Oral enzyme combination therapy reduces systemic inflammation, urinary CTXII and pain in knee osteoarthritis: a proof-of-mechanism, randomised, crossover, double-blind, placebo-controlled trial

Abstract

Objectives

Oral enzyme combination (OEC) therapy with bromelain, trypsin and rutoside reduces pain and improves function in patients with knee osteoarthritis (OA). Here, we investigated several potential biological mechanisms underlying the clinical effects of OEC therapy in patients with established knee OA with respect to innate immunity, systemic inflammation and cartilage turnover (EudraCT 2020-003154-80, NCT05038410).

Methods

Patients (age ≥40 years, body mass index (BMI) ≤35 kg/m²) with symptomatic knee OA were randomised to either placebo or OEC, administered 2×3 tablets/day, for 8 weeks before crossing over after a 4-week washout period. Different markers exploring innate immunity, inflammation and cartilage matrix degradation have been measured in the blood using immunoassays or cytometric methods. Data from the modified intention-to-treat population (mITT) were analysed using a generalised linear mixed model. No correction for multiple comparisons was made due to the exploratory nature of the study.

Results

Altogether, 45 patients were randomised; 43 completed both treatment sequences (mITT; mean age: 63.3 years; mean BMI: 27.4 kg/m²; mean global Knee injury and Osteoarthritis Outcome Score (KOOS): 48.7). OEC significantly increased levels of α2-macroglobulin (p=0.038) and interleukin-10 (p<0.0001) while decreasing urinary carboxyl-terminal cross-linked telopeptide of type II collagen (p=0.038). Patients administered OEC exhibited significant improvements in KOOS Pain (p=0.0464) and Symptoms (p=0.026) subdomains but not globally. OEC was well tolerated, with no serious related adverse events reported in either group.

Conclusions

One of the key findings of this proof-of-mechanism study is that OEC modulates IL-10 production, suggesting an anti-inflammatory effect in patients with knee OA. This main finding contributes to explaining the effects of OEC on pain and function in these patients.

Trial registration number

NCT05038410.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Currently, no disease-modifying treatment for knee osteoarthritis (OA) is available. Oral enzyme combination (OEC) has demonstrated efficacy in relieving pain in knee OA and other conditions.

WHAT THIS STUDY ADDS

• Here, we show that in patients with knee OA, OEC increased serum α2-macroglobulin, the anti-inflammatory cytokine IL-10 and decreased urinary CTX-II levels while improving patient algo-functional status.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• OEC has the potential to be a safe disease-modifying oral treatment for knee OA. Its use in clinical practice could facilitate physical activity, a core treatment of knee OA, and limit the consumption of non-steroidal anti-inflammatory drugs and analgesics.

Introduction

Osteoarthritis (OA), a chronic condition affecting the joint and its tissues, is clinically defined by pain, stiffness, deformity, reduced range of motion and muscle atrophy and dysfunction.¹,³ In OA, joint tissues undergo a progressive loss of cartilage, bone remodelling and synovium inflammation.^{1 2} A systematic analysis for the Global Burden of Disease Study 2021 for OA found that in 2020, the knee was the most common site of OA with a global age-standardised prevalence of over 4300 cases per 100 000.⁴ Patients with knee OA experience substantial disability and impaired quality of life, and OA-associated physical inactivity due to pain increases the risk of all-cause mortality, cardiovascular disease, osteoporosis, as well as depression and anxiety⁵ .^{1 6}

Current knee OA treatments aim to reduce pain and improve joint function. The Osteoarthritis Research Society International, American College of Rheumatology, American Academy of Orthopaedic Surgeons and European Alliance of Associations for Rheumatology guidelines include both non-pharmacological (exercise, education and weight reduction) and pharmacological (topical or oral non-steroidal anti-inflammatory drugs (NSAIDs) and intra-articular corticosteroids) approaches⁷ .⁸,¹¹ While pharmacological treatments improve symptoms, their clinical efficacy is weak to moderate, and their long-term use, particularly that of oral NSAIDs, may induce adverse effects.⁸,^{1012 13} Consequently, guidelines recommend that the consumption of oral NSAIDs be limited to the lowest dose and shortest time possible.⁸,¹⁰ As such, effective but safer alternative pharmacological options would benefit patients with OA.

