Infiltrations of plasma cells in synovium predict inadequate response to Adalimumab in Rheumatoid Arthritis patients

Abstract

Objective

Rheumatoid arthritis (RA) is a clinically heterogeneous and complex autoimmune disease, making the prediction of therapeutic responses a significant challenge. This study aims to assess the role of clinical and synovial biomarkers in predicting poor response to adalimumab treatment in RA patients.

Methods

This single-center prospective study included 56 RA patients who had an inadequate response to methotrexate (MTX). At baseline, comprehensive assessments including complete blood count, liver and kidney function tests, blood glucose levels, erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), rheumatoid factor (RF), anti-citrullinated protein antibody (ACPA), as well as counts of swollen and tender joints, Health Assessment Questionnaire (HAQ) score, pain visual analogue scale (VAS) scores, and DAS28-CRP scores were conducted. Synovial biopsies were performed, followed by an efficacy evaluation at 12 weeks of adalimumab treatment. Patients not meeting the ACR20 criteria were classified into the non-responder group, with the remainder categorized as the responder group.

Results

Out of the participants, 24 (42.9%) failed to achieve ACR20 with adalimumab treatment. Non-responders exhibited higher infiltration of plasma cells in the synovium. Multivariate logistic regression analysis identified the presence of plasma cells as an independent risk factor for inadequate response to adalimumab.

Conclusion

Inadequate responses to adalimumab in RA patients were associated with increased plasma cell infiltrations in the synovium. These findings suggest a promising target for tailored therapies in rheumatoid arthritis.

Introduction

Rheumatoid arthritis (RA) is a chronic autoimmune disorder characterized by persistent inflammation in the synovial lining of joints, which typically results in pain, stiffness, and joint damage. Globally, RA affects approximately 1% of the population, disproportionately impacting women, and often leads to both local joint deformities and systemic complications [1]. The management of RA has been revolutionized by the introduction of biologic treatments, which have shown substantial efficacy in reducing inflammation and slowing joint damage in patients who have either not responded to or cannot tolerate traditional Disease-Modifying Anti-Rheumatic Drugs (DMARDs) such as methotrexate [2].

Adalimumab, a tumor necrosis factor-alpha (TNF-α) inhibitor, has emerged as a cornerstone in the biologic treatment of RA. It is one of the most frequently prescribed biologic medications globally [3]. Clinical trials and real-world studies have consistently demonstrated that adalimumab, whether used alone or in combination with traditional DMARDs like methotrexate, can significantly reduce disease activity and improve patient outcomes [4, 5]. However, it is also noted that approximately 30–40% of RA patients may exhibit an inadequate response to adalimumab or other TNF-α inhibitors, due to factors such as individual variability, immunogenicity, disease severity, and presence of comorbidities [6].

Predicting the response to biologic treatments remains a complex and dynamic area of research. No single predictor or methodology can ensure accurate forecasts for all cases. A personalized medicine approach, combining multiple factors, is often essential to tailor treatments to individual needs and optimize outcomes with biological therapies. Despite significant advances and ongoing research into potential biomarkers, precise prediction of treatment responses in RA remains challenging.

In this context, we conducted a prospective study involving 56 patients with active RA who were eligible for adalimumab treatment. Our study focuses on examining synovial characteristics and immunoinflammatory biomarkers as predictors of the response to adalimumab, aiming to contribute to the refinement of treatment strategies in RA.

Patients and methods

This was a single-center, prospective, observational study that recruited patients diagnosed with rheumatoid arthritis (RA) according to the 2010 American College of Rheumatology/European League Against Rheumatism (ACR/EULAR) criteria [7]. All participants in this study were of Asian descent, and were 18 years of age or older at the time of RA diagnosis. We excluded pregnant or lactating women, individuals with active infections (e.g., hepatitis, pneumonia, pyelonephritis, or chronic skin infections), and those with a history of bleeding disorders.

