Sage Journals: Discover world-class research

Abstract

The senescence-associated secretory phenotype (SASP) contributes to tissue degeneration and inflammation, yet its role in osteoarthritis (OA)-related pain remains poorly understood. We hypothesized that circulating SASP markers would be associated with distinct OA-pain phenotypes, defined by pain impact and radiographic OA (ROA) severity. A subset of middle-to-older-aged adults (45–85 years) from a larger multi-site study (n = 169) self-reported pain impact – defined as the extent to which pain interferes with daily functioning – and underwent knee radiography and blood collection. Hierarchical cluster analysis was used to empirically identify OA-pain phenotypes based on combined pain impact and Kellgren-Lawrence (KL) grade. Plasma levels of four SASP markers (GDF-15, activin-A, Osteopontin (OPN), and IL-15) were quantified from whole blood samples. Among 169 participants, 35.5% reported high-impact chronic knee pain and 27.8% exhibited moderate-to-severe radiographic OA. Cluster analysis identified distinct ROA-pain phenotypes. GDF-15 levels were significantly elevated in non-Hispanic White females with early ROA and high-impact pain, with race- and sex-dependent differences. Activin-A levels were higher in non-Hispanic Black participants without pain or ROA and varied by sex in early ROA/low-impact pain phenotypes. Osteopontin levels were elevated in males compared to females within the same phenotype group. IL-15 levels showed no association with ROA-pain phenotypes but were higher in males and positively correlated with age. SASP factors, particularly GDF-15, Activin-A, and Osteopontin, demonstrated race- and sex-dependent associations with OA-pain phenotypes. These findings underscore the importance of demographic context in OA pathophysiology and support further investigation into SASP factors as potential biomarkers and therapeutic targets for OA-related pain.

Keywords

Cellular senescence SASP factors osteoarthritis chronic pain biological aging

Introduction

Osteoarthritis (OA) is a degenerative joint disease predominantly affecting the knees and hips, characterized by the breakdown of cartilage, leading to pain, stiffness, and reduced mobility.¹ It is the most common age-related disease and a leading cause of chronic pain and disability.² The condition predominantly affects older adults, with around 75% of those diagnosed being over the age of 55.³ Chronic pain associated with OA is often debilitating, contributing to limitations in daily activities and function, also known as high-impact chronic pain.⁴ Despite the high prevalence of OA, no disease-modifying osteoarthritis drugs are currently approved to halt or reverse the disease. Similarly, current pain management approaches are inadequate, particularly in older adults, in whom the use of opioids adds to concerns of polypharmacy.⁵ Studying the mechanisms of aging may provide insights into the pathologies of OA and its associated pain, as aging and OA share common biological pathways, including inflammation, mitochondrial dysfunction, and epigenetic alterations.⁶ Some have postulated that OA contributes to biological aging, in part due to pain and disability.⁷ Indeed, our work and others’ have shown that chronic pain, driven largely by high-impact chronic pain, is associated with markers of accelerated biological aging, including epigenetic aging,^8,9 and brain aging.¹⁰

Cellular senescence is involved in many age-related diseases, including OA.^11–14 It is characterized by a permanent proliferative arrest and apoptotic resistance.¹⁵ Senescent cells secrete pro-inflammatory cytokines, chemokines, growth factors, and proteases, collectively known as the senescence-associated secretory phenotypes (SASP).¹⁵ These SASP factors alter the surrounding cells and tissues, contributing to chronic inflammation, tissue degradation, and extracellular matrix remodeling, which are OA hallmarks and drive disease progression and pain.^15–18

Additionally, the transplantation of senescent cells into the knee joint area of mice induces pain-like behaviors and structural joint remodeling consistent with OA.^11,19 Studies have also shown that neuronal senescence in the dorsal root ganglion (DRG) plays a key role in pain modulation, with increased senescence markers (i.e. p16INK4a, p21) and SASP factors such as IL-6 observed after injury.²⁰ Aged DRG neurons exhibit these senescence markers, which are further exacerbated after peripheral nerve injury and correlate with nociception, suggesting that senescent DRG neurons contribute significantly to pain in aging and injury models.²⁰ Despite promising pre-clinical findings, the relationship between senescence markers and OA-related pain has not been explored in humans.

Therefore, in this study, we evaluated SASP factors in relation to clinical (i.e. self-reported) pain impact and radiographic OA (ROA) severity in middle-to-older aged adults. We selected the following SASP factors because of their strong relevance to senescent cells (SnCs) and the progression of OA. SnCs, through SASP, mediate cellular cross-talk and drive inflammation, extracellular matrix degradation and joint degeneration.^15,21 The factors we focused on are growth differentiation factor-15 (GDF-15), Activin-A, osteopontin (OPN), and interleukin-15 (IL-15); these are key mediators of these processes. GDF-15 is among the most upregulated genes in early OA and may drive chondrocyte senescence through MAPK14 activation.²² It also promotes angiogenesis, which contributes to inflammation and tissue degradation, indicating its role in synovial inflammation.²³ Activin-A is highly expressed in OA-affected cartilage compared to unaffected cartilage.²⁴ OPN induces the production of pro-inflammatory cytokines in human OA chondrocytes.²⁵ IL-15 has been identified as one of four circulating proteins associated with the onset of both knee and hand OA.²⁶

Because the discordance between clinical pain severity and OA severity has been well documented,^27–29 we employed cluster analysis to identify empirically derived radiographic OA-pain phenotypes. We hypothesized that SASP factors would be associated with radiographic OA-pain phenotypes, where individuals with the highest impact pain and greater radiographic OA severity would have the greatest circulating levels of SASP factors. Importantly, since cellular senescence levels are strongly associated with obesity and older age,^30–33 these were adjusted in our analysis along with additional biobehavioral factor (e.g. sex).

