Intra-Articular Injection of Human Bone Marrow–Derived Mesenchymal Stem Cells in Knee Osteoarthritis: A Randomized, Double-Blind, Controlled Trial

Abstract

To assess the impact of a single intra-articular (IA) injection of bone marrow–derived mesenchymal stem cells (BM-MSCs) in patients with knee osteoarthritis (OA), a randomized, double-blind, placebo-controlled study was conducted. The study included 24 patients with knee OA who were randomly assigned to receive either a single IA injection of BM-MSCs or normal saline. Changes in the visual analog scale (VAS), Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), and Knee Injury and Osteoarthritis Outcome Score (KOOS) after IA injection were assessed at 3, 6, 9, and 12 months. Magnetic resonance imaging (MRI) with T2 mapping sequences was conducted for knee cartilage assessment at baseline and at 3 and 12 months. The MSC group showed between-group improvement in WOMAC (–5.0 ± 3.6 vs. –0.1 ± 5.5, P = 0.02) and KOOS (23.9 ± 18.3 vs. 7.2 ± 15.9, P = 0.028) scores at 9 months compared with the control group. The MSC group exhibited a less sharp increase in the mean T2 value of the medial compartment than the control group at 12 months, with no serious adverse events observed during follow-up. A single IA injection of allogeneic BM-MSCs provided satisfactory pain relief for patients with knee OA compared with the control group at 9 months. Quantitative T2 MRI mapping of the cartilage showed that IA BM-MSCs could have a preventive effect on OA progression for 12 months. Our findings suggest the potential of allogeneic BM-MSCs IA injection as a pain-relieving and disease-modifying treatment for patients with knee OA in the outpatient setting.

Graphical Abstract.

Introduction

Knee osteoarthritis (OA) is a complex degenerative articular disorder characterized by the degeneration of articular cartilage, resulting in damage to the articular tissue, which consequently leads to impaired joint function. Because of its avascular nature and the limited self-renewal capacity of chondrocytes, adult articular cartilage presents limited repair capability, and the optimal treatment of knee OA remains unknown ¹ . Most conventional pharmacological therapies concentrate on relieving pain in the knee joint. For severe cases of knee OA, total knee arthroplasty (TKA) is the only therapeutic option ² . However, TKA is associated with the risk of general anesthesia and postoperative (surgery-related) complications. Therefore, identifying safe and effective alternative treatments for OA is essential.

Intra-articular (IA) injection into the knee joint may be suitable for individuals who are unable to take oral medications or wish to avoid surgical treatment. IA injections primarily provide short-term symptom relief and reduce pain and function. Molecules such as corticosteroids, hyaluronic acid (HA), and platelet-rich plasma (PRP) have been used as IA injection treatments ³ . However, these treatments currently have major limitations in terms of timing, frequency, and efficacy. Therefore, further studies are required to investigate new molecules and/or therapies associated with IA injections for knee OA.

Mesenchymal stem cells (MSCs), recognized for their extensive proliferative ability, exhibit immunomodulatory properties associated with their potential to facilitate cartilage repair and improve knee function. While numerous clinical trials have shown that using MSCs for treating knee OA is safe and effective, questions remain regarding its effectiveness, how it works, the best timing for injections, and the proper dosage of MSCs ⁴ .

Several studies have compared the effects of MSCs with those of a placebo or HA. Some of these studies found better overall functional improvement and magnetic resonance imaging (MRI)-identified cartilage quality improvement in the MSC group^5–7. Recently, several MRI techniques such as T2 mapping, diffusion-weighted imaging (DWI), gagCEST, and dGEMRIC have been used to perform compositional imaging of the cartilage and selectively visualize the glycosaminoglycan (GAG) and collagen fiber network of the extracellular matrix (ECM). However, studies on the IA injection of MSCs in OA using these MRI techniques are lacking ⁸ .

This double-blind, randomized, controlled clinical study examined the effect of a single IA injection of human allogeneic bone marrow (BM)-derived MSCs compared with that of normal saline (placebo) in knee OA. In this study, we evaluated the safety of IA injection of MSCs into the knee and the effects of symptom relief on knee OA. We also attempted to clarify the regenerative potential of MSCs by reporting pre- and post-injection MRI T2 quantitative mapping data. We hypothesized that pain relief in MSC-treated knees would continue for several months through the prevention of cartilage damage detected by MRI.

