RARRES1 MODULATION OF TUBULIN DEGLUTAMYLATION REGULATES METABOLISM AND CELL SURVIVAL
A Dissertation Submitted to the Faculty of the Graduate School of Arts and Sciences Of Georgetown University In partial fulfillment of the requirements for the Degree of Doctor of Philosophy In Cell Biology
By
Sara Maimouni, B.A.
Washington, D.C. August 20, 2018 Copyright 2018 by Sara Maimouni All Rights Reserved
ii RARRES1 Modulation of Tubulin Deglutamylation Regulates Metabolism and Cell Survival
Sara Maimouni, B.A.
Thesis Advisor: Stephen Byers, Ph.D.
ABSTRACT
One of the driving mechanisms of cancer progression is the reprogramming of metabolic pathways in intermediary metabolism. Cancers increase their energy expenditure by increasing ATP production for utilization in anabolic pathways to increase production of proteins, nucleic acids and lipids. The Warburg Effect, where cancer cells predominantly use aerobic glycolysis rather than oxidative phosphorylation to produce ATP, was long thought to be the main initiating pathway in increasing tumor burden. However, compelling new evidence shows that there exists metabolic heterogeneity among and within tumors. Mitochondrial respiration often plays a major role in tumor progression, as many different cancers contain a subpopulation of slow-cycling tumor-initiating cells that are multidrug-resistant and dependent on oxidative phosphorylation.
These cells represent a target for cancer therapy. However, the identification of endogenous regulators of mitochondrial respiration is understudied. Depletion of (RARRES1) occurs in many cancers, including melanoma, colon, prostate and breast cancers. Interestingly, in cancers associated with fibrosis, such as triple negative breast cancer, pancreatic and hepatocellular carcinomas, RARRES1 is overexpressed. Its role in metabolism might explain the duality of this protein in cancer. Our data show that RARRES1 and its target CCP2 regulate mitochondrial bioenergetics and subsequently alter energy homeostasis by modulating the function of the mitochondrial voltage-dependent anion channel (VDAC). The changes in energy homeostasis rewire glucose for biosynthetic pathways, such as de novo lipogenesis, that drive the pathogenicity
iii and survival of cancer. These data lay the foundation for metabo-therapy of the many tumor types that exhibit RARRES1 depletion or overexpression and may have the added benefit of targeting drug-resistant tumor-initiating cells.
i v Dedication
This thesis is dedicated to my parents, Ihsan Lemtouni and Rachid Maimouni, who have supported me and sacrificed everything to support my endeavors in the United States. I would have never gotten this far in my career without you. My grandparents, Aboubakr Lemtouni and Rabia Temsamani, who have always echoed to me that “No matter what, education comes first”. To my brother, Karim Maimouni and my sister, Dania Maimouni, who have always believed in me and reminded me why I chose this path in the first place.
Acknowledgments
I would like to thank Dr. Stephen Byers for his invaluable guidance during my thesis work and his advice and support in my career development.
Thank you to Dr. Jessica Jones, Dr. Rebecca Riggins and Dr. Kathryn Sandberg for their mentorship and support.
I would like to thank my thesis committee members, Dr. Aykut Uren, Dr. Partha Banerjee, Dr. Amrita Cheema and Dr. Maria Laura Avantaggiati for their guidance during my predoctoral training.
Thank you to my lab members and collaborators who have contributed to my research: Naiem Issa, Yong-Sik Bong, Ivana Peran, and Becky Hoxter.
Thank you to the students I have mentored and have contributed to my research: Selina Cheng and Asha Krishnakumar.
A special thank you to my college advisor, Dr. Kristina Blake-Hodek, who inspired me to pursue my Ph.D.
v Table of Contents
Introduction ...... 1 RARRES1- A Carboxypeptidase Inhibitor ...... 4 RARRES1 and CCP2, Regulators of the Tubulin Code ...... 10 The C-Terminal Tail of Tubulin Modulates Mitochondrial VDAC ...... 14 RARRES1 and CCP2 in Metabolic Diseases ...... 18 The Role of RARRES1 and CCCP2 in Tumorigenesis ...... 19 Statement of Purpose ...... 24 Specific Aims ...... 26
Chapter 1 Tumor Suppressor RARRES1 Links Tubulin Deglutamylation to Metabolic Reprogramming and Cell Survival ...... 27 Introduction ...... 27 Methods...... 28 Results ...... 35 Discussion ...... 62
Chapter 2 RARRES1, A Novel Regulator of Fatty Acid Metabolism ...... 67 Introduction ...... 67 Methods...... 68 Results ...... 81 Discussion ...... 98
Chapter 3 Endogenous Transcriptional Regulation of RARRES1 ...... 102 Introduction ...... 102 Methods...... 105 Results ...... 107 Discussion ...... 120
Conclusion ...... 125 RARRES1 and Stem Cell Metabolism...... 126 vi RARRES1 and Obesity ...... 127 Oncogene or Tumor Supressor?- RARRES1 and Its Role in Fibrosis...... 129 Role of RARRES1 and CCP2 in Cardiovascular Diseases...... 130 Other Targets of RARRES1?...... 132
Appendix: Chapter 3 Supplementary Figures ...... 134
Bibliography ...... 143
vii List of Figures
Figure 1.1. RARRES1 and Latexin Structure Modeling ...... 6 Figure 1.2. Phylogenetic Tree of RARRES1 Across All Species ...... 8 Figure 1.3 RARRES1 (TIG1A) and Its Variant RARRES1 Short (TIG1B) ...... 9 Figure 1.4. The Human Tubulin Isotypes ...... 11 Figure 1.5. The Tubulin Code ...... 12 Figure 1.6. The Post Translational Modification of GEEY Motif on Tubulin ...... 14
Figure 1.7. Multi-Substrate Transport and Gate of Death ...... 17𝛼𝛼 − Figure 1.8. Comparison of RARRES1 Across Twelve Analyses of Breast Cancer Stroma, TNBC, Pancreatic Cirrhosis/Cancer, and Liver Cirrhosis Datasets ...... 22
Figure 2.1. CCP2 Expression in Tumorigenic and Non-Tumorigenic Cell Lines ...... 37 Figure 2.2. CCP2 Splice Variants and Primer Profiling ...... 38 Figure 2.3. Immunocytochemistry of Polyglutamylated Tubulin ...... 38 Figure 2.4. RARRES1, CCP2 and Retinoic Acid Regulate Tubulin Glutamylation ...... 39 Figure 2.5. RARRES1and CCP2 Regulate Apoptosis but Not Cell Proliferation ...... 42 Figure 2.6. FACS Analysis of Annexin Stained Cells ...... 43 Figure 2.7. RARRES1 and CCP2 Regulate Anchorage Independent Cell Growth and CD49f ....46 Figure 2.8. RARRES1 and CCP2 Reciprocally Regulate Stem Cell Marker CD49f ...... 47 Figure 2.9. MitoTracker Staining ...... 48 Figure 2.10. RARRES1 Regulates Mitochondrial Membrane Potential (MMP) ...... 49 Figure 2.11. RARRES1/CCP2-Mediated Changes in Mitochondrial Respiration Require VDAC1. 52 Figure 2.12. VDAC1 and Its Effects on RARRES1 Depletion in PWR-1E Cells and KD Efficiency of VDAC1 and RARRES1 ...... 53 Figure 2.13. RARRES1 Depleted Cells are Less Dependent on Glycolysis ...... 55 Figure 2.14. RARRES1 Regulates Energy Metabolism...... 58 Figure 2.15. VDAC1 Depletion or Inhibition Reverses the Apoptotic Phenotype Seen in RARRES1- Depleted Cells ...... 61 Figure 2.16. FACS Analysis of Annexin Stained Cells ...... 62
viii Figure 3.1. RARRES1-YFP and RARRES1 siRNA Transfection Efficiency ...... 71 Figure 3.2. Non-Targeted LC-MS Analysis of Transient RARRES1 KD in MCF 10A Cells ...... 77 Figure 3.3. GC-MS Analysis ...... 78 Figure 3.4. RARRES1 Regulates Lipid Accumulation ...... 84 Figure 3.5. RARRES1 Regulates Lipid Content ...... 85 Figure 3.6. RARRES1 Increases Substrate Availability for Fatty Acid Oxidation ...... 88 Figure 3.7. Fatty Acid Oxidation Activity in RARRES1-Depleted Epithelial Cells ...... 89 Figure 3.8. RARRES1 Regulates De Novo Lipogenesis ...... 92 Figure 3.9. LC-MS of Stable RARRES1 Knockdown MCF 10A Cells ...... 93 Figure 3.10. Glycolytic Activity in RARRES1 Depleted Epithelial Cells ...... 93 Figure 3.11. FASN, ACC1 and ACC2 Expression in RARRES1 siRNA or RARRES1 Transfected MCF 10A Cells ...... 95 Figure 3.12. Differential Expression of RARRES1 in Cancers Correlates with Expression of Fatty Acid Metabolism Genes ...... 98 Figure 3.13. RARRES1 Depletion in Epithelial Cells Increases De Novo Lipogenesis and Improves Fatty Acid Oxidation during Starvation ...... 101
Figure 4.1. RARRES1 is Commonly Regulated by Retinoic Acid and Vitamin D ...... 104 Figure 4.2. Relative RARRES1 Expression at Different Densities ...... 108 Figure 4.3. RARRES1 Expression is Regulated by Growth Factors...... 109 Figure 4.4. RARRES1 Expression after Treatment with Glutamine or Glucose in Serum Starved MCF 10A Cells ...... 112 Figure 4.5. Inhibition of Mitochondrial Respiration Regulates RARRES1 and AGBL2 Expression………………………………………………………………………………………...113 Figure 4.6. YY1 Reverses Growth Factor Starvation Induced RARRES1 Expression ...... 117 Figure 4.7. PPAR Alpha and Gamma Regulate the Expression of RARRES1 Gene ...... 120
Figure S4.1. Signaling Pathways that Regulate RARRES1 Expression in the NURSA Database………………………………………………………………………………………...125
Figure 5.1. Proposed Effects of RARRES1 and CCP2 on Metabolic Reprogramming ...... 135
ix List of Tables
Table 1.1. Oncominetm Analysis: Cancer Summary for RARRES1 Expression ...... 21
Table 1.2. Oncominetm Analysis: Cancer Summary for AGBL2 Expression ...... 23
Table 2.1. Metabolites Significantly Altered in RARRES1 Stable Knockdown MCF 10A Cells Relative to Empty Vector ...... 59
Table 3.1. Common Pathways Identified in the Analysis of Genes Co-expressed with RARRES1 in Breast, Prostate, and Colorectal Cancers ...... 96
Table S4.1. Transcription Factors that Were Predicted to Bind to RARRES1 Promoter Region in PROMO...... 126 Table S4.2. Transcription Factors that Were Found Bound to RARRES1 Gene in CHEA (ENRICHR) Database...... 133
x Abbreviations
ACC1/2- acetyl-CoA carboxylase 1 /2
AGBL2- ATP/GTP Binding Protein Like 2
AMPK- Adenine monophosphate activated protein kinase
ATP- Adenine triphosphate
BCAA- Branched Chain Amino Acid
CAP-gly- cytoskeleton associated protein-glycine enriched proteins
CCP2- cytosolic carboxypeptidase 2
CTT- Carboxy-terminal tail
D2-tubulin- detyrosinated tubulin
DHA- Docosahexaenoic Acid
DNL- de novo lipogenesis
EGF-Epidermal growth factor
ETC- Electron transport chain
TCA- tricarboxylic acid cycle
FAO- fatty acid oxidation
FASN- fatty acid synthase
SCD- Stearoyl-CoA desaturase
GC-MS- Gas chromatography in combination with mass spectrometry
Glut-tubulin- glutamylated tubulin
HC- Hydrocortisone
HCC- Hepatocellular Carcinoma
xi IBC- Invasive breast cancer
LC-MS- Liquid chromatography in combination with mass spectrometry
MT-Microtubule
Mtor- mechanistic target of rapamycin
OXC-32- Ovocalyxin-32
PDAC- Pancreatic Ductal Adenocarcinoma
PPARA- peroxisome proliferator-activated receptor alpha
PPARG- peroxisome proliferator-activated receptor gamma
PPARGC1A- peroxisome proliferator-activated receptor gamma coactivator 1 alpha
PTM- Post translational modification
RA- Retinoic acid
RARRES1- Retinoic acid receptor responder 1
RNA- ribonucleic acid
SIRT1- sirtuin 1
SREBF1- sterol regulatory element-binding transcription factor 1
TCA cycle- tricarboxylic acid cycle
TIG1- Tazarotene induced gene 1
TNBC- Triple Negative Breast Cancer
TTCP- Tubulin tyrosine carboxypeptidase
TTL-Tubulin tyrosine ligase
TZDs- Thiazolidinediones
VASH1 and VASH2- Vasohibin 1 and 2
VDAC1- Voltage Dependent Anion Channel 1
xii Vit D- Vitamin D
YY1- Yin-Yang 1
x iii Introduction
Metabolism is the basis of any biological entity. From the cellular to the organismal level,
metabolic signaling is needed in all stages of the living organism. Metabolism is defined as
chemical reactions in living organisms that either break down molecules to produce energy or
consume energy to synthesize compounds (1). The molecules and compounds essential in energy
production and consumption consist of nutrients, like carbohydrates, fatty acids, and amino acids
(1). Together these pathways are called intermediary metabolism and are essential for energy homeostasis. Intermediary metabolism can be defined as 1. Synthesis of molecules or polymerization of molecules into complex metabolites (anabolism), and 2. Degradation of molecules to release energy (catabolism) (1). These pathways are crucial to maintain existence at a cellular and organismal level. Perturbation or reprogramming of energy homeostasis can alter the state of the cell.
Metabolic reprogramming was first appreciated in stem cell biology. Metabolic pathways
influence the commitment, fate and self-renewal of stem cells (2, 3). For example, several studies demonstrated higher levels of glycolytic activity in embryonic stem cells and induced pluripotent stem cells compared to more differentiated cells (3). The glycolytic phenotype is coupled with a
reduced number of mitochondria and oxidative capacity (4, 5). In agreement with the idea that
stem cells prefer glycolysis, ESC differentiation to cardiomyocytes involves upregulation of
oxidative phosphorylation and a decrease in glycolysis (4). But interestingly, the opposite was
observed for ESCs differentiating into neuronal cells (6). This indicates that the metabolic
preference is dependent on the type of cells in which the ESCs will differentiate into.
1 Perturbation of metabolic pathways is also seen in differentiated cells and can lead to a diseased
state. In fact, metabolic reprogramming is now considered a hallmark in a wide range of diseases.
This not only includes metabolic syndrome, diabetes and obesity but also progression of cancer,
autoimmune diseases, and chronic lung, hepatic and pancreatic fibrosis (7–10).
In the case of cancer, metabolic reprogramming has been recognized as one of the ten cancer
hallmarks (11). Solid tumors especially contain significant heterogeneity of perfusion, where tumor cells in the center of the tumor bulk inhabit a nutrient and oxygen poor environment while cells at the edges of the tumor receive higher levels of oxygen and nutrients (12). These cancer
cells, along with cancer-associated cells (such as fibroblasts, tumor-associated immune cells and
adipocytes), adapt to such changes in environment by rewiring their metabolic needs. For example,
the mitochondrial ETC in cancer cells can meet energy requirements with oxygen levels as low as
0.5% (13). The most marked alterations of tumor bioenergetics include increase in catabolic
pathways (such as glycolysis, glutamine metabolism, amino acid and lipid metabolism and
mitochondrial biogenesis) and regulation of anabolic pathways (like pentose phosphate pathway
and macromolecule biosynthesis) (11, 12). In the case of Kras-driven pancreatic cancer, cells
recycle proteins from the extracellular matrix to produce glutamine and other amino acids to feed
the TCA cycle for cell survival and growth in nutrient-depleted conditions (14). In the absence of
glucose, the amino acid, glutamine, can produce glutamate and subsequently α-ketoglutarate to
fuel the TCA cycle, a series of enzymatic reactions called glutaminolysis (1). Glutamine can also
provide acetyl-coa, a metabolite essential in the TCA cycle as well as de novo fatty acid synthesis,
to maintain cell survival when pyruvate oxidation is compromised (15). Other pathways have been
2 recognized to play a role in tumorigenesis. Fatty acid breakdown (beta oxidation) in the
mitochondria can generate acetyl-coa and the reducing equivalents NADH and FADH2, which are used by the ETC to produce mitochondrial ATP. In addition, recent studies have demonstrated branched chain amino acids (BCAAs) like isoleucine, valine, and leucine, can be converted into acetyl-coa and other organic molecules that also enter the TCA cycle (16). The number of
alternative metabolic pathways enables a striking plasticity in cancer cells and tumor associated
cells, to adapt to the changing microenvironment of the tumor.
Oncogenes, tumor suppressors, and the tumor microenvironment dictate the metabolic preference
of these cancer cells. Hypoxia inducible factor alpha (HIF-1 ), c-Myc, and p53 are the most
studied factors that regulate metabolism in the context of tumorigenesis𝛼𝛼 (11). HIF-1 is important
during hypoxia. This transcription factor is stabilized in hypoxic conditions and activates𝛼𝛼 genes
essential to survive in low oxygen environment (17). For example genes that are regulated by HIF-
1 are glycolytic genes; this enables an increase of glycolysis and renders these cells independent
of𝛼𝛼 mitochondrial respiration to produce ATP (17). P53, a tumor suppressor and transcription factor,
is best known for its function in the DNA damage response. Recent publications have suggested
that p53 also regulates the glycolysis and oxidative phosphorylation pathways in cells (18). P53
decreases the glycolytic rate by inhibiting the expression of glucose transporters, GLUT1 and
GLUT4, and increases mitochondrial respiration rate (19–21). C-Myc, one of the best studies
oncogenes is overexpressed or deregulated in at least 40% of cancers (22). This helix-loop-helix
leucine zipper transcription factor binds to regulatory regions in a sequence-dependent or
independent manner and represses or promotes transcription of genes (22). These transcriptional
changes support the production of intermediates for cell growth and proliferation, such as
3 increasing energy production and organic molecule synthesis(23). For example, recent studies have shown that c-Myc enhances glycolytic pathways, beta-oxidation, fatty acid synthesis, one- carbon metabolism and the pentose phosphate pathway (22, 24–26). These emerging studies reinforce the importance of oncogenes and tumor suppressors in actively reprogramming the metabolic needs of cancer cells to promote survival and invasion of tumors. There is more work to be done in order to identify all the players responsible for the striking metabolic plasticity of cancer cells. Recent studies have identified new potential players in metabolic reprogramming in cancer (27–30). One example is Retinoic acid receptor responder 1, RARRES1, a type II tumor suppressor. Not only is RARRES1 differentially expressed in cancer but also in hyperinsulinemia, cholestatic liver disease and obesity (31–35). A recent study has suggested that this protein can regulate mTOR and SIRT1 expression, two major regulators of energy homeostasis in prostate cancer cell lines (32). The capabilities of RARRES1 to metabolically transform cells is the subject of my dissertation.
RARRES1 – A Carboxypeptidase Inhibitor
RARRES1 was first discovered in psoriatic skin grafts that were treated with retinoic acid. The
hydrophobic profile of the protein resembled cyclic ADP ribose hydrolase (CD38) in that they
both consist of a single pass transmembrane, a long extracellular domain (the carboxy-terminal)
and a small intracellular domain (36). With its transmembrane domain, RARRES1 was
hypothesized to localize in membranes but its specific function and exact location was unknown.
Crystallization of the murine latexin protein revealed that the extracellular domain of RARRES1
has a 30.8% homology with the human protein latexin (37). The extracellular domain of
4 RARRES1 is now called the latexin domain. The latexin protein is a cytosolic carboxypeptidase
inhibitor which can inhibit the carboxypeptidase A family such as CPA1, CPA2, CPA3 and CPA4
(37–40) in non-pancreatic tissues, such as the brain. Latexin was found in complex with CPA4 and
localized in the same compartment as its target (38). Latexin and its targets have been suggested
to play a role in in the digestion and post-translational modification of proteins and inflammatory
responses (41). For example, latexin is found expressed constitutively at high basal levels in mouse
macrophages. Latexin can be further upregulated with a growth factor or proinflammatory stimuli
(37, 41–43). CPAs have also been shown to correlate with macrophage cytotoxic activity; they
regulate several proinflammatory mediators, such as endothelin-1 and leukotrienes(43, 44).
Comparing the electrostatic profiles between RARRES1 and latexin also indicate that these
proteins might function similarly (Figure 1.1). The electrostatics surface map of RARRES1 shows that one face is similar to that of latexin. This may indicate that RARRES1 is a carboxypeptidase inhibitor but might interact with peptidase families that do not interact with latexin. And recent studies have also suggested different subcellular locations of latexin and RARRES1. Latexin is located in the cytoplasm, while RARRES1 is localized in the endoplasmic reticulum and golgi compartments; this strongly suggests that these proteins could have different carboxypeptidase targets. Both proteins also consist of a two-fold symmetry with loops connecting the two domains
(37). The loops contribute to the dynamic structure of the proteins, and this allows them to bind
promiscuously to different carboxypeptidase families. Our group was the first to confirm that
RARRES1 is a carboxypeptidase inhibitor. We demonstrated that RARRES1 inhibits and binds to
cytosolic carboxypeptidase 2 (CCP2) or ATP/GTP Binding Protein Like 2 (AGBL2)(45).
5 Figure 1.1. RARRES1 and Latexin Structure Modeling. RARRES1 structure was built upon the crystalized human latexin protein. Amino acid charge is pictured in the figure. Blue is basic amino acids, red is acidic amino acids and white is non-polar amino acids. A. Amino acid alignment of RARRES1, depicted as Model_01, with the human latexin PDB structure, named 1wnh.1.A. B. The cartoon model of RARRES1 and latexin are depicted. The barrel, alpha and beta chains are modeled in the panels. B. Surface structure with the amino acid charge is depicted. Two faces of each protein was depicted. Swiss Prot was utilized to model the proteins (46). The structure is based on the cytoplasmic domain and not the n-terminus transmembrane domain.
Sequence comparison by NCBI’s blastp tool search found that the RARRES1 protein also shares 30-50% homology, depending on species, with the ovocalyxin-32 (OXC-32) protein found in birds (47). Phylogenetic analysis suggests that the RARRES1 is orthologous to bird OCX-32 and paralogous to the latexin gene (48). My phylogenetic study suggests that the RARRES1-related genes fall under the Gnathostomata clade, i.e. Jawed vertebrates. The earliest ancestor of the RARRES1 gene is a “lxn-like precusor” gene found in ghost shark (Callorhinchus milii), a cartilaginous fish but not in elasmobranchs (sharks). The ancestor then diverged into the Latexin- like gene found in extinct (Latimeria chalumnae-an early coelocanth), the ancestral bony fish and
6 in all non-extinct bony fish, such as Danio rerio (Figure 1.2). The generated phylogeny tree also suggests that the latexin-like gene might have duplicated and given rise to RARRES1 (mammalian and xenopous tropicalis)/OCX-32 (birds) and their paralog latexin (Figure 1.2); a study comparing the vertebrate genome shows that the region encompassing ovocalyxin-32/RARRES1 and LXN is conserved among bony fishes, X. Tropicalis, G. Gallus and mammals (48). This suggests that the tandem duplication giving rise to RARRES1/OXC-32 and latexin occurred before the tetrapod superclass (48). Although latexin and RARRES1 share the same ancestor with the ovacalyxin-32 protein, OCX-32 has distinct functions in birds. OCX-32 is expressed in the oviduct and eggshell of birds and is a matrix protein that acts as an antimicrobial. Latexin has been linked to hematopoietic stem cell differentiation and RARRES1 regulates tubulin post translational modification, tumor suppression and stem cell differentiation. As we continue to study these proteins in more depth, we may discover a common function between RARRES1/LXN gene and
OXC-32 gene in the future.
7 Figure 1.2. Phylogenetic Tree of RARRES1 Across All Species. Blastp in NCBI was used to generate a phylogenetic tree. The delta-blast algorithm was used to obtain ancestral genes of RARRES1. The algorithm constructs a position-specific scoring matrix using the results of a Conserved Domain Database search and searches a sequence database. The conserved domain in this case is the latexin domain.
There are two isoforms that translate to RARRES1. Isoform two is 294 amino acids while isoform 1 is 228 amino acids [1] (Figure 1.3). The two proteins differ at the C-terminal end, as shown in Figure 7. RARRES1 short (TIG1 B) form lacks part of the latexin domain, which is thought to retain the carboxypeptidase inhibiting activity.
8 Figure 1.3. RARRES1 (TIG1 A) and its Variant, RARRES1 Short (TIG1 B). RARRES1 has a variant called RARRES1 short or TIG1 short. The variant differs from RARRES1 at the carboxyl terminal. The variant lacks part of the Latexin domain, it lacks amino acids 229 to 294. The variant’s three last amino acids, underlined above, are not present in the long form (45).
These two variants’ structural differences cause functional alterations between the two
isoforms. Zaid et al. Demonstrated that the short isoform or TIG1 B is unable to bind efficiently to AGBL2, supporting the idea that the latexin domain retains the canonical carboxypeptidase
inhibiting activity (45). In colon cancer both isoforms retard the migration of cells but only the
long isoform regulates cell proliferation by delaying the progression from G1 to S phase (49). One
reason only RARRES1 long is able to delay G1 to S phase, is due to RARRRES1 long’s ability to
bind to AGBL2 and inhibit the process of detyrosination. As explained earlier, inhibiting
detyrosination from occurring will cause a halt in tubulin dynamics and thus a halt in cellular
growth, assuming the cell cycle checkpoints are not dysfunctional. RARRES1 short is unable to
bind to AGBL2 thus it will not be able to inhibit the microtubule dynamics.
RARRES1 short’s ability to play an important role in other protein interactions other than
its inability to inhibit carboxypeptidases has not been explored. Relative expression of the different
isoforms in varied cell types must be considered to understand their role. Mrna expression of the
protein is not enough to understand the physiological effects of RARRES1. Unfortunately,
availability of efficient antibodies against RARRES1 are not available. The development of high-
9 quality antibodies must be done to make the study of this protein more feasible. Developing antibodies that specifically target the long form of RARRES1 but not the short form, i.e. An
antibody that targets amino acids 229-294 (Figure 1.3), will enable scientists to understand the
effects of the two isotypes, their interacting partners and when each isoform is needed in certain
cell types and pathways.