Methods

The present study consists of a cross-sectional analysis of data from middle-to-older age adults (45–85 years) with and without symptomatic knee OA who took part in a larger, multi-site study conducted at the University of Florida (UF; Gainesville, FL, USA) and the University of Alabama-Birmingham (UAB; Birmingham, AL, USA). Participants were recruited from the surrounding areas at both study sites using community- and clinic-based recruitment methods. Exclusion criteria consisted of conditions that could confound knee OA-related outcomes, preclude consent, or prevent completion of study procedures, such as rheumatic disease (e.g. rheumatoid arthritis, fibromyalgia), history of surgery to the index (i.e. most painful) knee, uncontrolled hypertension, neurological disease, heart diseases, peripheral neuropathy, pregnant or nursing, had more severe pain in body sites other than the index knee, chronic opioid use, hospitalization for psychiatric disorders in the preceding year, and diminished cognitive function. For this investigation, we included a random subset of participants in whom an array of senescence biomarkers was completed. The study was approved by the Institutional Review Boards at both sites, and all participants provided written informed consent.

Assessments

Graded Chronic Pain Scale (GCPS): Self-reported pain impact was determined with the GCPS, which was modified to be specific for knee pain. The GCPS consists of seven items that measure pain intensity and interference over the past 6 months. The scale allows the classification of pain severity into no pain (grade 0), low-impact chronic pain (grades 1–2), and high-impact chronic pain (grades 3–4).³⁴

Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC): Participants also completed the WOMAC questionnaire to evaluate knee pain, stiffness, and physical function. Total WOMAC score was calculated by adding the subscale scores, that is, 0–20 for pain, 0–8 for stiffness, and 0–68 for physical function.

Knee radiographs: Knee radiographs with posterior-anterior and lateral views were completed for all participants and used to determine Kellgren-Lawrence (KL) scores by a rheumatologist (RS) at each site. Radiographs were assigned the following scores: 0 = normal (no OA), 1 = doubtful, 2 = mild, 3 = moderate, and 4 = severe.³⁵ We reduced these categories into no radiographic OA (ROA, grade 0), early ROA (grades 1–2), and late-stage ROA (grades 3–4).

Blood Collection: Venous blood was withdrawn from participants in the EDTA tubes and inverted 8–10 times and stored on ice (<30 min) until centrifugation. Samples were centrifuged at 2300 rpm for 15 min at 4°C and immediately placed on ice, and aliquoted into a 2 ml storage vial, Fisher Scientific Cat# 12567502.

SASP-related biomarkers: Plasma specimens were collected at baseline and immediately stored at −80℃. The concentration of Activin-A, Growth Differentiation Factor 15 (GDF15), Osteopontin (OPN) and Interleukin 15 (IL-15) in EDTA-plasma samples were determined with commercially available test kits as indicated, using a MILLIPLEX^® Analyzer 3.1 xPONENT System (Luminex^® 200™) and analyzed by the MILLIPLEX^® Analyst software: GDF15, (Millipore, Burlington, Ma; MILLIPLEX^® MAP Human Aging Magnetic Bead Panel, Cat.# HAGE1MAG-20K); OPN, IL15 (R&D Systems, Minneapolis, MN; Human Luminex^® Discovery Assay – Magnetic Bead Panel, Cat#: LXSAHM-06). The concentrations of Activin-A was determined by ELISA, using a Synergy HT multidetection microplate reader (Agilent Technologies Inc./BioTek, Santa Clara, CA): Activin-A (R&D Systems; Quantikine ELISA Kit, Cat#: DAC00B). All assays were performed according to the manufacturer’s instructions.

Additional biobehavioral factors: Relevant potentially confounding factors were included in the analysis, including age, sex, race, body mass index (BMI), and study site.

Statistical analysis

Hierarchical cluster analysis was employed to identify empirical clusters based on the combined impact of self-reported pain and KL grade, excluding controls with no pain/no ROA. After confirming that GCPS grade and KL scores were not highly correlated (r ≤ 0.51), we used Ward’s clustering method with squared Euclidean distances as the similarity measure. The optimal cluster solution was established through visual inspection of the dendrogram and changes to the agglomeration coefficients. We then examined the cluster group differences across the clustering variables to determine the appropriateness and internal validity of the cluster solution, identifying empirically derived ROA-pain phenotypes, summarized in Figure 1.

Figure 1.

Phenotype clusters based on radiographic osteoarthritis (OA) severity and graded-chronic pain scale pain-impact.

Sample characteristics are presented as means and standard deviations or counts and percentages where appropriate. The ROA-pain clusters were compared across demographic and pain characteristics using chi-square tests for categorical variables and analysis of variance (ANOVA) for continuous variables. We examined the independent associations between SASP factors and ROA-pain group classifications using one-way analysis of covariance (ANCOVA), adjusting for age, sex, race, BMI, and study site as covariates. Given the established relationships between pain and demographic factors, we also tested two-way interactions between ROA-pain group and both sex and race. Statistical significance was defined as an alpha level of ≤0.05, with Bonferroni corrections applied for post hoc comparisons.

Results

Participant characteristics are shown in Table 1. In total, 35.5% of participants (n = 60) reported high-impact chronic knee pain, and 56.2% reported low-impact chronic pain (n = 95). Forty-seven (27.8%) participants showed moderate-to-severe radiographic OA (ROA) with KL grades 3–4. Participants had a mean age of 58.1 ± 7.8 (mean ± SD), 100 (59.2%) were female, and 92 (54.4%) identified as non-Hispanic Black. The mean BMI was 31.5 ± 6.9 kg/m².

Table 1.

Participant characteristics.