Materials and Methods

Study Design

This single-center, randomized, controlled clinical trial was conducted between November 2019 and June 2022 at Seoul St. Mary’s Hospital (Seoul, Korea). Random allocation with a computer-assisted method was based on a 1:1 ratio between the intervention and control groups. The study used a permuted-block randomization method to ensure balanced allocation of participants across treatment groups, with blocks of varying sizes randomly assigned to minimize predictability. This approach helps maintain an equal distribution of participants throughout the study. The process of randomization and allocation was conducted by medical personnel who were not involved in any other aspects of the study. Throughout the study, neither the participants nor investigators were aware of the group allocations. The ClinicalTrials.gov identifier used was NCT04240873. This study was approved by the Institutional Review Board of Seoul St. Mary’s Hospital of the Catholic University of Korea (KC19CNSI0469). All participants provided written informed consent.

Participants

Patients diagnosed with primary knee OA, based on the American College of Rheumatology Classification Criteria, were selected for participation ⁹ . Patients were eligible if they (a) were between 20–80 years of age; (b) exhibited Kellgren–Lawrence (K–L) Grade I–IV confirmed via radiography; (c) had a visual analog scale (VAS) score exceeding 50 mm at baseline; (d) had not seen improvement after at least 3 months of non-surgical treatment; and (e) voluntarily provided informed consent for enrollment. Individuals were excluded if they (a) had received IA injections of corticosteroids or HA in the past 3 months; (b) could not ambulate; or (c) were concurrently using anti-inflammatory medications. The study sample size was determined to achieve a statistical significance threshold (α) of 0.05 and a power of 0.80.

Preparation of BM-MSCs

Human BM-MSCs, Catholic MASTER cells, were sourced from the Catholic Institute of Cell Therapy (CIC) in Seoul, Korea. Bone marrow (BM) was obtained from healthy individuals between the ages of 20 and 55 years via aspiration from the iliac crest, with the approval of the Institutional Review Board of Seoul St. Mary’s Hospital. A single donor’s BM was selected from various BM donations and sent to the CIC’s facility which adheres to Good Manufacturing Practice (GMP) standards for the isolation, growth, and quality assurance processes of BM-MSCs. The bone marrow was centrifuged at 4°C and 793 × g for 7 min to obtain the marrow pellet. After centrifugation, sterile distilled water was applied in a 10:1 ratio to eliminate red blood cells (RBCs). The RBC-depleted marrow pellets were then resuspended in an MSC growth medium provided by GE Healthcare (Seoul, Korea), enriched with 20% fetal bovine serum (also from GE Healthcare), and cultured in T-75 tissue culture flasks (Nunc, Rochester, NY, USA) at 37°C in a 5% CO₂ incubator. This growth medium was consistently used throughout cell expansion, with medium changes occurring biweekly. Once cell confluence reached approximately 70% to 90%, the cells were harvested and seeded at a density of 5–8 × 10³ cells/cm² for further expansion. This expansion occurred over 2–4 passages in a GMP-certified environment. Throughout the cell expansion phase, tests for mycoplasma and bacterial sterility, along with endotoxin levels (below 3 EU/ml), were conducted. Post-fourth passage, the cells were assessed for surface marker expression (CD90/CD73, over 95% positive; CD34/CD45, over 95% negative) and their ability to differentiate into multiple cell types.

Treatment

Twenty-six patients were recruited, and 24 patients were randomized (Fig. 1). An IA injection of MSCs (1 × 10⁸ cells/2 ml) or placebo (normal saline, 2 ml) was administered. After positioning the patient supine with a pillow under the knee to achieve approximately 30° of joint flexion, thorough antiseptic measures were applied. A high-frequency linear ultrasonography probe was used to scan the suprapatellar recess of the knees. Next, ultrasound-guided knee joint injection with a 20-gauge needle was performed by a rheumatologist (J.J.L.) who was skilled in knee injection.

Figure 1.

Flow chart of patient enrollment and follow-up.

