Overweight/obesity affects histological features and inflammatory gene signature of synovial membrane of Rheumatoid Arthritis

Overweight/obesity influence disease burden and clinical outcome of Rheumatoid Arthritis (RA). The impact of overweight/obesity on synovial tissue (ST) inflammation is largely unknown. Here, we investigated the histological and transcriptional signature of ST obtained from RA in different disease phases (disease onset, failure to first-line conventional DMARDs and in sustained clinical and ultrasound remission) finding that overweight/obese DMARDs naive RA showed higher likelihood of follicular synovitis, higher IHC scores for sublining inflammatory cells (CD68+, CD21+ and CD20+) and higher IL-1RA plasma levels than normal weight RA. Regardless to the synovitis pattern, overweight/obese DMARDs naive RA showed a worse clinical response to “Treat-to-target” (T2T) than normal weight RA at 6 and 12 months follow-up. Conversely, MTX-IR RA did not show significant differences in synovial inflammation based on BMI category. Overweight/obese RA in stable clinical and US remission showed higher degree of residual synovitis in terms of sublining CD68+, CD20+ cells and lining and sublining CD3+ compared to normal weight RA. Finally, gene expression profile analysis revealed that ST of overweight/obese DMARDs naive RA is enriched by CCL3 and MyD88 compared to normal weight RA in sustained disease remission, the latter correlating with BMI and IHC scores for synovial CD68+ cells. These findings suggest that indeed overweight/obese RA show higher degree of synovitis at disease onset and after remission achievement that influences the response rate to T2T and should be considered within the management of patients with RA.

There are increasing evidences that overweight and obesity are risk factors for the development of Rheumatoid Arthritis (RA) and that a high Body Mass Index (BMI) is associated with high disease activity and disability at disease onset 1-3 , being an independent factor of worse clinical response to RA treatment [4][5][6] . Despite this, pooling data derived from multiple imaging studies showed that BMI is inversely associated with joint inflammation compared to normal weight RA [7][8][9] .
BMI category influences the response rate to T2T strategy in naive to treatment RA independently from synovitis pattern. The analysis of overweight/obesity status effect on remission achievement after T2T scheme, revealed that naive RA with BMI ≥ 25 kg/m 2 showed significantly lower rate of DAS remission after 6 months (28.2%) and 12 months (37.8%) follow-up compared to naive RA with BMI < 25 kg/m 2 (57.1% and 67.9% RA in DAS remission after 6 and 12 months of follow-up; p = 0.02 and p = 0.01 respectively). Stratifying naive RA based on the synovitis pattern detected at the ST level, before the beginning of DMARDs treatment, RA with BMI ≥ 25 kg/m 2 and follicular synovitis showed significantly lower rate of DAS remission achievement after 6 (34.5%) and 12 months (37.8%) follow-up compared to RA with BMI < 25 kg/m 2 and diffuse synovitis [64.7% and 70.6% RA with BMI < 25 kg/m 2 and diffuse synovitis in DAS remission after 6 (p = 0.005) and 12 months (p = 0.03) follow-up] treated with the T2T (see Supplementary Fig. 2A,B). Moreover, considering the IHC scores for inflammatory synovial cells, there were no significant differences in terms of baseline IHC scores for CD68 + , CD21 + , CD20 + and CD3 + cells comparing RA reaching or not DAS remission after 6 and 12 months follow-up, regardless to the BMI category (data not shown). None of the naive RA experienced significant weight variation leading to change of the initial BMI category during the follow-up.

BMI does not influence IHC characteristics of RA patients inadequately responding to MTX.
Since overweight/obesity status affected the response rate to T2T strategy (see Supplementary Fig. 2A,B), we analysed histological synovial features of an independent cohort of MTX-IR RA, finding no significant difference in terms of ST inflammation stratifying patients based on the BMI category ( Fig. 3A-H). In particular, in MTX-IR RA there was a similar rate of follicular synovitis comparing patients with BMI ≥ 25 kg/m 2 (48.2%) and patients with BMI < 25 kg/m 2 (52.1%; p = 0.83) (Fig. 3I). Moreover, there were no significant differences in lining and sublining IHC scores for CD68 + , CD21 + , CD3 + and CD20 + cells in MTX-IR RA based on the BMI category ( Fig. 3J,M). Therefore, the inflammation driving the MTX-IR seems to reside in the joint not in the fat tissue.

Discussion
This study shows that overweight/obesity affects the histological features and the gene expression profile of ST of RA patients at the time of disease onset and at achievement of sustained clinical and ultrasound remission. In addition, overweight/obesity condition at the time of DMARDs beginning influences the rate of disease remission achievement in RA treated according to the T2T strategy regardless to the synovitis pattern found at ST level at disease onset.