Characteristic	Total	Cluster 1 (No pain/No ROA)	Cluster 2 (Pain/No ROA)	Cluster 3 (Pain/Late ROA)	Cluster 4 (High-impact pain/Early ROA)	Cluster 5 (Low-impact pain/Early ROA)	Probability
	Mean ± SD	Mean ± SD	Mean ± SD	Mean ± SD	Mean ± SD	Mean ± SD
	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)
N (%)	169	14 (8.3)	53 (31.4)	47 (27.8)	20 (11.8)	35 (20.7)
Age (years)	58.1 ± 7.8	58.1 ± 9.1	56.9 ± 7.0	60.7 ± 8.5	57.4 ± 7.8	56.9 ± 7.0	0.113^a
Gender
Male	69 (40.8)	4 (28.6)	23 (43.4)	18 (38.3)	12 (60.0)	12 (34.3)	0.308^b
Female	100 (59.2)	10 (71.4)	30 (56.6)	29 (61.7)	8 (40.0)	23 (65.7)
Race
Non-Hispanic Black	92 (54.4)	9 (64.3)	30 (56.6)	22 (46.8)	9 (45.0)	22 (62.9)	0.483^c
Non-Hispanic White	77 (45.5)	5 (35.7)	23 (43.4)	25 (53.2)	11 (55.0)	13 (37.1)
Education
High school degree or less	74 (43.8)	4 (28.6)	28 (52.8)	21 (44.7)	9 (45)	12 (34.2)	0.314^c
Two-year college degree	33 (19.5)	4 (28.6)	4 (7.5)	11 (23.4)	6 (30)	8 (22.9)
Five-year college degree	36 (21.3)	3 (21.4)	12 (22.6)	10 (21.3)	4 (20)	7 (20)
Master’s degree or higher	26 (15.3)	3 (21.4)	9 (17)	5 (10.6)	1 (5)	8 (22.9)
Study site
UF	121 (71.6)	14 (100)	53 (100)	26 (55.3)	8 (40)	20 (57.1)	<0.001^b
UAB	48 (28.4)	0	0	21 (44.7)	12 (60)	15 (42.9)
BMI (kg/m²)	31.5 ± 6.9	28.7 ± 5.5	28.8 ± 5.9	35.5 ± 7.4	30.9 ± 5.4	31.6 ± 6.5	<0.001^a
WOMAC
Pain	6.3 ± 4.8	0.2 ± 0.6	5.8 ± 4.7	8.3 ± 4.2	9.2 ± 4.5	5.1 ± 4.0	<0.001^a
Stiffness	2.7 ± 2.2	0.1 ± 0.5	2.5 ± 2.1	3.9 ± 1.9	3.7 ± 2.1	2.0 ± 1.7	<0.001^a
Physical function	20.0 ± 15.7	0.8 ± 2.2	18.2 ± 14.6	28.4 ± 14.9	28.3 ± 13.5	14.4 ± 12.5	<0.001^a
Total	29.1 ± 22.1	1.1 ± 3.2	26.4 ± 20.8	40.5 ± 20.3	41.1 ± 19.4	21.7 ± 17.7	<0.001^a

ANOVA.

Chi-Square Test.

Fisher’s Exact Test.

Table 1 and Figure 1 depict the ROA-pain groups represented by individuals in cluster 1 (no pain/no ROA), cluster 2 (pain/no ROA), cluster 3 (pain/late ROA), cluster 4 (high-impact pain/early ROA), and cluster 5 (low-impact pain/early ROA). The clusters did not significantly differ in most sociodemographic characteristics. Notably, cluster 1 (no pain/no ROA) did not include male participants because all individuals meeting these criteria were female. BMI was significantly higher in individuals in cluster 3 (pain/late ROA) compared to all other clusters (Bonferroni-corrected p < 0.05), except for cluster 4 (high-impact pain/early ROA). Additionally, WOMAC scores were highest among individuals in cluster 3 (pain/late ROA), and cluster 4 (high-impact pain/early ROA), whose ratings were significantly higher than those in cluster 1 (no pain/no ROA); (Bonferroni-corrected p < 0.05 for both) and cluster 5 (low-impact pain/early ROA) (Bonferroni-corrected p < 0.01 for both).

GDF-15

ROA-pain phenotypes were associated with GDF-15 levels in a race- and/or sex-dependent manner; shown in Figure 2. A significant three-way interaction was found between OA-pain phenotypes, race, and sex on GDF-15 levels (F (3, 149) = 2.92, p = 0.036). Among non-Hispanic White females, cluster 4 (early ROA/high-impact pain) had significantly higher GDF-15 levels compared to cluster 1 (no pain/no ROA, Bonferroni-corrected p = 0.022), cluster 2 (pain/no ROA, Bonferroni-corrected p = 0.024), cluster 3 (pain/late ROA, Bonferroni-corrected p = 0.006), and cluster 5 (low-impact pain/early ROA, Bonferroni-corrected p = 0.005). No significant associations between ROA-pain clusters and GDF-15 levels were measured among non-Hispanic Black participants (all p’s > 0.05). In cluster 4 (early ROA/high-impact pain), Non-Hispanic White females had significantly higher circulating GDF-15 levels compared to non-Hispanic Black females (Bonferroni-corrected p = 0.006) while non-Hispanic White females had significantly higher GDF-15 levels than non-Hispanic White males (Bonferroni-corrected p = 0.044). On the other hand, non-Hispanic White females in cluster 5 (early ROA/low-impact pain) had significantly lower GDF-15 levels than non-Hispanic White males (Bonferroni-corrected p = 0.010). Finally, GDF-15 levels were also positively correlated with age (F (1, 149) = 6.17, p = 0.014), and BMI (F (1, 147) = 5.86, p = 0.017.

Figure 2.

Association between OA-pain phenotypes with GDF-15 levels. The x-axis includes individuals in cluster 1 (no pain/no ROA), cluster 2 (pain/no ROA), cluster 3 (pain/late ROA), cluster 4 (high-impact pain/early ROA), and cluster 5 (low-impact pain/early ROA). The y-axis indicates the corresponding plasma GDF-15 levels. (a) Among non-Hispanic White females, individuals with high-impact pain and early OA exhibited significantly elevated GDF-15 levels compared to all other phenotype clusters (p = 0.006). (b) No significant associations between OA-pain phenotypes and GDF-15 were observed among non-Hispanic Black participants.