MSC: mesenchymal stem cell; MRI: magnetic resonance imaging.

Primary Outcomes

The primary outcomes were changes in patient-reported outcome measures (PROMs). Pain and knee joint function outcomes were assessed using a VAS of 0–100 mm, the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), and the Knee Injury and Osteoarthritis Outcome Score (KOOS). The WOMAC score includes 24 items of pain (0–20), stiffness (0–8), and physical function (0–68), with a total score of 96 ⁹ . The KOOS is a self-administered questionnaire used to evaluate participants’ opinions regarding their knee problems through five specific subscales: pain, symptoms, activities of daily living (ADL), sport and recreation, and quality of life (QOL) ¹⁰ . A single evaluator conducted assessments of all primary outcomes at baseline and again at 3, 6, 9, and 12 months.

Secondary Outcomes

Quantitative T2 mapping of the articular cartilage along with standard knee MRI sequences was performed at baseline and at 3 and 12 months. MR was performed using a 3 T scanner (Magnetom Vida; Siemens Healthcare, Erlangen, Germany). The imaging protocol consisted of intermediate-weighted two-dimensional turbo spin-echo (TSE) sequences with three orthogonal planes, a three-dimensional proton density-weighted TSE, and a multiecho spin-echo T2-weighted sequence for T2 mapping with the following parameters: field of view, 150 × 150 mm; pixel matrix, 384 × 230; repetition time ms/echo time ms, 3170/11.5, 23, 34.5, 46, 57.5, 69, 80.5, and 92; bandwidth, 241 Hz/pixel; and total acquisition time, 4 min 35 s. T2 maps were generated using commercial software (MapIt; Siemens Healthcare). Increased T2 values indicated a disorganized collagen microstructure and a resultant increase in the water content of the cartilage. Image registration among the three time points and region-of-interest (ROI) placement were performed using the ITK-SNAP software, version 3.8.0 (open source, http://www.itksnap.org/) ¹¹ . Cartilages were divided into 14 subregions based on the MRI Osteoarthritis Knee Score (MOAKS) system ¹² . The registered MR images from the three time points were uploaded side-by-side, and ROIs were manually located at the same location among the three time points by one musculoskeletal radiologist (J.Y.J.) with 13 years of clinical experience. Measurements were only performed when the cartilage retained more than 50% of its native thickness. After the measurements were repeated three times, all T2 values were recorded as representative and independent values for each subregion at each time point. Subregions were aggregated to establish three distinct compartments: the patellofemoral, incorporating both the patella and trochlea; the medial, composed of the medial femur’s central and posterior segments, alongside the medial tibia’s anterior, central, and posterior sections; and the lateral, consisting of the lateral femur’s central and posterior parts, in addition to the lateral tibia’s anterior, central, and posterior areas ¹³ .

For biochemical analyses, fasted venous blood samples were collected in the morning at the beginning of the study and after 3 and 12 months of treatment. The bone turnover markers osteocalcin (OC) and carboxy-terminal telopeptide of type I collagen (CTX-I) were measured along with inflammatory markers (C-reactive protein and erythrocyte sedimentation rate).

Safety Assessments

During the monitoring phase, the occurrence of adverse events (AEs) in the index knee, including heating sensation, redness, tenderness, joint effusion, and limitations in range of motion, were recorded. Patients were advised against using any prescribed or non-prescribed analgesics if they experienced no pain and were instructed to record any emergency medication utilized, including its use period and frequency. The safety assessment included vital signs, laboratory tests, and physical examination findings. AEs were categorized as serious, treatment-related, or discontinuation owing to AEs. The severity of AEs was assessed using the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI-CTCAE) ¹⁴ .