RARRES1 and CCP2, Regulators of the Tubulin Code
Microtubules are the most abundant cytoskeletal component in eukaryotic cells (50). These
filamentous components have a variety of different roles in the cell including, maintaining
structural integrity, cell differentiation, cell signaling, chromosome segregation during cell
division, organelle and vesicle transportation in the cells (51). The diverse roles of microtubules
are achieved in part due to post-translational modifications of tubulin and their interactions with
microtubule associated proteins (51). The microtubules consist of mainly -tubulin heterodimers.
and mo nomers are organized into three domains, the N-terminus𝛼𝛼𝛼𝛼 that interacts wi th
nuc𝛼𝛼 leotides,𝛼𝛼 the intermediate domain, and the C-terminal domain (52). Surprisingly, there are
multiple genes encoding for and tubulin (5 1, 53, 54) (Figure 1.4). The tubulin isotypes
assemble and create unique configurations𝛼𝛼 𝛼𝛼 and carry out distinctive functions in the cells (52). This was supported by studies that discovered that certain isotypes were specifically expressed in
specialized cells and tissues (55). The isotypes are also differentially expressed during development (56). Recently, mutations in tubulin isotypes were associated with neurodevelopmental disorders. These mutations can lead to dysregulation of tubulin assembly into
particular microtubule structures (57, 58). The isotypes differ from each other in their amino acid
sequences in the c-terminal tails (Figure 1.4) (52). The c-terminal tail of tubulin consists of a
10 negatively charged tail of about 20 amino acid residues. The c-terminus protrudes from the microtubules (51). This allows for post-translational modifications and variations among tubulin isotypes to dictate the function of microtubules (51). The PTMs are referred to as “tubulin code”
(Figure V).
Figure 1.4. The Human Tubulin Isotypes. The c-terminal tails of each known tubulin isotype is depicted. Each isotype has a unique c-terminus. (Adapted from Janke, C. & Kneussel, M., 2011 (54))
11 Figure 1.5. The Tubulin Code. The different tubulin post translational modifications that occur in alpha and beta tubulin are depicted in this diagram (52).
The “tubulin code” includes ubiquitous protein modifications such as phosphorylation and acetylation, while also consisting of modifications specific to tubulin, such as glutamylation (59)
(Figure 1.5). Tubulin glutamylation and polyglutamylation is especially abundant in neurons, cilia and centrioles (51). One of the most studied PTMs on the C-terminal tail (CTT) of tubulin is the
GEEY motif of -tubulin (Figure 1.4 & 1.5). This motif undergoes a dynamic tyrosination- detyrosination cycle𝛼𝛼 that is associated with microtubule stability but does not seem to directly control it (52). As shown in figure V, the cycle begins when tubulin tyrosine ligase (TTL) catalyzes the addition of a peptide bond between glutamate and tyrosine in the conserved sequence EE/Y of free alpha tubulins (60). Ttl-knockout mice die within 24 hours after birth and present with a drastic increase in detyrosinated tubulin (61, 62). Studies show the Ttl-null mice are also born with neuronal dysmorphism and respiratory problems (62). Such abnormalities and lethality occur due to the importance of tyrosinated and detyrosinated tubulin in axonal MTs (63). Specifically, detyrosination is associated with modulation of MAP-microtubule interactions (64).
Detyrosination increases kinesin-1 driven transport in neurons. Detyrosination also hinders microtubule plus-end localization of a subgroup of CAP-Gly domain proteins called +tips and inhibits microtubule-depolymerizing kinesin activity (65). To further support the importance of
12 tyrosination status in the polymerization of microtubules, detyrosinated tubulin is enriched in long- lived microtubules (66). In vitro, fibroblasts isolated from Ttl-null mice have impaired motility and less-polarized structure (65). These fibroblasts also have excessively detyrosinated microtubules that are abnormally long and hyper stable, and this phenotype is rescued by overexpression of kinesin-13, a microtubule-end depolymerizer (67, 68).
The detyrosination of tubulin was first identified 40 years ago, but the identity of the enzyme responsible for the activity has remained elusive. In 2012, our group demonstrated a potential candidate tubulin tyrosine carboxypeptidase (TTCP), CCP2 (45). Depletion of CCP2 resulted in a decrease in detyrosinated tubulin on the C-terminal EEY region of α-tubulin. While RARRES1 knockdown increased the level of tyrosination on the c-terminus of α-tubulin CCP2 did not appear to directly detyrosinate tubulin. Later work confirmed that CCP2 does not have any detyrosination enzymatic activity, and it is now known that VASH1 and VASH2 can detyrosinate alpha tubulin
(Figure 4), although the studies did identifty tubulin detyrosination activity that could not be explained through VASH (69, 70). In 2014, Janke showed that CCP2, along with CCP3, deglutamylate the penultimate glutamate on c-terminus of alpha tubulin (71). This deglutamylation renders the tubulin unable to reenter the tyrosination cycle (Figure 1.6). Thus it is plausible that
CCP2 indirectly regulates detyrosination by inhibiting tubulin from getting re-tyrosinated, thus decreasing the abundance of tyrosinated tubulin.
13 Figure 1.6. The Post Translational Modification of GEEY Motif on Tubulin. VASH 1 and VASH2 remove the ultimate tyrosine on the c-terminus of alpha tubulin, while tyrosine ligase reverses the activity, by re-tyrosinating the end of tubulin. CCP2 is𝜶𝜶 −suggested to hydrolyze glutamate from the c-terminus of tubulin. I hypothesize that RARRES1 inhibition of CCP2 activity regulates the ∆2 tubulin status of alpha tubulin.
The dynamic and intrinsically disordered carboxy-terminal tails (ctts) of tubulin are clearly
involved in regulating binding interactions of tubulin with microtubule associated protein to
regulate microtubule dynamics. But Recent studies have identified that the ctts are also important
in metabolic reprogramming. For example, tyrosinated tubulin dictates lipid droplet interaction
with mitochondria during glucose deprivation and and tubulins bind to the mitochondrial
voltage dependent anion channel (VDAC) to regulate𝛽𝛽 mitochondrial𝛼𝛼 energetics (72–74).
The C-Terminal Tail of Tubulin Modulates Mitochondrial VDAC
Oxidative phosphorylation requires transport of metabolites, including cytosolic ADP, ATP and citrate, across both mitochondrial membranes for F1F0-atpase to generate ATP in the matrix (75).
It has been established that VDAC is the central regulator of ATP and ADP flux for mitochondrial
function (76, 77). This channel is known as the gate of death. It regulates metabolite flux, ion flux
(i.e. Calcium and potassium), and regulates cytochrome c release from the mitochondria, inducing apoptosis (cell death) (73, 76, 77). There are three isoforms known, VDAC1, VDAC2 and
14 VDAC3; they share 70% identity (78). The three are expressed in most tissues, but VDAC1 is the
most abundant (79).
Mouse models of VDAC helped elucidate the distinct roles of each isoform. Although mvdac2 and
VDAC3 deficient mice were viable, VDAC2 was associated with neurodegenerative diseases and
VDAC3 knockout mice developed heart muscle defects due to attenuation in complex IV activity
(80, 81). VDAC3 knockout male mice were also infertile due to their sperm’s reduced flagellum motility (80). On the other hand, VDAC1 knockout mice, are partially embryonic lethal, a study generated mvdac1-/- by deleting exons 2-5. The resulting mutated allele lacks four coding exons.
The heterozygous (VDAC1+/-) mice had no obvious phenotype, but the VDAC1 +/- mice had a reduced number of homozygous deficient mice when mated, by utilizing the Mendelian ratio. That
ratio was further exasperated when VDAC1+/− mice were intercrossed. Around 40% of the
expected number of VDAC1−/− mice and 88% of VDAC1+/−mice were identified (82). The
surviving Vdac1-/- mice demonstrated that VDAC1 plays an important role in transporting
metabolites across outer mitochondrial membrane. For example, the VDAC1 knockout model displayed a decrease in ADP-stimulated oxygen consumption, defects in the electron transport chain, and dysmorphic mitochondria (83). All isoforms were associated with generation of reactive
oxygen species (ROS), steroidogenesis and mitochondria-ER pathways, such as calcium
homeostasis (77). VDAC1 and VDAC2 regulate cytochrome C release but VDAC3 does not (84–
86). In vitro studies have also demonstrated that VDAC1 plays a bigger role in metabolite transportation compared to VDAC2 and VDAC3 (79). For example, unlike VDAC1 and 2,
VDAC1 binds to hexokinase and regulates coupling of mitochondrial respiration and glycolysis
(87, 88)(Figure 1.7).
15 These diverse roles of VDAC require fine tuning and thus multiple factors to modulate its activity.
VDAC opening is regulated by, glutamate, NADH, hexokinase, and several kinases. PKA,
glycogen synthase 3β (GSK3β), and protein kinase C epsilon (pkcε), phosphorylate VDAC and
block or inhibit association of VDAC with other proteins, such as Bax and tbid (77, 79).
Maldanado et al. And other groups have recently shown in vitro, that tubulin physically blocks the
channel, thus potentially limiting substrate exchange between the cytosol and the mitochondria
(73, 74, 89, 90).
Carré et al. Were the first to show that tubulin was associated with the mitochondria (91).
Specifically, there was an enrichment of tyrosinated alpha tubulin in the mitochondria. They subsequently immunoprecipitated VDAC and alpha tubulin, in complex with the channel. Other groups have further confirmed this phenomenon and were able to show that the tubulin-blocked state is still highly ion conductive, e.g. Calcium, and but is impermeable to ADP and ATP (74,
92). The tubulin-blocked state of VDAC causes changes in ATP production and NADH generation. Sacket et al. Have suggested that the C-terminus of tubulin is the main site where interaction with VDAC occurs. Using C-termini’s of alpha or beta tubulin conjugated to BSA, detyrosinated alpha tubulin actively bound to VDAC in an electrophysiological setting, and that was reversed when the C-terminus was deglutamylated, forming a ∆2-tubulin (89). This contradicts Carré et al. Findings, where tyrosinated tubulin was enriched in the mitochondria.
Replicating the BSA experiments in cells is key to clarify tyrosinated tubulin and its interaction with VDAC. In the case of beta-tubulin, alanine on the c-terminus is needed to see physical blockage of VDAC. The blockage is reversed when alanine is removed and replaced with tyrosine.
16 This indicates that VDAC blockage of tubulin is dependent on the post-translational modification
of tubulin. In cells free tubulin, generated by depolymerizing tubulin, regulated mitochondrial
membrane potential and ADP/ATP ratio in the cytosol (73). These tubulin effects on VDAC
activity could potentially affect the metabolic status of cells and may therefore affect the health of
cells and organisms, although this has not been tested in cell or animal models. Since RARRES1
and CCP2 regulate PTM of tubulin and have been identified as potential players in metabolism
associated diseases, I hypothesize that these regulators of the tubulin code might play a role in
blocking the voltage dependent anion channel.
Figure 1.7. Multi-Substrate Transport and Gate of Death. A schematic diagram of the mitochondrial voltage dependent anion channel, situated on the outer mitochondrial membrane. VDAC1 is an important metabolic regulator as it regulates the transport of ATP/ADP and NAD+/NADH. Hexokinase, and other metabolic associated enzymes are mainly associated with VDAC1. Ion flow, ROS, and apoptosis are regulated by all isotypes of VDAC. (Adapted from Magri et al., 2018 (93))
17 RARRES1 and CCP2 in Metabolic Diseases
Gene array expression data demonstrate that RARRES1 is differentially expressed in metabolic
diseases and is associated with biological hallmarks that regulate metabolism (33, 35, 94). One of
the first studies to mention RARRES1 in metabolic disorders was an obesity study where
RARRES1 is the most up-regulated gene in subcutaneous fat from obese human subjects on a diet-
induced weight loss regimen (35). In the same study, RARRES1 was one of the most
downregulated gene (top 15) in adipose tissue during weight maintenance in the obese human
subjects (35). Another study looking at adipocyte differentiation in vitro showed that RARRES1
is a potential marker for adipocyte differentiation and dedifferentiation, where RARRES1
increased in dedifferentiating adipocytes and decreased during differentiation (95). Retinoic acid
signaling, an inducer of RARRES1 transcription, also negatively regulates differentiation of adipocytes. Out of the 530 genes regulated by retinoic acid only 15 were differentially expressed
in differentiated vs. Differentiated adipocytes and RARRES1 was one of the 15 genes. RARRES1
was consistently and markedly upregulated in visceral stem cells compared to differentiated
visceral adipocytes (96). RARRES1 also plays a role in liver fibrosis and liver metabolism. A
recent study has shown that RARRES1 expression is significantly decreased when treated with
fenofibrate, a PPARα agonist, in hepatocyte specific PPAR alpha knockout mouse, which develop liver steatosis, in comparison to wildtype mice (34). Studies manipulating the expression of
RARRES1 demonstrate that this protein is able to induce autophagy and modulate expression of important regulators of energy homeostasis, SIRT1 and mTOR (32, 97).
CCP2 can also regulate autophagy in cervical cancer and several studies have demonstrated genetic
variability in the locus of the AGBL2 gene (CCP2) in diabetic patients (98, 99). Furthermore, a
18 mutation in the AGBL2 gene was identified in a human sample study looking at insulin processing
and secretion in metabolic syndrome associated men (99). However, the mechanism whereby
RARRES1 and CCP2 regulate metabolism is unknown. I hypothesize that their regulation of tubulin PTMs might be the mechanism behind RARRES1 and CCP2 association with metabolic diseases.
The Role of RARRES1 and CCP2 in Tumorigenesis
RARRES1 was first implicated in tumorigenesis following studies that showed that RARRES1
expression was lost in prostate, skin, neck, acute myeloid leukemia, Wilms tumor and colon
cancers (31, 100–103). Loss of RARRES1 expression was strongly associated with RARRES1
hypermethylation in the promoter region of poorly differentiated colon adenocarcinoma, prostate
and nasopharyngeal cancers (100, 101, 104). Jing et al. Demonstrated that transfecting the
RARRES1 gene in PC3 (prostate cancer cells) cells decreased their tumorigenicity in nude mice
and inhibited proliferation and invasiveness in vitro (105). In colon cancer cells, the same
phenomenon was observed (106). RARRES1 was also expressed at low levels in prostate epithelial
cells enriched for stem cells compared to differentiated committed basal cells (94). These findings
define RARRES1 as a class II tumor suppressor in these cancers.
In contrast RARRES1 is overexpressed in hepatocellular carcinoma and pancreatic ductal
adenocarcinoma (107). Datasets in the Oncomine database, as shown in Table 1, show that in both
of these cancers RARRES1 expression is at least 2 FC higher compared to normal samples in at
least 2 publicly available datasets. However studies focusing on RARRES1 in hepatocellular carcinoma, show that RARRES1 is in fact silenced in malignant hepatocytes (33, 108) but is highly
expressed in the activated hepatic stellate cells. Stellate cells are a crucial component in both
19 hepatocellular carcinoma and pancreatic adenocarcinoma. A fibrotic tumor microenvironment is a hallmark for pancreatic and liver tumorigenesis (109). In the case of liver cancer, 80% of tumors are associated with the chronic inflammation that accompanies liver cirrhosis or fibrosis (109).
The acute inflammation activates the stellate cells which alters the extracellular matrix and leads to chronic inflammation and alterations in the extracellular matrix(110). Fibrosis is acknowledged as a protective response to acute injury, but the fibrotic component becomes chronic and dysfunctional when the cause of injury, such as cholestasis or viral infection, endures and cannot be eliminated (111). This explains why most of HCCs are a result from chronic liver diseases, i.e.
Alcoholic liver disease, nonalcoholic liver disease and chronic viral hepatitis (111). With the combination of hepatocyte death, fibrosis and inflammation the disease can progress into cancer.
One study identified RARRES1 as a marker for fibrosis and discovered that RARRES1 expression is especially elevated in activated stellate cells (33). This phenomenon is also seen in lung and kidney fibrosis (33). For pancreatic ductal adenocarcinoma (PDAC), fibrosis is associated with progression of PDAC, where stellate cells interact with the cancer cells to create a favorable niche for the tumor (112). Unfortunately, there are no data identifying RARRES1 as important in pancreatic stellate cells. Nevertheless, since RARRES1 was implicated in liver, lung, and kidney fibrosis, it is possible that RARRES1 expression is high in PDAC because of the level of fibrosis seen in pancreatic cancer etiology. We thus assessed whether pancreatitis, a disease where pancreatic fibrosis is a characteristic feature, and found that RARRES1 was significantly increased
(Figure 1.8). RARRES1 is also highly expressed in triple negative breast cancer (TNBC) (Table
1.1) (Figure 1.8) which is also associated with fibrosis. Several studies have shown that transforming growth factor (TGF- ), a pro-fibrogenic cytokine that promotes stellate cell activation, ligands are enriched𝛽𝛽 in the TNBC𝛽𝛽 tumor microenvironment (113). RARRES1 could be
20 overexpressed in the stromal area of the tumor rather than the breast cancer cells in TNBC. We
used Oncomine to address this speculation. Two datasets were identified that assessed gene
expression in the stroma of invasive breast carcinoma, and in both datasets RARRES1 is
significantly overexpressed (Figure 1.8). The publicly available data clearly point at a role for
elevated RARRES1 in the stromal environment and fibrosis. To the best of our knowledge, there
are no studies addressing the mechanism behind RARRES1 expression in fibrosis. RARRES1 and
its function in cancer associated fibroblasts must be assessed to understand the discrepancy of
RARRES1 expression in TNBC, HCC, and PDAC.
Table 1.1. Oncominetm Analysis: Cancer Summary for RARRES1 Expression.
The Oncomine database was queried for RARRES1 expression in the available datasets based on the following criteria: cancer type, cancer versus normal, cancer versus cancer, cancer subtype, cancer versus baseline, pathway and drug and outlier analyses. The 'red' cells represent RARRES1 overexpression and the 'blue’ cells represent RARRES1 underexpression. The levels of expression are based on the gene rank percentile. This disease summary was performed using a criterion of a 2-fold change for RARRES1 expression and a p-value of 1E-4. (Oncominetm platform (C).
21 Figure 1.8. Comparison of RARRES1 Across Twelve Analyses of Breast Cancer Stroma, TNBC, Pancreatic Cirrhosis/Cancer, and Liver Cirrhosis Datasets. 12 Different datasets were selected from Oncominetm that analyzed the expression of RARRES1 gene in invasive breast cancer with ERBB2/ER/PR Negative status, pancreatic adenocarcinoma, pancreatitis, and liver cirrhosis.
There are only two studies that suggest a role for CCP2, the target of RARRES1, in cancer. One
study identifies CCP2 as a promoter of oncogenesis by driving autophagy through aurora kinase
in hepatocellular carcinoma (98). The group suggests that CCP2 is highly expressed in
hepatocellular carcinoma and ectopic expression of CCP2 enhanced proliferation rate as well as
cell survival in vitro (98). Another study identified CCP2 as an oncogene in gastric cancer. CCP2,
along with RARRES1, regulated proliferation and chemotherapy resistance in cancer cells (114).
Examining AGBL2 gene expression in different cancers, there is no significant differential
expression of AGBL2 in cancers (Table 1.2). Since AGBL2 is an enzyme, gene expression might
not be the best approach to identify the clinical relevance of AGBL2 in cancer. Its enzymatic
activity needs to be assessed instead.
22 Table 1.2. Oncominetm Analysis: Cancer Summary for AGBL2 Expression.
The Oncomine database was queried for AGBL2 expression in the available datasets based on the following criteria: cancer type, cancer versus normal, cancer versus cancer, cancer subtype, cancer versus baseline, pathway and drug and outlier analyses. The 'red' cells represent AGBL2 overexpression and the 'blue’ cells represent AGBL2 underexpression. The levels of expression are based on the gene rank percentile. This disease summary was performed using a criterion of a 2-fold change for AGBL2 expression and a p-value of 1E-4. (oncominetm platform (C).
23 Statement of Purpose
Bioenergetic dynamics are emerging as a basis for understanding the pathology of cancer. As
malignancy progresses its metabolic pattern and mitochondrial respiration become more
dysfunctional. The pathways that regulate bioenergetic dynamics are poorly understood, and the
characterization of proteins implicated in those processes must be assessed. One understudied
protein and type II tumor suppressor is retinoic acid receptor responder 1 (RARRES1). RARRES1
is induced by retinoic acid and vitamin D (two major metabolic regulators) (45, 115) and was recently linked to metabolic associated diseases.
The biological function of RARRES1 and how that relates to its tumor suppressor traits is still
poorly understood. Recent findings in our laboratory show that RARRES1 binds to EB1, a mitotic- spindle associated protein. RARRES1 was also simultaneously bound to CCP2, a glutamate carboxypeptidase (45). These proteins are thought to be essential in regulating microtubule dynamics. Since CCP2 is a glutamate carboxypeptidase, it was found to remove glutamate residues at the c-terminal end of -tubulins, whereas RARRES1 inhibits this activity and increases glutamylation of the c-terminal𝛼𝛼 (Figure I). Glutamylation of -tubulin is a topic that is understudied, but some studies indicate that alpha and beta tubulin dimers𝛼𝛼 are important in binding to and blocking the voltage-dependent anion channel (VDAC). The removal of glutamate on the c-terminus of alpha tubulin reverses the VDAC blockage (89). VDAC is located in the outer
membrane of the mitochondria and regulates ATP and ADP substrate exchange between the
cytoplasm and the mitochondria. Thus regulation of VDAC blockage controls substrate
availability for bioenergetic pathways like mitochondrial respiration (74), and such regulation will
control the overall oxidative phosphorylation activity. We propose that CCP2 and RARRES1,
24 could attribute to the metabolic regulation by tubulin via modulating the activity of voltage dependent anion channel. Our studies might help elucidate mechanism whereby RARRES1 and
CCP2 regulate metabolic associated diseases and cancer.
25 Specific Aims
Aim 1 To characterize the role of glutamylation changes of tubulin by RARRES1 and CCP2 on the regulation of mitochondrial voltage dependent anion channel activity.
Aim 2 To characterize the role of RARRES1 in fatty acid metabolism in epithelial cells.
Aim 3 To identify mechanisms of endogenous transcriptional regulation of the RARRES1 gene.
26 Chapter 1-Tumor Suppressor RARRES1 Links Tubulin Deglutamylation to Metabolic Reprogramming and Cell Survival
Introduction
Cancer cells and stem cells share several common features including similar changes in
metabolism and resistance to mitochondria mediated cell death (116–118). Many
oncogenic pathways influence these processes, usually indirectly following multistep alterations
in metabolic enzymes (75). There is growing interest in the notion that directly targeting
glycolysis or mitochondrial metabolic function may influence cancer in much the same way that
metformin can be used to treat diabetes (9). Indeed metformin, a mitochondrial respiration
inhibitor, has quite striking anti-cancer activities, although these may involve additional
metabolic targets (117, 119). However, with the exception of LKB1, which regulates the cytoplasmic ATP sensor AMPK, no tumor suppressor genes are known to directly regulate glycolysis or mitochondrial metabolic activity (120).
RARRES1, a retinoic acid regulated tumor suppressor gene associated with ageing, metabolism and stem cell differentiation is among the most commonly methylated loci in multiple cancers
but has no known molecular function (20, 36, 104). Like LKB1, the RARRES1 homologue
latexin is essential for hematopoietic stem cell differentiation and ageing (121). RARRES1 and
latexin are putative carboxypeptidase inhibitors and we showed earlier that RARRES1
interacts with cytoplasmic carboxypeptidase 2 (CCP2/AGBL2) (45). Both RARRES1 and
CCP2 have been associated with metabolic diseases and several studies have identified the
proteins as important
27 regulators of autophagy (32, 97–99). CCP2 is a member of the CCP family of deglutamylases
important for the removal of glutamic acid residues from the C-terminal tail of several tubulin
isoforms (53, 71, 122, 123). Glutamylated and polyglutamylated tubulin is enriched in mitotic
spindles and other structures, such as axonemes/cilia that contain arrays of stable microtubules
(54, 71, 124). Although ccps have not been associated with cancer, the enzymes that modify
tubulin (TTL and ttlls) and detyrosinated tubulin have (123). Peptide mimics of the acidic C-
terminal tail of tubulin can also directly influence the activity of mitochondrial voltage dependent
anion channels (VDAC) and mitochondrial membrane potential, raising the possibility that
pathways that alter its acidic C-terminal tail could influence mitochondrial activity directly by
influencing VDAC function (74, 89, 90). We now show that the metabolic and tumor suppressor
effects of RARRES1 are mediated by its inhibition of CCP2 catalyzed tubulin deglutamylation, which in turn regulates mitochondrial bioenergetics and subsequently alters energy homeostasis
by modulating the function of the mitochondrial voltage-dependent anion channel 1 (VDAC1).
Methods
Cell Culture
PWR-1E, MCF10A, MDA-MB-231, MDA-MB-157, Human neonatal foreskin primary epidermal
keratinocyte (HFK) cells were maintained according to the recommendation of American Type
Culture Collection. Human pancreatic stellate cells (hpastec) were cultured and maintained
in Stellate Cell Medium (sciencecell Research Laboratories) according to supplier’s directions.
Antibodies and Reagents
All-trans-RA and Erastin were obtained from Sigma-Aldrich (St. Louis, MO). Primary antibodies
targeting the following antigens were used: rabbit anti-human RARRES1, mouse B3-
polyglutamylated tubulin, mouse anti-β actin, and mouse anti α-tubulin were from Sigma-Aldrich
28 and GT335-polyglutamylated tubulin was from Adipogen andΔ-2 tubulin antibody was from
Upstate. Phosphorylated AMPK and total AMPK antibodies were from Cell Signaling.
SirnasiRNA, Constructs,Expression Expression and TransfectionConstructs, and Transfection
Codon-optimized full-length human RARRES1 cloned into bglii and hindiii sites in the peyfp-N1
vector (Clonetech) and pglue vector were provided by Genscript. AGBL2 (CCP2; Myc-DDK-
tagged) construct was from Origene Technologies, Inc. (Cat #RC206949). ON-targetplus
smartpool Human RARRES1 (Cat #L-009925-00), sigenome smartpool Human AGBL2 (CCP2)
(Cat #M-012937-00) and Non-Targeting control sirna (Cat #D-001210-01-20) were from Thermo
Scientific Dharmacon. Cells were seeded at 100,000 cells per well in 6-well plastic tissue culture
dishes on day 0. On day 1, for each cell sample, 7.5μl of Lipofectamine 3000 (Invitrogen,
Carlsbad, CA) was added to 125μl of OPTI-MEM I (Gibco-Invitrogen, Grand Island, NY) in a
sterile eppendorf tube, and mixed. 2.5μg of DNA or 75pmol of siRNA in 125ul of OPTI-MEM were added with or without 5ul of P3000, mixed, and incubated for 5 minutes at room temperature. The mixture was then added dropwise to cells.
