POS0475 INTEGRATIVE CLINICAL, MOLECULAR AND COMPUTATIONAL ANALYSES ALLOW THE IDENTIFICATION OF DISTINCTIVE PHENOTYPES OF RHEUMATOID ARTHRITIS PATIENTS RELATED TO THE CLINICAL INVOLVEMENT AND THE RESPONSE TO TNF INHIBITORS

Background:TNF inhibitors (TNFi) represent an extraordinary advance in the management of Rheumatoid Arthritis (RA). Despite their benefits, there is a percentage of patients (20–40%) that do not achieve clinical improvement. Therefore, it is necessary to search for new and easily accessible biomarkers predictive of therapeutic response that might guide precision medicine.Objectives:1. To explore changes in the molecular profile of RA patients following TNFi therapy in serum samples. 2. To search for new and reliable biomarkers predictive of TNFi response, based on clinical and molecular profiles of RA patients, by using machine learning algorithms.Methods:In a prospective multicenter study, 79 RA patients undergoing TNFi and 29 healthy donors (HD) were enrolled. Twenty-two RA patients were further included as a validation cohort. Serum samples were obtained before and after 6 months of treatment, and therapeutic efficacy was evaluated. Patients’ response was determined following EULAR response criteria. Serum inflammatory profile was analyzed by a multiplex immunoassay, along with oxidative and NETotic profiles, evaluated by commercial kits. A circulating miRNA array was also performed by next-generation sequencing. Clustering analysis was carried out to identify groups of patients with distinctive molecular signatures. Then, clinical and molecular changes induced by TNFi were delineated after 6 months of therapy. Finally, integrative clinical and molecular signatures as predictors of response were assessed at baseline by supervised machine learning methods, using regularized logistic regressions.Results:Inflammatory, oxidative stress and NETosis-derived biomolecules were found altered in RA patients versus HD, closely interconnected and associated with several deregulated miRNAs. This altered molecular profile at baseline allowed the unsupervised division of three clusters of RA patients with distinctive clinical phenotypes, further linked to TNFi effectiveness. Cluster 1 included RA patients with low levels of pro-inflammatory cytokines, associated with a medium-low disease activity score and good clinical response. Clusters 2-3 comprised patients with high levels of pro-inflammatory cytokines, associated with a high disease activity and a non-response rate of 30%.After 6 months of therapy the molecular profile found altered in RA patients was reversed in responder patients, who achieved a molecular phenotype similar to HDs. However, non-responder patients’ molecular profile remained significantly deregulated, including alterations in inflammatory mediators (IL-6, L-8, TNFα, VEGF, IL-1RA, IL-5, IL-15, GMCSF, GCSF, FGFb), oxidative stress markers (LPO) and NETosis-derived products (Elastase), along with specific miRNAs (miR-199a-5p). These molecular changes further correlated with changes in disease activity score. Machine-learning algorithms identified clinical (Creatinine, IgM, Vitamin D, Swollen Joints, C4, Disease Duration and Tryglicerides) and molecular (Nucleosomes, IL-10, miR-106a-5p, IL-13, IL-12p70, IL-15 and LPO) signatures as potential predictors of response to TNFi treatment with high accuracy. Furthermore, the integration of both features in a combined model increased the predictive value of these signatures (AUC: 0.91). These results were further confirmed in an independent validation cohort.Conclusion:1. RA patients display distinctive altered molecular profiles directly linked to their clinical status and associated with TNFi effectiveness. 2. Clinical response was associated with a specific modulation of the inflammatory profile, the reestablishment of the altered oxidative status, the reduction of NETosis and the reversion of related altered miRNAs. 3. The integrative analysis of the clinical and molecular profiles using machine learning allows the identification of novel signatures as potential predictors of therapeutic response to TNFi therapy.Disclosure of Interests:None declared

Download Full-text

Integrative Clinical, Molecular, and Computational Analysis Identify Novel Biomarkers and Differential Profiles of Anti-TNF Response in Rheumatoid Arthritis

Frontiers in Immunology ◽

10.3389/fimmu.2021.631662 ◽

2021 ◽

Vol 12 ◽

Author(s):

Maria Luque-Tévar ◽

Carlos Perez-Sanchez ◽

Alejandra Mª Patiño-Trives ◽

Nuria Barbarroja ◽

Ivan Arias de la Rosa ◽

...

