Joint damage is amplified in rheumatoid arthritis patients with positive thyroid autoantibodies

Patients

Between February 2015 and August 2016, hospitalized patients who fulfilled the 1987 revised criteria of the ACR (Arnett et al., 1988) or the 2010 ACR/European League Against Rheumatism (EULAR) criteria for the classification of RA (Aletaha et al., 2010) were consecutively recruited in the Department of Rheumatology at Sun Yat-Sen Memorial Hospital, Guangzhou, China. All patients were aged ≥ 18 years. The exclusion criteria included the following: individuals overlapping with other autoimmune diseases (e.g., systemic lupus erythematosus, scleroderma, polyarteritis nodosa, dermatomyositis); with hypothalamus or pituitary disease; with serious infection, organ failure, or malignancy; with a history of surgical removal of the thyroid gland; or treated with iodine-containing drugs (e.g., amiodarone, lithium, carbamazepine, phenytoin sodium, interferon-alpha). Written informed consent was obtained from all patients. This study was approved by the Medical Ethics Committee of Sun Yat-Sen Memorial Hospital (identifier: SYSEC-2009-06 and SYSEC-KY-KS-011).

Clinical assessments

The clinical data of the RA patients were collected at enrollment, as described before (Ma et al., 2015), including the 28-joint tender and swollen joint counts (28TJC and 28SJC, both 0–28), patient and provider global assessments of disease activity (PtGA and PrGA, both 0–10 cm, 10 = worst status), pain visual analog scale (Pain VAS, 0–10 cm, 10 = most pain), Chinese-language version of the Stanford Health Assessment Questionnaire (HAQ, 0–3, 3 = most functional disability; functional limitation was defined as an HAQ score >1), erythrocyte sedimentation rate (ESR, mm/h, normal range: 0–20 mm/h (female), 0–15 mm/h (male)), level of C-reactive protein (CRP, mg/dl, normal range: 0–0.5 mg/dl), level of serum rheumatoid factor (RF, IU/ml, determined by nephelometry (Siemens Healthcare Diagnostics, Munich, Germany), normal range <20 IU/ml), and level of anti-cyclic citrullinated peptide antibody (ACPA, measured by ELISA (Version 2.0, Aesku Diagnostics, Wendelsheim, Germany), normal range <18 U/ml). Disease activity was assessed with DAS28-ESR (the disease activity score in 28 joints based on four variables, including the ESR), DAS28-CRP (the disease activity score in 28 joints based on four variables, including CRP), the simplified disease activity index (SDAI), the clinical disease activity index (CDAI), and the Routine Assessment of Patient Index Data 3 (RAPID3) (Anderson et al., 2012). Disease activity defined by DAS28-ESR was divided into four grades: >5.1 (high disease activity (HDA)), ≥3.2 and <5.1 (moderate disease activity (MDA)), ≥ 2.6 and <3.2 (low disease activity (LDA)), and <2.6 (remission).

Radiographic assessments

Plain radiographs of the bilateral hands and wrists (anteroposterior view) of all included patients were performed to determine the radiographic status at baseline. Joint damage, including the Sharp/van der Heijde-modified total Sharp score (mTSS), joint erosion (JE), and joint space narrowing (JSN), was scored by two experienced observers (MJD from the Department of Rheumatology and YZH from the Department of Radiology) who were blinded to the patients’ clinical data as described previously (Van der Heijde, 2000). Reliability and agreement were assessed based on the intra-class correlation coefficient (ICC): the mean ICC for interobserver agreement was 0.90. Bony erosion was defined as a cortical break by radiography (Van der Heijde et al., 2013). Patients with an mTSS score >10 were defined as having radiographic joint damage (RJD) (Baker et al., 2011).

Chemiluminescent immunoassay for detection of serum TAbs and thyroid function

Blood was obtained from all the included RA patients after overnight fasting, and serum specimens were separated and stored at −80 °C for further analysis. The levels of TPOAbs (normal range: 0–60 IU/ml), TgAbs (normal range: 0–60 IU/ml), free thyroxine (FT4, normal range: 11.5–22.7 pmol/L), free triiodothyronine (FT3, normal range: 3.5–6.5 pmol/L), and thyroid-stimulating hormone (TSH, normal range: 0.55–4.78 pmol/L) were measured using chemiluminescent immunoassay (CLIA, Siemens Healthineers, Massachusetts, USA) according to the manufacturer’s instructions. RA patients with positive TPOAbs or TgAbs were assigned to the TAbs-positive group, and patients with neither positive TPOAbs nor positive TgAbs were assigned to the TAbs-negative group.