Oral enzyme combination (OEC) therapy with bromelain, trypsin and the flavonoid rutoside has been reported to significantly improve pain, joint stiffness, functional disability and mobility in several painful musculoskeletal disorders, including knee OA.¹³ In clinical trials where OEC or diclofenac was administered to patients with knee OA, OEC was not less effective than diclofenac in reducing pain and/or improving knee function.¹⁴,¹⁸ This finding was recapitulated in a pooled reanalysis of individual patient-level data from six clinical trials, which further indicated that the safety profile of OEC was superior to that of diclofenac.¹⁹ The exact pharmacokinetics and pharmacodynamics of OEC and the mechanisms by which it exerts its clinical effects remain poorly understood, although previous studies have explored some potential mechanisms of action. For example, trypsin, a serine protease and bromelain, a cysteine protease, are both thought to modulate inflammatory and immune responses by interacting with the acute-phase protein α2-macroglobulin (A2M) to enable subsequent binding of pro-inflammatory cytokines.²⁰,²² The flavonoid rutoside has been suggested to have antioxidant and anti-inflammatory effects, which has led to its use in combination with proteolytic enzymes.²³ In preclinical OA models, rutoside reduced cartilage degradation scores and protected articular chondrocytes against hydrogen peroxide-induced oxidative stress.^{24 25} Regardless, a more complete understanding of the systemic biological mechanisms of action of OECs remains incomplete. Therefore, this clinical trial was designed to evaluate specifically, in patients with established knee OA (1) as main goal the biological effects of OEC on innate immunity, systemic inflammation and cartilage turnover and (2) as secondary outcomes, potential clinical effects.

Methods

Design

This study was a randomised, double-blind, placebo-controlled, single crossover, multicentre, exploratory clinical trial (WOBE-SMART, EudraCT 2020-003154-80 and NCT05038410) involving individuals with symptomatic knee OA (figure 1). Patients were recruited from five sites across Belgium from April 2021 to October 2022.

Figure 1. Trial design. This study was a randomised, double-blind, placebo-controlled, single crossover, multicentre, exploratory clinical trial. Treatment duration was 8 weeks with a 4-week washout period. BID, bis in die; BMI, body mass index; FIP, Fédération Internationale Pharmaceutique; OA, osteoarthritis; OEC, oral enzyme combination; V, visit; VAS, Visual Analogue Scale.

Eligible patients were centrally randomised at baseline to an intervention sequence in a 1:1 ratio using a pre-established randomisation list with block sizes of two or four. Randomisation was stratified by site. Study participants who withdrew during the active recruitment period were replaced.

After the first 8 weeks of treatment with either placebo or OEC, patients had a 4-week washout period before crossing over to their respective alternative arm of OEC or placebo for an additional 8 weeks. Primary outcome objectives were assessed following 8 weeks of intervention with placebo or OEC. Following the completion of the core 20-week study period, an additional 16-week extension was initiated—the results from this extension will be reported elsewhere.

Patients

Individuals were eligible if they were aged ≥40 years, had a body mass index ≤35 kg/m², had pain on the index knee at rest or walking (≥40 on a 0–100 Visual Analogue Scale (VAS)), and had symptomatic OA of the knee for more than 6 months (radiological Kellgren-Lawrence grade II–III). Key exclusion criteria included knee prosthesis, recent trauma or arthroscopy, and certain pharmacotherapies or diet (oral prednisolone, local corticosteroid injections, recent use of symptomatic slow-acting drugs for OA or dietary supplements). For full inclusion and exclusion criteria, please see the online supplemental file.

Interventions

The active investigational product was Wobenzym, an OEC containing bromelain (67.5–76.5 mg), trypsin (32–48 mg) and rutoside trihydrate (100 mg) (This formulation is approved as a drug under the name ‘Wobenzym’ in Germany, used in this trial, as well as under the name of ‘Phlogenzym’ in Czech Republic and some other Eastern European countries). To ensure blinding of the study medication, OEC and its matching placebo were provided from the same batch number and in identical tablet dosing forms. More specifically, the matching placebo was identical in form and shape and used the same enteric coating as the OEC, but the active ingredients were replaced with microcrystalline cellulose. Both treatments were dosed as three tablets twice a day, in the morning and the evening, taken at least 30 min before or 90 min after a meal.