Keyword(s):

Rheumatoid Arthritis ◽

Machine Learning ◽

Mixed Model ◽

Learning Algorithms ◽

Clinical Status ◽

Machine Learning Algorithms ◽

Supervised Machine Learning ◽

Molecular Profile ◽

Molecular Signatures ◽

Molecular Profiles

Background: This prospective multicenter study developed an integrative clinical and molecular longitudinal study in Rheumatoid Arthritis (RA) patients to explore changes in serologic parameters following anti-TNF therapy (TNF inhibitors, TNFi) and built on machine-learning algorithms aimed at the prediction of TNFi response, based on clinical and molecular profiles of RA patients.Methods: A total of 104 RA patients from two independent cohorts undergoing TNFi and 29 healthy donors (HD) were enrolled for the discovery and validation of prediction biomarkers. Serum samples were obtained at baseline and 6 months after treatment, and therapeutic efficacy was evaluated. Serum inflammatory profile, oxidative stress markers and NETosis-derived bioproducts were quantified and miRNomes were recognized by next-generation sequencing. Then, clinical and molecular changes induced by TNFi were delineated. Clinical and molecular signatures predictors of clinical response were assessed with supervised machine learning methods, using regularized logistic regressions.Results: Altered inflammatory, oxidative and NETosis-derived biomolecules were found in RA patients vs. HD, closely interconnected and associated with specific miRNA profiles. This altered molecular profile allowed the unsupervised division of three clusters of RA patients, showing distinctive clinical phenotypes, further linked to the TNFi effectiveness. Moreover, TNFi treatment reversed the molecular alterations in parallel to the clinical outcome. Machine-learning algorithms in the discovery cohort identified both, clinical and molecular signatures as potential predictors of response to TNFi treatment with high accuracy, which was further increased when both features were integrated in a mixed model (AUC: 0.91). These results were confirmed in the validation cohort.Conclusions: Our overall data suggest that: 1. RA patients undergoing anti-TNF-therapy conform distinctive clusters based on altered molecular profiles, which are directly linked to their clinical status at baseline. 2. Clinical effectiveness of anti-TNF therapy was divergent among these molecular clusters and associated with a specific modulation of the inflammatory response, the reestablishment of the altered oxidative status, the reduction of NETosis, and the reversion of related altered miRNAs. 3. The integrative analysis of the clinical and molecular profiles using machine learning allows the identification of novel signatures as potential predictors of therapeutic response to TNFi therapy.

Download Full-text

Seroprevalence of Antibodies against SARS-CoV-2 in Children with Juvenile Idiopathic Arthritis a Case-Control Study

Journal of Clinical Medicine ◽

10.3390/jcm10081771 ◽

2021 ◽

Vol 10 (8) ◽

pp. 1771

Author(s):

Violetta Opoka-Winiarska ◽

Ewelina Grywalska ◽

Izabela Korona-Glowniak ◽

Katarzyna Matuska ◽

Anna Malm ◽

...