RARRES1/CCP2 shRNA shRNA
Constructs for stable depletion of RARRES1 and CCP2 were obtained from the RNAi Consortium
(Moffat et al., 2006) via SIGMA-Aldrich (NM_002888 for RARRES1 and NM_024783 for
CCP2). Five pre-made constructs were obtained for both RARRES1 gene (Clone ID:
TRCN0000063373, TRCN0000063374, TRCN0000063375, TRCN0000063376,
TRCN0000063377) and CCP2 gene (Clone ID: TRCN0000073908, TRCN0000073909,
TRCN0000073910, TRCN0000073911, TRCN0000073912) and individually tested to identify those able to achieve efficient knockdown at the protein level. Negative control constructs in the
29 same vector system (vector alone, plko.1 puro) were obtained from SIGMA-Aldrich
(NM_003177).
Western Blot
Cells were washed with PBS and lysed in RIPA lysis buffer (0.25% sodium deoxycholate, 0.5%
NP-40, 1% Triton X-100, 1mm EDTA, 50mm Tris-HCL ph7.5, 150 mm nacl) for 30 minutes on
ice. Protein concentration was determined using protein microplate assay (Bio-Rad Laboratories).
5-25μg of cell lysates were separated on 4-12% SDS/PAGE, transferred to PVDF membrane
(Millipore), blocked with 5% BSA-PBS, and incubated in primary antibody in 1% BSA-PBS
overnight at 4°C. Blots were washed 3 times for 5 minutes each in PBS-0.1% tween 20, followed
by incubation in HRP-conjugated secondary antibody (KPL, Gaithersburg, MD) for 1 hour at room
temperature on an orbital. Blots were washed 3 times in PBS-0.1% tween 20, followed by 1 wash
in PBS. Detection of immunoreactive bands was carried out using chemiluminescence with ECL
Western Blotting Detection Reagents (Amersham, Piscataway, NJ).
Immunofluorescence
Cells were seeded at 1X105 and grown on glass coverslips for 24 hours. 24 or 48 hrs post-
transfection, cells were fixed with freshly prepared 3% paraformaldehyde for 20 minutes and
permeabilization of cells is then performed on ice, in PBS containing 0.5 % Triton X-100 for 3-5
min. Cells were blocked for 1 hour in blocking solution (B/B solution= 1:1 ratio of 2% BSA in
PBS and 1x Animal-Free Blocker (VECTOR Laboratories Cat # SP-5030) and then incubated in
primary antibodies diluted in B/B solution overnight at 4oc. Cells were washed 3 times in 1X PBS
and incubated in secondary antibodies conjugated to Alexa-488 and Alexa-568 fluorophores
(Molecular Probes-Invitrogen, Carlsbad, CA). For permanent mounting, use Molecular Probes,
30 Inc. “prolong Antifade Kit” and slides were observed and photographed using the Zeiss LSM 510
confocal microscope and Olympus BX 61 DSU Fluorescent microscope.
Cell Cycle Analysis and Flow Cytometry
Cells were fixed in 70% ethanol and stained in PBS containing 0.1% Triton X-100, 50 μg/ml
RNase, and 50 μg/ml propidium iodide. Samples were analyzed by flow cytometry and the
hypodiploid peak constituted the apoptotic population in this analysis. DNA content was measured
on a facsort flow cytometer (Becton Dickinson, Franklin Lakes, NJ), and data were analyzed using
modfit software (Verity Software House, Topsham, ME) in LCCC Flow Cytometry and Cell
Sorting Shared Resource. At least 1x106 cells were analyzed per sample. For Annexin V Labeling,
samples were stained with fluorescein-labeled Annexin V and propidium iodide (Trevigen) according to the manufacturer's protocol and analyzed by the LCCC Flow Cytometry and Cell
Sorting Shared Resource. The two Annexin V positive quadrants of the analysis were taken as the apoptotic fraction. For Anoikis Assays, confluent cells were trypsinized into a single cell suspension. 700,000 cells were plated in 150-mm tissue culture dishes coated with 0.8% agarose, to which they could not attach. At the various time points, the cells were collected, washed in PBS, and any cell aggregates were dispersed by trypsinization, and cells were then analyzed for apoptosis. To perform the analysis of cell surface markers (CD44+/CD24-, CD49+/CD24-),
MCF10A cells were detached with trypsin, washed in blocking buffer (PBS containing 3% FBS), then stained with anti-CD24-PE (BD Biosciences, San Diego, CA, USA) and anti-CD44-PE-Cy5
(BD Biosciences) or anti-CD49-FITC (BD Biosciences) using 1 ml of antibody per 106 cells, and
incubated at room temperature for 1 hr. Following incubation, cells were washed twice with 1 ml
PBS. Cells were re-suspended in 1 ml PBS and then were measured by LCCC Flow Cytometry
and Cell Sorting Shared Resource.
31 Soft Agar Growth Assay
5,000 cells were plated in 0.3% agar layered on top of 0.6% agar in 6 well plate. After 2 weeks,
colonies were stained with 0.005% crystal violet and pictures taken to count and measure colonies
size using metamorph Image analysis software 7.0.
Statistical Analysis
Results are shown as the mean ± S.E.M. Statistical significance was calculated by using Student's
t-test and P < 0.05 was accepted as significant value.
Zebrafish Animal Model
Zebrafish (Danio rerio) were raised, maintained and crossed as described previously (125) .
Development of embryos was at 28oc and staging was determined by both hpf and morphological
characteristics (126). All procedures were in accordance with NIH guidelines on the care and use
of animals and were approved by the Georgetown University Institutional Animal Care and
Use Committee. For Morpholino oligonucleotide (MO) injection experiments, two mos were
used to knockdown CCP2 (Agbl2); abgl2 MO1 was designed to block translation, and, agbl2
MO2 was designed to block splicing. The following mos were synthesized by Gene-Tools, LCC: ,
5’-AAGGATCGTCTGATTTCTTCTGCAT; agbl2(CCP2) MO2, 5’-
CCGTGTTGTCATTATAAATGAACGC; and tp53 MO, 5’-
GCGCCATTGCTTTGCAAGAATTG. The Standard Control MO from Gene Tools was used as control. Solutions consisting of 4 ng/nl MO plus 0.5% tetramethyl rhodamine dextran in dh20 were microinjected into one to four cell stage embryos. Acridine orange staining was done as previously described (127). Briefly, embryos were manually dechorionated, anesthetized in 0.016% tricaine
(Sigma), and then incubated in 50ng/ml acridine orange (Invitrogen, A1301) for 1hr. Following several washes in fish water, the embryos were mounted in 3% methyl cellulose and imaged.
32 Images were acquired using a Zeiss Axioplan2 microscope fitted with an axiocam camera using
axiovision software, or a Zeiss dissecting microscope fitted with a Cannon Powershot A590
digital camera. Whole-Mount In Situ Hybridization (WISH) was performed as previously
described using DIG-labeled riboprobe synthesized from zebrafish lxn cdna (genbank:
BC086837) (128).
MitoTracker and TMRM Assay
For MitoTracker staining, cells were seeded at 1X105 and grown on glass coverslips for 24 hours.
24 or 48 hrs post-transfection, cells were stained with MitoTracker@Deep Red FM reagent from
Life Technologies and more detailed the procedures were followed according to supplier’s
directions. For TMRM assay, cells were seeded at 1X104 and grown on chamber slides (Lab-Tek
II Chambered Coverglass W/Cover #1.5 Borosilicate Sterile, Nalge Nunc International) for 24
hours. 24 or 48 hrs post-transfection, Tetramethylrhodamine methylester (TMRM), a lipophilic
cationic fluorescent redistribution dye was applied to detect the mitochondrial membrane
potential (Δψm) fromTMRE & TMRM Assay Kits (immunochemistry Technologies). Cells in
complete growth medium were loaded for 20 min at 37oc with 70 nm of TMRM and shakingis
very important to have homogeneous staining of the images in this step. More detailed
procedures followed the supplier’s directions. TMRM-loaded cells were incubated in 1x Assay
o Buffer in humidified 5% CO2/air at 37 c and imaged with a Zeiss LSM 510 confocal
microscope (Thornwood, NY). Both Integrated Brightness of M itoTracker and TMRM assay
were measured and analyzed by Keyence BZ-X Analyzed software.
Extracellular Flux Assays
Bioenergetics profile of transient RARRES1 knockdown in MCF10A and PWR-1E cells
and stable RARRES1 knockdown MCF10A cells were measured using the xfe96
Extracellular Flux Analyzer. Oxygen consumption rates and response to mitochondrial stress
factors were 33 analyzed using the XF Cell Mito Stress Kit. 48 hours prior to analysis, MCF10A and PWR-1E
cells were transfected with RARRES1 siRNA or scramble siRNA in duplicates. The night before
the assay, the cells were seeded at an optimized cell density (MCF10A cells: 10,000 and
PWR-1E cells: 20,000) in the 96-well xfe plate and incubated at 37oc with 5% CO2 overnight.
o The day of analysis, the cells were incubated at 37 c in a CO2-free atmosphere incubator for 1
hour. Basal oxygen consumption rate and extracellular acidification rate were measured.
Subsequently, oxygen consumption rate (OCR) and extracellular acidification rate (ECAR)
responses were observed after separate injections of oligomycin (1 um), carbonyl cyanide-p-
trifluoromethoxyphenylhydrazone FCCP (0.5 um for MCF10A cells or 0.25 um for PWR-1E
cells), and a combination of rotenone and antimycin A (0.5 um) were respectively prompted in
the assay. For each injection, there was a total of 3 cycles, each one lasted 3 minutes and a
measurement was taken at the beginning of each cycle. Glycolysis was also analyzed, using the
XF Glycolysis Stress Test kit, in the xfe96 Extracellular Flux Analyzer. Cells were transiently
knocked down with CCP2 siRNA or scramble negative control siRNA and plated at the same optimal density as the XF Cell Mito Stress Test protocol. ECAR was measured at baseline and after adding glucose (10 mm), oligomycin (1um) and 2-deoxy-d-glucose (100 mm) respectively.
All experiments were analyzed in duplicates for at least two independent experiments and the results were normalized to their protein concentration. Glycolytic Rate Assay was also run on transiently CCP2 depleted cells and control cells (transfected with scrambled siRNA).
Cells were grown in media with 10 mm glucose prior to Seahorse extracellular flux assay.
During the flux assay, ECAR was measured at baseline and after injecting antimycin A and
rotenone (0.5 um) and 2-deoxy-d-glucose (100 mm).
34 LC-MS Analysis and Metabolite Identification Ultra-performance Liquid Chromatography-Time of Flight Mass Spectrometry (UPLC TOF MS) based metabolomic analyses was done on RARRES1 stable knockdown and the control empty vector MCF 10A cells. Refer to Maimouni et al., for the running, metabolite
extraction, data preprocessing, and statistical analyses methods. The significantly altered metabolites (Table 1) were putatively identified via accurate mass based search using the Madison Metabolomics Consortium Database (MMCD), the Human Metabolome Database (HMDB) and LIPID MAPS. The metabolite identifications were confirmed by comparing the retention time under the same chromatographic conditions and by matching the fragmentation
pattern of the parent ion from the biological sample to that of the standard metabolite using tandem mass spectrometry (UPLC-TOFMS/MS).
Results
RARRES1, CCP2 and Retinoic Acid Regulate Tubulin Glutamylation
RARRES1 interacts with AGBL2/CCP2 (CCP2), a member of the CCP family of
carboxypeptidases responsible for post-translational modifications of the C-terminal region of tubulin (45). Although ccps are most commonly associated with ciliated organs, non-ciliated cells
exhibit varying glutamylated forms of tubulin and CCP2 is expressed in many cancer cells (45).
Figure S1.1 shows that several human cancer and normal cells express significant CCP2 and
demonstrate its successful depletion. However CCP1, which is highly expressed in neuronal
tissues, was barely detectable in only MDA-MB-231 and BT549 breast cancer cells (129) (Figure
2.1). CCP2 has many splice variants, some of which do not contain the catalytic domain (Figure
2.2). The qPCR primers used in this study and our previous work only detect forms of CCP2 that
contain the catalytic domain (Figure 2.2(45)). CCP2 can remove the penultimate glutamate from
35 tubulin to form Δ2-tubulin, an isoform that can no longer be re-tyrosinated and which accumulates
in neurons and in cancer cells (130). Consequently CCP2 action could indirectly change the
relative ratio of tyrosinated and detyrosinated tubulin without actually acting as a detyrosinase (45,
71, 131). Alternatively, changes in glutamylation could alter microtubule-severing activity or prevent detyrosinated and deglutamylated tubulin from being retyrosinated (132, 133). In addition to forming Δ 2-tubulin CCP2 also removes glutamic acid residues from polyglutamylated side
chains containing one or more glutamic acids (71). Figure 1.1 shows for the first time that
RARRES1 and its major regulator, retinoic acid (RA), decrease the level of Δ 2-tubulin and
increase side chain glutamylation of tubulin in primary human keratinocytes and several different
normal and cancer cell lines by inhibiting CCP2 activity. Importantly the effect of RA on tubulin
glutamylation is dependent upon RARRES1. We used two poly-glutamylated tubulin antibodies,
GT335, which recognizes side chains containing one or more glutamic acids, and B3, which
detects side chains containing two or more glutamic acids (134). Both antibodies consistently
showed that CCP2 transient expression reduced glutamylated tubulin levels and that its depletion
increased levels. Consistent with a role for RARRES1 as a CCP2 inhibitor, its transient expression
increased glutamylated tubulin and its depletion reduced it. Similar results were obtained by
immunostaining of cells following RARRES1 or CCP2 depletion (Fig. 2.3). These data strongly
implicate RA and RARRES1 in the regulation of CCP2-mediated deglutamylation (Figure 2.4G).
36 Figure 2.1. CCP2 Expression in Tumorigenic and Non-Tumorigenic Cell Lines. (A) Relative mRNA expression of CCP1 and CCP2 in Non-tumorigenic cell lines, MCF10A, HFK, and PRWR- 1E. CCP2 expression is regulated by cell density in MCF10A cells. Depicted as (low) and (high). Low stands for low density and high stands for high density. (B) Relative mRNA expression of CCP2 and CCP2 in Tumorigenic cell lines, MDA-MB-157, MDA-MB-231, BT547, and HS578T (C) CCP2 knockdown in several cell lines. These results were confirmed by qPCR (data not shown).
37 Figure 2.2. CCP2 Splice Variants and Primer Profiling. The CCP2 gene contains over 13 variants six of which are translated (135, 136). The red and black asterisks are splice variant regions predicted by ENSEMBL, where red is the splice donor and the black signifies the splice acceptor. Our primer sets, we were able to detect four of the known translated variants. This includes the full length variant and the catalytically viable version with a truncated c-terminus (exons 19,18, and 17), three processed transcripts and one nonsense-mediated decay transcript. Our polyclonal antibody detects five out of the six translated splice variants.
Figure 2.3. Immunocytochemistry of Polyglutamylated Tubulin. (Antibody B3) in RARRES1 Depleted (A) and Expressing PWRE-1 Cells (B) are displayed in the immunocytochemistry images.
38 Figure 2.4. RARRES1, CCP2 and Retinoic Acid Regulate Tubulin Glutamylation. (A) Δ-2 tubulin levels are regulated by RARRES1 in MDA-10A, MDA-231, HFK, PWR-1E cells and in RARRES1-/- mouse skin. (B) RARRES1 and polyglutamylated tubulin is increased by retinoic acid (10-7M all-trans-RA). Western blots in HFK (human foreskin epidermal Keratinocytes), PWR-1E, and HPaSteC (Human Pancreatic Stellate Cell) cells. (C and D) Immunoblot for RARRES1 and polyglutamylated tubulin in MCF10A, MDA-MB-231, HFK and PWR-1E cells in which RARRES1 was exogenously expressed or depleted. (E) Immunoblot of CCP2 and polyglutamylated tubulin in CCP2 knockdown MCF10A and MDA-MB-231 cells. (F) Immunoblot of RARRES1 and polyglutamylated tubulin after RARRES1 manipulation with or without retinoic acid in HFK and PWR-1E cells. (G) Schematic illustrating the effects of RA and RARRES1 in CCP2 regulation of tubulin deglutamylation.
39 RARRES1 and CCP2 Reciprocally Regulate Sensitivity to a Variety of Apoptotic Stimuli
RA is a pleiotropic factor with growth inhibitory, stem cell differentiating and apoptotic properties in normal and cancer cells (137). The effects of RA on migration and invasion of prostate cancer cells is dependent on RARRES1 (94, 105). We next asked if RARRES1 influenced the proliferative and apoptotic effects of RA. RARRES1 manipulation did not affect MCF10A cell proliferation or cell cycle in the time frame of the experiment but cell death induced by high doses of RA was reversed by depletion of RARRES1 (Figures 2.5A, B and 2.6). Depletion of RARRES1 reduced apoptosis and exogenous expression of RARRES1 enhanced apoptosis even in the absence of RA (Figure 2.5C). Consistent with the function of RARRES1 as an inhibitor of CCP2, knockdown and expression of CCP2 had the opposite effect and increased and decreased apoptosis respectively (Figure 2.5D). As RARRES1 depletion inhibited both basal and RA induced cell death we wondered if it might be a global regulator of mitochondria mediated cell death. MCF10A cells are exquisitely sensitive to detachment-induced cell death (anoikis) and express both RARRES1 and CCP2 (Figures 2.4 and 2.1) (135) . Anoikis analyses confirmed that most control MCF10A cells died rapidly if they were not attached to a substrate (Figure 2.5E). In contrast survival of detached MCF10A cells in which RARRES1 was stably or transiently depleted was markedly enhanced (Figure 2.5E). Similar results were observed in MCF10A cells exogenously expressing
CCP2 pointing to a role for RARRES1 inhibition of CCP2-mediated tubulin glutamylation in the regulation of cell death. Detached RARRES1 depleted cells also formed spheroids while in suspension. Cell death induced by treatment with the microtubule stabilizing drug taxol was also inhibited by RARRES1 depletion and exogenous expression of CCP2 (Figure 2.5F). Importantly, the effects of RARRES1 depletion were abolished when CCP2 was also depleted, strongly suggesting that inhibiting CCP2 function mediated the effects of RARRES1 on cell survival.
40 CCP2 knockdown in zebrafish embryos resulted in transient cell death, slightly stunted embryos but no specific overt embryological abnormalities (Figure 2.5G). CCP2 MO1 was designed to block translation and CCP2 MO2 to block splicing. Injection of either MO resulted in identical dose dependent phenotypes. CCP2 knockdown resulted in a distinct cell-death phenotype. By 24 hpf, tissue became opaque and granular, particularly in the ventral midbrain and hindbrain regions suggesting significant cell-death (Figures 1.2G). The cell death phenotype was confirmed by staining apoptotic cells with acridine orange. Embryos were co-injected with the CCP2 MO1 and a tp53 MO since p53 knockdown is known to suppress potential MO induced off-target effects on apoptosis (139). Counts of acridine orange stained cells at 24 hpf indicated a 10X increase in apoptosis in CCP2 knockdown embryos compared to controls (Figure 2.5H). This phenotype was transient however, as by 48 hpf, tissues were no longer opaque and granular. Differentiation of the remaining cells appeared to be fairly normal as all tissue types were present, although the embryos were small and somewhat dysmorphic overall (Figure 2.5G). This phenotype is similar to that previously reported in a smaller number of CCP2 morphants (140). In our hands this was the dominant phenotype.
41 Figure 2.5. RARRES1 and CCP2 Regulate Apoptosis but not Cell Proliferation. (A) Cells were stained with propidium iodide (Sigma) and analyzed by flow cytometry. % S-phase cells are expressed as fold change in MCF10A cells exogenously expressing RARRES1 or following transient or stable knockdown. (B) RARRES1 depletion inhibits retinoic acid induced apoptosis in MCF10A cells. (C) RARRES1 exogenous expression, transient and stable depletion, regulate basal levels of apoptosis. Samples were stained with fluorescein-labeled Annexin V and propidium iodide and analyzed by flow cytometry. (D and E) RARRES1 and CCP2 have reciprocal effect on anoikis-induced apoptosis. Fold change (D) and phase-contrast photographs of colonies (E) of RARRES1 or CCP2 depleted cells following suspension-induced apoptosis (anoikis) in MCF10A cells. (F) Taxol-induced cell apoptosis. % cell death after RARRES1 or CCP2 manipulation, with, or without taxol (C-control, T-taxol treatment). (G) CCP2 knockdown causes transient cell death in zebrafish embryos. (1-10) Phenotype following injection of 8 ng control MO (1, 3, 5, 7, 9), or 8 ng agbl2 MO (2, 4, 6, 8, 10). (H) Quantification of CCP2 knockdown induced cell death measured by acridine orange staining.
42 Figure 2.6. FACS Analysis of Annexin Stained Cells. (A) RARRES1 knockdown MCF10A without or with retinoic acid in dose dependent manner. (B) MCF10A controls, RARRES1 transient knockdown, and RARRES1 stable knockdown, and RARRES1 overexpression. Samples were stained with fluorescein-labeled Annexin V and propidium iodide (Sigma) and analyzed by flow cytometry to measure apoptosis. Experiments were repeated three times with consistent and repeatable results.
43 RARRES1 and CCP2 Reciprocally Regulate Anchorage Independent Growth and Are Associated with Stem Cell Differentiation
Resistance to anoikis is a requirement for anchorage-independent colony formation, a cardinal
indication of transformation and is also a characteristic of many stem cells (141–143). To test if
RARRES1 could regulate colony formation in soft agar we first created stable RARRES1
knockdown cells (see Materials and Methods). Parental and PLKO control vector MCF10A cells
formed very few colonies after 2 weeks in soft agar. Loss of RARRES1 was associated with
enhanced colony growth within soft agar (Figures 2.7A, B and C). As control MCF10A cells do
not form colonies in soft agar we cannot use them to measure inhibition of colony formation.
Instead we chose a breast cancer cell MDA157, which forms many colonies in soft agar. These
cells express CCP2 but not RARRES1 and stable expression of RARRES1 in MDA157 and other
cells without endogenous RARRES1 causes cell death prior to establishment of stable
transfectants. However stable CCP2 knockdown reduced both the size and number of colonies in
MDA157 cells (Figure 2.7A, B and C).
RARRES1 and its homologue latexin have been implicated in the regulation of stem cell
differentiation and ageing (36, 94). Breast epithelial stem cells are characterized by, high CD49f
(α6-integrin) expression. Consistent with a role in the regulation of stem cell differentiation depletion of RARRES1 from MCF10A cells resulted in a 3-fold increase in cell surface CD49f
(Figure 2.7D, 2.8). Conversely, depletion of CCP2 reduced CD49f cell surface expression and
completely prevented the stimulatory effects of RARRES1 depletion. Taken together these data
demonstrate that RARRES1 and CCP2 reciprocally regulate, expression of stem cell marker
CD49f, anoikis, anchorage independent growth and sensitivity to a variety of apoptotic stimuli in
44 cells and zebrafish. Moreover depletion of CCP2 reversed the effects of RARRES1 depletion consistent with a role for RARRES1 as an inhibitor of CCP2 function. As CCP2 is a tubulin deglutamylase and RARRES1 both interacts with CCP2 and regulates tubulin glutamylation these data are consistent with a role for tubulin glutamylation in stem cell differentiation and the regulation of cell survival. How these functions might involve tubulin deglutamylation is not immediately obvious.
Several studies have pointed to an association of glutamylated tubulin, tubulin glutamylases (ttlls) and the CCP2 product Δ2-tubulin with cancer and resistance to treatment, but the molecular mechanism(s) or the result of these associations is completely unknown (53, 123). Taken together with its effects on apoptosis shown here and elsewhere, these data point to a role for RARRES1- mediated changes in tubulin glutamylation in the general regulation of metabolism and mitochondrial function. Consistent with this possibility are biophysical studies that show a marked effect of BSA conjugated to tubulin-like C-terminal peptides on the activity of the mitochondrial voltage dependent anion channel (VDAC) and mitochondrial membrane potential. This activity is reversed when both the C-terminal tyrosine and the penultimate glutamic acid are removed to form a Δ2-tubulin-like C-terminal or if cells are treated with erastin, which displaces tubulin from
VDAC (89). The studies complement multiple earlier demonstrations that certain isoforms of tubulin are often found associated with mitochondria(91). The regulation of the VDAC is likely to influence oxidative phosphorylation and energy balance, functions that are important in stem and cancer cell function as well as resistance to apoptosis. As the pleiotropic effects of RARRES1 involve the regulation of tubulin C-terminal modifications by CCP2 (including the formation of
Δ2-tubulin) we next asked if RARRES1 could alter mitochondrial membrane potential.
45 Figure 2.7. RARRES1 and CCP2 Regulate Anchorage Independent Cell Growth and CD49f. (A) RARRES1 depletion increases MCF10A colony number and size. (B) CCP2 depletion decreases MDA-157 colony number and size. (C) Phase-contrast images of colonies formed following RARRES1 or CCP2 manipulation. (D) RARRES1 and CCP2 reciprocally regulate stem cell marker, CD49f. Flow cytometry of CD49f labeled cells.
46 Figure 2.8. RARRES1 and CCP2 Reciprocally Regulate Stem Cell Marker CD49f. (A) Results for the flow cytometry for CD49f demonstrates increased cell surface CD49f following RARRES1 knockdown by siRNA and culture of MCF10A cells. (B) CCP2 knockdown reduces CD49f cell surface expression.
RARRES1 Regulates Mitochondrial Membrane Potential (MMP) and AMPK Activation
MMP like other aspects of mitochondrial function is finely tuned to the proliferative and energetic status of the cell; vdacs that are too active might lead to dangerous levels of reactive oxygen
species, vdacs that are not active enough could lead to a change in ATP/ADP ratio and a switch to
aerobic glycolysis (the Warburg effect). This “Goldilocks” like aspect of VDAC regulation
perhaps explains the existence of multiple pathways that regulate its activity (73, 89). We
wondered if the effects of RARRES1 on tubulin glutamylation might contribute to the regulation
of the VDAC and influence MMP. To test this, we first imaged RARRES1 and CCP2 manipulated
cells with mito-tracker. This dye changes color depending on the ph of the mitochondrial lumen
and provides an indirect (and rough) assessment of MMP. Figure 2.9A-D shows that alterations
in mitochondrial ph (and morphology) take place in RARRES1 depleted cells. To further
47 investigate this we labeled cells with the MMP-sensitive dye TMRM and quantified the effects of
RARRES1 depletion on MMP (Figure 2.10). One caveat of our depletion studies is that RARRES1 and CCP2 knockdowns are rarely 100% (usually in the range of 75-90%, Figure 2.1) and cannot address any graded response, such as that observed with the effects of glutamylation on microtubule severing (132). Nevertheless, these data confirm that depletion of RARRES1 significantly increased MMP likely indicative of an increase in oxidative phosphorylation. RA treatment also decreased MMP in a RARRES1 dependent manner. In contrast CCP2 depletion reduced MMP and its exogenous expression increased it. To confirm that RARRES1/CCP2 regulation of MMP involved VDAC we used erastin. Although originally discovered in a screen for inhibitors of ras signaling, erastin is now known to be a VDAC interactor that displaces acidic tubulin-like peptides from the VDAC (73, 144, 145). Co-treatment with erastin inhibited the decrease in MMP following CCP2 depletion providing strong evidence that the effects of
CCP2/RARRES1 on MMP are mediated by regulation of VDAC activity by tubulin glutamylation
(Figure 2.10H).
