Doublecortin-like kinase 1 as a novel drug target and biomarker in gastrointestinal and renal cancers

NATHANIEL WEYGANT

SCHOOL OF PHARMACY AND BIOMEDICAL SCIENCES

The thesis is submitted in partial fulfillment of the requirements for the award of the degree of Doctor of Philosophy of the

University of Portsmouth

JANUARY 2018

Doublecortin-like kinase 1

as a novel drug target and biomarker in

gastrointestinal and renal cancers

The thesis is submitted in partial fulfillment of the requirements for the

award of the degree of Doctor of Philosophy of the University of

Portsmouth.

Nathaniel Weygant

January 2018

i

I. Declarations Whilst registered as a candidate for the above degree, I have not been registered for any other research award. The results and conclusions embodied in this thesis are the work of the named candidate and have not been submitted for any other academic award.

Nathaniel Weygant Date

Word count: 10,975

10 January 2018

ii

II. Abstract

Although their existence was once controversial, cancer stem cells (CSCs) are now widely accepted as an important source of metastasis and drug resistance. Recent advances in genetic engineering allowing hierarchies of cell lineages to be traced in vivo have led to the discovery of specific markers of CSCs. Doublecortin-like kinase 1 (DCLK1) is one such marker that specifically identifies chemosensory cells known as tuft cells in normal conditions, that become CSCs in the presence of mutation and inflammation in mouse models of colon and pancreatic cancer.

Using in vitro, in vivo, and in silico techniques, the studies discussed in this PhD

investigation demonstrate four key aspects of DCLK1 biology in injury and cancer. First, DCLK1 is essential to the maintenance of the intestinal epithelium and inducing its loss in radiation injury and colitis leads to barrier disruption, dysregulation of key pathways, and exacerbated disease severity. Second, DCLK1 kinase inhibition using small molecule drugs demonstrates efficacy in models of colon and pancreatic cancer. Third, direct and surrogate markers of DCLK1 may have potential as prognostic markers in gastrointestinal cancers. Finally, DCLK1 may have a comparable CSC/tumor supportive role in kidney cancer and likely other cancers outside of the GI tract.

Despite an expanding body of knowledge concerning DCLK1, many aspects remain mysterious. It is clear from the literature that DCLK1 is a regulator of epithelial-mesenchymal transition and functional stemness, but a complete understanding of this complex molecule will require an understanding of the role of each of its primary isoforms, interacting partners, and upstream and downstream signaling. Clinically, the development of biomarkers and targeted therapies against DCLK1 is warranted and has the potential to improve outcomes in subsets of cancer patients.

iii

III. Table of Contents I. Declarations……………………………………………….……………………i II. Abstract………………………………………….…………………………....ii III. Table of Contents………………………………..………………………….iii

IV. List of Figures………………………………………………………….....vi V. Abbreviations……………………………………………………….………vii VI. Acknowledgements…………….…………..…………………………….viii VII. Dissemination……………………………….…….………….……….…ix

1. Introduction: Tumorigenesis and Metastasis…………………..1

1.1. Solid tumor cancers……………………………………….…….1 1.2. Inflammation, DNA damage, and cancer initiation..……..3 1.3. The tumor microenvironment………………………….………3

1.3.1. Cancer stem cells and EMT in the tumor microenvironment………..4

1.3.2. Endothelia and pericytes in the tumor microenvironment…………..5

1.3.3. Infiltrating immune cells in the tumor microenvironment……………6

1.3.4. Cancer-associated fibroblasts and stromal components……………6

1.4. Cancers of the gastrointestinal tract…….………………...7 1.4.1. Colorectal cancer…………………………………………...…7

1.4.2. Other gastrointestinal tract cancers…………….………………..9

1.4.3. Pancreatic cancer…………………………………………..……..9

1.5. Cancers of the urinary tract…………………………………….11 1.5.1. Renal cell carcinoma……………………………….…………..11

2. The Cancer Stem Cell Hypothesis………………………………...12 2.1. Normal stem cells………………………………………………...12

2.1.1. Pluripotency, differentiation, and self-renewal…………………..12

2.2. Cancer stem cells………………………………………………13 2.2.1. CSCs and tumor initiation……………………………………….15

iv

2.2.2. CSCs, resistance, and tumor progression……………..………..17

3. Doublecortin-like kinase 1: a Specific Marker of Cancer Stem Cells..................................................................................................19 3.1. The doublecortin-like kinase 1 (DCLK1) gene……....….….19 3.2. A unique lineage of DCLK1+ cells in the gastrointestinal

tract………………………………………………………..………..21 3.3. DCLK1 marks tumor-initiating cancer stem cells in the

gastrointestinal tract……………………………….…..………24 3.4. Functional implications of DCLK1 expression in cancers…26

4. Aims of the Thesis……..………………………………….…………...30 5. Results and Discussion………………………….….…….……….....32

5.1. Doublecortin-like kinase 1 deletion in tuft cells results in impaired epithelial repair after radiation injury………………………………….32 5.2. Ablation of doublecortin-like kinase 1 in the colonic epithelium

exacerbates dextran sulfate sodium-induced colitis…….…………..33 5.3. Small-molecule kinase inhibitor LRRK2-IN-1 demonstrates potent activity against colorectal and pancreatic cancer through inhibition of doublecortin-like kinase 1……………………………………………...35 5.4. Doublecortin-like kinase 1 is elevated serologically in pancreatic ductal adenocarcinoma and widely expressed on circulating tumor cells……………………………………………………………………...37 5.5. Survival of patients with gastrointestinal cancers can be predicted by a surrogate miRNA signature for cancer stem-like cells marked by doublecortin-like kinase 1…………………………………………...39 5.6. Doublecortin-like kinase 1 is a broadly dysregulated target against epithelial-mesenchymal transition, focal adhesion, and stemness in clear cell renal carcinoma……………………………………………….41

6. Conclusions and Future Directions……….………….………….43 7. References……………………………………………...………………..45

v

8. Appendix I………………………….……………………………………..58 8.1 Epidemiology and classification of select tumors…………58 8.2 Research Ethics Review Checklist……………………...…….60

vi

IV. List of Figures FIGURE 1. The complex web of proposed hallmarks of cancer…………..……..………2

FIGURE 2. Molecular, cellular, and functional characteristics of EMT.……………..….5

FIGURE 3. EMT predicts poorer survival in CRC………………………………………….8

FIGURE 4. Comparison of the stochastic and hierarchical models of tumorigenesis....14

FIGURE 5. Cre/flox lineage tracing models…………………………………..……..…….16

FIGURE 6. CSCs are a source of drug resistance and disease progression…………..18

FIGURE 7. The domain organizations of the primary isoforms of DCLK1.……………..19

FIGURE 8. Exon/intron organization and regulation of alternative splicing of DCLK1

primary transcripts by CTCF…...........................................................................20

FIGURE 9. IHC of DCLK1+ tuft cells…………………………...…………………………..23

FIGURE 10. DCLK1 associates with miR-200 and EMT in GI cancer………………….27

FIGURE 11. A hypothetical etiology of inflammation-associated intestinal tumors……29

vii

V. Abbreviations

ADM acinar-ductal metaplasia AML acute myeloid leukemia ALDH aldehyde dehydrogenase CSCs cancer stem cells CAFs cancer-associated fibroblasts CAR-T chimeric antigen receptor T cells chRCC chromophobe renal cell carcinoma CTCs circulating tumor cells ccRCC clear cell renal carcinoma CRC colorectal cancer CTL cytotoxic T lymphocyte DNA deoxyribonucleic acid DDR DNA damage response EMT epithelial-to-mesenchymal transition FAP familial adenomatous polyposis FACS fluorescence-activated cell sorting GI gastrointestinal HSC hematopoietic stem cell H&E hematoxylin and eosin HIFs hypoxia-inducible factors ILC2s tissue-resident group 2 innate lymphoid cells IM intestinal metaplasia MHC major histocompatibility complex MET mesenchymal-epithelial transition mCRC metastatic colorectal cancer miRNAs microRNAs MPO myeloperoxidase NSCs normal stem cells PDAC pancreatic ductal adenocarcinoma PanINs pancreatic intraepithelial neoplasia lesions pRCC papillary renal cell carcinoma ROS/RNS reactive oxygen/nitrogen species ROC receiver-operating characteristic RCC renal cell carcinoma RNA ribonucleic acid SNPs single nucleotide polymorphisms siRNAs small-interfering RNAs TF Transcription factor Th T helper cells TCGA The Cancer Genome Atlas TILs tumor infiltrating lymphocytes TME tumor microenvironment UT urinary tract VEGF vascular endothelial growth factor

viii

VI. Acknowledgements First and foremost, I want to express my sincere gratitude to Dr. Sassan Hafizi for his selfless efforts as the Director of Studies for this PhD project. One could not ask for a more patient and responsive supervisor than Dr. Hafizi, and his clarity and guidance has brought order to the chaos of preparing this Commentary. I owe a debt of gratitude to my colleague Dr. Dongfeng Qu, who was kind enough to edit this commentary for errors and with whom I have collaborated on many projects including those described in this thesis. I also want to thank Randal May, Dr. William L. Berry, and Dr. Parthasarathy Chandrakesan who have all contributed materially to my projects through the years. Finally, I want to thank Dr. Courtney Houchen for providing funding for the work discussed in this Commentary, allowing me to gain experience beyond my qualifications, and teaching me about the practical aspects of research.

Envisioning a research career would not have been possible without the influence of some great people that inspired me during my undergraduate studies. Especially, the late Dr. Linda L. Wallace who went out of her way to introduce me to critical thinking and the scientific method through our research of nitrogen fixation in Desmanthus illinoensis, Dr. Scott D. Russell who was always willing to take time out of his busy schedule to discuss science and give advice, and Dr. Thomas S. Ray

who opened my mind to the endless possibilities of scientific research.

Finally, I want to acknowledge my good friends and collaborators Dr. Yang Ge and Dr. Jiannan Yao who continually inspire me to do my best in research and in life, and my family – Jia Li, my sister and brother, and my parents – who support me unconditionally.

Many thanks to all of you and to everyone else who has helped me along the way.

ix

VII. Dissemination 1. May R, Qu D, Weygant N, Chandrakesan P, Ali N, Lightfoot SA, Li L, Sureban SM,

Houchen CW. Brief report: Dclk1 deletion in tuft cells results in impaired epithelial

repair after radiation injury. Stem Cells. 2014 Mar;32(3):822-7. DOI:

10.1002/stem.1566.

2. Qu D*, Weygant N*, May R, Chandrakesan P, Madhoun M, Ali N, Sureban SM, An

G, Schlosser MJ, Houchen CW. Ablation of Doublecortin-Like Kinase 1 in the

Colonic Epithelium Exacerbates Dextran Sulfate Sodium-Induced Colitis. PLoS One.

2015 Aug 18;10(8):e0134212. DOI: 10.1371/journal.pone.0134212.

3. Weygant N*, Qu D*, Berry WL, May R, Chandrakesan P, Owen DB, Sureban SM,

Ali N, Janknecht R, Houchen CW. Small molecule kinase inhibitor LRRK2-IN-1

demonstrates potent activity against colorectal and pancreatic cancer through

inhibition of doublecortin-like kinase 1. Molecular Cancer. 2014 May 6;13:103. DOI:

10.1186/1476-4598-13-103.

4. Qu D, Johnson J, Chandrakesan P, Weygant N, May R, Aiello N, Rhim A, Zhao L,

Zheng W, Lightfoot S, Pant S, Irvan J, Postier R, Hocker J, Hanas JS, Ali N, Sureban

SM, An G, Schlosser MJ, Stanger B, Houchen CW. Doublecortin-like kinase 1 is

elevated serologically in pancreatic ductal adenocarcinoma and widely expressed

on circulating tumor cells. PLoS One. 2015 Feb 27;10(2):e0118933. DOI:

10.1371/journal.pone.0118933.

5. Weygant N*, Ge Y*, Qu D, Kaddis JS, May R, Chandrakesan P, Bannerman-

Menson E, Vega KJ, Bronze MS, An G, Houchen CW. Survival of patients with

gastrointestinal cancers can be predicted by a surrogate microRNA signature for

cancer stem-like cells marked by DCLK1 kinase. Cancer Research. 2016 Jul

76(14):4090-4099. DOI: 10.1158/0008-5472.CAN-16-0029.

6. Weygant N, Qu D, May R, Tierney RM, Berry WL, Zhao L, Agarwal S,

Chandrakesan P, Chinthalapally HR, Murphy NT, Li JD, Sureban SM, Schlosser MJ,

Tomasek JJ, Houchen CW. DCLK1 is a broadly dysregulated target against

epithelial-mesenchymal transition, focal adhesion, and stemness in clear cell renal

carcinoma. Oncotarget. 2015 Feb 10;6(4):2193-2205. DOI:

10.18632/oncotarget.3059.

*Equally contributing authors.

1

1. Introduction: Tumorigenesis and Metastasis 1.1 Solid tumor cancers In the most simplistic view of solid tumor cancers, a normal cell is converted to a tumor-initiating cell by the accumulation of specific mutations with functional consequences to the cell, resulting from heredity, environmental conditions, or random chance1. This cell divides rapidly, avoiding roadblocks such as programmed cell death (apoptosis) that prevent normal cells from dividing indefinitely, leading to the formation of a solid mass (tumor) within the source tissue. Cells within this mass may differentiate into various cell types under selection pressure (as in acquired resistance to chemotherapy), in response to changes to the environment, or for other reasons. Often solid tumors tend towards the deadly acquisition of a motile phenotype leading to the colonization of receptive distant organs (metastasis). This concept, the seed-and-soil hypothesis, was first described in 1889 by London surgeon Stephen Paget who noted 241 of 735 breast cancer patients had liver metastases at necropsy, leading him to reject the role of random chance in where metastases form and conjure the imagery of seeds falling ubiquitously, but sprouting only in fertile soil2 – a straightforward observation that describes much of the basis of modern cancer research.

In the time since Paget’s hypothesis, a more complete understanding of solid tumor

cancer biology has emerged, fueled by advances such as the discovery of genetic transmission of information, the structure and function of deoxyribonucleic acid (DNA), and other discoveries leading to a fully sequenced human genome in the year 20003. In this genomic era, knowledge of the molecular and cellular complexities of cancer biology has increased exponentially. A prominent analysis by Hanahan and Weinberg identifies eight hallmarks and two key enabling characteristics that define cancer4. The accepted hallmarks include cellular immortalization, angiogenesis, cell death resistance, sustained proliferative signaling, evasion of growth suppressors, and activation of invasion and metastasis, while the emerging hallmarks are deregulated cellular metabolism and evasion of immune-mediated destruction and the enabling characteristics are genome instability and tumor-promoting inflammation4. These processes and their distinctive

2

subroutines such as epithelial-mesenchymal transition, pluripotency, growth factor signaling, immune checkpoint, and others form a complex web (FIGURE 1)4-31 that partly explains why effectively treating cancer remains a challenge.

Despite the challenge that cancer presents in its diversity, targeted cancer therapies including kinase inhibitors, monoclonal antibodies, immunotherapy, and emerging gene therapies have begun to improve patient outcomes. The first decrease in total cancer deaths was measured in the year 2005 with a relative survival rate of 68%

compared to 50% for 30 years previous (1975) and 35% for approximately 50 years previous (1953)3, marking the beginning of a new era of improved therapies, biomarkers, and other advances.

FIGURE 1. The complex web of proposed hallmarks of cancer. Through the years a variety of interconnected hallmarks have been proposed to explain cancer’s progression and intractability (figure adapted from Hanahan and Weinberg 20124).

3

1.2 Inflammation, DNA damage, and cancer initiation

Tumor-promoting inflammation is a key enabling characteristic of cancer and a central focus of this PhD investigation. Inflammation is an abnormality of tissue that results from injury and the subsequent immune response. This process plays a key role in cancer initiation and progression, and occurs in response to many environmental causative agents linked to cancer32. Common cancer-associated factors that induce inflammatory injury include alcohol, tobacco, red meat, bacteria (e.g. Helicobacter pylori), and viruses (e.g. HPV and hepatitis)5,7,32. Often inflammatory agents drive acute or chronic conditions that increase the risk of developing cancer32. For instance, excessive alcohol consumption may lead to

chronic pancreatitis or esophagitis which are linked to pancreatic and esophageal cancer respectively.

Inflammation is a primary source of DNA damage, particularly in more susceptible epithelial cells. The inflammatory process leads to DNA damage via production of reactive oxygen or nitrogen species (ROS, RNS) in epithelial and inflammatory cells5. ROS/RNS are persistent enough to breach multiple barriers and penetrate the nucleus where they react with DNA and cause mutations5. In response, the cell activates its primary safeguard – the DNA damage response (DDR) pathway. In DDR, double strand breaks are detected by a complex of MRE11, RAD50, and NBS1 which then initiate key DDR components including ATM/ATR which

phosphorylate g-H2AX, which in turn activates p53 leading to cell cycle arrest.

During arrest, DNA errors are repaired. If repairs fail, then the damaged cell is terminated through apoptosis6. If DDR fails to complete its task and the damaged

cell escapes apoptosis through the presence of p53 mutations or by other means, the cell may become tumorigenic. Alternative paths to somatic mutations (e.g. random chance, damaged DNA mismatch repair, etc.), cellular immortalization, and escape from apoptosis may similarly initiate tumorigenesis.

