carcinomas, the most recognized form of cellular plasticity is
epithelial-to-mesenchymal transition (EMT), whereby cancer cells
lose their epithelial features (such as the adherens junction protein
E-cadherin [ECAD]) and acquire the features of mesenchymal
lineages (e.g., fibroblasts or leukocytes), thereby facilitating invasion and metastasis. Along with this gain of motile features, EMT
has been reported to be associated with the acquisition of stem-celllike characteristics, including shifts in the vulnerability of these cells
to antineoplastic drugs [2]. Work over the past several years has
determined that EMT does not involve a single mechanism in
either normal physiology or tumors. Rather, it encompasses a
phenotypic spectrum with varying degrees of epithelial and mesenchymal characteristics [3–5]. This has led to the recognition of
intermediate epithelial-mesenchymal phenotypes, also referred to
as partial EMT (P-EMT) states, whereby tumor cells simultaneously express epithelial and mesenchymal features. Importantly,
these partial states may be associated with unique functional abilities, absent in either a pure epithelial state or a pure mesenchymal
state, which confer a metastatic advantage by promoting collective
migration or through other means [6, 7]. The P-EMT state may
also allow for the evasion of certain chemotherapies as another
mechanisms of resistance [8, 9].
Standard approaches to understand EMT in tumors have relied
on in vitro studies and exogenous factors such as Transforming
growth factor beta (TGF-β) [3]. A major challenge facing in vivo
studies has been the difficulty distinguishing carcinoma cells which
have undergone an EMT (therefore exhibit a fibroblast/mesenchymal morphology) from fibroblasts or other mesenchymal cells that
infiltrate and populate the tumor microenvironment [10]. We previously developed a lineage-traced mouse model of pancreatic cancer driven by an oncogenic Kras and loss of Trp53 (KPCY) that
provide a means of distinguishing these various populations based
on a yellow fluorescent protein (YFP) marker carried exclusively by
tumor cells [6, 11]. Here we describe a method to isolate these cell
populations using fluorescence-activated cell sorting (FACS) by
which we isolate YFP+/ECAD+ and YFP+/ECAD- cells to study
transcriptional changes that accompany EMT in vivo. These cells
can further be used for a plethora of applications, including functional studies of cells in culture, gene expression analysis (RNA
sequencing and RT-qPCR), DNA sequencing, epigenetic analysis,
tumor subtyping, western blotting, and immunohistochemistry
(Fig. 1a). We also highlight another method that can be used to
isolate EMT populations by magnetic-activated cell sorting
(MACS) and a dual antibody technique to identify intracellular
proteins that can identify tumor cells that have undergone partial
EMT both in vivo and in vitro. The following techniques are
broadly applicable to other tumor models and tumor cells outside
of pancreatic cancer that contain a lineage label.
316
Robert J. Norgard and Ben Z. Stanger
epithelial-to-mesenchymal transition (EMT), whereby cancer cells
lose their epithelial features (such as the adherens junction protein
E-cadherin [ECAD]) and acquire the features of mesenchymal
lineages (e.g., fibroblasts or leukocytes), thereby facilitating invasion and metastasis. Along with this gain of motile features, EMT
has been reported to be associated with the acquisition of stem-celllike characteristics, including shifts in the vulnerability of these cells
to antineoplastic drugs [2]. Work over the past several years has
determined that EMT does not involve a single mechanism in
either normal physiology or tumors. Rather, it encompasses a
phenotypic spectrum with varying degrees of epithelial and mesenchymal characteristics [3–5]. This has led to the recognition of
intermediate epithelial-mesenchymal phenotypes, also referred to
as partial EMT (P-EMT) states, whereby tumor cells simultaneously express epithelial and mesenchymal features. Importantly,
these partial states may be associated with unique functional abilities, absent in either a pure epithelial state or a pure mesenchymal
state, which confer a metastatic advantage by promoting collective
migration or through other means [6, 7]. The P-EMT state may
also allow for the evasion of certain chemotherapies as another
mechanisms of resistance [8, 9].
Standard approaches to understand EMT in tumors have relied
on in vitro studies and exogenous factors such as Transforming
growth factor beta (TGF-β) [3]. A major challenge facing in vivo
studies has been the difficulty distinguishing carcinoma cells which
have undergone an EMT (therefore exhibit a fibroblast/mesenchymal morphology) from fibroblasts or other mesenchymal cells that
infiltrate and populate the tumor microenvironment [10]. We previously developed a lineage-traced mouse model of pancreatic cancer driven by an oncogenic Kras and loss of Trp53 (KPCY) that
provide a means of distinguishing these various populations based
on a yellow fluorescent protein (YFP) marker carried exclusively by
tumor cells [6, 11]. Here we describe a method to isolate these cell
populations using fluorescence-activated cell sorting (FACS) by
which we isolate YFP+/ECAD+ and YFP+/ECAD- cells to study
transcriptional changes that accompany EMT in vivo. These cells
can further be used for a plethora of applications, including functional studies of cells in culture, gene expression analysis (RNA
sequencing and RT-qPCR), DNA sequencing, epigenetic analysis,
tumor subtyping, western blotting, and immunohistochemistry
(Fig. 1a). We also highlight another method that can be used to
isolate EMT populations by magnetic-activated cell sorting
(MACS) and a dual antibody technique to identify intracellular
proteins that can identify tumor cells that have undergone partial
EMT both in vivo and in vitro. The following techniques are
broadly applicable to other tumor models and tumor cells outside
of pancreatic cancer that contain a lineage label.
316
Robert J. Norgard and Ben Z. Stanger