Activin-A

ROA-pain phenotypes were associated with circulating Activin-A levels in a race- and/or sex-dependent manner. There was a significant two-way interaction between ROA-pain phenotype by race (F (4, 147) = 2.53, p = 0.043), where circulating Activin-A levels were significantly higher in non-Hispanic Black participants compared to non-Hispanic White participants within cluster 1 (no pain/no ROA, Bonferroni corrected p = 0.034, Figure 3). Also, there was a significant two-way interaction between ROA-pain phenotype by sex (F (4, 147) = 2.62, p = 0.037). While non-Hispanic White females in cluster 5 (early ROA/low-impact pain) had lower Activin-A levels compared to non-Hispanic White males (Bonferroni-corrected p = 0.05). Finally, Activin-A levels were also positively correlated with age (F (1, 149) = 9.44, p = 0.003) and BMI (F (1, 147) = 8.04, p = 0.005).

Figure 3.

Association between OA-pain phenotypes with Activin-A levels. The x-axis includes individuals in cluster 1 (no pain/no ROA), cluster 2 (pain/no ROA), cluster 3 (pain/late ROA), cluster 4 (high-impact pain/early ROA), and cluster 5 (low-impact pain/early ROA). The y-axis indicates the corresponding plasma activin-A levels. (a) Activin-A levels were significantly lower among individuals with low-impact pain and early OA compared to those with no pain and no OA, but only within female (p = 0.026), and (b) non-Hispanic Black (p = 0.037) subgroups.

Osteopontin

ROA-pain phenotypes were associated with circulating Osteopontin levels in a sex-dependent manner. There was a significant two-way interaction between ROA-pain phenotype by sex (F (4, 147) = 4.28, p = 0.003), where females in cluster 5 (early ROA/low-impact pain) had lower Osteopontin levels than males in the same cluster 5 (Bonferroni-corrected p = 0.003). Finally, Osteopontin levels were significantly associated with BMI (F(1, 147) = 4.36, p = 0.038; Figure 4).

Figure 4.

Association between OA-pain phenotypes with Osteopontin levels. The x-axis includes individuals in cluster 1 (no pain/no ROA), cluster 2 (pain/no ROA), cluster 3 (pain/late ROA), cluster 4 (high-impact pain/early ROA), and cluster 5 (low-impact pain/early ROA). The y-axis indicates the corresponding plasma osteopontin levels. A significant interaction between OA-pain phenotype and sex was also observed for Osteopontin levels (p = 0.003).

Il-15

There were no associations between ROA-pain phenotypes and circulating IL-15 levels (Supplemental Figure 1). However, there was a main effect of sex on IL-15 levels where males had significantly higher levels than females (F(1, 147) = 10.25, p = 0.002). Further, age was positively correlated with circulating IL-15 levels (F(1, 149) = 7.66, p = 0.006).

Discussion

Cellular senescence is linked to various age-related diseases, and senescent cells, along with their senescence-associated secretory phenotype (SASP) factors, significantly contribute to the progression of OA.³⁶ However, the relationship between senescence and OA-related pain has not been thoroughly investigated in humans. In this study, we examined whether circulating SASP factors were associated with OA-related pain by evaluating both clinical pain impact and radiographic knee OA severity in middle-aged and older adults with or at risk of knee OA. We identified distinct OA-pain phenotypes that differentiated individuals based on the level of pain impact (low- vs high-impact chronic pain) and the structural severity of knee OA, ranging from early to advanced OA disease, as determined by radiographs. These empirically derived phenotypes align with the well-documented discordance between clinical pain severity and radiographic OA findings.^27–29

Circulating levels of Growth Differentiation Factor-15 (GDF-15) varied among OA-pain phenotypes in a sex- and race-dependent manner, with the most striking finding being elevated levels in Non-Hispanic women with high-impact pain but early radiographic OA compared to all other groups, including controls and individuals with advanced radiographic disease. GDF-15, a stress-responsive cytokine member of the transforming growth factor-β (TGF-β) superfamily, is a key SASP protein strongly associated with aging and serves as a potent predictor of all-cause mortality,^37–46 including in persons with OA,⁴⁷ supporting its crucial role in biological aging processes. This sex-specific elevation is particularly relevant given that higher plasma GDF-15 levels predict accelerated decline in bone parameters in women but not men^48,49 and aligns with the well-established association between post-menopausal status and OA development, where hormonal changes may accelerate cellular senescence and inflammatory aging within joint tissues.^49,50 Mechanistically, GDF-15 drives OA progression through multiple pathways. Weng et al.²² demonstrated that GDF-15 is among the most upregulated genes in early OA and may drive chondrocyte senescence through MAPK14 activation. Beyond cartilage effects, GDF-15 facilitates angiogenesis, contributing to inflammation and degradation,²³ and is found in elevated concentrations in synovial fluid and membranes of OA patients,²² indicating its role in synovial inflammation. Studies have also shown that GDF-15 modulates nociception and neuronal excitability of dorsal root ganglia neurons in arthritic rats by regulating pain-related ion channels like Nav1.8,⁵¹ while elevated serum and muscle levels of GDF-15 are associated with chronic muscular pain, decreased skeletal muscle mass and function, and show inverse relationships with muscle strength in aging populations – all known risk factors for chronic muscle pain.⁵²

Further, elevated GDF-15 in women experiencing severe pain despite minimal radiographic changes may also reflect the limitations of X-ray assessments, which primarily visualize bony structures and fail to capture early cartilage degradation, synovial inflammation, or other soft tissue changes contributing to cellular stress. This finding underscores the critical importance of considering self-reported pain impact as an early alarm system indicating significant underlying pathological processes that require therapeutic intervention to prevent future joint damage and OA progression. Importantly, this GDF-15 elevation was not observed in groups with late radiographic OA, suggesting that in advanced disease stages cartilage involvement may differ, or that the senescence signal is obscured when high- and low-impact pain groups are combined, highlighting high-impact pain as a key differentiator of underlying senescence burden occurring before structural damage becomes radiographically apparent. Moreover, since inflammation of the synovium (i.e. synovitis) can be missed with radiographic assessments, the association between GDF-15 and pain observed here may reflect a more inflammatory phenotype of OA rather than structural disease severity. Ultimately, additional work incorporating more sensitive imaging and inflammatory measures will be necessary to clarify the underlying pathophysiologic drivers.

