Hyperbaric Oxygen Therapy Improves Subchondral Perfusion and Cartilage Integrity in Established Osteoarthritis: A Pharmacokinetic DCE-MRI and T2* Quantification Study

Abstract

Objective

This study investigated whether hyperbaric oxygen (HBO) therapy, initiated after disease establishment, exerts disease-modifying effects in a rat model of anterior cruciate ligament transection (ACLT)-induced post-traumatic osteoarthritis (PTOA).

Design

Thirty-six Sprague-Dawley rats underwent ACLT surgery. HBO therapy commenced at Week 11 (established OA phase) and continued for 6 weeks. Outcomes were assessed using Von Frey testing, multimodal 7T MRI (T2* quantification and pharmacokinetic DCE-MRI), and histopathology.

Results

By Week 10, ACLT induced significant mechanical allodynia, T2* prolongation, and SCB perfusion abnormalities (A increase; k _ep , k _el decrease). Following HBO intervention, the ACLT+HBO group exhibited a marked recovery in mechanical withdrawal threshold and significantly lower T2* values compared to the ACLT-only group (p < 0.01). DCE-MRI demonstrated that HBO restored clearance kinetics (k _el , k _ep ) and suppressed pathological blood volume (A), effectively remediating subchondral venous congestion. Histologically, HBO reduced OARSI scores and suppressed α-SMA-positive vascular invasion across the tidemark. These findings suggest that HBO therapy may alleviate PTOA by suppressing catabolic activity and normalizing pathological subchondral angiogenesis.

Conclusions

HBO therapy, initiated post-pathogenesis, effectively attenuates nociception and preserves cartilage integrity by restoring subchondral microcirculatory homeostasis. These findings establish HBO as a potent, non-invasive disease-modifying strategy for PTOA.

Introduction

Anterior cruciate ligament (ACL) injury is among the most common knee traumas in young, physically active populations, with an annual incidence of 60–80 per 100,000 persons.^1,2 While ACL reconstruction effectively restores mechanical stability, it frequently fails to arrest the onset of post-traumatic osteoarthritis (PTOA).^3,4 Longitudinal data indicate that over 50% of patients manifest radiographic evidence of joint degeneration within a decade of injury, representing a stark elevation in risk compared to uninjured cohorts.^5,6 The pathogenesis of PTOA is characterized by a deleterious crosstalk within the osteochondral unit, involving acute inflammation, cartilage matrix attrition, and aberrant subchondral bone remodeling. ⁷ Crucially, pathological angiogenesis—specifically the aberrant proliferation of CD31^hiEmcn^hi Type H vessels—has emerged as a pivotal mechanistic bridge coupling subchondral bone remodeling with articular cartilage degeneration in osteoarthritis. This neurovascular-bone crosstalk, often driven by elevated pro-angiogenic factors within the subchondral microenvironment, facilitates the transport of catabolic cytokines across the tidemark, thereby accelerating matrix failure.^8,9 These vessels are coupled with accelerated osteogenesis and have been implicated in a feedback cycle that drives articular cartilage degradation and joint structural failure.^10,11 Current management, however, remains symptom-oriented, with NSAIDs and intra-articular injections serving as the primary therapeutic options. There is a critical, unmet clinical need for disease-modifying osteoarthritis drugs (DMOADs) or interventions capable of stabilizing the joint microenvironment.^12,13

Hyperbaric oxygenation (HBO) offers a potent, non-invasive physical intervention. By delivering 100% oxygen at elevated atmospheric pressures, HBO substantially increases tissue oxygen tension, modulates inflammatory cascades, and promotes vascular normalization. Crucially, HBO is expected to modulate pathological angiogenesis rather than entirely suppress it. Mechanistically, elevating tissue oxygen tension upregulates prolyl hydroxylase domain enzyme 2 (PHD2), which accelerates the degradation of hypoxia-inducible factor-1α (HIF-1α), thereby downregulating vascular endothelial growth factor A (VEGFA) signaling. This pathway is anticipated to inhibit the excessive proliferation of Type H vessels (CD31^hiEmcn^hi) in the subchondral bone, thereby interrupting the vicious cycle between abnormal bone remodeling and cartilage degeneration. While HBO has shown promise in attenuating cartilage degradation and suppressing pathological angiogenesis,^8,14-16 its longitudinal protective effects of HBO in the context of established PTOA remain poorly defined. Specifically, previous studies have often relied on isolated, end-point outcome measures, failing to capture the multidimensional pathophysiology of the disease.