1.3 The tumor microenvironment

Although tumor initiation is a key process in cancers, the tumor microenvironment (TME), innate and adaptive resistance mechanisms, and the ability to invade locally

4

and spread to distant organs accounts for a large portion of their intractability. The TME concept describes the tumor as an organ unto itself with unique cell types and functions. The cell types present in solid tumors include cancer stem cells (CSCs), endothelial cells, pericytes, infiltrating immune cells, cancer-associated fibroblasts (CAFs), and stromal cells4. Understanding the functions and interplay between these cell types is spurring the development of a new generation of targeted therapies. PD-1 (e.g. PDCD1) targeted monoclonal antibodies (such as nivolumab) are an example of one such therapy that is demonstrating robust survival improvements in some patients33 by targeting the TME to allow improved immune infiltration.

1.3.1 Cancer stem cells and EMT in the tumor microenvironment

Although the existence of CSCs in solid tumors was once controversial, they are now accepted as key members of the TME that facilitate tumorigenesis, resistance, and metastasis. As the primary subject of this PhD investigation, they are covered extensively in 2.1 – 3.3. Importantly, CSCs may induce epithelial-to-mesenchymal transition (EMT), a process by which epithelial cells (expressing E-cadherin, claudins, and MUC1) begin to co-express specific transcription factors (TFs; ZEB1, ZEB2, Snail, Slug, and TWIST), leading to loss of polarity, remodeling of the cytoskeletal structure, and a phenotypic conversion to a transitional and then mesenchymal cell type (expressing vimentin, N-cadherin, and/or fibronectin) which promotes drug-resistance, intravasation through the tumor, and may ultimately lead to extravasation at a distant metastatic site15,34 (FIGURE 2).

EMT can be induced by various stimuli including loss of tumor suppressor miRNAs

such as miR-200a-c15,35,36 and activation of Tgf-b, Wnt, growth factor, and ECM-

integrin pathways15. Moreover, EMT may be reversed via mesenchymal-epithelial transition (MET) which can allow flexibility in the chaotic TME and in response to therapy, and may support invasion and clonal expansion at distant sites15. Finally, EMT supports other mechanisms in the TME including limiting immune infiltration via expression of immune checkpoint ligand PD-L119-21,23,37. Overall, EMT is activated in many cancer types and associated with poor survival, and anti-EMT

5

therapies have the potential to reverse drug resistance and limit dissemination15,17,38-40.

1.3.2 Endothelia and pericytes in the tumor microenvironment

The induction of angiogenesis is a hallmark of cancer dependent on the construction of tumor vasculature from endothelial cells with the support of specialized mesenchymal pericytes4, allowing increased nutrient uptake and likely supporting other functions. This TME component has been known for decades leading to effective therapies targeting vascular endothelial growth factor (VEGF)41, but the characteristics of tumor-associated endothelial cells are not fully elucidated4, and angiogenesis is not a uniform trait across cancer types. For example, renal tumors tend towards heavy vascularization while pancreatic tumors begin with angiogenesis but often become hypovascular during progression4. Finally, novel findings indicate that the presence of bone marrow-derived immune cells including

FIGURE 2. Molecular, cellular, and functional characteristics of EMT. EMT is driven by molecular and cellular changes including expression of specific transcription factors, loss of tight junction proteins and associated cell junctions, and other features that lead to increased stemness, invasion, dissemination, and metastasis.

6

macrophages and myeloid cells support the initiation of angiogenesis in the pre-cancer microenvironment, and some of these cells may penetrate the emerging tumor and convert to endothelia or pericytes to support angiogenesis4.

1.3.3 Infiltrating immune cells in the tumor microenvironment

Tumor infiltrating lymphocytes (TILs) are a diverse component of the TME, which include T cells, B cells, NK cells, and macrophages4. The classic view of TILs would suggest that their presence is an indicator that the immune system is attempting to prevent tumor progression. A straightforward example of this would be the process of major histocompatibility complex (MHC) class I presenting an antigen on the

surface of the tumor which is then detected by a CD8+ cytotoxic T lymphocyte (CTL) that proceeds to force apoptosis of the tumor cell42. Counterintuitively, recent findings show that tumors can hijack some immune cell types and use them to protect themselves from immune-mediated destruction. For example, myeloid derived suppressor cells can block CTL and NK cell activity, and M2 macrophages driven by IL-4 and IL-13 signaling can promote progression and suppress anti-tumor TILs43. Based on an improved understanding of immune cells in the TME, methods to promote or inhibit TIL mechanisms (immunotherapy) are now a major focus of clinical oncology.

1.3.4 Cancer-associated fibroblasts and stromal components

CAFs and tumor stroma form a milieu that provides structural support in the TME and has functional implications for tumor progression. CAFs come in the form of tissue-derived fibroblasts or myofibroblasts, and those recruited to the TME enhance progression by increasing proliferation, angiogenesis, and invasion4. Moreover, evidence indicates that they can program EMT in surrounding cells44, and their secretion of ECM is a cause of desmoplasia in some advanced cancers such as pancreatic ductal adenocarcinoma4. In addition to CAFs, stroma may be recruited to the TME from existing tumor stroma, adjacent normal tissue, or bone marrow4. A

7

better understanding of CAFs and tumor stroma will likely be necessary to overcome drug delivery hurdles.

1.4 Cancers of the gastrointestinal tract

The gastrointestinal (GI) tract is uniquely exposed to inflammatory injury from shifts in microbiota populations45, foreign pathogens, chemicals such as alcohol, and other sources. The need for the GI tract to absorb fluids, digest food, regulate barrier permeability, and monitor microbiome populations and the presence of foreign pathogens among other processes, has led to exquisite cellular specialization in GI epithelia and rapid cell turnover. Varieties of cell types produced from rapidly cycling

populations of stem cells perform absorptive, secretory, and other functions to maintain the complex balance required for homeostasis46,47. Consequently, cells in this dynamic system have an increased risk for DNA damage and are a major source of human cancer. According to the latest US statistics, GI tract cancers were the most category in 2017 accounting for 310,440 cases48. For individual cancers, colon cancer is the 4th most prevalent in the US48 and the UK49. These cancers are exceptionally deadly, with colon and pancreatic cancers ranking 2nd and 3rd in mortality respectively, behind lung cancer in the US48, and 2nd and 5th in the UK49. US and UK incidence, mortality, 5-year survival rates, tumor staging categories, TNM classification, prevalent subtypes, and select risk factors for colorectal, pancreatic, gastric, and esophageal cancer are tabulated in Appendix I.

1.4.1 Colorectal cancer

Colorectal cancer (CRC) develops from normal epithelial cells in stages beginning with an abnormal crypt which transitions to an adenoma and then adenocarcinoma. This process is driven by mutations and epigenetic alterations which may arise in the presence of inflammatory injury. The risk factors for sporadic CRC are partly modifiable and include obesity, alcohol, tobacco, and red meat50. CRC may also arise due to hereditary conditions including Lynch syndrome, which results from mutations in mismatch repair genes, or familial adenomatous polyposis (FAP) which

8

results from germline mutations in the APC gene50. APC is a tumor suppressor that is considered a CRC gate-keeper and mutated in up to 70% of CRCs51-53. The second most prevalent set of mutations in CRC are KRAS mutations, which are often single nucleotide polymorphisms (SNPs) affecting amino acids G12 or G13. Estimates for the prevalence of these mutations range from 30 – 60%54, and they are associated with increased mortality especially in metastatic cancer55,56. As in other cancers, p53 tumor suppressor mutations are also exceptionally common57. In addition to this mutational landscape that demonstrates additive disease severity58, CRC is characterized by an EMT phenotype59,60 linked to survival (FIGURE 3). As discussed previously, EMT is a source of metastatic

transformation61 associated with CSCs which are key players in colorectal tumorigenesis and progression59,60,62.

Treatment strategies for CRC vary by the location and disease severity. Where possible, surgical resection is favored, and following rectal cancer resection, neoadjuvant therapies can be used to improve outcomes. A variety of chemotherapies are available for CRC, but adjuvant therapies are typically not given in non-metastatic disease in the absence of significant risk factors. The mainstay for patients with metastatic CRC (mCRC) are combination treatments such as leucovorin, 5-fluorouracil, and either oxaliplatin (FOLFOX) or irinotecan (FOLFIRI).

A. B.

FIGURE 3. EMT predicts poorer survival in CRC. Analysis of TCGA’s colorectal cancer RNA-seq datasets (n=291) demonstrates that increased EMT marker expression (calculated as previously described61) predicts poor overall (A) and recurrence-free survival (B) with an approximately 30-month decrease in median survival for both comparisons.

* P=0.008 * P=0.014

9

Additionally, targeted therapies may be used against tumors with receptive genotypes, such as anti-EGFR monoclonal antibodies in RAS wild-type CRC50. Despite the availability of broad spectrum and targeted therapies, disease-specific mortality remains high at up to 50%50, and improvements in socioeconomic conditions worldwide will likely increase CRC incidence. However, an improved understanding of the molecular and cellular traits that define CRC, and the development of novel biomarkers and therapeutic strategies may lead to improvements in patient outcomes.

1.4.2 Other gastrointestinal tract cancers

Other common cancers of the GI tract include esophageal cancer, stomach adenocarcinoma, and hepatocellular carcinoma. Like the intestine, the esophagus and stomach have specialized cell types to aid in absorption, secretion, surveillance, and other processes. Both the stomach and esophagus are susceptible to the pre-neoplastic formation of pseudo-intestinal tissue characterized by an intestinal crypt type architecture and the presence of mucin-secreting goblet cells63,64. This intestinal metaplasia is linked to Helicobacter pylori and other inflammatory agents and affected tissue expresses intestinal stem cell markers suggesting its appearance is not coincidental63. Like colorectal and pancreatic cancers, esophageal and stomach cancers also express EMT markers which are capable of driving metastasis in their respective tumors65,66.

1.4.3 Pancreatic cancer

Pancreatic cancer, the most common subtype of which is pancreatic ductal adenocarcinoma (PDAC), is a relatively low incidence cancer with exceptionally high mortality. PDAC’s dismal <7% 5-year survival rate is partly due to limited symptoms in patients during pre-cancerous and early stage disease, aggressive invasion, early dissemination, and resistance to therapy67. Known PDAC risk factors are pancreatitis, age, tobacco use, heavy alcohol consumption, obesity, diabetes, and a family history of the disease. Persons from low risk locales who immigrate to

10

high risk locales adopt the high risk of disease within decades suggesting the importance of environment to PDAC67. Currently, surgery during early disease is the only effective intervention38, enabling an additional 11 months of survival (median) in stage I or II disease68. However, despite this advantage, the disease is so often fatal that global statistics show little difference between PDAC incidence and mortality69, and incidence continues to climb due to improved longevity.

Due to its various functions, the pancreas is comprised of a variety of cell types with one or both of two epithelial types (acinar and ductal) playing a key role in PDAC development. In genetically engineered mouse models, the introduction of KRAS activating mutations with the loss of p53 or another relevant tumor suppressor in

either cell type is sufficient to induce pancreatic ductal adenocarcinoma (PDAC)39,70,71. This dual cell origin may be explained by the process of acinar to ductal metaplasia (ADM), in which injury to the pancreas causes destruction of ducts, and acinar cells convert to compensate for the loss. ADM may also lead to the generation of pancreatic intraepithelial neoplastic lesions (PanINs) which are the forbears of PDAC72. Although the cellular processes leading to PDAC are now better understood, strategies for detecting and blocking these processes at the molecular level will be necessary to improve outcomes due to PDAC’s aggressiveness.

On the molecular level, PDAC shares many similarities with CRC. Although PDACs rarely demonstrate mutations to APC, KRAS G12/G13 and p53 mutations are present in most PDACs. By some estimates, KRAS mutations in PDAC may reach levels >95%. PDACs exhibit high levels of cellular heterogeneity and are characterized by an EMT phenotype which is linked to early metastatic progression39,73. These characteristics along with PDAC’s fibrotic microenvironment (desmoplasia) have complicated therapy. The most commonly used PDAC chemotherapy, gemcitabine, provides a small survival advantage and modestly improves quality of life. More recently, albumin-bound paclitaxel (Abraxane™), anti-EGFR therapy, and combination leucovorin, 5-FU, irinotecan and oxaliplatin (FOLFIRINOX) have come into use and demonstrate benefits beyond gemcitabine alone in suitable candidates67. However, despite these treatment options and

available surgical techniques, survivability remains dismal.

11

1.5 Cancers of the urinary tract

Like the GI tract, the urinary tract (UT) depends on specialized cell types to monitor its microbiome74 and maintain homeostasis, and evidence is beginning to accumulate that UT cancers are associated with changes in the UT microbiome74. Unlike infections of the GI tract, urinary tract infections (UTIs) often present with obvious symptoms allowing more efficient treatment, and epidemiological studies suggest that treatment of UTIs with antibiotics is inversely associated with bladder cancer75. This indirectly implicates inflammation as a key process in UT cancers, which is also supported in kidney cancer, which is associated with exposure to

carcinogens including tobacco, radiation, viral infection, and other inflammatory factors76. US and UK incidence, mortality, 5-year survival rates, tumor staging categories, TNM classification, prevalent subtypes, and select risk factors for kidney cancer are tabulated in Appendix I.

1.5.1 Renal cell carcinoma

Renal cell carcinoma (RCC) is the most common UT cancer and the most common subtype is clear cell RCC (ccRCC) followed by the papillary (pRCC) and chromophobe (chRCC) subtypes77. RCC is more common in males and increases in prevalence with age. Major risk factors for RCC include obesity, hypertension, and tobacco consumption. Partial or radical nephrectomy is the primary therapy for most patients with localized RCC because it rarely presents bilaterally, and in patients with non-localized disease, surgery can reduce progression. However, even radical nephrectomy with curative intent in localized RCC leads to a >30% chance of metastasis77. Despite the availability of advanced surgical options and many chemotherapies, survival of patients with RCC is relatively low and becomes dismal with minor local spread. Like the GI cancers discussed previously, RCCs are often characterized by an EMT phenotype40,78,79 which may explain their metastatic potential. Additionally, they maintain a hypoxic microenvironment and frequently develop resistance to chemotherapies.

12

On the molecular level, mutation to the Von-Hippel Lindau tumor suppressor gene (VHL) is the most common event in RCC, which in combination with other VHL loss mechanisms, represents a unifying factor in RCC development. As a consequence of VHL loss, hypoxia-inducible factors (HIFs) avoid ubiquitin-mediated degradation and drive hypoxia and angiogenesis77. Despite the importance of these processes, they are insufficient to initiate and maintain RCC on their own. Large scale sequencing projects to better understand other contributors have narrowed in on the PI3K/AKT pathway which is frequently mutated and global hypomethylation mediated by SETD2 and chromatin remodeling-associated factors80, but a more complete understanding of RCC remains elusive.

2 The Cancer Stem Cell Hypothesis 2.1 Normal stem cells

Stem cells are undifferentiated cells defined by their ability to originate all cell types of their substrate tissue (pluripotency) including themselves (self-renewal). They are best known for their role in embryonic development (embryonic stem cells), but also reside in adult tissue (somatic or normal stem cells; NSCs). Because most tissue requires specialized cells to function effectively, stem cells have a major impact on homeostasis, the response to environmental changes, and injury response.

2.1.1 Pluripotency, self-renewal, and differentiation

The modern concept of a stem cell, defined by self-renewal ability and pluripotency, originated with the discovery that bone marrow cells could rescue lethally irradiated mice from death by restoring the function of the hematopoietic system81. Additionally, the presence of transplantable colonies was observed in the spleens of the injured mice81. This finding led to the first complete stem cell lineage – the hematopoietic stem cell (HSC) and its progenitors82. HSCs differentiate into ever more specific cell types until reaching a terminal, fully-differentiated type83,84, and the pathways that drive their differentiation are now known in exquisite detail due to an accessible

13

niche (blood), cell-surface markers that allow lineage tracking, and a long history of experimental manipulations using fluorescence-activated cell sorting (FACS).

In contrast to HSCs, identification of stem cells and their lineages in solid tissue has been elusive81. This is partly because removing the cell from its niche and recreating a complicated microenvironment ex vivo is fraught with difficulties. However, NSCs have been identified in intestinal and hair follicle epithelia. Although stem cells have often been imagined as singular populations, findings in both of these tissues suggest the presence of multiple stem cell types – a rapidly cycling type and a slowly cycling (quiescent) type85. Both stem cell niches are structurally similar and support rapid regeneration necessary for response to damage from exposure to pathogens,

radiation, and other environmental factors. Notable research suggests that the quiescent cell type may repopulate the tissue in the event of significant injury which will selectively damage proliferating cells86,87. This combination of quiescence and stemness may be a key factor in the emergence of CSCs.

2.2 Cancer stem cells

Until recently, clonal evolution (the stochastic model) was the primary hypothesis for the functional and molecular variation (heterogeneity) found in tumors88. This model posits that most or all cells can potentially initiate tumorigenesis and do so due to the accumulation of mutations and other molecular changes leading to polyclonal subpopulations of cells over time89 (FIGURE 4A). The Vogelstein group first proposed this model in 1987 based on their studies of CRC progression. Specifically, they proposed that CRC begins with a molecular event (such as hypomethylation) leading to a benign tumor (adenoma), followed by oncogenic KRAS mutation, and then an alteration leading to subsequent tumor suppressor loss89. Since the cells arising from the original source in this model would acquire different mutations or other molecular alterations, this might explain the polyclonal nature of many tumor types.

A more recent hypothesis used to explain the polyclonal nature of tumors is the cancer stem cell hypothesis (hierarchical model)88,90. CSCs are largely expected to

function like NSCs which produce progeny that form the bulk of their source tissue.