Figure 2.9. MitoTracker Staining. (A and B) MitoTracker staining in PWR-1E control, RARRES1 and CCP2 manipulated cells. (C and D) Integrated brightness of MitoTracker measured by Keyence analysis software.
48 Figure 2.10. RARRES1 Regulates Mitochondrial Membrane Potential (MMP). (A-D) MMP measurement by, TMRM assay after RARRES1/CCP2 manipulation, retinoic acid or erastin treatment, in MCF10A, HFK, and PWR-1E cells. Fluorescent images of cells treated with the indicated reagents. Integrated brightness of TMRM assay measured by Keyence analysis software and converted to fold change.
49 RARRES1 and CCP2 Control Mitochondrial Respiration by Regulation of VDAC1- Mediated Intracellular Substrate Availability
To further test the idea that the RARRES1/CCP axis affects cellular energetics and metabolism we
examined the effects of RARRES1 manipulation on oxygen consumption rate, extracellular
acidification and metabolic flexibility. To determine whether RARRES1 and CCP2 regulate
mitochondrial activity and glycolytic potential, they were silenced and oxygen consumption rates
(OCR) and extracellular acidification rates (ECAR) analyzed. We were initially concerned that
any data we generated from RARRES depleted stable cell lines may reflect an indirect, adaptation response to long term RARRES1 knockdown. This was not the case however, as in all experiments the results from either stable or transient knockdowns were similar. Silencing RARRES1 in both
MCF-10A and PWR-1E cells improved their energetic status (Figure 2.11). Basal OCR increased suggesting that RARRES1 depleted cells rely more on mitochondrial respiration and increased
ATP-turnover (Figure 2.11A). The spare respiratory capacity increased, which indicates higher substrate availability with which to conduct mitochondrial respiration and electron transport chain activity. The uncoupling response (UR) OCR, measured after oligomycin injection, was significantly increased compared to controls. This indicates an increase in ATP production, most likely to supply the high energetic demand that RARRES1 depleted cells need to conduct anabolic reactions. To confirm that the interaction of RARRES1 with CCP2 contributes to its effects on mitochondrial respiration and glycolytic capacity, we depleted MCF10A and PWR-1E cells of
CCP2 (Figure 2.11B). Consistent with a role for RARRES1 as an inhibitor of CCP2, CCP2
knockdown displayed the opposite phenotype of RARRES1 depleted cells. In contrast to RARRES1
knockdown CCP2 depletion decreased the uncoupled respiration rate and spare respiratory
capacity and resulted in an overall decrease in the energetic status of cells (Figure 1.5B). We also
depleted cells of RARRES1 and CCP2 simultaneously and found that RARRES1 depletion could
50 not rescue the effects of CCP2 knockdown in cells (Figure 2.11C) indicating that the effect of
RARRES1 on mitochondrial respiration is a consequence of its regulation of CCP2, as it was in
cell survival (Figure 2.11F).
We next confirmed that the effects of CCP2/RARRES1 on mitochondrial metabolism required
VDAC. We focused on VDAC1 since it is the most studied isoform and is the dominant isoform
in our cell models. The increase in OCR that occurs when RARRES1 is depleted was completely
reversed when both RARRES1 and VDAC1 were simultaneously depleted in either MCF 10A
cells or PWR-1E cells (Figures 2.11E and 2.12). Thus, the effects of RARRES1 and CCP2 on mitochondrial respiration, are due to regulation of VDAC1 activity.
51 Figure 2.11. RARRES1/CCP2-Mediated Changes in Mitochondrial Respiration Require VDAC1. (A) Oxygen consumption rate (OCR) of RARRES1 depleted MCF10A and PWR-E cells profiled in the extracellular flux assay. The relative quantification of the area below the curves corresponding to the stages labeled as BR (Basal Rate), UR (Uncoupling Respiration Rate), and MR (Spare Respiratory Capacity) are shown as histograms. (B) Extracellular acidification rate (ECAR) during the uncoupling response compares the glycolytic capacity of RARRES1-depleted cells relative to control MCF10A cells (scrambled). (C) OCR in PWR-1E and MCF10A cells following transient CCP2 depletion. (D) OCR activity after simultaneous transient knockdown of CCP2 and RARRES1 in MCF 10A Cells (E) RARRES1 depleted MCF 10A cells were also simultaneously transfected with VDAC1 siRNA OCR was measured in all experimental groups. 52 Figure 2.12. VDAC1 and its Effects on RARRES1 Depletion in PWR-1E Cells and KD Efficiency of VDAC1 and RARRES1. A. PWR-1E cells were transfected with scrambled siRNA, RARRES1 siRNA or VDAC1 and RARRES1 siRNAs. The OCR after FCCP injection and oligomycin injection were quantified. B. VDAC1 and RARRES1 KD efficiency. qPCR was run to quantify KD efficiency.
RARRES1 and CCP2 Reciprocally Regulate Glycolytic Demand
It is well established that hexokinase II can bind to VDAC1 and alter coupling between mitochondrial respiration and glycolysis (88). The binding of HKII to VDAC1 facilitates ATP accessibility to HKII and helps the enzyme catalyze the first reaction in glycolysis (146). In
RARRES1 depleted PWR-1E and MCF 10A cells a significant increase in glycolytic capacity or compensatory glycolysis accompanied the ECAR shift that occurred after inhibition of mitochondrial respiration. Here we wanted to test if the effects of RARRES1 on glycolytic
53 capacity, is a result of its modulation of tubulin blockade of VDAC1. We thus assessed whether
the target of RARRES1, CCP2, can regulate coupling between glycolysis and oxidative
phosphorylation. We used an extracellular Flux Assay to monitor the glycolytic shift that occurs
following inhibition of mitochondrial respiration, a phenomenon called glycolytic capacity or
compensatory glycolysis. Uncoupling respiration with oligomycin increased the ECAR indicating
an improved glycolytic capacity in RARRES1 knockdown cells. The opposite is seen in CCP2
depleted cells (Figure 2.13A). To confirm that the ECAR shift seen in RARRES1 depleted cells
is mediated through VDAC1, we transiently depleted VDAC1 in RARRES1 depleted cells and
measured the ECAR shift after oligomycin treatment. The double knockdown reversed the effects
of RARRES1 depletion. Thus the glycolytic coupling seen in RARRES1 depleted cells is mediated
through modulation of VDAC1 activity (Figure 2.13B). To ensure the shift in ECAR is due to
glycolysis, we examined glycolytic activity of CCP2 depleted cells using the glycolytic rate assay
as well as the glycolysis stress test. We found that CCP2 depletion caused a significant decrease
in glycolytic capacity or compensatory glycolysis (Figure 2.13C).
To further examine a role for RARRES1 and CCP2 in glycolysis, we performed knockdown
studies in zebrafish embryos and monitored levels of enzymes important in glycolysis. Knockdown
of RARRES1 increased levels of enolase-1 (eno1), phosphoglycerate mutase 1A (pgam1),
phosphoglycerate kinase 1 (pgk1), and aldolase c (aldoc) (Figures 2.13D & 2.13E). In contrast, knockdown of CCP2 decreased levels of these enzymes. These reciprocal changes in the levels of enzymes important in metabolism indicate a function for RARRES1 in central carbon metabolism and points to a role for the RARRES1/CCP2 axis in the regulation of coupling between oxidative phosphorylation and glycolysis.
54 Figure 2.13. RARRES1 Depleted Cells are Less Dependent on Glycolysis. (A) RARRES1 and CCP2 depletion alters ECAR response to inhibition of mitochondrial respiration by oligomycin. (B) ECAR response to inhibition of mitochondrial respiration in VDAC1 and RARRES1 double knockdown MCF 10A cells was assessed. (C)Treatment with glucose, oligomycin and 2-deoxy-d- glucose determined the glycolytic usage, capacity in CCP2 depleted PWR-1E cells.PWR-1E cells transfected with scrambled siRNA or CCP2 siRNA were also subjected to glucose prior to the assay and a combination of antimycin A and rotenone and 2-deoxy-d-glucose were injected. The compensatory glycolysis rate was assessed. (D) Enzymes related to glycolysis were assessed in RARRES1 MO and CCP2 MO zebrafish. The transcript level of these genes were quantified using qPCR (E). Enzymes, byproducts and pathways related to glycolysis that increased (highlighted in red) or decreased (highlighted in blue) following RARRES1 depletion.
55 RARRES1 and CCP2 Depletion Regulate Energy Homeostasis
VDAC1 regulates ATP/ADP release from the mitochondria and its inhibition by itraconazole and knockout studies have shown that VDAC1 regulation of ATP/ADP transport regulates ATP availability in the cytoplasm and subsequently dictates AMP kinase activity. AMP kinase is a central regulator of metabolic reprogramming and is sensitive to nutrient availability and the
AMP/ATP ratio (147). Starvation or increased AMP results in phosphorylation and activation of
AMP kinase by liver kinase B1 (LKB1) (147). Since RARRES1 and CCP2 govern mitochondrial energetics through VDAC1, we examined whether they can affect ATP availability and AMP kinase activity. Levels of activated AMPK increase significantly upon exogenous expression of
RARRES1 or depletion of CCP2 and decreased upon RARRES1 depletion (Figure 2.14A). The
RARRES1 inducer retinoic acid also activated AMP kinase in an RARRES1-dependent manner in both HFK and PWR-1E cells. (Figure 2.14B).
AMP kinase activation regulates anabolic and catabolic homeostasis, in which glucose and lipid metabolism and cell growth are affected (147). AMP kinase directly phosphorylates a number of enzymes that mediate metabolism and growth, such as acetyl-coa carboxylase 1 and 2 and 6- phosphofructo-2-kinase/fructose-2,6-bisphosphatase 2/3. AMP kinase activation rewires metabolic pathways to promote energy production by inducing catabolic pathways, such as glycolysis and fatty acid oxidation, while inhibiting anabolic pathways such as fatty acid and protein synthesis (147). In this study we examined whether other pathways regulated by AMP kinase are also altered in RARRES1 depleted cells. Since RARRES1 decreases AMP kinase activity, we hypothesized that RARRES1 would reprogram cells to a more anabolic state. We validated and measured significantly altered metabolites in the RARRES1 depleted MCF 10A
56 cells. The metabolites were initially identified based on accurate mass and confirmed by
comparing the fragmentation pattern against standard compounds (Table 2.1). Many metabolites
participating in central carbon metabolism were altered in the RARRES1 knockdown. Pathway
enrichment analysis identified a number of metabolic pathways that were significantly changed
including; the citric acid cycle, glycolysis, fatty acid, glutamate and glutathione metabolism
(Figure 2.14 & Table 2.1). RARRES1 knockdown increased oxidized glutathione suggesting a cellular response to high oxidative stress induced by RARRES1 knockdown. We showed previously that basal glycolysis is decreased in RARRES1 depleted cells although their glycolytic reserve increased and that RARRES1 depletion redirected glucose metabolism for lipid synthesis.
In our non-targeted LC-MS analysis of stable RARRES1 knockdown, we also observed a 7-fold increase in UDP-N-acetylglucosamine suggesting that part of the glycolytic reserve is being used for the production of UDP-N-acetylglucosamine, the end product of the hexosamine pathway
(Figure 2.14A). The hexosamine pathway is a minor branch of glycolysis where fructose-6-
phosphate is converted to glucosamine-6-phosphate, catalyzed by fructose-6-
phosphate amidotransferase (GFAT). We also observed a 16-foldchange in UDP-glucose, a
precursor of glycogen (Table 2.1). Glycogenesis is another branch of glycolysis in which glucose
is converted to glucose-6-phosphate and subsequently catabolized to glycogen. The large (>50
fold) increase of glutathione along with UDP-glucose and UDP-galactose are also an indication of
an increase in anabolic pathways such as glycogenesis and the pentose phosphate pathway (Table
1.1). Taken together these data support the MMP, energy flux and AMP kinase activity assays,
and indicate that RARRES1 knockdown increased mitochondrial respiration and ATP production
to drive essential anabolic reactions by attenuating AMP kinase activity (Figure 2.14D).
57 Figure 2.14. RARRES1 Regulates Energy Metabolism. (A) Immunoblot for total and phosphorylated AMPK in MCF10A, MDA-MB-231 or in which RARRES1 or CCP2 was exogenously expressed or depleted. (B) Immunoblot for total and phosphorylated AMPK HFK and PWR-1E cells with or without RA after RARRES1 depletion. (C) Enriched pathways in RARRES1 stable knockdown MCF10A cells. The dashed line represents p=0.05. (D) Schematic of the biological effects of RARRES1 and CCP2.
58 Table 2.1. Metabolites Significantly Altered in RARRES1 Stable Knockdown MCF 10A Cells Relative to Empty Vector
The table lists features that were significantly altered in the KD cells via multivariate data analysis. These were putatively identified by accurate mass based search using Madison Metabolomics Consortium Database (MMCD) and Human Metabolome Database (HMDB). M/z denotes – mass in Daltons/charge; FC - Fold change; KD/C – ( knock down/Control). P-values were calculated by student t-tests. The formula and exact mass of the metabolites were also included.
Inhibition of VDAC1 Activity Re-Sensitizes RARRES1-Depleted Cells to Apoptosis
Our data shows that RARRES1 and CCP2 reciprocally regulate metabolism through their modulation of VDAC1 activity. However these data do not address whether their effects on
59 VDAC1 activity also mediate their effects on cell survival. Therefore to test if the effect of
RARRES1 on cell survival involves VDAC1, we assessed the “resistant” phenotype of RARRES1-
depleted cells when VDAC1 is also depleted or inhibited. VDAC1 depletion increased apoptosis
(Figure 2.15A), and apoptosis resistant RARRES1-depleted cells were re-sensitized to apoptosis
when VDAC1 was also depleted. We also treated cells with retinoic acid and as expected,
RARRES1 depletion decreased the rate of RA-induced apoptosis but when VDAC1 was depleted in these cells, the cells were resensitized to apoptosis (Figure 2.15B). This clearly indicates that
the RARRES1 effect on apoptosis is mediated through VDAC1. We next used itraconazole, an
anti-fungal drug that binds to and inhibits VDAC1 in mammalian cells (145). We first assessed the effects of itraconazole on mitochondrial respiration and glycolytic coupling. Itraconazole had the same effects on mitochondrial respiration as CCP2 siRNA (Figures 2.15C, 2.11B, 2.13 and
2.14A) and reversed the effects of RARRES1 depletion on apoptosis. As a repurposed non-toxic
drug it is feasible that Itraconazole could be used as treatment for cancers characterized by
RARRES1 depletion or other alterations in mitochondrial metabolism.
60 Figure 2.15. VDAC1 Depletion or Inhibition Reverses the Apoptotic Phenotype Seen in RARRES1-Depleted Cells. A. Flow cytometry was conducted. Cells were stained with Annexin V and propidium iodine (PI). VDAC1 depletion induced apoptosis and RARRES1 depleted cells are sensitive to VDAC1 depletion. B. MCF 10A cells were treated with retinoic acid or vehicle and subjected to annexin V and PI staining and subsequent flow cytometry analysis. The Retinoic Acid treated cells were normalized to vehicle. RARRES1-depleted cells were resistant to apoptosis whereas RARRES1-depleted cells with VDAC1 depletion, were sensitive to retinoic acid treatment. (Figure 2.16).
61 A
B
Figure 2.16. FACS Analysis of Annexin Stained Cells. (A) MCF10A controls, RARRES1 transient knockdown, VDAC1 siRNA and RARRES1 siRNA in combination with VDAC1 siRNA were treated with 10^-5 M retinoic acid or vehicle. Samples were stained with fluorescein-labeled Annexin V and propidium iodide (Sigma) and analyzed by flow cytometry to measure apoptosis. (B) Scrambled siRNA and RARRES1 siRNA were treated with itraconazole. Cells were stained with fluorescein-labeled Annexin V and propidium iodide (Sigma) and analyzed by flow cytometry to measure apoptosis.
Discussion
Although no molecular mechanism has been characterized for the pleiotropic effects of RARRES1
on cells and cancers of multiple origins, we recently demonstrated that RARRES1 modulates
expression of major metabolic regulators, mTOR and SIRT1 (32). We also showed previously that
RARRES1 interacts with CCP2, a tubulin C-terminal deglutamylase now known to be responsible
62 for the production of Δ 2-tubulin (45, 71). However, it was not clear how this activity was linked
to the anti-cancer or metabolic effects of RARRES1. Post-translational modifications of the acidic
C-terminal tail of tubulin have long been linked with modulation of microtubule function (130).
These modifications included glutamylation, tyrosination and glycylation and are often associated
with microtubule activities in cilia, spindles and other stable microtubule arrays associated with
mitosis. Although anti-microtubule drugs are effective anti-cancer agents there is growing
evidence that their primary therapeutic action does not involve direct effects on mitosis (148).
Consistent with this view RARRES1 and CCP2 manipulation does not change cell division
(Figure 2.5A). Biophysical studies have also demonstrated that tubulin-like C-terminal peptides can interact with the mitochondrial voltage-dependent anion channel (VDAC) to influence mitochondrial membrane potential (89). Here we propose a novel model for the action of a tumor suppressor and metabolic regulator gene. Our results show that tumor suppressive and metabolic effects (of the carboxypeptidase inhibitor RARRES1 in this case) can be mediated by inhibition of
CCP-catalyzed tubulin C-terminal deglutamylation which in turn directly regulates VDAC1
activity, an in turn regulates mitochondrial membrane potential (MMP), reprograms intermediary
metabolism and influences cell survival (Figure 2.14D). Our data also indicate that this
mechanism may be particularly important in the action of RA as well as the survival and drug
resistance of cancer stem-like populations (118).
The simplest molecular explanation consistent with our results is that high CCP2 activity removes
glutamic acid residues from tubulin and its side chains to prevent it from attenuating VDAC
activity. RA through RARRES1 blocks CCP2 to inhibit deglutamylation, inhibit VDAC, reduce
MMP and activate AMPK (Figure 2.14D). Reduction in RARRES1 increases MMP and oxidative
63 phosphorylation, which in turn alters ADP/ATP flux, and inactivates AMPK with concomitant
changes in multiple metabolic and transcriptional pathways to result in a more energetic and
anabolic phenotype. VDAC1 depletion reversed the apoptotic and metabolic phenotype of
RARRES1 depleted cells. Consistent with this, the RARRES1 gene is heavily methylated in
multiple cancers. Both tubulin and VDAC1 are highly abundant proteins although their exact
stoichiometry is not known. However it is clear that most mitochondrial vdacs normally exist in
the closed state, consistent with the likely high concentrations of tubulin available to interact with
them (74). Consequently, small, localized changes in VDAC permeability following CCP2-
mediated deglutamylation may have significant effects on anion transport and MMP.
We recognize that our data is not consistent with the Warburg model in which cancer cells are
hypothesized to reduce their dependence on oxidative phosphorylation and increase aerobic
glycolysis (149). However the metabolic and phenotypic changes induced by RARRES1 depletion
are characteristic of proliferative stem cells and likely activated cancer initiating cells (150). The
changes we observed are also consistent with the metabolic reprogramming that occurs following
anchorage independent growth and the switch from quiescent to proliferating mammary cells (21,
151) . The elevated OCR we observed following RARRES1 depletion might be expected to increase the production of ROS and our LC-MS studies demonstrate an increase in protective ROS
scavengers in RARRES1 depleted cells. Increased ROS following increased mitochondrial
respiration also inhibits glyceraldehyde-3-phosphate dehydrogenase (GAPDH) to activate
glycogenesis and hexosamine pathways (152, 153). Both pathways are increased by RARRES1
depletion, suggesting that increased oxidative phosphorylation decreased AMP kinase activity,
64 which in turn decreased anaerobic glycolysis and shifted glucose metabolism toward biosynthetic
pathways.
RARRES1/CCP2 also regulates coupling between glucose and mitochondrial metabolism.
Consistent with a role for RARRES1/CCP2 in regulating VDAC function, VDAC1 is also an
important regulator of the coupling between glycolytic and mitochondrial metabolism. Hexokinase
2 (HK2) localizes to the VDAC/adenine nucleotide translocase channel where it stimulates the
rapid phosphorylation of glucose using ATP that is translocated directly from the mitochondria
(154). HK2 also enhances ADP recycling between the cytoplasm and the mitochondria potentially
explaining how RARRES1 loss increased mitochondrial respiration as well as glycolytic capacity
while CCP2 had the opposite effect (146). VDAC1 depletion in RARRES1 depleted cells was able
to reverse the increase in glycolytic capacity. Taken together, these data support the concept that
increased VDAC activity as a result of RARRES1 depletion enhances the coupling of HK2 and
ATP synthase, while CCP2 depletion displaces HK2 from VDAC1 by mediating tubulin dependent blockage of the channel.
Since polyglutamylated tubulin is highly enriched in stable microtubule arrays such as cilia the
reason for the high abundance of enzymes that deglutamylate tubulin (the ccps) in ciliated cells
and in the ciliary proteome is not altogether clear (140, 155). Ciliary axonemes are also rich in
mitochondria, which provide the large amounts of ATP required for ciilary beating. Our data raise
the possibility that local CCP2-mediated formation of Δ2-tubulin may act to open VDAC and
promote local increases in oxidative phosphorylation to power ciliary beating, without
compromising the metabolism of the rest of the cell. In addition, other non-tubulin substrates for
65 ccps have been identified and it is possible that RARRES1 would also inhibit CCP activity toward
them (122).
Our data is consistent with growing evidence that stem cells and cancer initiating cells exhibit
quite different metabolic characteristics than their differentiated/tumor counterparts (118). Loss of
RARRES1 in multiple tumor types, its role in stem cell differentiation and in cell survival indicates
that the RARRES1/CCP2 axis could be a candidate for therapeutic intervention. In the many
cancers in which the RARRES1 gene is methylated, the unencumbered enzymatic activity of CCP2 could be inhibited by, new or repurposed drugs (156). Understanding how RARRES1 expression
is regulated could lead to interventions that increase its levels. Alternatively, one could imagine
that agents that mimic the effects of RARRES1 (i.e. Inhibit mitochondrial respiration), would
reverse the effects of RARRES1 loss (157). In this study, we were able to successfully reverse the
resistant phenotype of RARRES1-depleted cells by treating cells with itraconazole, an anti-fungal
drug that inhibits the activity of VDAC in mammalian cells. Finally, in an age of deep sequencing
and personalized medicine our data support the notion that treatments that target fundamental cell
functions such as, metabolism, immune rejection, DNA replication and mitosis, have an important
role in the treatment of drug resistant cancers characterized by multiple mutations or translocations
and/or a proliferative stem-like phenotype.
66 Chapter 2- RARRES1, A Novel Regulator of Fatty Acid Metabolism
Introduction
Retinoic acid receptor responder element 1 (RARRES1) or tazarotene induced gene 1 (TIG1) was
initially identified as a novel retinoic acid receptor regulated gene in the skin(36). RARRES1 is
also regulated by vitamin D (45, 94, 115). Roy et al. In prostate cancer cells that RARRES1 is able to induce autophagy, decrease mTOR and increase SIRT1, two important regulators of energy homeostasis (32). RARRES1 also induces autophagy in cervical cancer cells(97). The ability of
RARRES1 to regulate the expression of metabolic master regulators and induce autophagy
suggests that it might function to reprogram metabolism in cells.
RARRES1 is differentially expressed in metabolic diseases and is associated with biological
hallmarks that require metabolic reprogramming. For example, RARRES1 is among the most up-
regulated genes in subcutaneous fat from obese human subjects on a diet-induced weight loss and
among the most downregulated genes during weight maintenance (35). In adipocyte differentiation,
in which metabolic reprogramming is crucial, RARRES1 is increased during dedifferentiation and
decreased during differentiation (95, 96, 158). RARRES1 is also differentially expressed in mouse
models of hepatic steatosis and cholestatic liver disease(152, 159–163).
Although RARRES1 is among the most commonly methylated genes in multiple cancers, it is
actually increased in basal-like hormone receptor negative breast cancer and in liver cirrhosis, a risk
factor for hepatocellular cancer(31, 33, 100–103, 106, 164). Metabolic reprogramming is now considered a hallmark of cancer etiology(11). For decades, aerobic glycolysis (Warburg metabolism) was considered to be the most important energetic pathway that cancer cells use to
67 survive and proliferate. However, it is now clear that oxidative phosphorylation and fatty acid
metabolism play a major role in cancer progression and drug resistance(118, 165–167).
Here we demonstrate that RARRES1 reprograms metabolism by modulating a switch from aerobic
glycolysis to glucose dependent de novo lipogenesis (DNL). RARRES1 modulated DNL in turn, enabled normal breast and prostate epithelial cells, to improve fatty acid oxidation (FAO) activity during starvation by utilizing stored endogenous fatty acids. We also show that RARRES1 expression correlates with that of fatty acid metabolism genes in breast, colorectal and prostate cancers. These findings open up a new avenue for metabolic reprogramming and identify RARRES1 as a potential target for cancers and other diseases with impaired fatty acid metabolism.
Methods
Cell-culture
HEK293T was cultured in DMEM (Gibco-Invitrogen, Grand Island, NY) supplemented with 5% fetal bovine serum. The cells were maintained in a humidified modular incubator at 37° C and
5% CO2. MCF 10A cells were maintained in DMEM/F12 (50:50 mix) supplemented with 5%
horse serum, 10 mm HEPES, 10 ug/ml insulin, 20 ng/ml epidermal growth factor, 0.5 ug/ml
hydrocortisone, 100 ng/ml cholera toxin. Immortalized human prostate epithelial PWR-1E cells
(a gift from Dr. S.C. Chauhan, University of South Dakota) were maintained according to
ATCC’s recommendation.
Antibodies and Reagents
Primary antibodies targeting the following antigens were used: rabbit anti-human RARRES1
(TIG1) (Sigma-Aldrich, St. Louis, MO), GAPDH (Fitzgerald Industries International, Acton, MA)
68 and alpha-tubulin (Sigma-Aldrich, St. Louis, MO). C75 compound was obtained from Cayman
Chemicals. Oleic acid was purchased from Cayman Chemicals.
