In this paper, we present a formal and operational framework for multiscale modeling and simulation. We focus on scale transfer viewed as the coupling between two models of the same system, one at the microscopic scale and the other at the macroscopic scale. We consider that some parameters of the macroscopic model are the result of microscopic processes described in the microscopic model. The main idea is to use an "experimenter model" which performs experiments on the microscopic model during the simulation and compute these parameters in line. To formalize the experimental design of the experimenter model, we define an experimental frame of transfer based on Discret Event System Specification. As an example, we present a multi scale model in epidemiology.We specify it using our framework, and we implement and simulate it within the Virtual Laboratory Environment. Finally, we discus some of the scale transfer issues and how we want to use the developed framework to address them.