Keyword(s):

Juvenile Idiopathic Arthritis ◽

Disease Activity ◽

Disease Activity Score ◽

Juvenile Arthritis ◽

Activity Score ◽

Control Group ◽

Healthy Children ◽

Serum Samples ◽

History Of ◽

Novel Coronavirus

There is limited data on the effect of the novel coronavirus disease (COVID-19) caused by severe acute respiratory syndrome-coronavirus-2 (SARS-CoV-2) on pediatric rheumatology. We examined the prevalence of antibodies against SARS-CoV-2 in children with juvenile idiopathic arthritis (JIA) and a negative history of COVID-19 and the correlation of the presence of these antibodies with disease activity measured by juvenile arthritis disease activity score (JADAS). In total, 62 patients diagnosed with JIA, under treatment with various antirheumatic drugs, and 32 healthy children (control group) were included. Serum samples were analyzed for inflammatory markers and antibodies and their state evaluated with the juvenile arthritis disease activity score (JADAS). JIA patients do not have a higher seroprevalence of anti-SARS-CoV-2 antibodies than healthy subjects. We found anti-SARS-CoV-2 antibodies in JIA patients who did not have a history of COVID-19. The study showed no unequivocal correlation between the presence of SARS-CoV-2 antibodies and JIA activity; therefore, this relationship requires further observation. We also identified a possible link between patients’ humoral immune response and disease-modifying antirheumatic treatment, which will be confirmed in follow-up studies.

Download Full-text

Tissue-guided LASSO for prediction of clinical drug response using preclinical samples

10.1101/724310 ◽

2019 ◽

Author(s):

Edward W Huang ◽

Ameya Bhope ◽

Jing Lim ◽

Saurabh Sinha ◽

Amin Emad

Keyword(s):

Machine Learning ◽

Drug Response ◽

Drug Sensitivity ◽

Machine Learning Algorithms ◽

Learning Models ◽

Large Databases ◽

Tissue Of Origin ◽

Molecular Profiles ◽

Clinical Drug ◽

Machine Learning Models

ABSTRACTPrediction of clinical drug response (CDR) of cancer patients, based on their clinical and molecular profiles obtained prior to administration of the drug, can play a significant role in individualized medicine. Machine learning models have the potential to address this issue, but training them requires data from a large number of patients treated with each drug, limiting their feasibility. While large databases of drug response and molecular profiles of preclinical in-vitro cancer cell lines (CCLs) exist for many drugs, it is unclear whether preclinical samples can be used to predict CDR of real patients.We designed a systematic approach to evaluate how well different algorithms, trained on gene expression and drug response of CCLs, can predict CDR of patients. Using data from two large databases, we evaluated various linear and non-linear algorithms, some of which utilized information on gene interactions. Then, we developed a new algorithm called TG-LASSO that explicitly integrates information on samples’ tissue of origin with gene expression profiles to improve prediction performance. Our results showed that regularized regression methods provide significantly accurate prediction. However, including the network information or common methods of including information on the tissue of origin did not improve the results. On the other hand, TG-LASSO improved the predictions and distinguished resistant and sensitive patients for 7 out of 13 drugs. Additionally, TG-LASSO identified genes associated with the drug response, including known targets and pathways involved in the drugs’ mechanism of action. Moreover, genes identified by TG-LASSO for multiple drugs in a tissue were associated with patient survival. In summary, our analysis suggests that preclinical samples can be used to predict CDR of patients and identify biomarkers of drug sensitivity and survival.AUTHOR SUMMARYCancer is among the leading causes of death globally and perdition of the drug response of patients to different treatments based on their clinical and molecular profiles can enable individualized cancer medicine. Machine learning algorithms have the potential to play a significant role in this task; but, these algorithms are designed based the premise that a large number of labeled training samples are available, and these samples are accurate representation of the profiles of real tumors. However, due to ethical and technical reasons, it is not possible to screen humans for many drugs, significantly limiting the size of training data. To overcome this data scarcity problem, machine learning models can be trained using large databases of preclinical samples (e.g. cancer cell line cultures). However, due to the major differences between preclinical samples and real tumors, it is unclear how accurately such preclinical-to-clinical computational models can predict the clinical drug response of cancer patients.Here, first we systematically evaluate a variety of different linear and nonlinear machine learning algorithms for this particular task using two large databases of preclinical (GDSC) and tumor samples (TCGA). Then, we present a novel method called TG-LASSO that utilizes a new approach for explicitly incorporating the tissue of origin of samples in the prediction task. Our results show that TG-LASSO outperforms all other algorithms and can accurately distinguish resistant and sensitive patients for the majority of the tested drugs. Follow-up analysis reveal that this method can also identify biomarkers of drug sensitivity in each cancer type.

