Living cell has active skeletons which consist of fibrous assembly of unit proteins accompanied with dynamic interaction of various proteins. These so-called cytoskeletal proteins such as actin and microtubule can dynamically assemble and disassemble in the network structures. Motor proteins which can walk on the specific cytoskeletal proteins generate driving forces, which lead to deformation at macroscopic, i.e. cell scale, through their cooperative behaviors assisted by various binding proteins which allow modulation of stability and elasticity of the network structure. Therefore, the cell can be regarded as a highly functional active gel which can autonomously reassemble and deform. According to such the dynamic nature, cells show various responses also in mechanical manners against various environmental stimuli. Our understanding on the behaviors of cells against mechanical environments must be necessary for comprehensive elucidation of the mechanism on cellular dynamics which play pivotal roles in important phenomena such as cancer metastasis and tissue organization. To understand the response of cells for mechanical input, methods to give mechanical stimuli have been reported such as point-wise stimuli with use of atomic force microscope, hydrodynamic stimuli by flowing liquid medium, stretching of elastic substrates for adhering cells and so on. Considering a cell that is surrounded by other cells in living organism, it can be assumed that deformations of the surrounding cells are working as mechanical stimuli, since individual cells in the group may deform and/or migrate autonomously. There, mechanical stimuli which are multidirectional at cell scale and act in a multi-point simultaneous manner should exist, however method to give such a mechanical stimulus of micro-fluctuation to cells has been lacked. Thus, we had been motivated to realize a method to give such mechanical stimuli to live cells. Assuming that the group of cells that is providing a dynamic mechanical environment can be regarded as a kind of active gel, we had decided to employ cytoskeletal protein and motor protein to build up active gels, expecting that the physical properties such as power, velocity and elasticity are comparable to cells. Microtubule and kinesin are representative cytoskeletal protein and the motor protein of cells, and method to reproduce movement of these proteins in vitro has been well established as a technique of biophysics. To realize multidirectional movements at cell scale, the microtubules were chemically crosslinked to gel form to integrate the driving force of kinesins on a solid substrate. As the networked microtubule has successfully showed fluctuating motions at micrometer scale, we modified the liquid condition to allow the coexistence of living cells. By seeding metastatic cancer cell on to the substrate modified with this dynamic microtubule network, we had found that the mechanically stimulated cells showed unique morphology which was reminiscent of metastatic behavior of the cancer cells, however, the quantitative analysis was limited due to the technical difficulty in getting desirable success rate for sample preparation. With manual operation of the liquids, flow rate of the sample into chamber was largely fluctuated and it often caused failure with bubbles. To overcome this problem, here we built up an automated pump controlled multi-channel flow system using common and cheap materials, and it allowed quantitative analysis. Detailed condition and the results will be discussed on our presentation in this symposium. Conventional cell culture dish that is made of solid and transparent materials has been allowed us to observe cellular behavior precisely in vitro under microscope. However, the environment in vivo surrounding cells has many different aspects compared to such the dishes. To reduce the gap, attempts to giving a cell-like property to cell culturing substrates have been reported focusing on the passive and active physical properties of the cellular environments. Our approach with microtubule gel could be a candidate for a new type of method to characterize dynamic nature of cells quantitatively, which is beneficial to diagnose metastatic cancer cells