Phase objects can become visible by slightly defocusing an optical microscope, a technique seldom used as a useful tool. We revisited the theory of defocusing and apply it to our optical microscope with optics corrected at infinity. In our approximation, we obtain that the image contrast is proportional to the two-dimensional (2D) Laplacian of the phase difference introduced by the phase object. If the index of refraction of the phase object is uniform the image obtained from defocusing microscopy is the image of curvature (Laplacian of the local thickness) of the phase object, while standard phase-contrast microscopy gives information about the thickness of the object. We made artificial phase objects and measured image contrasts with defocusing microscopy. Measured contrasts are in excellent agreement with our theoretical model. We use defocusing microscopy to study curvature fluctuations (ruffles) on the surface of macrophages (cell of the innate immune system), and try to correlate mechanical properties of macrophage surface and phagocytosis. We observe large coherent propagating structures: Their shape, speed, density are measured and curvature energy estimated. Inhomogeneities of cytoskeleton refractive index, curvature modulations due to thermal fluctuations and/or periodic changes in cytoskeleton-membrane interactions cause random fluctuations in image contrast. From the temporal and spatial contrast correlation functions, we obtain the decay time and correlation length of such fluctuations that are related to their size and the viscoelastic properties of the cytoskeleton. In order to associate the dynamics of cytoskeleton with the process of phagocytosis, we use an optical tweezers to grab a zymosan particle and put it into contact with the macrophage. We then measure the time for a single phagocytosis event. We add the drug cytochalasin D that depolymerizes the cytoskeleton F-actin network: It inhibits the large propagating coherent fluctuations on the cell surface, increases the relaxation time of cytoskeleton fluctuations, and increases the phagocytosis time. Our results suggest that the methods developed in this work can be of utility to assess the importance of cytoskeleton motility in the dynamics of cellular processes such as phagocytosis exhibited by macrophages.