Download Full-text

A Multibiomarker Disease Activity Score for Rheumatoid Arthritis Predicts Radiographic Joint Damage in the BeSt Study

The Journal of Rheumatology ◽

10.3899/jrheum.131412 ◽

2014 ◽

Vol 41 (11) ◽

pp. 2114-2119 ◽

Cited By ~ 30

Author(s):

Iris M. Markusse ◽

Linda Dirven ◽

Marianne van den Broek ◽

Casper Bijkerk ◽

K. Huub Han ◽

...

Keyword(s):

Rheumatoid Arthritis ◽

Disease Activity ◽

Disease Activity Score ◽

Radiographic Progression ◽

Joint Damage ◽

Activity Score ◽

Serum Samples ◽

Radiographic Damage ◽

Damage Progression ◽

Increased Risk

Objective.To determine whether a multibiomarker disease activity (MBDA) score predicts radiographic damage progression in the subsequent year in patients with early rheumatoid arthritis.Methods.There were 180 serum samples available in the BeSt study (trial numbers NTR262, NTR 265): 91 at baseline (84 with radiographs available) and 89 at 1-year followup (81 with radiographs available). Radiographs were assessed using the Sharp/van der Heijde Score (SvdH). Twelve serum biomarkers were measured to determine MBDA scores using a validated algorithm. Receiver-operating curves and Poisson regression analyses were performed, with Disease Activity Score (DAS) and MBDA score as independent variables, and radiographic progression as dependent variable.Results.At baseline, MBDA scores discriminated more between patients who developed radiographic progression (increase in SvdH ≥ 5 points) and patients who did not [area under the curve (AUC) 0.767, 95% CI 0.639–0.896] than did DAS (AUC 0.521, 95% CI 0.358–0.684). At 1 year, MBDA score had an AUC of 0.691 (95% CI 0.453–0.929) and DAS had an AUC of 0.649 (95% CI 0.417–0.880). Adjusted for anticitrullinated protein antibody status and DAS, higher MBDA scores were associated with an increased risk for SvdH progression [relative risk (RR) 1.039, 95% CI 1.018–1.059 for baseline MBDA score; 1.037, 95% CI 1.009–1.065 for Year 1 MBDA score]. Categorized high MBDA scores were also correlated with SvdH progression (RR for high MBDA score at baseline 3.7; low or moderate MBDA score as reference). At 1 year, high MBDA score gave a RR of 4.6 compared to low MBDA score.Conclusion.MBDA scores predict radiographic damage progression at baseline and during disease course.

Download Full-text

The Relationship of Cytokines IL-13 and IL-17 with Autoantibodies Profile in Early Rheumatoid Arthritis

Journal of Immunology Research ◽

10.1155/2016/3109135 ◽

2016 ◽

Vol 2016 ◽

pp. 1-10 ◽

Cited By ~ 12

Author(s):

Isabela Siloşi ◽

Mihail Virgil Boldeanu ◽

Manole Cojocaru ◽

Viorel Biciuşcă ◽

Vlad Pădureanu ◽

...