RA is a chronic disease in which inflammatory cells aberrantly migrate within the ST contributing to joint inflammation and bone damage 13,14 . Obesity incidence is increasing in the general population 15 , and multiple studies confirmed that obesity is a risk factor associated with RA development [16][17][18] . Several lines of evidence have proven that adipose tissue is an endocrine organ acting not only on metabolism but even on immune and inflammatory processes by releasing inflammatory molecules promoting the chronicity of the inflammatory response in the target tissue 19 . It has been shown that the overweight/obesity status is associated with a higher degree of systemic inflammation and disease burden in RA patients at disease onset 3,20 , and with an inferior response to treatment 1,6,21-23 . Adipose tissue in normal weight subjects is mostly composed by mature adipocytes, their precursors, fibroblasts, endothelial cells and scarce immune cells, mainly macrophages 24 . During the progressive and excessive fat accumulation occurring in overweight/obese subjects, there is a substantial increase in the number of immune cells and a change in their phenotype towards pro-inflammatory cells contributing to the development of obesity-related local and systemic inflammation (the so called low grade inflammation) 25 . In particular, within the adipose tissue of obese individuals, macrophages are the most abundant resident immune cells characterized by a pro-inflammatory (M1) phenotype, active in the secretion of inflammatory cytokines 10,26 . In the present study, for the first time, we evaluated the histological features of ST of RA patients enrolled at different disease phases (i.e. naive to treatment, inadequately responder to conventional DMARDs and at the time of stable clinical and US remission after MTX + TNF-i combined therapy) stratified according to the BMI category. Considering RA at onset, before any DMARDs treatment, we found that overweight/obese RA showed higher rate of follicular synovitis and higher IHC scores for resident synovial inflammatory cells (i.e. CD68 + , CD21 + and CD20 + ) compared to normal weight ones, with a direct correlation between BMI value and IHC scores for resident inflammatory cells suggesting a tight link between body weight and degree of ST inflammation at RA onset. These findings are in line with Kim et al.'s study which investigated the early effect of obesity in RA using preclinical animal models, finding that obese Collagen Induced Arthritis (CIA) mice have an earlier disease onset compared to the lean ones due to a faster immune cells migration within the joint tissue 10 .
To date, despite the direct association between high BMI and disease activity in RA, bone damage is less likely associated with obesity in RA 11,12 , suggesting that protective factors may be present in obese patients. It's well known that adipose tissue is a major source of IL-1RA 27 , whose release was found to be increased in human obesity 28 . In our cohort, we found that naive overweight/obese RA showed higher IL-1RA plasma levels compared to normal weight RA regardless to the synovitis subtype directly related to IL-6 plasma levels possibly due to the inflammatory status. Moreover, despite we observed no differences at baseline in erosion score comparing overweight/obese and normal weight naïve to treatment RA patients, overweight/obese naive RA without erosive disease showed the lowest IL-6/IL-1RA ratio supporting the concept of an unbalance between pro-inflammatory and osteoclastogenic via OPG/RANK pathway, such as IL-6 29 , and bone protective (as IL-1RA) soluble factors in naive RA based on BMI category. Interestingly, at the time of sustained remission achievement, IL-1RA is significantly higher in overweight/obese RA who did not developed erosive disease, supporting the hypothesis that overweight/obese RA patients may have been protected through IL-1RA pathway from bone damage.
Finally, to investigate the role of overweight/obesity status on ST composition during the remission phase of the disease, we included in the study RA patients in stable sustained clinical and ultrasound remission under combination therapy with MTX + TNF-i as previously described 30 . We found that overweight/obese RA patients reaching stable clinical and ultrasound remission showed higher degree of residual synovitis in terms of synovial CD68 + , CD20 + and CD3 + cells whose IHC scores directly correlated with the BMI value at the time of remission achievement. Interestingly, IL-6 plasma levels, despite significantly reduced at the time of remission achievement compared to patients at diagnosis, directly correlates with the IHC score for sublining CD68 + cells only in overweight/obese RA in sustained clinical and ultrasound remission, supporting the notion that fat excess can promote IL-6 release, contributing to the persistence of residual synovial inflammation, despite good clinical response. These findings are in line with previous data from animal models showing that obese arthritic mice show a significant delay in remission achievement compared to lean ones due to the aberrant polarization, within the synovial tissue, of macrophages towards a pro-inflammatory (M1) phenotype 10 . p = 0.63 for CD20 + and R = 0.30, p = 0.14 for CD3 + cells respectively) in RA in stable clinical and ultrasound remission under MTX + TNF-i; (M) Correlation between IL-6 PB levels and IHC of sublining CD68 + cells (R = 0.43; p = 0.04) in RA in stable clinical and ultrasound remission under MTX + TNF-i (red dots indicate RA with BMI ≥ 25 kg/m 2 and green dots indicate RA with BMI < 25 kg/m 2 , respectively); RA: Rheumatoid Arthritis; IHC: Immunohistochemistry; BMI: Body Mass Index; CD: Cluster designation; MTX: Methotrexate; TNF-i: Tumor Necrosis factor inhibitor.
