**3. Conclusion**

In this chapter, we described the main models used in breast cancer research in order to obtain results of high scientific quality. In summary, we can say that BCCL models allow repeatable experiments with simple material and methods. They are inevitable models for basic studies and mechanistic explorations, but their use is still controversial owing to their approximate representativeness of breast tumours in human and to the existence of misidentified cell lines.

Cultures of cancerous tissues preserve the tumour architecture and the cell diversity of a tumour but this model suffers of limited reproducibility and cannot be easily maintained for a long time. Co-culture systems offer an alternative with reproducible long term culture systems, and offer the possibility to study the relations between different types of cells in tumour, but this model suffers from the same controversies as BCCL as it mainly relies on their use.

3-dimensional systems allow the mimicking of the tumour architecture and microenvironment, but very few cell lines are able to form spheroids under specific conditions.

Considering the advantages and drawbacks of these models, the xenografts appear to be good alternative models as they enable to take into account the tumour structure, its microenvironment, the role of the metabolism and they preserve the cell diversity of the tumour. But as other models, they also have drawbacks principally due to the metabolic and physiological differences existing between human and rodents, and to the fact that the role of the immune system against tumour is not taken into account with the immunodeficient rodent models used for xenografts.

Application of the 3Rs principle leaded to the development of all these models, but we showed that none of them is sufficient by itself and able to perfectly mimic breast cancer in human. However it clearly appears that all these models are essential to accumulate data and information to fight breast cancer.
