This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/authorsrights

Author's personal copy

Joint Bone Spine 80 (2013) 165–170

Available online at

www.sciencedirect.com

Original article

Blockade of intra-articular TNF in peripheral spondyloarthritis: Its relevance toclinical scores, quantitative imaging and synovial fluid and synovial tissuebiomarkers

Ugo Fioccoa,∗, Paolo Sfrisoa, Francesca Olivieroa, Francesca Lunardib, Fiorella Calabreseb,Elena Scagliori a,b, Luisella Cozzia, Antonio Di Maggiob, Roberto Nardacchionec, Béatrice Molenaa,Mara Felicetti a, Katia Gazzolaa, Roberto Stramareb, Léopoldo Rubaltelli b, Benedetta Accordid,Luisa Costae, Pascale Roux-Lombarde,f, Leonardo Punzia, Jean-Michel Dayerg

a Department of Medicine, University of Padova, Via Giustiniani, 2, 35128 Padova, Italyb Department of Medical-Diagnostic Science and Special Therapies, University of Padova, Padova, Italyc Department of Orthopedics, Leonardo Foundation, Abano Terme General Hospital, Abano Terme (PD), Italyd Oncohematology Laboratory, Department of Pediatrics, University of Padova, Padova, Italye Rheumatology Research Unit, Department of Clinical and Experimental Medicine, University Federico II, Naples, Italyf Immunology and Allergy Division, Geneva University Hospitals and University of Geneva, 4, rue Gabrielle Perret-Gentil, CH-1211 Geneva, Switzerlandg Faculty of Medicine, CMU, 1, rue Michel-Servet, CH-1211 Geneva, Switzerland

a r t i c l e i n f o

Article history:Accepted 27 June 2012Available online 3 August 2012

Keywords:Peripheral spondyloarthritisIntra-articular TNF blockadeContrast-enhanced magnetic resonanceimagingUltrasonographySynovial fluid biomarkerSynovial tissue biomarkers

a b s t r a c t

Objectives: This open-label study is based on a translational approach with the aim of detecting changes inthe clinical condition as well as in imaging and synovial biological markers in both synovial fluid (SF) andsynovial tissue (ST) in peripheral spondyloarthritis (SpA) patients following intra-articular (IA) blockadeof TNF-� by serial etanercept injections.Methods: Twenty-seven SpA patients with resistant knee synovitis underwent four biweekly IA injectionsof etanercept (E) (12.5 mg). The primary outcome of Thompson’s Knee Index (THOMP), and secondaryoutcomes of Knee Joint Articular Index (KJAI), C-reactive protein (CRP), HAQ-Disability Index (HAQ-DI),maximal synovial thickness (MST) according to ultrasonography (US) and contrast-enhanced magneticresonance (C+MR) imaging, ST-CD45+ mononuclear cells (MNC) and ST-CD31+ vessels, IL-1�, IL-1Ra andIL-6 levels in SF were assessed at baseline and at the end of the study.Results: At the study end, clinical and imaging outcomes as well as ST and SF biological markers weresignificantly reduced compared to baseline. There were significant correlations between clinical, imagingand biological markers (CRP with either THOMP, or KJAI, or HAQ-DI or SF-IL-1Ra; US-MST with KJAI, ST-CD45+ with either THOMP, or KJAI, or ST-CD31+, or SF-IL-1�; SF-IL-6 with either THOMP, or KJAI, orSF-IL-1�, or IL-1Ra).Conclusions: The proof of concept study revealed early improvement either in local and systemic clinicalscores, in synovial thickness measures by C+MR and US, or expression of synovial biological markers.CD45+, CD31+ in ST and IL-6 and IL-1� in SF may be considered potential biomarkers of the peripheralSpA response to IA TNF-� blocking.

© 2012 Société franc aise de rhumatologie. Published by Elsevier Masson SAS. All rights reserved.

Abbreviations: ACR-SJC, ACR 68 swollen joint number; ACR-TJC, ACR 66 tender joint number; ASAS, Assessment of Spondylo Arthritis International Society; C+MR-MST,contrast-enhanced magnetic resonance maximal synovial thickness; CD, cluster differentiation; CK, cytokine; CRP, C-reactive protein; DMARD, disease-modifying anti-rheumatic drugs; E, etanercept; HAQ-DI, Health Assessment Questionnaire-Disability Index; IA, intra-articular; IL, interleukin; IL-1Ra, IL-1 receptor antagonist; KJAI, KneeJoint Articular Index score; KJS, knee joint synovitis; LCA, leucocyte-common antigen; MNC, mononuclear cell; PECAM, pan endothelial cell adhesion molecule; P, placebo;SpA, spondyloarthritis; SF, synovial fluid; ST, synovial tissue; THOMP, Thompson’s Knee Index score; TNF, tumor necrosis factor; US-MST, ultrasonography maximal synovialthickness; VEFG, vascular endothelial growth factor.

