbiome interplay assessment by Principal Component Analysis (PCA) on day 3 (A) and 7 postoperative (B). Blue ellipse represents DH group and red ellipse the UH group. Host response mediators and specific species vectors (e.g., Porphyromonas gingivalis, Treponema sp HMT 257, TIMP-1, IL-1β and IL-6) related to UH group on day 3 may indicate the impact of certain species on host response during wound healing.

On day 7, a more distinct immune-microbial interaction pattern could be seen between groups. TIMP-1, EGF, IL-1β, MCP-1, and PDGF were correlated with Prevotella sp._HMT_472, Prevotella sp._HMT_443, Or bacterium sp._HMT_078, and Treponema pectinovorum in dimension 1 (UH group) whereas Burkholderia cepacian, Cupriavidus gilardii, Prevotella sp._HMT_396, Ralstonia pickettii, and Aquamicrobium lusatiense made a substantial contribution in PCA dimension 2 (DH group), together with MMP-2, IL-6, FGF-2, and TNF-α (Figure 5B).

DISCUSSION

This study focused on microbiome and host response profiles associated with healing of palatal open wounds. Based on the outcomes, it seems that differences in microbiome composition and interplay between inflammatory biomarkers and bacteria can modulate and clinically impair healing.

The UH and DH group displayed a change in the beta diversity profile on days 3–7 and 7 after surgery, respectively, returning to baseline levels after complete EPT. This early shift in the UH group can be attributed to the wound’s inflammatory environment¹⁹, and a dynamic colonization²⁸ is expected in this stage with the interplay between commensal and pathogens²⁹. The interaction between EPT rates throughout healing and microbiome shift returning to baseline diversity on day 30 shows that microbiota impacts epithelium migration or delayed repair modifies the local environment, leading to microbiome change. A tendency of a distinguished microbiome was observed 7 days after surgery in desired wounds compared to undesired. In this early critical timeframe for the healing process, several variables may influence the repair, such as host immunological aspects and microbiome composition³⁰. The literature has shown that inflammation in oral mucosa can negatively impact the microbiome diversity³¹. This crosstalk may provide a harsh environment for cell activity, leading to an unexpected WC.

DH group presented greater abundance of commensal microorganisms over time. Granulicatella sp and Gemella sp are part of the oral microbiome core whereas Gemella haemolysans has been frequently identified in the buccal mucosa³². Not only is Gemella haemolysans a commensal in the microflora, but it also is capable of suppressing P.gingivalis, which may modulate the oral microbiome against harmful oral pathogens³³. Additionally, Granulicatella elegans is usually associated with a healthy microbiome in the hard palate, along with Veillonella sp³⁴. LEfSe analysis showed Prevotella nanceiensis and Haemophilus sp._HMT_036 are important species associated with DH at baseline. Both microorganisms have been identified as commensal under healthy oral conditions in the tongue dorsum and keratinized mucosa³⁵. Conversely, the UH group displayed a microbiome profile comprising microorganisms with potentially harmful action against healing.

The relative abundance of Rothia aeria was greater in UH at baseline. Despite its benefits against periodontopathogens³⁶, Rothia is known to produce acetaldehyde^37,38, which may impair DNA maintenance, leading to epithelium disfunction and high exfoliation rates. Tannerella sp, Treponema sp, Selenomonas sp, and P. gingivalis were persistently more abundant in the UH group microbiome profile throughout the early and delayed healing stages. Species from these genera have the ability of tissue invasion and trigger a pro-inflammatory host response due to interaction with TLRs and pathogen-associated molecular patterns expressed by the oral epithelium that could be impairing the healing process^39–42. In fact, P. gingivalis is capable of downregulating genes from oral epithelial cells associated with cell proliferation and migration⁴³. These species may be a factor in the delayed/undesired healing response in this group.

Even though UH displayed higher rates of bacteria with pathogenic activity on day 7, the DH microbiome profile also presented low but detectable frequency of periodontopathogens at the same time point (e.g., Treponema denticola and Treponema socranskii). Perhaps, the host inflammatory response associated with wound healing also could be similarly impacting bacterial species in both groups⁴⁴. Neutrophils, as a first line of defense, promote oxygen depletion due to their intense activity, which may elicit more anaerobic pathogens in early healing phases⁴⁵. Fretibacterium fastidiosum increased in relative abundance in UH from 3 days to 14 days after surgery. Although this microorganism is a newly identified oral pathogen⁴⁶, the literature has shown that this bacterium promotes dysbiosis and upregulates the microorganism virulence factor (e.g., flagellar system)⁴⁷. Therefore, it seems that this species may impact biofilm composition resulting in new challenges for tissue repair.