There were significant differences in Activin-A levels between ethnic/racial groups among the control participants, as well as sex differences observed in non-Hispanic White participants with early radiographic osteoarthritis (ROA) and low-impact pain. Previous studies have consistently reported that activin-A exerts detrimental effects on knee joints.^24,53 Zou et al.²⁴ found elevated serum levels of activin-A in OA patients compared to controls. Activin-A levels in both serum and synovial fluid were significantly associated with OA severity based on KL grading. Additionally, Alexander et al.⁵⁴ reported that activin-A is highly expressed in OA-affected cartilage compared to unaffected cartilage. However, previous studies have not considered a combination of radiographic OA and pain impact as in our study. Thus, it is possible that the relationship between activin-A and OA-pain phenotypes is not uniform across populations. Future larger studies are needed for clarification.

Males in the low-impact pain and late-stage osteoarthritis (OA) group exhibited higher levels of Osteopontin (OPN) compared to females. OPN promotes the production of pro-inflammatory cytokines, such as IL-6 and IL-8, in human OA chondrocytes, which may contribute to OA development.^25,55 Additionally, elevated OPN levels in both plasma and synovial fluid have been linked to greater radiographic severity and increased pain intensity in knee OA.^31,55–57 Further research is needed to explore how OPN levels vary between sexes during OA progression, particularly in relation to OA-related pain.

Although our study revealed new insights into OA-pain phenotypes linked to SASP factors, we did not find a significant association between IL-15 levels and OA-pain phenotypes. In contrast, Sun et al.³² reported that serum IL-15 levels were independently and positively associated with increased pain severity in knee OA patients, though no link was found with radiographic OA severity. Additionally, a proteomic case-control study identified IL-15 as one of four circulating proteins connected to both established OA and the onset of knee and hand OA.²⁶ These conflicting findings highlight the need for further research to clarify the role of IL-15 in OA-related pain.

As expected, we found significant associations between SASP factors and chronological age, although most of these associations lost significance after adjusting for other covariates. Our results also support the well-established link between SASP factors and body mass index (BMI). Obesity is known to increase certain SASP factors, potentially exacerbating osteoarthritis (OA) and related pain.⁵⁸ A multicenter study further showed that chronically obese patients face a higher risk of medial knee cartilage damage, which accelerates disease progression.⁵⁹

This study has several strengths. Firstly, we measured established SASP factors in a well-characterized cohort,^60,61 enabling meaningful comparisons across different studies and diverse populations. Another key strength is the use of cluster analysis to identify OA-pain phenotypes, which provided an empirical basis for differentiating individuals according to both pain impact and OA radiographic severity – reflecting the known disconnect between pain experience and radiographic findings. On the other hand, the cellular source of circulating SASP factors cannot be determined from our study, as these were measured in blood. Several factors, including obesity and biological aging, and comorbidities, may have overlapping systemic influences that share identical proteins to the reported SASP factors^30,62–64; as such, our models were adjusted for age and BMI. In addition, non-senescent cells, such as tissue resident immune cells, can also secrete overlapping SASP factors.^65,66 The direction of the relationship between circulating SASP factors and knee OA pain cannot be determined within our cross-sectional study design. Future research with larger cohorts is necessary to validate these findings and investigate the potential overlap between circulating and local (i.e. knee synovial fluid) SASP factors. Additionally, controlled preclinical models are essential to better understand how inflammatory SASP profiles arising from underlying conditions like age-related comorbidities intersect with specific tissues undergoing cellular senescence.

Conclusion

Joint pain is a hallmark of osteoarthritis (OA), though it does not always reflect the extent of tissue remodeling. While the mechanisms driving joint pain remain unclear, local inflammation is likely a critical contributor. Senescent cells may play a key role in OA pain, with evidence showing that clearing these cells before structural tissue changes can lead to pain relief.^67,68 Pain-inducing cytokines from synovial and other cell types may also contribute to this process. Understanding how senescence-associated secretory phenotype (SASP) factors cause pain and the involvement of synovial, immune, and other cells will deepen our knowledge of OA’s clinical features and support the development of targeted therapeutics.¹⁶ Our findings highlight that SASP factors, especially GDF-15, may be indicative of females experiencing high-impact pain in early OA, but not males, with these effects also varying by race. This underscores the importance of considering demographic factors in OA pathophysiology. Continued research is essential to validate GDF-15 as a potential biomarker and therapeutic target for OA-related pain.

Supplemental Material

sj-docx-1-mpx-10.1177_17448069261422994 – Supplemental material for Circulating cellular senescence biomarkers in persons with chronic knee osteoarthritis pain: An exploratory study

Supplemental material, sj-docx-1-mpx-10.1177_17448069261422994 for Circulating cellular senescence biomarkers in persons with chronic knee osteoarthritis pain: An exploratory study by Muhammad Abbas, Javier A Tamargo, Carlos J Cruz, Yutao Zhang, Stephanie Wohlgemuth, Kevin Wu, Li Chen, Roland Staud, Roger B Fillingim, Christiaan Leeuwenburgh and Yenisel Cruz-Almeida in Molecular Pain

Footnotes

Acknowledgements

We want to thank our research teams and participants for their efforts to this research.

Author contributions

Conceptualization: YCA, MA; Writing – original draft: YCA, MA, JAT; Investigation: MA, SW, KW, CL, RS; Formal analysis: YCA, MA, YZ, LC; Writing – Review & Editing: all authors; Supervision: YCA, CL.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Institute of Arthritis and Musculoskeletal and Skin Diseases of the National Institutes of Health under award number R90AR085802, the National Institute on Aging T32AG049673 and L32AG096730, and the University of Florida Pain Research & Intervention Center of Excellence.