Outcomes

The primary outcome objectives were to explore the effect of OEC treatment on (1) the innate immune response, (2) markers of systemic inflammation and (3) markers of cartilage turnover. Markers of innate immune response included A2M, macrophage 1/macrophage 2 (M1/M2) polarisation and nucleotide-binding domain, leucine-rich repeat containing family, pyrin domain-containing-3 (NLRP3) inflammasome. Systemic inflammatory markers included interleukin-10 (IL-10), interleukin-1β (IL-1β), interleukin-4 (IL-4), interleukin-6 (IL-6) and tumour necrosis factor alpha (TNFα). Cartilage biomarkers included urinary carboxyl-terminal cross-linked telopeptide of type II collagen (uCTXII), serum Coll2-1 (s-Coll2-1), serum Coll2-1NO2 (s-Coll2-1NO2), serum cartilage oligomeric matrix protein and serum N-propeptide of collagen IIA. Further details on assay techniques, sampling, processing and storage conditions can be found in the online supplemental file.

Secondary outcome objectives included the serum concentrations of transforming growth factors (TGF-β1, TGF-β2, TGF-β3) and the Knee injury and Osteoarthritis Outcome Score (KOOS). The KOOS is a patient-reported outcome measure used in knee injury and/or OA that has a global score as well as five subdomain scores: Pain, Symptoms, Activities of Daily Living, Sport and Recreation Function and Knee-Related Quality of Life.^{26 27} Each subdomain is scored from 0 (extreme knee problems) to 100 (no knee problems).²⁶ Other secondary outcome objectives included patient-reported knee pain at rest and while walking as measured on a VAS of 0–100 with 0 representing ‘no pain’ and 100 representing ‘pain as bad as it could possibly be’; use of rescue medication and cumulative analgesic consumption scores²⁸; and clinical chemistry measurements using immunoassays, which measured the concentration of myokines (eg, myostatin), haemoglobin, haematocrit, erythrocyte count, triglyceride (TG), high-density lipoprotein, glucose, high-sensitivity C reactive protein (hsCRP), uric acid, coagulation markers and key hepatic enzymes. Post hoc exploratory analyses were performed for hsCRP and interleukin-1 receptor antagonist (IL-1RA). Safety outcomes included discontinuation rates and adverse events (AEs) as coded using the Medical Dictionary for Regulatory Activities.

Statistical analyses

This clinical study was explorative in nature; thus, sample sizes were not formally calculated. However, to inform the study size, we applied an approximate calculation based on an expected difference of 10% between OEC and placebo in the collagen biomarker Coll2-1NO2 and a type I error (alpha) of 5%, assuming a 15% drop-out. The expected difference of 10% was selected as an estimate of clinically relevant changes. Coll2-1NO2 was chosen as a surrogate marker because of its association with collagen degradation and local inflammatory processes, its correlation with disease severity and progression, and its robustness to environmental and experimental conditions.^{29 30} To ensure a power of 80% to 90%, we targeted 20 randomised patients per treatment arm.

The modified intention-to-treat (mITT) population included all randomised patients with at least one post baseline assessment for any primary efficacy endpoint. The patients were included in the group they were randomised to irrespective of the actual study intervention received. Data for the mITT population were analysed using a generalised linear mixed model (GLMM) after 8 weeks of treatment based on the changes from the baseline of each period before and after crossover. The GLMM accounted for the presence of repeated measures as well as the crossover design, and no multiplicity adjustments were implemented. The GLMM included fixed effects for treatment, period and sequence, with subject as a random effect. Baseline values were included as covariates when relevant. An unstructured covariance matrix was used to model within-subject correlations, consistent with best practices for crossover designs.

Mean changes between baseline and the end of the treatment period for the primary outcome assessments are provided for both treatment periods for OEC and placebo. The means are based on the combination of the identical treatment arms during the crossover periods. Accordingly, each treatment period was compared with its specific baseline (ie, data from V0 (visit 0) served as baseline for the first treatment period, while data from V3 (visit 3) served as baseline for the second treatment period). Treatment effects are given as the ratio of the geometric mean (G_mean) of the percent change from baseline of OEC over placebo (G_mean OEC/G_mean placebo) with a 95% CI.