∗ Corresponding author. Tel.: +39 04 98 21 21 90; fax: +39 04 98 21 21 91.E-mail address: ugo.fiocco@unipd.it (U. Fiocco).

1297-319X/$ – see front matter © 2012 Société franc aise de rhumatologie. Published by Elsevier Masson SAS. All rights reserved.doi:10.1016/j.jbspin.2012.06.016

Author's personal copy

166 U. Fiocco et al. / Joint Bone Spine 80 (2013) 165–170

1. Introduction

Synovial immunopathology has been well characterized inspondyloarthropathies (SpA) [1], by delineation of disease-specificfeatures, of the similarity between different SpA subtypes,including psoriatic arthritis (PsA), as well by highlighting thehistopathological characteristics induced by effective treatment inSpA synovitis, such as down-regulation of leucocyte infiltration andhypervascularity [2]. Nevertheless, only limited data on synovialbiomarkers of peripheral joint involvement in SpA are available [3].

Peripheral arthritis in SpA is often resistant to intra-articular (IA)drugs or surgical therapies. Consequently, therapies that specifi-cally target affected joints would constitute a welcome alternative[4,5]. So far, a few small randomized studies on serial TNF-�-blocking injections in individual joints have been described in theliterature [6–8]. Likewise, very few studies providing “proof ofconcept” have addressed the issue of the local assessment of inflam-mation at single joint level [9] to detect early changes resulting fromthe blockade of intra-articular (IA) TNF-�.

A randomized, blind-controlled, single-centre study on seriallow dosage injections of IA etanercept (E) in resistant knee jointsynovitis (KJS) has shown early slow down of KJS inflammation,along with the good patient compliance and safety profile. Eitherthe superiority of low-dose of IA etanercept to placebo, or the drug’sshort term effect after the cross-over, has been reflected in a higherthan expected drop-out rate of IA-P treated knees and a lowerpower of the study (Fig. 1) [10,11]. The open-label extension ofthe IA blind study was designed to analyse the response to seriallow dosage injections of IA E in resistant SpA KJS (Fig. 1) [10,11]. Atranslational approach was used, based on specific tools such as bio-logical synovial markers, composite clinical scores and ultrasound(US) and contrast-enhanced magnetic resonance (C+MR) imaging,and quantitative measures of the synovitis response to IA TNF-�blockade.

Synovial hypervascularity along with leucocyte ST infiltration isa distinctive feature of SpA synovitis. However, only limited dataare available as to their association with either clinical disease

Fig. 1. Flow-chart of patients in the blind-controlled study on serial IA etanercept(E) injections and in the open-label extension study.

activity or with the response to systemic TNF-blocking therapy inperipheral SpA [2,3]. Among the immunocytochemical markers, thesynovial tissue expression of CD45+, the leucocyte-common anti-gen (LCA) by mononuclear cells (MNC), along with that of the CD31+platelet endothelial cell adhesion molecule (PECAM)-1 by maturesynovial vessels, were assessed.

The cytokine (CK) patterns observed so far in synovial fluid (SF)of patients with SpA were disparate [12,13]. Further, the poten-tial role of synovial fluid inflammatory CK as biomarkers of theresponse to anti-TNF therapy has recently been suggested [13,14].

The proof of concept study showed early improvement either inlocal and systemic clinical scores, synovial thickness measures byC+MR- and US, or in the expression of CD45+-, CD31+-ST and IL-6-SF synovial markers that are significantly interactive and correlatewith clinical measures.

2. Methods

2.1. Study design

The study represents the prospective open-label extension ofa randomized, placebo-controlled, single-centre study protocol onserial IA etanercept (E) injections that was approved by the localethics committee (Etanercept/TNR-001:n.878P; ClinicalTrials.govIdentifier: NCT00678782).

All patients signed informed consent statements beforeentering the study. Joints were assessed before each IA injectionand 2 weeks after the final one. The blind study protocol consistedof the administration either of IA-E (12.5 mg/0.5 mL) or IA placebo(P) (saline 0.5 mL), injected once every 2 weeks to single kneejoints affected by active synovitis (KJS) (characterized by pain,tenderness, and effusion) for an 8-week period. The patients werecrossed-over after 2 weeks. Patients who suffered of with KJS flaresthroughout the blind study, were invited to enter the open-labelextension, during which four IA-E injection (12.5 mg/0.5 mL) wereadministered after synovial fluid aspiration, once every 2 weeks,for an 8-week period (Fig. 1) [10].

2.2. Main inclusion and exclusion criteria of the blind study

Patients eligible to participate in the trial were:

• at least 18 years of age;• patients who met the generally accepted criteria for PsA and SpA

or the 1987 American Rheumatism Association (ARA) revised cri-teria for RA;

• patients who suffered from persistent, active gonarthritis(characterized by pain, tenderness, and effusion) which hadproved resistant to repeated IA corticosteroid therapy, unrespon-sive to surgical treatment and/or to a minimum of a 6 monthssecond-line drug therapy with DMARD-, and/or anti-TNF� bio-logic agent monotherapy and/or systemic combination therapy;

• patients who received a stable, appropriate dose of DMARD, etan-ercept and less or equal to 10 mg prednisone as monotherapy or inassociation during the 4 weeks before the screening with/withoutnon-steroidal anti-inflammatory drugs (NSAID) and were conti-nued at stable doses throughout the whole study.