14

Although progeny originate from a single source, differentiation leads to a variety of different cellular phenotypes (FIGURE 4B). Like NSCs, CSCs self-renew and produce every cell type that forms the tumor. The concept of CSCs was first proposed in blood cancers during the 1960s-70s, but compelling evidence of CSCs has only accumulated within the past 20 years88. The fusion of new knowledge of CSCs, clonal evolution, the TME, and other factors has led to a better understanding of tumor initiation.

FIGURE 4. Comparison of the stochastic and hierarchical models of tumorigenesis. A. The stochastic model assumes that every cell type can become tumorigenic under certain conditions. This begins with a deleterious event such as a mutation, but a single event is unlikely to initiate tumorigenesis. Instead, an accumulation of events is thought responsible. Moreover, this model predicts homogeneous and heterogeneous tumor types. B. The hierarchical model predicts that tumors arise when a tumorigenic event affects a specific cell type which gains the ability to self-renew and continuously produces the progeny that form the tumor bulk (e.g. a CSC). This cell can be a hijacked normal stem cell or other cell type. The assumptions of this model support heterogeneity as a characteristic of all CSC-driven tumors. The assumptions of the stochastic and hierarchical models together, predict metastatic progression and the possibility of multiple CSC types within the same tumor (adapted from Bradshaw et al, 201690).

15

2.2.1 CSCs and tumor initiation

The first experimental proof of CSCs came in the mid-1990s when it was determined that CD34+/CD38- acute myeloid leukemia (AML) cells could be transplanted into immune deficient mice, populate the bone marrow, and disseminate consistent with human disease, while other populations could not88,91,92. This subpopulation demonstrated self-renewal and the ability to give rise to other populations of leukemia cells as expected of a CSC93. Moreover, since CD34 is a marker of HSCs, this suggested that AML could originate with the direct transformation of HSCs.

Following findings of CSCs in AML and other blood cancers, the search for CSCs in solid tumors began. The first solid tumor CSC population identified was

CD44+/CD24- breast cancer cells which were able to form subcutaneous tumors in immunocompromised mice from <1000 cells88,92,94. Soon, highly potent CD44+ CSC populations were discovered in prostate, colon, pancreatic, and other cancers88,92 – and a second marker, CD133, was found to mark CSCs in brain, colorectal, and lung cancer92. Ultimately, the discovery of these CSC subpopulations was a function of convenience as CD44 and CD133 are extracellular and easily targeted for FACS cell sorting which could be followed by in vitro or in vivo transplantation. Further investigation led to the discovery of new CSC markers such as aldehyde dehydrogenase (ALDH), drug transporter ABCG2, and focal adhesion kinase (FAK). Although all of these enriched solid tumor subpopulations demonstrated significant CSC properties, none could rightfully be described as a pure population of CSCs92 – which could theoretically give rise to spheroids in vitro and tumors in mice from a single cell.

In recent years, genetically engineered, Cre-recombinase driven lineage tracing mouse models have revolutionized CSC discovery (FIGURE 5). The Clevers group demonstrated the value of this technology in 2009 by deleting the Apc tumor suppressor gene in a previously identified population of intestinal NSCs marked by G-protein coupled receptor Lgr5 (Lgr5CreApcflox/flox)14. The induced loss of Apc in the Lgr5+ NSC in this model was sufficient to initiate tumorigenesis. Although the entire tumor originated from Lgr5+ cells, FACS analysis of tumors 36-days post Apc

deletion demonstrated a 6.5% population of Lgr5+ cells14. This and later studies

16

would confirm that Lgr5+ cells not only initiate tumors in conditions of Apc loss, but also reproduce themselves and the additional cell types that maintain tumorigenesis13,14. These and similar studies demonstrated for the first time that a single event in a receptive cell type could consistently drive CSC transformation resulting in solid cancer tumorigenesis.

FIGURE 5. Cre/flox lineage tracing models. On the gene level, Cre is often inserted upstream of an internal ribosome entry site (IRES) which allows Cre and the expression of a reporter (e.g. GFP) to be concurrently driven by a gene promoter of interest. When resulting Cre recombinase encounters two loxP sites in another construct, it deletes the contents between them by recombination. This can activate a persistent reporter under the control of Rosa26 (lineage trace), knockout a gene, or activate a knock-in mutation. On the organism level, this is achieved in mice by crossing a Cre mouse model with a loxP (flox) mouse model. On the cellular level, the Cre/flox system activates an event of interest under the control of the gene promoter in a time-dependent manner, as shown for the reporter/lineage tracing combination above. Depending on the Cre construct, the resulting activity can be constitutive or chemically inducible.

17

2.2.2 CSCs, resistance, and tumor progression

Aside from their established role in initiating cancer, CSCs are central factors in drug resistance and disease progression88,92. Unlike cells that form the bulk of the tumor, CSCs may not be highly proliferative while many drug therapies primarily target rapidly dividing cells. For example, taxanes target microtubules which are accessible during mitosis, and consequently would be less effective against quiescent CSCs. Moreover, CSCs are often molecularly empowered to avoid destruction and may express multi-drug resistance factors and immune checkpoint proteins allowing drug efflux before damage and shielding their presence from the immune system. On a more basic level, CSCs drive tumor heterogeneity which

complicates all non-surgical therapies. Following therapy, most tumor cells may be effectively destroyed, but remaining CSCs are left behind and may be able to initiate recurrence (FIGURE 6).

CSCs are consistently linked to EMT in solid tumor cancers. As discussed previously, EMT fuels cancer migration and invasion, and in combination with CSC properties of self-renewal, pluripotency, and drug-resistance, is thought to be key in the establishment of metastases92. Though subject to debate, EMT itself may confer a CSC identity to otherwise differentiated tumor cells (dedifferentiation)95. Finally, both EMT and CSCs are associated with a TME that includes angiogenesis, hypoxia, increased stromal elements, and the presence of extracellular matrix88, which may be key in the CSC response to therapy and subsequent progression. Overall, identification of specific CSC targets, factors controlling their self-renewal, properties of their niche, checkpoints for their rate of proliferation, and biomarkers for determining patient candidates for therapy will be essential to progress in this area.

18

FIGURE 6. CSCs are a source of drug resistance and disease progression. A. CSCs are a subpopulation of tumor cells defined by their ability to reproduce themselves and all tumor progeny. They are often tumor-initiating cells and are slowly-dividing (quiescent). B. A majority of 1st line chemotherapies are preferentially targeted towards rapidly dividing cells and may destroy the tumor bulk while CSCs resist damage or are unable to be drugged due to off-target toxicity to more slowly dividing normal cells. C. The majority of targeted therapies are not specific to CSCs, although they may remove 1st line therapy-resistant tumor bulk. D. CSCs that resist therapy are primed to become deadly through enhanced drug resistance and invasive capabilities, often mediated by EMT, and as a result of selection.

19

3 Doublecortin-like kinase 1: a Specific Marker of

Cancer Stem Cells 3.1 The doublecortin-like kinase 1 (DCLK1) gene

Human doublecortin-like kinase 1 (DCLK1) was first identified in fetal brain tissue,

localized to chromosome 13, and found to encode four isoforms termed a-Long

(Isoform 2), a-Short (Isoform 1), b-Long (Isoform 4), and b-Short (Isoform 3) based

on the location of their promoter (a or b) and C-terminus length (Long or Short)96.

DCLK1’s a-promoter driven isoforms are comprised of two N-terminal microtubule-

binding domains with high homology to doublecortin (DCX), a serine/proline-rich region suspected to participate in autophosphorylation, a kinase domain with

homology to the CAMK kinase family, and a C-terminus region96,97. DCLK1’s b-

promoter driven isoforms are nearly identical to the a-promoter isoforms but lack N-

terminal DCX domains96 (FIGURE 7).

DCLK1-Long (AL/BL) isoforms result from alternative splicing between exons 16 and 17. Interestingly, this is a binding site for CTCF, a TF known to guide transcription by binding to effectively create a “roadblock” to exon skipping98 (FIGURE 8). To test if CTCF regulates DCLK1 alternative splicing, RNA-Seq

FIGURE 7. The domain organizations of the primary isoforms of DCLK1. a-promoter driven isoforms of DCLK1 (AL/AS or 2/1) share microtubule-binding (DCX) domains at their N-terminus and have a molecular weight of 82 kDa. b-promoter driven isoforms (BL/BS or 4/3) share a shortened N-terminus and have a molecular weight of 52 kDa. Isoforms are described as “Long” or “Short” based on the length of their C-terminus and all isoforms share a portion of the serine/proline-rich (SP-Rich) domain and the complete kinase domain (Source: Uniprot).

DCLK1:O15075

Long

Short

20

expression data generated by TCGA for ccRCC80, pRCC99, PDAC100, colon adenocarcinoma57, lung adenocarcinoma101, and lung squamous cell carcinoma102 was downloaded from the UCSC Cancer Genome Browser103, each cancer type was grouped by CTCF expression, and percentages of expression for each available DCLK1 transcript (AL, BL, AS) were compared. In all cases, DCLK1-AL/BL percentages increased proportionally with the CTCF loss. Moreover, the DCLK1-L/S expression ratio in ccRCC increased with CTCF loss and tumors with the lowest CTCF levels had a 160:1 long/short expression ratio (FIGURE 8). This suggests that CTCF binding and related regulatory processes are factors in DCLK1 alternative splicing.

FIGURE 8. Exon/intron organization and regulation of alternative splicing of DCLK1 primary transcripts by CTCF. Top Panel: The a-promoter encodes a transcript from 18 exons (AS) and the b-promoter encodes a transcript from 14 exons (BS). Also, due to an alternative splicing (exon skip) event in the vicinity of a CTCF binding site, the a-promoter encodes a separate transcript from 17 exons (AL) and the b-promoter encodes a similar alternative transcript from 13 exons (Source: ENSEMBL). Bottom Panel: Analysis of RNA-Seq data from TCGA suggests that DCLK1-Long transcripts are produced in kidney, colon, pancreatic, and lung cancers when transcription factor CTCF is unable to bind its site between DCLK1 exons 16 and 17.

7.6 kB

8.5 kB

7.5 kB

8.4 kB

21

The earliest studies of DCLK1 indicated its expression in the brain, but demonstrated no comparable expression elsewhere96,97. As a result, it was originally thought to be a brain-specific gene, and was studied purely in this context for almost a decade after its discovery. Studies utilizing in vitro models with a focus on neural development soon demonstrated molecular properties for DCLK1 consistent with its protein domains including: kinase activity104,105; microtubule-binding, bundling, and elongation activity104; autophosphorylation and phosphorylation of substrates104,106; and other activities104-107. Initial studies of DCLK1’s function in the brain characterized the protein as a regulator of mitosis and differentiation in neural progenitor cells and axonal transport in neurons108-110. Perhaps as an evolutionary

redundancy, DCLK1 and DCX are able to substitute for one another in stabilizing microtubules and modulating axon genesis and cortical neuron migration109,110. Moreover, loss or gain of DCLK1 function in neural progenitors leads to differentiation suggesting that DCLK1 is a determinant in this process108. Finally,

global knockout of DCLK1’s a-promoter isoforms is not lethal or damaging in mouse

models unless combined with concomitant perturbation of DCX111,112.

Given previous northern and western blotting results suggesting minimal to no expression of DCLK1 outside of the brain, it was not reported in other contexts until 2006 when it was identified as a marker of specific cells in the base of the intestinal crypts of Lieberkühn and the gastric stem cell niche by transcriptome-wide analysis of micro-dissected tissue. These cells were often localized to the suspected location of the multipotent reserve stem cell in the intestine and displayed a poor ability to uptake the proliferative marker bromodeoxyuridine in both the intestine and stomach

indicating quiescence113. Given these reported GI stem cell-like properties114, these findings formed the basis for investigations of the role of DCLK1 in the GI tract.

3.2 A unique lineage of DCLK1+ cells in the gastrointestinal tract

To better understand the role of the DCLK1+ cell in the gut, studies were undertaken to assess its role at homeostasis and in the response to DNA-damaging injury using the well-characterized total body irradiation model. To test if the DCLK1+ cells in the intestinal crypt stem cell region were resistant to radiation injury, mice were

22

irradiated at a highly damaging dose of 6 Gy. Although the DCLK1+ cells demonstrated hallmarks of DNA damage 6 hours after irradiation, they did not demonstrate apoptotic features like surrounding DCLK1- cells. Moreover, 24 hours after irradiation they were found to be either apoptotic or mitotic115 suggesting that some of the DCLK1+ cells could avoid the mechanisms that designated DNA damaged cells for destruction. However, at a lethal dose of 12 Gy, which is sufficient to destroy most intestinal crypts through stem cell eradication, regenerated crypts demonstrated no expression of DCLK1 protein115. These findings suggested that while DCLK1 marked a unique cell population with resistance to apoptosis after DNA-damaging radiation, it was not a marker of NSCs at homeostasis leading to

the conclusion that it might mark a reserve stem cell that functioned during catastrophic injury.

Further studies were conducted to test the reserve stem cell hypothesis and demonstrated strong similarities in pancreas116 and stomach117. Dclk1+ intestinal and pancreatic epithelial cells isolated from normal mouse demonstrated clonogenic capacity in vitro and in vivo – an important property of stem and progenitor cells115,116. However, despite stem-like functionality in the Dclk1+ cell, this concept was effectively refuted by methodical demonstration that Dclk1 marked a novel lineage of terminally differentiated, secretory tuft cells46,118 (FIGURE 9)119. The intestine had been previously defined in the literature by markers of four fully differentiated lineages of Paneth (antibiotic secretions), enteroendocrine (hormone secretions), and goblet (mucus secretions) cells and absorptive enterocytes46. Previous studies focused on Dclk1 had attempted to co-localize markers for these known cell types and failed. Therefore, it was incorrectly concluded that Dclk1 marked undifferentiated cells because the potential for a fifth lineage of cells had not been adequately considered. Building on the findings of the new lineage, reports of apparent Dclk1+ tuft cells in the stomach117,120 and pancreas70,121 emerged suggesting their presence throughout the GI tract.

23

Tuft cells, also known as brush cells, have been found in various tissues and were first described in 1956, but despite knowledge of the existence of this cell type, the lack of a specific marker had hindered research for over 50 years46. With confirmation of a new lineage of GI tuft cells defined by the expression of Dclk1, studies to understand their functionality at homeostasis, in injury, and in neoplasia began. Tuft cells, whose identity had previously been muddled with the secretory enteroendocrine cell lineage, had already been suspected of secreting endogenous opioids122 and were quickly confirmed to secrete endorphins46. Moreover, having a specific tuft cell marker allowed the confirmation of additional chemo-signaling/sensory characteristics long suspected. In particular, Dclk1+ tuft cells express taste receptors (TRPM5, TAS1R1/3) and cyclooxygenase enzymes (COX-1/2) which may contribute to chemosensory function28. However, the biggest leap in understanding tuft cell function was reported in late 2015 with the publication of concurrent papers outlining the mediator role of the Dclk1+ tuft cell lineage in response to intestinal parasite infection123,124. Both papers demonstrated that Dclk1+

tuft cells maintain resident innate immune cells at homeostasis via IL-25 secretion and increase secretion in response to infection to activate tissue-resident group 2 innate lymphoid cells (ILC2s). The activated ILC2s respond by recruiting type II helper T cells which secrete IL-4 and IL-13. The secreted IL-4/IL-13 induces rapid

FIGURE 9. IHC of DCLK1+ tuft cells‡.

A. IHC staining of DCLK1 in normal duodenal epithelium displaying the characteristic apical “tuft” structure.

B-C. IHC staining of DCLK1 in normal adjacent pancreatic epithelium and pancreatic cancer epithelium respectively demonstrating the presence of DCLK1+ tuft cells in both conditions.

D-E. IHC staining of DCLK1 in cancerous breast duct-like epithelium showing tuft-like DCLK1+ cells.

‡Sources: Human Protein Atlas119 (A,D-E) and an unpublished pancreatic cancer tissue microarray stained in the Houchen laboratory (B-C).

24

tuft cell differentiation from the stem/progenitor zone leading to hyperplasia123,124. The result is an expansion of sensory tuft cells which can presumably sense progress in the immune response to infection and activate or deactivate that response as necessary. Interestingly, a subsequent study found that knocking out tuft cell-specific taste receptor TRPM5 blocked their secretion of IL-25 in response to infection125, suggesting that tuft cells activate the immune system after “tasting” the presence of infection. These findings highlight the importance of tuft cells in the inflammatory response which is a key factor in tumor-initiation and progression in GI and other cancers126,127.

3.3 DCLK1 marks tumor-initiating cancer stem cells in the gastrointestinal tract

The first investigation of DCLK1 in cancer was conducted in 2008 and focused on the ApcMin/+ mouse model of intestinal polyposis115 – a model relevant to human CRC. As discussed previously, APC is the most commonly mutated gene in CRC patient tumors128, frequently leading to loss of APC and potential intestinal tumorigenesis. Immunohistochemistry revealed that DCLK1 was expressed in isolated cells in ApcMin/+ adenomas. Notably, these cells were found to be quiescent as measured by proliferative marker PCNA. Additionally, while in normal tissue

DCLK1+ cells demonstrated cytoplasmic expression of b-catenin – a key protein

that supports CSCs and cooperates with APC in colonic tumorigenesis and metastasis115 – in ApcMin/+ adenomas DCLK1+ cells presented with nuclear

translocation of activated b-catenin115. Therefore, it was hypothesized that DCLK1

marked CSC or CSC-like cells in intestinal cancers.

