Multiomics and Machine Learning Accurately Predict Clinical Response to Adalimumab and Etanercept Therapy in Patients With Rheumatoid Arthritis

OBJECTIVE

To predict response to anti-tumor necrosis factor (anti-TNF) prior to treatment in patients with rheumatoid arthritis (RA), and to comprehensively understand the mechanism of how different RA patients respond differently to anti-TNF treatment.

METHODS

Gene expression and/or DNA methylation profiling on peripheral blood mononuclear cells (PBMCs), monocytes, and CD4+ T cells obtained from 80 RA patients before they began either adalimumab (ADA) or etanercept (ETN) therapy was studied. After 6 months, treatment response was evaluated according to the European League Against Rheumatism criteria for disease response. Differential expression and methylation analyses were performed to identify the response-associated transcription and epigenetic signatures. Using these signatures, machine learning models were built by random forest algorithm to predict response prior to anti-TNF treatment, and were further validated by a follow-up study.

RESULTS

Transcription signatures in ADA and ETN responders were divergent in PBMCs, and this phenomenon was reproduced in monocytes and CD4+ T cells. The genes up-regulated in CD4+ T cells from ADA responders were enriched in the TNF signaling pathway, while very few pathways were differential in monocytes. Differentially methylated positions (DMPs) were strongly hypermethylated in responders to ETN but not to ADA. The machine learning models for the prediction of response to ADA and ETN using differential genes reached an overall accuracy of 85.9% and 79%, respectively. The models using DMPs reached an overall accuracy of 84.7% and 88% for ADA and ETN, respectively. A follow-up study validated the high performance of these models.

CONCLUSION

Our findings indicate that machine learning models based on molecular signatures accurately predict response before ADA and ETN treatment, paving the path toward personalized anti-TNF treatment.

Dysregulated miRNome of plasmacytoid dendritic cells from patients with Sjögren's syndrome is associated with processes at the centre of their function

Objective

A considerable body of evidence supports a role for type-I IFN in the pathogenesis of primary SS (pSS). As plasmacytoid dendritic cells (pDCs) are a major source of type-I IFN, we investigated their molecular regulation by measuring expression of a large set of miRNAs.

Methods

pDCs were isolated from peripheral blood of pSS patients (n = 30) and healthy controls (n = 16) divided into two independent cohorts (discovery and replication). Screening of 758 miRNAs was assessed by an OpenArray quantitative PCR-based technique; replication of a set of identified miRNAs was performed by custom array. Functional annotation of miRNA targets was performed using pathway enrichment. Novel targets of miR-29a and miR-29c were identified using a proteomic approach (stable isotope labelling with amino acids in cell culture).

Results

In the discovery cohort, 20 miRNAs were differentially expressed in pSS pDCs compared with healthy control pDCs. Of these, differential expression of 10 miRNAs was confirmed in the replication cohort. The dysregulated miRNAs were involved in phosphoinositide 3-kinase-Ak strain transforming and mammalian target of rapamycin signalling, as well as regulation of cell death. In addition, a set of novel protein targets of miR-29a and miR-29c were identified, including five targets that were regulated by both miRs.

Conclusion

The dysregulated miRNome in pDCs of patients with pSS is associated with aberrant regulation of processes at the centre of pDC function, including type-I IFN production and cell death. As miR-29a and miR-29c are pro-apoptotic factors and several of the novel targets identified here are regulators of apoptosis, their downregulation in patients with pSS is associated with enhanced pDC survival.

Cartilage Quality (dGEMRIC Index) Following Knee Joint Distraction or High Tibial Osteotomy

OBJECTIVE

High tibial osteotomy (HTO) and knee joint distraction (KJD) are treatments to unload the osteoarthritic (OA) joint with proven success in postponing a total knee arthroplasty (TKA). While both treatments demonstrate joint repair, there is limited information about the quality of the regenerated tissue. Therefore, the change in quality of the repaired cartilaginous tissue after KJD and HTO was studied using delayed gadolinium-enhanced magnetic resonance imaging of cartilage (dGEMRIC).

DESIGN

Forty patients (20 KJD and 20 HTO), treated for medial tibiofemoral OA, were included in this study. Radiographic outcomes, clinical characteristics, and cartilage quality were evaluated at baseline, and at 1- and 2-year follow-up.

RESULTS

Two years after KJD treatment, clear clinical improvement was observed. Moreover, a statistically significant increased medial (Δ 0.99 mm), minimal (Δ 1.04 mm), and mean (Δ 0.68 mm) radiographic joint space width (JSW) was demonstrated. Likewise, medial (Δ 1.03 mm), minimal (Δ 0.72 mm), and mean (Δ 0.46 mm) JSW were statistically significantly increased on radiographs after HTO. There was on average no statistically significant change in dGEMRIC indices over two years and no difference between treatments. Yet there seemed to be a clinically relevant, positive relation between increase in cartilage quality and patients' experienced clinical benefit.