This study was approved by the Ethical Review Board of the First Affiliated Hospital of Nanchang University, with the ethics approval number IIT[2023] Clinical Ethics Review No. 011. All participants provided informed consent. At baseline, we collected comprehensive clinical data, including the baseline Disease Activity Score (DAS28), pain assessments using the Visual Analog Scale (VAS), tender joint count (TJC), swollen joint count (SJC), Health Assessment Questionnaire (HAQ) scores, baseline age (i.e., the participant’s age at baseline), and disease duration (i.e., the time from the initial diagnosis to baseline). Additionally, we collected data on concomitant medications, primarily glucocorticoids, methotrexate, and hydroxychloroquine. Laboratory tests were also performed to assess inflammatory markers, such as C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR), as well as autoantibody levels, including rheumatoid factor (RF) and anti-citrullinated protein antibody (ACPA).

Following the initial assessments, patients underwent a synovial biopsy and commenced treatment with adalimumab according to standard clinical guidelines. Throughout the adalimumab treatment period, patients were regularly monitored. Follow-up assessments comprised clinical evaluations, imaging studies (if applicable), and repeated laboratory tests to assess disease activity, evaluate treatment responses, and monitor for potential adverse events.

Treatment response was evaluated based on predefined criteria, which included the American College of Rheumatology (ACR) response criteria and the Disease Activity Score (DAS). According to the ACR20 criteria, patients were required to have at least a 20% reduction in either the Tender Joint Count (TJC) or Swollen Joint Count (SJC), and at least a 20% improvement in three of the following criteria: the patient’s pain score (assessed by the Visual Analog Scale, VAS), the Health Assessment Questionnaire (HAQ) score, erythrocyte sedimentation rate (ESR), or C-reactive protein (CRP). Patients were then categorized as ‘responders’ or ‘non-responders’ based on whether they met the ACR20 response criteria.

ACPA and RF detection methods

Fasting blood samples (5 mL) were collected from study participants in the morning, centrifuged at 3000 rpm for 10 min, after which the serum was separated and stored at -80 °C for future analysis. Serum levels of ACPA were measured using an automated chemiluminescence immunoassay (CIA), targeting the CCP2 antigen [13]. The assay kit was provided by Yuhuilong Biotechnology Co., Ltd., and the YQMY218 fully automated chemiluminescence analyzer from the same company was used, with the ACPA threshold set at 5 U/mL. RF was detected using the rate scattering turbidimetry method. The assay kit was also sourced from Yuhuilong Biotechnology Co., Ltd., and the analysis was conducted using the Beckman Coulter IMMAGE 800 specific protein analysis system. The criteria for determining positivity were ACPA levels > 5 U/mL and RF levels > 20 U/mL. All procedures were performed strictly according to the instructions provided with the assay kits and equipment manuals.

Synovial biopsy and tissue processing

Prior to the initiation of adalimumab therapy, synovial biopsies were performed on the enrolled patients to assess synovial tissue pathology. Synovial tissue samples were obtained from affected joints using minimally invasive procedures with an improved Parker Pearson biopsy needle(as shown in the Fig. 1). Each sampling procedure at least six fragmented tissue pieces, each measuring approximately 2 × 3 mm(as shown in the Fig. 2). Immediately following collection, the tissue samples were immersed in a formaldehyde solution and stored at 4 °C for subsequent histological analysis.

Hematoxylin and eosin (H&E) staining was employed to examine the tissue for signs of inflammation and structural changes. Histopathological assessments were conducted, focusing on synovial lining layer hyperplasia, activation of resident stromal cells, and the extent of inflammatory infiltration. Each histological feature was semi-quantitatively scored on a scale from 0 to 3.

Immunohistochemistry was used to detect immune cells expressing specific lineage markers, including CD3, CD20, CD68, and CD138. All antibodies were purchased from Beijing Zhongshan Golden Bridge Biotechnology Co., Ltd.Paraffin sections were deparaffinized, dehydrated, subjected to microwave heating, and underwent antigen retrieval in 10 mM citrate buffer (pH 6.0) for 15 min. Endogenous peroxidase activity was quenched with 3% H2O2. Sections were then incubated overnight at 4 °C with primary antibodies, followed by incubation with corresponding secondary antibodies. Slides were developed using 3,3’-diaminobenzidine, counterstained with hematoxylin, and finally mounted in a non-aqueous mounting medium. Negative control slides (without primary antibody incubation) were included in each staining run.