In the coming years, additional studies provided support for the DCLK1+ CSC hypothesis129-135, but confirmation was elusive. Concurrently, the use of inducible Cre recombinase models, which allow specific tracing of cell lineages (FIGURE 5),

was increasing. In 2012, Nakanishi et al reported a tamoxifen-inducible DCLK1 a-

promoter-driven Cre recombinase model (Dclk1Cre-ERT2) that allowed DCLK1+ cell

lineage tracing. Crossing this model with the ApcMin/+ model of intestinal cancer confirmed the hypothesis that DCLK1 is specifically expressed in CSCs in conditions

25

of Apc loss. Specifically, intestinal NSCs (Lgr5+) give rise to DCLK1+ cells which produce progeny to form the ApcMin/+ polyps. Moreover, diphtheria-toxin mediated ablation of DCLK1+ cells in this model results in polyp collapse and returns surrounding tissue to an approximately normal architecture136. Overall, these findings suggested that DCLK1 not only marks CSCs in conditions of Apc loss, but that targeting the DCLK1+ cell for destruction may partly reverse colon cancer progression. Soon after, Westphalen et al reported a separate DCLK1-driven Cre recombinase model utilizing artificial chromosome technology (Dclk1Cre-ERT)10. Unlike the Nakanishi et al model, the engineering in this model does not result in the deletion of a DCLK1 allele via the knock-in process. When crossed with an

Apcflox/flox mouse, which results in specific loss of APC in DCLK1+ cells globally, this model did not appear to result in tumor initiation. To investigate further, Westphalen et al induced conditions mimicking inflammation (colitis) using dextran sulfate sodium (DSS) – a chemical irritant that cannot initiate cancer on its own. With the addition of DSS, tumors formed rapidly in the Dclk1Cre-ERTApcflox/flox mice in agreement with the Nakanishi et al findings. Together these findings confirm that DCLK1 may serve as a specific CSC marker in colon cancer10. However, the application of these concepts outside of inflammation-associated colon cancer should be approached cautiously because the ApcMin/+ model utilized by Nakanishi et al demonstrates significantly reduced tumorigenesis in germ-free conditions137.

KRAS is the most frequently mutated gene in pancreatic cancer and SNPs present at its 12th or 13th amino acid (G12D, G12V, G13D, G13D) result in constitutive over-activation of the protein which signals downstream to a large number of pro-oncogenic pathways including PI3K, AKT, mTOR, anti-apoptosis, and others138,139. Tantalizing evidence for a direct link between DCLK1 and KRAS came from CRISPR-based mutagenesis studies in SW48 KRAS wild-type colon cancer cells. When KRAS G12D was introduced, DCLK1 was determined to be the most overexpressed protein in the proteome by mass spectrometry. Although DCLK1 had been linked to KRAS previously70,140,141, this study was unique because CRISPR allowed the insertion of an activating mutation without unnatural levels of

overexpression, demonstrating that the presence of KRAS mutation leading to constitutive KRAS activation, as it occurs in human cancer development,

26

preferentially induces DCLK1. To investigate the hypothesis that DCLK1 might also mark CSC-like cells in KRAS-mutant pancreatic cancer, Westphalen et al crossbred their existing Dclk1Cre-ERT mouse model with a model that inserts a Kras G12D point mutation in Kras exon 1 and a stop codon upstream resulting in constitutive activation in the presence of Cre recombinase (KrasLSL-G12D). Similar to their Apc-loss colon model10, introducing KRAS G12D in the DCLK1+ cell did not result in spontaneous pancreatic cancer142. However, inducing pancreatitis in the same model using the inflammatory peptide cerulein, led to the formation of pancreatic cancer that could be specifically traced from the DCLK1+ cell142. These findings demonstrated that the DCLK1+ cell may serve a CSC-like role in pancreatic cancer

similar to its role in colon cancer.

3.4 Functional implications of DCLK1 expression in cancers

Gene manipulation studies have demonstrated that beyond its CSC marker role downstream of oncogenic mutations, DCLK1 also regulates key cancer traits including EMT, stemness/pluripotency, and expression of tumor suppressor microRNAs (miRNAs). DCLK1 downregulation was originally reported to result in upregulation of tumor suppressor miRNAs Let-7a, miR-144, miR-145, and miR-200130-132,140. More recent reports have linked DCLK1 to the regulation of other miRNAs or vice-versa, including miR-137, miR-363, miR-424, miR-448, and miR-613143-147. Downregulation of pluripotency and EMT factors is concurrent with suppression of miRNAs, and the effect appears to abrogate xenograft tumor growth130-132,140. Despite these suggestive studies, the mechanism by which DCLK1 is able to effectively regulate these processes remains unknown.

One possible explanation for DCLK1’s broad effects on EMT, stemness, and miRNAs, is a signaling cascade set off by DCLK1’s regulation of either EMT inhibitor miR-200 or ZEB1, which form a positive feedback loop35,36,148. EMT and hypothesized CSC transdifferentiation15,149 could occur downstream of this loop, explaining many functional effects of DCLK1 overexpression or downregulation. Analysis of TCGA RNA-Seq and miRNA-Seq data provide support for a strong

correlation between DCLK1 and EMT’s molecular signature in GI cancers57,100,150-

27

152, with the exception of liver cancer (FIGURE 10A). In colon adenocarcinoma, which shows the most correlation, expression of EMT’s signature is mutually exclusive with low expression of DCLK1 (FIGURE 10B), and inverse correlations with each miR-200 isoform and DCLK1, as well as corresponding positive and negative correlations with EMT and epithelial markers respectively, are present (FIGURE 10C).

Another possible explanation for DCLK1’s functional activity in cancer is its involvement in RAS signaling, which could have implications in pancreatic, colorectal, and lung adenocarcinoma, and other cancers. Not only does KRAS activation specifically upregulate DCLK1, but downregulating the expression of KRAS with shRNA results in a dose-dependent decrease in DCLK1153. Although

FIGURE 10. DCLK1 associates with miR-200 and EMT in GI cancer. A. Analysis of DCLK1 and EMT signature expression shows strong correlation in all GI cancers with the exception of hepatocellular carcinoma. B. Alluvial plot demonstrating that in colon adenocarcinoma, high EMT levels (>75th percentile) are mutually exclusive with low DCLK1. C. DCLK1 is inversely correlated with the miR-200 EMT inhibitor family (a-c) and shows expected positive and negative associations with EMT/mesenchymal and epithelial genes respectively.

28

this KRAS-mediated regulation is driven transcriptionally, a recent study demonstrated that DCLK1 interacts with mutant KRAS and found significant homology between its N-terminal domain and RAS effector RalGDS142. If confirmed, these findings may describe a KRAS-DCLK1 feedback loop that explains regulation of DCLK1 expression by members of MAPK/ERK pathway downstream of KRAS142,154-157. Given a recent study demonstrating that KRAS activation suppresses the miR-200 EMT inhibitor family158, it is tempting to speculate that DCLK1’s interactions with the aforementioned miR-200-ZEB1-EMT axis and RAS/MAPK may be directly connected.

Perhaps the most important aspect of DCLK1’s function that remains to be fully

understood is the activity of its kinase. This is complicated by the multiple isoforms of DCLK1 with varying subcellular localizations and functions. For isoforms containing a microtubule-binding domain, some evidence suggests that their kinase activity and phosphorylation may impair their ability to polymerize microtubules159,160. Additionally, although a few substrates including MAP7D1, synapsin I and II, and JDP2161,162 have been identified, a role for these in DCLK1’s tumor function is unknown. Going forward, determining the specific function of DCLK1’s kinase activity may elucidate its role in cancer163 and the recently solved crystal structure for this domain159 may aid in this task.

Overall, DCLK1 and the DCLK1+ tuft cell’s homeostatic function, lineage characteristics, response to injury and DNA damage, demonstrated role in tumor initiation and maintenance, and ability to drive EMT and invasion are logically connected and deserve continued investigation (FIGURE 11). Opportunities to target either DCLK1 or the DCLK1+ tuft cell are likely to emerge in the form of targeted antibodies, gene therapies, kinase inhibitors, and other innovative therapies. To fully realize these opportunities, an improved understanding of DCLK1 and the tuft cell will be necessary.

29

FIGURE 11. A hypothetical etiology of inflammation-associated intestinal tumors. Intestinal stem cells (ISCs) (A) produce every cell in the intestinal epithelium – antibiotic secreting paneth cells (B), hormone secreting enteroendocrine cells (C), mucin secreting goblet cells (D), absorptive enterocytes (E), and chemosensory and IL-25 and opioid-secreting tuft cells (F) which are marked by the expression of DCLK1. In the presence of significant injury (G), DNA damage is inflicted on the cells of the crypt (H) in response to which, tuft cells demonstrate significant resistance (I), and are tasked with orchestrating the response of the innate immune system. In this capacity, tuft cells secrete IL-25 which ultimately recruits immune cells that initiate tuft cell hyperplasia (J) through secretion of IL-4/13. Due to the resistance that makes this process possible and their quiescence and long-lived nature, tuft cells are exposed to DNA-damage and may escape apoptosis leading to mutation. A mutation to the tuft cell in the presence of inflammation is sufficient to transform the tuft cell into a cancer stem cell (K) that is able to produce the lineages of the primary tumor (L). As the tumor progresses, the expression of DCLK1 is a major contributing factor to EMT and may support intravasation (M) as evidenced by the expression of DCLK1 in circulating tumor cells.

30

4 Aims of the PhD investigation Given the significance of DCLK1 and the DCLK1+ tuft cell in GI injury and tumorigenesis, the following Aims were developed to further understanding of the role of DCLK1 in these processes, and are the subject of this PhD study:

1) Determine if Dclk1 can foster resistance to DNA damage-inducing epithelial injury in the gastrointestinal tract. Dclk1 is specifically expressed in tuft cells in the intestinal epithelium which resist destruction following injury115. Because epithelial injury is an important source of oncogenic transformation, determining whether Dclk1 is a functional mediator of this resistance, rather than just a marker of the resistant cell, is key to understanding how tuft cells become CSCs. [Papers 5.1, 5.2]

2) Determine if inhibitors of DCLK1 kinase activity have DCLK1-specific therapeutic potential in gastrointestinal cancers. DCLK1 is expressed in all gastrointestinal cancers10,70,120,136,140,157,164 and specifically marks tumor-initiating, CSCs in the colon10,12 and pancreas142. RNA-interference targeting DCLK1 shows promising effects in vitro and in vivo130-132,140, but gene therapies for solid tumors are in the early stages of development and it

remains to be seen whether this therapeutic concept will become clinically feasible. Conversely, kinase inhibitors are a class of drugs that are extensively used in cancer therapy with beneficial results. [Paper 5.3]

3) Assess DCLK1’s potential as a biomarker in gastrointestinal cancers. Biomarkers that can predict the presence of or severity of disease are desperately needed in clinical practice. As a specific CSC marker in colon and pancreatic cancer, DCLK1 has the potential to predict the predilection for recurrence or metastasis, response to therapy, and other traits associated with CSCs.

31

[Papers 5.4, 5.5]

4) Assess DCLK1 as a potential cancer stem cell marker or therapeutic target in cancers outside of the gastrointestinal tract. The functional significance of DCLK1 in gastrointestinal cancers is established, but there is limited knowledge about its role in other cancers. A number of tissue types demonstrate apparent tuft cells and the sensory and secretory function that DCLK1 contributes to in the gastrointestinal tract may be important in organs such as the lungs, breast, and urinary tract. [Paper 5.6]

32

5 Results and Discussion 5.1 Doublecortin-like kinase 1 deletion in tuft cells results in impaired epithelial repair after radiation injury

Background

Radiation injury is a risk factor for some GI cancers, and experimentally it is valuable as a direct method for studying the effects of DNA damage and DDR – a process that results in the rapid accumulation of mutations and eventually may give rise to tumor-initiating cells. In the intestine, the radiation injury model is pertinent to the study of the origin and fate of cells, and is especially advantageous because the kinetics of the intestinal radiation response are well known165. The first study to assess DCLK1’s role outside of the brain made extensive use of this model115 and although it led to incorrect conclusions about the nature of DCLK1+ cells in the intestine, the findings of this study may explain how they eventually become CSCs. When exposed to sub-lethal doses of radiation (e.g. 6 Gy), DCLK1+ cells accumulate DNA damage as evidenced by the expression of phosphorylated histone H2AX. However, unlike surrounding DNA-damaged cells, some DCLK1+ cells escape apoptosis. This effect is overcome at lethal doses of radiation (>8 Gy)115. Therefore, understanding DCLK1’s role in the intestinal epithelial response to DNA-damage may have implications for its role in the tuft cell and in the development of cancer.

Cre/flox mouse models allow specific events to be driven at the promoter level including gene knockout, activation, and cellular lineage tracing (FIGURE 5). Compared to conventional knock-in or knockout mouse models, this allows monitoring of events that occur at the organ, tissue, or cellular level to be examined specifically in vivo. Although conventional constitutive global knockouts of Dclk1 have previously been reported in the literature and assessed from the standpoint of neurobiology, tissue compartment specific models of Dclk1 knockout had not previously been established.

33

Summary of Results

Cross-breeding of intestinal epithelial-specific VillinCre mice with mice with loxP

flanked Dclk1 exon 3 (Dclk1flox/flox)109, which is present in a-promoter isoforms of

DCLK1 (AL/AS; FIGURE 7), led to excision of Dclk1 as confirmed by PCR of isolated

intestinal epithelium. RNA-sequencing of isolated intestinal epithelial cells from the Dclk1 knockout mice demonstrated broad dysregulation of pathways including stem cell, notch signaling, tight junction, WNT, and sensory/secretory signaling. Induction of radiation injury after loss of Dclk1 in the intestinal epithelium led to dramatic consequences including loss of tight junctions, increased membrane permeability, decreased stem cell marker expression, impaired regeneration, and reduced survival times (approximately 5 days vs. 10 days for controls).

Conclusion/Significance

These results support a sensory/signaling role for tuft cell Dclk1 in maintaining the epithelium in injurious conditions and implicate it in GI cancer signaling pathways (stem cell, WNT, Notch). In combination with previous findings of intestinal epithelial Dclk1+ cell resistance to apoptosis following DNA damage115, a hypothetical pathway towards tumorigenesis is apparent (FIGURE 11).

5.2 Ablation of doublecortin-like kinase 1 in the colonic epithelium exacerbates dextran sulfate sodium-induced colitis

Background

The response to epithelial injury is key to restoring normal tissue architecture and homeostasis in many organs. This is of particular importance in the crypt-villus axis of the intestine where stem cells that reside at the base of the crypt have the ability to rapidly replace all lineages of the epithelium. The tuft cell lineage of the intestine, marked by DCLK1, is able to resist radiation injury115 and is essential to survival afterwards as described in manuscript 5.1. However, acute and chronic

34

inflammation are far more common human diseases of the intestine, and conditions such as ulcerative colitis and Crohn’s disease predispose patients to CRC. Considering the DCLK1+ tuft cell’s transformation to a tumor-initiating, CSC under conditions of Apc loss10,136, determining its contribution to the development and progression of pre-cancerous inflammatory disease may provide insights into colorectal tumorigenesis.

Colitis is linked to a number of factors including lifestyle, genetics, and infection. The response of the immune system to these events is perhaps the most important factor in the development of chronic colitis, precancerous lesions, and eventually cancer126,166. The DCLK1+ tuft cell is able to sense intestinal infection (perhaps via

taste receptors125) and secrete IL-25 to recruit ILC2s and Th cells to the site of infection. Th cells then secrete IL-13 to recruit macrophages and IL-4 which results in tuft cell hyperplasia123,124. This process may form a surveillance feedback loop which is programmed to diminish when infection is eradicated. Given these aspects of the DCLK1+ tuft cell’s function, better understanding the role of the DCLK1 gene/protein and its impact on inflammation and disease severity in colitis may improve our understanding of inflammation-associated intestinal tumorigenesis.

Summary of Results

Baseline characterization of the intestinal epithelial-specific knockout mouse model was in agreement with the results of manuscript 5.1. Following treatment with dextran sulfate sodium (DSS) to induce colitis, the knockout mice demonstrated decreased expression of epithelial/tight junction proteins and mucosal barrier integrity was compromised compared to control mice. To assess the consequences of the weakened barrier, direct and indirect inflammatory markers were assessed in the colon including myeloperoxidase (MPO) activity (a surrogate for neutrophil infiltration), cytokines, chemokines, and NF-kB tissue expression. Results revealed exacerbated inflammation and histological analysis showed decreased proliferation and increased apoptosis in Dclk1-knockout mice suggesting impaired regeneration after injury. Phenotypic assessment of the mice demonstrated more severe disease

as measured by fecal blood, diarrhea, and weight loss. Finally, Dclk1-knockout mice

35

had poor survival compared to control mice which did not die through the duration of the study.

Conclusion/Significance

These findings provide further support for the conclusions of manuscript 5.1 in a more relevant model of injury. The Dclk1+ tuft cell’s sensory role in response to inflammation including immune recruitment and hyperplasia to increase surveillance may be key to maintaining mucosal barrier integrity. Conversely, the necessity of resistance in this cell and the tendency towards hyperplasia in DNA-damaging conditions117,123,124 is a potential source of tumorigenesis that remains to be fully

elucidated.