The safety population included all patients who received one or more dose of OEC. Statistical analyses were performed using SAS software (V.9.4, SAS Institute) without replacement of missing data. Subjects with at least one post baseline value were included in the modified ITT (mITT) analysis. No imputation was done; missing data were handled via mixed models (GLMM/GEE) on observed data only. Nominal p values <0.05 were defined as statistically significant, and no correction for multiple comparisons was made.

Patient and public involvement

Patients and/or the public were not involved in the design, conduct, reporting or dissemination plans of this research.

Results

Patients

Between April 2021 (first participant signed informed consent) and October 2022 (trial end), 45 individuals were randomised, and 43 completed both treatment sequences (mITT population). The five investigational sites recruited between 5 and 14 patients each. Of these 43 patients, 21 patients started with a placebo and thereafter crossed over to OEC (Placebo/OEC group), whereas 22 patients were initially on OEC and thereafter received a placebo (OEC/Placebo group). An overview of patient disposition is depicted in figure 2. Overall, major protocol deviations were infrequent, and a detailed summary is presented in the online supplemental file.

Figure 2. Patient disposition. Numbers represent the mITT population. mITT, modified intention-to-treat; OEC, oral enzyme combination.

Baseline demographics and characteristics were generally similar across both treatment groups, except that VAS knee pain intensity at walking was significantly higher in patients assigned to OEC/placebo than in patients assigned to placebo/OEC (table 1). Additionally, OEC/placebo patients had lower KOOS global scores than placebo/OEC patients, although this difference did not reach statistical significance.

Table 1. Baseline demographics and characteristics by intervention sequence.

Characteristic	All (N=43)	Placebo/OEC (n=21)	OEC/Placebo (n=22)
Age, years, mean (SD)	63.3 (10.4)	64.1 (12.2)	62.5 (8.5)
Sex, n (%)
Male	19 (44.2)	11 (52.4)	8 (36.4)
Female	24 (55.8)	10 (47.6)	14 (63.6)
Body mass index, kg/m2, mean (SD)	27.4 (3.5)	27.6 (3.3)	27.3 (3.7)
Blood pressure, mm Hg, mean (SD)
Systolic	128.4 (19.7)	129.1 (20.3)	127.8 (19.6)
Diastolic	80.1 (10.9)	78.2 (10.9)	81.9 (10.8)
Radiological Kellgren-Lawrence grade II or III, n (%)	43 (100)	21 (100)	22 (100)
VAS knee pain intensity, mm, mean (SD)
At rest	38.3 (21.7)	39.2 (16.7)	37.4 (25.9)
At walking	56.9 (13.3)	51.0 (11.2)	62.5 (12.9)
KOOS global score, %, mean (SD)	48.7 (12.9)*	52.6 (12.1)†	45.0 (12.9)‡

Patients assigned to OEC/placebo had significantly higher baseline VAS knee pain intensity at walking compared to patients assigned to placebo/OEC. OEC/placebo patients also had lower KOOS global scores than placebo/OEC patients, although this difference was not statistically significant.

^*n=39.

^†n=19.

^‡n=20.

KOOS, Knee injury and Osteoarthritis Outcome Score; NA, not applicable; OEC, oral enzyme combination; VAS, Visual Analogue Scale.

Across all patients, the mean (SD) age was 63.3 (10.4) years; most patients were female (n=24, 55.8%). All patients had a Kellgren-Lawrence grade of II or III, and their mean (SD) KOOS global score was 48.7 (12.9).

Primary outcome objectives

Markers of innate immune response

After the 8-week treatment period, baseline mean A2M increased from 1.60 (SD 0.37) to 1.62 (SD 0.36) g/L in the OEC group and decreased from 1.68 (SD 0.35) to 1.66 (SD 0.37) g/L in the placebo group. The observed total +3% between-group difference in change of A2M from baseline was statistically significant and in favour of OEC (G_mean OEC/G_mean placebo: 103%; 95% CI: 100% to 105%; p=0.038; figure 3A,B). No significant between-group differences were found for other innate immunity markers (p>0.05; online supplemental table S1).