To address these gaps, a multimodal diagnostic framework is necessary to integrate quantitative imaging with functional and structural assessments. Without incorporating advanced MRI biomarkers—specifically T2 quantification for collagen integrity and pharmacokinetic dynamic contrast enhanced magnetic resonance imaging (DCE-MRI) for subchondral perfusion—the therapeutic window and mechanistic impact of HBO remain difficult to fully characterize. This lack of standardized evaluation has hindered the clinical translation of HBO as a potential DMOAD strategy.^17,18 In the present study, we utilized a rat ACL-transection model to reflect the progression of PTOA. We hypothesized that HBO, initiated post-injury, would exert disease-modifying effects by stabilizing subchondral microvascular function and preserving cartilage ultrastructure. Our primary objective was to evaluate these effects longitudinally at baseline, 10 weeks, and 16 weeks, using an integrated approach of T2 quantification, DCE-MRI perfusion analysis, behavioral pain testing, and histopathological validation. This study aims to provide a reliable scientific foundation for HBO as a non-invasive strategy for joint preservation in PTOA.

Materials and Methods

Animal Models and Experimental Design

Thirty-six 8 weeks old male Sprague-Dawley rats (285–310 g) were utilized to establish a PTOA model via anterior cruciate ligament transection (ACLT). ¹⁹ After one week acclimatization period, all animals were randomly allocated into four cohorts: Sham (n = 6), Sham + HBO (n = 6), ACLT (n = 12), and ACLT + HBO (n = 12). Sample size calculations were performed a priori using G*Power 3.1. Based on preliminary data, a sample size of n = 6 per sham-control group yielded over 80% power to detect a substantial effect size (Cohen’s d > 1.5) between the Sham and ACLT groups. Furthermore, allocating n = 12 per ACLT-derived group ensured a robust statistical power (>90%) to identify moderate-to-large therapeutic effects of HBO intervention within the disease model, accounting for potential biological variability. All protocols were approved by the National Defense Medical University IACUC (No. IACUC-21-105) and adhered to ARRIVE guidelines. PTOA was induced via a medial parapatellar approach under general anesthesia. In ACLT groups, the ligament was completely transected, with joint instability confirmed intraoperatively by a positive anterior drawer test. Sham groups underwent arthrotomy without transection. Following layered wound closure, all animals were permitted unrestricted weight-bearing and cage ambulation to facilitate degenerative progression.

Hyperbaric Oxygen Therapy (HBOT)

Starting at postoperative week 11—a timeframe corresponding to established PTOA—Groups 2 and 4 (Sham+HBO, ACLT+HBO) underwent HBOT in a specialized small-animal chamber. The protocol utilized 100% oxygen at 2.5 atmospheres absolute (ATA) for 90 min/session (including 15-min compression/decompression phases), administered 5 days/week for 6 weeks (30 sessions total). ¹⁴ Sham and ACLT groups were maintained under normobaric conditions (1 ATA) for equivalent durations. The Sham+HBO group served as a biological control for hyperoxia and pressure independent of joint pathology.

Pain Behavioral Assessment

Mechanical allodynia was longitudinally evaluated at baseline (week 0), week 10, and week 16. Following a 15-min acclimation to a wire-mesh environment, the 50% paw withdrawal threshold (PWT) was determined using von Frey filaments (Stoelting, Wood Dale, IL, USA; range: 0.41–15.10 g) via the up-down method. ²⁰ Filaments were applied perpendicularly to the ipsilateral hindpaw for 5 s; brisk withdrawal or flinching was recorded as a positive response. Assessments were performed by two independent observers blinded to group allocation to ensure objective quantification and reproducibility.

