ATRPred: A machine learning based tool for clinical decision making of anti-TNF treatment in rheumatoid arthritis patients.

Rheumatoid arthritis (RA) is a chronic autoimmune condition, characterised by joint pain, damage and disability, which can be addressed in a high proportion of patients by timely use of targeted biologic treatments. However, the patients, non-responsive to the treatments often suffer from refractoriness of the disease, leading to poor quality of life. Additionally, the biologic treatments are expensive. We obtained plasma samples from N = 144 participants with RA, who were about to commence anti-tumour necrosis factor (anti-TNF) therapy. These samples were sent to Olink Proteomics, Uppsala, Sweden, where proximity extension assays of 4 panels, containing 92 proteins each, were performed. A total of n = 89 samples of patients passed the quality control of anti-TNF treatment response data. The preliminary analysis of plasma protein expression values suggested that the RA population could be divided into two distinct molecular sub-groups (endotypes). However, these broad groups did not predict response to anti-TNF treatment, but were significantly different in terms of gender and their disease activity. We then labelled these patients as responders (n = 60) and non-responders (n = 29) based on the change in disease activity score (DAS) after 6 months of anti-TNF treatment and applied machine learning (ML) with a rigorous 5-fold nested cross-validation scheme to filter 17 proteins that were significantly associated with the treatment response. We have developed a ML based classifier ATRPred (anti-TNF treatment response predictor), which can predict anti-TNF treatment response in RA patients with 81% accuracy, 75% sensitivity and 86% specificity. ATRPred may aid clinicians to direct anti-TNF therapy to patients most likely to receive benefit, thus save cost as well as prevent non-responsive patients from refractory consequences. ATRPred is implemented in R.

Anti-tumour necrosis factor-alpha response associated with combined CD226 and HLA-DRB1[*]0404 haplotype in rheumatoid arthritis.

OBJECTIVES: Predicting response to anti-tumour necrosis factor alpha (anti-TNFα) drugs at baseline remains an elusive goal in rheumatoid arthritis (RA) management. The purpose of this study was to determine if baseline genetic variants of PTPRC, AFF3, myD228, CHUK, MTHFR1, MTHFR2, CD226 and a number of KIR and HLA alleles could predict response to anti-TNF-α in rheumatoid arthritis patients. METHODS: Peripheral blood samples were collected from 238 RA patients treated with anti-TNFα drugs. Genotyping was performed using biochip array technology by Randox Laboratories Ltd. and sequence specific polymerase chain reaction. Linear regression analysis was performed to investigate the role of these genotypes in predicting response to treatment, as defined by European League Against Rheumatism (EULAR) response classification and absolute change in disease activity score (DAS28). RESULTS: Of 238 RA patients analysed, 50.4% received adalimumab, 29.7% received etanercept, 14.8% received infliximab, 3.4% certoluzimab and 1.7% golimumab. The MTHFR1 variant rs1801133 was significantly associated with the EULAR response, p=0.044. Patients with the HLA-DRB1*0404 allele displayed a significantly larger reduction in DAS28 compared to non-carriers (mean -2.22, -1.67 respectively, p=0.033). CD226 rs763361 was the only SNP variant significantly associated with ΔDAS28 (p=0.029). CONCLUSIONS: This study has investigated individual allele associations with reductions in DAS28 across a range of anti-TNFα treatments. A combined predictive model indicates that patients with the HLA-DRB1*0404 allele and without the CD226 rs763361 polymorphism exhibit the largest reduction in DAS28 after anti-TNF-α treatment.

Killer immunoglobulin-like receptor and human leukocyte antigen-C genotypes in rheumatoid arthritis primary responders and non-responders to anti-TNF-α therapy.

The identification of patients who will respond to anti-tumor necrosis factor alpha (anti-TNF-α) therapy will improve the efficacy, safety, and economic impact of these agents. We investigated whether killer cell immunoglobulin-like receptor (KIR) genes are related to response to anti-TNF-α therapy in patients with rheumatoid arthritis (RA). Sixty-four RA patients and 100 healthy controls were genotyped for 16 KIR genes and human leukocyte antigen-C (HLA-C) group 1/2 using polymerase chain reaction sequence-specific oligonucleotide probes (PCR-SSOP). Each patient received anti-TNF-α therapy (adalimumab, etanercept, or infliximab), and clinical responses were evaluated after 3 months using the disease activity score in 28 joints (DAS28). We investigated the correlations between the carriership of KIR genes, HLA-C group 1/2 genes, and clinical data with response to therapy. Patients responding to therapy showed a significantly higher frequency of KIR2DS2/KIR2DL2 (67.7% R vs. 33.3% NR; P = 0.012). A positive clinical outcome was associated with an activating KIR-HLA genotype; KIR2DS2 (+) HLA-C group 1/2 homozygous. Inversely, non-response was associated with the relatively inhibitory KIR2DS2 (-) HLA-C group 1/2 heterozygous genotype. The KIR and HLA-C genotype of an RA patient may provide predictive information for response to anti-TNF-α therapy.

CD169+ Monocyte and Regulatory T Cell Subsets Are Associated with Disease Activity in Rheumatoid Arthritis.