Cell extracts. Cells were washed with PBS and lysed in RIPA lysis buffer (1% sodium
deoxycholate, 0.1% SDS, 1% Triton X-100, 10mm Tris-HCL ph8, 150 mm nacl) for 15 minutes
at 4°C. Protein concentration was determined using a protein microplate assay (Biorad
Laboratories).
RNA Extraction and Real-time Quantitative PCR Analysis
Total RNA was isolated from cells by using Trizol reagent (Invitrogen) according to the
instructions of the manufacturer. RNA was isolated using the rneasy kit (Qiagen) according to the
instructions of the manufacturer. Reverse transcription was done using Invitrogen Reverse
Transcription Kit. The cdna samples were used to quantify RARRES1 expression using the fast
real-time PCR kit (Thermo Fisher) with the appropriate taqman probes (Thermo Fisher: RARRES1
Hs00894859_m1 and 18s Hs03003631_g1) in a stepone (Applied Biosystems) compact
qPCR machine. Transcript levels were normalized to the level of 18S.
Plasmid and siRNA Constructs
RARRES1 were directionally cloned into the bglii and hindiii sites in the peyfp-N1 vector
(Clontech). Full-length RARRES1 (+1 - +897) were cloned into the pglue vector as codon
optimized versions by Genscript (Piscataway, NJ). RARRES1 (cat# L-009925-00-0005 5 nmol)
and non-targeting control siRNAs (cat#D-001210-01-20) were from Dharmacon (Lafayette, CO).
Plasmid DNA was introduced by lipofectamine 3000 and siRNA constructs were transfected using
RNAi Max. Constructs for stable depletion of RARRES1 were obtained from the RNAi
Consortium (168) via SIGMA-Aldrich (NM_002888). For the RARRES1 gene, five pre-made
constructs were obtained and individually tested to identify those able to achieve efficient
69 knockdown at the protein level. Negative control constructs in the same vector system (vector
alone, plko.1 puro) were obtained from SIGMA-Aldrich (NM_003177). Next, MCF10A cells were
infected with shRNA lentiviruses. To do this, the cells were plated at sub confluent densities. The next day, the cells were infected with a cocktail of 100 ul virus-containing medium, 1 ml regular
medium, and 8 g/ml Polybrene. The medium was changed 1-day post-infection, and selection medium was added 2 days post-infection (5 g/ml puromycin for MCF10A). After 3 days of puromycin selection, the mock-infected cells had all died. Stably infected pooled clones were studied.
Oleic Acid Treatment
Oleic acid treatment was used as recommended by Cayman Chemicals; MCF 10A and HEK 293
T cells were treated at a 1:5000 dilution for 4 hours or overnight respectively.
Western Blot Analysis
To confirm RARRES1-YFP expression or RARRES1 knockdown in cells, samples were subjected
to western blot analysis (Figure 3.1). 50 g of cell lysate was separated on 4-12% SDS/PAGE,
transferred to nitrocellulose membrane (Amersham Pharmacia Biotech), blocked with 5% milk-
PBS, and incubated in primary antibody in 5% milk-PBS overnight at 4°C. Blots were washed 3
times for 5 minutes each in PBS-0.5% tween, followed by incubation in HRP-conjugated
secondary antibody (KPL, Gaithersberg, MD) for 1 hour at room temperature on an orbital. Blots
were washed 2 times in PBS-0.5% tween, followed by 1 wash in PBS. Detection of
immunoreactive bands was carried out using chemiluminescence with ECL Western Blotting
Detection Reagents (Amersham, Piscataway, NJ).
70 S8 Figure
A B
MCF 10A
RARRES1 Scramble siRNA
RARRES1 33 kDa
36 kDa GAPDH
C
1.5 D MCF 10A
CT) * Δ (
Δ 1.0
0.5 Fold Change
0.0
Scramble
RARRES1 siRNA
PWR-1E
E MCF 10A Cells
Figure 3.1. RARRES1-YFP and RARRES1 siRNAS-12 Transfection Efficiency. (A) RARRES1- YFP (expected band ~ 60 kda) expression was validated in HEK 293 T cells. Tubulin or GAPDH was used as a loading control. The image was cropped, and lanes were juxtaposed; black line is drawn to describe the boundary. RARRES1-YFP overexpression was also confirmed in MCF 10A cells. (B) Western blot was done to confirm the transient RARRES1 knockdown efficiency in MCF 10A cells. (C) qPCR was run to validate RARRES1 knockdown in PWR-1E cells. Transcript levels were normalized to the level of 18S. In our previous study4, we validated through western blot that PWR-1E cells is endogenously expressed in PWR-1E cells. (D) Stable RARRES1 knockdown in MCF 10A cells were validated through RT-PCR. Beta-actin was used as a loading control. (E) Stable RARRES1 knockdown was also confirmed through western blot. Alpha-tubulin was used as a loading control.
Nile Red Lipophilic Staining and Oil Red O Staining
Following RARRES1 or scrambled siRNA transfection, MCF 10A cells were fixed with 10 %
formaldehyde (Protocols) for 15 minutes at room temperature and washed 3 times with 1X PBS.
To stain with Nile Red (Sigma, St. Louis, MO, USA), 1 L of a 1 mg/ml stock solution was added
71 to 10 ml of 150 mm NaCl in PBS to make a Nile Red solution. The Nile Red solution was added
to the cells and incubated for 10 minutes in the dark. Cells were counterstained with DAPI(169).
Cells were imaged with the Keyence BZ-X microscope (Keyence Corporation, Osaka, Japan).
MCF 10A cells transfected with siRNA were grown for 48 hours. HEK 293 T cells and MCF 10A
cells transfected with YFP or RARRES1-YFP plasmids were grown for 24 hours then treated with
oleic acid for 4 hours or overnight before fixation. Cells were fixed in 10% formalin (Electron
Microscopy Science, Hatfield, PA), for 10 min. At room temperature. Cells were stained for
neutral lipids using the oil red O staining kit (American Master Tech Scientific, Lodi, CA). Briefly,
cells were washed four times with ddh20 (EMD Chemicals Inc., San Diego, CA), incubated for
10 min RT with oil red O, and counterstained with DAPI (Thermo-Scientific). Cells were imaged with the Keyence BZ-X microscope. For both staining methods, oleic acid treatment was used as a positive control. Intensity of lipid staining was quantified through imagej62.
Metabolic Analysis Using Extracellular Flux Assays
Bioenergetics profile of transient RARRES1 knockdown in MCF10A and PWR-1E cells and
stable RARRES1 knockdown MCF10A cells were measured using the xfe96 Extracellular Flux
Analyzer. Oxygen consumption rates and response to mitochondrial stress factors were analyzed by using the XF Cell Mito Stress Kit. 48 hours prior to analysis, MCF10A and PWR-1E cells were transfected with RARRES1 siRNA or scramble siRNA in duplicates. The night before the assay, the cells were seeded at an optimized cell density (MCF10A cells: 10,000 and PWR-1E cells:
20,000) in the 96-well xfe plate and incubated at 37°C with 5% CO2 overnight. The day of analysis, the cells were incubated at 37°C in a CO2-free atmosphere incubator for 1 hour. Basal oxygen
consumption rate and extracellular acidification rate were measured. Subsequently, oxygen
72 consumption rate (OCR) and extracellular acidification rate (ECAR) responses were observed after separate injections of oligomycin (1 M), carbonyl cyanide-p- trifluoromethoxyphenylhydrazone FCCP (0.5 M for MCF10A cells or 0.25 M for PWR-1E cells), and a combination of rotenone and antimycin A (0.5 M) were respectively prompted in the assay. For each injection, there was a total of 3 cycles, each one lasted 3 minutes and a measurement was taken at the beginning of each cycle. Glycolysis was also analyzed, using the
XF Glycolysis Stress Test kit, in the xfe96 Extracellular Flux Analyzer. Cells were transiently knocked down with RARRES1 siRNA or scramble negative control siRNA and plated at the same optimal density as the XF Cell Mito Stress Test protocol. ECAR was measured at baseline and after adding glucose (10 mm), oligomycin (1 M) and 2-deoxy-d-glucose (100 mm) respectively.
Fatty Acid Oxidation measurements were done as recommended by Agilent Seahorse. Cells in nutrient-rich media, were treated with BPTES (a glutaminolysis inhibitor), UK-5099 (a glycolysis inhibitor) and etomoxir (a fatty acid oxidation inhibitor). The cells were assessed for their dependency on each pathway and their plasticity when each pathway is inhibited. To measure fatty acid oxidation in a starved state, cells were starved with minimal substrate DMEM for 24 hours.
The minimal substrate media included 1% serum, 1 mm glutamine, 0.5 mm carnitine, and 0.5 mm of glucose. Insulin, EGF and hydrocortisone were excluded from the media. The day of the assay, starved cells were washed and incubated with 1X KHB (111 mm nacl, 4.7 mm kcl, 1.25 mm cacl2
, 2 mm mgso4, 1.2 mm nah2po4 and supplemented with 2.5 mm glucose, 0.5 mm carnitine, and 5 mm) in a non-CO2 37°C incubator. 15 minutes prior to the assay, 40 M etomoxir was added to the cells to measure endogenous fatty acid uptake for FAO. To measure exogenous fatty acid uptake for FAO, Palmitate-BSA and BSA (vehicle) were added to cells right before treating with etomoxir and adding the cells to the Flux Analyzer. All experiments were analyzed in duplicates
73 for at least two independent experiments for each cell line and the results were normalized to their
protein concentration or cell number.
ATP Detection
Cells were transfected with RARRES1 siRNA or scrambled siRNA by RNAi Max in a 6-well plate. These cells were left to grow for 24 hours and subsequently replated in a 96-well plate. The cells were then treated with 150 M of C75 for 2, 3 and 5 hours. The cells were then harvested, and ATP was quantified by luminescence according to the manufacturer’s instructions (Cayman
Chemicals, Ann Arbor, MI, USA). ATP standards were included (0.01 M to 10 M).
LC-MS and GC-MS Metabolite Extraction. Cell samples (greater than 107 confluency) were processed using 150 μl 100% water. Samples were plunged in to dry ice for 30 seconds and heat shocked in a water bath for 90 sec. At 37 ⁰C. The freeze thaw process was repeated twice and the samples sonicated for 30 sec. 5 μl of each sample was transferred into a 0.7 ml Eppendorf tube and kept at -20 ⁰C for Bradford protein analysis. The remainder of the sample was processed using 600
μl of methanol containing internal standard. The sample was vortexed and allowed to incubate on ice for 15 min. Chloroform was added to each sample (600 μl) and centrifuged (13,000 rpm) at
4⁰C for 20 min. The two phases were extracted and kept in different Eppendorf tubes. Chilled
Acetonitrile (600 μl) was added to each phase and allowed to incubate at -20 ⁰C to further precipitate cellular debris and proteins. Incubated samples were centrifuged (13,000 rpm) at 4⁰C for 20 min. Supernatants from the incubated samples were transferred to another Eppendorf tube, dried under vacuum, and stored at -80°C until analysis. For analysis, the dried samples were reconstituted in 100 μl of 50% Methanol in water, and both phases combined in a Mass Spec
Sample tube for LCMS or GCMS analysis.
74 Ultra-Performance Liquid Chromatography-Time of Flight Mass Spectrometry (UPLC TOF MS)
mm based metabolomic analyses. Each sample (2 L) was injected onto a reverse-phase 50 × 2.1
BEH 1.7 m C18 column using an Acquity UPLC system (Waters Corporation, USA). The gradient mobile phase comprised of: solvent A- 100% water + 0.1% formic acid; solvent B- 100% acetonitrile + 0.1% formic acid; solvent D- 90% isopropanol and 10% acetonitrile + 0.1% formic acid (all containing 0.1% Formic Acid). Each sample was resolved for 13 min at a flow rate of 0.4 ml/min. The gradient consisted of 95% A and 5% B for .50 minutes, then a ramp of curve 6 to 2%
A and 98% B from 0.5 min. To 8.0 min., then a ramp of curve 6 to 2% B and 98% D to 9.0 min., a hold of 2% B and 98% D up to 10.5 min., then a ramp of curve 6 to 50% A and 50% B to 11.5 min., then a ramp of curve 6 to 95% A and 5% B to 12.5 min., and a hold of 95% A and 5% B to
13.0 min. The column eluent was introduced directly into the mass spectrometer by electrospray.
Mass spectrometry was performed on a quadrupole-time-of-flight mass spectrometer operating in either negative or positive electrospray ionization mode. Positive mode has a capillary voltage of
3.0 kv, a sampling cone voltage of 30 V, and a source offset of 80 V. Negative mode has a capillary voltage of 2.75 kv, a sampling cone voltage of 20 V, and a source offset of 80 V. The de-solvation gas flow was 600 L/hr. And the temperature was set to 500⁰C. The cone gas flow was 25 L/h, and the source temperature was 100⁰C. The data were acquired in the Sensitivity and MS Mode with a scan time of 0.1 seconds, and inter-scan delay at 0.08 seconds. Accurate mass was maintained by infusing Leucine Enkephalin (556.2771 m/z) in 50% aqueous acetonitrile (1.0 ng/ml) at a rate of 10 μl/min via the lock-spray interface every 10 seconds. Data were acquired in centroid mode from 50- 1200 m/z mass range for TOF-MS scanning. Pooled sample injections at the beginning and end of the run (one pool was created by mixing a set aliquot from all samples) were used as
75 quality controls (qcs) to assess inconsistencies that can be particularly evident in large batch
acquisitions in terms of retention time drifts and variation in ion intensity over time.
Data Pre-Processing and Metabolite Identification and Validation
Centroided and integrated UPLC-TOFMS data were pre-processed using the XCMS software
normalized to the ion intensity of respective internal standards as well as to the total protein
concentration. The data were log transformed and multivariate data analyses were performed to
delineate significantly altered metabolites in the two groups. These metabolites were putatively
identified via accurate mass-based search using the Madison Metabolomics Consortium Database
(MMCD), simlipid (Premier Biosoft), the Human Metabolome Database (HMDB) and LIPID
MAPS(170–172). Selected metabolites in transient RARRES1 KD and scrambled siRNA cells
were validated using tandem mass spectrometry. The daughter and parent ions for the metabolites
were matched with the MS/MS spectra available in HMDB, SimLpid and LIPID MAPS. MSE
results were also subjected to additional lipid validations through SimLipid software V6.01
(Premier Biosoft, Palo Alto, CA, USA). This validation method has been reported by others(173–
176). It is important to note that our non-targeted LC-MS method and validation approach did not
allow for the identification of the specific lipid that corresponds to each ion, but we were able to
identify the lipid classes. Citrate was confirmed by comparing the retention time under the same chromatographic conditions and by matching the fragmentation pattern of the parent ion from the biological sample to that of the standard metabolite using tandem mass spectrometry (UPLC-
TOFMS/MS) (Figures 3.2 and 3.3).
76 Supplementary Information
S1 Figure
A A B
C
24 Figure24 3.2. Non-targeted LC-MS Analysis of Transient RARRES1 KD in MCF 10A Cells. (A) Heat map of all metabolites significantly altered were depicted. (B) Important features in
transient RARRES124 KD MCF 10A cells were selected24 by volcano plot with fold change threshold 24 24 Log2 (FC) Log2 (x) 2 and t-tests threshold (y) 0.05. The red circles represent features above the threshold. Not the fold changes are log transformed. The further its position away from (0,0), the more significant Log2 (FC) Log2 Log2 (FC) Log2
-4 -2 0 the-4 feature -2 0 is.
Peaks(mz/rt) -4 -2 0 -4 -2 0 -4 -2 0 -4 -2 0
Peaks(mz/rt) Peaks(mz/rt)
S-1
77 A B
Peak True - sample "201706_Sara_Cell_14_9_1", peak 344, at 1062.8 s Oleic Acid 75 1000 55 117 500 199 264 339
50 100 150 200 250 300 350 400 450 500 550 600 Library Hit - similarity 906, "9-Octadecenoic acid, (E)-, TMS derivative"
73 1000 117 55 500 339 185 222 264
50 100 150 200 250 300 350 400 450 500 550 600
C D
3 0.5117
2
1 Palmitic Acid Palmitic (Fold Change)
0
Scramble RARRES1 KD
PWR-1E - RARRES1 siRNA Human Hepatocytes- RARRES1 siRNA Scrambled siRNA RARRES1 siRNA
MCF 10A
Figure 3.3. GC-MS Analysis. (A) Important features in transient RARRES1 KD MCF 10A cells Were selected by volcano plot with fold change threshold (x) 2 and t-tests threshold (y) 0.05. The red circles represent features above the threshold. Not the fold changes are log transformed. The further its position is away from (0,0), the more significant the feature is. (B) Oleic acid GC-MS fragmentation pattern peaks in transient RARRES1 KD and scramble MCF 10A cells were aligned against the fragmentation pattern peaks available in NIST database. (C) Detected palmitic acid was quantified and normalized against the peak intensity of the scramble control. (D) Stearic acid, myristic acid and cholesterol were detected in GC-MS and quantified in transient RARRES1 KD PWR-1E and primary human hepatocytes. The fold change is in terms of the peak intensity of the corresponding metabolites in the appropriate scrambled siRNA transfected control cells.
LC-Mass Spectrometry Statistical Analyses
Raw metabolomics data was analyzed using analysis functionalities of Metaboanalyst 4.0 (177).
Data were log transformed before performing t- statistics to identify significant metabolites.
Metabolites with an adjusted p-value of less than0.05 and a fold change either less than or equal
to 0.5 or greater than 1.5 were used to create volcano and Partial Least Squares –Discriminant
Analysis (PLS-DA) plots (figures1.1). GC-TOF Single Injections. Processed samples were transferred (100 L) to a GC vial with a salinized insert, dried in by speed-vac at room temp, caped
78 and kept at 20 degrees for derivatization. The dry residue was dissolved in 15 L of 20.0 mg/ml O-
Methoxyamine-Hydrochloride in pyridine. The dissolved sample was heated to 40 °C with
constant agitation for 30 min. N-Methyl-N-(trimethylsilyl) trifluoroacetamide: MSTFA/BSTFA
WITH 1% TMCS (50:50) was added (60 L) and the sample heated 40 °C with constant agitation
for 30 min. The sample was then allowed to cool to 20 °C and incubated for 4 hrs. Before injection
for GC-TOF analysis After TMS derivatization, A 5.0 L aliquot of the derivatized solution was
injected in (1:5) split mode into an Agilent 7890B GC system (Santa Clara, CA, USA) that was
coupled with a Pegasus HT TOF-MS (LECO Corporation, St. Joseph, MI, USA). Separation was
achieved on an Rtx-5 w/Integra-Guard capillary column (30 m x 0.25 mm ID, 0.25 m film
thickness; Restek Corporation, Bellefonte, PA, USA), with helium as the carrier gas at a constant
flow rate of 1.0 ml/min. The temperature of injection, transfer interface, and ion source was set to
150, 270, and 230 °C, respectively. The GC temperature programming was set to 0.2 min. Of
isothermal heating at 70 °C, followed by 10 °C/min oven temperature ramps to 270 °C, a 4.0 min.
Isothermal heating of 270 °C, 20 °C/min to 320 °C, and a 2.0 min. Isothermal heating of 320 °C.
Electron impact ionization (70 ev) at full scan mode (m/z 40–600) was used, with an acquisition rate of 30 spectra per second in the TOF/MS setting.
GC-MS Analysis and Metabolite Validation
Raw data files were pre-processed through Leco Statistical Compare software to generate an excel output. The Excel output data was normalized by internal standard (4- Nitrobenzoic acid). In
Metaboanalyst 4.0(177), the data was further normalized by log transformation and Pareto scaling and a Univariate Analysis was performed. For data analysis, a T-Test and fold change was performed to generate a volcano plot. The unique peaks and their respective retention indices were used to identify metabolites. Peak detection with background subtraction and subsequent matching
79 of the resulting mass spectra to the Agilent Fiehn GC/MS Metabolomics RTL Library using NIST
MS searches was done.
Bioinformatics Analysis
Oncominetm (Compendia Bioscience, Ann Arbor, MI, USA) was used to identify cancers where
RARRES1 gene expression was significantly upregulated or downregulated. These cancers
included breast, cervical, colorectal, esophageal, gastric, head and neck, kidney, liver, lung, lymphoma, pancreatic and prostate (Table I from oncominetm). Of the identified cancers, three
(breast, colorectal and prostate) were selected for further analysis given the number of datasets found to have significant RARRES1 gene expression changes as well as the literature known about
RARRES1 and its molecular biology within these cancers. Within the cancers selected, individual datasets were chosen given their large sample sizes and subtype analyses. Thus, the The Cancer
Genome Atlas (TCGA) Breast Cancer, TCGA Colorectal Cancer, and Grasso Prostate datasets were selected. The Grasso Prostate dataset was also uniquely selected for its metastatic vs. Primary analysis (178). For the TCGA Breast Cancer dataset, subtype analysis was stratified into ER+ vs.
ER- status; PR+ vs. PR- status and Triple Negative (PR-, HER2- & ER-) vs. Non-Triple
Negative(179). Metastatic vs. Primary site analysis was performed in the Grasso Prostate dataset.
All colorectal cancer types (rectal mucinous adenocarcinoma, colonadenocarcinoma, rectal adenocarcinoma, colon mucinous adenocarcinoma and cecum adenocarcinoma) identified in the
TCGA dataset were also analyzed in comparison to their respective normal samples (180).
Genes positively or negatively co-expressed with RARRES1 were further identified using oncominetm with respect to each dataset of tumor vs. Normal. The sets of co-expressed genes with
80 respect to each dataset were pipelined into DAVID to identify enriched pathways (adjusted
P<0.05) found in the Reactome and Kyoto Encyclopedia of Genes and Genomes (KEGG)
databases(181, 182). Significantly enriched pathways common across all datasets were then
identified. In addition, genes important in fatty acid metabolism pathway were further assessed for co-expression with RARRES1 with respect to cancer subtypes (e.g. ER+ vs ER-) via correlation
coefficients. Selected genes included the following: PPARG, PPARA, PPARGC1A, SREBF1,
FASN, SCD1 and CPT1A.
In Vitro Data Analysis
Results are shown as the mean ± SD. Statistical significance was calculated by using Student's
t-test and P < 0.05 was accepted as significant value (***, p < 0.001; **, p < 0.01; *, p < 0.05).
At least three biological replicates were done to confirm the results.
Results
RARRES1 regulates lipogenesis and lipid droplet accumulation
Since RARRES1 expression is associated with metabolism-associated diseases and its exogenous
expression regulates the expression of two important players in metabolic reprogramming (mTOR
and SIRT1)(32), we examined whether RARRES1 has any functional significance in metabolic
reprogramming. As RARRES1 is highly expressed in differentiated epithelial cells, we used
selected epithelial cells as our model. We subjected transient RARRES1-depleted mammary
epithelial cells to non-targeted LC-MS (Figures 3.4A & 3.5) to get an overview of the metabolic
changes that occur after decreasing the expression of the gene. We noticed a significant increase
in neutral lipids, triacylglycerols and the cholesterol derivative 24,25-epoxy-cholesterol.
Phospholipids, such as phosphatidylinositol (PI), phosphatidylethanolamine (PE),
81 phosphatidylserines and phosphatidylcholines and the eicosanoid substrate, eicosatrienoic acid
were also increased, as were three metabolites important in the synthesis of sphingolipids,
sphinganine, niacinamide and sphingosine (Figures 3.4A & 3.2). These data indicate a global
increase in lipid synthesis. We next used GC-MS to measure free monounsaturated and saturated
fatty acids, since they are a good indicator of an increase in fatty acid synthesis(183). We found
an increase in oleic acid, the principle product of SCD1 (SCD gene), but no change in palmitate
(Figures 3.4B, 3.3B & 3.3C)(184). However palmitate, a saturated fatty acid, is highly toxic to the cell and is usually converted to palmitoleate or oleic acid(185). We also examined lipid changes in RARRES1 in normal prostate epithelial cells and primary hepatocytes to ensure that this change in lipid content is epithelial specific and not cell line specific, by GC-MS. In both cell lines there was increase in cholesterol and saturated fatty acids (stearic acid and myristic acid)
(Figure 3.3D). Taken together these data demonstrate that manipulation of RARRES1 alters global fatty acid metabolism.
Since RARRES1 depletion increased several classes of lipids, including neutral lipids, we validated the findings by staining lipid droplets in cells. Consistent with the LC/GC-MS data, transient RARRES1 depletion in normal breast epithelial MCF 10A cells, lead to significant accumulation of lipid droplets, verified by Oil Red O and Nile Red staining (Figures 3.4C &
3.5). We then transiently overexpressed RARRES1-YFP in MCF 10A and HEK-293T cells. We treated the cells with oleic acid for 4 hours or overnight, respectively, to induce lipid droplets.
Cells that exogenously expressed RARRES1 had a striking decrease in lipid droplet accumulation compared to YFP expressing cells and the surrounding cells that did not take up the RARRES1-
YFP plasmid (Figures 3.1D & 3.5B). This indicates that higher levels of RARRES1 leads to the
82 degradation or oxidation of exogenous fatty acids rather than their storage in lipid droplets.
Conversely, depletion of RARRES1 leads to the accumulation of lipids. Consistent with this observation, RARRES1 expression in prostate and cervical cancer cells induces autophagy, a cellular response that sequesters and degrades lipid droplets during starvation, a mechanism called lipophagy(186,187).
83 A B
Sphingosine Sphinganine Eicostrienoic Acid 10 Niacinamide ***
24,25-epoxy-cholesterol 8
TG(62:6) TG(54:5) 6 TG(52:5) TG(60:3) 4
TG(58:5) Acid Oleic LysoPC(16:0) (Fold Change) 2 GPI(16:0) PC(O-18:1) 0 PE(20:2) PE(34:4) PE(P-16:0) PS(38:5) PS(44:5) Scrambled siRNA RARRES1 siRNA PI(36:4) PI(34:1)
C Scrambled siRNA-Oil Red O DAPI RARRES1 siRNA-Oil Red O DAPI 500 ***
400
300
200
Overlay Overlay 100 % Lipid Droplet Area % Lipid Droplet 0
Scrambled siRNA RARRES1 siRNA D
YFP Oil Red O DAPI RARRES1-YFP Oil Red O DAPI
250
200 Overlay Overlay
150
100 % of Total Cells % of Total
50
0
YFP Brightfield
RARRES1-YFP Lipid Droplet Formation
No Lipid Drople Formation
E TAGs
Sphingolipids LPA, PA Eiscosanoids
Monounsaturated FA Oleic Acid TCA
Citrate Acetyl-CoA
Cholesterol/Cholesterol Esters RARRES1 Depletion
Figure 3.4. RARRES1 Regulates Lipid Accumulation. A. Classes of metabolites measured by non-targeted LCMS analysis of transient RARRES1 knockdown MCF10A cells and scrambled control cells. B. GC-MS was run in transient RARRES1 knockdown MCF10A cells and monounsaturated fatty acids were analyzed. Oleic acid is represented in the bar graph. D. RARRES1-YFP was exogenously expressed and the transfected cells were treated with oleic acid and stained with Oil Red O to localize lipid droplets. The numbers of cells with YFP or RARRES1- YFP transfection displaying lipid droplets or no lipid droplets were counted in both HEK 293T cells and MCF 10A cells. E. A schematic representation of metabolites that were upregulated in RARRES1-depleted cells. C. RARRES1 was transiently knocked down in MCF10A cells and Oil Red O staining was used to stain for lipid accumulation. The bar graph represents the intensity of lipid droplet staining. Each point in the graph represents a biological replicate.