www.nature.com/scientificreports www.nature.com/scientificreports/ These concepts are strengthened by the findings obtained from the gene expression analysis conducted, in the exploratory cohort, on ST from RA at different disease phases stratified by BMI category, revealing that ST from naive overweight/obese RA is enriched by inflammatory genes as CCL3 and MyD88 compared to normal weight RA in sustained disease remission. Interestingly, ST biopsies from overweight/obese RA in sustained remission showed a persistent over-expression of MyD88 compared to normal weight RA in sustained disease remission. MyD88 (myeloid differentiation primary response gene 88) is a central adaptor molecule for the majority of Toll-like receptors (TLRs), which are the most studied pathogen recognition receptors 31,32 . TLR are www.nature.com/scientificreports www.nature.com/scientificreports/ trans-membrane receptors that play a crucial role in pathogen recognition and immune response by activating various inflammatory signaling pathways, including MyD88, which then leads to an activation of NF-kB signaling activity 32 . High fat diet was demonstrated to induce an increased expression of TLR in murine adipose tissue causing the activation of MyD88 signaling cascades 33 , and MyD88 expression was found to be increased in both peripheral blood mononuclear cells and subcutaneous adipose tissue of overweight/obese subjects compared to normal weight controls 34 . Moreover, MyD88 deficient mice are partially protected to high fat diet induced obesity 35 . Yu et al. demonstrated that MyD88 signaling in myeloid cells participates in the initiation and progression of obesity-induced systemic low grade inflammation since MyD88 deficiency in myeloid cells inhibits macrophage recruitment to adipose tissue and their switch to an M1-like phenotype 36 . These findings are in line with MyD88 expression in ST of RA patients, directly correlating with BMI value regardless to RA phase and with the IHC scores of lining and sublining CD68 + cells in ST biopsies supporting the notion that an aberrant persistent over-expression of MyD88, dependent from fat mass excess, is involved in the promotion of residual synovitis persistence in ST of overweight/obese RA, despite sustained disease control.
Therefore, based on our findings we conclude that the early and the resolution phase of RA are influenced by overweight/obesity status in terms of ST inflammation and bone remodelling. Moreover, the ST histological composition found at RA onset strongly supports the lower rate of treatment response, in terms of DAS28 reduction, found in obese patients belonging to early RA cohorts 37 . Therefore, these results suggest that weight control is a crucial aim along the whole disease course in RA and interventional studies, including body weight reduction, are necessary to definitively confirm the biological effect of adipose tissue on the entity of RA inflammation and to envision a personalized approach for obese RA.

Methods
Patients recruitment. One hundred and thirty-eight consecutive patients fulfilling the American College of Rheumatology 2010 revised criteria for RA 38 , were enrolled. RA patients were divided into naïve to treatment (n = 70), inadequately responder to Methotrexate (MTX-IR) (n = 43) and patients in stable remission under combination of MTX + TNF-inhibitor (TNF-i) (n = 25). All MTX-IR RA were taking stable dose of MTX (mean dose: 13.8 ± 5.7 mg/week). All RA in sustained clinical (DAS44 < 1.6 for at least 6 months) and ultrasound remission were selected based on the published protocol 29 . For each enrolled RA, clinical and laboratory evaluations included the number of tender and swollen joints on 44, erythrocyte sedimentation rate (ESR), C-reactive protein (CRP) and Disease Activity Score (DAS). Peripheral blood (PB) samples were tested for IgA and IgM-RF (Orgentec Diagnostika, Bouty-UK) and ACPA (Menarini Diagnostics-Italy) using commercial Enzyme-Linked Immunosorbent Assay (ELISA) and ChemiLuminescence Immunoassay (CLIA) methods respectively. At study entry, for each enrolled RA, Body Mass Index (BMI) was assessed 39 , and patients were stratified according to the following cut-off value references: <25 as normal weight; 25-29.9 as overweight; ≥30 as obese, respectively. After study enrolment, all naïve RA started MTX 10 mg/weekly for 2 subsequent administrations and 15 mg/weekly afterwards, according to RA management recommendations 40 . Therefore, each RA was followed every 3 months for at least 12 months in an outpatient setting and DAS value was recorded to assess treatment response and the BMI category was registered. All methods were carried out in accordance with the declaration of Helsinki. All the study experimental protocols were approved by the Ethic Committee of the Università Cattolica del Sacro Cuore (Protocol number: 19526/17) and all subjects provided signed informed consent.
