Research ArticlePsoriatic Arthritis
Open Access

Candidate Biomarkers for Response to Treatment in Psoriatic Disease

Rachel Offenheim, Omar F. Cruz-Correa, Darshini Ganatra and Dafna D. Gladman
The Journal of Rheumatology December 2024, 51 (12) 1176-1186; DOI: https://doi.org/10.3899/jrheum.2024-0396

Abstract

Objective To determine whether biologic therapy alters serum C-X-C motif chemokine ligand 10 (CXCL10), matrix metalloproteinase 3 (MMP3), S100 calcium-binding protein A8 (S100A8), acid phosphatase 5 (ACP5), and C-C motif chemokine ligand 2 (CCL2) levels in patients with psoriatic arthritis (PsA) and cutaneous psoriasis without arthritis (PsC), and whether baseline levels of these proteins predict response to treatment for PsA.

Methods We included (1) patients with PsA taking tumor necrosis factor inhibitors (TNFi), interleukin 17 inhibitors (IL-17i), methotrexate (MTX), and those who were untreated with bDMARDs or csDMARDs; (2) patients with PsC taking bDMARDs; and (3) matched patients with PsC who were not treated with bDMARDs or csDMARDs. Serum samples at baseline and at the 3- to 6-month follow-up visit were retrieved from the biobank. Protein levels were quantified using a Luminex multiplex assay. We compared follow-up vs baseline protein levels within groups and change in levels between groups. For the predictive potential of the biomarkers, we developed logistic regression classification models. Response to treatment was defined as (1) achieving low disease activity or remission (according to the Disease Activity Index for Psoriatic Arthritis); (2) ≥ 75% reduction in Psoriasis Area and Severity Index; and (3) ≥ 50% reduction in actively inflamed joint count.

Results In PsA, TNFi reduced serum levels of all 5 proteins, IL-17i increased ACP5 and CCL2, and MTX reduced MMP3. Changes in MMP3 and S100A8 levels were significantly different between untreated PsA and matched biologic-treated PsA (P < 0.05). There were no significant differences between treated or untreated patients with PsC. Baseline levels of CXCL10, MMP3, S100A8, and ACP5 had good predictive value (area under the curve > 0.80) for response to biologics in patients with PsA.

Conclusion Treatment with biologics and MTX affect serum CXCL10, MMP3, S100A8, ACP5, and CCL2 levels in patients with PsA. MMP3, S100A8, ACP5, and CXCL10 have potential use as serum biomarkers to predict response to treatment for PsA.

Key Indexing Terms:

Psoriasis is an inflammatory skin disease mediated by activation of T lymphocyte and innate immune cells leading to keratinocyte hyperproliferation.1,2 Activated T cells produce tumor necrosis factor-α (TNF-α) and interleukins (IL; IL-23, IL-12, and IL-6).2 Subsequently, keratinocytes cause more inflammation through the release of cytokines and chemokines including chemokine ligand 20 (CCL20), C-X-C motif chemokine ligand 5 (CXCL5), CXCL8, CXCL8, and CXCL10.3

Psoriatic arthritis (PsA), affecting roughly 30% of patients with psoriasis,4 is an inflammatory arthritis that affects both peripheral joints and the axial skeleton.5 PsA causes increased vascularization and inflammation of the synovial membrane.6 Release of proinflammatory molecules from the invading immune cells causes the activation of fibroblast-like synoviocytes that invade adjacent cartilage and bone.7

Treatment options for PsA include nonsteroidal antiinflammatory drugs (NSAIDs), conventional disease-modifying antirheumatic drugs (csDMARDs) such as methotrexate (MTX), and biologic DMARDs (bDMARDs) that inhibit cytokines such as TNF-α, IL-17A, and IL-23.8

Several biomarkers that may aid in the prediction of treatment response of PsA have been proposed. Among them, CXCL10 was significantly higher in patients with psoriasis who develop PsA compared to those that do not.9 Matrix metalloproteinase 3 (MMP3) serum levels are higher in patients with PsA compared to healthy controls.10 S100 calcium-binding protein A8 (S100A8) showed predictive ability of response to tumor necrosis factor inhibitors (TNFi) in synovial tissue of patients with PsA.11 Acid phosphatase 5 (ACP5, also known as tartrate-resistant acid phosphatase [TRAP]) is differentially expressed between TNFi responders and nonresponders.12 C-C motif chemokine ligand 2 (CCL2) is increased in patients with PsA and promotes the migration of T cells into the synovium contributing to the joint damage seen in PsA.13,14

The objectives of our study were to determine whether bDMARD therapy and MTX treatment for psoriatic disease alter the levels of CXCL10, MMP3, S100A8, ACP5, and CCL2, and whether baseline levels of CXCL10, MMP3, S100A8, ACP5, and CCL2 predict response to treatment in patients with PsA. These biomarkers have been shown previously to be predictive for drug response and could be multiplexed in our assay system.

METHODS

Study subjects. Patients with PsA and cutaneous psoriasis without arthritis (PsC) were recruited from a prospective study, and data analysis was performed retrospectively. The research was conducted in accordance with Good Clinical Practice standards and the guidelines of the Declaration of Helsinki. The study protocol was approved by the University Health Network Research Ethics Board in Toronto, Canada. All patients with PsA were diagnosed by a rheumatologist and fulfilled the Classification for Psoriatic Arthritis (CASPAR) criteria.15 All patients with PsC were diagnosed by a dermatologist and were reviewed by a rheumatologist to exclude PsA. Serum was collected from patients with PsA at baseline (prior to treatment) and at a follow-up visit 3 to 6 months after beginning biologic treatment (n = 108) or MTX (n = 23). Patients not receiving treatment with biologics or MTX were also included (n = 75). Serum was also collected from patients with PsC not receiving treatment (n = 28) and those receiving biologic treatment (n = 28), matched 1:1 on age, sex, and disease duration. Serum samples were aliquoted and stored in the serum biobank at −80 °C.