84 A
Scramble RARRES1 KD Oleate Treatment
B
C YFP-Overexpression RARRES1-YFP
Figure 3.5. RARRES1 Regulates Lipid Content. (A) Transient RARRES1 KD MCF 10A cells were stained with Nile Red. Oleic acid treatment was used as a positive control. (B) RARRES1- YFP was overexpressed or YFP (negative control) in oleic acid treated MCF 10A cells and droplets were stained with Oil Red O. (C) DAPI staining of RARRES1-YFP and YFP overexpression in HEK 293T cells (Figure 3.4B).
RARRES1 Depletion Regulates Fatty Acylcarnitine Content and Fatty Acid Oxidation (FAO)
We next assessed FAO activity in RARRES1-depleted epithelial cells. Acylcarnitines are involved in the first committed step of FAO. Long chain fatty acids are catabolized to acetyl- coenzyme A, a major fuel for mitochondrial respiration (especially during starvation), acyl-coas must then be converted to acylcarnitine to cross the outer mitochondrial membrane(188).
Disorders of fatty acid metabolism are typically associated with primary and secondary forms of
85 carnitine deficiency (189). Thus monitoring carnitine levels is an accurate way to assess transportation of long chain fatty acids into mitochondria FAO. Acylcarnitines were detected and identified in our non-targeted LC-MS analysis of MCF 10A cells in which RARRES1 was transiently depleted. Isobutyryl-L-carnitine (AC-4:0) and L-palmitoylcarnitine (AC-6:0) were increased in the transient knockdowns (Figure 3.6A).
The data above suggests that RARRES1 depletion increases the uptake of fatty acids into the mitochondria, either for beta oxidation or mitochondrial fatty acid elongation. To address this, we examined fatty acid oxidation activity. Oxygen consumption rate (OCR) of RARRES1-depleted cells was monitored in nutrient rich media. As carnitine palmitoyltransferase 1 (CPT1) is essential for the beta oxidation of long chain fatty acids, cells were treated with etomoxir, a CPT1 inhibitor, to measure FAO dependent OCR. There was no significant change in fatty acid oxidation dependency between control and RARRES1 depleted cells grown in nutrient rich media (Figure
3.7A)(190). In addition, there was no difference between the control and RARRES1 depleted cells in ATP production until we triggered FAO by treating cells with C75, an irreversible inhibitor of fatty acid synthase (FAS), an enzyme important in de novo fatty acid synthesis. Other studies have established that treatment with C75, at a concentration of 30 to 60 µg/ml, for 1 to 2 hours decreases incorporation of acetate into fat and increases ATP production by oxidizing palmitate(190, 191). After 2 hours of treatment with 40 µg/ml C75, ATP production was significantly increased in RARRES1-depleted cells compared to C75-treated control cells
(Figures 3.6B). The RARRES1- depleted cells were also capable of increasing ATP content after
3 hours of treatment unlike the control cells, in which a significant decrease in ATP concentration occurred after this time. This indicates that although the RARRES1-depleted cells have more fatty
86 acids available to oxidize for FAO the cells prefer to store the fatty acids in nutrient-rich conditions until FAO is triggered by C75.
Another way to trigger fatty acid oxidation is through glucose and serum starvation. We next monitored the mitochondrial activity of transient RARRES1 depleted MCF 10A and PWR-1E epithelial cells in starved conditions and inhibited CPT with etomoxir. MCF 10A and PWR-1E cells were used to ensure that the changes in fatty acid oxidation are due to epithelial specific effects rather than MCF10A specific changes. We first monitored the utilization of exogenous fatty acids for FAO by treating the cells with exogenous palmitate or BSA (as a control). The
OCR of RARRES1-depleted cells was indistinguishable from control cells after treatment with exogenous fatty acids (Figure 3.7B). We next examined the utilization of endogenous fatty acids for FAO during starvation. RARRES1-depleted PWR-1E and MCF 10A cells had a distinct oxygen consumption pattern compared to control cells. RARRES1 depleted MCF 10A cells had an increase in mitochondrial respiration that was reversed by treatment with etomoxir (Figure
3.6C). The normal metabolic needs of prostate epithelial cells are different than mammary epithelial cells. For example, they do not produce endogenous fatty acids through DNL; in turn, they rely on exogenous fatty acids to induce fatty acid oxidation. Specifically during glucose starvation, they rely solely on exogenous fatty acids to produce ATP through mitochondrial respiration(192). However, even in these cells, RARRES1 depletion significantly reversed the phenotype (Figure 3.7C).
These data demonstrate that RARRES1 depletion increases availability of endogenous fatty acids for FAO, thereby enabling these cells to improve their mitochondrial respiration for ATP
87 production during starvation. The data also suggests that RARRES1 does not affect the activity of fatty acid oxidation but rather the pool of fatty acids available for degradation and ATP production. As transient starvation is a common phenomenon that occurs during nutrient deprivation and in the center of large tumors, our data suggest that cells may adapt to these harsh environments by reducing RARRES1 levels.
A B 15
** 8000 ** * 10 31.1% 6000 n.s. 10.3% * ** 4000
Fold change 5 63%
2000 relative to scrambled siRNA relative to scrambled siRNA ATP Concentration (nM) 0 0
AC-04:0 AC-16:0
Vehicle-MCF 10A Vehicle-MCF 10A KD C C75-MCF C75-MCF10A 2 Hours 10A 3 Hours C75-MCF10AC75-MCF KD 2 10AHours KD 3 Hours
300 Rot./Ant. A.
80 *** *** 200 Oligomycin 60 ** *** 60 ** *
FCCP 40 40 100
OCR (pmol/min) 20 20 OCR (pmol/min) ATP Production OCR (pmol/min) Basal Respiration
0 0 0 0 20 40 60 80
Time (minutes) KD Vehicle KD Vehicle KD Etomoxir KD Etomoxir Scramble Vehicle Scramble-Vehicle Scramble-Etomoxir Scramble Vehicle Scramble Etomoxir Scramble Etomoxir KD-Vehicle KD-Etomoxir
Figure 3.6. RARRES1 Increases Substrate Availability for Fatty Acid Oxidation. A. All detected Acylcarnitines, AC-4:0 and AC-16:0, were quantified and validated in the non-targeted LC-MS results of RARRES1 siRNA transfected MCF 10A cells (Fig. S3). The fold change was normalized to scrambled siRNA transfected MCF 10A cells. B. ATP content was quantified in transient RARRES1 KD MCF 10A cells and scrambled siRNA MCF 10A cells after 2 and 3 hours of 40 ug/ml C75 treatment. C. MCF 10A were starved and treated with etomoxir to measure fatty acid oxidation rate dependent on endogenous fatty acids. Scrambled siRNA and transient RARRES1 siRNA were measured and compared. Basal respiration and ATP production were quantified.
88 S5 Figure
A 100 90 80 70 60 50 40 Oxidation) 30 20 Fatty Acid Oxidation (% of GLC + GLN + FA 10 0 Empty PLKO Vector RARRES1 TIG1 KD shRNA Dependency Flexibility
B Scramble Palmitate-Vehicle-MCF10A 250 Scrambled siRNA Palmitate-Vehicle-MCF10A 250 KD Palmitate-Vehicle-MCF10A ScrambledScramble Palmitate-Etomoxir-MCF10AsiRNA Palmitate-Etomoxir-MCF10A RARRES1 siRNA Palmitate-Vehicle-MCF10A 200 ScrambledScramble BSA-Vehicle-MCF10AsiRNA BSA-Vehicle- MCF10A 200 KDRARRES1 Palmitate-Etomoxir-MCF10A siRNA Palmitate-Etomoxir-MCF10A KDRARRES1 BSA-Vehicle-MCF10A siRNA BSA-Vehicle- MCF10A 150 150
100 100 OCR (pmol/min) 50 OCR (pmol/min) 50
0 0 0 20 40 60 80 0 20 40 60 80 Time (minutes) Time (minutes)
C Oligomycin 60 L)
g/ µ 15 µ 40 L)
g/ µ * µ *** 10
20 5 Basal Respiration OCR (pmol/min)/( OCR (pmol/min)/( 0 0 0 20 40 60 80 Time (minutes)
Scrambled siRNA-Vehicle Scrambled siRNA-Etomoxir RARRES1 siRNA-Vehicle RARRES1 siRNA-Etomoxir RARRES1siRNA-Vehicle Scrambled siRNA-Vehicle Scrambled siRNA-EtomoxirRARRES1 siRNA-Etomoxir
15 ** *** L)
g/ µ 10 µ
5 ATP Production (pmol/min)/(
0
Scrambled siRNA-Vehicle RARRES1 siRNA-Vehicle Scrambled siRNA-EtomoxirRARRES1 siRNA-Etomoxir
Figure 3.7. Fatty Acid Oxidation Activity in RARRES1-Depleted Epithelial Cells. (A) Fatty acid oxidation rate in RARRES1-depleted cells was measured in nutrient rich media. FAO dependency and flexibility of the cells were calculated. Initial inhibition of FAO by etomoxir measures how dependent the cells are on that particular fuel source to meet energy demand. Using a combination of glycolytic, glutaminolytic and FAO inhibitors, the cells' capacity and flexibility in the meeting energy demand were calculated in terms of oxygen consumption rate. (B) Fatty acid oxidation dependent mitochondrial respiration was quantified and the effects of exogenous fatty acids on OCR in RARRES1-depleted and scrambled siRNA MCF 10A cells in glucose and serum starved media. (C) PWR-1E cells were starved and treated with etomoxir to measure fatty acid oxidation rate dependent on endogenous fatty acids. Scrambled siRNA and transient RARRES1 siRNA were measured and compared. Basal respiration and ATP production were quantified.
89 RARRES1-Depleted Epithelial Cells Enhance De Novo Lipogenesis
Because RARRES1-depleted cells increased lipid content and enhanced fatty acid oxidation
activity due to improved utilization of endogenous fatty acids, we next examined the effects of
RARRES1 on endogenous fatty acid synthesis. De novo fatty acid synthesis occurs when sugars,
such as glucose, get converted to citrate and subsequently acetyl-coa to produce cholesterol and
the fatty acid metabolites that were enriched in RARRES1-depleted MCF 10A cells (Figure 3.4A,
& 3.4B)(193). Consistent with either a high conversion rate of citrate into acetyl-coa and complex
fatty acids or to decreased DNL, steady state citrate levels were decreased in stable RARRES1- depleted MCF 10A cells (Figures 3.8A & 3.9). As RARRES1- depleted cells improve their utilization of endogenous fatty acids (Figure 3.6) but not exogenous palmitate, this shows that the cells increase endogenous fatty acid synthesis, i.e. DNL.
Since glucose is a major source for DNL, we assessed if RARRES1-depleted PWR-1E and MCF
10A cells undergo glycolytic reprogramming to redirect glucose usage for de novo fatty acid synthesis(194). Aerobic glycolysis was examined in RARRES1-depleted epithelial cells using the extracellular flux assay- Glycolysis Stress Test kit (Seahorse, Agilent Technologies, Santa Clara,
CA, USA). Glycolytic usage was determined by measuring the extracellular acidification rate reached by a given cell after the addition of saturating amounts of glucose. The glycolytic capacity was determined after oligomycin injection. The capacity depicts the maximum rate of conversion of glucose to pyruvate or lactate that can be accomplished immediately by a cell(195). The difference between basal glycolysis and the glycolytic capacity determines the glycolytic reserve, which is glucose that is not used in the basal state. The glycolytic reserve can provide insight on whether glucose is stored or shuttled to other pathways besides aerobic glycolysis. RARRES1-
90 depleted MCF 10A cells had a marked decrease in glucose usage but their capacity to induce
glycolysis was higher than the control cells (Figure 3.10A), which indicates that there is no glycolytic dysfunction. OCR after glucose injection was assessed to examine if glucose was instead used for mitochondrial respiration. There was no change in OCR (Figure 3.10B). The
glycolytic capacity phenomenon was also observed in PWR-1E cells (Figure 3.10C). Thus, in
both cell lines their glycolytic reserve increased (Figure 3.8B). This indicates that in RARRES1-
depleted cells, glucose is either stored or being reprogrammed for other pathways instead of being
used for aerobic glycolysis or oxidative phosphorylation.
The changes in citrate levels, lipid synthesis and glycolytic reprogramming in RARRES1-
depleted epithelial cells suggest that glucose is mostly likely used for glucose dependent DNL
(Figures 3.4 & 3.8A). To test this, we used the fatty acid synthase inhibitor C75. At
concentrations of 10µg/ml (40 µm) and 30 minutes to 2 hours of incubation, C75 treatment
markedly decreases lipid content and increases citrate levels(196–199). We next examined
whether C75 treatment can reverse the glycolytic reprogramming seen in RARRES1-depleted
cells. We treated RARRES1-depleted MCF 10A cells with C75 for 1 hour or 2 hours at a concentration of 40 µm. In C75 treated MCF 10A cells, both control and RARRES1-depleted cells
increased glycolytic usage and decreased their glycolytic reserve compared to the vehicle treated
groups (Figure 3.8C). These observations indicate that MCF 10A cells have a significant basal level of DNL. Importantly, treatment of RARRES1-depleted cells with C75 completely reversed the decrease in glucose usage seen in the vehicle-treated cells and they had higher aerobic
glycolytic activity (Figure 3.8C). C75 also reversed the changes in glycolytic reserve seen in the
transient knockdown, as the glycolytic reserve is indistinguishable from the control group treated
91 with C75 (Figure 3.8C). This indicates that RARRES1-depleted cells use glucose for fatty acid synthesis instead of aerobic glycolysis and that inhibition of FAS by C75 reverses this phenotype
(Figure 3.8D).
Glycolytic A B Capacity
2.0 Glycolysis 2-DG 120 ** 300 **** 1.5 100 231%
Glycolytic
Reserve 80 O 200 1.0 * 60 G 31% Citrate 40 100 (Fold Change) 0.5
20 ECAR (mpH/min)
0.0 ECAR (mpH/min)/1000 Cells 0 0
0 20 40 60 80 Glycolytic Reserve (%Scramble) Time (Min) Scrambled siRNA - MCF 10A
Empty Vector RARRES1 KD RARRES1 siRNA - MCF 10A Scramble - PWR-1E Scramble - MCF 10A RARRES1 KD - PWR-1E Scrambled siRNA-PWR-1ERARRES1 siRNA-PWR-1E RARRES1 KD -MCF 10A Scrambled siRNA-RARRES1 MCF 10A siRNA- MCF 10A C Glycolytic Capacity
2-DG O 100 Glycolysis ** 80 300 ***
Glycolytic
Reserve ** 80 *** 60 200 60 ** 40 40 G *** 100 20 20 Glycolysis (mpH/min) Glycolytic Reserve (%)
ECAR (mpH/min)/1000 cells 0 0 0
1.33 8.95 16.6324.3432.0139.6347.3855.0062.6570.4878.1785.85 Time (Min) Scrambled siRNA (C-75 Treatment) RARRES1 siRNA (C-75 Treatment)
Scrambled siRNA - Vehicle RARRES1 siRNA - Vehicle Scrambled siRNA - Vehicle RARRES1 siRNA - Vehicle
D Scrambled siRNA - C-75 TreatmentRARRES1 siRNA - C-75 Treatment Scrambled siRNA - C-75 TreatmentRARRES1 siRNA - C-75 Treatment
Figure 3.8. RARRES1 Regulates De Novo Lipogenesis. A. Citrate levels in stable RARRES1 knockdown MCF10A cells were calculated using MS/MS method and compared to empty vector MCF10A cells. B. Glycolytic activity was assessed in MCF 10A and PWR-1E cells. The glycolytic reserve was calculated for both cell lines. C. Transient RARRES1 KD and scramble control were treated with C75 inhibitor and glycolytic activity was assessed. Glycolytic usage and reserve were assessed in vehicle and C75 treated groups D. A schematic diagram of glycolysis and glucose dependent de novo lipogenesis pathway in RARES1 depleted epithelial cells and the effect of C75 on these cells is depicted. G: Glucose; O: Oligomycin; 2-DG: 2-Deoxy-D-glucose.
92 S6 Figure
A
Figure 3.9. LCS7- MSFigure of Stable RARRES1 Knockdown MCF 10A Cells. (a) MS/MS LC-MS was done to validate and quantify citrate in stable knockdown cells. A B 150 150 1.26e-007
100 100 7.994e-014
50 50 ECAR (%Scr) ECAR (%Scr) Glycolytic Usage Glycolytic Capacity
0 0
MCF 10A - Scramble MCF 10A - Scramble MCF 10A - RARRES1 KD MCF 10A - RARRES1 KD
C 100
80 Scramble - PWR-1E RARRES1 KD - PWR-1E 60 RARRES1 KD - PWR-1E
40 n.s. 80 40 30 *** ECAR (mpH/min) 60
20 40 20
10 20 ECAR (mpH/min) ECAR (mpH/min) Glycolysitic Usage Glycolytic Capacity
0 0 0 0 20 40 60 80 Time (Min)
Scramble - PWR-1E Scramble - PWR-1E RARRES1 KD - PWR-1E RARRES1 KD - PWR-1E
Figure 3.10. Glycolytic activity in RARRES1-depleted Epithelial Cells. (A) Glycolytic usage and capacity was quantified in transient RARRES1 knockdown in MCF 10A cells by using the Glycolysis Stress Test. (B) Oxygen consumption rate measurement was assessed after glucose injection in the Seahorse XF Flux machine. (C) Transient glycolytic activity was assessed in PWR- 1E cells with transient RARRES1 knockdown by using the Seahorse Glycolysis Stress Test. Glycolytic usage and capacity was quantified.
S-10
93
S-11 RARRES1 and Its Effects on Lipid Metabolism Enzymes
Fatty acid synthase catalyzes the formation of long chain fatty acids (complex lipids) from malonyl
coa and acetyl-coa (191). Since RARRES1 depletion increases the complex fatty acid content, we
thus assessed if RARRES1 can alter fatty acid synthase (FAS) protein level. We either depleted or overexpression RARRES1 in but we did not see any significant changes in protein level of FAS
(Figure 3.11A).
Our previous study has established that RARRES1 and its target can regulate AMP kinase activity
(Figure 3.14A & 2.14B). One of the most studied functions of AMP kinase is its regulation of
lipid metabolism (200). AMP kinase regulates lipogenesis through phosphorylation of acetyl-coa
carboxylase 1 (ACC1) at Ser79and ACC2 at Ser212 and in turn inhibits the production of malonyl-
coa, a substrate for fatty acid synthase (FAS) and precursor for the de novo synthesis of palmitate
(201, 202). We thus assessed the phosphorylation status of Ser79 in ACC1 and Ser212 in ACC2.
RARRES1 depletion decreased phosphorylation but the protein level of ACC1 and 2 also
decreased (Figure 3.11B). We also overexpressed RARRES1 in MCF 10A cells, we noticed a
significant increase in a lower molecular weight band (Figure 3.11C). The band was present when immunoblotting for ACC2 protein as well as the phosphorylated ACC2. This could indicate an increase in the degradation of ACC2. Ubiquitination status of the ACC2 protein must be assessed.
ACC2 phosphorylation and degradation after overexpression of RARRES1 must be further studied through proteomics and simple ubiquitination immunoblotting after overexpressing RARRES1.
94 Figure 3.11. FASN, ACC1 and ACC2 Expression in RARRES1 siRNA or RARRES1 Transfected MCF 10A Cells. A. FASN immunoblot was assessed in RARRES1 overexpressed and RARRES1-depleted MCF 10A cells. Tubulin was used as a loading control. B. Immunoblot of phosphorylated Ser79 ACC1 and ACC1 protein is displayed. RARRES1 expression and GAPDH (as a loading control) was immunoblotted. C. Phosphorylated Ser212 ACC2 and ACC2 protein were assessed through western blot after RARRES1-YFP overexpression. Tubulin was used as a positive control and RARRES1-YFP overexpression was confirmed.
95 Table 1 Table 3.1. Common Pathways Identified in the Analysis of Genes Co-expressed with RARRES1 in Breast, Prostate, and Colorectal Cancers.
Number of Genes Benjamini (Breast, (Breast; Colorectal; Prostate)/ P-Value (Breast; Colorectal, Pathways Enriched in all Datasets (Total Genes in Pathway) Colorectal; Prostate) Prostate)
ECM-receptor interaction (40; 28; 48)/ 87 9.32E-09; 0.0035; 1.10E-08 8.85E-07; 0.032; 5.38E-07 Retinol metabolism (27; 29; 30)/65 0.000034; 4.35E-06; 6.09E-04 0.00064; 9.05E-05;0.0043 Salivary secretion (31; 28; 40)/86 0.000175; 0.0029; 4.80E-05 0.0022; 0.022; 4.85E-04
PPAR signaling pathway (25; 21; 32)/67 4.88E-04; 0.017; 1.79E-04 0.004; 0.091; 0.00164
Steroid hormone biosynthesis (24; 25; 24)/58 0.00011; 4.66E-05; 0.012 0.0016; 6.17E-04; 0.054
Regulation of lipolysis/ Fat Digestion (24; 15; 24)/56 0.0000581; 0.0081; 0.0076 0.00092; 0.052; 0.38
PI3K-Akt signaling pathway (98; 83; 124)/345 0.000003; 0.0085; 1.17E-05 0.000078; 0.054; 1.71E-04 0.00000001381; 6.14E-09; 0.00000363; 4.47E-07; Protein digestion and absorption (40; 41; 41 )/80 3.61E-05 4.06E-04
Pathways that were significantly enriched and shared among the three cancer datasets (TCGA breast, TCGA Colorectal and Grasso Prostate Cancers) of genes co-expressed with RARRES1 were listed in the table. The number of genes identified in the pathways out of the total genes identified in the pathway was listed. Statistical analysis was applied for all datasets and the P-value and Benjamini- Hochberg value determined for each cancer dataset are included in the table. DAVID platform was used to identify the pathways relevant in the cancer datasets.
RARRES1 Gene Expression in Common Solid Tumors Correlates with Fatty Acid Metabolism Genes In addition to its role in hepatosteatosis and other metabolic disorders, RARRES1 is among the most commonly methylated genes in multiple human tumors (31, 102, 104, 105). As our data indicate that RARRES1 has a role in fatty acid metabolism, we wondered if the extensive transcriptomic data available for multiple cancers also points to a relationship with fatty acid metabolism pathways. We used publicly available cancer datasets to assess whether RARRES1 is 30 co-expressed with fatty acid metabolism genes. OncomineTM (Compendia Bioscience, Ann
Arbor, MI, USA) was used to identify cancers in which RARRES1 is significantly increased or decreased (Table 1.1 from OncomineTM). Of the identified cancers, three (breast, colorectal and
96 prostate) were selected for further analysis given the number of datasets found to have significant
RARRES1 gene expression changes. Interestingly, subtype analysis revealed differential
RARRES1 expression for more aggressive tumor phenotypes. For example, hormone-negative
breast cancers (e.g. TNBC vs. Non-TNBC, ER+ vs. ER-, PR+ vs. PR-) and metastatic prostate
cancer have differential RARRES1 gene expression (Figure 3.12A).
We then sought to determine important pathways associated with RARRES1 expression. Genes
co-expressed with RARRES1 were identified via oncominetm (Compendia Bioscience, Ann Arbor,
MI) with respect to each cancer dataset (original data file available upon request). The sets of co-
expressed genes were pipelined into DAVID(203, 204) for pathway enrichment analysis (adjusted
P<0.05). Enriched pathways common across all three cancer datasets were identified (Table 3.1).
As expected, the retinol metabolism and protein digestion pathways were enriched across all
datasets indicating a role of RARRES1 in retinoic acid biochemistry(36) and cellular
autophagy(32, 97). Importantly, pathways involved in lipid metabolism, synthesis and signaling
were identified (Table 3.1). These include PPAR signaling, steroid hormone biosynthesis and
regulation of lipolysis/fat digestion.
We next sought to assess the correlation of RARRES1 with important genes implicated in PPAR
signaling and fatty acid metabolism. These genes include PPARG, PPARGC1A, PPARA, SCD,
FASN and SREBF1. SREBF1 or Sterol regulatory element-binding transcription factor 1
(SREBP-1) protein is a transcription factor activated by mTOR which subsequently upregulates genes including SCD or stearoyl-coa desaturase 1 (SCD1) enzyme and FASN or fatty acid synthase
(FAS) that are essential in lipogenesis(205). The interaction of peroxisome proliferator-activated receptor alpha (pparα) protein or PPARA gene with peroxisome proliferator-activated receptor
97 gamma coactivator 1-alpha (PGC1α) protein or PPARGC1A gene induces the transcription of genes that are important in fatty acid oxidation(206). These genes correlate with RARRES1 gene expression changes when all cancers are considered (Figure 3.12A). For example, PPARGC1A and PPARA positively correlated with RARRES1 whereas FASN and SCD negatively correlated with RARRES1 gene expression. This correlation pattern was further recapitulated in all individual subtype analyses (e.g. TNBC vs. Non-TNBC) (Figure 3.12B). RARRES1 expression therefore is positively correlated with fatty acid oxidation genes (PPARGC1A, PPARA) and negatively correlated with lipogenesis genes (FASN, SCD). Taken together these data demonstrate the association of RARRES1 and fatty acid metabolism in the context of cancer.
