Pioglitazone ameliorates neuronal damage after traumatic brain injury via the PPARg/NF-kB/IL-6 signaling pathway

Traumatic brain injury(TBI)is the major cause of high mortality and disability rates worldwide.Pioglitazone is an activator of peroxisome proliferator-activated receptor-gamma(PPARγ)that can reduce inflammation following TBI.Clinically,neuroinflammation after TBI lacks effective treatment.Although there are many studies on PPARγin TBI animals,only few could be converted into clinical,since TBI mechanisms in humans and animals are not completely consistent.The present study,provided a potential theoretical basis and therapeutic target for neuroinflammation treatment after TBI.First,we detected interleukin-6(IL-6),nitric oxide(NO)and Caspase-3 in TBI clinical specimens,confirming a presence of a high expression of inflammatory factors.Western blot(WB),quantitative real-time PCR(qRTPCR)and immunohistochemistry(IHC)were used to detect PPARγ,IL-6,and p-NF-kB to identify the mechanisms of neuroinflammation.Then,in the rat TBI model,neurobehavioral and cerebral edema levels were investigated after intervention with pioglitazone(PPARγactivator)or T0070907(PPARγinhibitor),and PPARγ,IL-6 and p-NF-kB were detected again by qRT-PCR,WB and immunofluorescence(IF).The obtained results revealed that:1)increased expression of IL-6,NO and Caspase-3 in serum and cerebrospinal fluid in patients after TBI,and decreased PPARγin brain tissue;2)pioglitazone could improve neurobehavioral and reduce brain edema in rats after TBI;3)the protective effect of pioglitazone was achieved by activating PPARγand reducing NF-kB and IL-6.The neuroprotective effect of pioglitazone on TBI was mediated through the PPARγ/NF-kB/IL-6 pathway.

Kernel and graph： Two approaches for nonlinear competitive learning clustering

Competitive learning has attracted a signif- icant amount of attention in the past decades in the field of data clustering. In this paper, we will present two works done by our group which address the nonlin- early separable problem suffered by the classical com- petitive learning clustering algorithms. They are ker- nel competitive learning （KCL） and graph-based multi- prototype competitive learning （GMPCL）, respectively. In KCL, data points are first mapped from the input data space into a high-dimensional kernel space where the nonlinearly separable pattern becomes linear one. Then the classical competitive learning is performed in this kernel space to generate a cluster structure. To real- ize on-line learning in the kernel space without knowing the explicit kernel mapping, we propose a prototype de- scriptor, each row of which represents a prototype by the inner products between the prototype and data points as well as the squared length of the prototype. In GM- PCL, a graph-based method is employed to produce an initial, coarse clustering. After that, a multi-prototype competitive learning is introduced to refine the coarse clustering and discover clusters of an arbitrary shape. In the multi-prototype competitive learning, to gener- ate cluster boundaries of arbitrary shapes, each cluster is represented by multiple prototypes, whose subregions of the Voronoi diagram together approximately charac- terize one cluster of an arbitrary shape. Moreover, we introduce some extensions of these two approaches with experiments demonstrating their effectiveness.