Were ineligible for study inclusion:

• pregnant or potentially pregnant women not using a medicallyacceptable form of contraceptive and breastfeeding mothers;

• patients who present relevant comorbidities includingchronic or active infections, hematologic/hemorrhagic disease,

Author's personal copy

U. Fiocco et al. / Joint Bone Spine 80 (2013) 165–170 167

Table 1Clinical and demographic features of a cohort of peripheral SpA patients treatedduring the study.

CharacteristicsKnee joint synovitis, n 27Bilateral involvement, n (%) 2 (7.40)Patients age (years), mean ± SD 41.8 ± 14.9Female, n (%) 16 (59.26)Disease duration (years), mean ± SD 6.3 ± 4.6CRP, mg/dL, mean ± SD 1.2 ± 1.7HAQ-DI, score, mean ± SD 0.65 ± 0.54

Systemic treatment at study entryDMARD, n (%) 24 (88.8)Etanercept, n (%) 4 (14.8)Prednisolonea, n (%) 12 (44.4)

Intra-articular (IA) treatmentPrevious IA steroid injections: n, mean ± SD 2.3 ± 1.5IA etanercept injection: n, mean ± SD 4 ± 0.2

CRP: C-reactive protein; DMARD: disease-modifying anti-rheumatic drugs; HAQ:health assessment questionnaire; IA: intra-articular.

a Daily prednisolone dose less or equal to 10 mg.

malignancies, decompensatio cordis (New York Heart Associationclassification III and IV), multiple sclerosis;

• patients affected by (latent) tuberculosis, on the basis of thepatient’s clinical history, physical examination, chest X-ray, andsubcutaneous tuberculin skin test;

• patients who had used experimental, biologic or cytotoxic drugsin the 6 months preceding the trial or who had undergone IAinjection during the 6 weeks before the study.

Of the 34 patients (17 injected with E and 17 with P) enrolledin the blind study 27 (26 IA-P and one IA-E), who had withdrawndue to worsening/non-improvement of KJS subjective and/or pri-mary outcome, entered the prospective open-label extension study(Table 1). The highest cumulative IA-E dose injected into any oneknee during the whole (blind and open-label) study was 62.5 mg.The demographic and clinical characteristics of the patients enter-ing the open-label study, who met the ASAS classification criteriafor peripheral SpA [15], are listed in Table 1.

2.3. Efficacy assessment

The primary outcome was determined by means of the Thomp-son Knee Index (THOMP), a modified Lansbury index assessing kneetenderness and swelling and providing severity indices. Previouslongitudinal studies demonstrated that the THOMP score, a sum ofscores for pain on movement (0–3), soft tissue swelling (0–3) andwarmth (0–3) (range 0–9) [16], is more sensitive to alterations indisease activity and is more closely correlated to laboratory vari-ables than the Ritchie articular index or the swollen joint score.The secondary local disease activity endpoint, assessed at baselineand at the end of the study, included: the local Knee Joint ArticularIndex (KJAI) score which comprises either local disease activity andrange of motion measures. It consists of a sum of scores (0–14) grad-ing tenderness (0–3), joint swelling (0–3), patellar ballottement or“bulge sign” (0–2), range of knee joint flexion (0–3), and exten-sion (0–3) [17]. Systemic secondary endpoints included: the ACR66/68 swollen and tender joint number (ACR-SJC, ACR-TJC), levelsof serum C-reactive protein (CRP) (less or equal to 5 mg/L) and theHealth Assessment Questionnaire-Disability Index (HAQ-DI) (range0–3).

Patients who complained of subjective worsening/non-improvement of SpA KJS, confirmed by the increase/no changeon the THOMP index score that required IA re-treatment, jointaspiration and/or IA injection, were free at any time to withdrawfrom the blind study phase.

2.4. Ultrasound

Ultrasonographic evaluation was carried out using a highfrequency linear transducer (10 MHzElegra, Siemens, Erlangen,Germany) and the scans were assessed by two independent spe-cialized radiologists (Elena Scagliori and Roberto Stramare) blindedto information about patient medication. Standardized anatomi-cal guidelines concerning scans of the suprapatellar, lateral, andmedial parapatellar recesses (SPR, LPPR, MPPR) were adopted, aspreviously described, to assess synovial thickness [17]. At the studyonset, each knee was evaluated as a whole, and the area wherethickening was worse (US-MST) was measured.

