Clinical predictors of response to methotrexate in patients with rheumatoid arthritis: a machine learning approach using clinical trial data

The fibroid phenotype of biological naïve patients with rheumatoid arthritis are less likely to respond to anti-IL-6R treatment

Type III collagen gene expression is upregulated in the synovium of patients with rheumatoid arthritis (RA) presenting the fibroid phenotype. The soluble type III collagen formation biomarker, PRO-C3, is known to measure fibrogenesis in fibrotic diseases. In this exploratory study, we aimed to investigate the association between fibrogenesis (PRO-C3) and the disease- and treatment response in patients with RA. We measured PRO-C3 in subsets of two clinical trials assessing the effect of the anti-interleukin-6 (IL-6) receptor treatment tocilizumab (TCZ) as monotherapy or polytherapy with methotrexate. PRO-C3 levels had weak or very weak correlations with the clinical parameters (Spearman's). However, when the patients were divided into Disease Activity Score-28 groups characterized by the erythrocyte sedimentation rate (DAS28-ESR), there was a statistical difference between the PRO-C3 levels of the different groups (p < 0.05). To determine the response in relation to PRO-C3, a cut-off based on PRO-C3 levels and patients in remission (DAS28-ESR ≤ 2.6) was identified. This showed that a reduction in PRO-C3 after treatment initiation was associated with decreased DAS28-ESR and a higher response rate in patients with low PRO-C3 levels than in those with high PRO-C3 levels. This indicates that a fibrotic component affects the responsiveness of patients.

Prediction of ineffectiveness of biological drugs using machine learning and explainable AI methods: data from the Austrian Biological Registry BioReg

OBJECTIVES

Machine learning models can support an individualized approach in the choice of bDMARDs. We developed prediction models for 5 different bDMARDs using machine learning methods based on patient data derived from the Austrian Biologics Registry (BioReg).

METHODS

Data from 1397 patients and 19 variables with at least 100 treat-to-target (t2t) courses per drug were derived from the BioReg biologics registry. Different machine learning algorithms were trained to predict the risk of ineffectiveness for each bDMARD within the first 26 weeks. Cross-validation and hyperparameter optimization were applied to generate the best models. Model quality was assessed by area under the receiver operating characteristic (AUROC). Using explainable AI (XAI), risk-reducing and risk-increasing factors were extracted.

RESULTS

The best models per drug achieved an AUROC score of the following: abatacept, 0.66 (95% CI, 0.54-0.78); adalimumab, 0.70 (95% CI, 0.68-0.74); certolizumab, 0.84 (95% CI, 0.79-0.89); etanercept, 0.68 (95% CI, 0.55-0.87); tocilizumab, 0.72 (95% CI, 0.69-0.77). The most risk-increasing variables were visual analytic scores (VAS) for abatacept and etanercept and co-therapy with glucocorticoids for adalimumab. Dosage was the most important variable for certolizumab and associated with a lower risk of non-response. Some variables, such as gender and rheumatoid factor (RF), showed opposite impacts depending on the bDMARD.

CONCLUSION

Ineffectiveness of biological drugs could be predicted with promising accuracy. Interestingly, individual parameters were found to be associated with drug responses in different directions, indicating highly complex interactions. Machine learning can be of help in the decision-process by disentangling these relations.

Machine learning application in autoimmune diseases: State of art and future prospectives

Autoimmune diseases are a group of disorders resulting from an alteration of immune tolerance, characterized by the formation of autoantibodies and the consequent development of heterogeneous clinical manifestations. Diagnosing autoimmune diseases is often complicated, and the available prognostic tools are limited. Machine learning allows us to analyze large amounts of data and carry out complex calculations quickly and with minimal effort. In this work, we examine the literature focusing on the use of machine learning in the field of the main systemic (systemic lupus erythematosus and rheumatoid arthritis) and organ-specific autoimmune diseases (type 1 diabetes mellitus, autoimmune thyroid, gastrointestinal, and skin diseases). From our analysis, interesting applications of machine learning emerged for developing algorithms useful in the early diagnosis of disease or prognostic models (risk of complications, therapeutic response). Subsequent studies and the creation of increasingly rich databases to be supplied to the algorithms will eventually guide the clinician in the diagnosis, allowing intervention when the pathology is still in an early stage and immediately directing towards a correct therapeutic approach.

