DPSCs are a type of mesenchymal stem cell.37 To obtain these cells, healthy teeth were collected from adult donors aged 18–35 years at the Dental Clinic of Beijing Stomatological Hospital, following written informed consent for participation in this study. DPSCs were extracted from fresh, caries-free, pulpitis-free, and non-necrotic third molars or from intact permanent teeth extracted for orthodontic purposes.38 All extracted teeth had intact roots. Donors with infections caused by hepatitis B virus, hepatitis C virus, human immunodeficiency virus, syphilis, cytomegalovirus, Epstein–Barr virus, or human T-lymphotropic virus were excluded from the final donor pool. (For detailed protocols on DPSC preparation, see the Supplementary Fig. 1 in Supplementary Materials. The donor informed consent forms can be found in the Supplementary Materials).

To evaluate the safety and efficacy of DPSC injection, we investigated whether this technique, when combined with standard periodontal therapy, could safely and effectively promote periodontal tissue regeneration in humans. Two randomized, placebo-controlled clinical trials were conducted at separate clinical centers using consistent protocols for medication, injection method (periodontal tissue injection), inclusion and exclusion criteria, and outcome assessments.

The first trial was an investigator-initiated, randomized controlled study designed to evaluate the safety and efficacy of single-dose DPSC therapy, conducted at Beijing Stomatological Hospital from May 22, 2020, to February 16, 2023 (Ethical code: CMUSH-IRB-KJ-PJ-2019-14). The second trial, conducted at Peking University Third Hospital (Ethical code: D2020188), was a Phase I clinical trial—a randomized, double-blind, controlled study aimed at assessing the safety and preliminary efficacy of DPSC therapy at varying dose concentrations, carried out from July 31, 2021, to December 24, 2022.

Patients aged 18–65 years with chronic periodontitis (probing pocket depth of 4–8 mm) were eligible to participate. However, individuals with unstable blood pressure; systemic diseases (e.g., cancer, diabetes, heart disease, recent myocardial infarction within 6 months, recent symptoms of angina pectoris, or congenital heart disease); systemic infections; prior surgical treatment near the affected tooth; or current smoking habits exceeding 10 cigarettes per day were excluded.12 (For further details, see Table 1).

In the initial treatment stage, periodontal infections in all patients were managed using oral hygiene instruction and mechanical debridement. When necessary, antiseptics were applied to disinfect the area.39 Mechanical debridement involved the removal of plaque, calculus, and damaged tissue. During debridement, the root surfaces and bone defects were carefully scaled to remove residual mineralized deposits while preserving the root cementum. Bone defects were assessed using a needle injection immediately after scaling and root planning. Following these initial treatments, 0.6 mL of DPSCs were injected at the root surface to overfill the defect prior to the start of the experimental procedure (Fig. 3).

A total of 132 participants were enrolled across both trials: 96 patients in the investigator-initiated trial and 36 patients in the Phase I trial. In both studies, either DPSCs (experimental group) or saline (control group) was injected directly into the local periodontal bone defects immediately after deep cleaning, including scaling and root planning. The control procedure was identical to the experimental procedure, except for the omission of DPSCs (Fig. 3). A flow chart outlining the step-by-step sampling design for both trials is provided in Fig. 1.

In the investigator-initiated trial, patients underwent two stages of randomization using the block randomization method, with random group codes generated by SAS software version 9.4. In the first stage, participants were randomly assigned in a 1:1 ratio to receive either DPSC injection or saline injection based on their assigned codes. In the second stage, sealed envelopes were used to further allocate participants into one of three groups: a single DPSC injection (1 × 10⁷ cells/injection), a double DPSC injection (1 × 10⁷ cells/injection, administered twice at a 30-day interval), or a saline injection. Allocation was conducted in a 1:1:1 ratio according to the randomization codes, resulting in 33 patients in the saline group, 33 in the single-injection group, and 30 in the double-injection group.

The investigator-initiated trial adopted a complete randomization method. Randomization tables and the corresponding groups of the randomization tables were generated using SAS software version 9.4. And randomization numbers were assigned using the Interactive Web Response System (IWRS). Each group was randomly assigned separately. The subjects in each group were randomly assigned to the experimental group or the control group at a ratio of 3:1. Each subject was assigned a random number from small to large according to the screening number on Day 1 and entered the corresponding group.

For the sample size calculation of Phase I trial, the objective was to assess safety; therefore, the sample size was not determined based on statistical assumptions and was not formally calculated. A total of 36 participants were included, and each participant was randomly assigned to a distinct dose group, receiving a single dose of DPSC injection. The sample sizes for each group were as follows: nine individuals received no cells for one tooth (saline injection group); three individuals received 1 × 106 cells for one tooth; six individuals received 5 × 106 cells for one tooth; six individuals received 1 × 107 cells for one tooth; six individuals received 2 × 107 cells for two teeth; and six individuals received 3–4 × 107 cells for three or four teeth. In each dose group, controls were randomly selected. Four participants were enrolled in the 1 × 106 cells/one tooth group and randomly assigned to the experimental group (three cases) or the saline injection group (one case), while for the remaining four dose groups, eight participants were enrolled in each group and randomly assigned to the experimental group (six cases) or the saline injection group (two cases).