2.5. Magnetic resonance imaging

MRI of the knee was performed at baseline, 10 days beforeand 10 days after each IA injection. MRI images were obtainedusing a 0.2 T Magnetom unit (ESAOTE ArtroScan C MR Scanner).The following scan protocol was utilized: axial, coronal and sagittalT1-weighted spin-echo, coronal T1-weighted gradient-echo, coro-nal and sagittal short time inversion recovery (STIR). T1-weightedspin-echo contrast-enhanced MRI (C+MR) [18]. Two of the authors(Antonio Di Maggio and AD) independently reviewed and gradedthe patients’ C+MR images and the distribution of synovial mem-brane and site-specific C+MR-MST measurements expressed inmillimetres. In the event of disagreement, a third author (ES) pro-vided an independent evaluation of the subjective variables.

2.6. Synovial biopsy and immunohistochemistry

Eleven synovial tissue (ST) specimens were collected for biopsyduring arthroscopy while the patients were under anaesthesia.The specimens were taken from areas of intense synovial hyper-emic proliferation. The patients’ serial biopsy samples were storedin paraformaldehyde and embedded in bloc in paraffin [13]. Syn-ovial mononuclear cell infiltration and vessels were evaluated afterimmunostaining using CD45/LCA (clone 2B11 and PD7-26) anti-bodies, labelling mononuclear cells (MNC), as monocytes, DC andlymphocytes, and CD31/PECAM-1 (clone JC70A) antibodies (DakoCytomation, Glostrup, Denmark) respectively. CD45+ and CD31+expression were measured by computer-assisted morphometricanalysis (Image Proplus version 5) and a 2-mm2 area was evalu-ated. The average value over four to six samples per knee joint wasused for analysis.

2.7. Synovial fluid cytokines

SF samples were aspirated from 19 knee joints (out of 27 treatedknees) before the first IA-E injection (baseline) and after IA-Einjections (“post-SF” samples). The SF samples were centrifugedat 1000 g to remove cells and debris and collected and frozen at–80 ◦C. IL-1�, IL-1 receptor antagonist (IL-1Ra) and IL-6 were mea-sured using a commercially available multiplex bead immunoassaybased on a Luminex platform (Fluorokine MAP7 Multiplex HumanCytokine Panel A, R&D Systems, Minneapolis, USA) following themanufacturer’s instructions. As no “normal” synovial fluids wereavailable to determine normal levels, the serum values found in 50healthy blood donors were considered normal levels [13].

2.8. Statistics

Means and standard deviations were used as descriptive statis-tics. The non-parametric Mann-Whitney U test was used to assessdifferences between the distinct treatment groups at different timepoints as well as the correlations between changes in the outcomes

Author's personal copy

168 U. Fiocco et al. / Joint Bone Spine 80 (2013) 165–170

Table 2Comparison of local and systemic disease activity, imaging indexes and biologicalsynovial markers before and after intra-articular TNF-� blockade in SpA knee jointsynovitis. Significance determined by Wilcoxon rank test.

Knee joint synovitis in SpA Baseline (n = 27) End of thestudy (n = 27)

P

Biological and clinical outcomesCRP, mg/dL, mean ± SD 1.2 ± 1.6 0.5 ± 0.9 <0.001THOMP, score, mean ± SD 6.74 ± 1.7 2.03 ± 1.97 <0.001KJAI, score, mean ± SD 9.11 ± 1.84 3.18 ± 3 <0.001HAQ-DI, score, mean ± SD 0.65 ± 0.54 0.41 ± 0.45 <0.01

Imaging outcomesC+MR-MST, mm, mean ± SD 8.76 ± 2.56 7.65 ± 2.33 <0.05US-MST, mm, mean ± SD 7.70 ± 2.82 5.17 ± 2.63 <0.01

Biological synovial markersSynovial tissue (mean ± SD) (n = 9) (n = 9)

CD45+ (cell/2 mm2) 1188.2 ± 657.1 576.3 ± 248.2 <0.01CD31+ (cell/2 mm2) 97.7 ± 28.4 60.3 ± 43.6 =0.050

Synovial fluid (mean ± SD) (n = 17) (n = 17)a

IL-1Ra, pg/mL 9722 ± 7976 4996 ± 6226 < 0.01IL-6, pg/mL 4208 ± 4918 1883 ± 3424 < 0.05IL-1�, pg/mL 6.35 ± 3.75 4.29 ± 0.58 < 0.01

Immunohistochemistry computer-assisted morphometric analysis of a 2-mm2 area.C+MR-MST: contrast-enhanced magnetic resonance maximal synovial thick-ness score; CD: cluster differentiation; CRP: C-reactive protein; HAQ-DI: HealthAssessment Questionnaire-Disability Index; IL: interleukin; IL-1Ra: IL-1 receptorantagonist; KJAI: Knee Joint articular Index score; SpA: spondyloarthritis; THOMP:Thompson’s Knee Index score; US-MST: ultrasonography maximal synovial thick-ness.

a Last synovial fluid available for aspiration after intra-aticular etanercept injec-tions.

of KJS response between pre- vs post-IA injections. The correla-tions between different outcome measures were assessed usingSpearman’s rho test. The Bonferroni correction method was usedfor multiple comparisons.