Efficient delivery of the lncRNA LEF1-AS1 through the antibody LAIR-1 (CD305)-modified Zn-Adenine targets articular inflammation to enhance the treatment of rheumatoid arthritis

BACKGROUNDS

Rheumatoid arthritis (RA) is a chronic inflammatory autoimmune disease characterized by synovial hyperplasia. Maintaining a balance between the proliferation and apoptosis of rheumatoid arthritis synovial fibroblasts (RASFs) is crucial for preventing the erosion of bone and cartilage and, ultimately, mitigating the progression of RA. We found that the lncRNA LEF1-AS1 was expressed at low levels in the RASFs and inhibited their abnormal proliferation by targeting PIK3R2 protein and regulating the PI3K/AKT signal pathway through its interaction with miR-30-5p. In this study, we fabricated a nano-drug delivery system for LEF1-AS1 using Zn-Adenine nanoparticles (NPs) as a novel therapeutic strategy against RA.

METHODS

The expression levels of LEF1-AS1, miR-30-5p, PIK3R2, p-PI3K, and p-AKT were detected in the primary RASFs and a human fibroblast-like synovial cell line (HFLS). Zn-Adenine nanoparticles (NPs) were functionalized with anti-CD305 antibody to construct (Zn-Adenine)@Ab. These NPs were then loaded with LEF1-AS1 to form (Zn-Adenine)@Ab@lncRNA LEF1-AS1. Finally, the (Zn-Adenine)@Ab@lncRNA LEF1-AS1 NPs were locally injected into a rat model with collagen-induced arthritis (CIA). The arthritic injuries in each group were evaluated by HE staining and other methods.

RESULTS

LEF1-AS1 was expressed at low levels in the primary RASFs. High expression levels of LEF1-AS1 were detected in the HFLS cells, which corresponded to a significant downregulation of miR-30-5p. In addition, the expression level of PIK3R2 was significantly increased, and that of p-PI3K and p-AKT were significantly downregulated in these cells. The (Zn-Adenine)@Ab@lncRNA LEF1-AS1 NPs significantly inhibited the proliferation of RASFs and decreased the production of inflammatory cytokines (IL-1β, IL-6, TNF-α). Intra-articular injection (IAI) of (Zn-Adenine)@Ab@lncRNA LEF1-AS1 NPs significantly alleviated cartilage destruction and joint injury in the CIA-modeled rats.

CONCLUSIONS

LEF1-AS1 interacts with miR-30-5p to inhibit the abnormal proliferation of RASFs by regulating the PI3K/AKT signal pathway. The (Zn-Adenine)@Ab NPs achieved targeted delivery of the loaded LEF1-AS1 into the RASFs, which improved the cellular internalization rate and therapeutic effects. Thus, LEF1-AS1 is a potential target for the treatment of RA.

Autoantibody-negative rheumatoid arthritis: still a challenge for the rheumatologist

Increased research over the past 30 years has greatly improved the understanding of the pathophysiological mechanisms and clinical aspects of autoantibody-positive rheumatoid arthritis, resulting in improved management and outcomes. In contrast, the subset of rheumatoid arthritis that does not have autoantibodies (such as rheumatoid factor and anti-citrullinated protein autoantibodies) remains less well defined in its pathogenic mechanisms. Autoantibody-negative rheumatoid arthritis continues to pose diagnostic challenges, might respond differently to therapies, and appears to be burdened with different comorbidities and outcomes. The clear separation of rheumatoid arthritis according to serotypes is still a subject of uncertainty and controversy, and studies specifically focused on comparing rheumatoid arthritis and rheumatoid arthritis-like arthritides that do not have autoantibodies remain scarce. The purpose of this Review is to summarise the peculiarities that make autoantibody-negative rheumatoid arthritis different from its autoantibody-positive counterpart, with the aim of generating debate and stimulating further research on this challenging condition.

