Search

Background: Multiple sclerosis is an autoimmune disease in which disease heterogeneity is poorly understood. Improving our understanding of pathways that determine inter-patient variation generates new insights that can advance the therapeutic landscape with tailored treatment options. Objective: By focusing on imaging, immunology and genetics, we aim to unravel how patients differ from one another and how mechanistically different treatments can target similar pathways. Results: Using flow cytometry, we found that carrying the CD40 risk allele that increases susceptibility to multiple sclerosis lowers the cell-surface expression of CD40 in all tested B cell subsets. Moreover, the CD40 risk allele resulted in lower total CD40 expression but with an increased proportion of alternative splice-forms leading to decoy receptors. In addition, the CD40 risk allele was associated with decreased levels of IL-10. Intriguingly, the CD40 risk allele conveys protection to other antibody-driven autoimmune diseases. Furthermore, we showed that the magnetization transfer ratio (MTR, a semi-quantitative measure for myelin content) and volumetric brain MRI traits are primarily influenced by unidentified patient-specific factors. Variation over time and demographical variables such as age, gender and disease duration explained much less of the observed variance. Along this line, also the genetic risk score of multiple sclerosis susceptibility variants was not convincingly associated with MTR or volumetric indices. The latter implies that susceptibility and heterogeneity are not driven by the same genetic factors. Conversely, we identified two independent novel associations (rs6982453; rs3729856) on chr8p23 near FDFT1 and CTSB with susceptibility to demyelination across three tissues (lower peak height MTR in normal-appering white matter, normal-appearing grey matter and lesions). In addition, we looked at how two different multiple sclerosis treatments induce similar peripheral B cell changes. Both interferon-beta (IFNB) and fingolimod induce B cell-activating factor (BAFF) without affecting the proportion of the antagonizing spliceform. Importantly, BAFF-induced BAFF receptor signalling might redirect the B cell compartment towards a more immature state. This is the only known shared mechanism across IFNB, fingolimod and many other established multiple sclerosis treatments. Although suggested by multiple sclerosis animal models, BAFF-induced IL-10 signalling, CD40-induced IL-10 signalling and IL-35 were not involved in the IFNB- or fingolimod-induced regulatory immune state. Conclusion: Our data emphasize that multiple sclerosis patients differ in remyelinating potential as well as genetic and immunological composition. Baseline differences likely shape the response to treatment and different treatments ultimately target the same pathways. Understanding these processes allows evolving towards heterogeneity-based patient management. status: published

Dit Standpunt is een aanvulling op de bevindingen en aanbevelingen van de buitenlandse onafhankelijke Denkers die in opdracht van de Academie Vlaanderen hebben bezocht en met betrokken academische, industriële en politieke actoren overleg hebben gepleegd. Op basis van een ruim overleg met de hele watersector en de beleidsstructuren zijn aansluitend actielijnen uitgewerkt die de transitie naar een toekomstgericht waterbeheer in Vlaanderen mogelijk moeten maken.

This work studies the identification and control of circulation fluidized bed (CFB) boilers. The CFB boiler under investigation shows strong nonlinearity due to big changes of steam load. A linear parameter varying (LPV) model is used to represent the process dynamics and used in control. The steam flow is used as the working-point variable (scheduling variable) of LPV model. Multivariable plant tests were carried out and a LPV model and a linear model were identified. The identification results of the industrial CFB boiler show that the LPV model has much higher accuracy than the linear model. The simulation studies for the MPC control of the CFB boiler have been performed using both the LPV model and the linear model. The simulation results show that the MPC using the LPV model performs better than the MPC using the linear model.

In this abstract we introduce velocity control for our 2D dynamic walking robot Dribbel [1] (figure 1) and show that, by ‘closing the loop’, this automatically leads to increased robustness. Takuma and Hosoda [2] also have used active gait adaptation by using step-by-step feedback, but their research was largely based on experimental results, whereas this research is more theoretical.