The changes over time in selected outcomes in the two groupswere evaluated by Wilcoxon Rank test. P values less or equal to 0.05were considered significant. The inter-observer reliability betweenthe two US and C+MR readers was measured by calculatingCronbach’s alpha coefficient and the agreement between theseparameters was assessed by the Bland-Altman method. All datawere processed by SPSS software (version 15.0).

3. Results

All 27 SpA KJS completed the IA-E open-label prospective study.At baseline of the blind study, IA-E and IA-P groups showedno statistically significant differences in both the ACR-SJC score(2.353 ± 2.09 vs 3.765 ± 3.666; p:ns) and in the ACR-TJC score(4.765 ± 4.711 vs 3.706 ± 2.932; p:ns).

Neither severe nor moderate adverse effects and no IA drug-related adverse events were reported during the whole study.

3.1. KJS response before and after open-label IA TNF blockade

3.1.1. Clinical measuresAt end of the open-label extension study, compared to the base-

line, systemic disease activity CRP index was significantly reduced(P < 0.001); both the IA-E and IA-P groups showed no statisticallysignificant changes in the ACR-SJC as well in the ACR-TJC scores.

The SpA KJS which underwent open-label IA-E injectionsshowed a significant reduction in either the local composite THOMP(P < 0.001) and KJAI (P < 0.001), or the systemic HAQ-DI score(P < 0.01) (Table 2).

Fig. 2. Correlation between: a: the composite knee Thompson score (THOMP) andCD45+ mononuclear cells in synovial tissue; b: between THOMP and log IL-6 levelin synovial fluid.

3.1.2. Imaging measuresInter-observer reliability of the US-MST, C+MR-MST measure-

ments were found to have an adequate level of concordance(Cronbach’s alpha: 0.810 and 0.817, respectively). The Bland-Altman analysis detected a bias of –1.06 mm (3.12 SD) in the twoprocedures, with 95% concordance limits between –7.3 and 5.2.

Compared to the baseline, both US-MST and C+MR-MST mea-sures at the end of the open-label IA extension study showed asignificant reduction (P < 0.01; P < 0.05) (Table 2).

3.1.3. Synovial biological parametersThere was a significant reduction in IL-1�, IL-1Ra, IL-6 levels in

“post”-SF as compared to basal SF values (P < 0.01; P < 0.01; P < 0.05)(Table 2).

ST-CD45+MNC infiltration and ST-CD31+ vessels were all signi-ficantly lower than baseline values (P = 0.05) (Table 2 and Fig. 3).

3.2. Correlations between clinical, imaging and biologicalparameters of the KJS response

There were significant correlations between clinical, imag-ing and biological markers of the knee joint synovitis response,assessed by considering baseline and end of the extension studyas a whole: CRP with either THOMP (P < 0.001), or KJAI (P < 0.01),or HAQ-DI (P < 0.05) or SF-IL-1Ra (P < 0.05); US-MST with KJAI(P < 0.05); ST-CD45+ with either THOMP (P < 0.001) (Fig. 2), orKJAI (P < 0.01), or ST-CD31+ (P = 0.001) (Figs. 3 and 4), or SF-IL-1� (P < 0.05); SF-IL-6 with either THOMP (P < 0.05) (Fig. 2), or KJAI

Fig. 3. Correlation between CD45+ mononuclear cells and CD31+ vessels in synovialtissue.

Author's personal copy

U. Fiocco et al. / Joint Bone Spine 80 (2013) 165–170 169

Fig. 4. Immunohistochemical staining of infiltrating mononuclear cells (CD45) andendothelial cells (CD31) in synovial tissue before and after intra-articular etanercept(IA-E) treatment of the knee. Staining of CD45 (a and c) and CD31 (b and d) onsynovial biopsies, taken before and after IA-E treatment, from the same biopsy areaand patient. Note the reduced CD45 and CD31 staining after treatment (c and d).

(P < 0.05), or SF-IL-1� (P < 0.0001), or IL-1Ra (P < 0.0001); SF-IL-1�with SF-IL-1Ra (P < 0.0001).

4. Discussion

This open-label extension study, compared to the previousblind study, allows a more accurate longitudinal evaluation of thelocal effect of the new IA therapeutic procedure, since the defi-nite number of IA drug injections for each knee joint of peripheralSpA affected patients and the single joint translational approach,including a wider range of imaging, clinical, and synovial biologicalmeasures.

The early significant improvement in the clinical compositeinflammation and functional indexes of local disease activity, aswell in the US and MRI imaging measures suggests that multipleIA-E injections may exert certain cumulative effects at the jointlevel. A single etanercept dose of 12.5 mg was indeed insufficientfor obtaining a lasting clinical effect [6,7] or to induce early US orMRI alterations after the IA injection [19]. As to the imaging for kneesynovial thickness measurements by MRI and US, both internal andexternal responsiveness have been reported, despite the lack ofvalidated scoring systems [20]. We found high concordancebetween gray-scale US with C+MRI, in agreement with observa-tions by Song, who recently reported that synovial thickness mea-surements by US had a good level of concordance with C+MRI [21].