Two senior pathologists, blinded to the relevant clinical data, independently scored the intimal and subintimal layers of the synovium. A detailed microscopic examination of the subsynovial layer was performed using a 20x high-power microscope. A semi-quantitative four-point scoring system was used to assess the density of lymphocytes, macrophages, and plasma cells in the tissue, with the following scale: 0 = 1–9 cells (no infiltration), 1 = 10–99 cells (mild infiltration), 2 = 100–999 cells (moderate infiltration), and 3 = more than 1000 cells (severe infiltration). In this study, a plasma cell score of 0 (1–9 cells) was defined as “plasma-poor,” and a score of 1 or higher (10 or more cells) was defined as “plasma-rich.” Discrepancies between observers were resolved through joint review, reaching consensus in all cases.

Data analysis

Statistical analysis was performed using SPSS 26.0 and GraphPad Prism 9 software. The Kolmogorov-Smirnov test was used to assess the normality of the data. For normally distributed continuous variables, the mean and standard deviation (SD) were reported, while for non-normally distributed continuous variables, the median and interquartile range (IQR) were used. Group differences for normally distributed data were assessed using unpaired two-tailed t-tests, whereas non-normally distributed data were analyzed using the Mann-Whitney U test. Categorical data were presented as the number and percentage of participants, and group differences in categorical variables were analyzed using Fisher’s exact test or the chi-square test. Univariate logistic regression analysis was performed to identify potential risk factors associated with an inadequate response to adalimumab treatment. Multivariate logistic regression analysis was subsequently conducted to evaluate the independent effects of these factors, adjusting for age, disease duration, and DAS28. A P-value of less than 0.05 was considered statistically significant. Data visualization and figure generation were conducted using GraphPad Prism 9 to illustrate group differences and correlation results.

Results

Baseline characteristics of all patients with different ACR response

Between Jan. 1, 2023, and Nov 30, 2023, 56 active RA patients were enrolled. All patients had completed 12 weeks of adalimumab continuous treatment. 24(42.9%) failed ACR 20 response. Baseline characteristics, disease activity, and clinical data are reported in Table 1.

Table 1 Demographic and baseline disease characteristics of RA patients in two groups

Synovial characteristics in RA patients with different response to adalimumab

No significant differences were observed between the two groups regarding synovial hyperplasia, stromal activation, neovascularization, and inflammatory infiltration. However, compared to the responder group, the non-responder group exhibited significantly higher infiltration of plasma cells (P < 0.05) (see Fig. 3). Immunohistochemical staining revealed no significant differences between the responder and non-responder groups in the presence of CD3 + T-cells, CD20 + B-cells, and CD68 + macrophages. However, a significant difference was observed in the presence of CD138 + plasma cells between the two groups (See Fig. 4).

Fig. 3

Comparative Analysis of Synovial Features and Immune Responses in RA Patients.Data for CD3 + T-cells and CD20 + B-cells were presented as part of the lymphocytes, as no statistically significant differences were observed between the two groups

Fig. 4

Images of Synovial Histology for the Responder and Non-Responder Groups: HE staining shows significant plasma cell infiltration in the synovium of the non-responder group. Immunohistochemical staining shows no significant differences between the responder and non-responder groups in CD3, CD20, and CD68, but there is a significant difference in CD138 + positive cells

Clinical and synovial characteristics in RA patients with synovial plasma cells

Although there were no significant differences observed in clinical parameters between the two groups of adalimumab responders and non-responders regarding stromal activation and inflammatory infiltration, after dividing the plasma cells into two groups, the plasma cell rich group and the plasma cell poor group, the plasma cell rich group showed significant stromal activation and inflammatory cell infiltration compared to the plasma cell poor group, especially macrophages (see Fig. 5) (P < 0.05).