Multiplex protein quantification assay. Levels of 5 protein biomarkers (CXCL10, MMP3, S100A8, ACP5, and CCL2) were measured in baseline and follow-up serum samples from 205 patients with PsA and in baseline serum samples from 56 patients with PsC. The serum samples were diluted 1:1 using the diluent provided by the Luminex Discovery Assay kit (R&D Systems). The CXCL10, MMP3, S100A8, CCL2, and ACP5 levels were quantified using a Luminex Discovery Assay following the manufacturer’s protocols. The duplicated samples were run and quantified relative to a 3-fold serially diluted standard, using a 5-parameter logistic regression curve.

Statistical analysis. Analysis for this study was performed separately for 4 cohorts: (1) biologics (comprising 92 TNFi-treated patients with PsA and 16 IL-17i), (2) MTX (comprising 22 MTX-treated and 22 paired biologic-treated patients with PsA from cohort 1), (3) untreated PsA (75 untreated and 75 paired biologic-treated patients with PsA from cohort 1), and (4) PsC (28 untreated and 28 paired biologic-treated patients with PsC).

The statistical analysis was performed using the software R, version 4.1.0 (R Foundation for Statistical Computing). Protein measurements out of range were adjusted to the minimum and maximum value observed across all cohorts (specifically, MMP3 levels were adjusted to a maximum of 89,140.93 pg/mL and S100A8 levels to a minimum of 0.67 pg/mL). A Shapiro-Wilk test was used to evaluate the data for significant deviations from the normal distribution. Then, a paired (or unpaired) t test or Mann-Whitney U test was selected as appropriate to compare (1) follow-up vs baseline protein levels within groups of patients with PsA, (2) changes in protein levels following treatment between PsA patient groups and matched bDMARD-treated patients, and (3) patients with PsC who were untreated with bDMARDs or csDMARDs and paired bDMARD-treated patients with PsC (in which protein levels were measured only once). We compared the protein changes between patients stratified by sex and performed an ANCOVA analysis with sex as a covariate to assess the effect of sex on protein biomarker levels between treatment groups. All analyses were considered significant when P ≤ 0.05.

Correlation with disease activity. At each visit, patients undergo a complete assessment, which includes the number of tender and swollen joints as well as the Psoriasis Area and Severity Index (PASI).16 Patients also complete patient-reported outcome questionnaires.17 Disease activity measures included the following: (1) Disease Activity Index for Psoriatic Arthritis (DAPSA)—which consists of tender and swollen joint counts (of 68 and 66 joints, respectively), C-reactive protein levels, patient global assessment (0-10), and patient assessment of pain (0-10)—for which a score < 14 reflects low disease activity (LDA) and achieving this was considered response18; (2) ≥ 75% reduction in PASI score (PASI75); and (3) actively inflamed joint count (out of 68 tender and 66 swollen joints).19 The association between protein levels and disease activity was assessed by calculating the Spearman correlation coefficient.

Prediction of treatment response. Treatment response in patients with PsA was defined according to 3 criteria: (1) reduction of DAPSA resulting in LDA (DAPSA < 14) or remission (DAPSA < 4), (2) PASI75, and (3) reduction of ≥ 50% in actively inflamed joint count (tender and/or swollen joints; Supplementary Table S1, available with the online version of this article).

The predictive performance of baseline protein levels was assessed by logistic regression classification models. Logistic regression models were trained using 80% of the samples, keeping the other 20% as an independent set for evaluating the prediction of treatment response. Models were built using the natural logarithm of the baseline protein levels alone and in combination with age and sex. To assess the performance of each model, we calculated the area under the receiver-operating characteristic curve (AUC). An AUC ≥ 0.80 was considered to be good predictive performance.

RESULTS

The clinical characteristics of the patients with PsA and PsC are shown in Table 1 and Table 2, respectively. These were representative of patients with psoriatic disease, with a mean age of close to 50 years, mostly male, and a disease duration of 22 years for PsC and 11 years for PsA. Baseline and follow-up samples from a total of 205 patients with PsA were included in the study, including 92 treated with TNFi, 16 with IL-17i, 22 treated with MTX, and 75 patients with PsA not treated with bDMARDs or csDMARDs. Another 28 patients with PsC who were not treated with bDMARDs and csDMARDs and 28 matched biologic-treated patients with PsC were also included in the study.

View this table:
Table 1.

Characteristics of patients with PsA at baseline.

View this table:
Table 2.

Characteristics of patients with PsC at baseline.

Biomarker levels in biologic-treated PsA. In the cohort of patients with biologic-treated PsA (TNFi and IL-17i combined), there were significant decreases in CXCL10 (P < 0.01), MMP3 (P < 0.001), S100A8 (P < 0.001), and ACP5 (P < 0.05) at follow-up, after treatment with biologics. Upon stratification by biologic treatment type, CXCL10 (P < 0.001), MMP3 (P < 0.001), S100A8 (P < 0.001), ACP5 (P < 0.001), and CCL2 (P < 0.05) showed significant decreases in TNFi-treated patients; conversely, ACP5 (P < 0.01) and CCL2 (P < 0.05) showed significant increases in IL-17i–treated patients (Figure 1; Supplementary Table S2, available with the online version of this article). A direct comparison of protein changes between IL-17i– and TNFi-treated patients highlighted significant differences for CXCL10 (P < 0.05), S100A8 (P < 0.05), ACP5 (P < 0.001), and CCL2 (P < 0.01).