Figure 3. Mean absolute changes during V0–V5 and mean per cent change from baseline in A2M (A, B), IL-10 (C, D) and uCTXII (E, F) in patients treated with OEC and placebo. A2M, α2-macroglobulin; G_mean, geometric mean; IL-10, interleukin-10; OEC, oral enzyme combination; uCTXII, urinary carboxyl-terminal cross-linked telopeptide of type II collagen; V, visit.

Markers of systemic inflammation

After the 8-week treatment period, baseline mean IL-10 increased from 208.5 (SD 0.57) to 302.3 fg/mL (SD 0.83) in the OEC group and from 208.0 (SD 0.53) to 213.6 (SD 0.45) fg/mL in the placebo group. The observed total +41% between-group difference in change of IL-10 from baseline was statistically significant and in favour of OEC (G_mean OEC/G_mean placebo: 141%; 95% CI: 122% to 162%; p<0.0001; figure 3C,D). No significant between-group differences were found for other markers of systemic inflammation (p>0.05; online supplemental tables S2 and S3).

Markers of cartilage turnover

After the 8-week treatment period, baseline mean uCTXII decreased from 493.2 (SD 0.63) to 488.3 (SD 0.83) ng/mL in the OEC group but increased from 485.4 (SD 0.79) to 556.6 (SD 0.70) ng/mL in the placebo group. The observed total –11% between-group difference in change of uCTXII from baseline was statistically significant and in favour of OEC (G_mean OEC/G_mean placebo: 89%; 95% CI: 80% to 99%; p=0.0381; figure 3E,F). No significant between-group differences were found for other serum collagen markers (p>0.05; online supplemental tables S4 and S5).

Secondary outcome objectives

Transforming growth factor-β

OEC treatment after 8 weeks increased TGF-β1 and TGF-β3 levels compared with placebo, although only the increase in TGF-β3 was significant (increase in the OEC group from baseline 26.23 (SD 0.48) to 30.74 (SD 0.66) ng/mL and in the placebo group a decrease from 27.32 (SD 0.42) to 26.17 (SD 0.44) ng/mL; +100% between-group difference, Gmean OEC/Gmean placebo: 200%; 95% CI: 165% to 244%; p=0.0012; figure 4; online supplemental tables S6 and S7). By contrast, TGF-β2 levels were decreased following OEC and placebo treatment, with the magnitude of the decrease significantly smaller in patients treated with OEC (p=0.0273).

Figure 4. Mean absolute changes during V0–V5 and mean percent change from baseline in TGF-β1 (A, B), TGF-β2 (C, D), and TGF-β3 (E, F) in patients treated with OEC and placebo. G_mean, geometric mean; OEC, oral enzyme combination; TGF, transforming growth factor; V, visit.

Knee injury and Osteoarthritis Outcome Score

After 8 weeks of treatment, scores in the KOOS Pain (increase in the OEC group from baseline 54.4 (SD 14.95) to 58.9 (SD 17.62)% and in the placebo group from 56.1 (SD 14.41) to 56.2 (SD 19.18)%; +3% between-group difference, Change in response mean OEC/Gmean placebo: 103%; 95% CI: 100% to 107%; p=0.0464) and Symptoms (increase in the OEC group from baseline 60.4 (SD 13.94) to 65.1 (SD 16.36)% and in the placebo group a decrease from 64.6 (SD 12.85) to 64.4 (SD 18.18)%; +4% between-group difference, Change in response mean OEC/Gmean placebo: 104%; 95% CI: 100% to 107%; p=0.0257) subdomains were significantly increased in the OEC group relative to placebo (figure 5), indicating improvements. No significant between-group differences were observed in the global score or other KOOS subdomains (p>0.05; online supplemental tables S8 and S9).

Figure 5. Mean per cent change from baseline in KOOS global and subdomain scores in patients treated with OEC and placebo. ADL, activities of daily living; KOOS, Knee injury and Osteoarthritis Outcome Score; OEC, oral enzyme combination; QoL, quality of life.

