Some remarks on the Gelfand-Cetlin system