Figure 1.
Figure 1.

Biomarker protein levels at baseline and follow-up in patients with PsA treated with (A) TNFi, (B) IL-17i, (C) MTX, and (D) no bDMARDs or csDMARDs. * P < 0.05. ** P < 0.01.*** P < 0.001. **** P < 0.0001. bDMARD: biologic disease-modifying antirheumatic drug; csDMARD: conventional synthetic disease-modifying antirheumatic drug; IL-17i: interleukin 17 inhibitor; MTX: methotrexate; ns: not significant; TNFi: tumor necrosis factor inhibitor.

In TNFi-treated patients with PsA, the reductions in protein levels were significantly larger in male vs female individuals for S100A8 (P < 0.05) and ACP5 (P = 0.01; Supplementary Table S3, available with the online version of this article). When corrected for sex, reductions in protein levels at follow-up remained significant for S100A8 (sex F[1,181] = 37.48, P < 0.001; timepoint F[1,181] = 23.06, P < 0.001) and ACP5 (sex F[1,181] = 12.94, P < 0.001; timepoint F[1,181] = 6.63, P = 0.01). In IL-17i–treated patients with PsA, the changes in protein levels were nonsignificantly different between male and female individuals for all proteins.

Comparing the change in protein levels between IL-17i- and TNFi-treated patients, there was a significant effect of sex on MMP3 change (F[1,105] = 4.22; P < 0.05), and S100A8 change (F[1,105] = 7.53; P < 0.01). When adjusted for these effects, the decrease in S100A8 levels still showed significant differences between IL-17i- and TNFi-treated patients (F[1,105] = 4.99; P < 0.05), whereas the decrease in MMP3 was not significantly different between treatment groups after adjustment for sex (F[1,105] = 0.91; P = 0.34).

Biomarker levels in MTX-treated patients with PsA. MMP3 protein levels showed a significant decrease (P < 0.05) at follow-up after treatment with MTX. There were no significant differences between the baseline and follow-up levels of CXCL10, S100A8, ACP5, and CCL2 (Figure 1; Supplementary Table S2, available with the online version of this article). In MTX-treated patients with PsA, the changes in protein levels were significantly different between sexes, increasing in male patients and decreasing in female patients for CCL2 (P < 0.01), but were nonsignificant for all other proteins (Supplementary Table S3). When correcting for sex, MMP3 levels were not significantly different between baseline vs follow-up (sex F[1,41] = 16.85; P < 0.001; timepoint F[1,41] = 0.95; P = 0.34); as well, CCL2 levels were not significantly different at follow-up after correcting for sex (sex F[1,41] = 0.002, P > 0.99; timepoint F[1,41] = 1.47, P = 0.23).

A direct comparison of protein changes between MTX-treated and paired biologic-treated patients showed no significant differences between treatment groups. These results indicate that MTX treatment reduces the levels of all proteins comparably to biologics treatment. Further, when comparing the change in protein levels between MTX-treated and matched biologic-treated patients, there was no significant effect of sex on the change in levels for all proteins.

Biomarker levels in patients with PsA and PsC not treated with bDMARDs or csDMARDs. In the no bDMARD or csDMARD PsA cohort, S100A8 (P < 0.05) and ACP5 (P < 0.001) were significantly decreased, whereas CCL2 (P < 0.05) protein levels were significantly increased at follow-up compared to baseline.

In the matched biologic-treated PsA cohort, MMP3 (P < 0.001), S100A8 (P < 0.001), and ACP5 (P < 0.05) were significantly decreased at follow-up compared to baseline (Figure 1; Supplementary Table S2, available with the online version of this article). The decreases in MMP3 (P < 0.01) and S100A8 (P < 0.05) were significantly larger in the biologic-treated cohort compared to the patients with PsA not treated with bDMARDs and csDMARDs, whereas these were not significantly different for CXCL10, ACP5, and CCL2. In patients with PsA not treated with bDMARDs and csDMARDs, the changes in protein levels were nonsignificantly different between male and female individuals for all proteins (Supplementary Table S3), and there was no significant effect of sex on the comparison of changes in levels of all proteins between patients with PsA not treated with bDMARDs or csDMARDs and matched biologic-treated patients.

There were no significant differences in CXCL10, MMP3, S100A8, ACP5, and CCL2 levels between patients with PsC who were not treated with bDMARDs or csDMARDs and matched biologic-treated patients with PsC (Figure 2; Supplementary Table S2, available with the online version of this article). In biologic-treated patients with PsC, protein levels were significantly different between male and female individuals for MMP3 (P < 0.001), and ACP5 (P < 0.001), but nonsignificant for CXCL10 and CCL2. In patients with PsC not treated with bDMARDs or csDMARDs, protein levels were significantly different between male and female individuals for MMP3 (P < 0.001), but nonsignificant for CXCL10, ACP5, and CCL2 (Supplementary Table S4).

Figure 2.
Figure 2.

Biomarker protein levels in patients with PsC treated with biologics and no bDMARDs or csDMARDs. (A) CXCL10, (B) MMP3, (C) ACP5, and (D) CCL2. ACP5: acid phosphatase 5; bDMARD: biologic disease-modifying antirheumatic drug; CCL2: C-C motif chemokine ligand 2; csDMARD: conventional synthetic disease-modifying antirheumatic drug; CXCL10: C-X-C motif chemokine 10; MMP3: matrix metalloproteinase 3; ns: not significant; PsC: cutaneous psoriasis; S100A8: S100 calcium binding protein A8.