Functional implications of rs9373441 with FOXP3+Treg and Tr1 for the clinical effectiveness of csDMARDs in rheumatoid arthritis

Rheumatoid arthritis (RA) is characterized by a deficiency in regulatory T cells (Treg), which play a crucial role in immune regulation. While conventional synthetic disease-modifying antirheumatic drugs (csDMARDs) are widely used, there remains a challenge as efficacy varies among patients. In this genome-wide association study (GWAS) involving 410 RA patients, rs9373441 emerged as the most significantly linked single-nucleotide polymorphism (SNP) to csDMARDs response. This non-coding variant functions as a cis-acting regulatory element within the UTRN gene, which is associated with cortical erosion and osteoporosis. Particularly, individuals with the TT allele at rs9373441 exhibited a more favorable response, characterized by a significant increase in FOXP3 + Treg and Type 1 regulatory T cells (Tr1) (p = 0.04, 0.02) and a decrease in Effector T helper cells (Effector Th) (p = 0.03). The GATA3-GCM2-PTH and GATA3-FOXO1-FOXP3 pathways were implicated. RNA-sequencing (RNA-seq) analysis revealed increased expression levels of UTRN, PTH2R, FOXO1, and FOXO3 in good and moderate responders (p = 0.01, 0.03, 0.0005, and 0.02). Notably, the change in FOXP3 + Treg and Tr1 was positively correlated with UTRN expression (both p = 0.03). These findings underscore the critical link between rs9373441 and the response to csDMARDs, empowering clinicians to tailor treatments for enhanced outcomes in patients with RA.

Rheumatoid Arthritis Has Won the Battle but Not the War: How Many Joints Will We Save Tomorrow?

Rheumatoid arthritis refers to joint diseases of unclear etiology whose final stages can lead to unbearable pain and complete immobility of the affected joints. As one of the most widely known diseases of the joints, it serves as a study target for a large number of research groups and pharmaceutical companies. Modern treatment with anti-inflammatory drugs, including janus kinase (JAK) inhibitors, monoclonal antibodies, and botanicals (polyphenols, glycosides, alkaloids, etc.) has achieved some success and hope for improving the course of the disease. However, existing drugs against RA have a number of side effects which push researchers to elaborate on more selective and effective drug candidates. The avant-garde of research, which aims to develop treatment of rheumatoid arthritis using antisense oligonucleotides along with nonsteroidal drugs and corticosteroids against inflammation, increases the chances of success and expands the arsenal of drugs. The primary goal in the treatment of this disease is to find therapies that allow patients with rheumatoid arthritis to move their joints without pain. The main purpose of this review is to show the victories and challenges for the treatment of rheumatoid arthritis and the tortuous but promising path of research that aims to help patients experience the joy of freely moving joints without pain.

Machine learning identifies a profile of inadequate responder to methotrexate in rheumatoid arthritis

OBJECTIVES

Around 30% of patients with RA have an inadequate response to MTX. We aimed to use routine clinical and biological data to build machine learning models predicting EULAR inadequate response to MTX and to identify simple predictive biomarkers.

METHODS

Models were trained on RA patients fulfilling the 2010 ACR/EULAR criteria from the ESPOIR and Leiden EAC cohorts to predict the EULAR response at 9 months (± 6 months). Several models were compared on the training set using the AUROC. The best model was evaluated on an external validation cohort (tREACH). The model's predictions were explained using Shapley values to extract a biomarker of inadequate response.

RESULTS

We included 493 therapeutic sequences from ESPOIR, 239 from EAC and 138 from tREACH. The model selected DAS28, Lymphocytes, Creatininemia, Leucocytes, AST, ALT, swollen joint count and corticosteroid co-treatment as predictors. The model reached an AUROC of 0.72 [95% CI (0.63, 0.80)] on the external validation set, where 70% of patients were responders to MTX. Patients predicted as inadequate responders had only 38% [95% CI (20%, 58%)] chance to respond and using the algorithm to decide to initiate MTX would decrease inadequate-response rate from 30% to 23% [95% CI: (17%, 29%)]. A biomarker was identified in patients with moderate or high activity (DAS28 > 3.2): patients with a lymphocyte count superior to 2000 cells/mm3 are significantly less likely to respond.

CONCLUSION

Our study highlights the usefulness of machine learning in unveiling subgroups of inadequate responders to MTX to guide new therapeutic strategies. Further work is needed to validate this approach.