5.3 Small-molecule kinase inhibitor LRRK2-IN-1 demonstrates potent activity against colorectal and pancreatic cancer through inhibition of doublecortin-like kinase 1

Background

Downregulation of DCLK1 gene expression with specific small-interfering RNAs (siRNAs) leads to reduced expression of pluripotency and EMT TFs130,131,140. On the functional level this manifests in decreased spheroid formation, cell invasion, cell migration, and other anti-cancer properties. In vivo delivery of DCLK1 siRNAs in nanoparticles results in highly decreased tumorigenesis in colon, pancreatic, and liver cancer xenografts131,132,140. Although these findings are promising, RNA interference technology has not been highly successful in the clinic to date. In contrast to RNA interference, kinase inhibition is an established therapeutic tool in oncology and widely used drugs (sunitinib, imatinib, erlotinib, etc.) function via this mechanism. Despite limited research into the role of DCLK1’s kinase activity in neurobiology, its potential as a kinase inhibitor target in cancer had not been previously assessed as inhibitors were not available. However, a kinome-wide

36

screen of the kinase inhibitor LRRK2-IN-1, which was developed to target the LRRK2 Parkinson’s disease kinase, demonstrated comparable affinity for DCLK1 (approximately 5 nM Kd)167.

Summary of Results

An in vitro kinase assay confirmed that LRRK2-IN-1 was a DCLK1 kinase inhibitor with a low nanomolar IC50. Treatment of pancreatic (AsPC-1) and colon (HCT116) cancer cells with LRRK2-IN-1 resulted in apoptosis and cell death with an EC50 of

1.69-1.73 µM as determined by MTT, LIVE/DEAD, and Caspase-3/7 assays.

Moreover, immunocytochemistry of treated AsPC-1 cells demonstrated increased phosphohistone H3 expression and cell cycle analysis showed potent dose-dependent G2/M arrest. LRRK2-IN-1 treatment also downregulated DCLK1 gene and protein expression which was hypothesized to be downstream of one of LRRK2-IN-1’s other targets, MAPK7 (ERK5). In support of this hypothesis, treatment with the MEK1/2 inhibitor U0126 led to decreased expression of DCLK1. LRRK2-IN-1 also demonstrated anti-metastatic effects including downregulation of EMT factor expression and inhibition of migration and invasion. To investigate if some of these anti-cancer effects could be ascribed to inhibition of DCLK1 kinase activity, we overexpressed wild-type DCLK1, kinase-dead DCLK1, and vector control (GFP) in cells and performed proliferation and colony formation assays. In both assays kinase-dead DCLK1 cells responded to LRRK2-IN-1 comparable to vector control cells while wild-type DCLK1 cells demonstrated resistance. Finally, to determine if LRRK2-IN-1 had potential as an in vivo anti-cancer agent, AsPC-1 xenografts were

treated with a dose of 100 mg/kg on a fixed schedule which resulted in statistically significant tumor growth inhibition.

Conclusion/Significance

This was the first study to assess a DCLK1 kinase inhibitor as an anti-cancer therapy. Although the results were promising and new research with LRRK2-IN-1 is forthcoming168, significant off-target affinity for ERK5 complicates assigning specific activities to DCLK1. We overcame this with the use of a kinase-dead DCLK1 mutant

37

with moderate results. The finding that MEK1/2 inhibitor U0126 treatment results in downregulation of DCLK1 suggests that DCLK1’s role in this pathway should be further investigated and other groups have found indications that may suggest a feedback loop between DCLK1, MAPK/ERK pathway, and KRAS142,155. Given the findings of this and other studies with both LRRK2-IN-1 and analogue XMD8-92, there is a major need for the development of DCLK1 kinase inhibitors with less off-target activity.

5.4 Doublecortin-like kinase 1 is elevated serologically in pancreatic ductal adenocarcinoma and widely expressed

on circulating tumor cells

Background

CSCs are known to influence drug and radiation resistance, recurrence and metastasis, and other tumor processes that impact survival and quality of life17,92,169,170. If it was possible to track the presence or activity of CSCs with direct or surrogate techniques, it may be possible to predict disease progression and survival. Moreover, monitoring CSCs using biomarkers may help determine whether or not therapy is necessary, the type and course of therapy, and the patient’s response to therapy. However, despite the potential benefits, biomarker development focused on CSCs has been limited.

Immunohistochemical biomarkers are the most common markers used in the clinic. Examples include the widely-used proliferation marker Ki-67, breast cancer subtyping markers ER, PR and HER2/Neu, and EGFR, which are used to assess susceptibility to targeted therapies. However, although tissue biomarkers are useful, blood biomarkers are more desirable in most cases because they are minimally invasive and allow the monitoring of therapeutic efficacy and progression on a regular basis. Carbohydrate antigen 19-9 (CA19-9) is the most widely used serum biomarker in pancreatic cancer. However, research has shown that CA19-9 may be

present in non-cancerous conditions which limits its use in the clinic171. Therefore,

38

a novel blood-based biomarker of sufficient specificity and accuracy in pancreatic cancer may have significant clinical utility.

Summary of Results

DCLK1 was upregulated in early (Stage I – II) pancreatic cancer but not in advanced disease. Receiver-operating characteristic (ROC) analysis of Stage I/II serum showed a modestly significant area under the curve of 0.74 suggesting that ELISA of serum DCLK1 was neither sensitive nor specific enough for use in this fashion. Western blotting of the serum samples confirmed those results. In comparison, CA19-9 serum levels were elevated in Stages II – IV but not Stage I in the same

samples. Since the ELISA and western blotting techniques used to measure serum DCLK1 were denaturing, a related analysis was performed on circulating tumor cells (CTCs) from KPCY mice (PdxCreKrasLSL-G12DTp53LSL-R172HR26YFP) which mimic PDAC progression and metastasis and include a reporter (YFP) to track dissemination. YFP+ CTCs from the KPCY mice expressed Dclk1 approximately 40 – 60% of the time, indicating a possibility that serum DCLK1 levels could be detected as a result of DCLK1+ CTCs. Moreover, tissue staining demonstrated DCLK1/YFP++ populations in metastatic tissues.

Conclusion/Significance

Overall, the findings of this study demonstrate that: 1) DCLK1 is detectable in human serum; 2) DCLK1 is upregulated in Stage I-II PDAC patient serum but indistinguishable from normal patients in Stage III-IV; 3) DCLK1 is expressed on PDAC CTCs; 4) DCLK1 may not be suitable as a diagnostic/prognostic serum biomarker for PDAC and the identification of relevant patient subgroups or better markers of its expression/activity are likely needed.

39

5.5 Survival of patients with gastrointestinal cancers can be predicted by a surrogate miRNA signature for cancer stem-like cells marked by doublecortin-like kinase 1

Background

The decoding of the human genome and resulting microarray and sequencing technologies provide an immense opportunity for advances in biomarker development. Whereas previous technologies were dependent on detecting a mutation or the expression of a single target, high throughput technologies make the development of multi-target biomarkers more feasible. Examples of these

technologies in clinical use include the Oncotype DX™, PAM50 (Prosigna™), and Mammaprint™ array technologies which are used to predict recurrence in hormone-receptor positive breast cancers172. Patients with a low score in these tests may be able to avoid prolonged chemotherapies which have unpleasant side-effects, including significant long-term risks in some cases.

Multigene biomarkers may be particularly suitable for detection of CSC activity. Since CSCs are defined by their ability to self-renew and produce all progeny that comprise the tumor, it is feasible that a tumor with low CSC marker expression may produce more progeny than a comparable tumor with high CSC marker expression. Counterintuitively, the tumor with low CSC marker expression could potentially have more CSC activity and therefore more aggressiveness and adaptability to treatment than its counterpart. The Cancer Genome Atlas (TCGA) offers a novel platform to test this concept in terms of the DCLK1+ GI CSC.

Summary of Results

DCLK1 was expressed and strongly correlated to EMT across gastrointestinal cancer types (colon, esophagus, pancreatic, stomach, and rectal). However, in liver it was only correlated with mesenchymal but not epithelial gene expression. As a result, and due to the lack of similarities with other GI tumors, we excluded liver from further analysis. With the remaining cancers, correlation analysis was performed for

DCLK1 and each gene in the transcriptome. The resulting consensus significant

40

genes between all cancer types were too numerous, which is consistent with overlapping functions for DCLK1+ cells across GI cancers, but impractical for deriving a biomarker. For further analysis, we turned to miRNAs which are stable analytes and limited in number. Correlation analysis for miRNAs identified 15 significantly associated species including miR-200a/b which were previously reported to be regulated by DCLK1. For 5 of the miRNAs, real-time PCR analysis of miRNA expression demonstrated the predicted effect in both directions (e.g. DCLK1 downregulation upregulates miR-200a, while DCLK1 does the reverse). Using standardized miRNA expression data, a signature score was calculated by subtracting the sum of downregulated miRNAs from the sum of upregulated miRNAs.

The signature predicted overall survival in colon, pancreatic, and stomach cancers and recurrence in colon and stomach cancers. Finally, subgroup analysis identified patient groups that might benefit from prognostic use of the biomarker including early stage colon cancer patients who have survival consistent with advanced disease when presenting with high signature expression, and stomach cancer patients receiving radiotherapy who have significantly reduced survival with high signature expression.

Conclusion/Significance

The overall significance of this study falls into three different categories. Primarily, it identifies a surrogate biomarker for DCLK1 and consequently its activity. The import of this arises from the concept of CSCs which counterintuitively may diminish their own signal by producing diverse progeny in the tumor. The use of a surrogate signature may overcome insufficient sensitivity and specificity inherent to a singular CSC marker. Another significant aspect of this study is the methodology used to derive the marker itself. Instead of focusing on a single tumor type, a signature was derived across 5 tumor types that are known or suspected to contain DCLK1+ sensory epithelial cells and in which DCLK1 is correlated to EMT. Similar procedures may be applied to biomarker development for other targets using data from other platforms. Finally, the signature predicted survival in a number of cancer

41

types (colon, gastric, pancreatic) and in specific patient subgroups (e.g. early-stage colon cancer), and further development may yield a viable biomarker.

5.6 Doublecortin-like kinase 1 is a broadly dysregulated target against epithelial-mesenchymal transition, focal adhesion, and stemness in clear cell renal carcinoma

Background

The urinary and gastrointestinal tracts are regions that are susceptible to microbial and chemical inflammation. Both tracts require extensive sensory, secretory, and

absorptive capabilities to maintain the sensitive balance required for homeostasis74,173,174. An analysis of data available from TCGA suggested that of all cancers, DCLK1 was most overexpressed in ccRCC. This finding was of interest as DCLK1 was previously thought not to be expressed in the kidney175, suggesting potential tumor specificity. Moreover, RCC is characterized by its slow growth, hypoxic microenvironment, EMT, and potent resistance to chemotherapy40,176-178 – characteristics linked to CSCs16,17,149,179,180. However, the role of CSCs in kidney cancer has only been investigated to a limited extent.

Summary of Results

Analysis of TCGA’s ccRCC dataset revealed overexpression of DCLK1 in tumors compared to normal and matched adjacent normal tissue. Immunohistochemical staining showed a similar result with increased expression of DCLK1 in Stages II-III compared to Stage I or normal tissue. Hypomethylation of DCLK1 promoter regions in tumors versus matched normal tissue supported epigenetic dysregulation as a likely source of overexpression. DCLK1 downregulation led to significantly reduced expression of EMT factors, migration, and invasion. During migration analysis, increased space between cells was noted in cells transfected with DCLK1 siRNA. Image thresholding calculations, staining for focal adhesions (vinculin) and cell structure (filamentous actin), and western blotting revealed a loss of focal adhesion

42

mediated by downregulation of focal adhesion kinase (FAK/PTK2). Molecular analysis demonstrated downregulation of pluripotency factors and stem cell marker ALDH1A1, and functional stemness was significantly impaired as evidenced by much fewer and smaller spheroids after DCLK1 knockdown.

Conclusion/Significance

This was the first study to investigate DCLK1 outside of the GI tract or neuroblastoma. Previous studies had claimed incorrectly that DCLK1 was not expressed in the normal kidney, but immunohistochemical tissue staining showed it present in some tubules where there is some potential for a sensory role, and in the

kidney stroma. The results of this study demonstrate that DCLK1 can have a comparable oncogenic role (e.g. EMT, stemness, etc.) to its GI cancer role in other tumor types, and necessitate further study of whether DCLK1 may also mark CSCs or CSC-like cells in kidney and other urinary tract cancers.

43

6 Conclusions and Future Directions The work presented here covers novel findings in four specific areas of research and comprises a body of work focused on DCLK1. Since the publication of these works, DCLK1 has emerged as a prominent CSC-related marker in many cancers

including non-GI tract cancers. Although Dclk1 was discredited as a normal or reserve stem cell marker at homeostasis or in injury, the work in gut biology outlined in this thesis (5.1 & 5.2)181,182 helped establish Dclk1 as a sensory factor and regulator of the response to epithelial injury – a finding that provides additional support for studies that show the Dclk1+ tuft cell orchestrating the immune response to parasitic invasion123-125 and implicate Dclk1 in inflammation-mediated cancer10,142. In the area of cancer therapies, the work presented here was the first to demonstrate the potential for DCLK1-targeted kinase inhibitors against gastrointestinal cancers183 (5.3). Improving on these existing inhibitors for DCLK1 selectivity may be a promising avenue for the development of novel therapies with significant clinical potential. Success in this area may also necessitate further development of DCLK1-based biomarkers. The data reported here suggest that DCLK1 serological levels may not be useful as clinical diagnostic or prognostic biomarkers in GI tract cancers184 (5.4), though exploring these levels in select patient subgroups including early-stage patients may yet prove promising. However, the work presented here utilizing RNA-Seq data shows that surrogate multi-gene biomarkers may be able to fill this role185 (5.5). Finally, findings of DCLK1 as an overexpressed drug target and potential CSC marker in kidney cancer186 may be relevant to research in RCC and other non-GI cancers (5.6). Although previous work had suggested oncogenic

properties of DCLK1 in neuroblastoma, this work was the first to look at its role in EMT and stemness in non-GI cancers. Overall, DCLK1 is rapidly emerging as a deleterious factor in many non-GI cancers133,143,145,155,187,188.

On a technological level, advances in single-cell RNA sequencing technologies have already begun to revolutionize our understanding of tuft cells, revealing distinct origins for the cell type in the colon and small intestine189, suggesting the presence of a separate immune-regulatory tuft cell type190, and providing an atlas that can be used for confirming results and cross-referencing other data. Whether these and

44

other murine studies will have the expected relevance in human systems will likely be revealed as the Human Cell Atlas begins progress in characterizing the cells of the GI tract under conditions of homeostasis, infection, and neoplasia. These advances may also improve our understanding of CTCs which DCLK1 marks in several cancers184,188,191-193.

Although studies have proven the potential of epigenetic biomarkers, only a single marker is currently approved in the US (SEPT9 for CRC diagnosis)194. Analyses of DCLK1 methylation suggests that it may have potential as a diagnostic

biomarker186,188,195-197, but the hypermethylation of its a-promoter has also been a

source of controversy as it relates to the nature of DCLK1’s isoforms. This has led some to propose a tumor suppressor function for some isoforms and an oncogenic/CSC function for others195. Studies to methodically characterize the activity of each isoform will ultimately be necessary to decode DCLK1’s interactions and its upstream and downstream signaling. It is notable that virtually all studies in

the GI tract or cancer have focused on a-promoter isoforms, and the activity of the

b-promoter driven isoforms are not well known. To overcome this issue, the

development of a b-promoter DCLK1-driven Cre or comparable mouse model may

be necessary.

Finally, although DCLK1 is a CSC marker in mouse models, it will be important to determine how applicable this finding is to human patients. Based on findings using the stem cell/CSC marker LGR5198, the development of DCLK1-specific organoid culture from human colorectal, pancreatic, and other tumors may be feasible and

could provide advantages over other models, or be used as a personalized therapy platform. However, the most important work to determine the relevance of DCLK1 to human cancer is likely the development of DCLK1-targeted therapies. This may be in the form of monoclonal antibodies that take advantage of the unusual membrane localization of some of its isoforms199, specific kinase inhibitors, novel gene therapies such as chimeric antigen receptor T cells (CAR-T), or other modalities that may improve outcomes in patients.

45

7 References

1 Tomasetti, C., Li, L. & Vogelstein, B. Stem cell divisions, somatic mutations, cancer etiology, and cancer prevention. Science 355, 1330-1334, doi:10.1126/science.aaf9011 (2017).

2 Paget, S. The distribution of secondary growths in cancer of the breast. 1889. Cancer Metastasis Rev 8, 98-101 (1989).

3 DeVita, V. T., Jr. & Rosenberg, S. A. Two hundred years of cancer research. N Engl J Med 366, 2207-2214, doi:10.1056/NEJMra1204479 (2012).

4 Hanahan, D. & Weinberg, R. A. Hallmarks of cancer: the next generation. Cell 144, 646-674, doi:10.1016/j.cell.2011.02.013 (2011).

5 Kawanishi, S., Ohnishi, S., Ma, N., Hiraku, Y. & Murata, M. Crosstalk between DNA Damage and Inflammation in the Multiple Steps of Carcinogenesis. Int J Mol Sci 18, doi:10.3390/ijms18081808 (2017).

6 Kalisperati, P. et al. Inflammation, DNA Damage, Helicobacter pylori and Gastric Tumorigenesis. Front Genet 8, 20, doi:10.3389/fgene.2017.00020 (2017).

7 Samraj, A. N. et al. A red meat-derived glycan promotes inflammation and cancer progression. Proc Natl Acad Sci U S A 112, 542-547, doi:10.1073/pnas.1417508112 (2015).