In comparing patients with PsC not treated with bDMARDs or csDMARDs and matched biologic-treated patients, there was no significant effect of sex on the differences in protein levels for CXCL10 (F[1,53] = 2.02; P = 0.16) and CCL2 (F[1,53] = 0.07; P = 0.80). However, there was a significant effect of sex for MMP3 (F[1,53] = 46.47; P < 0.001) and ACP5 (F[1,53] = 8.19; P < 0.01). When corrected for this effect, differences in MMP3 (F[1,53] = 3.26; P = 0.08) and ACP5 (F[1,53] = 2.11; P = 0.15) between patients not treated with bDMARDs or csDMARDs and biologic-treated patients with PsC were not significant.

Prediction of treatment response. We considered a good predictive performance to be when AUC ≥ 0.80 (Figure 3; Figure 4). We present the results of the prediction using only baseline protein levels, as the addition of age and sex did not improve the performance of the models.

Figure 3.
Figure 3.

Performance of logistic regression models predicting response to biologics in patients with PsA (achieving DAPSA < 14, ≥ 75% reduction in PASI, or ≥ 50% reduction in actively inflamed joints). Models based on the natural logarithm of baseline serum levels of (A) CXCL10, (B) MMP3, (C) S100A8, (D) ACP5 and (E) CCL2. Active joints: tender and/or swollen joints; ACP5: acid phosphatase 5; AUC: area under the curve; CCL2: C-C motif chemokine ligand 2; CXCL10: C-X-C motif chemokine 10; DAPSA: Disease Activity Index for Psoriatic Arthritis; MMP3: matrix metalloproteinase 3; PASI: Psoriasis Area Severity Index; S100A8: S100 calcium binding protein A8.

Figure 4.
Figure 4.

Performance of logistic regression models predicting response in patients with PsA not treated with bDMARDs or csDMARDs (achieving DAPSA < 14, ≥ 75% reduction in PASI or ≥ 50% reduction in actively inflamed joints). Models based on the natural logarithm of baseline serum levels of (A) CXCL10, (B) MMP3, (C) S100A8, (D) ACP5, and (E) CCL2. Active joints: tender and/or swollen joints; ACP5: acid phosphatase 5; AUC: area under the curve; bDMARD: biologic disease-modifying antirheumatic drug; CCL2: C-C motif chemokine ligand 2; csDMARD: conventional synthetic disease-modifying antirheumatic drug; CXCL10: C-X-C motif chemokine 10; DAPSA: Disease Activity Index for Psoriatic Arthritis; MMP3: matrix metalloproteinase 3; PASI: Psoriasis Area Severity Index; S100A8: S100 calcium binding protein A8.

According to the DAPSA response criteria for achieving LDA or remission (DAPSA < 14), high baseline levels of ACP5 (AUC = 0.80) were predictive of response in patients with PsA treated with biologics, and low baseline levels of MMP3 (AUC = 0.80) predicted response in patients with PsA not treated with bDMARDs or csDMARDs. High baseline levels of CXCL10 (> 22.12 pg/mL; AUC = 0.87) and S100A8 (> 27.96 pg/mL; AUC = 0.88) were predictive of a PASI response (PASI75) in patients with PsA treated with biologics, whereas high baseline levels of MMP3 (AUC = 0.93) and ACP5 (AUC = 0.93) and low baseline levels of S100A8 (AUC = 0.86) were predictive of a PASI response in patients with PsA not treated with bDMARDs or csDMARDs. High baseline levels of MMP3 (AUC = 0.82) and S100A8 (> 20.32 pg/mL; AUC = 0.84), and low baseline levels of ACP5 (AUC = 0.90) were predictive of a response in actively inflamed joint count (50% reduction in actively inflamed joint count) in patients with PsA treated with biologics.

DISCUSSION

In this study, we aimed to determine the alterations in serum levels of CXCL10, MMP3, S100A8, ACP5, and CCL2 due to biologic therapy and MTX treatment, and whether the baseline levels of these proteins predict response to treatment. We found that TNFi reduced serum levels of all 5 proteins, IL-17i increased ACP5 and CCL2, and MTX reduced MMP3. Changes in MMP3 and S100A8 levels were significantly different between patients with PsA not receiving bDMARDs or csDMARDs and matched biologic-treated patients with PsA (P < 0.05). There were no significant differences between patients with PsC who were taking biologics and not taking bDMARDs or csDMARDs. Baseline levels of CXCL10, MMP3, S100A8, and ACP5 had good predictive value (AUC > 0.80) for response to biologics in PsA.

CXCL10 is a chemokine that attracts monocytes and T lymphocytes to areas of inflammation.20 We found that CXCL10 levels decrease after treatment with TNFi in patients with PsA. TNF-α increases the production of CXCL10 from THP-1 monocytes through the activation of MEK/ERK and NF-kB.21 CXCL10 likely decreases after TNFi treatment as the inhibition of TNF-α results in reduced activation of MEK/ERK and NF-kB, thus reducing production of CXCL10 and reducing inflammation. CXCL10 is higher in patients with psoriasis who will develop PsA (converters) compared to patients with PsC who will not develop PsA (nonconverters).9 Levels of CXCL10 decrease in converters soon after PsA development but remain constant at low levels in nonconverters.22 This is supported by our results as we showed a decrease in CXCL10 in patients with PsA who had been receiving TNFi for 3 to 6 months, but no change was seen in the cohort that had not been treated with biologics or MTX for the same duration. However, it is important to note that this significant decrease in CXCL10 levels in the TNFi-treated PsA cohort was not observed in the combined biologic-treated cohort (TNFi and IL-17i) when matched to the untreated PsA cohort. We did not observe any differences in CXCL10 levels between the treated and untreated PsC cohort. CXCL10 levels > 22.12 pg/mL predicted a reduction in PASI score in the biologic-treated cohort. To our knowledge, our study is the first to show the predictive potential of PsA treatment response using CXCL10 baseline levels.