After TNF-� blockade, the changes in the KJAI scores confirmedthe significant reduction in the HAQ-DI, signalling an importantfunctional improvement in patients, that was more pronouncedthan the one reported after systemic etanercept [11].

The present study has certain limitations: its single-centre,open-label design, the small number of sample, and the shortfollow-up time.

Systemic CRP levels correlated significantly with SF levels of IL-1Ra, that represents a host response directed at tempering chronicinflammation [22]. Improvement in the CRP levels following IA-E injections could mean that there was a rapid change in disease

activity in large single joints, or that IA-E was distributed systemi-cally, outside the joint.

The expression of inflammatory IL-6, IL-1�, TNF-� CK, both in SFand in synovial tissue (ST) of peripheral SpA, was indeed observed[13,23]. IL-6 is involved either in the induction of monocyte-specific chemokines by endothelial cells or in the differentiationof macrophages or of effector T cells [24]. In particular, IL-1� andIL-6 take also part in the differentiation of TH IL-17-producing cells[25]. Further, IL-1� and TNF-� act in synergy with IL-6 in produc-ing vascular endothelial growth factor (VEFG) [26]. The associationof CD31+ with CD45+ expression in synovial tissue is consistentwith MNC interacting with the marked expression of IL-6 by CD31+endothelial cells previously reported in lesional psoriatic skin [27].

In keeping with the alterations due to systemic TNF-� block-ade such as the down-regulation of CD68+ monocytes and CD3+T cells [3,28–30], the reduced infiltration of CD45+ leucocytes insynovial tissue supports the role of MNC as candidate biomarkersof SpA and PsA synovitis [3,13,30]. This finding is also con-sistent with the recent evidences of the involvement of thecellular components of the innate immune response in SpA[31,32].

The reduction of vascularity, already observed after systemicinfliximab treatment [33], was found to be only moderate in SpAsynovium [29] or not detectable after 12 weeks of systemic etan-ercept treatment in psoriatic synovitis [30], underlining the localselective effect of IA TNF blockade on synovial vessel depletion [34].Angiogenesis and inflammation are inversely related to synovialhypoxia [35]. Inflammatory CK and in particular a higher IL-6 secre-tion by MNC, which is involved in the mechanisms of resistance toTNF-�-blocking therapy [35], may be induced by synovial hypoxia[32].

Previous studies by us and others on the effect of systemic or IAetanercept treatment in psoriasis, PsA and RA showed a reductionof the synthesis of either the T helper 1 (TH-1) cytokines or of theIL-6/17 axis cytokines such as IL-6, IL-17, IL-22 and IL-23, or of theCC chemokine such as CCL3 or CCL20 and of IL-8, both at systemicand joint levels [13,36,37].

The therapeutic action of etanercept has been shown to involvethe down-regulation of TNF-� signalling via either NF�B and cAMP,or the oxidative stress response in RA [38]. In psoriasis, etanercepthas been found to affect either TNF early innate response genes(IL-1 � and IL-8), or late IL-23 p19 and p40 subunit inflammatorypathway genes [39]. These findings indicate that the deactivationof antigen-presenting cells/T cell-multiple gene expression may beimportant for the mechanism underlying etanercept therapy [40].

The functional dynamics in vivo of TNF-� and IL-6 have indeedemerged to be highly crucial to the outcome of etanercept therapy[13,38]. The association of IL-6 level in SF to joint inflammationunderlines a role for IL-6 in the mechanism of persistent SpAsynovitis. New evidences indicate that the concentration of IL-6 inSF may predict the response to TNF inhibitors in RA [14], suggestingthat the approach of assessing biological markers at local jointlevel can be clinically meaningful.

Disclosure of interest

Ugo Fiocco has received speaking fees and/or research grantsfrom Wyeth Lederle, Schering Plough and Bristol-Myers Squibb.

Leonardo Punzi has received speaking fees and/or researchgrants from Wyeth Lederle, Schering Plough, Bristol-Myers Squibb,Abbott International, Rottapharm, Fidia Farmaceutici and Roche.

Funding: This work was in part supported by a joint researchgrant from the Padova University Hospital and Wyeth Lederle SpA(Wyeth Pharmaceuticals, USA).

Author's personal copy

170 U. Fiocco et al. / Joint Bone Spine 80 (2013) 165–170

The work of Roberto Nardacchione was in part supported by theLeonardo Foundation — Abano General Hospital.