MRI Acquisition and Pharmacokinetic Analysis

To longitudinally monitor PTOA progression and therapeutic response, MRI was performed at baseline (Week 0) and postoperatively at Weeks 10 and 16 (Figure 1). MRI was performed using a 7.0 T preclinical system (Bruker, Germany) with a volume excitation coil and a quadrature knee surface receiver coil. Under isoflurane anesthesia (1–2% maintenance), the right knee was centrally aligned to ensure B₀ homogeneity. To assess cartilage integrity, a sagittal multi-echo gradient-recalled echo (ME-GRE) sequence was employed (FOV=30×30 mm²; matrix=192×256; TR/FA=800 ms/30°; 5 TEs: 4.0–20.0 ms). T2* relaxation maps were generated via voxel-wise mono-exponential fitting using MIStar® software (Apollo Medical Imaging, Australia),^21,22 with ROIs manually delineated on the LFC (lateral femoral condyle), MFC (medial femoral condyle), LTP (lateral tibial plateau), and MTP (medial tibial plateau) by a blinded radiologist, strictly excluding synovial fluid artifacts (Figure 2). Subchondral perfusion was further evaluated via DCE-MRI using a sagittal spoiled gradient-recalled echo sequence (TR/TE=150/3.0 ms; 80 frames/slice). Gadobutrol (0.2 mmol/kg) was administered via tail vein bolus. Contrast concentration–time curves C(t) were fitted to the Brix two-compartment model to derive three primary parameters: Amplitude (A), reflecting subchondral blood volume; Permeability constant (k _ep ), representing contrast reflux; and Elimination constant (k _el ), quantifying systemic clearance.^23,24 Mean values of the tibial plateau and femoral condyle were utilized to characterize the subchondral bone (SCB) and hyaline cartilage status for each compartment (medial and lateral).

Figure 1.

Experimental design and timeline of the study. Schematic of the experimental protocol and assessment schedule. Sprague-Dawley rats (n=36) were assigned to four cohorts: Sham, Sham+HBO, ACLT, and ACLT+HBO. ACLT or sham surgery was performed at baseline, followed by hyperbaric oxygen (HBO) therapy from week 11 to 16. Mechanical allodynia (50% PWT) was longitudinally evaluated at weeks 0, 10, and 16 before MRI experiment. The study concluded at week 16 with terminal MRI (T2* quantification/DCE-MRI) and histological analyses. Data are mean ± SD; statistical significance was assessed via two-way RM ANOVA with Sidak’s post-hoc test. *p < 0.05, **p < 0.01, ***p < 0.001 vs. Sham group

Figure 2.

Representative sagittal MRI and ROI segmentation of the rat knee joint. Representative sagittal MR images illustrating the standardized segmentation of regions of interest (ROIs) for quantitative analysis. Yellow solid lines delineate the ROIs within the subchondral bone (SCB), while white dashed lines define the overlying hyaline cartilage (H.C.) layers. ROIs were manually segmented across four anatomical compartments: (A) the lateral femoral condyle (LFC) and lateral tibial plateau (LTP); and (B) the medial femoral condyle (MFC) and medial tibial plateau (MTP). These adjacent SCB and H.C. zones were identified as Lat. (H.C.) and Med. (H.C.) to facilitate the assessment of osteochondral crosstalk. All ROIs were consistently localized to the central weight-bearing regions to ensure reproducible volumetric and parametric quantification

Histopathological and Immunohistochemical Evaluation

Knee joints harvested at week 16 were fixed in 4% paraformaldehyde and decalcified in 10% EDTA (pH 7.4) for 6–8 weeks. Following paraffin embedding, coronal sections of 5 μm were cut at 100-μm intervals, referenced to the intercondylar eminence. Sections were processed with H&E stain, Safranin-O/Fast Green, and Masson’s trichrome to evaluate morphology and matrix distribution. Cartilage degeneration was quantified using the OARSI grading system. ²⁵ For vascular assessment, sections underwent heat-induced antigen retrieval and were incubated overnight with anti-α-SMA antibody (Abcam, 1:200). Signal was visualized using AEC substrate with hematoxylin counterstaining. For each animal, three non-consecutive sections (separated by at least 200 μm) were selected for analysis. Subchondral bone vascularization (0–500 μm below the tidemark) was quantified across five 400× fields using ImageJ software (NIH, USA). Two blinded observers determined vessel density (vessels/mm², defined as α-SMA⁺ vessels with clear luminal structures) and percent positive area (%) to ensure objective quantification.