Disease activity in rheumatoid arthritis (RA) is influenced by activation of circulating and synovial immune cells. Regulatory T cells (Tregs) and monocytes are key cells that drive inflammation in RA. This study investigated if a relationship exists between disease activity in RA and circulating Treg and monocyte numbers and phenotypes. A potential sialic acid (Sia) mediated link between Tregs and monocytes was also probed in vitro. Peripheral blood mononuclear cells (PBMCs) were isolated from RA patient (n = 62) and healthy control (n = 21) blood using density gradient separation. Flow cytometry was used to count and phenotype Treg and monocyte subsets, and to sort healthy control Tregs for Sia cell culture experiments. The effects of Sia on activated Treg FoxP3 and NFκB expression was assessed by flow cytometry and concentrations of secreted TNFα, IL-10 and IFNγ determined by ELISA. High disease activity RA patients who were unresponsive to disease modifying anti-rheumatic drugs (n = 31), have significantly lower relative numbers (percentages) of CD4+CD25+CD127− Tregs (p < 0.01) and memory CD45RA−FoxP3+ Tregs (p < 0.01), compared to low disease activity responders (n = 24). Relative numbers of non-classical CD169+ monocytes are associated with disease activity in RA (p = 0.012). Sia reduced Treg expression of FoxP3, NFκB and cytokines in vitro. A strong association has been identified between non-classical CD169+ monocytes and post-treatment disease activity in RA. This study also indicates that Sia can reduce Treg activation and cytokine release. We postulate that such a reduction could be mediated by interaction with sialyted proteins captured by CD169+ monocytes.

A Novel Processing-Free Method for RNAseq Analysis of Spontaneous Sputum in Chronic Obstructive Pulmonary Disease.

Background: Assessments of airways inflammation in patients with chronic obstructive pulmonary disease (COPD) require semi-invasive procedures and specialized sample processing know-how. In this study we aimed to set up and validate a novel non-invasive processing-free method for RNA sequencing (RNAseq) of spontaneous sputum samples collected from COPD patients. Methods: Spontaneous sputum samples were collected and stabilized, with or without selection of plugs and with or without the use of a stabilizer specifically formulated for downstream diagnostic testing (PrimeStore® Molecular Transport Medium). After 8 days storage at ambient temperature RNA was isolated according to an optimized RNAzol® method. An average percentage of fragments longer than 200 nucleotides (DV200) >30% and an individual yield >50 ng were required for progression of samples to sequencing. Finally, to assess if the transcriptome generated would reflect a true endotype of COPD inflammation, the outcome of single-sample gene-set enrichment analysis (ssGSEA) was validated using an independent set of processed induced sputum samples. Results: RNA extracted from spontaneous sputum using a stabilizer showed an average DV200 higher than 30%. 70% of the samples had a yield >50 ng and were submitted to downstream analysis. There was a straightforward correlation in terms of gene expression between samples handled with or without separation of plugs. This was also confirmed by principal component analysis and ssGSEA. The top ten enriched pathways resulting from spontaneous sputum ssGSEA were associated to features of COPD, namely, inflammation, immune responses and oxidative stress; up to 70% of these were in common within the top ten enriched pathways resulting from induced sputum ssGSEA. Conclusion: This analysis confirmed that the typical COPD endotype was represented within spontaneous sputum and supported the current method as a non-invasive processing-free procedure to assess the level of sputum cell inflammation in COPD patients by RNAseq analysis.

Siglec-1 and -2 as potential biomarkers in autoimmune disease.

Autoimmune diseases (ADs) are currently treated with anti-inflammatory and immunosuppressive drugs, aimed at reducing symptoms of disease in order to improve quality of life for patients. However, for a significant number of patients these therapies are ineffective, leading to an increased risk of irreversible damage and eventual disability in certain cases. Growing evidence has implicated glycosylated proteins and their cognate receptors in modulation of the autoimmune response. This review will summarize these findings with particular focus on sialic acid-binding immunoglobulin-like lectin (Siglec)-1 and Siglec-2 involvement in AD. Fluctuations in these glycosylation-dependent pathways could act as sentinels of disease activity or drug responses. If validated, protein modification and cellular response markers could help clinicians achieve remission earlier.

Current and future trends in biomarker discovery and development of companion diagnostics for arthritis.

Musculoskeletal diseases such as rheumatoid arthritis are complex multifactorial disorders that are chronic in nature and debilitating for patients. A number of drug families are available to clinicians to manage these disorders but few tests exist to target these to the most responsive patients. As a consequence, drug failure and switching to drugs with alternate modes of action is common. In parallel, a limited number of laboratory tests are available which measure biological indicators or 'biomarkers' of disease activity, autoimmune status, or joint damage. There is a growing awareness that assimilating the fields of drug selection and diagnostic tests into 'companion diagnostics' could greatly advance disease management and improve outcomes for patients. This review aims to highlight: the current applications of biomarkers in rheumatology with particular focus on companion diagnostics; developments in the fields of proteomics, genomics, microbiomics, imaging and bioinformatics and how integration of these technologies into clinical practice could support therapeutic decisions.

Camera phone-based quantitative analysis of C-reactive protein ELISA.

We demonstrate the use of a camera phone as a low-cost optical detector for quantitative analysis of a high-sensitivity C-reactive protein (hs-CRP) enzyme-linked immunosorbent assay (ELISA). The camera phone was used to acquire images of the ELISA carried out in a conventional 96 well plate. Colorimetric analysis of the images was used to determine a standard curve that exhibited excellent agreement with a fitted 4-parameter logistic model (R²=0.998). The limit of detection (LOD) for this approach was determined to be 0.026 ± 0.002 µg/ml (1.035 ± 0.079 µM) CRP. Furthermore, these results were found to be in very close agreement with measurements obtained for the same assay using a laboratory-based instrument. These findings indicate the basic technology to enable low-cost quantitative home-based monitoring of an important clinical biomarker of inflammatory disease may already be present in the patient's home.