Identification and validation of metabolism-related genes signature and immune infiltration landscape of rheumatoid arthritis based on machine learning

Rheumatoid arthritis (RA) causes irreversible joint damage, but the pathogenesis is unknown. Therefore, it is crucial to identify diagnostic biomarkers of RA metabolism-related genes (MRGs). This study obtained transcriptome data from healthy individuals (HC) and RA patients from the GEO database. Weighted gene correlation network analysis (WGCNA), the least absolute shrinkage and selection operator (LASSO), and random forest (RF) algorithms were adopted to identify the diagnostic feature biomarker for RA. In addition, biomarkers were verified by qRT-PCR and Western blot analysis. We established a mouse model of collagen-induced arthritis (CIA), which was confirmed by HE staining and bone structure micro-CT analysis, and then further verified the biomarkers by immunofluorescence. In vitro NMR analysis was used to analyze and identify possible metabolites. The correlation of diagnostic feature biomarkers and immune cells was performed using the Spearman-rank correlation algorithm. In this study, a total of 434 DE-MRGs were identified. GO and KEGG enrichment analysis indicated that the DE-MRGs were significantly enriched in small molecules, catabolic process, purine metabolism, carbon metabolism, and inositol phosphate metabolism. AKR1C3, MCEE, POLE4, and PFKM were identified through WGCNA, LASSO, and RF algorithms. The nomogram result should have a significant diagnostic capacity of four biomarkers in RA. Immune infiltration landscape analysis revealed a significant difference in immune cells between HC and RA groups. Our findings suggest that AKR1C3, MCEE, POLE4, and PFKM were identified as potential diagnostic feature biomarkers associated with RA's immune cell infiltrations, providing a new perspective for future research and clinical management of RA.

Specific recommendations to improve the design and conduct of clinical trials

There are many reasons why the majority of clinical trials fail or have limited applicability to patient care. These include restrictive entry criteria, short duration studies, unrecognized adverse drug effects, and reporting of therapy assignment preferential to actual use. Frequently, experimental animal models are used sparingly and do not accurately simulate human disease. We suggest two approaches to improve the conduct, increase the success, and applicability of clinical trials. Studies can apply dosing of the investigational therapeutics and outcomes, determined from animal models that more closely simulate human disease. More extensive identification of known and potential risk factors and confounding issues, gleaned from recently organized "big data," should be utilized to create models for trials. The risk factors in each model are then accounted for and managed during each study.

Efficacy and gastrointestinal tolerability of methotrexate in late-onset rheumatoid arthritis patients: a prospective cohort study

Background

The proportion of the late-onset forms of disease is growing in rheumatoid arthritis (RA) population. Concerns about comorbidities and drug adverse events lead to delay or ineffective treatment in these patients. The aim of this study is to analyze the tolerability and efficacy of methotrexate therapy in late-onset RA (LORA) patients and compare the baseline characteristics, efficacy, and gastrointestinal (GIT) adverse effects of methotrexate treatment between LORA and young-onset RA patients (YORA).

Results

Patients whose symptoms began after 65 years or older were classified as LORA. Baseline characteristics, rheumatoid factor (RF) and anti-citrullinated protein antibody (ACPA) status, C-reactive proten (CRP) levels, disease activity scores, and radiographs of hands and feet were recorded. Patients were started to methotrexate therapy and followed for 6 months. Female gender was predominant in both LORA and YORA. LORA patients had less seropositivity (RF or anti-CCP), higher CRP levels, and higher DAS 28 scores. More than half of the patients (58%) had large joint involvement. Remission rates were higher in LORA patients, and total remission and low disease activity rates were similar. Methotrexate withdrawal due to gastrointestinal adverse events (nausea and vomiting) was lower than YORA patients. Logistic regression analysis demonstrated that DAS 28 score was the only predictor for disease remission (p: 0.000), and no predictive factor was found for methotrexate-related adverse events.

Conclusion

Methotrexate-related gastrointestinal adverse events do not increase in LORA patients, and nausea-vomiting is seen lower than YORA. Methotrexate is well tolerated and effective in LORA patients, and a large amount of patients achieve treatment targets after 6 months of treatment with MTX. Methotrexate should be started immediately in LORA without additional concerns on adverse effects.