Authors’ contributions: Ugo Fiocco was responsible for the studyconcept and design, analysis and interpretation, and drafting themanuscript. Luisa Costa participated in the clinical diagnostic eva-luation of the patients. Paolo Sfriso participated in the design of thestudy and performed the statistical analysis. Francesca Oliviero per-formed the statistical analysis and helped to draft the manuscript.Elena Scagliori participated in the assessment of the patients.Roberto Nardacchione carried out the arthroscopy and synovialbiopsies. Luisella Cozzi assessed the patients’ response to therapy.Antonio Di Maggio and Roberto Stramare performed magnetic reso-nance imaging. Fiorella Calabrese was involved in the pathologicaldiagnosis and Francesca Lunardi in immunohistochemical charac-terization. Benedetta Accordi and Béatrice Molena helped to carryout the immunohistochemistry. Mara Felicetti and Katia Gazzolaparticipated in the assessment of the patients. Léopoldo Rubaltelliperformed the ultrasonography. Pascale Roux-Lombard carried outthe immunoassays. Jean-Michel Dayer participated in the design ofthe study and revised the manuscript. Leonardo Punzi was respon-sible for the study concept and revising the manuscript. All authorsread and approved the final manuscript.

References

[1] Braun J, Sieper J. Ankylosing spondylitis. Lancet 2007;369:1379–90.[2] Baeten D, Kruithof E, De Rycke L, et al. Diagnostic classification of

spondylarthropathy and rheumatoid arthritis by synovial histopathol-ogy: a prospective study in 154 consecutive patients. Arthritis Rheum2004;50:2931–41.

[3] Kruithof E, De Rycke L, Vandooren B, et al. Identification of synovial biomarkersof response to experimental treatment in early-phase clinical trials in spondy-larthritis. Arthritis Rheum 2006;54:1795–804.

[4] Fisher BA, Keat A. Should we be using intra-articular tumor necrosis factorblockade in inflammatory monoarthritis? J Rheumatol 2006;33:1934–5.

[5] Fiocco U, Punzi L. Are there any evidences for using the intra-articular TNF-�blockade in resistant arthritis? Joint Bone Spine 2011;78:331–4.

[6] Osborn TG. Intra-articular etanercept vs saline in rheumatoid arthritis: a sin-gle injection double-blind placebo-controlled study. Arthritis Rheum 2002;46Suppl.:S518.

[7] Bliddal H, Terslev L, Qvistgaard E, et al. A randomized, controlled study of asingle intra-articular injection of etanercept or glucocorticosteroids in patientswith rheumatoid arthritis. Scand J Rheumatol 2006;35:341–5.

[8] van der Bijl AE, Teng YK, van Oosterhout M, et al. Efficacy of intra-articular inflix-imab in patients with chronic or recurrent gonarthritis: a clinical randomizedtrial. Arthritis Rheum 2009;61:974–8.

[9] Keen HI, Bingham 3rd CO, BradLey LA, et al. Assessing single joints in arthritisclinical trials. J Rheumatol 2009;36:2092–6.

[10] Fiocco U, Sfriso P, Nardacchione R, et al. Evaluation of the efficacy and safety ofrepeated intra-articular etanercept injections in resistant knee joint synovitis.Arthritis Rheum 2007;56:S592.

[11] Fiocco U, Sfriso P, Scagliori E, et al. Soft tissue changes induced by repeatedserial intra-articular etanercept injections in the knee joints of patients withresistant knee joint synovitis. Arthritis Rheum 2008;58:S588.

[12] Singh R, Aggarwal A, Misra R. Th1/Th17 cytokine profiles in patientswith reactive arthritis/undifferentiated spondyloarthropathy. J Rheumatol2007;34:2285–90.

[13] Fiocco U, Sfriso P, Oliviero F, et al. Synovial effusion and synovial fluid biomark-ers in psoriatic arthritis to assess intra-articular tumor necrosis factor-alphablockade in the knee joint. Arthritis Res Ther 2010;12:R148.

[14] Wright HL, Bucknall RC, Moots RJ, et al. Analysis of SF and plasma cytokines pro-vides insights into the mechanisms of inflammatory arthritis and may predictresponse to therapy. Rheumatology (Oxford) 2012;51:451–9.

[15] Rudwaleit M, van der Heijde D, Landewé R, et al. The Assessment of SpondyloArthritis International Society classification criteria for peripheral spondy-loarthritis and for spondyloarthritis in general. Ann Rheum Dis 2011;70:25–31.

[16] Thompson PW, Silman AJ, Kirwan JR, et al. Articular indices of joint inflam-mation in rheumatoid arthritis. Correlation with the acute-phase response.Arthritis Rheum 1987;30:618–23.

[17] Fiocco U, Ferro F, Vezzu M, et al. Rheumatoid and psoriatic knee synovitis:clinical, grey scale, and power Doppler ultrasound assessment of the responseto etanercept. Ann Rheum Dis 2005;64:899–905.

[18] Rhodes LA, Grainger AJ, Keenan AM, et al. The validation of simple scoringmethods for evaluating compartment specific synovitis detected by MRI in kneeosteoarthritis. Rheumatology 2005;44:1569–73.

[19] Boesen M, Boesen L, Jensen KE, et al. Clinical outcome and imaging changesafter intra-articular (IA) application of etanercept or methylprednisolone inrheumatoid arthritis: magnetic resonance imaging and ultrasound-Dopplershow no effect of IA injections in the wrist after 4 weeks. J Rheumatol 2008;35:584–91.