8 Huang, S. et al. p53 modulates acquired resistance to EGFR inhibitors and radiation. Cancer research 71, 7071-7079, doi:10.1158/0008-5472.CAN-11-0128 (2011).

9 Aggarwal, P. et al. Nuclear accumulation of cyclin D1 during S phase inhibits Cul4-dependent Cdt1 proteolysis and triggers p53-dependent DNA rereplication. Genes Dev 21, 2908-2922 (2007).

10 Westphalen, C. B. et al. Long-lived intestinal tuft cells serve as colon cancer-initiating cells. J Clin Invest 124, 1283-1295, doi:10.1172/JCI73434 (2014).

11 Blanpain, C. Tracing the cellular origin of cancer. Nat Cell Biol 15, 126-134, doi:10.1038/ncb2657 (2013).

12 Nakanishi, Y. et al. Dclk1 distinguishes between tumor and normal stem cells in the intestine. Nature genetics 45, 98-9103 (2013).

13 Schepers, A. G. et al. Lineage tracing reveals Lgr5+ stem cell activity in mouse intestinal adenomas. Science 337, 730-735, doi:10.1126/science.1224676 (2012).

14 Barker, N. et al. Crypt stem cells as the cells-of-origin of intestinal cancer. Nature 457, 608-611, doi:10.1038/nature07602 (2009).

15 Shibue, T. & Weinberg, R. A. EMT, CSCs, and drug resistance: the mechanistic link and clinical implications. Nat Rev Clin Oncol 14, 611-629, doi:10.1038/nrclinonc.2017.44 (2017).

16 Wang, S. S., Jiang, J., Liang, X. H. & Tang, Y. L. Links between cancer stem cells and epithelial-mesenchymal transition. Onco Targets Ther 8, 2973-2980, doi:10.2147/OTT.S91863 (2015).

17 Mitra, A., Mishra, L. & Li, S. EMT, CTCs and CSCs in tumor relapse and drug-resistance. Oncotarget 6, 10697-10711, doi:10.18632/oncotarget.4037 (2015).

18 Jolly, M. K., Ware, K. E., Gilja, S., Somarelli, J. A. & Levine, H. EMT and MET: necessary or permissive for metastasis? Mol Oncol 11, 755-769, doi:10.1002/1878-0261.12083 (2017).

46

19 Noman, M. Z. et al. The immune checkpoint ligand PD-L1 is upregulated in EMT-activated human breast cancer cells by a mechanism involving ZEB-1 and miR-200. Oncoimmunology 6, e1263412, doi:10.1080/2162402X.2016.1263412 (2017).

20 Tsutsumi, S. et al. PD-L1 expression at tumor invasive front is associated with EMT and poor prognosis in esophageal squamous cell carcinoma. Cancer Sci, doi:10.1111/cas.13237 (2017).

21 Lou, Y. et al. Epithelial-Mesenchymal Transition Is Associated with a Distinct Tumor Microenvironment Including Elevation of Inflammatory Signals and Multiple Immune Checkpoints in Lung Adenocarcinoma. Clinical cancer research : an official journal of the American Association for Cancer Research 22, 3630-3642, doi:10.1158/1078-0432.CCR-15-1434 (2016).

22 Kim, S. et al. PD-L1 expression is associated with epithelial-to-mesenchymal transition in adenocarcinoma of the lung. Hum Pathol 58, 7-14, doi:10.1016/j.humpath.2016.07.007 (2016).

23 Ock, C. Y. et al. PD-L1 expression is associated with epithelial-mesenchymal transition in head and neck squamous cell carcinoma. Oncotarget 7, 15901-15914, doi:10.18632/oncotarget.7431 (2016).

24 Alsuliman, A. et al. Bidirectional crosstalk between PD-L1 expression and epithelial to mesenchymal transition: significance in claudin-low breast cancer cells. Molecular cancer 14, 149, doi:10.1186/s12943-015-0421-2 (2015).

25 Fantozzi, A. et al. VEGF-mediated angiogenesis links EMT-induced cancer stemness to tumor initiation. Cancer Res 74, 1566-1575, doi:10.1158/0008-5472.CAN-13-1641 (2014).

26 Chiba, K. et al. Cancer-associated TERT promoter mutations abrogate telomerase silencing. Elife 4, doi:10.7554/eLife.07918 (2015).

27 Hoffmeyer, K. et al. Wnt/beta-catenin signaling regulates telomerase in stem cells and cancer cells. Science 336, 1549-1554, doi:10.1126/science.1218370 (2012).

28 Steele, S. P., Melchor, S. J. & Petri, W. A., Jr. Tuft Cells: New Players in Colitis. Trends Mol Med 22, 921-924, doi:10.1016/j.molmed.2016.09.005 (2016).

29 Cantor, J. R. & Sabatini, D. M. Cancer cell metabolism: one hallmark, many faces. Cancer Discov 2, 881-898, doi:10.1158/2159-8290.CD-12-0345 (2012).

30 Vander Heiden, M. G., Cantley, L. C. & Thompson, C. B. Understanding the Warburg effect: the metabolic requirements of cell proliferation. Science 324, 1029-1033, doi:10.1126/science.1160809 (2009).

31 Basu, N. et al. Telomeres and telomere dynamics: relevance to cancers of the GI tract. Expert Rev Gastroenterol Hepatol 7, 733-748, doi:10.1586/17474124.2013.848790 (2013).

32 Grivennikov, S. I., Greten, F. R. & Karin, M. Immunity, inflammation, and cancer. Cell 140, 883-899, doi:10.1016/j.cell.2010.01.025 (2010).

33 Motzer, R. J. et al. Nivolumab versus Everolimus in Advanced Renal-Cell Carcinoma. N Engl J Med 373, 1803-1813, doi:10.1056/NEJMoa1510665 (2015).

34 Lamouille, S., Xu, J. & Derynck, R. Molecular mechanisms of epithelial-mesenchymal transition. Nat Rev Mol Cell Biol 15, 178-196, doi:10.1038/nrm3758 (2014).

35 Brabletz, S. et al. The ZEB1/miR-200 feedback loop controls Notch signalling in cancer cells. EMBO J 30, 770-782, doi:10.1038/emboj.2010.349 (2011).

47

36 Wellner, U. et al. The EMT-activator ZEB1 promotes tumorigenicity by repressing stemness-inhibiting microRNAs. Nat Cell Biol 11, 1487-1495, doi:10.1038/ncb1998 (2009).

37 Kim, S. et al. PD-L1 expression is associated with epithelial-to-mesenchymal transition in adenocarcinoma of the lung. Hum Pathol, doi:10.1016/j.humpath.2016.07.007 (2016).

38 Krantz, S. B., Shields, M. A., Dangi-Garimella, S., Munshi, H. G. & Bentrem, D. J. Contribution of epithelial-to-mesenchymal transition and cancer stem cells to pancreatic cancer progression. J Surg Res 173, 105-112, doi:S0022-4804(11)00749-9 [pii] 10.1016/j.jss.2011.09.020 (2012).

39 Rhim, A. D. et al. EMT and dissemination precede pancreatic tumor formation. Cell 148, 349-361, doi:10.1016/j.cell.2011.11.025 (2012).

40 Harada, K., Miyake, H., Kusuda, Y. & Fujisawa, M. Expression of epithelial-mesenchymal transition markers in renal cell carcinoma: impact on prognostic outcomes in patients undergoing radical nephrectomy. BJU Int 110, E1131-1137, doi:10.1111/j.1464-410X.2012.11297.x (2012).

41 Hoeben, A. et al. Vascular endothelial growth factor and angiogenesis. Pharmacol Rev 56, 549-580, doi:10.1124/pr.56.4.3 (2004).

42 Angell, T. E., Lechner, M. G., Jang, J. K., LoPresti, J. S. & Epstein, A. L. MHC class I loss is a frequent mechanism of immune escape in papillary thyroid cancer that is reversed by interferon and selumetinib treatment in vitro. Clin Cancer Res 20, 6034-6044, doi:10.1158/1078-0432.CCR-14-0879 (2014).

43 Mantovani, A., Marchesi, F., Malesci, A., Laghi, L. & Allavena, P. Tumour-associated macrophages as treatment targets in oncology. Nat Rev Clin Oncol 14, 399-416, doi:10.1038/nrclinonc.2016.217 (2017).

44 Yu, Y. et al. Cancer-associated fibroblasts induce epithelial-mesenchymal transition of breast cancer cells through paracrine TGF-beta signalling. Br J Cancer 110, 724-732, doi:10.1038/bjc.2013.768 (2014).

45 Zhu, Q., Gao, R., Wu, W. & Qin, H. The role of gut microbiota in the pathogenesis of colorectal cancer. Tumour Biol 34, 1285-1300, doi:10.1007/s13277-013-0684-4 (2013).

46 Gerbe, F. et al. Distinct ATOH1 and Neurog3 requirements define tuft cells as a new secretory cell type in the intestinal epithelium. J Cell Biol 192, 767-780, doi:10.1083/jcb.201010127 (2011).

47 Tian, H. et al. A reserve stem cell population in small intestine renders Lgr5-positive cells dispensable. Nature 478, 255-259, doi:10.1038/nature10408 (2011).

48 Siegel, R. L., Miller, K. D. & Jemal, A. Cancer Statistics, 2017. CA Cancer J Clin 67, 7-30, doi:10.3322/caac.21387 (2017).

49 Cancer Statistics for the UK, <http://www.cancerresearchuk.org/health-professional/cancer-statistics-for-the-uk> (

50 Kuipers, E. J. et al. Colorectal cancer. Nat Rev Dis Primers 1, 15065, doi:10.1038/nrdp.2015.65 (2015).

51 Luchtenborg, M. et al. APC mutations in sporadic colorectal carcinomas from The Netherlands Cohort Study. Carcinogenesis 25, 1219-1226, doi:10.1093/carcin/bgh117 (2004).

52 Miyoshi, Y. et al. Germ-line mutations of the APC gene in 53 familial adenomatous polyposis patients. Proceedings of the National Academy of Sciences of the United States of America 89, 4452-4456 (1992).

48

53 Powell, S. M. et al. APC mutations occur early during colorectal tumorigenesis. Nature 359, 235-237, doi:10.1038/359235a0 (1992).

54 Kodaz, H. et al. Association between specific KRAS mutations and the clinicopathological characteristics of colorectal tumors. Molecular and clinical oncology 3, 179-184, doi:10.3892/mco.2014.448 (2015).

55 Xu, J. M. et al. KRAS mutations in tumor tissue and plasma by different assays predict survival of patients with metastatic colorectal cancer. Journal of experimental & clinical cancer research : CR 33, 104, doi:10.1186/PREACCEPT-2062328391321780 (2014).

56 Phipps, A. I. et al. KRAS-mutation status in relation to colorectal cancer survival: the joint impact of correlated tumour markers. British journal of cancer 108, 1757-1764, doi:10.1038/bjc.2013.118 (2013).

57 Cancer Genome Atlas, N. Comprehensive molecular characterization of human colon and rectal cancer. Nature 487, 330-337, doi:10.1038/nature11252 (2012).

58 Moon, B. S. et al. Role of oncogenic K-Ras in cancer stem cell activation by aberrant Wnt/beta-catenin signaling. Journal of the National Cancer Institute 106, djt373, doi:10.1093/jnci/djt373 (2014).

59 Fan, F. et al. Overexpression of Snail induces epithelial-mesenchymal transition and a cancer stem cell-like phenotype in human colorectal cancer cells. Cancer Med 1, 5-16, doi:10.1002/cam4.4 (2012).

60 Wang, Y. et al. Rapamycin inhibits FBXW7 loss-induced epithelial-mesenchymal transition and cancer stem cell-like characteristics in colorectal cancer cells. Biochem Biophys Res Commun 434, 352-356, doi:10.1016/j.bbrc.2013.03.077 (2013).

61 Salt, M. B., Bandyopadhyay, S. & McCormick, F. Epithelial-to-mesenchymal transition rewires the molecular path to PI3K-dependent proliferation. Cancer discovery 4, 186-199, doi:10.1158/2159-8290.CD-13-0520 (2014).

62 Mani, S. A. et al. The epithelial-mesenchymal transition generates cells with properties of stem cells. Cell 133, 704-715, doi:S0092-8674(08)00444-3 [pii] 10.1016/j.cell.2008.03.027 (2008).

63 Jang, B. G., Lee, B. L. & Kim, W. H. Intestinal Stem Cell Markers in the Intestinal Metaplasia of Stomach and Barrett's Esophagus. PLoS One 10, e0127300, doi:10.1371/journal.pone.0127300 (2015).

64 Souza, R. F., Krishnan, K. & Spechler, S. J. Acid, bile, and CDX: the ABCs of making Barrett's metaplasia. Am J Physiol Gastrointest Liver Physiol 295, G211-218, doi:10.1152/ajpgi.90250.2008 (2008).

65 Li, T. T. et al. Evaluation of epithelial-mesenchymal transitioned circulating tumor cells in patients with resectable gastric cancer: Relevance to therapy response. World J Gastroenterol 21, 13259-13267, doi:10.3748/wjg.v21.i47.13259 (2015).

66 Wen, J. et al. The epithelial-mesenchymal transition phenotype of metastatic lymph nodes impacts the prognosis of esophageal squamous cell carcinoma patients. Oncotarget 7, 37581-37588, doi:10.18632/oncotarget.9036 (2016).

67 Kleeff, J. et al. Pancreatic cancer. Nat Rev Dis Primers 2, 16022, doi:10.1038/nrdp.2016.22 (2016).

68 Chakraborty, S. & Singh, S. Surgical resection improves survival in pancreatic cancer patients without vascular invasion- a population based study. Ann Gastroenterol 26, 346-352 (2013).

49

69 Ilic, M. & Ilic, I. Epidemiology of pancreatic cancer. World J Gastroenterol 22, 9694-9705, doi:10.3748/wjg.v22.i44.9694 (2016).

70 Bailey, J. M. et al. DCLK1 marks a morphologically distinct subpopulation of cells with stem cell properties in preinvasive pancreatic cancer. Gastroenterology 146, 245-256, doi:10.1053/j.gastro.2013.09.050 (2014).

71 Habbe, N. et al. Spontaneous induction of murine pancreatic intraepithelial neoplasia (mPanIN) by acinar cell targeting of oncogenic Kras in adult mice. Proc Natl Acad Sci U S A 105, 18913-18918, doi:10.1073/pnas.0810097105 (2008).

72 Giroux, V. & Rustgi, A. K. Metaplasia: tissue injury adaptation and a precursor to the dysplasia-cancer sequence. Nat Rev Cancer 17, 594-604, doi:10.1038/nrc.2017.68 (2017).

73 Tuveson, D. A. & Neoptolemos, J. P. Understanding metastasis in pancreatic cancer: a call for new clinical approaches. Cell 148, 21-23, doi:S0092-8674(11)01574-1 [pii] 10.1016/j.cell.2011.12.021 (2012).

74 Aragon, I. M. et al. The Urinary Tract Microbiome in Health and Disease. Eur Urol Focus, doi:10.1016/j.euf.2016.11.001 (2016).

75 Vermeulen, S. H. et al. Recurrent urinary tract infection and risk of bladder cancer in the Nijmegen bladder cancer study. Br J Cancer 112, 594-600, doi:10.1038/bjc.2014.601 (2015).

76 Pascual, D. & Borque, A. Epidemiology of kidney cancer. Adv Urol, 782381, doi:10.1155/2008/782381 (2008).

77 Hsieh, J. J. et al. Renal cell carcinoma. Nat Rev Dis Primers 3, 17009, doi:10.1038/nrdp.2017.9 (2017).

78 Mikami, S. et al. Expression of TNF-alpha and CD44 is implicated in poor prognosis, cancer cell invasion, metastasis and resistance to the sunitinib treatment in clear cell renal cell carcinomas. Int J Cancer 136, 1504-1514, doi:10.1002/ijc.29137 (2015).

79 Wang, Y. et al. PD-L1 induces epithelial-to-mesenchymal transition via activating SREBP-1c in renal cell carcinoma. Medical oncology 32, 212, doi:10.1007/s12032-015-0655-2 (2015).

80 Cancer Genome Atlas Research, N. Comprehensive molecular characterization of clear cell renal cell carcinoma. Nature 499, 43-49, doi:10.1038/nature12222 (2013).

81 Houghton, J., Morozov, A., Smirnova, I. & Wang, T. C. Stem cells and cancer. Semin Cancer Biol 17, 191-203, doi:10.1016/j.semcancer.2006.04.003 (2007).

82 Daley, G. Q. Stem cells and the evolving notion of cellular identity. Philos Trans R Soc Lond B Biol Sci 370, 20140376, doi:10.1098/rstb.2014.0376 (2015).

83 Wahlestedt, M., Pronk, C. J. & Bryder, D. Concise review: hematopoietic stem cell aging and the prospects for rejuvenation. Stem Cells Transl Med 4, 186-194, doi:10.5966/sctm.2014-0132 (2015).

84 Bryder, D., Rossi, D. J. & Weissman, I. L. Hematopoietic stem cells: the paradigmatic tissue-specific stem cell. Am J Pathol 169, 338-346, doi:10.2353/ajpath.2006.060312 (2006).

85 Goodell, M. A., Nguyen, H. & Shroyer, N. Somatic stem cell heterogeneity: diversity in the blood, skin and intestinal stem cell compartments. Nat Rev Mol Cell Biol 16, 299-309, doi:10.1038/nrm3980 (2015).