Fig. 5

Comparison of Stromal Activation, Inflammatory Cell Infiltration, and Macrophage Presence Between Plasma Cell-Rich and Plasma Cell-Poor Groups in Synovial Tissue.Data for CD3 + T-cells and CD20 + B-cells were presented as part of the lymphocytes, as no statistically significant differences were observed between the two groups

Evaluation of clinical and synovial characteristics as predictors of Adalimumab response in RA

Univariate logistic regression analysis revealed that ESR, ACPA, VAS, synovial lining hyperplasia, lymphocytes, and plasma cells were significantly associated with an inadequate response to adalimumab treatment. After adjusting for baseline age, disease duration (i.e., time from diagnosis to baseline), and baseline DAS28 in the univariate analysis, plasma cells, ACPA, and VAS remained associated with a poor treatment response. Finally, when plasma cells, ACPA, and VAS were simultaneously included in the multivariate logistic regression analysis, and further adjustments were made for baseline age, disease duration, DAS28, and ESR, the presence of plasma cells was identified as an independent risk factor for an inadequate response to adalimumab (see Table 2).

Table 2 Logistic regression analysis of factors influencing response to Adalimumab

Discussion

Tumor necrosis factor α (TNFα), a pro-inflammatory cytokine, is central to the pathogenesis of RA. It is predominantly produced by macrophages and T cells and plays a critical role in driving the inflammatory and immune responses that characterize RA. These include promoting the release of other inflammatory cytokines, enhancing leukocyte migration, and inducing the expression of adhesion molecules in the synovium. Such activities contribute significantly to the sustained inflammation and pannus formation observed in affected joints [8].

Adalimumab, a TNFα inhibitor, is designed to target and neutralize TNFα, thereby reducing its pathological impacts in RA. Clinical trials have consistently demonstrated adalimumab’s significant efficacy in reducing symptoms, improving physical function, and slowing the progression of joint damage. Key trials have highlighted that treatment with adalimumab leads to improved response rates, with up to 75% of patients achieving an ACR20 response, while about 58% and 38% reach ACR50 and ACR70 response rates, respectively [9]. These results underscore the therapeutic potential of directly inhibiting TNFα in managing RA.

Despite the significant roles of TNFα in RA, the pathogenesis of the disease involves a complex network of cellular pathways and a variety of cytokines, extending beyond the influence of a single cytokine. Consequently, the clinical efficacy of TNFα inhibitors(TNFi) like adalimumab is limited and often characterized by a ‘ceiling effect.’ Approximately 30–40% of patients do not achieve a sufficient therapeutic response [10]. This limitation highlights the complexity of RA and suggests that TNFα, although critical, may not be the central pathological factor in certain RA subtypes. Our study findings indicate that nearly 50% of patients with RA did not meet the ACR20 criteria after 12 weeks of adalimumab combination therapy. This observation is consistent with the known heterogeneity in treatment response among RA patients and suggests that factors beyond TNFα may play a crucial role in the pathogenesis of certain subtypes of RA.

The quest to predict the efficacy of TNFα inhibitors is paramount in optimizing treatment strategies for RA. Identifying reliable predictors of response can significantly enhance patient outcomes by tailoring treatments to those most likely to benefit, thereby minimizing exposure to ineffective therapies and their potential side effects. This discourse is especially relevant given the substantial variability in patient responses to TNFi.

Numerous factors have been linked to poor responses to TNFi, including smoking, older age, high body mass index (BMI), the presence of RF and ACPA, extended disease duration, and elevated disease activity [11]. Research has consistently shown that patients with higher baseline disease activity scores are more likely to experience significant reductions in disease activity when treated with TNFi [12]. This observation underscores the paradox that patients with more severe disease manifestations may derive more pronounced benefits from TNF inhibition. Aligning with these findings, our study further corroborates that responders to adalimumab typically exhibit longer disease duration, higher ESR, and elevated DAS28 scores.