[20] Keen HI, Mease PJ, Bingham 3rd CO, et al. Systematic review of MRI, ultra-sound, and scintigraphy as outcome measures for structural pathology ininterventional therapeutic studies of knee arthritis: Focus on responsiveness. JRheumatol 2011;38:142–54.

[21] Song IH, Althoff CE, Hermann KG, et al. Contrast-enhanced ultrasound in mon-itoring the efficacy of a bradykinin receptor 2 antagonist in painful kneeosteoarthritis compared with MRI. Ann Rheum Dis 2009;68:75–83.

[22] Chizzolini C, Dayer JM, Miossec P. Cytokines in chronic rheumatic diseases: iseverything lack of homeostatic balance? Arthritis Res Ther 2009;11:246.

[23] van Kuijk AW, Reinders-Blankert P, Smeets TJ, et al. Detailed analysis of the cellinfiltrate and the expression of mediators of synovial inflammation and jointdestruction in the synovium of patients with psoriatic arthritis: implicationsfor treatment. Ann Rheum Dis 2006;65:1551–7.

[24] Assier E, Boissier MC, Dayer JM. Interleukin-6: from identification ofthe cytokine to development of targeted treatments. Joint Bone Spine2010;77:532–6.

[25] Acosta-Rodriguez EV, Napolitani G, Lanzavecchia A, et al. Interleukins 1betaand 6 but not transforming growth factor-beta are essential for the differenti-ation of interleukin 17-producing human T helper cells. Nat Immunol 2007;8:942–9.

[26] Cohen T, Nahari D, Cerem LW, et al. Interleukin 6 induces the expression ofvascular endothelial growth factor. J Biol Chem 1996;271:736–41.

[27] Goodman WA, Levine AD, Massari JV, et al. IL-6 signaling in psoriasis pre-vents immune suppression by regulatory T cells. J Immunol 2009;183:3170–6.

[28] Baeten D, Kruithof E, van den Bosch F, et al. Immunomodulatory effects ofanti-tumor necrosis factor-therapy on synovium in spondylarthropathy: histo-logic findings in eight patients from an open-label pilot study. Arthritis Rheum2001;44:186–95.

[29] Kruithof E, De Rycke L, Roth J, et al. Immunomodulatory effects of etanercepton peripheral joint synovitis in the spondyloarthropathies. Arthritis Rheum2005;52:3898–909.

[30] Pontifex EK, Gerlag DM, Gogarty M, et al. Change in CD3 positive T cellexpression in psoriatic arthritis synovium correlates with change in DAS28and magnetic resonance imaging synovitis scores following initiation of bio-logic therapy – a single-centre, open-label study. Arthritis Res Ther 2011;13:R7.

[31] Appel H, Maier R, Wu P, et al. Analysis of IL-17+ cells in facet joints of patientswith spondyloarthritis suggests that the innate immune pathway might be ofgreater relevance than the Th17-mediated adaptive immune response. Arthri-tis Res Ther 2011;13:R95.

[32] Moran EM, Heydrich R, Ng CT, et al. IL-17A expression is localised toboth mononuclear and polymorphonuclear synovial cell infiltrates. PLoS One2011;6:e24048.

[33] Canete JD, Pablos JL, Sanmartí R, et al. Antiangiogenic effects of anti-tumornecrosis factor-alpha therapy with infliximab in psoriatic arthritis. ArthritisRheum 2004;50:1636–41.

[34] Izquierdo E, Canete JD, Celis R, et al. Immature blood vessels in rheumatoidsynovium are selectively depleted in response to anti-TNF therapy. PLoS One2009;4:e8131.

[35] Kennedy PA, Ng CT, Chang TC, et al. TNF-blocking therapy alters joint inflam-mation and hypoxia. Arthritis Rheum 2011;63:923–32.

[36] Caproni M, Antiga E, Melani L, et al. Serum levels of IL-17 and IL-22 are reducedby etanercept, but not by acitretin, in patients with psoriasis: a randomized-controlled trial. J Clin Immunol 2009;29:210–4.

[37] Kageyama Y, Ichikawa T, Nagafusa T, et al. Etanercept reduces the serum levelsof interleukin-23 and macrophage inflammatory protein-3 alpha in patientswith rheumatoid arthritis. Rheumatol Int 2007;28:137–43.

[38] Koczan D, Drynda S, Hecker M, et al. Molecular discrimination of respondersand non-responders to anti-TNF-alpha therapy in rheumatoid arthritis by etan-ercept. Arthritis Res Ther 2008;10:R50.

[39] Zaba LC, Suárez-Farinas M, Fuentes-Duculan J, et al. Effective treatment of pso-riasis with etanercept is linked to suppression of IL-17 signaling, not immediateresponse TNF genes. Allergy Clin Immunol 2009;124:1022–30.

[40] Gottlieb AB, Chamian F, Masud S, et al. TNF inhibition rapidLy down-regulates multiple proinflammatory pathways in psoriasis plaques. J Immunol2005;175:2721–9.