Statistical Analysis

All statistical analyses were conducted in GraphPad Prism (version 9.0; San Diego, CA, USA); continuous variables are reported as mean ± SD. Longitudinal transitions in pain thresholds, T2* relaxation times, and DCE-MRI parameters were evaluated using two-way repeated measures(RM) ANOVA, with Group as the between-subjects factor and Time as the within-subjects factor. The Greenhouse–Geisser correction was applied when the assumption of sphericity was violated. Significant main effects or interactions were further analyzed using Sidak’s post-hoc test for multiple comparisons. Between-group differences at single time points were assessed via independent-samples Student’s t-tests. To ensure measurement reliability, inter- and intra-rater consistency for T2* and DCE-MRI quantification were evaluated using intraclass correlation coefficients (ICC; two-way mixed model, absolute agreement), with ICC > 0.75 defined as excellent reproducibility. ²⁶ A two-tailed p < 0.05 was considered statistically significant.

Results

Model Validation and Procedural Safety

No adverse events or procedure-related complications were recorded throughout the study. The PTOA model was successfully established via ACLT, as confirmed by intraoperative anterior tibial subluxation and persistent joint instability. In contrast, sham-operated controls maintained intact joint architecture. Assessment of measurement reliability yielded excellent inter- and intra-rater concordance. For DCE-MRI, the inter-rater intraclass correlation coefficient (ICC) was 0.979 (95% CI: 0.973–0.984), while T2* quantification reached 0.959 (95% CI: 0.935–0.974). Intra-rater reliability was similarly robust (DCE-MRI: 0.986, 95% CI: 0.981–0.989; T2*: 0.969, 95% CI: 0.950–0.980), ensuring high reproducibility across all imaging datasets.

Longitudinal Attenuation of Mechanical Allodynia

The 50% PWT demonstrated a significant time × group interaction (p < 0.001, Figure 3A). At baseline (Week 0), PWT values were comparable across all cohorts (ranging from 6.73 ± 0.89 g to 7.33 ± 1.63 g; all p > 0.926, Supplemental Table S1), establishing a uniform sensory baseline. By Week 10, both ACLT and ACLT+HBO groups exhibited a precipitous decline in PWT compared to sham controls (all p < 0.01), confirming the early manifestation of marked mechanical allodynia post-ligamentous injury. By Week 16, this sensory deficit was further exacerbated in the untreated ACLT group, with thresholds dropping to 1.71 ± 0.44 g—significantly lower than both sham and sham+HBO cohorts (p < 0.001). Notably, HBO therapy facilitated a significant preservation of sensory function; the ACLT+HBO group maintained a threshold of 4.61 ± 1.18 g, representing a superior outcome compared to the ACLT-only group (p = 0.007). While the ACLT+HBO threshold remained below sham levels, the data indicate that HBO significantly mitigates the progression of PTOA-induced mechanical allodynia (Figure 3A, Table 1). Sham surgery, regardless of HBO intervention, exerted no long-term influence on nociceptive sensitivity.

Figure 3.

Longitudinal quantitative analysis of mechanical allodynia and cartilage T2 relaxation times. (A) Mechanical allodynia assessed by the 50% paw withdrawal threshold (PWT) across experimental cohorts at weeks 0, 10, and 16. (B), (C) Longitudinal quantification of T2 relaxation times (msec) in the (B) lateral hyaline cartilage (Lat. (H.C.)) and (C) medial hyaline cartilage (Med. (H.C.)) derived from MRI T2* mapping. ACLT-induced osteoarthritis resulted in a progressive decline in PWT and a corresponding elevation in T2 values by week 16, both of which were significantly mitigated by hyperbaric oxygen (HBO) therapy

Table 1.

Summary of Sidak Post Hoc Comparisons Between ACLT and ACLT+HBO Groups at Week 16

Outcome	Mean difference	95% CI	p value
Paw withdrawal threshold	-2.90	-4.83 to -0.96	0.007
Lateral cartilage T2*	2.79	2.01 to 3.57	<0.001
Medial cartilage T2*	2.67	1.75 to 3.60	<0.001
Lateral SCB A	0.19	0.09 to 0.29	0.002
Medial SCB A	0.19	0.10 to 0.28	<0.001
Lateral SCB kel	-0.16	-0.24 to -0.08	<0.001
Medial SCB kel	-0.17	-0.24 to -0.09	<0.001
Lateral SCB kep	-1.76	-2.38 to -1.14	<0.001
Medial SCB kep	-1.58	-2.28 to -0.87	<0.001