ORCID iDs

Muhammad Abbas

Javier A Tamargo

Carlos J Cruz

Christiaan Leeuwenburgh

Yenisel Cruz-Almeida

Availability of data and materials

The data supporting this study’s findings are available from the corresponding author, YCA, upon reasonable request.*

Supplemental material

Supplemental material for this article is available online.

References

Cross

Smith

Hoy

Nolte

Ackerman

Fransen

Bridgett

Williams

Guillemin

Hill

Laslett

Jones

Cicuttini

Osborne

Vos

Buchbinder

Woolf

March

. The global burden of hip and knee osteoarthritis: estimates from the global burden of disease 2010 study. Ann Rheum Dis 2014; 73: 1323–1330.

GBD 2021 Diseases and Injuries Collaborators. Global incidence, prevalence, years lived with disability (YLDs), disability-adjusted life-years (DALYs), and healthy life expectancy (HALE) for 371 diseases and injuries in 204 countries and territories and 811 subnational locations, 1990-2021: a systematic analysis for the Global Burden of Disease Study 2021. Lancet 2024; 403: 2133–2161.

Institute for Health Metrics and Evaluation (IHME). GBD results, https://vizhub.healthdata.org/gbd-results/ (2024, accessed 26 November 2024).

Von Korff

Scher

Helmick

Carter-Pokras

Dodick

Goulet

Hamill-Ruth

LeResche

Porter

Tait

Terman

Veasley

Mackey

. United States National Pain Strategy for population research: concepts, definitions, and pilot data. J Pain 2016; 17: 1068–1080.

Nawai

. Chronic pain management among older adults: a scoping review. SAGE Open Nurs 2019; 5: 2377960819874259.

Loeser

Collins

Diekman

. Ageing and the pathogenesis of osteoarthritis. Nat Rev Rheumatol 2016; 12: 412–420.

Valdes

Stocks

. Osteoarthritis and ageing. Eur Med J 2018; 3: 116–123.

Tamargo

Strath

Cruz-Almeida

. High-impact pain is associated with epigenetic aging among middle-aged and older adults: findings from the health and retirement study. J Gerontol A Biol Sci Med Sci 2024; 79: glae149.

Cruz-Almeida

Johnson

Meng

Sinha

Rani

Yoder

Huo

Foster

Fillingim

. Epigenetic age predictors in community-dwelling adults with high impact knee pain. Mol Pain 2022; 18: 17448069221118004.

10.

Johnson

Buchanan

Laffitte Nodarse

Valdes Hernandez

Huo

Cole

Buford

Fillingim

Cruz-Almeida

. Cross-sectional brain-predicted age differences in community-dwelling middle-aged and older adults with high impact knee pain. J Pain Res 2022; 15: 3575–3587.

11.

Han

Wang

Ding

Zhang

. Cross-talk of inflammation and cellular senescence: a new insight into the occurrence and progression of osteoarthritis. Bone Res 2024; 12: 69.

12.

Mansfield

Ramponi

Gupta

Stevenson

Mathew

Barinda

Herbstein

Morsli

. Emerging insights in senescence: pathways from preclinical models to therapeutic innovations. NPJ Aging 2024; 10: 53.

13.

Zhang

Pitcher

Yousefzadeh

Niedernhofer

Robbins

Zhu

. Cellular senescence: a key therapeutic target in aging and diseases. J Clin Invest 2022; 132: e158450.

14.

Kaur

Farr

. Cellular senescence in age-related disorders. Transl Res 2020; 226: 96–104.

15.

Abbas

Cruz

Tamargo

Lipat

Fareed

Deaton

Gordon

Johnson

Sharma

Leeuwenburgh

Allen

Cruz-Almeida

. Exploring senescence markers as potential drivers of osteoarthritis pain in aging adults. Exp Gerontol 2025; 212: 112950.

16.

Jeon

David

Campisi

Elisseeff

. Senescent cells and osteoarthritis: a painful connection. J Clin Invest 2018; 128: 1229–1237.

17.

Coppe

Patil

Rodier

Sun

Munoz

Goldstein

Nelson

Desprez

Campisi

. Senescence-associated secretory phenotypes reveal cell-nonautonomous functions of oncogenic RAS and the p53 tumor suppressor. PLoS Biol 2008; 6: 2853–2868.

18.

Gil

Zheng

Lee

Shin

Hwang

Jang

Jeon

. Senolytic drugs relieve pain by reducing peripheral nociceptive signaling without modifying joint tissue damage in spontaneous osteoarthritis. Aging (Albany NY) 2022; 14: 6006–6027.

19.

Bradley

Weivoda

Hwang

Pirtskhalava

Decklever

Curran

Ogrodnik

Jurk

Johnson

Lowe

Tchkonia

Westendorf

Kirkland

. Transplanted senescent cells induce an osteoarthritis-like condition in mice. J Gerontol A Biol Sci Med Sci 2017; 72: 780–785.

20.

Donovan

Brewer

Bond

Laslavic

Pena Lopez

Colman

Jordan

Hansen

González

Pujari

de Lecea

Quarta

Kauer

Tawfik

. Aging and injury drive neuronal senescence in the dorsal root ganglia. Nat Neurosci 2025; 28: 985–997.

21.

Liu

Zhang

. Senescence in osteoarthritis: from mechanism to potential treatment. Arthritis Res Ther 2022; 24: 174.

22.

Weng

Pikatan

Setiawan

Yadav

Fong

Hsu

Yeh

Lee

. Role of GDF15/MAPK14 axis in chondrocyte senescence as a novel senomorphic agent in osteoarthritis. Int J Mol Sci 2022; 23: 7043.

23.

Henrotin

Pesesse

Lambert

. Targeting the synovial angiogenesis as a novel treatment approach to osteoarthritis. Ther Adv Musculoskelet Dis 2014; 6: 20–34.

24.

Zou

Liu

. Correlation of concentrations of activin A with occurrence and severity of knee osteoarthritis. J Musculoskelet Neuronal Interact 2018; 18: 320–322.