However, our multivariable regression analysis did not identify a significant correlation between factors such as age, disease duration, baseline DAS28, ACPA, and ESR, and poor response to adalimumab. While these markers may reflect clinical characteristics associated with poor response, they are not necessarily direct causes. Rather, they may be more indicative of disease severity than independent determinants of treatment outcomes.Several factors could explain these discrepancies. One possible reason is the limitation in our sample size, which may have reduced the statistical power to detect certain associations. Additionally, we used ACR20 as the primary outcome measure, while some other studies may have employed the EULAR criteria. Differences in these outcome measures may contribute to variations in the predictive power of these markers. Therefore, relying solely on markers like ESR to predict treatment response may not be sufficient. In clinical practice, a more comprehensive evaluation incorporating additional variables and indicators is necessary for more accurate prediction and management of treatment responses.Future research should focus on identifying new biomarkers or underlying mechanisms to better understand the true causes of poor TNFi response, thereby providing stronger support for personalized treatment approaches.Notably, in our non-responder group, the levels of RF and ACPA were significantly lower than expected, contrasting with conventional expectations where higher levels might predict poorer outcomes. However, some studies have shown that B cells and plasma cells in synovial tissue are not associated with the status of autoantibodies in serum [13].Although our logistic regression analysis did not establish a direct correlation between these clinical indicators and treatment response, it suggests that high titers or positivity for RF and ACPA are not reliable predictors of poor response to TNFi. The conflicting nature of these results underscores the complexity of predicting treatment responses in RA and highlights the necessity for the development of more reliable biomarkers and additional large-scale studies.

Recent studies underscore the potential of synovial biomarkers in predicting therapeutic responses to various biologic agents in RA. Molecular analyses of gene expression profiles in synovial tissues from RA patients have identified four distinct histopathological phenotypes: lymphoid-like, myeloid-like, low-inflammatory, and fibroid-like [14]. Notably, individuals responding to anti-TNF therapy typically exhibit a myeloid-like phenotype, closely associated with the circulating biomarker ICAM1. In contrast, the lymphoid-like phenotype, which correlates with the circulating biomarker CXCL13, shows a more favorable response to anti-IL6 treatments [15]. A small sample study has also confirmed that adalimumab treatment can effectively reduce synovial CD68 + cell infiltration, while not affecting synovial CD20 + B-cells [16]. Additionally, research has shown that in RA patients who do not respond well to TNFi therapy and have CD20 + positive B cell-rich synovial tissue, both tocilizumab and rituximab have comparable efficacy, However, for patients with B cell-poor synovial tissue, tocilizumab may be more effective than rituximab [17]. Our study did not find a significant relationship between synovial CD20 + positive B-cells, CD68 + cells, and the efficacy of adalimumab. This discrepancy may be due to the differences in our study methods, which employed analyses based on H&E staining and immunohistochemical pathology. Currently, there are two main approaches for predicting non-response to anti-TNFα therapy with synovial biopsy: one based on histopathological morphology and the other on RNA sequencing. By homogenizing and pooling multiple biopsy samples, RNA sequencing can provide a more comprehensive measurement of pathobiological processes compared to synovial histopathological analysis [17]. However, histopathology has advantages in directly visualizing cell morphology and localization, with simpler procedures, lower costs, and more specific results relative to sequencing. In our study, we analyzed at least six tissue samples and six different sections of synovial tissue per sampling session, focusing on cell counting in the most severely affected areas. This approach aimed to reduce sampling errors and improve data comparability.

We observed that in the non-responder group, there was significantly greater plasma cell infiltration in the synovium, as identified by both H&E and CD138 staining. Our analyses have revealed that plasma cell infiltration in the synovium is a significant indicator of poor response to adalimumab, independent of B cell presence, with an odds ratio of 6.612 (P = 0.037)—a novel finding to our knowledge. These findings suggest that simple H&E staining can also effectively identify synovial predictors for adalimumab efficacy.