Quantitative T2* Analysis of Articular Cartilage

Longitudinal T2* relaxation time analysis revealed that ACLT induced progressive degenerative alterations, while HBO therapy exerted a robust protective effect by Week 16 (Figure 3B–C, Table 1). At baseline (Week 0), T2* values were comparable across all cohorts (Lateral: 7.99–8.08 ms; Medial: 7.85–8.09 ms; all p > 0.433, Supplemental Tables S2-S3), establishing a uniform structural starting point. Two-way repeated measures ANOVA confirmed a highly significant time × group interaction for both compartments (p < 0.01). Notably, no statistically significant differences were observed between the lateral and medial hyaline cartilage metrics across all experimental groups (p > 0.05). By Week 10, T2* values in the ACLT and ACLT+HBO groups were significantly elevated compared to sham controls (p < 0.001), reflecting early-stage matrix degradation and increased water content post-injury. At this intermediate stage, no significant divergence was observed between the ACLT and ACLT+HBO groups (p > 0.05). However, by Week 16, a marked longitudinal divergence emerged. In the untreated ACLT group, T2* values further deteriorated to 12.35 ± 0.48 ms (lateral) and 12.28 ± 0.54 ms (medial), representing a profound loss of collagen integrity compared to sham levels (p < 0.001). Conversely, concomitant HBO therapy significantly attenuated this pathological T2* prolongation, with the ACLT+HBO group exhibiting substantially lower T2* values than the ACLT-only group (both p < 0.001, Table 1). These findings demonstrate that while ACLT triggers progressive articular cartilage degeneration, HBO intervention effectively stabilizes the extracellular matrix and mitigates long-term structural failure (Figures 3B,C, Tables S2 and S3).

Pharmacokinetic DCE-MRI Analysis of Subchondral Bone

Longitudinal analysis of subchondral bone (SCB) perfusion demonstrated that ACLT-induced microvascular dysfunction was significantly mitigated by HBO therapy by Week 16 (Figure 4, Table 1). At baseline (Week 0), all pharmacokinetic parameters (A, k _ep , k _el ) were comparable across all cohorts (p > 0.844, Supplemental Tables S4–S9), and no statistically significant differences were observed between the lateral and medial SCB compartments (p > 0.05). Two-way repeated measures ANOVA confirmed a highly significant time × group interaction for all three parameters (p < 0.001). By Week 10, both ACLT and ACLT+HBO groups exhibited a sharp elevation in amplitude (A) compared to sham controls (p < 0.01), reflecting pathological increases in subchondral blood volume. Concurrently, elimination (k _el ) and permeability (k _ep ) rate constants were markedly reduced (p <0.05), indicating impaired contrast clearance and potentially stagnant microcirculation. At this intermediate stage, no significant divergence was noted between the ACLT and ACLT+HBO groups (p > 0.05). By Week 16, a profound divergence emerged. In the untreated ACLT group, parameter A further deteriorated to 0.695 ± 0.06 (lateral) and 0.699 ± 0.06 (medial) (p < 0.001 vs. sham), accompanied by sustained deficits in k _el and k _ep (all p < 0.01). Conversely, HBO therapy significantly attenuated these pathological shifts. The ACLT+HBO group exhibited markedly lower A values (p = 0.002; p < 0.001 respectively) and substantially restored k _el and k _ep levels compared to the ACLT-only group (all p < 0.001). While these parameters had not fully normalized to sham levels, the progression of microvascular congestion was effectively curtailed.

Figure 4.

Pharmacokinetic analysis of subchondral bone perfusion via DCE-MRI. Longitudinal quantification of dynamic contrast-enhanced MRI (DCE-MRI) perfusion parameters derived from the Brix model in the subchondral bone (SCB). (A) The signal amplitude (A, a.u.), representing the peak contrast enhancement and integrated vascular-interstitial volume; (B) the elimination rate constant (k _el , msec^-1); and (C) the transfer rate constant (k _ep , msec^-1), reflecting microvascular permeability and exchange efficiency. Upper panels correspond to the lateral SCB (Lat. SCB), and lower panels to the medial SCB (Med. SCB). ACLT-induced osteoarthritis triggered a significant elevation in amplitude (A) and a concomitant reduction in k _el and k _ep by week 16, indicating pathological hyperperfusion and impaired clearance. Conversely, HBO therapy effectively normalized these perfusion deficits, restoring microvascular hemodynamics to levels comparable to the sham cohorts

