Use of patient-derived models, in which part of a human tumor is transplanted into mice (also known as patient-derived xenografts, or PDXs), is gaining traction as a method to investigate tumor behavior such as “response to therapies”. Human-derived tumor models are becoming popular in the context of “personalized medicine”, but [consistent with what we had previously warned, Regul Toxicol Pharmacol 2o16; 75: 1–4] a new study [see attached full article + editorial] shows that these models could be less representative of primary tumors than previously thought –– particularly when using late passages of the cells in culture.
Many studies agree that PDXs overcome several limitations of more common and established models –– such as human cell lines, including their homogeneity and lack of human stromal microenvironment. To develop PDXs, tumors must be grown in immunodeficient mice, and then transplanted (sequentially) over several generations. Each of these transplantations is commonly referred to as “a passage” and provides the advantage of amplifying the amount of tissue available from a single patient biopsy. In the attached full-length paper, however, Beroukhim, Golub and colleagues questioned the assumption that PDX models remain representative of the original human tumor during passaging. They systematically analyzed previously published data on more than a thousand samples. By tracking DNA copy-number alterations (CNAs) through different passages, they report that PDXs start diverging relatively early from the primary tumor, in addition to which they find data supporting mouse-specific positive selection of pre-existing tumor clones. In other words, clones with minor representation in the human tissue graft can gain a fitness advantage during PDX passaging. Authors argue that this is due to the different evolutionary constraints posed by the mouse environment on human cells. An alternative hypothesis would entail random, non-adaptive genetic drift –– caused by a series of population bottlenecks and expansions at each stage of transplantation.
Cancer research relies on interrogating model systems that mirror the biology of human tumors. Cell lines cultured from human tumors have been the workhorse of cancer research, but marked differences between the cell culture environment and the in vivo tumor environment raise concerns that these lines may not be representative of human tumors. Recently, there have been increasing efforts to use PDXs as models to study drug response; these in vivo models are assumed to capture the cellular and molecular characteristics of human cancer better than simpler cell-line-based models.
Authors [see attached] monitored the dynamics of CNAs in 1,110 PDX samples across 24 cancer types. They observed rapid accumulation of CNAs during PDX passaging, often due to selection of preexisting minor clones. CNA acquisition in PDXs was correlated with tissue-specific levels of aneuploidy and genetic heterogeneity observed in primary tumors. However, the particular CNAs acquired during PDX passaging differed from those acquired during tumor evolution in patients. Several CNAs recurrently observed in primary tumors gradually disappeared in PDXs –– indicating that events undergoing positive selection in humans can become dispensable during propagation in the mouse. Notably, the genomic stability of PDXs was associated with their response to chemotherapy and targeted drugs. These findings have major implications for PDX-based modeling of human cancers of various types.
Nat Genet Nov 2o17; 49: 1557–1575 [full article] + pp. 1565-6 [News’N’Views editorial]