Statistical Analysis

Statistical analysis was performed using GraphPad Prism (GraphPad Software, Inc. La Jolla, CA, USA) and R version 3.1.0 (R Foundation for Statistical Computing, Vienna, Austria). The evaluation was conducted using an intention-to-treat approach, incorporating all data from participants randomly assigned to their respective groups. A significance threshold (P-value) was established at 0.05 to test all hypotheses. For categorical data, both absolute (n) and relative (%) frequencies were calculated. Normally distributed quantitative data are reported as mean values along with their standard deviations, and 95% confidence intervals (CIs) for these means were also derived. Fisher’s exact test was used to assess the relationship between baseline categorical variables. The distribution normality of quantitative data was assessed using the Kolmogorov–Smirnov and Shapiro–Wilk tests. Changes from baseline to each subsequent time point within groups were described using means and 95% CIs. For comparative analyses, we utilized the unpaired Student’s t-test, the Mann–Whitney U test, the paired Student’s t-test, and the Wilcoxon matched-pairs signed-ranks test, depending on the context.

Results

The study adhered to the Consolidated Standards of Reporting Trials (CONSORT) statement, and a flow diagram is shown in Fig. 1. After screening 26 patients, 24 patients were included in the study and randomized into one of the two study groups. During the follow-up period, 1 participant withdrew from the trial, and 23 participants, including 12 in the control group and 11 in the MSC group, completed the 12-month sessions and underwent pre-post assessments.

Participant Characteristics

The basic demographic data of the randomized patients are presented in Table 1. The study population was predominantly female (83.3%, 20 of 24 patients), with a mean age of 67.5 years (standard deviation, 7.5 years; 71.2 ± 5.7 years in the MSC group and 63.8 ± 8.6 years in the control group). Baseline characteristics were well balanced between the groups in terms of sex, K–L grade, body mass index (BMI), and knee index. The MSC group showed poorer results in several PROMs of knee OA, including total, pain, and physical function scores of the WOMAC and sport/reaction scores of the KOOS.

Table 1.

Demographic and Baseline Characteristics of Both Groups.

Variables	Total (n = 24)	Control group (n = 12)	MSC group (n = 12)
Age, years	67.5 ± 8.1	63.8 ± 8.6	71.2 ± 5.7
Sex, female (%)	20 (83.3%)	12 (100%)	8 (66.7%)
BMI, kg/m2	25.0 ± 3.2	24.1 ± 3.0	25.8 ± 3.2
Index knee, no. (%)
Right	12 (50.0%)	6 (50.0%)	6 (50.0%)
Left	12 (50.0%)	6 (50.0%)	6 (50.0%)
K-L grade, no. (%)
II	9 (37.5%)	5 (41.7%)	4 (33.3%)
III	12 (50.0%)	6 (50%)	6 (50%)
IV	3 (12.5%)	1 (8.3%)	2 (16.7%)
Pain VAS (mm)	72.3 ± 9.5	70.8 ± 5.0	73.8 ± 12.6
WOMAC, total	53.3 ± 20.0	44.8 ± 15.1	61.9 ± 21.1
WOMAC, pain	10.7 ± 4.5	8.8 ± 4.2	12.5 ± 4.3
WOMAC, stiffness	4.9 ± 2.0	4.7 ± 1.6	5.2 ± 2.3
WOMAC, physical function	37.8 ± 14.4	31.2 ± 10.8	44.2 ± 14.9
KOOS, pain	42.7 ± 18.3	49.1 ± 16.3	36.3 ± 18.5
KOOS, symptom	45.8 ± 19.0	48.5 ± 17.7	43.2 ± 20.7
KOOS, activities of daily living	50.5 ± 19.1	57.5 ± 15.9	43.5 ± 20.2
KOOS, sport/recreation	16.9 ± 16.7	23.8 ± 18.8	10.0 ± 11.1
KOOS, quality of life	33.1 ± 9.5	34.4 ± 9.0	31.8 ± 10.1

KOOS: Knee Injury and Osteoarthritis Outcome Score; MSCs: mesenchymal stem cells; WOMAC: Western Ontario and McMaster Universities Osteoarthritis Index; VAS: visual analog scale.

Patient-Reported Outcome Measures

The mean differences in PROMs between baseline and each visit for each group are reported in Table 2. In the MSC group, significant improvements in WOMAC and KOOS pain scores were observed between baseline and most of the time points. However, the WOMAC and KOOS pain scores in the control group showed no significant improvement from baseline at any time point. In comparison of difference from baseline at each time point between the groups, compared with the control, MSC treatment was associated with a significantly greater improvement in the mean value of difference from baseline for WOMAC (–5.0 vs. –0.1 points; P = 0.02) and KOOS (+23.9 vs. +7.2 points; P = 0.028) pain scores at 9 months (Table 2). Figure 2 illustrates the comparison of changes in all patient-reported outcomes between the control and MSC groups from baseline to follow-up visits.