In summary, ACLT triggers a cascade of subchondral perfusion abnormalities characterized by hyper-permeability (increased A) and circulatory stasis (decreased k _ep , k _el ). HBO intervention effectively modulates these microcirculatory dynamics by Week 16, thereby stabilizing the subchondral environment and delaying PTOA progression. Pixel-by-pixel parametric maps (Figure 5) further corroborated these quantitative findings, providing visual evidence of HBO-mediated microvascular stabilization.

Figure 5.

Representative spatial parametric maps of the knee joint at the 16-week endpoint. Longitudinal pseudocolor mapping of the ACLT-induced PTOA model in Sprague-Dawley rats. The panels (arranged vertically) display T2 relaxation time (ms), amplitude (A, arbitrary units), elimination rate constant (k _el , min^-1), and transfer rate constant (k _ep , min^-1). The pseudocolor scale transitions from warm colors (red/yellow; higher values) to cool colors (blue/green; lower values). Note the HBO-mediated restoration of microvascular kinetics, characterized by a reduction in amplitude (A) and a concomitant increase in k _el and k _ep , reflecting improved efferent clearance and normalized subchondral perfusion

Histopathological and Immunohistochemical Findings

Histopathological analysis corroborated that ACLT induced profound articular cartilage degeneration alongside pathological subchondral bone angiogenesis, both of which were significantly rescued by HBO therapy (Figure 6A). Safranin-O/Fast Green staining in the ACLT group revealed transmural proteoglycan depletion, surface fibrillation, and vertical fissuring extending to the tidemark. Conversely, HBO-treated joints exhibited robust preservation of proteoglycan content and superior structural integrity. H&E and Masson’s Trichrome staining further characterized these deficits as chondrocyte disorganization and collagenous network collapse in the ACLT group, whereas HBO intervention stabilized the extracellular matrix and maintained a more ordered cellular architecture. IHC staining for α-SMA identified a marked surge in subchondral vascularization following ACLT, characterized by pathological vessel penetration across the tidemark. In contrast, HBO therapy significantly suppressed this vascular hyperplasia, maintaining vessel density at levels comparable to the sham cohorts. Quantitative histomorphometry confirmed these qualitative observations (Figure 6B). The OARSI score was significantly elevated in the ACLT group (14.1 ± 4.3) compared to sham controls (3.3 ± 1.5; p < 0.0001). Notably, HBO therapy resulted in a significantly lower OARSI score (8.2 ± 3.3; p < 0.001 vs. ACLT), reflecting a ∼42% reduction in structural severity. Similarly, SCB vascular metrics—vessel number (82.7 ± 14.6 vs. 57.3 ± 0.6 vessels/mm²) and vessel density (3.32 ± 0.34 vs. 2.42 ± 0.22 %)—were significantly reduced in the ACLT+HBO group compared to the ACLT-only group (all p < 0.05). Vessel diameter was significantly increased in the ACLT group compared to the sham group, and was subsequently reduced following HBOT (all p < 0.001). In summary, while ACLT triggers a cascade of cartilage degradation and aberrant angiogenesis, HBO therapy effectively fortifies the cartilage matrix, improves OARSI outcomes, and normalizes the subchondral vascular microenvironment, demonstrating a clear chondroprotective effect.

Figure 6.

Histological and immunohistochemical characterization of osteochondral integrity at week 16. (A) Representative histological profiles of the knee joints. Micrographs of the medial tibia plateau illustrate the structural and molecular changes across the four experimental cohorts. Rows from top to bottom display: Safranin-O/Fast Green staining for proteoglycan content (red); H&E for cellular morphology and tidemark integrity; Masson’s Trichrome for collagen network organization (blue); and α-SMA immunohistochemistry for subchondral bone (SCB) neoangiogenesis. In the ACLT group, extensive proteoglycan depletion, surface fibrillation, and pathological α-SMA⁺ vessel invasion (black arrows) across the tidemark are evident. These osteochondral deficits were markedly attenuated in the ACLT+HBO group, which exhibited superior structural preservation and suppressed vascular hyperplasia. Scale bars = 200 μm. High-magnification views of α-SMA⁺ staining for the ACLT and ACLT+HBO groups were presented in the bottom insets to illustrate vascular details within the subchondral bone/cartilage interface (red-boxed regions). Scale bars = 50 μm. (B) Quantitative histological and vascular analysis. OARSI histological scores representing the severity of articular cartilage degeneration. Quantitative assessment of SCB vascularization, including vessel number (vessels/mm²), vessel density (% positive area) and diameters (μm)