Patient-reported knee pain as measured by VAS

Changes in patient-reported knee pain from baseline to 8 weeks of treatment, both at rest and while walking, were comparable between the OEC and placebo groups (p>0.05; online supplemental tables S10 and S11).

Rescue medication

Compared with placebo, fewer individuals in the OEC group used rescue medication for general and/or knee pain, although this difference was not statistically validated (online supplemental table S12). Cumulative analgesic consumption scores, which integrate qualitative (potency) and quantitative (dose measurement) information to evaluate analgesic consumption,²⁸ did not differ between placebo and OEC-treated patients (p>0.05; online supplemental table S13).

Post hoc exploratory analyses

To further explore the effect of OEC treatment on the anti-inflammatory cytokine IL-10, we performed additional subgroup analyses by grouping patients according to baseline (V0) hsCRP tertiles (low, medium and high baseline hsCRP). Systemic IL-10 levels were significantly increased in patients treated with OEC whose baseline hsCRP levels were >0.9 mg/L but not in those with low baseline hsCRP (online supplemental table S14 and 15). Moreover, the ratio of a pro-inflammatory to anti-inflammatory systemic milieu (ie, IL-10/hsCRP) was significantly increased after OEC treatment vs placebo, suggesting a potential anti-inflammatory shift. Additional details of the exploratory analyses are available in the online supplemental file.

Safety and tolerability

Overall, treatment with OEC for 8 weeks was well tolerated, with no serious related AEs reported in either group. The number of patients experiencing at least one treatment-emergent AE (TEAE) was comparable between OEC (29/45) and placebo (38/45). TEAEs related to the investigational product were reported in 8 and 4 patients in the OEC and placebo groups, respectively (table 2).

Table 2. Summary of overall TEAEs in the safety population (N=45).

	Placebo (N=45)	OEC (N=45)
Patients with at least one AE, n (%)
TEAE	38 (84.4)	29 (64.4)
Mild TEAE	27 (60.0)	24 (53.3)
Moderate TEAE	18 (40.0)	14 (31.1)
Severe TEAE	3 (6.7)	1 (2.2)
TESAE	1 (2.2)	0 (0)
TEAE related to investigational product*	4 (8.9)	8 (17.8)

^*TEAE related to investigational product is defined as an AE that is ‘probably related’ or ‘related’ to the treatment.

AE, adverse event; OEC, oral enzyme combination; TEAE, treatment-emergent AE; TESAE, treatment-emergent serious AE.

The most frequent TEAE related to the investigational product in the OEC group was abdominal pain upper (table 3).

Table 3. TEAEs related to investigational product in the safety population by Preferred Term (N=45).

	Placebo	OEC
Total number of TEAEs related to investigational product	7	13
Abdominal distension	1	1
Abdominal pain upper	1	3
Arthralgia	0	1
Diarrhoea	3	1
Dyspepsia	1	0
Flatulence	1	1
Gastrointestinal disorder	0	1
Gastrointestinal pain	0	2
Increased appetite	0	1
Oral fungal infection	0	1
Pain in extremity	0	1

TEAE related to investigational product is defined as an AE that is ‘probably related’ or ‘related’ to the treatment.

AE, adverse event; OEC, oral enzyme combination; TEAE, treatment-emergent AE.

Discussion

This study explored the biological effects of OEC on markers of innate immunity, systemic inflammation and cartilage turnover, thereby fostering our understanding of the clinical effects of OEC in OA. Compared with placebo, OEC significantly elevated levels of A2M with a difference of 3% and the anti-inflammatory cytokine IL-10 with a difference of 41%, and significantly reduced levels of the urinary cartilage biomarker CTXII with a difference of −11%. Furthermore, OEC significantly increased TGF-β3 levels relative to placebo (100% difference), and while TGF-β2 levels were reduced in both OEC and placebo, the magnitude of reduction was significantly smaller in OEC (7% difference). Clinically, patients treated with OEC experienced greater improvements in the KOOS Symptoms (4% difference) and Pain (3% difference) subdomains. These differences persisted despite both treatment groups having used comparable amounts of rescue analgesia and reporting similar pain levels as measured by VAS, although it should be noted that baseline VAS knee pain intensity at walking was significantly higher in patients assigned to OEC/placebo than patients assigned to placebo/OEC. Overall, OEC was well tolerated with a comparable safety and tolerability profile to placebo.