MMP3 is a proteinase that breaks down the extracellular matrix. MMP3 serum levels are higher in patients with PsA compared to healthy controls and patients with PsC, causing extracellular matrix degradation and the destruction of the joint.10 MMP3 expression is induced by TNF-α and plays a key role in TNF-α–mediated collagen degradation in the skin.23 We observed a reduction in MMP3 levels after TNFi treatment, likely due to reduced TNF-α levels, which in turn reduce the induction of MMP3 expression. Our findings are supported by other studies that also show a reduction in MMP3 after treatment with TNFi.24,25 MTX exhibits its antiinflammatory effects in PsA by inhibiting neutrophil function through the release of adenosine.26 MTX has also been shown to reduce angiogenesis by reducing expression of vascular endothelial growth factor (VEGF).27 We demonstrated that MTX reduced the levels of MMP3 but not the other proteins included in this study. It is likely that MTX exerts its antiangiogenic effects through altering MMP3 promotor methylation, as MMP3 is associated with angiogenesis and VEGF expression.28,29 Our results demonstrate that the reduction in MMP3 was significantly larger in the biologic-treated cohort compared to the untreated cohort. Given the role of TNF-α in the production of MMP3,23 it is likely that without treatment, MMP3 levels would not be altered.

S100A8 is a monomer of the heterodimer of S100A8/S100A9 (calprotectin), which is a calcium-binding protein that contributes to inflammation by acting as a chemoattractant, recruiting neutrophils and monocytes. S100A8 protein is higher in patients with PsA and psoriasis compared to healthy controls,30 and likely also plays a role in inflammation by increasing the expression of the C3 complement protein in the complement cascade, leading to greater inflammation.31 S100A8 was significantly reduced in patients with PsA after treatment with TNFi, and this observation is supported by previous reports showing S100A8 expression is significantly higher in wild-type mice compared to TNF knockout mice.32 We demonstrated that S100A8 was significantly reduced in both the biologic-treated cohort and the untreated cohort, with the reduction in the treated cohort being significantly larger than the reduction in the untreated cohort. The decrease in S100A8 seen in the untreated cohort could be explained by NSAIDs that have general antiinflammatory effects, with the larger decrease in S100A8 in the biologic cohort due to the role of TNF-α in the production of S100A8.32

ACP5 is a histochemical marker for osteoclasts and likely plays a role in bone erosion in PsA as it is upregulated in the synovial membrane of patients with PsA.33,34 TNF-α increases the number of osteoclast precursors (OCPs), and given that ACP5 is a marker of osteoclasts, TNF-α likely increases the amount of ACP5.35 TNFi have also been shown to decrease OCPs, likely contributing to a decrease in ACP5, which supports our findings that ACP5 decreases after treatment with TNFi in patients with PsA.36 IL-17 induces osteoclastogenesis, indicating that IL-17 can increase the amount of ACP5 expression.37,38 Our results indicate that IL-17i significantly increases the amount of ACP5. Other studies show in vitro IL-17i antibodies reduce the IL-17–induced production of TRAP-positive osteoclasts, indicating that ACP5 should decrease with IL-17i treatment, as TRAP is the product of the ACP5 gene.37

CCL2 is a monocyte chemoattractant protein.13 It is significantly increased in patients with PsA compared to healthy controls and likely causes T lymphocyte migration into the synovium, causing joint damage.14 TNF-α induces the expression of CCL2 through the mitogen-activated protein kinase pathway.39 We found that CCL2 significantly decreased after treatment with TNFi, which decreases CCL2 expression in monocytes through epigenetic modification.40 We also found that IL-17i increased CCL2, which contrasts previous evidence that shows that IL-17 increases the expression of CCL2.41,42

Correlations between protein levels and disease activity measures were generally low, but significant in some cases. Logistic regression classification models based on baseline protein levels that achieved a good predictive performance (AUC ≥ 0.80) of response to treatment included CXCL10, MMP3, S100A8, and ACP5. Of note, CXCL10 levels > 22.12 pg/mL predicted a reduction in PASI score in the biologic-treated PsA cohort, and to the best of our knowledge, this study is the first to show the predictive potential of CXCL10 baseline levels on PsA treatment response.

MMP3 may be a promising predictor of response to treatment and reduction in disease activity, as we showed that it can predict a reduction in DAPSA score, PASI score, and actively inflamed joint count. MMP3 levels < 8087.56 pg/mL predict a DAPSA response in patients with PsA who are not receiving treatment, whereas MMP3 levels > 7989.08 pg/mL predict an active joint response in patients with PsA treated with biologics. Although further validation in independent cohorts is still warranted, these results indicate that MMP3 can be used as a predictive biomarker to determine if biologic treatment would be useful and necessary for a given patient. Other studies have also shown that MMP3 can be used as a predictive biomarker for response to TNFi treatment according to PASI score and actively inflamed joint count.43 MMP3 likely acts as a good predictive biomarker for response to TNFi because MMP3’s role in the pathogenesis of both skin and joint degradation is modulated by TNF-α.23

S100A8 can also predict a reduction in PsA disease activity. S100A8 levels > 27.96 pg/mL can predict a PASI response and levels > 20.32 pg/mL can predict an active joint response in patients with PsA treated with biologics. S100A8 levels < 65.87 pg/mL predicted PASI response in those not treated with biologics or MTX. S100A8 has been shown to be a predictive biomarker for response to TNFi in synovial fluid.11 S100A8 is upregulated in psoriatic skin lesions31 and along with its ability to predict a reduction in PASI score, it likely plays a role in the development of psoriatic lesions.

