Metal nanoparticles find wide application potential in the biological research due to their unique optical, magnetic, electrical properties. Utilization of the nanoparticles prepared by chemical synthesis for biological research is limited by its solubility in water and toxicity. One potential solution to overcome this limitation is encapsulation of metal nanoparticles into protein or lipidic shell. Apoferritin is highly symmetric 12 nm protein cage composed of 24 apoferritin monomers which forms hollow structure with 8 nm cavity. Protein shell specifically interacts with the receptor on the cell membrane which facilitates uptake of the apoferritin from the extra-cellular environment into the cell. Due to the high stability of the protein cage and inherent capacity to encompass metal nanoparticles, apoferritin is widely focused as a potentionally optimal and general system for delivery of the metal nanoparticles into the organism. In addition, apoferritin can be also used as a nanoreactor for the nanoparticle production. We have generated a set of different nanoparticles encapsulated in apoferritin cage which we have intended to use as a label for structural and cellular biology research by cryo-electron microscopy. Here, we show that despite the 8 nm cavity, single 6-8 nm nanoparticle is not formed inside apoferritin cage. Instead, larger number of 2 nm or smaller nanoparticles is present inside apoferritin reducing overall number of atoms which can be incorporated into the molecule and thus the attainable contrast during electron microscopy imaging. Despite its wide utilization in life-science research, we conclude, that apoferritin derived nanoparticle system is not an optimal labeling probe for cryo-electron microscopy.