Adaptability of quadruped animals is not solely reached by brain control, but by the interaction between its body, environment, and control. Especially, morphology of the body is supposed to contribute largely to the adaptability. We have tried to understand quadrupedal locomotion by building a bio-inspired quadruped robot named ”Pneupard”, which has a feline-like muscular-skeletal structure. In our previous study, we successfully realized alternative gait of hindlimbs by reflex control based on the sole touch information, which is called an unloading rule, and that of forelimbs as well. In this paper, we finally connect forelimbs and hindlimbs by a rigid spine, and conduct 3D walking experiments only with the simple unloading rule. Through several preliminary experiments, we realize that the touch information on the sole is the most critical for stable 3D walking.