ACP5 shows predictive potential for DAPSA, PASI, and actively inflamed joint response in patients with PsA. In the biologic-treated cohort, ACP5 levels > 6222.81 pg/mL predicted DAPSA response, whereas ACP5 levels < 5033.58 pg/mL predicted an active joint response. In the untreated cohort, ACP5 levels > 8003.35 pg/mL predicted PASI response. ACP5 is differentially expressed in TNFi responders and nonresponders12 and given the relationship between ACP5 and osteoclastogenesis,34 ACP5 likely plays a role in contributing to active joint numbers affecting both the actively inflamed joint count and DAPSA score.

In our study, we performed a direct comparison of 5 protein biomarker levels between patients with PsA and PsC receiving different treatment options; however, the sample size for some of these groups, including those treated with MTX and IL-17i, was limited. For this reason, in the biologic-treated cohort, we combined both TNFi- and IL-17i–treated patients. Since we saw different trends in the proteins, such as a reduction in all proteins after TNFi as well as an increase in ACP5 and CCL2 after IL-17i, it is possible that combining both treatments into a single biologic-treated cohort skewed the results. However, we performed the analysis for each biologic agent separately as well, even if the number of included IL-17i–treated patients was limited and the skewing effect may not be as drastic. The biological reason for the differences observed between the 2 biologic agents is not clear.

One important limitation of this study is that we could not include matching pretreatment and posttreatment samples for the PsC cohort, so the untreated patients with PsC were matched based on age, sex, and disease duration to patients with PsC who were treated with biologics. We did not find any significant differences between the groups, although further longitudinal studies may be needed to observe significant differences between cohorts.

Additionally, cohorts were not matched on disease activity. Patients not treated with biologics or MTX tend to have lower disease activity, which may be related to lower biomarker levels at baseline. However, we expect bias to be relatively small, as the untreated cohort was matched to biologic-treated patients based on age, sex, and disease duration.

In conclusion, we determined the effect of biologic treatment (TNFi and IL-17i) and MTX on serum CXCL10, MMP3, S100A8, ACP5, and CCL2 in patients with PsA. Our results highlight that MMP3, S100A8, ACP5 and CXCL10 have potential use as serum biomarkers to predict response to treatment in patients with PsA.

Footnotes

  • This study was partially funded by a grant from the Canadian Institute of Health Research (grant no. 23-5177). The Psoriatic Disease Research Program is funded by a grant from the Krembil Foundation.

  • The authors declare no conflicts of interest relevant to this article.

  • Accepted for publication August 14, 2024.

This is an Open Access article, which permits use, distribution, and reproduction, without modification, provided the original article is correctly cited and is not used for commercial purposes.