25.

Yang

Gao

Zhang

Luo

Xue

Lei

. Effects of osteopontin on the expression of IL-6 and IL-8 inflammatory factors in human knee osteoarthritis chondrocytes. Eur Rev Med Pharmacol Sci 2014; 18: 3580–3586.

26.

Scanzello

Umoh

Pessler

Diaz-Torne

Miles

Dicarlo

Potter

Mandl

Marx

Rodeo

Goldring

Crow

. Local cytokine profiles in knee osteoarthritis: elevated synovial fluid interleukin-15 differentiates early from end-stage disease. Osteoarthritis Cartilage 2009; 17: 1040–1048.

27.

Bedson

Croft

. The discordance between clinical and radiographic knee osteoarthritis: a systematic search and summary of the literature. BMC Musculoskeletal Disorders 2008; 9: 116.

28.

Finan

Buenaver

Bounds

Hussain

Park

Haque

Campbell

Haythornthwaite

Edwards

Smith

. Discordance between pain and radiographic severity in knee osteoarthritis: findings from quantitative sensory testing of central sensitization. Arthritis Rheum 2013; 65: 363–372.

29.

Sethi

Anand

Della Pasqua

. Clinical assessment of osteoarthritis pain: contemporary scenario, challenges, and future perspectives. Pain Ther 2024; 13: 391–408.

30.

Palmer

St Sauver

Fielding

Atkinson

White

McGree

Weston

LeBrasseur

. The influence of body mass index on biomarkers of cellular senescence in older adults. J Gerontol A Biol Sci Med Sci 2025; 80: glae251.

31.

Mohammed

Abd El-Azeem

KamalElDin

. Plasma and synovial fluid osteopontin levels in patients with knee osteoarthritis: relation to radiological grade. Egypt Rheumatol 2012; 34: 131–136.

32.

Sun

Liu

Shi

. Serum interleukin-15 levels are associated with severity of pain in patients with knee osteoarthritis. Dis Markers 2013; 35: 203–206.

33.

Spinelli

Baboota

Gogg

Beguinot

Blüher

Nerstedt

Smith

. Increased cell senescence in human metabolic disorders. J Clin Invest 2023; 133: e169922.

34.

Von Korff

Ormel

Keefe

Dworkin

. Grading the severity of chronic pain. Pain 1992; 50: 133–149.

35.

Kohn

Sassoon

Fernando

. Classifications in brief: Kellgren-Lawrence classification of osteoarthritis. Clin Orthop Relat Res 2016; 474: 1886–1893.

36.

Maurer

Kirsch

Ruths

Brenner

Riegger

. Senolytic therapy combining Dasatinib and Quercetin restores the chondrogenic phenotype of human osteoarthritic chondrocytes by the release of pro-anabolic mediators. Aging Cell 2025; 24: e14361.

37.

Rothenbacher

Dallmeier

Christow

Koenig

Denkinger

Klenk

; ActiFE study group. Association of growth differentiation factor 15 with other key biomarkers, functional parameters and mortality in community-dwelling older adults. Age Ageing 2019; 48: 541–546.

38.

Wallentin

Zethelius

Berglund

Eggers

Lind

Lindahl

Wollert

Siegbahn

. GDF-15 for prognostication of cardiovascular and cancer morbidity and mortality in men. PLoS One 2013; 8: e78797.

39.

Daniels

Clopton

Laughlin

Maisel

Barrett-Connor

. Growth-differentiation factor-15 is a robust, independent predictor of 11-year mortality risk in community-dwelling older adults: the Rancho Bernardo Study. Circulation 2011; 123: 2101–2110.

40.

Wiklund

Bennet

Magnusson

Eriksson

Lindmark

Yaghoutyfam

Marquis

Stattin

Pedersen

Adami

Gronberg

Breit

Brown

. Macrophage inhibitory cytokine-1 (MIC-1/GDF15): a new marker of all-cause mortality. Aging Cell 2010; 9: 1057–1064.

41.

Wang

Wollert

Larson

Coglianese

McCabe

Cheng

Fradley

Ghorbani

Xanthakis

Kempf

Benjamin

Levy

Vasan

Januzzi

. Prognostic utility of novel biomarkers of cardiovascular stress: the Framingham Heart Study. Circulation 2012; 126: 1596–1604.

42.

Schaub

Reichlin

Twerenbold

Reiter

Steuer

Bassetti

Stelzig

Wolf

Winkler

Haaf

Meissner

Drexler

Mueller

. Growth differentiation factor-15 in the early diagnosis and risk stratification of patients with acute chest pain. Clin Chem 2012; 58: 441–449.

43.

Dallmeier

Brenner

Mons

Rottbauer

Koenig

Rothenbacher

. Growth differentiation factor 15, its 12-month relative change, and risk of cardiovascular events and total mortality in patients with stable coronary heart disease: 10-year follow-up of the KAROLA Study. Clin Chem 2016; 62: 982–992.

44.

Wollert

Kempf

Peter

Olofsson

James

Johnston

Lindahl

Horn-Wichmann

Brabant

Simoons

Armstrong

Califf

Drexler

Wallentin

. Prognostic value of growth-differentiation factor-15 in patients with non-ST-elevation acute coronary syndrome. Circulation 2007; 115: 962–971.

45.

Hagstrom

Held

Stewart

Aylward

Budaj

Cannon

Koenig

Krug-Gourley

Mohler

3rd Steg

Tarka

Ostlund

White

Siegbahn

Wallentin

; STABILITY Investigators. Growth differentiation factor 15 predicts all-cause morbidity and mortality in stable coronary heart disease. Clin Chem 2017; 63: 325–333.

46.