A B
5 1.0 ER +/- -4.9 -2.9 -2.3 1.2 1.9 1.8 1.7 RARRES1 1.0 0.8 0.9 -0.3 -0.7 -0.8 -0.8
Breast Cancer -3.5 -2.4 -1.9 1.2 1.9 1.5 1.5 Progesterone +/- Correlation Coefficient (R -2.8e 0.8 1.0 0.9 -0.5 -0.8 -0.9 PPARGC1A -002 Triple Negative 5.5 3.1 2.5 -1.2 -2.2 -2.2 -2.2 Fold Change 0.5 PPARA 0.9 0.9 1.0 -0.2 -0.7 -0.9 -0.9 Prostate Cancer Metastatic Site/ Primary Site -7.8 -2.4 -1.5 -1.2 1.3 2.5 4.0 0 -2.8e 3.5e- Rectal Mucinous Adenocarcinoma -2.4 -2.7 -2.0 -1.5 1.0 1.6 3.1 PPARG -0.3 -0.2 1.0 0.5 -0.2 -002 002 0 Colon Adenocarcinoma -2.4 -5.3 -1.9 -1.6 1.2 1.8 3.7 SREBF1 -0.7 -0.5 -0.7 0.5 1.0 0.7 0.5
Colorectal Cancer Rectal Adenocarcinoma -2.4 -6.3 -1.8 -1.5 -1.3 1.8 3.7 2) 3.5e- -5 -0.8 -0.8 -0.9 0.7 1.0 0.8 FASN 002 -0.5 Colon Mucinous Adenocarcinoma -1.6 -2.6 -1.9 -1.9 1.5 1.6 3.9
Cecum Adenocarcinoma -1.9 -2.6 -1.7 -1.5 1.2 3.8 2.0 SCD -0.8 -0.9 -0.9 -0.2 0.5 0.8 1.0 SCD SCD FASN FASN PPARA PPARG PPARA PPARG SREBF1 SREBF1 RARRES1 RARRES1 PPARGC1A PPARGC1A
Figure 3.12. Differential Expression of RARRES1 in Cancers Correlates with Expression of Fatty Acid Metabolism Genes. Cancers where RARRES1 gene was differentially expressed were chosen and correlative analysis was done. FASN, SCD, PPARG, SREBF1, PPARGC1A and PPARA, were chosen as candidate genes important in fatty acid oxidation and lipogenesis. A. Three subtypes of breast cancer, 5 types of colorectal cancer and metastatic versus primary sites of prostate cancer were analyzed and fold change difference (with p-value <0.01) was plotted for each gene. B. The correlative score (calculated using Pearson correlation formula) was calculated between two genes in all cancers analyzed.
18 Discussion
In this study, we show that RARRES1 regulates fatty acid metabolism in epithelial cells. First, depletion of RARRES1 increased lipid accumulation along with altered fatty acylcarnitine
98 content. When RARRES1 was expressed in the presence of oleic acid, lipid accumulation was
hindered most likely by lipid degradation through activation of autophagy. Second, ATP
production was higher in RARRES1-depleted cells after treatment with a CPT1 agonist (FAO inducer). Third, triggering endogenous FAO by glucose and serum starvation improved mitochondrial respiration due to increased storage of endogenous fatty acids; a phenotype that was reversed by inhibiting fatty acid oxidation with etomoxir. Thus RARRES1-depleted epithelial cells increase lipid synthesis in a nutrient-rich state and have an improved bioenergetic phenotype during starvation due to an increase in the oxidation of stored fatty acids (Figure 3.13).
RARRES1 expression was also contextually correlated to fatty acid metabolism genes in multiple cancer types. This phenotype might be critical for the survival of cells in a nutrient deprived tumor environment or stem cell niche (207).
RARRES1 is silenced in colorectal cancer, prostate cancer, nasopharyngeal cancer, Wilms tumor and leukemia (100, 101, 104). Colorectal cancer and prostate cancer cell lines transfected with
RARRES1 were less invasive and more apoptotic(100, 106). These findings support the idea of
RARRES1 as a tumor suppressor, although this has not been formally demonstrated in animals.
However, RARRES1 is increased in some mesenchymal-like cancers such as triple negative breast cancer(33, 102, 164), (Figure 3.12A). The present study demonstrates an important role for RARRES1 in fatty acid metabolism and may explain the duality of RARRES1 in cancer etiology. The role of fatty metabolism on cancer progression is also contextual. During nutrient
deprivation, fatty acid degradation is necessary and fatty acid synthesis, specifically DNL, is
increased during metastasis, consistent with the decrease of RARRES1 expression in metastatic
prostate cancer(15, 166, 208). There is also evidence of changes in fatty acid metabolism
99 preference in different subtypes of breast cancer, in which RARRES1 is differentially expressed.
For example, fatty acid oxidation is disproportionately dysregulated in triple negative breast cancers compared to other subtypes of breast cancer (24, 209, 210). Hormone receptor status in breast cancer also correlates with changes in lipid metabolism (211, 212).
In summary, this study demonstrates that RARRES1 is a novel regulator of fatty acid metabolism and points to an important role in diseases in which lipid metabolism is a hallmark of disease progression. Treatments that regulate RARRES1 expression or activity could have utility in obesity, cholestatic liver disease, heart disease (in which RARRES1 expression is markedly decreased in hypertrophic and dilated cardiomyopathy) and cancers where lipogenesis is crucial for the progression of certain tumors (213–215).
100 Figure 5
Nutrient-Rich Starvation RARRES1 Depletion RARRES1 Depletion
Glucose Lipids
Fatty Acid Oxidation Pyruvate CPT1A
TCA TCA Lactate
CPT1A ETC Acylcarnitine
Citrate FASN ATP Lipids
Figure 3.13. RARRES1 Depletion in Epithelial Cells Increases De Novo Lipogenesis and Improves Fatty Acid Oxidation During Starvation. ETC: Electron transport chain; TCA: tricarboxylic acid cycle.
32
101 Chapter 3- Endogenous Transcriptional Regulation of RARRES1
Introduction
RARRES1 and CCP2 have a clear role in reprogramming metabolism, stem cell differentiation and cell survival with potential implications in diseases such as cancer, liver cirrhosis, pancreatitis, obesity and hyperinsulinemia (33, 35, 94, 95, 103, 216). In cancer,
RARRES1 expression is significantly down regulated (31, 32, 101–103). RARRES1 is considered a class II tumor suppressor. Class II tumor suppressors are functionally active in cancers but their transcription is downregulated, while class I tumor suppressor genes are often deleted or gain mutations and subsequently lose their functional activity (217). The gene is downregulated in these cancers due to hypermethylation of the promoter region in tumors
(31, 101, 102). RARRES1 is also differentially expressed in other diseases, such as obesity, where RARRES1 transcription is significantly downregulated, and fibrosis, where RARRES1 expression is significantly upregulated (33, 35). Thus identifying transcriptional regulation of RARRES1 is important as the gene is very often turned off or highly active in a disease context.
RARRES1 was first identified as a RA-responsive gene (36). Treatment with all-trans RA and 9- cis RA induces the expression of the gene. A series of events occur within the cell to shuttle the signaling molecule into the nucleus, where it induces transcription of retinoic acid- sensitive genes. First, all-trans-retinoic acid or 9-cis-retinoic acid binds to CRABPII or FABP5 and is transported into the nucleus. These retinoic acid binding proteins transport the retinoic acid into the nucleus and transfer the ligand to the nuclear receptors retinoic acid receptors (RARs) and retinoid X receptors (RXRs) (137). All-trans retinoic acid and likely 9-cis-retinoic acid bind to one or more of three retinoic acid receptors (RAR) and 9 cis-retinoic acid binds to one or
10 2 more of three retinoid X receptors (218). RARRES1 was initially found to be induced only by
retinoid analogues that activate RARs, specifically RAR beta and gamma, rather than RXRs
(36). Several publications have supported the idea of inducing RARRES1 with all-trans
retinoic acid treatment, which triggers RAR components rather than RXRs (45, 94, 219).
Others were able to prompt the expression of RARRES1 through 9-cis-retinoic acid, but as
mentioned earlier, 9-cis-RA can also induce RARs (45).
RARRES1 is also regulated by vitamin D3, an agent that promotes differentiation of colon, breast,
prostate and squamous carcinoma and myeloid leukemia cancer cells. The anti-tumorigenic effects
of vitamin D in cancer is similar to RA. We treated SKBR3 cells, a breast cancer cell line, with
retinoic acid and compared our microarray data with publicly available microarray data on vitamin
D treatment on SKBR3 and caco2 cell-lines (Figure 4.1). Very few genes are commonly regulated
by RA and vitamin D; one of these genes is RARRES1. We confirmed these findings in normal
prostate epithelial cells where both RA and vitamin D induced RARRES1 expression but its
expression did not change in prostate cancer cell lines in which the gene is heavily methylated
(45). Several studies showed the same trend in colorectal cancer cell-lines and keratinocytes (115,
220). RA and vitamin D both have anti-cancer properties, by promoting differentiation and inducing apoptosis in malignant cells (115, 137, 221). The fact that vitamin D and RA both upregulate RARRES1, indicates that this RA/Vit D target gene might be the main player in RA
and Vit D tumor suppressive effects.
103 Figure 4.1. RARRES1 is Commonly Regulated by Retinoic Acid and Vitmain D. A venn diagram was created to identify overlapping genes regulated by RA and vitamin D. The table on the right lists genes were commonly regulated in all cell-lines by retinoic acid and vitamin D (The data was analyzed and illustrated by Yun Ji).
Microarray profiling of cytokine cell stimulation reported direct induction of the RARRES1 gene following inflammation. IL-10 and IL-6 are cytokines known to date to induce such response. In one study, IL-10-treatment of macrophages increased RARRES1 expression 10-fold (222). In
contrast when IFN -gamma was used to activate monocytes into macrophages and later treated with
IL-10, RARRES1 expression decreased by 11-fold (222)I . L-10 induces homodimerization of
STAT3 whereas IFN-gamma with IL-10 induces the activation of STAT1 (222, 223). This indicates that RARRES1 is induced by the STAT3 transcription factor rather than STAT1 and that
the gene is induced by IL-10 but inhibited by IFN-gamma, a cytokine that has anti-inflammatory and immunosuppressive effects. IL-10 mainly targets antigen-presenting cells and is induced as a
negative feedback response to inflammation. IL-10 regulates the duration and strength of the
inflammatory response (224). The literature clearly demonstrates that RARRES1 is directly
104 regulated by retinoic acid signaling and by IL-10 through STAT3 activation. In this study we
intend to identify other transcriptional regulators of the RARRES1 gene.
Methods
Cell Culture
Primary HH (#HUFS1M, Lot#HUM4132, Triangle Research Labs, Durham, NC) were cultured
in hepatocyte medium (#5201, sciencell Research Laboratories, Carlsbad, CA). Triangle
Research Labs provides freshly isolated human hepatocytes, which undergo a proprietary isolation
procedure that guarantees 96-99% purity, and only healthy and viable cells, determined through
trypan blue staining, are provided. (Wobser H, Dorn C, Weiss TS, Amann T, Bollheimer C,
Buttner R, et al. Lipid accumulation in hepatocytes induces fibrogenic activation of hepatic stellate
cells. Cell Res. 2009 Aug;19(8):996-1005. Pubmed PMID: 19546889.) MCF10A cells were maintained according to the recommendation of American Type Culture Collection (ATCC).
Starvation and Treatment
MCF 10A cells were starved of 5% serum or of insulin, HC, and EGF, concentrations that are
recommended for the culturing of MCF 10A cells, overnight. Cells starved overnight were then
supplemented with HC, insulin or EGF at the recommended concentrations for MCF 10A cell
culture. The concentrations can be found in the ATCC website.
siRNA Transfection
YY1 silencerr (Cat # AM16708) and negative control siRNA (Cat # AM4611) were from Thermo
Scientific Dharmacon. Cells were seeded at 100,000 cells per well in 6-well plastic tissue culture
dishes on day 0. On day 1, for each cell sample, 7.5μl of Lipofectamine 3000 (Invitrogen, Carlsbad,
CA) was added to 125μl of OPTI-MEM I (Gibco-Invitrogen, Grand Island, NY) in a sterile
105 eppendorf tube, and mixed. 2.5μg of DNA or 75pmol of siRNA in 125ul of OPTI-MEM were added with or without 5ul of P3000, mixed, and incubated for 5 minutes at room temperature. The mixture was then added dropwise to cells. DHA Treatment 761 mm DHA stock was diluted to 1 mg/ml in 1XPBS and vortexed until dissolved. The DHA working solution was added and conjugated to 1% BSA at a 1:10 dilution. The solution was incubated for 1 hour at 37℃. The solution was then diluted in the cells’ media to the desired concentration. The cells were incubated for 30 min, 1 hour, 2 hours, or overnight (17 hours) at a concentration of 50 μM or 200 μM. The cells were then harvested in Trizol after incubation time. Mitochondrial Inhibitors Treatment MCF 10A cells were treated with 0-10 mm metformin (diluted in water), 100 nm antimycin A (diluted in DMSO) and 10 μM oligomycin (diluted in DMSO) overnight. Cells were harvested and RNA was extracted with Trizol. Extracellular Flux assay, Cell Mito Stress Test, was utilized to confirm decrease in mitochondrial respiration after treatment with drugs. NURSA Data Mining, Harmonizome and PROMO Analyses Promoter and non-coding regions of RARRES1 gene were obtained from Ensembl and UCSC genome browser. The promoter regions from both sources were then pipelined into PROMO to predict transcription factors that can bind to the promoter region. We then utilized Harmonizome database and examined publicly available data on chip sequencing and chip-chip to identify whether the predicted transcription factors were found physically bound to the promoter region of RARRES1 gene in cells or organisms. NURSA was then used to assess whether RARRES1 was regulated by signaling pathways that activate the transcription factors predicted.RARRES1 gene in cells or organisms. NURSA was then used to assess whether RARRES1 was regulated by signaling pathways that activate the transcription factors predicted.
106 Results Density Dependent RARRES1 Expression
Huebner et al. Used a trophoblast model and discovered that RARRES1 expression is density dependent. Where increasing the density of these cells increased the mRNA levels of RARRES1
(225). We thus assessed whether this phenomenon is relevant in mammary epithelial cells,
MCF 10A cells. We also noticed the same phenomenon in MFC 10A cells, where RARRES1 expression is unchanged on a per cell basis at sub-confluent levels but markedly increases once the cells reach confluence (Figure 3.2). Contact inhibition, a phenomenon which forces the cells into cell cycle arrest between G0/G1 phase, occurs when non-transformed cells are 100% confluent. This process also inhibits cells from migrating by changing polarity of the cells. In malignant cancer cells, contact inhibition is lost thus enhancing cell proliferation and invasiveness of cells. We showed earlier that RARRES1 depletion does not affect cell cycle kinetics (Figure 2.3A). However, RARRES1 does regulate the invasiveness of prostate cancer cells and colorectal cancer cells. Thus, RARRES1 might be essential during contact inhibition to attenuate invasion in cells. Serum starvation has also been used as a means to trigger cells to synchronize to G0/G1 phase. We thus assessed whether RARRES1 responds to complete serum starvation. We removed growth factors, epidermal growth factor (EGF), hydrocortisone
(HC) and insulin, and 5% horse serum that’s supplemented in the media. We saw a significant increase in RARRES1 expression in a time-dependent manner (Figure 4.3A).
107 Figure 4.2. Relative RARRES1 Expression at Different Densities. MCF 10A cells were plated at different densities and RARRES1 mRNA levels were quantified using qPCR 18s was used as a loading control. X-axis represents the different densities and the densities were numbered 1-7. 1 being the lowest density and 7 being the highest.
108 Figure 4.3. RARRES1 Expression is Regulated by Growth Factors. RARRES1 expression was assessed in all experiments using qPCR. 18S was used as the house keeping gene (endogenous control). A. MCF 10A cells were plated at sub-confluent levels and serum starved. RARRES1 expression was assessed after 18 hours and 40 hours of serum starvation. B. Growth factors, insulin and EGF, were removed from media and RARRES1 expression was assessed. Cells were serum starved and growth factors added to the media and RARRES1 expression was assessed. C. Cells were treated with EGF and serum starved and RARRES1 expression was assessed. D. Cells were serum starved and treated with insulin and RARRES1 expression was examined. E. Cells were serum starved and treated with hydrocortisone; subsequently RARRES1 expression was assessed. F. Cells were treated with Rapamycin or cells were serum starved; simultaneous rapamycin treatment with serum starvation could not furthe r increase RARRES1 expression. Insulin and EGF were used to evaluate whether Rapamycin can reverse the effects of EGF or insulin treatment. Rapamycin was able to partially reverse the effects of EGF treatment but not insulin. qPCR was run to assess RARRES1 expression in all conditions. 18S was used as the endogenous control.
109 RARRES1 is Regulated by Growth Factors and mTOR
Serum contains many components, that enhance the survival of cells by promoting energy
homeostasis and cell proliferation (226). We assessed which components within the media regulated the expression of RARRES1. We first examined whether the growth factors used to supplement the media, EGF and insulin, have any effect on RARRES1 transcription. We were able to mimic the effects of “complete serum starvation” on RARRES1 expression by removing growth factors insulin and EGF from the media and leaving 5% horse serum (Figure 4.3B). We also added
each factor separately in the media when cells were starved of 5% horse serum and growth factors.
Insulin, hydrocortisone and EGF significantly decreased RARRES1 expression during complete
serum starvation (Figure 4.3B-4.3E)
Insulin and EGF drive pro-survival signaling pathways and it is known that there is crosstalk
between these pathways (227). The common target of these signaling pathways is the mammalian
target of rapamycin (mTOR) (227, 228). mTOR is a central regulator of cell metabolism,
proliferation, growth and survival (229). mTOR forms complexes with different proteins and
activates distinct signaling pathways dependent on the complex formed (229). The best studied
mtor complex is mTOR complex 1 (mTORC1). mTORC1 is a master regulator of cell survival
and proliferation; the complex positively regulates growth by promoting anabolic processes, i.e.
Fatty acid synthesis, protein synthesis and restricts catabolic activity, such as autophagy (29, 229,
230). We have demonstrated that RARRES1 can regulate energy homeostasis; specifically, its
depletion increases anabolic pathways such as lipid synthesis. We thus assessed whether mTOR
can regulate the expression of RARRES1. Rapamycin inhibits the activation of mTOR and thus
110 decreases anabolic pathways while inducing a catabolic state in the cells (231). We treated cells
with rapamycin without starving the cells of growth factors or serum. Rapamycin treatment
induced RARRES1 expression within 48 hours. We used c-Myc, which is induced by rapamycin,
as a positive control to ensure rapamycin was in fact working (Figure 4.3F). We thus assessed
whether we can reverse the inhibitory effects of EGF and insulin on RARRES1. Rapamycin
was not able to reverse the effects of insulin but we saw a partial increase in RARRES1
expression when we treated cells with EGF and rapamycin. Thus the effects of EGF on RARRES1
expression is partly through mTOR activation (Figure 4.3F). The data above clearly
demonstrates a role for RARRES1 during depletion of growth factors.
RARRES1 Expression during Hypoxia and Nutrient Deprivation
mTOR has the ability to sense a variety of essential nutrients and respond to changes in metabolic
pathways. For example, the activation of TOR complex 1 is tightly regulated to the availability of
nutrients, like glutamine, glucose and oxygen. Since RARRES1 appears to be a target of rapamycin
and reprograms metabolism in the cell, we examined the effects of nutrient availability on
RARRES1 expression. We first sought the effects of glucose and glutamine, two important
metabolic molecules for TCA cycle, on RARRES1 expression after starvation. Glucose and
glutamine starvation did not significantly affect the expression of RARRES1 in MCF 10A cells
(Figure 4.4).
111 Figure 4.4. RARRES1 Expression after Treatment with Glutamine or Glucose in Serum Starved MCF 10A Cells. A. 10 mm glucose was added in glucose free treatment. Cells were treated with glucose for 24 hours. RARRES1 expression was assessed using qPCR. B. RARRES1 expression was assessed, using qPCR after glutamine treatment at a concentration of 2 mm or 8 mm in serum starved MCF 10A cells.
Mitochondrial inhibitors such as oligomycin, mimic oxygen deprivation in the cell. Oligomycin
blocks phosphorylation of p70s6k, a downstream target of mTOR, similar to hypoxic conditions
(231). We thus examined whether RARRES1 transcription can be regulated by inhibiting
mitochondrial respiration. We first utilized metformin, an FDA approved drug that is used for type
II diabetes (119). One of its mechanisms of action is inhibition of mitochondrial respiration.
Studies have shown in vitro that metformin binds to Complex I of the electron transport chain and inhibits mitochondrial respiration (119). We confirmed this by treating MCF 10A cells with metformin for 6 hours with concentrations of 0.5 mm, 1 mm, 5 mm and 10 mm. As described
previously in the literature a decrease in mitochondrial respiration occurred when cells were treated
with 1 mm of metformin (Figure 4.5A). This concentration is around 35 times higher than the
concentration found in the plasma of animals or humans that have taken metformin. Although this
mechanism of action might not be physiologically relevant, at these high concentrations,
metformin does decrease oxygen consumption. We then measured RARRES1 expression after metformin treatment. RARRES1 expression decreased markedly only at higher concentrations of metformin, when mitochondrial respiration is drastically attenuated. We then treated cells
112 with potent mitochondrial inhibitors to ascertain if RARRES1 expression changes as a consequence is of mitochondrial inhibition. Antimycin A and oligomycin, at concentrations known to inhibit mitochondrial respiration, decrease RARRES1 expression significantly
(Figure 4.5B). CCP2 expression significantly increased after mitochondrial respiration is inhibited (Figure 4.5B). As RARRES1 and CCP2 have reciprocal effects on mitochondrial respiration his suggests that cells might decrease RARRES1 and increase CCP2 expression after mitochondrial inhibition as a coping mechanism to rescue mitochondrial activity.
Figure 4.5. Inhibition of Mitochondrial Respiration Regulates RARRES1 and CCP2 Expression: A. Extracellular flux assay was utilized to measure oxygen consumption in MCF 10A cells after metformin treatment. The Mito Stress Test was used to measure ATP production rate and basal oxygen consumption rate. B. qPCR analysis was used to measure RARRES1 and CCP2 mRNA expression after treatment with antimycin A, oligomycin or metformin. 18S was used as an endogenous control.
113 Prediction of Transcription Factors and Signaling Pathways that Regulate RARRES1 Transcription
To determine the factors that regulate RARRES1 transcription, we utilized NURSA and PROMO
databases (232, 233). NURSA is a database that allows one to explore a relationship between
nuclear receptors and genes of interest (232). The platform integrates all publicly available data of
tissue-specific nuclear receptor transcriptomes based on published transcriptomics studies in the
field of nuclear receptor signaling (232). This platform will enable us to identify potential nuclear
receptor signaling pathways that can regulate RARRES1 transcription. PROMO is a
comprehensive bioinformatics approach to predict transcription factor binding sites within a genome region (233). In our search for predicted transcription factors, we limited our analysis to the RARRES1 promoter region and selected transcription factors predicted with matrix dissimilarity of less than 15%. We combined both databases and chose signaling pathways that are most likely to regulate RARRES1 in our system, by identifying transcription factors predicted to bind to the promoter region of RARRES1 (Table S4.1) and identifying if the transcription factor is linked to any signaling pathways in which RARRES1 is differentially expressed in the NURSA database (Figure S4.1). We also included insulin and EGF signaling as factors/signaling pathways that regulate RARRES1 as Figure 3.3 shows that these factors are important RARRES1 regulators.
In the PROMO platform, there were 30 transcription factors predicted to bind to the promoter region (Table S4.1). In addition to the retinoic acid receptors that are established as inducers of the RARRES1 gene, many transcription factors that are known to regulate metabolism, such as necrotic factor-kappa-light-chain-enhancer of activated B cells (NF-Kappa B), glucocorticoid
receptor-beta (GR-beta), Tumor protein 53 (TP53 or P53), pair boxed 5 (Pax-5), peroxisome
114 proliferator-activated receptor (PPAR) alpha and gamma, RXR-alpha and Yin Yang 1 (YY1)
transcription were identified.
YY1 Can Reverse the Effects of Growth Factor Starvation-Induced RARRES1 Transcription
Several binding sites for Yin Yang 1 (YY1) were found in the introns and promoter region of
RARRES1 (Figure 4.6A and Table S4.1). YY1 is a ubiquitous transcription factor. The C2H2- type zinc-finger transcription factor has several roles in embryogenesis, differentiation, replication and cellular proliferation (234). It can act as an activator or repressor of gene transcription (235).
YY1 is required for mitochondrial gene expression to regulate the structural integrity
and energetics of mitochondria in vivo and in vitro (234). YY1 deficiency in skeletal muscles
protects against rapamycin induced diabetic-like symptoms and was suggested to repress
insulin-IGF signaling (236). YY1 also plays a role in hematopoietic stem cell differentiation (237,
238). Similar to RARRES1, YY1 modulates mitochondrial energetics, responds to
rapamycin, insulin and starvation and regulates stem cell differentiation. We thus examined
whether YY1 regulates the transcription activity of RARRES1. We transiently depleted cells of YY1 for 24 hours and subsequently starved the cells for another 24 hours. We then harvested cells and ran quantitative PCR for mRNA analysis. RARRES1 induction by starvation was reversed by YY1 depletion (Figure 4.6A). Rapamycin and its target, mTOR
regulate YY1 transcription factor activity (235). Our previous publication showed that
RARRES1 regulates mTOR protein levels and reprograms metabolism in epithelial cells,
where RARRES1 depletion rendered cells more anabolic. Importantly we showed in
Figure 3.6B that RARRES1 is activated by rapamycin suggesting that rapamycin induction of
RARRES1 expression might be through YY1. However, YY1 depletion did not reverse the
effects of rapamycin (Figure 4.6B). Studies have identified Aurora kinase
115 families, A, B and C, as regulators of YY1 activity (239, 240). These kinases are essential for cell
proliferation and survival cells especially in the context of cancer. Looking at the phosphorylation
status of YY1 during starvation will enable us to identify whether RARRES1 can be regulated
through Aurora kinases. Other pathways also regulate YY1 activity, one example is OCT4 induced
pluripotency in stem cells differentiation (241–243). We have previously shown that RARRES1
can regulate the stem cell marker CD49f, in breast epithelial cells. CD49f is known to induce the
expression of OCT4 to enhance multipotency (244). OCT4-dependent YY1 activation of genes can be another possible mechanism of YY1 regulation of RARRES1 expression.
We showed earlier that RARRES1 expression is inhibited by insulin and EGF in the context of
serum starvation. We thus assessed which growth factor inhibits the induction of RARRES1
expression via YY1. Several studies have demonstrate an association between YY1 and insulin
and EGF signaling (236, 245, 246). EGF treatment alone combined with YY1 depletion further
decreased RARRES1 expression indicating that EGF inhibits RARRES1 expression through a
different pathway. We treated cells with insulin and found that YY1 depletion did not further
decrease RARRES1 expression (Figure 4.6C). This suggests that insulin signaling inhibits YY1
activation of RARRES1 transcription, while EGF regulation of RARRES1 is independent of YY1.