REFERENCES

  1. 1.↵
    1. Griffiths CE,
    2. Barker JN.
    Pathogenesis and clinical features of psoriasis. Lancet 2007;370:263-71.
    OpenUrlCrossRefPubMed
  2. 2.↵
    1. Grän F,
    2. Kerstan A,
    3. Serfling E,
    4. Goebeler M,
    5. Muhammad K.
    Current developments in the immunology of psoriasis. Yale J Biol Med 2020;93:97-110.
    OpenUrlPubMed
  3. 3.↵
    1. Rauschenberger T,
    2. Schmitt V,
    3. Azeem M, et al.
    T cells control chemokine secretion by keratinocytes. Front Immunol 2019;10:1917.
    OpenUrlCrossRefPubMed
  4. 4.↵
    1. Alinaghi F,
    2. Calov M,
    3. Kristensen LE, et al.
    Prevalence of psoriatic arthritis in patients with psoriasis: a systematic review and meta-analysis of observational and clinical studies. J Am Acad Dermatol 2019;80:251-265.e19.
    OpenUrlCrossRefPubMed
  5. 5.↵
    1. Ritchlin CT,
    2. Colbert RA,
    3. Gladman DD.
    Psoriatic arthritis. N Engl J Med 2017;376:957-70.
    OpenUrlCrossRefPubMed
  6. 6.↵
    1. Veale D,
    2. Yanni G,
    3. Rogers S,
    4. Barnes L,
    5. Bresnihan B,
    6. Fitzgerald O.
    Reduced synovial membrane macrophage numbers, ELAM-1 expression, and lining layer hyperplasia in psoriatic arthritis as compared with rheumatoid arthritis. Arthritis Rheum 1993; 36:893-900.
    OpenUrlCrossRefPubMed
  7. 7.↵
    1. Veale DJ,
    2. Fearon U.
    The pathogenesis of psoriatic arthritis. Lancet 2018;391:2273-84.
    OpenUrlCrossRefPubMed
  8. 8.↵
    1. Singh JA,
    2. Guyatt G,
    3. Ogdie A, et al.
    Special article: 2018 American College of Rheumatology/National Psoriasis Foundation guideline for the treatment of psoriatic arthritis. Arthritis Rheumatol 2019;71:5-32.
    OpenUrlCrossRefPubMed
  9. 9.↵
    1. Abji F,
    2. Pollock RA,
    3. Liang K,
    4. Chandran V,
    5. Gladman DD.
    Brief report: CXCL10 is a possible biomarker for the development of psoriatic arthritis among patients with psoriasis. Arthritis Rheumatol 2016;68:2911-16.
    OpenUrlCrossRefPubMed
  10. 10.↵
    1. Jadon DR,
    2. Nightingale AL,
    3. McHugh NJ,
    4. Lindsay MA,
    5. Korendowych E,
    6. Sengupta R.
    Serum soluble bone turnover biomarkers in psoriatic arthritis and psoriatic spondyloarthropathy. J Rheumatol 2015;42:21-30.
    OpenUrlAbstract/FREE Full Text
  11. 11.↵
    1. Ademowo OS,
    2. Hernandez B,
    3. Collins E, et al.
    Discovery and confirmation of a protein biomarker panel with potential to predict response to biological therapy in psoriatic arthritis. Ann Rheum Dis 2016;75:234-41.
    OpenUrlAbstract/FREE Full Text
  12. 12.↵
    1. Rahmati S,
    2. O’Rielly DD,
    3. Li Q, et al.
    Rho-GTPase pathways may differentiate treatment response to TNF-alpha and IL-17A inhibitors in psoriatic arthritis. Sci Rep 2020;10:21703.
    OpenUrlCrossRefPubMed
  13. 13.↵
    1. Bose S,
    2. Cho J.
    Role of chemokine CCL2 and its receptor CCR2 in neurodegenerative diseases. Arch Pharm Res 2013;36:1039-50.
    OpenUrlCrossRefPubMed
  14. 14.↵
    1. Ross EL,
    2. D’Cruz D,
    3. Morrow WJ.
    Localized monocyte chemotactic protein-1 production correlates with T cell infiltration of synovium in patients with psoriatic arthritis. J Rheumatol 2000;27:2432-43.
    OpenUrlPubMed
  15. 15.↵
    1. Taylor W,
    2. Gladman D,
    3. Helliwell P, et al.
    Classification criteria for psoriatic arthritis: development of new criteria from a large international study. Arthritis Rheum 2006;54:2665-73.
    OpenUrlCrossRefPubMed
  16. 16.↵
    1. Fredriksson T,
    2. Pettersson U.
    Severe psoriasis – oral therapy with a new retinoid. Dermatologica 1978;157:238-44.
    OpenUrlCrossRefPubMed
  17. 17.↵
    1. Gladman DD,
    2. Chandran V.
    Observational cohort studies: lessons learnt from the University of Toronto Psoriatic Arthritis Program. Rheumatology 2011;50:25-31.
    OpenUrlCrossRefPubMed
  18. 18.↵
    1. Schoels MM,
    2. Aletaha D,
    3. Alasti F,
    4. Smolen JS.
    Disease activity in psoriatic arthritis (PsA): defining remissions and treatment success using the DAPSA score. Ann Rheum Dis 2016;75:811-8.
    OpenUrlAbstract/FREE Full Text
  19. 19.↵
    1. Duarte-García A,
    2. Leung YY,
    3. Coates LC, et al.
    Endorsement of the 66/68 joint count for the measurement of musculoskeletal disease activity: OMERACT 2018 psoriatic arthritis workshop report. J Rheumatol 2019;46:996-1005.
    OpenUrlAbstract/FREE Full Text
  20. 20.↵
    1. Taub DD,
    2. Lloyd AR,
    3. Conlon K, et al.
    Recombinant human interferon-inducible protein 10 is a chemoattractant for human monocytes and T lymphocytes and promotes T cell adhesion to endothelial cells. J Exp Med 1993;177:1809-14.
    OpenUrlAbstract/FREE Full Text
  21. 21.↵
    1. Qi X-F,
    2. Kim D-H,
    3. Yoon Y-S, et al.
    Essential involvement of cross-talk between IFN-γ and TNF-α in CXCL10 production in human THP-1 monocytes. J Cell Physiol 2009;220:690-97.
    OpenUrlCrossRefPubMed
  22. 22.↵
    1. Abji F,
    2. Lee KA,
    3. Pollock RA, et al.
    Declining levels of serum chemokine (C-X-C motif) ligand 10 over time are associated with new onset of psoriatic arthritis in patients with psoriasis: a new biomarker? Br J Dermatol 2020;183:920-27.
    OpenUrlCrossRefPubMed
  23. 23.↵
    1. Mirastschijski U,
    2. Lupše B,
    3. Maedler K, et al.