Synovial tissues and immunohistochemistry

Synovium was collected by closed Parker-Pearson needle biopsy from one actively inflamed knee joint in RA patient (Schumacher & Kulka, 1972). As described previously (Ma et al., 2014; Zou et al., 2013), serial sealed sections of qualified synovial samples were stained with hematoxylin and eosin (H&E) and also immunohistochemically stained with the commercial mouse monoclonal antibodies (Invitrogen, Carlsbad, CA, USA) according to standard staining protocols: anti-CD3 (clone PS1, T cells), anti-CD15 (clone My1, neutrophils), anti-CD20 (clone L26, B cells), anti-CD34 (clone QB End/10, vascular endothelial cells), anti-CD38 (clone SPC32, plasma cells), and anti-CD68 (clone KP1, macrophages). Irrelevant isotype was used as a negative control. Appropriate positive controls were included to rule out the possibility of an absence of staining due to technical failure.

Synovitis assessments

At least three tissue pieces that contained well-defined synovial lining and sublining areas were included in the analysis for each specimen. Histological changes in H&E-stained sections were graded according to Krenn’s synovitis score (Krenn et al., 2006). The three-component synovitis score included enlargement of the synovial lining layer, increased density of sublining resident cells, and inflammatory infiltration. Each feature was scored from 0–3, and the total synovitis score ranged from 0–9. RA patients were then divided into high-grade (>4) and low-grade (≤4) synovitis groups according to the synovitis score. Synovitis assessments were performed using a Leica DM2500 microscope (Leica Corp., Heidelberg, Germany) by two independent trained investigators (MYQ and LJZ, both from the Department of Rheumatology) who were blinded to the clinical details of the specimens. Differences between the two investigators were resolved by mutual agreement. The densities of cells with positive staining for CD3, CD15, CD20, CD38, and CD68 and the microvessel counts (MVCs, as confirmed based on the presence of CD34-positive endothelial cells and vessel lumen diameter ≤8 red blood cell diameter) were determined by manual counting. Because each high-power field revealed a synovial area of 0.11740 mm², the densities of cells with positive staining for CD3, CD15, CD20, CD38, CD68, and MVCs were assessed in nine different fields at a magnification of × 400 for a total area of nearly 1 mm²; the densities are given as cells per mm².

Statistical analysis

IBM SPSS Statistics version 20.0 software (IBM, Armonk, NY, USA) was used for the statistical analyses. For continuous variables, descriptive statistics (median, interquartile range (IQR)) were calculated, and the Mann–Whitney analysis of variance on ranks between two groups was used. For categorical variables, indicators are presented as frequencies and percentages, and the Chi-square test or Fisher’s exact test was used. Univariate and multivariate logistic regression analyses were performed, and odds ratios (ORs) and 95% confidence intervals (CIs) were calculated to identify risk factors for RJD. Variables were included in the equation when p < 0.05 or removed when p > 0.10 following the step-forward selection rule. Spearman’s rank order correlation test was used to assess the relationship between TAb levels and RA duration. A two-tailed p < 0.05 was considered statistically significant.

Demographic characteristics of RA patients

There were 141 RA patients who fulfilled the inclusion criteria of this study. One patient who was being treated with amiodarone, five patients who were simultaneously diagnosed with malignancies (one hepatocellular carcinoma, one esophageal carcinoma, one breast carcinoma, one nasopharyngeal carcinoma, and one osteosarcoma), and 10 patients who overlapping with other autoimmune diseases (four systemic lupus erythematosus, three scleroderma, two polyarteritis nodosa, one dermatomyositis) were excluded. Finally, a total of 125 RA patients were included in the statistical analysis, of whom 96 (77%) were women. The median disease duration was 60 months, ranging from 12–120 months. Among the included patients, 72% were RF positive, 71% were ACPA positive, and 93% of the patients had bony erosion at baseline. In total, 59 (47%) patients were glucocorticosteroids (GCs) or disease-modifying antirheumatic drugs (DMARDs) therapy naive since six months before enrollment (Table 1).

