Page 16 - Read Online

and normal cells and tissues at the DNA, RNA, protein of the histological, biochemical, molecular and genomic
and microRNA levels. Multiple approaches of various characteristics of the models. [14-16] As many of the
through-put have been developed to identify differentially technologies have become more effi cient and affordable,
expressed genes and proteins. [7,8] Recent advances in whole-genome or transcriptome sequencing is increasingly
transcriptomics, proteomics, genomics, functional genomics, being used to replace traditional microarray-based gene
epigenomics and metabolomics have signifi cantly expanded expression profi ling and copy number variation studies.
the scope and depth of novel targets as well as utility Next generation sequencing (NGS) approaches such as
of existing targets. [6,9-11] Although cell lines have been exome sequencing or whole genome sequencing also
traditionally used due to their availability and accessibility, provide information on mutations and chromosomal
most recent efforts have been focused on patient samples, aberrations such as duplication, deletion and translocation,
tumor biopsies and resections, for example, for their clinical many of which identify tumor suppressors or oncogenic
[17]
relevance and heterogeneity. Once a potential candidate drivers and potentially predict drugs likely to be
target is identifi ed, the next key step is to functionally effi cacious in particular patient subgroups. [18]
validate the target in the context of relevant patient A number of studies were carried out to study the
population. The routinely employed approaches include impact of successive passages on the gene expression,
tool compound, blocking antibody, dominant negative chromosomal stability and copy number variation.
and RNA interference/short hairpin RNA. In addition, it Although not defi nitive and most likely model-dependent,
is imperative to investigate whether the target identifi ed the general consensus in the fi eld is that PDX models
in a small set of cells and tissues are refl ected in a larger should be used at early passages. At relatively low
[19]
population ideally identifi able with selective biomarkers. passage, the histological features, gene expression profi le,
To this end, a collection of large number of clinically copy numbers and chromosomal stability remains very
collected tumor samples and patient-derived tumor models similar to the matching tumor directly harvested from
are critical to ensure translatability from target to drug and patient. [20-23] On the other hand, with each passage to a
from laboratory to clinic.
new mouse host, subsequent genetic changes may occur
Although cancer cell lines are the most widely used at different tendencies intrinsic to individual tumors,
starting material as they are readily available and although the extent and impact of these alterations
propagated to provide suffi cient material for in vitro remain unclear. [24]
manipulation and in vivo tumor growth, most of In reality, each cancer patient’s tumor is heterogeneous
them have been established long time ago and have and unique. And within each of the tumor indications
been selected and cultured under nonphysiological mainly defi ned by anatomic locations of tumor
conditions. In contrast, the least manipulated samples incident (e.g. lung cancer, breast cancer), many
are those directly obtained from patients through surgical subtypes can be identifi ed by histopathology and
procedures or needle biopsies. However, one of the immunohistochemistry (IHC) of an abbreviated panel of
major challenges of using primary patient tumors is their markers. Although these approaches have been widely
limited “shelf-life” and very low quantity in most cases. used to describe and categorize tumors, they have
Compared with cell line models and patient tissues, largely failed to capture the variation of disease within
patient-derived xenografts (PDXs) provide a practical indications. Recently, gene expression profi ling and NGS
solution by both preserving the fi delity of clinical have helped further refi ne the models via molecular
characteristics and providing tumor supply suffi cient for subtyping within individual cancer indications. [25-29]
most target identifi cation and validation strategies. [12,13] Such molecular subtyping can be particularly helpful
Another signifi cant benefi t of using PDX for target in delineating subtypes that can be challenging to
identifi cation and validation is that the process from distinguish with routine histopathology or IHC. For
target identifi cation to validation and then to effi cacy example, traditionally, breast cancer subtyping is mainly
screening can be streamlined around the same models, based on histology fi ndings of IHC staining of selected
therefore, offering a complete circle from patient to markers. Recent molecular profi ling has identifi ed six
mouse and then back to patient.
distinct subtypes (luminal A, luminal B, human epidermal
Patient-derived Xenograft Model growth factor receptor 2, basal-like, claudin-low, and
Characterization a normal-like) with clinically signifi cant differences
in risk factors, incidence, prognosis, and treatment
Typically, when patient samples are obtained for response. [30-33] A similar approach has also been used
establishing PDX models, basic patient information in lung cancer to defi ne clinically relevant subtypes to
(such as age, sex, ethnicity, clinical diagnosis) with the which targeted therapy can be applied to achieve optimal
exception of patient identity will be provided. Once the effi cacy. In lung cancer, especially in non-small cell
tumors are established in immune-compromised mice, lung cancer (NSCLC), recurrent oncogenic drivers such
comprehensive characterization at DNA, RNA and protein as epidermal growth factor receptor, KRAS, anaplastic
levels will be carried out to gain detailed understanding lymphoma kinase, as well as their related pathways can

Journal of Cancer Metastasis and Treatment ¦ Volume 1 ¦ Issue 1 ¦ April 15, 2015 ¦ 9