Plasma cells are indispensable components of the adaptive immune response, differentiating from B cells that exhibit high specificity against antigens [18]. Predominantly originating in the bone marrow, plasma cells can also be found in various tissues. Their primary role is to produce antibodies that bind specifically to antigens, facilitating the recognition and elimination of pathogens marked by these antibodies. This is vital for maintaining immune memory and long-term immune protection.

In the context of RA, plasma cells play a pivotal role through their capacity to produce antibodies, including RF and ACPA. These antibodies are crucial in RA pathogenesis. They contribute to the formation of immune complexes that incite inflammatory processes within synovial joints [19].

While previous research has posited a positive correlation between synovial plasma cells and Anti-CCP Antibody Positivity based on gene expression analyses [20], our findings do not support this link(see Supplementary Table 1). Interestingly, we observed that non-responders often present lower levels of RF and ACPA despite a higher prevalence of synovial plasma cells. This discrepancy suggests complex roles for plasma cells within the synovial environment.

Indeed, our study further reveals that the presence of synovial plasma cells typically accompanies enhanced stromal activation, robust infiltration of inflammatory cells, and an increased number of macrophages. Thus, synovial plasma cells are implicated not only in antibody secretion but also as significant contributors to synovial inflammation. We hypothesize that some of these cells may be special plasma cells with non-secreting specific antibody, potentially crucial in mediating the lack of response to TNF inhibitors in RA. A more detailed comparative analysis of the expression and functionality of surface markers on both synovial and serum plasma cells could provide deeper insights into this complex interplay.

The infiltration of plasma cells in the synovial membrane is a hallmark of RA. These cells are found in higher numbers in the synovial tissue of RA patients compared to those with osteoarthritis or healthy controls, suggesting a specific role in RA pathophysiology [21]. Plasma cells in the synovium are not only sources of autoantibodies but also produce cytokines and growth factors that can perpetuate inflammation and tissue damage. Recent literature underscores the significance of synovial plasma cell infiltration in RA. Studies have shown that the degree of plasma cell presence correlates with disease severity and therapeutic response [20, 22].

In conclusion, our study reveals that synovial plasma cells do not show a direct correlation with disease activity markers such as DAS28-CRP, or systemic inflammation indicators (ESR, CRP), nor with serum antibodies (RF and ACPA). However, they are significantly associated with localized synovial inflammation and demonstrate potential as effective biomarkers for predicting poor responses to TNF inhibitors. Continued research into the pathobiological mechanisms and interactions with other cells in the synovium is crucial. Such studies will undoubtedly deepen our understanding and facilitate the development of more precise and effective treatment strategies for rheumatoid arthritis. Due to the limited number of cases included in this study, the findings must be interpreted with caution. Future research with a larger cohort and refined methodological approaches is necessary to validate and expand upon these preliminary findings.

Fig. 1

The structure of the new synovial biopsy device and Procedure of the Novel Synovial Biopsy Device for Joint Tissue Sampling. The novel synovial biopsy device for joint tissue sampling consists of the following key components: a Sleeve Needle (1), Sleeve Needle Core (2), Sampling Needle (3), Sampling Needle Core (4), Rebound Device (5), Slot (6), Sampling Port (7), and a Negative Pressure Connector Port (8). The device is designed to puncture into the joint cavity using the puncture sheath needle and the puncture inner core (a). After removing the puncture inner core, the sampling needle is inserted, and the firing device is pressed, placing the sampling needle in a ready-to-fire state (b). The tail of the biopsy needle tube is connected to a negative pressure device, which creates a vacuum inside the closed-end hollow section of the biopsy needle tube, ensuring that synovial tissue enters the groove of the sampling needle under the effect of negative pressure (c). Coordinating with the firing device of the biopsy outer sheath, the biopsy needle tube rapidly springs out, and the hooked end at the head of the biopsy needle tube cuts the synovial tissue (d)

Fig. 2

Microscopic features of the RA synovial membrane. (Original magnification x 4)