LEfSe showed that the DH community was correlated with oral health (Veillonella Parvula)⁴⁸ and disease (Streptococcus parasanguinis clade 411)³⁰ species on day 14, which would be consistent with eubiosis in the wound environment. The literature on gut microbiome reported that these two genera can metabolically interact, filling an immunomodulatory role that guarantees homeostasis⁴⁹. Conversely, the UH profile presented several microorganisms correlated with oral disease, such as periodontitis and mucosa ulcers. Acinetobacter johnsonii has been identified as a potential marker of aphthous stomatitis, inhibiting epithelial cell proliferation and promoting cytotoxicity in the mucosa⁵⁰. Lower EPT rates on day 14 in UH can be associated with these microorganisms’ characteristics.

Host response mediators’ role in tissue healing is time dependent and, based on their sustained or transient release, can either benefit or jeopardize WC. An increase in IL-6, MCP-1, and MIP-1α was observed in UH over time. Although these biomarkers are important to start cell recruitment and provide a response to injury, their persistent production may not be beneficial⁵¹. The literature has shown that MCP-1 depletion is correlated with macrophage phenotype (M2)^52,53, which is an important step in pro-resolution and tissue healing. It seems reasonable to infer that greater expression of MCP-1 in UH may have contributed to delayed EPT and WC outcomes. DH displayed high levels of MIP-1α and MMP-2 3–7 days after surgery. Studies reported MMP-2 activity in epithelium migration and its important role in tissue EPT after its evaluation in knockout mice^54,55. MIP-1α depletion in murine wound models provided evidence of this chemokine’s role targeting important steps in tissue repair, such as angiogenic activity and collagen deposition^56,57. PCA showed that MMP-2 and MIP-1α were associated with bacteria highly abundant in DH on day 3 whereas MCP-1 made a considerable contribution along with bacterial abundance related to UH on day 7. These correlations provide an insight into a potential interplay between these host and microbiome factors during healing.

Caution should be exercised with the outcomes presented in this study. This investigation is a pilot study with a limited sample size, which may impact the gathering of new information regarding microbiome and immunological differences between groups. Different types of samples, i.e., saliva, may add more information in the repair process considering proteins, metabolism products, and microorganisms. Metagenomics assessment could provide more informative data regarding other oral microorganisms (e.g., viruses, fungi) that may enhance or impair the oral healing process. Future studies validating the beneficial and detrimental role of some of the bacterial species detected in DH and UH are warranted. Identification of beneficial oral bacterial species and host factors could provide new potential therapeutic alternatives to improve oral healing. Moreover, further clinical studies should be carried out also recording patient centered parameters to correlate evidence regarding number of analgesic and discomfort with these molecular and microbiological findings. Finally, a clear definition regarding the parameters used to define desired and undesired healing have to be established.

CONCLUSIONS

With this study’s limits, it can be suggested that a distinguished oral microbiome bacterial species and host inflammatory proteins characterize desired and undesired palatal-wound healing. Caution should be exercised due to the limited sample size of this pilot study.

Supplementary Material

Supinfo1

Supplementary Figure 1. Representative image of epithelialization analysis using a disclosure solution.

Supinfo2

Supplementary Figure 2. Microbiome composition (LEfSe analysis) between DH and UD at baseline time-point (A) and 7 days (B). The bar graphic represents the fold change (>2 and p-value≤0.05) for each group (Red – UH group; Blue- DH group).

Supinfo3

Supplementary Figure 3. Microbiome composition (LEfSe analysis) between DH and UD on day 14 (A) and 30 (B). The bar graphic represents the fold change (>2 and p-value≤0.05) for each group (Red – UH group; Blue- DH group).

CLINICAL RELEVANCE.

Background.

Tissue repair can be influenced by systemic and local factors, host response characteristics, and bacterial composition that may interplay during this process. Focusing on new insights for healing improvement, this pilot study evaluated the bacterial and inflammatory profiles in patients with either normal/desired (DH) or bad/undesired healing (UD) of open wounds in the palatal mucosa.

Added Value of the Study.

Bacterial species related to disease conditions were more frequent in patients with undesired healing until late healing phases. Moreover, differences in the inflammatory process were also observed and more pro-inflammation markers were observed in patients with slow healing process.

Clinical Implications.

Identifying oral bacterial species and host factors that can negatively influence the healing process may provide new potential therapeutic alternatives to improve tissue repair.

Data availability statement:

Raw sequencies are available in SRA. Clinical data that support the findings of this study are available from the corresponding author upon request.