Background

Xenograft models have served as a robust method for investigating the genetic drivers of cancer and determining the potential efficacy of cancer therapeutics. The ability to propagate human cancer cells and tissues in mice was drastically advanced with the discovery of T-cell deficient athymic nude (nu/nu) mice and T- and B-cell deficient severe combined immunodeficient (scid/scid) mice (Flanagan, 1966; Bosma and Carroll, 1991). Since these discoveries additional immunocompromised mouse models have become available including, recombination-activating gene 2 (Rag2)-knockout mice, non-obese (NOD)-scid mice, and NOD-scid IL2Rgamma(null) mice (also known as NSG mice) (Shinkai et al., 1992; Prochazka et al., 1992; Shultz et al., 2005). These immunocompromised mouse models have enabled the development of numerous and diverse in vivo models of human cancer.

There are several options that should be considered when developing a xenograft model including the site of injection or implantation. Subcutaneous xenografts are often used in in vivo studies due to tumor accessibility for growth measurement and imaging; however a significant limitation of this model is the lack of a normal stromal microenvironment for most cancer cells. Orthotopic xenografts offer a complementary stromal microenvironment; however there are also disadvantages to this route depending on the orthotopic site, including more complex surgical procedures, difficulty of measuring tumor growth or response, and the limitations of rodent stroma (Talmadge et al., 2007). The use of orthotopic xenografts has been used extensively in many cancer studies, especially breast cancer research. Injecting into the mammary fat pad is a relatively simple procedure that allows for the visible and measurable growth of breast cancer cells. Even though the mammary fat pad offers a complementary tissue site for breast cancer cells, it is important to note there are distinct differences between the human and rodent mammary stroma and hormonal environment.

Xenograft models have been used extensively as predictive models of cancer therapeutic efficacy. For preclinical studies, it is essential to evaluate drug efficacy and potential toxicities in vivo. Even though in vivo preclinical studies are valuable, the results have not consistently translated to the clinic and the significance of these studies are debated (Talmadge et al., 2007; Sausville and Burger, 2006). There are several variables that need to be considered when designing drug studies such as the appropriate cell lines (or PDX models), dosage and dosing schedules, and statistical analysis. Each of these factors should be carefully considered in order to most closely mimic human cancer progression and treatment response.

Recently there have been significant advances in the development of patient-derived tumor xenografts (PDX). PDX models have the advantage of maintaining the molecular and histological heterogeneity of the original tumor (DeRose et al., 2011). Moreover, they have been shown to be superior at predicting drug response compared to standard cell culture xenograft models (Hait, 2010; Fruchter et al., 1990; Voskoglou-Nomikos et al., 2003; Gao et al., 2015). Recent studies have advanced the success of establishing breast cancer PDX models that recapitulate the molecular, stromal, and phenotypic heterogeneity that exists in breast cancer (DeRose et al., 2013). Overall, cell line-based and patient-derived xenografts are essential models for investigating cancer initiation, progression, and treatment response. Here, we describe the protocols for developing cell-based xenograft models, patient-derived xenograft models from breast cancer tissue, serial tumor transplantation, tumor measurement, and drug treatment.