Table 2.

Difference Between Values in Each Patient-Reported Outcome From the Baseline to Each Follow-Up Visit and Results for Between-Group Differences.

Patient-report outcomes	Control group	MSC group	Between-group analysis
	MD (95% CI) a	MD (95% CI) a	P-value b
Pain VAS score (mm)
3 months	−14.1 (–25.9 to −2.3)	−17.9 (–31.4 to −4.4)	0.640
6 months	−19.8 (–31.2 to −4.4)	−20.5 (–31.2 to −8.3)	0.915
9 months	−12.8 (–23.1 to −2.6)	−19.8 (–35.6 to −4.0)	0.412
12 months	−14.7 (–26.3 to −3.0)	−21.2 (–37.2 to −5.1)	0.469
WOMAC, total
3 months	−8.1 (–20.7 to 4.6)	−21.7 (–36.4 to −7.0)	0.132
6 months	−10.8 (–22.0 to 0.4)	−23.1 (–39.7 to −6.4)	0.183
9 months	−4.8 (–16.8 to 7.1)	−19.5 (–33.5 to −5.5)	0.091
12 months	−7.1 (–21.5 to 7.3)	−17.9 (–35.5 to −0.3)	0.301
WOMAC, pain
3 months	−1.6 (–4.5 to 1.3)	−5.0 (–5.1 to 0.6)	0.078
6 months	−2.3 (–5.1 to 0.6)	−4.9 (–7.9 to −1.9)	0.172
9 months	−0.1 (–3.6 to 3.4)	−5.0 (–7.4 to −2.6)	0.02
12 months	−0.6 (–4.1 to 2.9)	−4.3 (–7.8 to −0.8)	0.113
WOMAC, stiffness
3 months	−1.6 (–2.6 to −0.6)	−1.6 (–2.8 to −0.5)	1.000
6 months	−1.8 (–3.0 to −0.6)	−1.5 (–3.0 to 0.1)	0.676
9 months	−1.3 (–2.7 to 0.2)	−2.1 (–3.3 to −0.9)	0.345
12 months	−1.3 (–2.8 to 0.3)	−1.5 (–3.0 to −0.1)	0.759
WOMAC, physical function
3 months	−4.9 (–14.1 to 4.3)	−15.1 (–26.2 to −3.9)	0.131
6 months	−6.8 (–14.6 to 1.1)	−16.7 (29.2 to −4.2)	0.141
9 months	−3.5 (–11.1 to 4.1)	−12.4 (–23.2 to −1.5)	0.148
12 months	−5.3 (–15.1 to 4.6)	−12.1 (–25.1 to 0.9)	0.358
KOOS, pain
3 months	11.6 (–1.1 to 24.2)	27.1 (15.5 to 38.7)	0.060
6 months	13.2 (–1.7 to 28.2)	29.5 (16.5 to 42.6)	0.086
9 months	7.2 (–2.9 to 17.3)	23.9 (11.7 to 36.2)	0.028
12 months	12.0 (–2.0 to 26.1)	23.2 (6.5 to 39.9)	0.266
KOOS, symptom
3 months	10.4 (–2.3 to 23.1)	23.4 (8.3 to 38.5)	0.158
6 months	14.9 (–0.9 to 30.7)	21.4 (3.5 to 39.3)	0.548
9 months	12.5 (0.9 to 24.1)	19.8 (4.2 to 35.4)	0.409
12 months	15.5 (1.2 to 29.8)	24.7 (8.8 to 40.6)	0.350
KOOS, activities of daily living
3 months	10.0 (–0.5 to 20.6)	16.0 (5.9 to 26.2)	0.378
6 months	10.9 (–1.6 to 23.4)	22.1 (10.3 to 33.8)	0.168
9 months	7.5 (–2.7 to 17.6)	14.0 (4.8 to 23.3)	0.304
12 months	10.0 (–0.6 to 20.7)	15.0 (7.0 to 22.9)	0.430
KOOS, sport/recreation
3 months	10.8 (–1.0 to 22.6)	27.3 (11.0 to 43.5)	0.081
6 months	15.0 (4.8 to 25.2)	28.6 (8.5 to 48.7)	0.183
9 months	12.1 (1.6 to 22.5)	27.3 (9.4 to 45.1)	0.111
12 months	15.8 (4.9 to 26.7)	25.5 (7.0 to 43.9)	0.320
KOOS, quality of life
3 months	7.8 (0.6 to 15.0)	13.1 (–1.0 to 27.2)	0.458
6 months	11.5 (1.8 to 21.2)	14.2 (3.6 to 24.8)	0.676
9 months	9.4 (2.9 to 15.8)	15.9 (6.6 to 25.2)	0.207
12 months	12.0 (0.7 to 23.3)	11.9 (1.1 to 22.8)	0.995