Keyword(s):

Rheumatoid Arthritis ◽

Disease Activity ◽

Early Rheumatoid Arthritis ◽

Disease Activity Score ◽

High Specificity ◽

Serum Levels ◽

Activity Score ◽

Serum Concentrations ◽

Serum Samples ◽

And Control

Aims.In the present study, we aimed to assess the concentrations of IL-13 and IL-17 in serum of patients with early rheumatoid arthritis (eRA), the investigation of correlation between the concentrations of these cytokines and disease activity score, and the concentration of some autoantibodies and the evaluation of the utility of IL-13 and -17 concentration measurements as markers of disease activity.Materials and Methods. Serum samples were collected from 30 patients and from 28 controls and analysed parameters.Results. The serum concentrations of IL-13, IL-17, anti-CCP, and IgM-RF were statistically significantly higher in patients with eRA, compared to the controls. IL-13 concentrations in the severe and moderate groups with eRA were statistically higher than in the mild and control groups. Also, in the case of IL-17, serum concentrations increased proportionally with the disease activity of eRA. We observe that concentrations of IL-13 and -17 did not correlate with autoantibodies. IL-17 concentration significantly positively correlated with CRP, while IL-13 concentration significantly negatively correlated with CRP. Disease activity score, DAS28, was strongly positively correlated with levels of ESR and weakly positively correlated with concentrations of anti-RA33 autoantibodies. IL-13 has a higher diagnostic utility than IL-17, CRP, ESR, IgM-RF, and anti-CCP as markers of disease activity.Conclusions. The presence of higher IL-13 and IL-17 serum levels in patients, compared with those of controls, confirms that these markers, found with high specificity, might be involved in the pathogenesis of eRA. IL-13 and IL-17 might be of better usefulness in the prediction of eRA activity status than IgM-RF and anti-CCP.

Download Full-text

Machine Learning Identifies the miR-196b and miR-34c-5p as the Chemotherapy Response Biomarkers of Lung Adenocarcinoma

10.21203/rs.3.rs-54016/v1 ◽

2020 ◽

Author(s):

Yuanmin Lan ◽

Jiqing Xue ◽

Liang Chen ◽

Dongwei Wu ◽

Wei Wang ◽

...

Keyword(s):

Machine Learning ◽

Lung Adenocarcinoma ◽

Response Prediction ◽

Predictive Biomarker ◽

Machine Learning Algorithms ◽

Chemotherapy Response ◽

Serum Samples ◽

Predictive Capacity ◽

Platinum Based Chemotherapy ◽

Response To Chemotherapy

Abstract BackgrooundLung adenocarcinoma is one of the most common malignant lung cancers. Although platinum-based chemotherapy is the first-line adjuvant treatment for middle and late stage lung adenocarcinoma, the response of chemotherapy varies between patients. Moreover, there are no effective biomarkers that could predict chemotherapy response in clinical practice. MiRNAs that are stable in all types of body fluid have demonstrated their diagnostic and prognostic capacity in variety of cancers. Here, we utilized three different machine learning algorithms to identify miRNA signatures specific to chemotherapy response in lung cancer. MethodsThrough a public dataset, a panel of miRNAs for response to chemotherapy was identified by Machine Learning. The predictive capacity was determined by the receiver operation curve. A cohort involving 30 patients with lung adenocarcinoma was utilized for validate the miRNA panel. ResultsMachine Learnings identified five chemotherapy response featured miRNAs (miR-196b, miR-34c-5p, miR-181b, miR-27b and miR-26a). The putative targets of these miRNA signatures are enriched in the biosynthesis. Two of these miRNA signatures (miR-196b and miR-34c-5p) were validated for their chemotherapy response prediction in our 30 serum samples of lung adenocarcinoma with the accuracy of 0.90 and 0.93, respectively. ConclusionsOur study demonstrates circulating miRNA could potentially be predictive biomarker for chemotherapy response in lung adenocarcinoma.

Download Full-text

Profiling of oral microbiota and cytokines in COVID-19 patients

10.1101/2020.12.13.422589 ◽

2020 ◽

Author(s):

Valerio Iebba ◽

Nunzia Zanotta ◽

Giuseppina Campisciano ◽

Verena Zerbato ◽

Stefano Di Bella ◽

...