50

86 Clevers, H. The intestinal crypt, a prototype stem cell compartment. Cell 154, 274-284, doi:10.1016/j.cell.2013.07.004 (2013).

87 Barriga, F. M. et al. Mex3a Marks a Slowly Dividing Subpopulation of Lgr5+ Intestinal Stem Cells. Cell Stem Cell 20, 801-816 e807, doi:10.1016/j.stem.2017.02.007 (2017).

88 Sugihara, E. & Saya, H. Complexity of cancer stem cells. Int J Cancer 132, 1249-1259, doi:10.1002/ijc.27961 (2013).

89 Fearon, E. R., Hamilton, S. R. & Vogelstein, B. Clonal analysis of human colorectal tumors. Science 238, 193-197 (1987).

90 Bradshaw, A. et al. Cancer Stem Cell Hierarchy in Glioblastoma Multiforme. Front Surg 3, 21, doi:10.3389/fsurg.2016.00021 (2016).

91 Lapidot, T. et al. A cell initiating human acute myeloid leukaemia after transplantation into SCID mice. Nature 367, 645-648, doi:10.1038/367645a0 (1994).

92 Fulawka, L., Donizy, P. & Halon, A. Cancer stem cells--the current status of an old concept: literature review and clinical approaches. Biol Res 47, 66, doi:10.1186/0717-6287-47-66 (2014).

93 Bonnet, D. & Dick, J. E. Human acute myeloid leukemia is organized as a hierarchy that originates from a primitive hematopoietic cell. Nat Med 3, 730-737 (1997).

94 Al-Hajj, M., Wicha, M. S., Benito-Hernandez, A., Morrison, S. J. & Clarke, M. F. Prospective identification of tumorigenic breast cancer cells. Proc Natl Acad Sci U S A 100, 3983-3988, doi:10.1073/pnas.0530291100 (2003).

95 Friedmann-Morvinski, D. & Verma, I. M. Dedifferentiation and reprogramming: origins of cancer stem cells. EMBO Rep 15, 244-253, doi:10.1002/embr.201338254 (2014).

96 Omori, Y. et al. Expression and chromosomal localization of KIAA0369, a putative kinase structurally related to Doublecortin. J Hum Genet 43, 169-177, doi:10.1007/s100380050063 (1998).

97 Sossey-Alaoui, K. & Srivastava, A. K. DCAMKL1, a brain-specific transmembrane protein on 13q12.3 that is similar to doublecortin (DCX). Genomics 56, 121-126, doi:10.1006/geno.1998.5718 (1999).

98 Kornblihtt, A. R. CTCF: from insulators to alternative splicing regulation. Cell Res 22, 450-452, doi:10.1038/cr.2012.22 (2012).

99 Cancer Genome Atlas Research, N. et al. Comprehensive Molecular Characterization of Papillary Renal-Cell Carcinoma. N Engl J Med 374, 135-145, doi:10.1056/NEJMoa1505917 (2016).

100 Cancer Genome Atlas Research Network. Electronic address, a. a. d. h. e. & Cancer Genome Atlas Research, N. Integrated Genomic Characterization of Pancreatic Ductal Adenocarcinoma. Cancer Cell 32, 185-203 e113, doi:10.1016/j.ccell.2017.07.007 (2017).

101 Cancer Genome Atlas Research, N. Comprehensive molecular profiling of lung adenocarcinoma. Nature 511, 543-550, doi:10.1038/nature13385 (2014).

102 Cancer Genome Atlas Research, N. Comprehensive genomic characterization of squamous cell lung cancers. Nature 489, 519-525, doi:10.1038/nature11404 (2012).

103 Goldman, M. et al. The UCSC Cancer Genomics Browser: update 2015. Nucleic Acids Res 43, D812-817, doi:10.1093/nar/gku1073 (2015).

51

104 Lin, P. T., Gleeson, J. G., Corbo, J. C., Flanagan, L. & Walsh, C. A. DCAMKL1 encodes a protein kinase with homology to doublecortin that regulates microtubule polymerization. J Neurosci 20, 9152-9161 (2000).

105 Burgess, H. A. & Reiner, O. Cleavage of doublecortin-like kinase by calpain releases an active kinase fragment from a microtubule anchorage domain. J Biol Chem 276, 36397-36403, doi:10.1074/jbc.M105153200 (2001).

106 Burgess, H. A. & Reiner, O. Alternative splice variants of doublecortin-like kinase are differentially expressed and have different kinase activities. J Biol Chem 277, 17696-17705, doi:10.1074/jbc.M111981200 (2002).

107 Shang, L., Kwon, Y. G., Nandy, S., Lawrence, D. S. & Edelman, A. M. Catalytic and regulatory domains of doublecortin kinase-1. Biochemistry 42, 2185-2194, doi:10.1021/bi026913i (2003).

108 Shu, T. et al. Doublecortin-like kinase controls neurogenesis by regulating mitotic spindles and M phase progression. Neuron 49, 25-39, doi:10.1016/j.neuron.2005.10.039 (2006).

109 Koizumi, H., Tanaka, T. & Gleeson, J. G. Doublecortin-like kinase functions with doublecortin to mediate fiber tract decussation and neuronal migration. Neuron 49, 55-66, doi:10.1016/j.neuron.2005.10.040 (2006).

110 Deuel, T. A. et al. Genetic interactions between doublecortin and doublecortin-like kinase in neuronal migration and axon outgrowth. Neuron 49, 41-53, doi:10.1016/j.neuron.2005.10.038 (2006).

111 Weimer, J. M. & Anton, E. S. Doubling up on microtubule stabilizers: synergistic functions of doublecortin-like kinase and doublecortin in the developing cerebral cortex. Neuron 49, 3-4, doi:10.1016/j.neuron.2005.12.016 (2006).

112 Tanaka, T., Koizumi, H. & Gleeson, J. G. The doublecortin and doublecortin-like kinase 1 genes cooperate in murine hippocampal development. Cereb Cortex 16 Suppl 1, i69-73, doi:10.1093/cercor/bhk005 (2006).

113 Giannakis, M. et al. Molecular properties of adult mouse gastric and intestinal epithelial progenitors in their niches. J Biol Chem 281, 11292-11300, doi:10.1074/jbc.M512118200 (2006).

114 Barker, N., van de Wetering, M. & Clevers, H. The intestinal stem cell. Genes Dev 22, 1856-1864, doi:10.1101/gad.1674008 (2008).

115 May, R. et al. Identification of a novel putative gastrointestinal stem cell and adenoma stem cell marker, doublecortin and CaM kinase-like-1, following radiation injury and in adenomatous polyposis coli/multiple intestinal neoplasia mice. Stem Cells 26, 630-637, doi:10.1634/stemcells.2007-0621 (2008).

116 May, R. et al. Identification of a novel putative pancreatic stem/progenitor cell marker DCAMKL-1 in normal mouse pancreas. Am J Physiol Gastrointest Liver Physiol 299, G303-310, doi:10.1152/ajpgi.00146.2010 (2010).

117 Saqui-Salces, M. et al. Gastric tuft cells express DCLK1 and are expanded in hyperplasia. Histochem Cell Biol 136, 191-204, doi:10.1007/s00418-011-0831-1 (2011).

118 Gerbe, F., Brulin, B., Makrini, L., Legraverend, C. & Jay, P. DCAMKL-1 expression identifies Tuft cells rather than stem cells in the adult mouse intestinal epithelium. Gastroenterology 137, 2179-2180; author reply 2180-2171, doi:10.1053/j.gastro.2009.06.072 (2009).

119 Uhlen, M. et al. Proteomics. Tissue-based map of the human proteome. Science 347, 1260419, doi:10.1126/science.1260419 (2015).

52

120 Nam, K. T. et al. Gastric tumor development in Smad3-deficient mice initiates from forestomach/glandular transition zone along the lesser curvature. Lab Invest 92, 883-895, doi:10.1038/labinvest.2012.47 (2012).

121 Delgiorno, K. E. et al. Identification and manipulation of biliary metaplasia in pancreatic tumors. Gastroenterology 146, 233-244 e235, doi:10.1053/j.gastro.2013.08.053 (2014).

122 Kokrashvili, Z. et al. Release of endogenous opioids from duodenal enteroendocrine cells requires Trpm5. Gastroenterology 137, 598-606, 606 e591-592, doi:10.1053/j.gastro.2009.02.070 (2009).

123 von Moltke, J., Ji, M., Liang, H. E. & Locksley, R. M. Tuft-cell-derived IL-25 regulates an intestinal ILC2-epithelial response circuit. Nature 529, 221-225, doi:10.1038/nature16161 (2016).

124 Gerbe, F. et al. Intestinal epithelial tuft cells initiate type 2 mucosal immunity to helminth parasites. Nature 529, 226-230, doi:10.1038/nature16527 (2016).

125 Howitt, M. R. et al. Tuft cells, taste-chemosensory cells, orchestrate parasite type 2 immunity in the gut. Science 351, 1329-1333, doi:10.1126/science.aaf1648 (2016).

126 Mladenova, D. & Kohonen-Corish, M. R. Review: Mouse models of inflammatory bowel disease--insights into the mechanisms of inflammation-associated colorectal cancer. In Vivo 26, 627-646 (2012).

127 Greer, J. B. & Whitcomb, D. C. Inflammation and pancreatic cancer: an evidence-based review. Curr Opin Pharmacol 9, 411-418, doi:10.1016/j.coph.2009.06.011 (2009).

128 Markowitz, S. D. & Bertagnolli, M. M. Molecular origins of cancer: Molecular basis of colorectal cancer. N Engl J Med 361, 2449-2460, doi:10.1056/NEJMra0804588 (2009).

129 Sureban, S. M. et al. DCLK1 regulates pluripotency and angiogenesis via microRNA dependent mechanisms in pancreatic cancer. PloS one In Press (2013).

130 Sureban, S. M. et al. DCAMKL-1 regulates epithelial-mesenchymal transition in human pancreatic cells through a miR-200a-dependent mechanism. Cancer research 71, 2328-2338, doi:0008-5472.CAN-10-2738 [pii] 10.1158/0008-5472.CAN-10-2738 (2011).

131 Sureban, S. M. et al. Nanoparticle-based delivery of siDCAMKL-1 increases microRNA-144 and inhibits colorectal cancer tumor growth via a Notch-1 dependent mechanism. J Nanobiotechnology 9, 40-40 (2011).

132 Sureban, S. M. et al. Selective blockade of DCAMKL-1 results in tumor growth arrest by a Let-7a MicroRNA-dependent mechanism. Gastroenterology 137, 649-659, 659 e641-642, doi:S0016-5085(09)00748-3 [pii] 10.1053/j.gastro.2009.05.004 (2009).

133 Kadletz, L. et al. Role of cancer stem-cell marker doublecortin-like kinase 1 in head and neck squamous cell carcinoma. Oral Oncol 67, 109-118, doi:10.1016/j.oraloncology.2017.02.007 (2017).

134 Li, L. & Bellows, C. F. Doublecortin-like kinase 1 exhibits cancer stem cell-like characteristics in a human colon cancer cell line. Chin J Cancer Res 25, 134-142, doi:10.3978/j.issn.1000-9604.2013.03.02 (2013).

135 Bailey, J. M. et al. DCLK1 Marks a Morphologically Distinct Subpopulation of Cells with Stem Cell Properties in Pre-invasive Pancreatic Cancer. Gastroenterology, doi:10.1053/j.gastro.2013.09.050 (2013).

53

136 Nakanishi, Y. et al. Dclk1 distinguishes between tumor and normal stem cells in the intestine. Nat Genet 45, 98-103, doi:10.1038/ng.2481 (2013).

137 Tomkovich, S. et al. Locoregional Effects of Microbiota in a Preclinical Model of Colon Carcinogenesis. Cancer Res 77, 2620-2632, doi:10.1158/0008-5472.CAN-16-3472 (2017).

138 Risberg, K. et al. Pro-survival responses to the dual inhibition of anti-apoptotic Bcl-2 family proteins and mTOR-mediated signaling in hypoxic colorectal carcinoma cells. BMC Cancer 16, 531, doi:10.1186/s12885-016-2600-y (2016).

139 Okamoto, K., Zaanan, A., Kawakami, H., Huang, S. & Sinicrope, F. A. Reversal of Mutant KRAS-Mediated Apoptosis Resistance by Concurrent Noxa/Bik Induction and Bcl-2/Bcl-xL Antagonism in Colon Cancer Cells. Mol Cancer Res 13, 659-669, doi:10.1158/1541-7786.MCR-14-0476 (2015).

140 Sureban, S. M. et al. DCLK1 regulates pluripotency and angiogenic factors via microRNA-dependent mechanisms in pancreatic cancer. PloS one 8, e73940, doi:10.1371/journal.pone.0073940 (2013).

141 Zhou, B. et al. Exome sequencing and digital PCR analyses reveal novel mutated genes related to the metastasis of pancreatic ductal adenocarcinoma. Cancer Biol Ther 13, 871-879, doi:20839 [pii] 10.4161/cbt.20839 (2012).

142 Westphalen, C. B. et al. Dclk1 Defines Quiescent Pancreatic Progenitors that Promote Injury-Induced Regeneration and Tumorigenesis. Cell Stem Cell 18, 441-455, doi:10.1016/j.stem.2016.03.016 (2016).

143 Shan, C. et al. miR-448 is a novel prognostic factor of lung squamous cell carcinoma and regulates cells growth and metastasis by targeting DCLK1. Biomed Pharmacother 89, 1227-1234, doi:10.1016/j.biopha.2017.02.017 (2017).

144 Quan, X. et al. Potential Mechanism of Neurite Outgrowth Enhanced by Electrical Stimulation: Involvement of MicroRNA-363-5p Targeting DCLK1 Expression in Rat. Neurochem Res 42, 513-525, doi:10.1007/s11064-016-2100-0 (2017).

145 Wu, X., Ruan, Y., Jiang, H. & Xu, C. MicroRNA-424 inhibits cell migration, invasion, and epithelial mesenchymal transition by downregulating doublecortin-like kinase 1 in ovarian clear cell carcinoma. Int J Biochem Cell Biol 85, 66-74, doi:10.1016/j.biocel.2017.01.020 (2017).

146 Wang, W., Zhang, H., Wang, L., Zhang, S. & Tang, M. miR-613 inhibits the growth and invasiveness of human hepatocellular carcinoma via targeting DCLK1. Biochem Biophys Res Commun 473, 987-992, doi:10.1016/j.bbrc.2016.04.003 (2016).

147 Sakaguchi, M. et al. miR-137 Regulates the Tumorigenicity of Colon Cancer Stem Cells through the Inhibition of DCLK1. Mol Cancer Res 14, 354-362, doi:10.1158/1541-7786.MCR-15-0380 (2016).

148 Korpal, M., Lee, E. S., Hu, G. & Kang, Y. The miR-200 family inhibits epithelial-mesenchymal transition and cancer cell migration by direct targeting of E-cadherin transcriptional repressors ZEB1 and ZEB2. J Biol Chem 283, 14910-14914, doi:C800074200 [pii] 10.1074/jbc.C800074200 (2008).

149 Liu, X. & Fan, D. The epithelial-mesenchymal transition and cancer stem cells: functional and mechanistic links. Curr Pharm Des 21, 1279-1291 (2015).

54

150 Cancer Genome Atlas Research, N. Comprehensive molecular characterization of gastric adenocarcinoma. Nature 513, 202-209, doi:10.1038/nature13480 (2014).

151 Cancer Genome Atlas Research Network. Electronic address, w. b. e. & Cancer Genome Atlas Research, N. Comprehensive and Integrative Genomic Characterization of Hepatocellular Carcinoma. Cell 169, 1327-1341 e1323, doi:10.1016/j.cell.2017.05.046 (2017).

152 Cancer Genome Atlas Research, N. et al. Integrated genomic characterization of oesophageal carcinoma. Nature 541, 169-175, doi:10.1038/nature20805 (2017).

153 Hammond, D. E. et al. Differential reprogramming of isogenic colorectal cancer cells by distinct activating K-Ras mutations. Journal of proteome research, doi:10.1021/pr501191a (2015).

154 Watanabe, K. et al. Latent process genes for cell differentiation are common decoders of neurite extension length. Journal of cell science 125, 2198-2211, doi:10.1242/jcs.097709 (2012).

155 Wang, H. et al. DCLK1 is correlated with MET and ERK5 expression, and associated with prognosis in malignant pleural mesothelioma. Int J Oncol 51, 91-103, doi:10.3892/ijo.2017.4021 (2017).

156 Kuribara, M. et al. ERK-regulated double cortin-like kinase (DCLK)-short phosphorylation and nuclear translocation stimulate POMC gene expression in endocrine melanotrope cells. Endocrinology 152, 2321-2329, doi:en.2011-0067 [pii] 10.1210/en.2011-0067 (2011).

157 Ikezono, Y. et al. Pancreatic Neuroendocrine Tumors and EMT Behavior are Driven by the CSC Marker DCLK1. Molecular cancer research : MCR, doi:10.1158/1541-7786.MCR-16-0285 (2017).

158 Zhong, X. et al. Suppression of MicroRNA 200 Family Expression by Oncogenic KRAS Activation Promotes Cell Survival and Epithelial-Mesenchymal Transition in KRAS-Driven Cancer. Mol Cell Biol 36, 2742-2754, doi:10.1128/MCB.00079-16 (2016).

159 Patel, O. et al. Biochemical and Structural Insights into Doublecortin-like Kinase Domain 1. Structure 24, 1550-1561, doi:10.1016/j.str.2016.07.008 (2016).