A2M is a protease inhibitor and acts as an acute-phase protein.^{21 22} Here, we observed that OEC increased serum A2M levels—these results are consistent with previous data demonstrating that OEC consumption triggers the formation of intermediate forms of A2M.²⁰ Whether this systemic increase contributes to the clinical effects of OEC in OA is unknown, as we did not detect reductions in key pro-inflammatory cytokines nor did we measure intra-articular A2M levels. Moreover, the role of A2M in OA itself is unclear, as evidence indicating that A2M levels are reduced in synovial fluid or serum is lacking.³¹ However, it is thought to protect cartilage by binding to inflammatory cytokines and reducing levels of cartilage catabolic cytokines and enzymes including IL-1β, TNF-α and matrix metalloproteinases (MMPs).³¹ In fact, a fabricated form of simplified A2M has been investigated as a potential TNF-α inhibitor.³² While the numeric increase in A2M may seem modest, its statistical significance and biological role in immune regulation suggest that it could have meaningful implications for the treatment outcomes in the context of this study. Further research may be needed to explore the clinical significance of this change in A2M levels.

The significant increase in blood IL-10 levels relative to placebo following OEC treatment reported here is novel and indicates a potential systemic anti-inflammatory effect of OEC. IL-10, which has both anti-inflammatory and chondroprotective effects, is currently a target for gene therapy in OA.^{33 34} IL-10 in vitro can increase proteoglycan synthesis and decrease excess production of IL-1β and TNFα.³⁴ As TNFα is a prototypical pro-inflammatory cytokine, TNFα and IL-10 maintain a pro-inflammatory vs anti-inflammatory cytokine balance.³⁵ Accordingly, in patients at risk of developing knee OA secondary to ligament trauma, circulating IL-10 concentrations are compromised.³⁵ Moreover, compared with patients without OA, the regulatory T cells of patients with OA secrete IL-10 at reduced amounts, and patients at more advanced stages of the disease display further reductions in IL-10 levels.³⁶

uCTXII is an established biological marker for OA that reflects cartilage breakdown.³⁷ CTXII, a fragment originating from the telopeptide region of type II collagen, diffuses from the joint into the blood and is excreted through urine.^{38 39} Higher uCTXII levels are significantly associated with risk of hand, hip and knee OA; the incidence and progression of knee OA; and clinical symptoms and radiographic findings in knee OA.^{37 39 40} Its levels decrease following therapies such as intra-articular corticosteroid injections, risedronate, calcitonin or strontium ranelate.^{38 39} The reductions in uCTXII observed here suggest that OEC treatment reduces cartilage turnover, although we did not detect changes in other cartilage metabolism biomarkers. Given that uCTXII release involves a different set of MMPs than those of other type II collagen peptides such as Coll2-1, our results suggest that OEC may act differentially on distinct collagen degradation pathways and/or MMPs.^{41 42}

TGF-β2 and -β3 are essential growth factors for articular cartilage homeostasis, and their synthesis is dysregulated in OA.^{43 44} Accordingly, the significant effects in TGF-β2 and TGF-β3 reported in OEC-treated patients here may reflect a restoration of cartilage homeostasis in OA.

Improvements in the KOOS Pain and Symptoms subdomains were found in patients treated with OEC. The Pain subdomain includes information regarding the frequency and severity of pain experienced by patients during certain activities (eg, walking, bending their knee, standing upright).²⁶ By contrast, we did not detect OEC-related improvements in patient-reported pain VAS scores. These null results may be due to the small sample size, which may not have provided enough power to capture the minimally clinically important difference, considered to be 17 –28 mm, in VAS.⁴⁵ The observed improvements in the KOOS Sports/Recreation subdomain, which reflects the degree of difficulty in squatting, running, jumping and turning/pivoting on the injured knee,²⁶ and the increased use of rescue analgesics in the placebo group suggest that patients treated with OEC increased their activity levels without the need for these drugs. This observation is crucial, given the serious side effects of these drugs and the importance of physical activity in managing chronic pain.^{2 12 13}

OEC was generally well tolerated, with no serious related AEs reported in either OEC or placebo. The overall safety and tolerability profile of OEC is consistent with previous reports, with the most frequent TEAEs being gastrointestinal in nature.^{15 16 19}

Limitations

This trial was exploratory in nature and aimed to explore potential biological mechanisms of OEC in patients with OA.