Immunohistochemistry for CD68 + , CD21 + , CD20 + and CD3 + cells in synovial tissue. All enrolled RA underwent ultrasound guided ST biopsy of the knee following the published protocol 41 . Briefly, sections were stained for CD68 mouse anti-human monoclonal antibody (514H12) (for macrophages) or CD20, mouse anti-human monoclonal antibody (L26) (for B lymphocytes) or CD3 mouse anti-human monoclonal antibody (LN10) (for T lymphocytes) or CD21 mouse anti-human monoclonal antibody (2G9) (for follicular dendritic cells) 42,43 (all by Leica Biosystem, Newcastle-UK) by immunostainer BOND MAX III (Leica). Double Immunohistochemical staining for CD21/CD68 and CD20/CD3 was performed as previously described 30 . Slides were examined using a light microscope (Leica DM 2000) and classified as diffuse or follicular based on the immunostaining of CD68, CD21, CD20 and CD3 positive cells 42 . Specific lymphoid features of the cellular aggregates were assessed by staining 3-4 μm-thick FFPE consecutive sections for CD3, CD20 and CD21 as previously defined 42,43 . All tissues were evaluated using a numerical score based on the number of positive cells in the lining and sublining areas of the section (three different fields in each section), with a score of 0 indicating no positive cells; 1 indicating <10% positive cells; 2 indicating 10-50% positive cells; and 3 indicating >50% positive cells 30 . Inter-rater agreement coefficients for CD68, CD21, CD20 and CD3 IHC scores were assessed (see Supplementary Table 3).
ELISA assay for IL-6 and IL-1RA plasma levels evaluation. Naive to treatment RA and RA patients in stable clinical and ultrasound remission were tested for IL-6 and IL-1RA PB levels using commercial Enzyme-Linked ImmunoSorbent Assay (ELISA) kits (all by R&D Systems, United Kingdom). The sensitivity of the test was 0.70 pg/ml for IL-6 and 18.3 pg/ml for IL-1RA respectively.
Gene expression profile of synovial tissue of RA patients with naive active disease and in sustained clinical and ultrasound remission based on BMI category. Total RNA was isolated from synovial tissue of 15 RA patients (8 from RA with naive active disease and 7 from RA in sustained clinical and ultrasound remission as previously described) using the miRneasy kit (Qiagen). RNA was reverse transcribed using a cDNA conversion kit (Qiagen). The cDNA was used on the real-time RT 2 Profiler PCR Array (QIAGEN, Cat. no. PAHS-077Z) in combination with RT2 SYBR ® Green qPCR Mastermix (Cat. no. 330529). A set of controls was included on each plate which enabled data analysis using ΔΔCt method of relative quantification, assessment of reverse transcription performance and assessment of PCR performance. The RT² Profiler PCR Array www.nature.com/scientificreports www.nature.com/scientificreports/ enables SYBR Green-based real-time PCR analysis using Biorad iQ5 real-time PCR system as follows: 95 °C for 15 min; 40 cycles of 94 °C for 15 s; 55 °C for 30 s; and 70 °C for 30 s. The relative expression was calculated using the ΔΔCt method (relative gene expression = 2(ΔCt test − ΔCt control)] and is presented in fold increase relative to control. The Web-based GeneGlobe Data Analysis Center was used to analyse the real-time PCR data (Qiagen). Statistical analysis. Statistical analysis was performed using SPSS V. 20.0 (SPSS. Chicago, Illinois, USA) and Prism Software (GraphPad, San Diego, California, USA). Categorical and quantitative variables were described as frequencies, percentage and mean ± standard deviation (SD). Data on demographic and clinical features were compared between patients by the non-parametric Mann-Whitney U test or χ 2 test, as appropriate. Spearman's rank correlation test was used for correlation in all analyses. For the gene expression profile, data analysis was performed using the supplied software (http://www.qiagen.com/it/shop/genes-and-pathways/data-analysis-center-overview-page/), based on Student's t-test of the replicate 2 (−ΔCt) values for each gene in the tested group and in the control group. Fold-change values > 1 imply an upregulation while fold-change <1 imply down-regulation. using RT 2 Profiler PCR array, a t-test was used to identify significant differences in gene expression profiles between overweight/obese and normal weight naive active RA, overweight/obese RA in sustained remission and normal weight RA in sustained remission used as control. A value of p ≤ 0.05 was considered statistically significant.