^aObserved intra-group mean difference and 95% CI, unadjusted values (baseline as reference).

^bP-value for inter-group comparisons of difference from baseline.

MD: mean difference; CI: confidence interval; KOOS: Knee Injury and Osteoarthritis Outcome Score; MSCs: mesenchymal stem cells; WOMAC: Western Ontario and McMaster Universities Osteoarthritis Index; VAS: visual analog scale.

Figure 2.

Comparison of changes in patient-reported outcomes between groups from baseline to follow-up visits. (a-e) Changes in WOMAC scores; (f-j) Changes in KOOS scores.

KOOS: Knee Injury and Osteoarthritis Outcome Score; MSC, mesenchymal stem cell; WOMAC: Western Ontario and McMaster Universities Osteoarthritis Index; VAS: visual analog scale.

Inflammatory and Bone Turnover Markers

Changes in inflammatory and bone turnover markers in both groups between baseline and 3 and 12 months are shown in Fig. 3. The mean ± standard deviation of change in CTX-I levels from baseline for the MSC group was −0.1 ± 0.13 at 3 months, increasing further to 0.05 ± 0.18 at 12 months. The difference in CTX-I levels between the control and MSC groups was statistically significant only at 3 months (P = 0.049). In contrast, there were no significant differences in the mean changes in other inflammatory and bone turnover markers between the two groups.

Figure 3.

Comparison of the mean differences and standard deviations from baseline for inflammatory (A, ESR; B, CRP) and bone turnover markers (C, Osteocalcin; D, CTX-I) at each time point between the two groups (*P < 0.05).

CRP: C-reactive protein; CTX-I: carboxy-terminal telopeptide of type I collagen; ESR: erythrocyte sedimentation rate.

Quantitative T2 Cartilage Mapping MRI

Figure 4 details the change of T2 values using quantitative T2 cartilage mapping MRI at baseline and 3 and 12 months following placebo or MSC IA injection in three compartments of the knee cartilage (medial, lateral, and patellofemoral). Upon examining the changes in T2 values within the medial compartment over the study period, distinct patterns emerged between the control and MSC groups. Specifically, the mean T2 values (ms) increased progressively over time, starting from a baseline mean of 38.83 (95% CI: 36.87, 40.79). This upward trajectory continued at 3 months, with a mean T2 value of 41.62 (95% CI: 39.62, 43.61), culminating in a further increase to 44.04 (95% CI: 41.97, 46.11) at 12 months. Conversely, the MSC group demonstrated a slightly different pattern. Starting from a baseline mean T2 value of 40.04 (95% CI: 37.53, 42.55), there was a modest increase to 40.91 (95% CI: 38.46, 43.36) at 3 months. At the 12-month interval, the mean T2 value for the MSC group was 41.81 (95% CI: 39.31, 44.31), indicating a less pronounced increase over time compared with the control group. At 12 months, the MSC group showed a moderate increase in T2 values, which was statistically significant (P = 0.013) compared with that of the control group. In contrast, as observed in Fig. 4, unlike the medial compartment, no statistical significance was found in the lateral (P = 0.364) and patellofemoral (P = 0.189) compartments. Figure 5 presents representative images of quantitative T2 mapping, showing increased T2 values in the articular cartilage of the medial femoral condyle over 12 months in a control group patient, while the MSC group patient exhibits only a mild increase.