    Matrix metalloproteinase-3 is key effector of TNF-α-induced collagen degradation in skin. Int J Mol Sci 2019;20:5234.
    OpenUrlCrossRefPubMed
  24. 24.↵
    1. Ramonda R,
    2. Modesti V,
    3. Ortolan A, et al.
    Serological markers in psoriatic arthritis: promising tools. Exp Biol Med 2013;238:1431-36.
    OpenUrlCrossRefPubMed
  25. 25.↵
    1. van Kuijk AWR,
    2. DeGroot J,
    3. Koeman RC, et al.
    Soluble biomarkers of cartilage and bone metabolism in early proof of concept trials in psoriatic arthritis: effects of adalimumab versus placebo. PLoS ONE 2010;5:e12556.
    OpenUrlCrossRefPubMed
  26. 26.↵
    1. Cronstein BN,
    2. Sitkovsky M.
    Adenosine and adenosine receptors in the pathogenesis and treatment of rheumatic diseases. Nat Rev Rheumatol 2017;13:41-51.
    OpenUrlCrossRefPubMed
  27. 27.↵
    1. Shaker OG,
    2. Khairallah M,
    3. Rasheed HM, et al.
    Antiangiogenic effect of methotrexate and PUVA on psoriasis. Cell Biochem Biophys 2013;67:735-42.
    OpenUrlCrossRefPubMed
  28. 28.↵
    1. Celik ZB,
    2. Cankara FN,
    3. Gunaydin C.
    Alterations in the matrix metalloproteinase-3 promoter methylation after common chemotherapeutics: in vitro study of paclitaxel, cisplatin and methotrexate in the MCF-7 and SH-SY5Y cell lines. Mol Biol Rep 2020;47:8987-95.
    OpenUrlCrossRefPubMed
  29. 29.↵
    1. Saleh A,
    2. Stathopoulou MG,
    3. Dadé S, et al.
    Angiogenesis related genes NOS3, CD14, MMP3 and IL4R are associated to VEGF gene expression and circulating levels in healthy adults. BMC Med Genet 2015;16:90.
    OpenUrlCrossRefPubMed
  30. 30.↵
    1. Hansson C,
    2. Eriksson C,
    3. Alenius GM.
    S-calprotectin (S100A8/S100A9): a potential marker of inflammation in patients with psoriatic arthritis. J Immunol Res 2014;2014:696415.
    OpenUrlPubMed
  31. 31.↵
    1. Schonthaler HB,
    2. Guinea-Viniegra J,
    3. Wculek SK, et al.
    S100A8-S100A9 protein complex mediates psoriasis by regulating the expression of complement Factor C3. Immunity 2013; 39:1171-81.
    OpenUrlCrossRefPubMed
  32. 32.↵
    1. Sade-Feldman M,
    2. Kanterman J,
    3. Ish-Shalom E,
    4. Elnekave M,
    5. Horwitz E,
    6. Baniyash M.
    Tumor necrosis factor-α blocks differentiation and enhances suppressive activity of immature myeloid cells during chronic inflammation. Immunity 2013;38:541-54.
    OpenUrlCrossRefPubMed
  33. 33.↵
    1. Dolcino M,
    2. Ottria A,
    3. Barbieri A, et al.
    Gene expression profiling in peripheral blood cells and synovial membranes of patients with psoriatic arthritis. PLoS ONE 2015;10:e0128262.
    OpenUrlCrossRefPubMed
  34. 34.↵
    1. Burstone MS.
    Histochemical demonstration of acid phosphatase activity in osteoclasts. J Histochem Cytochem 1959;7:39-41.
    OpenUrlCrossRefPubMed
  35. 35.↵
    1. Yao Z,
    2. Li P,
    3. Zhang Q, et al.
    Tumor necrosis factor-alpha increases circulating osteoclast precursor numbers by promoting their proliferation and differentiation in the bone marrow through up-regulation of c-Fms expression. J Biol Chem 2006;281:11846-55.
    OpenUrlAbstract/FREE Full Text
  36. 36.↵
    1. Perpétuo IP,
    2. Caetano-Lopes J,
    3. Rodrigues AM, et al.
    Effect of tumor necrosis factor inhibitor therapy on osteoclasts precursors in rheumatoid arthritis. BioMed Res Int 2017;2017:2690402.
    OpenUrlPubMed
  37. 37.↵
    1. Kotake S,
    2. Udagawa N,
    3. Takahashi N, et al.
    IL-17 in synovial fluids from patients with rheumatoid arthritis is a potent stimulator of osteoclastogenesis. J Clin Invest 1999;103:1345-52.
    OpenUrlCrossRefPubMed
  38. 38.↵
    1. Kim K-W,
    2. Kim H-R,
    3. Kim B-M,
    4. Cho M-L,
    5. Lee S-H.
    Th17 cytokines regulate osteoclastogenesis in rheumatoid arthritis. Am J Pathol 2015;185:3011-24.
    OpenUrlCrossRefPubMed
  39. 39.↵
    1. Ho AWY,
    2. Wong CK,
    3. Lam CWK.
    Tumor necrosis factor-alpha up-regulates the expression of CCL2 and adhesion molecules of human proximal tubular epithelial cells through MAPK signaling pathways. Immunobiology 2008;213:533-44.
    OpenUrlCrossRefPubMed
  40. 40.↵
    1. Lin YC,
    2. Lin YC,
    3. Huang MY, et al.
    Tumor necrosis factor-alpha inhibitors suppress CCL2 chemokine in monocytes via epigenetic modification. Mol Immunol 2017;83:82-91.
    OpenUrlCrossRefPubMed
  41. 41.↵
    1. Shahrara S,
    2. Pickens SR,
    3. Dorfleutner A,
    4. Pope RM.
    IL-17 induces monocyte migration in rheumatoid arthritis. J Immunol 2009;182:3884-91.
    OpenUrlAbstract/FREE Full Text
  42. 42.↵
    1. Shahrara S,
    2. Pickens SR,
    3. Mandelin AM II, et al.
    IL-17-mediated monocyte migration occurs partially through CC chemokine ligand 2/monocyte chemoattractant protein-1 induction. J Immunol 2010;184:4479-87.
    OpenUrlAbstract/FREE Full Text
  43. 43.↵
    1. Chandran V,
    2. Shen H,
    3. Pollock RA, et al.
    Soluble biomarkers associated with response to treatment with tumor necrosis factor inhibitors in psoriatic arthritis. J Rheumatol 2013;40:866-71.
    OpenUrlAbstract/FREE Full Text

DATA AVAILABILITY

The data underlying this article will be shared on reasonable request to the corresponding author.

Back to top

In this issue

Print
Download PDF
Article Alerts
Email Article
Citation Tools

 Request Permissions

Bookmark this article

Jump to section

Keywords

BIOLOGICS
BIOMARKERS
PSORIASIS
PSORIATIC ARTHRITIS

Keywords