Wallentin

Lindholm

Siegbahn

Wernroth

Becker

Cannon

Cornel

Himmelmann

Giannitsis

Harrington

Held

Husted

Katus

Mahaffey

Steg

Storey

James

; PLATO study group. Biomarkers in relation to the effects of ticagrelor in comparison with clopidogrel in non-ST-elevation acute coronary syndrome patients managed with or without in-hospital revascularization: a substudy from the Prospective Randomized Platelet Inhibition and Patient Outcomes (PLATO) trial. Circulation 2014; 129: 293–303.

47.

Arnold

Rehm

Buchele

Peter

Brenner

Gunther

Brenner

Koenig

Rothenbacher

. Growth differentiation factor-15 as a potent predictor of long-term mortality among subjects with osteoarthritis. J Clin Med 2020; 9: 3107.

48.

Osawa

Tanaka

Semba

Fantoni

Moaddel

Candia

Simonsick

Bandinelli

Ferrucci

. Plasma growth and differentiation factor 15 predict longitudinal changes in bone parameters in women, but not in men. J Gerontol A Biol Sci Med Sci 2022; 77: 1951–1958.

49.

Gilmer

Iijima

Hettinger

Jackson

Bergmann

Bean

Shahshahan

Creed

Kopchak

Wang

Houston

Franks

Calderon

St Croix

Thurston

Evans

Ambrosio

. Menopause-induced 17β-estradiol and progesterone loss increases senescence markers, matrix disassembly and degeneration in mouse cartilage. Nat Aging 2025; 5: 65–86.

50.

Tian

Zhang

. The association between sex hormones and prevalence of OA in US adults. Front Med (Lausanne) 2024; 11: 1425210.

51.

Kim

Kang

Lim

. Growth differentiation factor-15 as a biomarker for sarcopenia in aging humans and mice. Exp Gerontol 2020; 142: 111115.

52.

León-González

Ortolá

Carballo-Casla

Sotos-Prieto

Buño-Soto

Rodríguez-Sánchez

Pastor-Barriuso

Rodríguez-Artalejo

García-Esquinas

. Growth differentiation factor 15 as a biomarker of cardiovascular risk in chronic musculoskeletal pain. J Gerontol A Biol Sci Med Sci 2024; 79: glae163.

53.

Tarabeih

Kalinkovich

Shalata

Higla

Livshits

. Pro-inflammatory biomarkers combined with body composition display a strong association with knee osteoarthritis in a community-based study. Biomolecules 2023; 13: 1315.

54.

Alexander

Watt

Sawaji

Hermansson

Saklatvala

. Activin A is an anticatabolic autocrine cytokine in articular cartilage whose production is controlled by fibroblast growth factor 2 and NF-κB. Arthritis Rheum 2007; 56: 3715–3725.

55.

Min

Shi

Han

Chen

Shi

Teng

Jiang

. Serum levels of leptin, osteopontin, and sclerostin in patients with and without knee osteoarthritis. Clin Rheumatol 2021; 40: 287–294.

56.

Honsawek

Tanavalee

Sakdinakiattikoon

Chayanupatkul

Yuktanandana

. Correlation of plasma and synovial fluid osteopontin with disease severity in knee osteoarthritis. Clin Biochem 2009; 42: 808–812.

57.

Gao

Zeng

Lei

. Elevated osteopontin level of synovial fluid and articular cartilage is associated with disease severity in knee osteoarthritis patients. Osteoarthritis Cartilage 2010; 18: 82–87.

58.

Wei

Umar

Zhu

Speakman

Chen

Tong

Chen

. Risk of metabolic abnormalities in osteoarthritis: a new perspective to understand its pathological mechanisms. Bone Res 2023; 11: 63.

59.

Voinier

Neogi

Stefanik

Guermazi

Roemer

Thoma

Master

Nevitt

Lewis

Torner

White

. Using cumulative load to explain how body mass index and daily walking relate to worsening knee cartilage damage over two years: the MOST Study. Arthritis Rheumatol 2020; 72: 957–965.

60.

Cruz-Almeida

Cardoso

Riley

3rd Goodin

King

Petrov

Bartley

Sibille

Glover

Herbert

Bulls

Addison

Staud

Redden

Bradley

Fillingim

. Physical performance and movement-evoked pain profiles in community-dwelling individuals at risk for knee osteoarthritis. Exp Gerontol 2017; 98: 186–191.

61.

Cardoso

Riley

3rd Glover

Sibille

Bartley

Goodin

Bulls

Herbert

Addison

Staud

Redden

Bradley

Fillingim

Cruz-Almeida

. Experimental pain phenotyping in community-dwelling individuals with knee osteoarthritis. Pain 2016; 157: 2104–2114.

62.

Azamar-Llamas

Hernández-Molina

Ramos-Ávalos

Furuzawa-Carballeda

. Adipokine contribution to the pathogenesis of osteoarthritis. Mediators Inflamm 2017; 2017: 5468023.

63.

Frasca

Blomberg

Paganelli

. Aging, obesity, and inflammatory age-related diseases. Front Immunol 2017; 8: 1745.

64.

Chaib

Tchkonia

Kirkland

. Obesity, senescence, and senolytics. Handb Exp Pharmacol 2022; 274: 165–180.

65.

Chung

Cesari

Anton

Marzetti

Giovannini

Seo

Carter

Leeuwenburgh

. Molecular inflammation: underpinnings of aging and age-related diseases. Ageing Res Rev 2009; 8: 18–30.

66.

Yue

Nie

Zhao

Liao

Wang

. Senescence-associated secretory phenotype and its impact on oral immune homeostasis. Front Immunol 2022; 13: 1019313.

67.

Jeon

Kim

Laberge

Demaria

Rathod

Vasserot

Chung

Kim

Poon

David

Baker

van Deursen

Campisi

Elisseeff

. Local clearance of senescent cells attenuates the development of post-traumatic osteoarthritis and creates a pro-regenerative environment. Nat Med 2017; 23: 775–781.

68.

Ruan

Jin

Sun

Jinghu

Lin

Zhang

Chen

. IGF signaling pathway in bone and cartilage development, homeostasis, and disease. FASEB J 2024; 38: e70031.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.31 MB