Discussion

This study provides the first longitudinal evidence that hyperbaric oxygen therapy, even when initiated post-establishment of ACLT-induced PTOA (Week 11), exerts significant disease-modifying protective effects. A critical clinical finding is the observed therapeutic window: while mechanical allodynia and cartilage matrix degradation were already pronounced by Week 10, a six-week HBO regimen facilitated a substantial recovery of the mechanical withdrawal threshold and stabilized the extracellular matrix by Week 16. This potent analgesic effect likely stems from HBO’s ability to suppress pathological vascular invasion across the tidemark. As subchondral angiogenesis is intrinsically linked to the innervation of sensory nerves into the osteochondral junction, the inhibition of these vessels directly contributes to pain alleviation. ²⁷ Furthermore, the reduction in iNOS expression following HBO intervention, as observed in human OA chondrocytes, ²⁸ suggests a potent anti-inflammatory mechanism that stabilizes the joint environment and further mitigates chronic pain. ¹⁷

The observed restoration of k _el and k _ep parameters via HBO intervention provides a novel hemodynamic explanation for the mitigation of mechanical allodynia. In the osteoarthritic state, microvascular dysregulation often leads to venous stasis and a subsequent rise in intraosseous pressure, which is clinically recognized as a primary source of deep-seated bone pain. ²⁹ Our findings suggest that HBO therapy may alleviate this intraosseous hypertension by facilitating efferent microvascular clearance, as evidenced by recovered k _el constants. By reducing the accumulation of acidic metabolic byproducts and pro-nociceptive factors within the subchondral microenvironment, HBO effectively deactivates the sensory nerve fibers densely distributed at the osteochondral junction, thereby exerting a potent and sustained analgesic effect that is superior to purely symptomatic treatments. Although exercise benefits vascular function, its application in early PTOA is hindered by pain and instability. HBO provides a distinct advantage by directly increasing tissue oxygenation and resolving venous stasis without requiring physical exertion, thereby ensuring more rapid stabilization of the subchondral microenvironment.

The therapeutic efficacy of HBO is further underscored by the multimodal synergy between biochemical imaging and microvascular kinetics. T2* quantification revealed that HBO effectively curtailed pathological water content increases and matrix failure (p < 0.01). This preservation of the collagenous framework is directly supported by molecular evidence that HBO suppresses catabolic enzymes like MMP-3 while up-regulating anabolic factors including TIMP-1, type-II collagen, and aggrecan. ²⁸ By shifting the metabolic balance toward anabolism—including enhanced collagen synthesis and ECM regulation—HBO retards the progressive biochemical degradation that precedes macroscopic structural damage. ¹⁴ Simultaneously, DCE-MRI provided deeper insights into the osteochondral interface, where HBO reversed established perfusion abnormalities. The restoration of k _el and k _ep constants suggests that HBO remediates microvascular dysregulation and potential venous stasis, a hallmark of OA pathogenesis. ³⁰

Mechanistically, our results suggest that HBO modulates the hypoxia-driven vicious cycle within the subchondral microenvironment. By elevating tissue oxygen tension (PO ₂ ), HBO suppresses HIF-1α stabilization and downregulates VEGFA-driven pathological angiogenesis. This aligns with the burgeoning paradigm of H-type vessel modulation (CD31^hi, EMCN^hi) in OA pathogenesis. Pathological neoangiogenesis across the tidemark, driven by hypoxia, serves as a conduit for catabolic cytokines that accelerate cartilage matrix failure.^8,15 While conventional pharmacological agents often fail to target this deep-tissue vascular remodeling, the elevated PO ₂ delivered by HBO serves as a physiological antagonist to HIF-1α. This ‘oxygen-driven signaling’ may mimic the effect of localized VEGF-inhibition therapy, but through a non-invasive, systemic route, possibly decoupling the pathological neurovascular–bone crosstalk that characterizes late-stage PTOA progression. ³¹ This is consistent with recent mechanistic investigations identifying that inhibiting H-type vessel proliferation at the subchondral interface significantly ameliorates OA progression. ³² By normalizing the subchondral vascular microenvironment, HBO promotes a stable environment conducive to chondrocyte survival and collagen synthesis.