Figure 4.

Longitudinal analysis of T2 values in three compartments (A, medial; B, lateral; C, patellofemoral) of knee cartilage across MSC and control groups. Each line represents the mean T2 value for each group. The shaded areas delineate the 95% confidence intervals.

MSC: mesenchymal stem cell.

Figure 5.

Representative images of quantitative T2 mapping of articular cartilage from control and MSC groups over 12-month follow-up.

(A-D) Proton density–weighted images (A) and corresponding T2 maps at baseline (B), 3 months (C), and 12 months (D) from a patient in the control group. The articular cartilage of the medial femoral condyle shows an increase in T2 values over time. (E-H) Proton density–weighted images (E) and corresponding T2 maps at baseline (F), 3 months (G), and 12 months (H) from a patient in the MSC group. The articular cartilage of the medial femoral condyle shows only a mild increase in T2 values over time.

Safety

As shown in Table 3, AEs occurred in nine patients (75.0%) in the MSC group. All were MSC treatment-related AEs, including post-injection pain or effusion within 1 month. Three patients experienced mild pain in the index knee, which was relieved without any medical treatment. Grade 2 AEs occurred in four patients with decreased joint function in the MSC group, and these patients were treated with only oral medication. Grade 3 AEs requiring IA corticosteroid injection and interfering with patients’ ADLs occurred in three patients in the MSC group. Five patients in the control group experienced skin rashes, upper respiratory tract infections, traumatic fractures, dizziness, and herpes simplex infections. No treatment-related AEs were observed in the control group. All AEs in the MSC group resolved without hospitalization or study discontinuation.

Table 3.

Adverse Events Reported During the Study.

Variables	Control group (n = 12)	Treatment group (n = 11)
Patients with AE, n (%)	5 (45.5%)	9 (75.0%)
Treatment related, n (%)	0 (0%)	9 (75.0%)
Patients with SAE	0 (0%)	0 (0%)
AEs by grade, n
Grade 1	0 (0%)	2 (16.7%)
Grade 2	5 (45.5%)	4 (33.3%)
Grade 3	0 (0%)	3 (25%)
Grade 4	0 (0%)	0 (0%)
Grade 5	0 (0%)	0 (0%)

AE: adverse event; SAE: serious adverse event.

Discussion

The results of the current study demonstrated that IA injection of allogenic BM-MSCs (Catholic MASTER cells) was safe and effective for the clinical improvement of knee OA. Structural changes in knee OA evaluated using quantitative T2 cartilage mapping MRI showed a significant preventive effect on the degenerative progression of the cartilage at 12 months. To the best of our knowledge, this study is the first to show the potential efficacy of a single-dose administration of allogeneic BM-MSCs via IA injection, supported by a comprehensive serial analysis through quantitative T2 cartilage mapping MRI evaluations.

The presence of an inflammatory environment, such as synovitis, within the joint plays a pivotal role in the development of OA, subsequently leading to the advancement of joint dysfunction ¹⁵ . MSCs exhibit a remarkable ability to alleviate inflammatory conditions by modulating immune cell properties. The regulatory effect of MSCs on the immune system has captured significant interest because of their potential allogeneic applications ¹⁶ . Although the therapeutic potential of autologous BM-MSCs has been widely investigated, harvesting these cells is an invasive procedure. More critically, autologous BM-MSCs harvested from older patients exhibit limitations in both quantity and quality ¹⁷ . Our research used Catholic MASTER cells obtained from the bone marrow of young and healthy donors. These cells may possess superior immunomodulatory functions compared with senescent MSCs, and several studies have reported their effects in this regard^18,19. The superiority of the MSC source might have played a role in the positive therapeutic outcomes observed in our study.