160 Nagamine, T., Shimomura, S., Sueyoshi, N. & Kameshita, I. Influence of Ser/Pro-rich domain and kinase domain of double cortin-like protein kinase on microtubule-binding activity. J Biochem 149, 619-627, doi:10.1093/jb/mvr013 (2011).

161 Koizumi, H. et al. DCLK1 phosphorylates the microtubule-associated protein MAP7D1 to promote axon elongation in cortical neurons. Dev Neurobiol 77, 493-510, doi:10.1002/dneu.22428 (2017).

162 Nagamine, T., Nomada, S., Onouchi, T., Kameshita, I. & Sueyoshi, N. Nuclear translocation of doublecortin-like protein kinase and phosphorylation of a transcription factor JDP2. Biochem Biophys Res Commun 446, 73-78, doi:10.1016/j.bbrc.2014.02.075 (2014).

163 Westphalen, C. B., Quante, M. & Wang, T. C. Functional implication of Dclk1 and Dclk1-expressing cells in cancer. Small GTPases 8, 164-171, doi:10.1080/21541248.2016.1208792 (2017).

55

164 Vega, K. J. et al. Identification of the putative intestinal stem cell marker doublecortin and CaM kinase-like-1 in Barrett's esophagus and esophageal adenocarcinoma. J Gastroenterol Hepatol 27, 773-780 (2012).

165 Potten, C. S., Merritt, A., Hickman, J., Hall, P. & Faranda, A. Characterization of radiation-induced apoptosis in the small intestine and its biological implications. Int J Radiat Biol 65, 71-78 (1994).

166 Meira, L. B. et al. DNA damage induced by chronic inflammation contributes to colon carcinogenesis in mice. The Journal of clinical investigation 118, 2516-2525, doi:10.1172/JCI35073 (2008).

167 Deng, X. et al. Characterization of a selective inhibitor of the Parkinson's disease kinase LRRK2. Nat Chem Biol 7, 203-205, doi:nchembio.538 [pii] 10.1038/nchembio.538 (2011).

168 Kawamura, D. et al. Enhancement of cytotoxic effects of gemcitabine by Dclk1 inhibition through suppression of Chk1 phosphorylation in human pancreatic cancer cells. Oncol Rep 38, 3238-3244, doi:10.3892/or.2017.5974 (2017).

169 Pattabiraman, D. R. & Weinberg, R. A. Tackling the cancer stem cells - what challenges do they pose? Nature reviews. Drug discovery 13, 497-512, doi:10.1038/nrd4253 (2014).

170 Scheel, C. & Weinberg, R. A. Cancer stem cells and epithelial-mesenchymal transition: concepts and molecular links. Semin Cancer Biol 22, 396-403, doi:10.1016/j.semcancer.2012.04.001 (2012).

171 Mayerle, J. et al. Metabolic biomarker signature to differentiate pancreatic ductal adenocarcinoma from chronic pancreatitis. Gut 67, 128-137, doi:10.1136/gutjnl-2016-312432 (2018).

172 Gyorffy, B. et al. Multigene prognostic tests in breast cancer: past, present, future. Breast Cancer Res 17, 11, doi:10.1186/s13058-015-0514-2 (2015).

173 Merrill, L., Gonzalez, E. J., Girard, B. M. & Vizzard, M. A. Receptors, channels, and signalling in the urothelial sensory system in the bladder. Nat Rev Urol 13, 193-204, doi:10.1038/nrurol.2016.13 (2016).

174 Blaine, J., Chonchol, M. & Levi, M. Renal control of calcium, phosphate, and magnesium homeostasis. Clin J Am Soc Nephrol 10, 1257-1272, doi:10.2215/CJN.09750913 (2015).

175 Matsumoto, N., Pilz, D. T. & Ledbetter, D. H. Genomic structure, chromosomal mapping, and expression pattern of human DCAMKL1 (KIAA0369), a homologue of DCX (XLIS). Genomics 56, 179-183, doi:10.1006/geno.1998.5673 (1999).

176 Minardi, D. et al. Prognostic role of tumor necrosis, microvessel density, vascular endothelial growth factor and hypoxia inducible factor-1alpha in patients with clear cell renal carcinoma after radical nephrectomy in a long term follow-up. Int J Immunopathol Pharmacol 21, 447-455 (2008).

177 Wang, K. et al. Increased expression of ALDH1A1 protein is associated with poor prognosis in clear cell renal cell carcinoma. Medical oncology 30, 574, doi:10.1007/s12032-013-0574-z (2013).

178 Ueda, K. et al. Aldehyde dehydrogenase 1 identifies cells with cancer stem cell-like properties in a human renal cell carcinoma cell line. PloS one 8, e75463, doi:10.1371/journal.pone.0075463 (2013).

179 Philip, B., Ito, K., Moreno-Sanchez, R. & Ralph, S. J. HIF expression and the role of hypoxic microenvironments within primary tumours as protective sites driving cancer stem cell renewal and metastatic progression. Carcinogenesis 34, 1699-1707, doi:10.1093/carcin/bgt209 (2013).

56

180 Hu, Y. & Fu, L. Targeting cancer stem cells: a new therapy to cure cancer patients. American journal of cancer research 2, 340-356 (2012).

181 Qu, D. et al. Ablation of Doublecortin-Like Kinase 1 in the Colonic Epithelium Exacerbates Dextran Sulfate Sodium-Induced Colitis. PloS one 10, e0134212, doi:10.1371/journal.pone.0134212 (2015).

182 May, R. et al. Brief report: Dclk1 deletion in tuft cells results in impaired epithelial repair after radiation injury. Stem cells 32, 822-827, doi:10.1002/stem.1566 (2014).

183 Weygant, N. et al. Small molecule kinase inhibitor LRRK2-IN-1 demonstrates potent activity against colorectal and pancreatic cancer through inhibition of doublecortin-like kinase 1. Molecular cancer 13, 103, doi:10.1186/1476-4598-13-103 (2014).

184 Qu, D. et al. Doublecortin-like kinase 1 is elevated serologically in pancreatic ductal adenocarcinoma and widely expressed on circulating tumor cells. PloS one 10, e0118933, doi:10.1371/journal.pone.0118933 (2015).

185 Weygant, N. et al. Survival of Patients with Gastrointestinal Cancers Can Be Predicted by a Surrogate microRNA Signature for Cancer Stem-like Cells Marked by DCLK1 Kinase. Cancer research 76, 4090-4099, doi:10.1158/0008-5472.CAN-16-0029 (2016).

186 Weygant, N. et al. DCLK1 is a broadly dysregulated target against epithelial-mesenchymal transition, focal adhesion, and stemness in clear cell renal carcinoma. Oncotarget 6, 2193-2205 (2015).

187 Liu, Y. H. et al. Doublecortin-like kinase 1 expression associates with breast cancer with neuroendocrine differentiation. Oncotarget 7, 1464-1476, doi:10.18632/oncotarget.6386 (2016).

188 Powrozek, T. et al. Methylation of the DCLK1 promoter region in circulating free DNA and its prognostic value in lung cancer patients. Clin Transl Oncol 18, 398-404, doi:10.1007/s12094-015-1382-z (2016).

189 Herring, C. A. et al. Unsupervised Trajectory Analysis of Single-Cell RNA-Seq and Imaging Data Reveals Alternative Tuft Cell Origins in the Gut. Cell Syst, doi:10.1016/j.cels.2017.10.012 (2017).

190 Haber, A. L. et al. A single-cell survey of the small intestinal epithelium. Nature 551, 333-339, doi:10.1038/nature24489 (2017).

191 Chandrakesan, P., Weygant, N., May, R., Sureban, S.M., Qu, D., and Houchen, C.W. in AACR 2015. AM2015-4220 (Cancer Research).

192 Kantara, C. et al. Methods for detecting circulating cancer stem cells (CCSCs) as a novel approach for diagnosis of colon cancer relapse/metastasis. Lab Invest 95, 100-112, doi:10.1038/labinvest.2014.133 (2015).

193 Mirzaei, A. et al. Upregulation of circulating cancer stem cell marker, DCLK1 but not Lgr5, in chemoradiotherapy-treated colorectal cancer patients. Tumour Biol 36, 4801-4810, doi:10.1007/s13277-015-3132-9 (2015).

194 Leygo, C. et al. DNA Methylation as a Noninvasive Epigenetic Biomarker for the Detection of Cancer. Dis Markers 2017, 3726595, doi:10.1155/2017/3726595 (2017).

195 O'Connell, M. R. et al. Epigenetic changes and alternate promoter usage by human colon cancers for expressing DCLK1-isoforms: Clinical Implications. Sci Rep 5, 14983, doi:10.1038/srep14983 (2015).

196 Vedeld, H. M., Skotheim, R. I., Lothe, R. A. & Lind, G. E. The recently suggested intestinal cancer stem cell marker is an epigenetic biomarker for colorectal cancer. Epigenetics 9 (2014).

57

197 Andresen, K. et al. Novel target genes and a valid biomarker panel identified for cholangiocarcinoma. Epigenetics 7, 1249-1257, doi:10.4161/epi.22191 (2012).

198 van de Wetering, M. et al. Prospective derivation of a living organoid biobank of colorectal cancer patients. Cell 161, 933-945, doi:10.1016/j.cell.2015.03.053 (2015).

199 Wibowo, A., Peters, E. C. & Hsieh-Wilson, L. C. Photoactivatable glycopolymers for the proteome-wide identification of fucose-alpha(1-2)-galactose binding proteins. J Am Chem Soc 136, 9528-9531, doi:10.1021/ja502482a (2014).

58

8 Appendix I. 8.1 Epidemiology and classification of select tumors

Table I. World Health Organization/IARC Guidelines for TNM Staging (T)1-2.

Table II. World Health Organization/IARC Guidelines for TNM Staging (N)1-2.

Table III. World Health Organization/IARC Guidelines for TNM Staging (M)1-2.

T Classification T1 T2 T3 T4

Colorectal Tumor invading the colorectal submucosa

Tumor invading the muscularis propria

Tumor invading through the subserosa or non-peritoneal surrounding tissues

Tumor invading any other organs or tissue structures or invading through to the peritoneal cavity

Pancreatic Pancreas localized tumor < 2 cm at longest diameter

Pancreas localized tumor >2cm at longest diameter

Tumor invading the duodenum, bile duct, or any other surrounding tissue

Tumour invading directly into the stomach, spleen, colon, or nearby large vessels

Gastric Tumor invades gastric lamina propria or submucosa

Tumor invades muscularis propria or subserosa

Tumor invading through serosa without invading adjacent tissue Tumor invading adjacent tissue

Esophageal Tumor invades esophageal lamina propria or submucosa

Tumor invades muscularis propria Tumor invades adventitia Tumor invading adjacent tissue

Kidney Tumor < 7 cm at longest diameterTumor > 7 cm but kidney localized

Tumor invading major veins, adrenal gland, or surrounding tissue in renal/adrenal compartment

Tumor invading beyond the renal/adrenal compartment ("Gerota fascia")

N Classification N1 N2 N3

ColorectalMetastasis in 1-3 local lymph nodes

Metastasis in >3 local lymph nodes N/A

PancreaticMetastasis in 1 (N1a) or more (N1b) local lymphnodes N/A N/A

GastricMetastasis in 1-6 local lymph nodes

Metastasis in 7-15 local lymph nodes

Metastasis in >15 local lymph nodes

Esophageal Metastasis in local lymph node N/A N/A

Kidney Metastasis in 1 local lymph node Metastasis in >1 local lymph node N/A

M Classification M1

Colorectal Distant metastasis

Pancreatic Distant metastasis

Gastric Distant metastasis

EsophagealDistant metastasis in celiac lymph node (M1a) or anywhere else (M1b)

Kidney Distant metastasis

59

Table IV. World Health Organization/IARC Guidelines for TNM Staging (Stage)1-2.

Table V. Incidence, Mortality, and Characteristics of Select Tumors1-4.

8.1 References. 1. Eble J.N., Sauter G., Epstein J.I., Sesterhenn I.A. (Eds.): World Health Organization

ClassificationofTumours.PathologyandGeneticsofTumoursoftheUrinarySystemand

MaleGenitalOrgans.IARCPress:Lyon2004.

2. HamiltonS.R.,AaltonenL.A.(Eds.):WorldHealthOrganizationClassificationofTumours.

PathologyandGeneticsofTumoursoftheDigestiveSystem.IARCPress:Lyon2000.

3. HowladerN,NooneAM,KrapchoM,MillerD,BishopK,KosaryCL,YuM,RuhlJ,Tatalovich

Z,MariottoA,LewisDR,ChenHS,FeuerEJ,CroninKA(eds).SEERCancerStatisticsReview,

1975-2014, National Cancer Institute. Bethesda,

MD,https://seer.cancer.gov/csr/1975_2014/, based on November 2016 SEER data

submission,postedtotheSEERwebsite,April2017.

4. Cancer Statistics for the UK, <http://www.cancerresearchuk.org/health-

professional/cancer-statistics-for-the-uk>

Staging Stage I Stage II Stage III Stage IV

Colorectal T1-2 N0 M0 T3-4 N0 M0 T1-4 N1-2 M0 T-4 N0-2 M1

Pancreatic T1-2 N0 M0 T3 N0 M0 T1-3 N1 M0 T4 N0-1 M0 or T1-4 N0-1 M1

Gastric T1-2 N0-1 M0 T1/N2 or T2/N1 or T3/N0 + M0

T2/N2 or T3/N1 or T4/N0 + M0

T4 N1-3 or T1-3 N3 + M0/T1-4 N0-3 M1

Esophageal T1 N0 M0 T2-3 N0 M0 T1-3 N1 M0 or T4 N0-1 M0 T1-4 N0-1 M1

Kidney T1 N0 M0 T2 N0 M0 T3 N0 M0 or T1-3 N1 M0

T4 N0-1 or T1-4 N2 + M0/T1-4 N0-2 M1

Category Locale Colorectal Pancreatic Gastric Esophageal KidneyIncidence (per 100,000 persons) US

40.1 12.5 7.3 4.2 15.6

UK 85.0 19.7 14.1 18.7 25.7

Mortality (per 100,000 persons) US

14.8 10.9 3.2 4.1 3.9

UK 33.2 18.1 9.8 16.4 9.3

5-Year Survival (%) US 64.9 8.2 30.6 18.8 74.1

UK 59 3 19 15 56

Common Subtypes

Adenocarcinoma, Mucinous Adenocarcinoma

Adenocarcinoma, Acinar Cell, Intraductal Papillary Mucinous Neoplasm

Adenocarcinoma Adenocarcinoma, Squamous Cell

Clear Cell, Papillary, Chromophobe

Major Risk Factors

Tobacco, Radiation, Red Meat, Alcohol, Physical Inactivity, Obesity , IBD

Tobacco, Obesity, Chronic Pancreatitis, Chemicals, Diabetes

Tobacco, Radiation, H. pylori , Preserved Foods, Chemicals

Tobacco, Radiation, Chemical Exposure, Alcohol, Obesity, HPV, Esophagitis

Tobacco, Obesity, Hypertension

60

8.2 Research Ethics Review Checklist

UPR16 – August 2015

FORM UPR16 Research Ethics Review Checklist

Please include this completed form as an appendix to your thesis (see the Postgraduate Research Student Handbook for more information

Postgraduate Research Student (PGRS) Information

Student ID:

872999

PGRS Name:

Nathaniel Weygant

Department:

Pharmacy & Biomedical Sciences

First Supervisor:

Dr. Sassan Hafizi

Start Date: (or progression date for Prof Doc students)

1 June 2017

Study Mode and Route:

Part-time

Full-time

MPhil

PhD

MD

Professional Doctorate

Title of Thesis:

Doublecortin-like kinase 1 as a novel drug target and biomarker in gastrointestinal and renal cancers

Thesis Word Count: (excluding ancillary data)

10,975

If you are unsure about any of the following, please contact the local representative on your Faculty Ethics Committee for advice. Please note that it is your responsibility to follow the University’s Ethics Policy and any relevant University, academic or professional guidelines in the conduct of your study

Although the Ethics Committee may have given your study a favourable opinion, the final responsibility for the ethical conduct of this work lies with the researcher(s).

UKRIO Finished Research Checklist: (If you would like to know more about the checklist, please see your Faculty or Departmental Ethics Committee rep or see the online version of the full checklist at: http://www.ukrio.org/what-we-do/code-of-practice-for-research/)

a) Have all of your research and findings been reported accurately, honestly and within a reasonable time frame?

YES NO

b) Have all contributions to knowledge been acknowledged?

YES NO

c) Have you complied with all agreements relating to intellectual property, publication and authorship?

YES NO

d) Has your research data been retained in a secure and accessible form and will it remain so for the required duration?

YES NO

e) Does your research comply with all legal, ethical, and contractual requirements?

YES NO

Candidate Statement:

I have considered the ethical dimensions of the above named research project, and have successfully obtained the necessary ethical approval(s)

Ethical review number(s) from Faculty Ethics Committee (or from NRES/SCREC):

N/A

If you have not submitted your work for ethical review, and/or you have answered ‘No’ to one or more of questions a) to e), please explain below why this is so:

This work is completed in regards to a PhD by publication project. All relevant studies are published. All work was done in accordance with the relevant ethics approvals from respective institutions.

UPR16 – August 2015

Signed (PGRS):

Date: 10 January 2018

UPR16 – August 2015

Signed (PGRS):

Date: 10 January 2018