We have therefore chosen to use an established biomarker, linked to pathophysiology of OA, to inform sample size considerations for this trial. As such, the study was not formally powered but had a sample size estimation based on a difference of 10% in Coll2-1NO2 values. To underscore our prestudy approach, we have performed a post hoc power calculation for the observed effect of OEC on IL-10, a key finding of this study. Based on the present study sample size and the difference in IL-10 between OEC and placebo, our study has 91% power to detect the observed effect size. Nevertheless, the study could be considered too small to detect effects on several clinical parameters, such as knee pain VAS as discussed above. In addition, we did not correct for multiple testing, which, in an adequately powered trial, would have been required to assess multiple endpoints. Finally, the clinical significance of the changes in A2M, IL-10, TGF-β2 and TGF-β3 levels is unclear.

Despite randomisation, patients between the two intervention sequences had significantly different knee pain VAS at walking at the first baseline. Moreover, KOOS global scores were lower in patients assigned to OEC/placebo versus those in placebo/OEC, although this difference was not statistically significant. Whether these slight differences in knee pain VAS and/or KOOS global scores impacted the results despite the crossover design is unknown as various knee OA phenotypes and endotypes may respond differently to treatment,⁴⁶ which we did not explicitly account for. This possibility is further emphasised by the results from the exploratory analysis, in which patients were stratified according to initial hsCRP levels, and the overall effects of OEC on IL-10 appeared to be mainly driven by patients with high baseline inflammation. However, caution should be taken when interpreting results from this subgroup analysis due to the small sample sizes.

Finally, this study was performed in patients with knee OA, and although many similarities exist in their pathogenesis, caution should be taken when extrapolating the findings from this study to hand or hip OA.⁴⁷

Future studies

The findings of this study suggest potential novel mechanisms of action of OEC in OA, and future studies may elucidate how these translate into clinical benefits. Further research should evaluate the role of A2M and IL-10 in OA and investigate whether OEC-induced elevations in TGF-β levels confer a protective effect in OA. Additionally, the net effects of increased A2M, IL-10 and TGF-β levels on joint tissue structure should be confirmed by MRI investigation.

We performed exploratory analyses to provide insights into which OA patient subgroups would benefit the most from OEC treatment. A large international initiative to explore OA phenotypes using clinical, imaging and biochemical markers is underway, and models to predict individuals whose OA will progress are encouraging.^{48 49} Our finding that baseline hsCRP may predict immunological responses (ie, IL-10 levels) to OEC may be clinically significant and should be further explored in patients whose OA is also likely to progress.

A key question is the impact of OEC on circulating pro-inflammatory cytokines as we did not observe changes in their levels following OEC treatment. Given that trypsin is endogenously produced by the pancreas, the effects of OEC reported here raise the question of whether trypsin levels are altered in knee OA.⁵⁰ Evaluating trypsin levels and/or its activity in patients with knee OA could help answer this question.

Appropriately powered studies for individual biomarkers could clarify if the absence of statistically significant differences between OEC and placebo in this study is due to sample size limitations or the lack of a biological effect. Finally, the reduction in uCTXII that we observed may indicate disease modification, but this reduction should be further investigated through long-term clinical trials.

Conclusions

This exploratory study identified biological effects of OEC that contribute to explain its potential clinical efficacy, adding to the literature the anti-inflammatory effect of OEC in OA with proposed potential mechanisms of action through IL-10 and TGF-β. The association between OEC treatment and reduced uCTXII suggests an effect on cartilage breakdown, although further confirmation is needed. The reported improvements in KOOS and the placebo-like safety profile of OEC highlight the promising clinical efficacy of OEC therapy in reducing pain and improving daily functioning in patients with OA. Larger, long-term clinical trials are required to explore if OEC could be a disease-modifying therapy for knee OA and which individuals are more likely to respond to OEC therapy.

Data availability statement

Data are available on reasonable request.