Thyroid abnormalities in RA patients

The profile of thyroid abnormalities is demonstrated in Table 2. Among the patients, 35 (28%) were positive for TPOAbs, and 27 (22%) were positive for TgAbs, with 44 (35%) patients having either positive TPOAbs or positive TgAbs and 17 (14%) being positive for both. TPOAbs positivity in women was 31%, and TgAbs positivity was 26%. Additionally, significantly higher prevalences of TPOAbs and TgAbs were found in patients with seropositive RF versus those with seronegative RF (36% vs. 9% and 27% vs. 9%, respectively; both p < 0.05). Five (4%) patients were diagnosed with hypothyroidism (two (2%) had clinical hypothyroidism, and three (2%) had subclinical hypothyroidism), and five (4%) patients had hyperthyroidism (three (2%) showed clinical manifestations, and two (2%) showed subclinical hyperthyroidism). All patients with thyroid disorders were in the TAbs-positive group except for 2 patients (1 subclinical hyperthyroidism and 1 subclinical hypothyroidism) in the negative group. Accordingly, FT3, FT4, and TSH abnormalities were all significantly more frequent in the TAbs-positive group than in the TAbs-negative group (21% vs. 6%, 14% vs. 3%, and 18% vs. 3%, respectively; all p < 0.05; Table 2).

Comparison of RA characteristics between patients with and without positive TAbs

There were 44 (35%) patients in the TAbs-positive group. A significantly greater percentage of patients with RJD was observed in the TAbs-positive group versus the TAbs-negative group (68% vs. 42%, p = 0.005; Table 1). Compared with the TAbs-negative group, patients with positive TAbs had RA onset at a significantly younger age (42(35–49) years vs. 47(41–55) years, p = 0.025). RF positivity and disease activity indicators, including TJC28, DAS28-ESR, and CDAI, were significantly higher in the TAbs-positive group (all p < 0.05), with borderline significant differences in SJC28, PrGA, DAS28-CRP, and SDAI as well as in the percentage of functional limitation. Similarly, a significantly higher percentage of patients with HDA was seen in the TAbs-positive group versus the TAbs-negative group (68% vs. 49%, p = 0.043). However, there was no significant difference in other RA clinical features between the two groups, including gender, disease duration, smoking status, and previous medications used since 6 months before enrollment (all p > 0.05; Table 1).

Comparison of synovitis between patients with and without positive TAbs

A total of 22 patients had qualified synovial tissues, of whom eight were patients with positive TAbs. Notably, RF positivity and ACPA positivity were both 100% in the eight patients, and were 71% and 79% respectively in the 14 patients with negative TAbs. Significantly more pronounced infiltration of CD38-positive plasma cells was observed in the TAbs-positive synovium (1,354(847–2,096) cells/mm²) than in the TAbs-negative control (274(109–1,252) cells/mm²) (p = 0.048; Table 3). The percentage of patients with high-grade synovitis was also higher in the TAbs-positive group than in the TAbs-negative group (63% vs. 36%), but the difference was not significant (p = 0.221). Representative images of H&E and immunohistochemical staining for CD38 in the synovium of RA patients with and without positive TAbs are shown in Fig. 1. No significant difference between the two groups was observed in the densities of other infiltrated inflammatory cells, including CD3-positive T cells, CD15-positive neutrophils, CD20-positive B cells, CD68-positive macrophages, or MVCs (all p > 0.05; Table 3).

Risk factors for RJD in RA patients

Logistic regression analysis was performed to investigate the risk factors for RJD (Fig. 2). Univariate logistic regression analysis performed on demographic and disease characteristics as well as TAbs status showed that medication-naive status was a protective factor (OR 0.446, 95% CI [0.218–0.913]; p = 0.027), whereas a longer disease duration (OR 1.013, 95% CI [1.006–1.019]; p < 0.001), a higher HAQ score (OR 1.969, 95% CI [1.125–3.448]; p = 0.018), positive RF (OR 2.623, 95% CI [1.162–5.918]; p = 0.020), and positive TAbs (OR 2.962, 95% CI [1.368–6.415]; p = 0.006) were risk factors for RJD. Multivariate logistic regression analysis was performed to control for confounding factors. The result revealed that positive TAbs (OR 2.999, 95% CI [1.301–6.913]; p = 0.010) and disease duration (OR 1.013, 95% CI [1.006–1.019]; p < 0.001) were independently associated with RJD. No significant correlations were found neither between TPOAb levels and disease duration (r = 0.082, p = 0.366) nor TgAb levels and disease duration (r =  − 0.057, p = 0.530).

Further analysis of positive TAbs yielded an OR of 2.845 (95% CI [1.062–7.622]) for RJD when comparing women in the TAbs-positive group (n = 37) with those in the TAbs-negative group (n = 59) (p = 0.038). Subgroup studies revealed that a high titer of ACPA (defined as an ACPA level ≥ 3 times the upper limit of the normal range (ULN)) was also a risk factor for RJD (OR 2.107, 95% CI [1.025–4.328]; p = 0.043). After adjusting for confounding factors, as mentioned above for the univariate logistic regression analysis, a high titer of ACPA remained significantly associated with RJD (OR 2.467, 95% CI [1.076–5.655]; p = 0.033).