CONCLUSIONS

Treatment of knee OA by either HTO or KJD leads to clinical benefit, and an increase in cartilage thickness on weightbearing radiographs for over 2 years posttreatment. This cartilaginous tissue was on average not different from baseline, as determined by dGEMRIC, whereas changes in quality at the individual level correlated with clinical benefit.

Explorative analyses of protein biomarkers in patients with early rheumatoid arthritis achieving sustained drug-free remission after treatment with tocilizumab- or methotrexate-based strategies: from transcriptomics to proteomics

OBJECTIVES

Previously, we identified networks of co-expressed genes related to achieving sustained drug-free remission (sDFR). The aim of the present exploratory analysis was to identify inflammatory proteins associated with achieving sDFR and their enriched biological pathways, and compare these pathways with those found in the previous transcriptomic analyses.

METHODS

Serum samples were used from 60 patients who participated in the U-Act-Early trial and were treated-to-target with tocilizumab plus methotrexate, or tocilizumab or methotrexate; 37 achieved sDFR (≥3 months drug-free) and 23 did not (controls). Luminex® multi-analyte profiling (xMAP)® was used to measure 85 proteins. Partial least square discriminant analyses (PLSDA) identified proteins associated with achieving sDFR within each strategy arm, which were thereafter used for pathway analyses.

RESULTS

PLSDA identified 9, 14 and 13 relevant proteins in the tocilizumab plus methotrexate, tocilizumab and methotrexate arm, respectively and pathway analyses thereafter identified respectively 49, 88 and 117 significantly enriched gene ontology (GO) terms. When comparing these terms with those previously found in the transcriptomic analyses, corresponding pathways were related in the tocilizumab arm to activity of leukocytes; in the methotrexate arm to response of stimuli and regulation of the Janus kinase signal transducer and activator of transcription (JAK-STAT) pathway. In the tocilizumab plus methotrexate arm, no corresponding enriched pathways were found.

CONCLUSIONS

Multiple proteins were associated with achieving sDFR and several biological pathways corresponded, mainly in the methotrexate arm, with our previous transcriptomic findings potentially providing further insights into gene expression and protein translation in newly diagnosed RA patients.

RNA sequencing to predict response to TNF-α inhibitors reveals possible mechanism for nonresponse in smokers

BACKGROUND

Several studies have employed microarray-based profiling to predict response to tumor necrosis factor-alpha inhibitors (TNFi) in rheumatoid arthritis (RA); yet efforts to validate these targets have failed to show predictive abilities acceptable for clinical practice.

METHODS

The eighty most extreme responders and nonresponders to TNFi therapy were selected from the observational BiOCURA cohort. RNA sequencing was performed on mRNA from peripheral blood mononuclear cells (PBMCs) collected before initiation of treatment. The expression of pathways as well as individual gene transcripts between responders and nonresponders was investigated. Promising targets were technically replicated and validated in n = 40 new patients using qPCR assays.

RESULTS

Before therapy initiation, nonresponders had lower expression of pathways related to interferon and cytokine signaling, while also showing higher levels of two genes, GPR15 and SEMA6B (p = 0.02). The two targets could be validated, however, additional analyses revealed that GPR15 and SEMA6B did not independently predict response, but were rather dose-dependent markers of smoking (p < 0.0001).

CONCLUSIONS

The study did not identify new transcripts ready to use in clinical practice, yet GPR15 and SEMA6B were recognized as candidate explanatory markers for the reduced treatment success in RA smokers.

Can baseline serum microRNAs predict response to TNF-alpha inhibitors in rheumatoid arthritis?

BACKGROUND

In rheumatoid arthritis, prediction of response to TNF-alpha inhibitor (TNFi) treatment would be of clinical value. This study aims to discover miRNAs that predict response and aims to replicate results of two previous studies addressing this topic.

METHODS

From the observational BiOCURA cohort, 40 adalimumab- (ADA) and 40 etanercept- (ETN) treated patients were selected to enter the discovery cohort and baseline serum profiling on 758 miRNAs was performed. The added value of univariately selected miRNAs (p < 0.05) over clinical parameters in prediction of response was determined by means of the area under the receiver operating characteristic curve (AUC-ROC). Validation was performed by TaqMan single qPCR assays in 40 new patients.

RESULTS

Expression of miR-99a and miR-143 predicted response to ADA, and miR-23a and miR-197 predicted response to ETN. The addition of miRNAs increased the AUC-ROC of a model containing only clinical parameters for ADA (0.75 to 0.97) and ETN (0.68 to 0.78). In validation, none of the selected miRNAs significantly predicted response. miR-23a was the only overlapping miRNA compared to the two previous studies, however inversely related with response in one of these studies. The reasons for the inability to replicate previously proposed miRNAs predicting response to TNFi and replicate those from the discovery cohort were investigated and discussed.

CONCLUSIONS

To date, no miRNA consistently predicting response to TNFi therapy in RA has been identified. Future studies on this topic should meet a minimum of standards in design that are addressed in this study, in order to increase the reproducibility.