A standout feature of this study is the efficacy of HBO when initiated during the established phase of PTOA (Week 11). Most preclinical investigations focus on prophylactic interventions immediately following joint injury; however, our data demonstrate that HBO can effectively ‘rescue’ the joint even after the onset of macroscopic cartilage fibrillation and significant mechanical sensitivity. Compared to intra-articular injections—which often exhibit rapid clearance and limited subchondral penetration—HBO facilitates an oxygen-driven volumetric therapeutic effect, simultaneously addressing cartilage surface integrity and subchondral microcirculatory health. ³³ Consequently, this study establishes a solid scientific foundation for HBO as a late-intervention adjunctive therapy with high translational potential for established human PTOA.

Despite the robust longitudinal evidence, several limitations warrant consideration. First, while the rat ACLT model recapitulates key features of human PTOA, inherent differences in joint biomechanics and metabolic rates may affect direct clinical translatability. Second, the HBO regimen was fixed (2.5 ATA, 90 min); thus, the dose-response relationship and the optimal timing for later-stage interventions remain to be elucidated. Third, while advanced imaging modalities were integrated, direct molecular-level analyses—including the expression of MMP-3, iNOS, and the HIF-1α pathway—were not performed in the current cohort and should be explored in future studies to confirm the signaling cascades. Fourth, our study lacked a sham-HBO control group (chamber isolation without pressure). However, the significant functional improvements in subchondral bone perfusion observed via DCE-MRI and the histological changes in microvasculature suggest that the therapeutic effects are primarily driven by HBO rather than the temporary immobilization during treatment. Finally, the 16-week observation period precludes assessment of long-term radiographic and functional outcomes.

In conclusion, this study demonstrates that HBO therapy, initiated during the established phase of PTOA, effectively attenuates pain, cartilage degeneration, and subchondral vascular hyperplasia. By integrating behavioral assessment, multimodal MRI (T2* quantification and pharmacokinetic DCE-MRI), and histopathology, we provide a solid scientific foundation for HBO as a late-intervention, disease-modifying strategy. Furthermore, the translational value of these findings lies in the use of high-field MRI to bridge the common gap between subjective patient-reported outcome measures (PROMs) and structural changes in OA. Since PROMs can be inconsistent, quantitative 7T MRI biomarkers—such as the DCE-MRI parameters used here—offer objective surrogate endpoints that can detect therapeutic responses earlier than clinical symptoms. We envision targeting the early-to-mid stage PTOA population with a history of joint trauma, where subchondral microvascular remodeling remains a reversible pathological driver. The clinical feasibility of this approach is supported by established protocols in wound care, where daily HBO sessions are well-tolerated by adult patients. Unlike neonatal populations, the risk of oxidative or neurological damage in adults is minimized through standardized regimes and the use of ‘air breaks.’ While the daily treatment schedule presents a logistical consideration, it aligns with current clinical paradigms for musculoskeletal recovery. These findings deepen our understanding of HBO’s regulatory role in the osteoarthritic microenvironment and support its translational potential as a non-invasive adjunctive therapy for human PTOA.

Supplemental Material

Supplemental Material - Hyperbaric Oxygen Therapy Improves Subchondral Perfusion and Cartilage Integrity in Established Osteoarthritis: A Pharmacokinetic DCE-MRI and T2* Quantification Study

Supplemental Material for Hyperbaric Oxygen Therapy Improves Subchondral Perfusion and Cartilage Integrity in Established Osteoarthritis: A Pharmacokinetic DCE-MRI and T2* Quantification Study by Guo-Shu Huang, Shih-Wei Chiang, Chiao-Chi Chen, Yi-Jen Peng, Yu-Juei Hsu, Kun-Lun Huang, Yu-Ching Chou, Ming-Huang Lin, Ying-Chun Liu, Chao-Ying Wang in CARTILAGE