Keyword(s):

Distinct Species ◽

Bacterial Consortium ◽

Magnetic Bead ◽

Oral Microbiota ◽

Serum Samples ◽

Inflammatory Profile ◽

First Time ◽

Neisseria Perflava ◽

Microbiota Structure ◽

Pro Inflammatory Cytokines

ABSTRACTSARS-CoV-2 presence has been recently demonstrated in the sputum or saliva, suggesting how the shedding of viral RNA outlasts the end of symptoms. Recent data from transcriptome analysis show that oral cavity mucosa harbors high levels of ACE2 and TMPRSS2, highlighting its role as a double-edged sword for SARS-CoV-2 body entrance or interpersonal transmission. In the present study, for the first time, we demonstrate the oral microbiota structure and inflammatory profile of COVID-19 patients. Hospitalized COVID-19 patients and matched healthy controls underwent naso/oral-pharyngeal and oral swabs. Microbiota structure was analyzed by 16S rRNA V2 automated targeted sequencing, while oral and sera concentrations of 27 cytokines were assessed using magnetic bead-based multiplex immunoassays. A significant diminution in species richness was observed in COVID-19 patients, along with a marked difference in beta-diversity. Species such as Prevotella salivae and Veillonella infantium were distinctive for COVID-19 patients, while Neisseria perflava and Granulicatella elegans were predominant in controls. Interestingly, these two groups of oral species oppositely clustered within the bacterial network, defining two distinct Species Interacting Group (SIGs). Pro-inflammatory cytokines were distinctive for COVID-19 in both oral and serum samples, and we found a specific bacterial consortium able to counteract them, following a novel index called C4 firstly proposed here. We even introduced a new parameter, named CytoCOV, able to predict COVID-19 susceptibility for an unknown subject at 71% of power with an AUC equal to 0.995. This pilot study evidenced a distinctive oral microbiota composition in COVID-19 subjects, with a definite structural network in relation to secreted cytokines. Our results would pave the way for a theranostic approach in fighting COVID-19, trying to enlighten the intimate relationship among microbiota and SARS-CoV-2 infection.

Download Full-text

Two-dimensional multiplexed assay for rapid and deep SARS-CoV-2 serology profiling and for machine learning prediction of neutralization capacity

10.1101/2021.08.03.454782 ◽

2021 ◽

Author(s):

Akiko Koide ◽

Tatyana Panchenko ◽

Chan Wang ◽

Sara A Thannickal ◽

Larizbeth A Romero ◽

...

Keyword(s):

Machine Learning ◽

Receptor Binding ◽

Neutralizing Antibodies ◽

Machine Learning Algorithms ◽

Two Dimensional ◽

Serum Samples ◽

Antibody Isotype ◽

Neutralization Capacity ◽

Binding Interface ◽

Neutralizing Capacity

Antibody responses serve as the primary protection against SARS-CoV-2 infection through neutralization of viral entry into cells. We have developed a two-dimensional multiplex bead binding assay (2D-MBBA) that quantifies multiple antibody isotypes against multiple antigens from a single measurement. Here, we applied our assay to profile IgG, IgM and IgA levels against the spike antigen, its receptor-binding domain and natural and designed mutants. Machine learning algorithms trained on the 2D-MBBA data substantially improve the prediction of neutralization capacity against the authentic SARS-CoV-2 virus of serum samples of convalescent patients. The algorithms also helped identify a set of antibody isotype-antigen datasets that contributed to the prediction, which included those targeting regions outside the receptor-binding interface of the spike protein. We applied the assay to profile samples from vaccinated, immune-compromised patients, which revealed differences in the antibody profiles between convalescent and vaccinated samples. Our approach can rapidly provide deep antibody profiles and neutralization prediction from essentially a drop of blood without the need of BSL-3 access and provides insights into the nature of neutralizing antibodies. It may be further developed for evaluating neutralizing capacity for new variants and future pathogens.

Download Full-text