The findings of this study demonstrate the safety of IA injection of BM-MSCs into osteoarthritic knees, despite prior concerns regarding the potential AEs associated with MSC IA injections. In the early phase following treatment, the most prevalent AEs associated with the procedure in the MSC group were pain and swelling of the knee, which were observed in 75.0% of cases. Similarly, recent clinical studies on IA injection of MSCs have consistently reported these AEs, and none of the participants experienced SAEs or AEs leading to study discontinuation ²⁰ . Another recent study indicated that the surgical implantation of MSCs derived from allogenic umbilical cords was safe, with no treatment-related SAEs or immune rejection over 7 years, even considering their allogeneic application ²¹ . These outcomes align with our results, showing the safety of IA injections of MSCs. However, additional studies are required to assess potential AEs over the mid- to long-term.

Numerous studies have evaluated the outcomes of IA injections of MSCs versus normal saline, providing a consistent and encouraging finding: IA administration of MSCs in patients with knee OA results in significant pain reduction and enhancement of function in the short term^22–25. Our study also suggests that IA injection of allogeneic BM-MSCs results in remarkable pain relief in patients with knee OA. Over a 12-month period, at all time points, MSCs induced an improvement in the pain index compared with baseline, while the control solution did not. This single injection of BM-MSCs led to remarkably greater reductions in the WOMAC and KOOS pain scores at 9 months compared with those in the control group. However, these differences did not persist after 12 months. This suggests that repeated IA injections at intervals of several months could produce long-term and continuous pain relief.

Another interesting finding from this study is the decrease in serum CTX-I levels at 3 months in the MSC group compared with that in the control group. Cytokine-driven inflammatory reactions, involving molecules such as IL-1, IL-6, and TNF-α, promote bone loss by regulating the differentiation and function of osteoblasts and osteoclasts. This activation in bone resorption, in turn, causes a rise in bone turnover and a reduction in bone mineral density ²⁶ . In our study, there was no change in serum inflammatory markers after the IA injection of MSCs, but a decrease in bone resorption markers was observed. These findings suggest that IA injection of MSCs induces a local, rather than a systemic, immunomodulatory effect. Although the number of participants in our study was small, this finding suggests that this local anti-inflammatory effect can have a positive effect on bone turnover and mineral density.

The capacity of MSCs to repair and regenerate cartilaginous tissues in joints is currently unclear. Multi-compositional MRI sequences offer a means to assess the progression of cartilage damage by analyzing the serial compositional changes in knee cartilage. This imaging modality can provide valuable insights for evaluating treatment efficacy and determining the appropriate course of treatment for patients. T2 mapping is an innovative technique in molecular imaging crafted for the early detection of OA because it identifies biochemical changes before any morphological changes occur. Numerous studies have established a link between cartilage deterioration and T2 values^27,28, with elevated T2 values being indicative of articular cartilage in individuals at risk for OA ²⁹ . We aimed to measure the difference in the T2 values at baseline and after IA injection of MSCs to quantitatively evaluate the effect of MSCs on knee cartilage. In our quantitative T2 MRI cartilage analysis, the medial compartment of the knee cartilage in the MSC group showed lower rates of increase in the T2 values between baseline and 12 months than that in the control group. In knee OA, the medial compartment is more commonly affected and often more severely affected than the lateral compartment ³⁰ . These results reinforce the concept that IA injection of MSCs slows cartilage degeneration in knee OA.

Our study had some limitations. First, the study’s limited sample size restricts its ability to discern differences, raising the risk of a false-negative outcome. Second, the small cohort size precluded the possibility of conducting subgroup analyses, such as distinct evaluations of groups based on the index knee and baseline K–L grade. Third, the use of permuted-block randomization led to an imbalance in the baseline characteristics of the samples. Finally, owing to ethical considerations, the potential effects of MSCs on cartilage could not be assessed through cartilage biopsy, which would have shed light on the mechanisms underlying pain alleviation. Conversely, a major advantage of this study is its selection of a homogenous study population based on BMI and K–L grade. Furthermore, as a single-center study supported by the CIC, consistency in the preparation and injection of MSCs was ensured.

Conclusion

In conclusion, a single IA injection of allogeneic BM-MSCs safely provided pain relief in patients with knee OA and prevented the radiological progression of degenerative knee cartilage. These findings suggest the potential use of IA injection of allogeneic BM-MSCs as a pain-relieving and disease-modifying treatment for patients with knee OA in the outpatient setting.