For the investigator-initiated trial, the unit of statistical analysis was the tooth. Based on historical data, attachment loss improved by 2.0 mm 6 months after surgery alone, while improvement reached 3.4 mm following DPSC injection, with a standard deviation of 1.7 mm. The single- and double-injection DPSC groups were compared with the saline group using a significance level of α = 0.025. Assuming a 1:1:1 group allocation ratio, a sample size of 29 teeth per group was required to achieve 80% power to detect between-group differences. Accounting for a 10% loss to follow-up, 32 teeth per group were needed.12,40

Clinical safety and efficacy evaluations were conducted over a 6-month follow-up period (180 ± 14 days).41 All adverse events were documented. Visit details are provided in Supplementary Table 2. Following treatment, all participants were monitored in the hospital for 24 h to assess clinical safety prior to discharge. The primary safety outcome was the occurrence of any serious adverse events.

To evaluate the efficacy of DPSC injection, both soft and hard tissue prognostic indicators of periodontitis were assessed at each follow-up visit through clinical examination.42 The primary efficacy outcome was the level of tooth AL at 6 months post-treatment. Secondary efficacy outcomes included periodontal probing depth (PD), gingival recession (GR), tooth mobility (TM), bleeding on probing (BOP), and bone defect depth (BDD) at 6 months post-treatment. BDD was evaluated based on cone-beam computed tomography (CBCT) imaging, and the specific measurement method is illustrated in Fig. 4. Briefly, the bone defect depth of the target tooth was measured in the CBCT data. The observation line was adjusted in the axial position to be consistent with the mesial-distal direction of the tooth. The observation line was adjusted in the oblique sagittal position to be consistent with the long axis direction of the tooth. Then the key data were measured respectively at the buccal 1/3, median and lingual 1/3 layer of the target tooth, including A: The distance from the cementoenamel junction (CEJ) to the lowest point of alveolar bone defect, B: The distance from the CEJ to the vertex of the alveolar crest. The BDD (C) in this layer was calculated according to such a formula: C = A − B. And the average values measured in all layers indicated the BDD of this tooth.

Schematic illustration of the BDD measurement based on CBCT results. a, b The analysis of each target tooth was accomplished based on a comprehensive assessment in three layers, including the buccal 1/3, median, and lingual 1/3 layer. c In each layer, the BDD (C) was calculated according to the formula: C = A – B. A represents the distance from the cementoenamel junction (CEJ) to the lowest point of alveolar bone defect; B represents the distance from the CEJ to the vertex of the alveolar crest. The final analyzed BDD of this target tooth was the average of the measured values in three layers (buccal 1/3, median, and lingual 1/3 layer). Figure created with BioRender.com. CEJ cementoenamel junction

For the separate analysis of results from the investigator-initiated study and the Phase I trial, the database was locked and loaded into SAS version 9.4 (Statistical Analysis Software; SAS Institute, Cary, NC, USA). All statistical tests were conducted on a two-sided basis (unless otherwise specified), and a P value of less than 0.05 was considered statistically significant.39 Statistical significance (P < 0.05) was determined by Student’s t-test.

In addition to the separate analyses, a combined analysis was conducted to assess the response to DPSC injection in patients with varying degrees of periodontitis, as well as to evaluate the efficacy of the 1 × 107 dose of DPSC injection. The combined treatment group consisted of teeth that received the same injection dose (1 × 107 cells/0.6 mL) in both trials: 33 teeth from the investigator-initiated study and 18 teeth from the Phase I trial. Similarly, data from the saline injection groups in both trials were pooled, comprising 29 teeth from the investigator-initiated study and 34 teeth from the Phase I trial. Based on baseline AL levels, patients were categorized into two groups: stage II periodontitis (AL < 5 mm) and stage III periodontitis (AL ≥ 5 mm). In addition to AL, all other efficacy outcomes were compared between these two periodontitis stages. The combined dataset was analyzed using SAS version 9.4, employing the same statistical methods as those used in the separate analyses. Statistical significance (P < 0.05) was determined by Student’s t-test.

Similarly, a combined analysis was conducted to assess the response to DPSC injection in teeth with single root or multiple roots. According to the position of the treated teeth, the anterior teeth and anterior molars are classified as single-root teeth, while the molars are classified as multiple-root teeth. Multiple-root teeth were further divided into two subgroups: those with furcation involvement and those without, based on their clinical examination results before the treatments. All efficacy outcomes were assessed on the day 180 post-treatment, and the combined dataset was analyzed using SAS version 9.4, employing the same statistical methods as those used in the separate analyses. Statistical significance (P < 0.05) was determined by Student’s t-test.