116 Figure 4.6. YY1 Reverses Growth Factor Starvation-Induced RARRES1 Expression. A. Predicted Binding sites of YY1 on the promoter region of RARRES1. Cells were serum and growth factor starved and depleted of YY1. RARRES1 mRNA expression was assessed through qPCR. B. Identification of the growth factor regulating YY1 activation of RARRES1 transcription during starvation. C. Insulin and EGF treatment were used to assess whether the growth factors regulate the activity of YY1 on RARRES1 transcription.
PPARs Regulate RARRES1 Expression
One of the top signaling pathways predicted to regulate RARRES1 transcription is PPAR signaling
(Figure 4.7A). The pathway was also identified as one of the most significantly co-expressed
pathways with RARRES1 in solid tumors (Table 3.1) (Table S4.1). PPAR signaling is especially
important in activating fatty acid oxidation and gluconeogenesis during starvation and increasing
117 lipogenesis in a fed state (247). PPARs can repress or induce transcription of genes (248, 249). As we previously showed a role of RARRES1 in these processes, we next assessed whether PPARs can regulate the transcription of RARRES1. ALGEN PROMO was used to predict all the binding sites of PPAR gamma and alpha in the non-coding regions of RARRES1 gene (Figure 4.7A and
Table S4.1) We also utilized the chip-X Enrichment Analysis (CHEA) transcription factor targets database, which identifies target genes of transcription factors from published chip-chip, chip-seq and other transcription factor binding profiling studies to further examine whether PPAR and
PPAR physically bind to the promoter region of RARRES1 in cells or animal models (250).𝛼𝛼 Both
PPAR𝛾𝛾α and PPARγ were identified as transcription factors that bind to the promoter region of
RARRES1 in vitro (Table S4.2). We then used hepatocytes and MCF10A cells that have high expression of RARRES1 to assess whether PPARs regulate RARRES1 transcription. RARRES1 responded to endogenous agonists of PPARs, docoshexanoic acid (DHA) (247) (Figure 4.7B).
Within 30 minutes of DHA treatment, RARRES1 expression decreased. This indicates RARRES1 is likely a direct target of this endogenous agonist of PPARs. We also treated cells with synthetic agonists of PPARγ and PPARα (rosiglitazone and pioglitazone and fenofibrate respectively). We observed induction of RARRES1 transcription by PPARγ agonists, rosiglitazone and pioglitazone, and reduction in RARRES1 expression with fenofibrate, a PPARα agonist, (Figures 4.7C and
4.7D) after 48 hours of treatment. Although our experiments only examined activation at 48 hours, the CHEA database does indicate that RARRES1 gene is a direct target of PPAR gamma and
PPAR alpha. For example, the mammalian ChIP-X database indicates that in several supplementary ChIP-Seq data in mice and human studies, PPAR and are found bound to the promoter region of RARRES1 (Table S4.2). The late induction of RARRES1𝛼𝛼 𝛾𝛾 could be to attenuate
118 the immediate metabolic effects of these PPARs. Rosiglitazone and pioglitazone are known to induce lipid
Droplets whereas fenofibrate increases fatty acid oxidation and decreases lipid droplet content (34,
247, 251–253). RARRES1 could be regulated to attenuate the metabolic effects that are induced by these agonists. PPAR-α is a key regulator of beta-oxidation (254). For example, fenofibrate, the
PPAR-α agonist, decreases dyslipidemia in metabolic syndrome (253). PPAR- α is also activated during fasting and enhances fatty acid oxidation for the generation of ketone bodies (255). PPAR-γ is especially important in adipocyte differentiation. PPAR- γ activation increases fatty acid storage and enhances glucose metabolism (256). Studies have repeatedly shown that PPAR- γ agonists, thiazolidinediones (TZDs), such as rosiglitazone and pioglitazone, improve insulin resistance, a number of genes involved in glucose and lipid metabolism are regulated by PPAR-γ (257–260). On top of the changes in lipid metabolites and fatty acid oxidation activity (Figures 3.1-3.6), the observations on PPAR γ and -α regulation of RARRES1 gene, place RARRES1 as an important player in the reprogramming of lipid metabolism.
119 Figure 4.7. PPAR Alpha and Gamma Regulate the Transcription of RARRES1 Gene. A. Binding sites of PPAR alpha and PPAR gamma on the RARRES1 promoter region. B. MCF 10A cells were treated with 50 M or 200 M for 30 min to 17 hours. C. MCF 10A cells were treated with 10 M rosiglitazone, 1 M pioglitazone or 30 M fenofibrate for 48 hours. D. Hepatocytes were treated𝜇𝜇 with 10 𝜇𝜇 M rosiglitazone, 1 M pioglitazone or 30 M fenofibrate for 48𝜇𝜇 hours. 𝜇𝜇 𝜇𝜇 𝜇𝜇 𝜇𝜇 𝜇𝜇
Discussion Our previous work demonstrated that RARRES1 is a master regulator of energy homeostasis. RARRES1 is a downregulated in cancer cells, upregulated in fibrosis and downregulated in obesity (32, 33, 35, 95, 102, 106). Thus transcriptional regulation of RARRES1 in disease context is crucial. Understanding how RARRES1 is regulated, may well identify how to manipulate the expression of RARRES1 in a disease context. In this study, we used multiple approaches to identify signaling pathways and transcription factors that regulate RARRES1 expression. We identified several metabolism related regulators of RARRES1 expression in
120 addition to the known regulators of RARRES1 expression, retinoic acid, vitamin D and the
inflammatory response.
Cell-cell contact inhibition is a phenomenon that occurs when normal cells contact each other. In
this contact inhibition phase, proliferation and differentiation is altered in a cell-density dependent
manner (261). This stage during proliferation is lost in cancer cells (138). A recent study has
suggested that RARRES1 expression is upregulated as the density of cells increases in a
trophoblast model (225). We see the same phenomenon in MCF 10A cells. RARRES1 does not
affect cell proliferation, we thus focused on how cell density can alter nutrient availability, we
were able to demonstrate that RARRES1 expression is dependent on growth factors, insulin, HC,
EGF, availability and responds to mitochondrial inhibitors.
In this study we found that insulin negatively regulates the expression of RARRES1 in mammary
epithelial cells. The microarray data of a previously published study focusing on hyperinsulinemia
confirms our findings. One of the genes significantly regulated after insulin treatment in human
skeletal muscles was RARRES1 gene (216). We also identified the mechanism behind how insulin
can regulate the expression of RARRES1. Insulin prevents YY1 from activating the expression of
RARRES1. By starving cells of insulin, the RARRES1 gene is re-expressed at high levels via
YY1.
EGF is a known pro-survival growth factor (262–264). It enhances growth of cells partly through
mTOR activation (227, 228). Here we show that RARRES1 expression is regulated by EGF partly
121 through mTOR activation by showing rapamycin can partially reverse the inhibitory effects of
EGF on RARRES1 transcription. We have demonstrated that RARRES1 depletion enhances cell
survival; the negative regulation of RARRES1 expression by EGF might be one mechanism the
growth factor affects survival and growth of cells. Future studies must assess whether EGF exerts
its effects on cell survival via RARRES1 by manipulating RARRES1 expression while treating
cells with EGF and assessing apoptosis.
RARRES1 and its target CCP2 regulate mitochondrial energetics and in this study we show that
RARRES1 and CCP2 are sensitive to mitochondrial inhibitors. RARRES1 expression decreased
after inhibiting mitochondrial respiration while CCP2 expression increased. We suggest that this
phenomenon is a coping mechanism by the cells to alleviate the consequence of acute inhibition
of mitochondrial respiration, as RARRES1 is a negative regulator of mitochondrial respiration
while CCP2 is a positive modulator. We need to assess the effects of RARRES1 and CCP2 genes
in hypoxic chambers to further elucidate their roles in low oxygen conditions.
Our previous study, has suggested that RARRES1 regulates lipogenesis and is co-expressed with
PPAR regulated genes (Figures 3.1-3.6 & Table 3.1). NURSA, Harmonizome and PROMO have all predicted PPAR gamma and alpha to regulate RARRES1. Publicly available chip-seq
supplementary data on PPAR alpha knockout mouse models and PPAR agonists confirmed this
speculation; we found that the RARRES1 promoter region is a target of PPAR and PPAR in the
datasets (250, 265). Here we show that in both MCF 10A and primary hepatocytes,𝛼𝛼 endogenous𝛾𝛾
and synthetic agonists of the PPAR family regulate the expression RARRES1. The mechanism behind how PPAR-alpha and -gamma can regulate the expression of RARRES1 is still elusive.
122 PPAR gamma and PPAR alpha have been repeatedly shown to activate as well as transrepress gene expression through different mechanisms. PPARs heterodimerize with retinoid X receptors,
RXR, and activate genes by binding to the specific DNA response elements in the their target genes (247). PPAR can also transrepress genes. Transrepression is when one protein represses the expression of the other through a protein-protein interaction (266). For example, PPARs repress transcription in a ligand-dependent manner by disrupting the activities of nuclear factor-κb (NF-
κb) and activator protein-1 (AP-1) (256, 267, 268). A chip analysis and DNA electrophoretic mobility shift assay must be run in the future to assess the PPAR complexes formed after TZD and fenofibrate treatment in the RARRES1 promoter gene. These PPAR agonists also have PPAR- independent effects on transcription by activating other nuclear receptors (269, 270). It is important to knockdown PPAR gamma and PPAR alpha in cells to confirm the effects of the agonists on
RARRES1 is through PPARs. PPARs can indirectly regulate expression of genes by upregulating or down-regulating the expression of co-activators or co-repressors that directly affect the expression of the genes (256). RARRES1 expression must be assessed after inhibition of protein synthesis during TZD and fenofibrate treatment to identify whether RARRES1 is regulated directly by PPARs or by PPAR targets. Regardless of the mechanism behind how these PPAR agonists regulate RARRES1 expression, the direct regulation of RARRES1 by docosahexaenoic acid and PPAR agonists strengthen our findings on RARRES1 being a novel regulator of fatty acid metabolism in epithelial cells.
In conclusion, our previous studies have identified RARRES1 as a novel regulator of energy homeostasis and identified that the protein is important in tumorigenesis and metabolism associated diseases. This study identified several pathways that regulate the expression of
123 RARRES1. In the future, we can utilize these pathways to find novel mechanisms to treat diseases with irregular expression of RARRES1.
124 Conclusion
Metabolic reprogramming in the context of cancer, diabetes and obesity is an important phase in
the pathology and progression of these diseases. Identifying important players in metabolic
rewiring is important to understand the disease etiology and recognize novel therapeutic targets.
We have shown that RARRES1 and its target, CCP2, are novel regulators of metabolic reprogramming. These two proteins have a unique mechanism to redirect metabolism and regulate
cell survival in epithelial cells; the interaction between RARRES1 and CCP2 regulates acidity of
the c-terminus of alpha tubulin, which in turn regulates the blockage of voltage dependent anion
channel. To the best of our knowledge, we are the first group to identify an endogenous inhibitor,
RARRES1, of microtubule associated carboxypeptidases and link their effects on the tubulin code
to mitochondrial VDAC and metabolism.
We have demonstrated that RARRES1 and CCP2 regulate the energetics of mitochondria and in turn affect coupling between mitochondria and glycolysis. For example, our data show that
RARRES1 depletion renders cells more energetic and redirects their energy to anabolic pathways while CCP2 and VDAC1 depletion has the opposite effect (Figure 5.1). Anabolic pathways that
RARRES1 affects include fatty acid synthesis and the hexosamine pathway. Three separate groups
have also identified a role for RARRES1 and CCP2 in autophagy. RARRES1 overexpression
induced autophagy, which supports the idea that higher levels of RARRES1 promote cell
catabolism. Microarray and RNA-Seq data show RARRES1 is differentially expressed in
metabolism associated diseases, such as obesity, hyperlipidemia and hyperinsulinemia (35, 95,
159, 162). Our findings on RARRES1 and CCP2 limiting substrate availability in epithelial cells
through VDAC1, explains how RARRES1 and CCP2 play an important role in cell survival,
125 metabolism associated diseases and biological hallmarks that are dependent on metabolic
pathways.
Figure 5.1. Proposed Effects of RARRES1 and CCP2 on Metabolic Reprogramming.
RARRES1 and Stem Cell Metabolism
We have identified RARRES1 and CCP2 as important players in differentiation of epithelial cells.
Depletion of RARRES1 in mammary epithelial cells increased CD49f, a stem cell marker for
epithelial cells (244), and our zebrafish in-situ hybridization of the RARRES1 orthologue gene,
demonstrate that the gene is localized in the intermediate cell mass and subventricular zone, areas
where hematopoiesis and neuronal differentiation occur respectively. We are currently developing
a RARRES1 knockout mouse model and our preliminary results suggest that RARRES1 plays an
important role in hematopoietic stem cell (HSC) differentiation and HSC niche localization during
development (data not shown). Oldridge et al. Have also demonstrated that RARRES1 and its homologue, latexin, are highly expressed in differentiated prostate epithelial cells, while their
126 expression is undetectable in progenitor prostate cells and stem cells (94). A recent study has
demonstrated that CCP2 is highly expressed in breast cancer stem cells, supporting our hypothesis
on CCP2 playing a role in stem cell function (271).
Metabolic reprogramming is a key feature in stem cell biology (3). Besides growth factors
and morphogens, the fate of stem cell differentiation, renewal and proliferation is dictated by
metabolic pathways (3). Glycolysis and oxidative phosphorylation are crucial in
differentiation and self-renewal of stem cells (3, 150, 207, 272–274). Although glycolysis is the
primary energy pathway used in stem cell cells, as stem cells divide and differentiate, their
oxidative phosphorylation capabilities increase (3, 118). We have shown that RARRES1
and its target CCP2 are both important in metabolic reprogramming. RARRES1 and CCP2
rewire glucose metabolism and regulate mitochondrial energetics. It is important to assess the
roles of RARRES1 and CCP2 in stem cells and progenitor cells. This could be assessed by
looking at the effects of self-renewal and expansion of these stem cells in the presence or
absence of RARRES1 and CCP2, and further assess whether stem cell RARRES1 and CCP2
reprogram their metabolic status to dictate their fate.
RARRES1 and Obesity
Several publications have identified both RARRES1 and CCP2 as potential players in obesity and
metabolic syndrome (35, 97–99, 216). A recent publication identified RARRES1 as a potential marker for adipocyte differentiation and demonstrated RARRES1 gene is induced during dedifferentiation of adipocytes, while RARRES1 expression decreases during differentiation of
preadipocytes (95). In one study RARRES1 was discovered to be the most up-regulated gene in
127 subcutaneous fat from obese human subjects on a diet-induced weight loss (35). In the same study,
RARRES1 was one of the most downregulated gene (top 15) in adipose tissue during weight
maintenance in the obese human subjects (35). Retinoic acid signaling, an inducer of RARRES1 transcription, was also demonstrated to negatively regulate differentiation of adipocytes. In the same study, microarray data showed RARRES1 was consistently and markedly upregulated in visceral stem cells compared to differentiated visceral adipocytes (96). RARRES1 has been associated with liver fibrosis and liver metabolism (33, 34, 152, 162, 163). RARRES1 gene expression is also significantly altered in a study focusing on the transcriptomic profile of hepatocyte specific PPAR knockout mice compared to wildtype mice. The hepatocyte pparα knockout mouse model had𝛼𝛼 lipid accumulation in the liver and developed steatosis(34).
Adipose tissues have a plastic energetic phenotype. For example, adipocytes respond to the
energetic and hormonal availability in the environment (275). In fact, adipocytes are the source for
energy-rich fatty acids when nutrient availability in the organism is depleted and imbalanced (276).
Adipocytes reduce their lipid uptake and decrease lipid storage during starvation (276, 277). When
nutrient is available in the environment, the cells expand and increase lipid uptake as well as
storage by increasing lipid droplet (275). Metabolic reprogramming is also a hallmark during
adipocyte differentiation (275). Pre-adipocytes tend to avoid storing lipid droplets and increase their catabolic pathways by increasing activation of AMP kinase (278). Differentiation of these progenitor cells increases lipogenesis and lipid uptake from the environment, enabling their expansion (279). We have shown that RARRES1 depletion in epithelial cells increase lipid droplets and phospholipids in the cells through changes in AMP kinase activity and substrate availability. Our work is consistent with the differential expression of RARRES1 in pre-adipocytes
128 and mature adipocytes (95). Future studies should focus on the effects of manipulating RARRES1
expression in adipocytes. Would increasing RARRES1 expression in adipocytes decrease lipid
content by enhancing catabolic activity? RARRES1 could be a potential target for obesity
and diseases related to dysfunction in adipocyte maturation.
Oncogene or Tumor Suppressor? – RARRES1 and Its Role in Fibrosis
Although RARRES1 is one of the most highly methylated genes in multiple cancers our analysis
on RARRES1 expression in different subtypes of solid common tumors has revealed a unique pattern in triple negative breast cancers. Re-examination of Oncomine and TCGA databases,
revealed that RARRES1 was also overexpressed in hepatocellular carcinoma and pancreatic cancers (Table 1). All these cancers are associated with extensive fibrotic pathology and it is likely that it is the fibrotic component of these cancers that makes high levels of RARRES1. As
mentioned earlier, RARRES1 was suggested to play an important role in activated stellate cells.
RARRES1 expression might be important in stellate cells due to its potential to
reprogram metabolism. Stellate cells have a unique metabolic status compared to resident
epithelial cells in tissue. When stellate cells are activated their lipid droplets disappear (280,
281). In the process of releasing these droplets, nutrients are also released into the environment
(282–284). Recent studies have shown pancreatic stellate cells promote growth of pancreatic
cancer cells by metabolic-cross talk (29, 282, 285, 286). Alanine released from stellate cells is
taken up by pancreatic cancer cells and promotes mitochondrial respiration and lipid
biosynthesis (282). There is also a close link between lipid metabolism and stellate cell
activation in liver cirrhosis (287). RARRES1 and its target do induce a catabolic state in
cells, RARRES1 induces autophagy. There is growing
129 evidence that autophagy is important for the loss of lipid droplets in stellate cells (288–
290). Attenuating autophagy in stellate cells, partially hinders stellate cell transitioning into myofibroblast-like cells (289, 291).
As RARRES1 appears to be strongly associated with fibrosis; future studies should elucidate the role of RARRES1 in stellate cells. Does its presence promote stellate cell activation? Does
it reprogram metabolism in stellate cells and promote cancer growth by enhancing
metabolic crosstalk between cancer cells and stellate cells?
Role of RARRES1 and CCP2 in Cardiovascular Diseases
ATP is essential for sustaining heart contractility. Central to the coordinated energy
transduction is the mitochondrion (292). The energy hub organelle generates more than 95% of
ATP used by the cardiomyocytes (293). Mitochondria also regulate calcium homeostasis, signaling and cell death (292). Mitochondrial VDAC1 is essential for the proper regulation of all major pathways that are directed by the mitochondria (79, 294). It is accepted that fatty acids
are the predominant substrate used in cardiomyocytes (189). However, other substrates such as
glucose do become favorable in different circumstances, such as exercise and prolonged fasting
(295). Such metabolic flexibility is crucial for the normal function of heart. RARRES1 regulates
fatty acid metabolism and modulates VDAC1 activity through CCP2, thus it is conceivable that
RARRES1 and CCP2 may play a major role in cardiac activity.
Microtubules are also an important factor in maintaining normal contractility in cardiomyocytes
(296). Along with its well-documented role as a roadway for cargo transport, the microtubule
130 cytoskeleton is linked to diverse structural and signaling roles in the cardiac myocyte (297). MTs can facilitate the rapid transmission of mechanical signals to intracellular effectors, a process called mechanotransduction (297). Recent evidence suggests that tubulin detyrosination is essential in cardiac mechanotransduction (298). One study elucidated how detyrosinated tubulin can regulate MT mechanical loads (69, 299, 300). During contraction, MTs must alter their geometry to accommodate the altered shape of myocytes. MT deform into sinusoidal buckled configuration and is reversed into a resting configuration after each beat. When the cardiomyocyte is compressed, systolic contraction occurs and the MTs buckle under load. The detyrosinated MTs are mechanically anchored to the sarcomere, the fundamental structure in muscles that enables contractility, and buckling during the contraction (297, 299). When detyrosination is attenuated, interaction between the sarcomere and MT is disrupted (297, 299). Thus the sacromeres are shortened and stretched and become less resistant (297). Detyrosination has been associated with cardiac disease, where detyrosination increased in patients that suffered cardiomyopathy (298,
299, 301). And suppression of detyrosination has been shown to improve cardiomyocyte function
(301). Intriguingly, in databases submitted by several human studies RARRES1 is highly downregulated in hypertrophic and dilated cardiomyopathy (215, 302). RARRES1 and its downregulation in failing cardiomyocytes could be explained through its regulation of PTM in MT mechanics, by increasing the glutamylation status of the MTs, or its regulation of metabolic activity in the cells, by modulating mitochondrial energetics. According to the Human Protein
Atlas, RARRES1 is expressed in the heart muscle (303). Therefore RARRES1 manipulation in vitro using cardiomyocytes as a model and assessing any metabolic changes through isotope tracing and extracellular flux assay and examining contractility and microtubule organization in these cells will help elucidate the specific roles of RARRES1 in healthy and diseased
131 cardiomyocytes. Other Targets of RARRES1?
We have demonstrated that RARRES1 targets CCP2, a tubulin deglutamylase. CCP2 is one of 6 members if the M14 CCP protein family. All ccps show the main determinants of substrate binding are conserved. All ccps preferentially bind to and cleave acidic amino acids. Thus
RARRES1 could potentially target other ccps and inhibit their activity. Exogenous co- expression of RARRES1 and all CCP members in HEK 293T and subsequent co- immunoprecipitation must be done to confirm this.
There are certain functions of RARRES1 that we cannot explain through CCP2. For example,
RARRES1 overexpression attenuates invasion of cancer cells, whereas CCP2 does not affect invasion (32, 94). As a putative carboxypeptidase inhibitor, RARRES1 can potentially regulate the activity of other enzymes. Endogenous co-immunoprecipitation of RARRES1 and subsequent proteomics in different tissues is essential to identify other potential targets of
RARRES1. Another phenomenon that cannot be explained through CCP2 is RARRES1 orthologue knockout in zebrafish is embryonic lethal while CCP2 knockout causes transient cell death. Thus RARRES1 most likely has other functions. We do have to keep in mind that zebrafish have one endogenous RARRES1 homologue (they do not express latexin), unlike mammals that express latexin and RARRES1 (Figure 1.2). Thus the orthologue knockout is not an ideal representative of the physiological effects of RARRES1. A full RARRES1 knockout in mice is also essential to understand the physiological implications of RARRES1. On the other hand, zebrafish do express all CCP isotypes found in humans (140). Thus the CCP2 knockout
132 might be a good representation of the physiological effects of CCP2 depletion. A CCP2 and CCP3 knockout mouse models were generated by the Janke group and the mice were viable and fertile (71).
They created a double AGBL2/AGBL3 knockout and these mice were still viable. AGBL2 and AGBL3 enzymatic activity on polyglutamylation is specific to the testes, no other organ showed detectable changes in tubulin polyglutamylation (71). Our RARRES1 knockout has a significant increase in polyglutamylated tubulin in the skin (Figure 3.3), which suggests RARRES1 most likely regulates other CCP isotypes. It would be interesting to compare CCP knockout models to conclude whether RARRES1 could regulate other carboxypeptidases and play a role in other pathways besides targeting CCP2.
It is clear that RARRES1 and CCP2 play a major role in metabolic reprogramming but how these changes apply in stem cell biology, obesity, fibrosis and overall organismal physiology is still poorly understood. Addressing the role of RARRES1 in stem cells, adipocytes, stellate cells and generating a whole knockout RARRES1 mouse model will elucidate the importance of RARRES1 in metabolic reprogramming and metabolism associated diseases.
133 Appendix: Chapter 3 Supplementary Figures
Figure S4.1. Signaling Pathways that Regulate RARRES1 Expression in the NURSA Database. Publicly available transcriptomics studies that show a significant change in RARRES1 expression are identified in this graph. The fold change cutoff is -2 FC or 2 FC. The p-value cutoff was 0.05. The red dots symbolize a decrease in RARRES1 expression and the blue dots signify an increase in expression. The detailed sample dataset information≤ ≥ can be accessed in the NURSA≤ database (free online access).
134 Table S4.1. Transcription Factors that Were Predicted to Bind to RARRES1 Promoter Region in PROMO.
135 136 137 138 139 140 Transcription factors with a matrix dissimilarity of less than 15% were chosen. The exact binding site sequence is displayed for each predicted transcription factor.
141 Table S4.2. Transcription Factors that Were Found Bound to RARRES1 Gene in CHEA (ENRICHR) Database.
The data were extracted from the supporting material of the published studies. Each study had a peak at the RARRES1 promoter region. (265).
142 Bibliography
1. Deberardinis, R. J., and Thompson, C. B. (2012) Cellular metabolism and disease: What
do metabolic outliers teach us? Cell. 148, 1132–1144
2. Suda, K. I. And T. (2014) NIH Public Access. 15, 243–256
3. Burgess, R. J., Agathocleous, M., and Morrison, S. J. (2014) Metabolic regulation of stem
cell function. J. Intern. Med. 276, 12–24
4. Chung, S., Dzeja, P. P., Faustino, R. S., Perez-terzic, C., Behfar, A., and Terzic, A. (2011)
NIH Public Access. 4, 1–12
5. Embryology, M., Group, A., and Group, E. M. (2010) E MBRYONIC S TEM C ELLS / I
NDUCED P LURIPOTENT S TEM C ELLS The Senescence-Related Mitochondrial /
Oxidative Stress Pathway is Repressed in Human Induced Pluripotent Stem Cells.
10.1002/stem.404
6. Birket, M. J., Orr, A. L., Gerencser, A. A., Madden, D. T., Vitelli, C., Swistowski, A.,
Brand, M. D., and Zeng, X. (2011) A reduction in ATP demand and mitochondrial activity
with neural differentiation of human embryonic stem cells. 10.1242/jcs.072272
7. Zhao, H., Dennery, P. A., and Yao, H. (2018) Metabolic reprogramming in the
pathogenesis of chronic lung diseases , including BPD , COPD , and pulmonary fibrosis.
10.1152/ajplung.00521.2017
8. Reprogramming, M., and Disease, L. (2016) HHS Public Access. 15, 1–11
9. Tennant, D. A, Durán, R. V, and Gottlieb, E. (2010) Targeting metabolic transformation
for cancer therapy. Nat. Rev. Cancer. 10, 267–277
10. Mauro, C. (2017) Similarities in the Metabolic Reprogramming of immune System and
endothelium. 10.3389/fimmu.2017.00837
143 11. Pavlova, N. N., and Thompson, C. B. (2016) The Emerging Hallmarks of Cancer
Metabolism. Cell Metab. 23, 27–47
12. De Berardinis, R. J., and Chandel, N. S. (2016) Fundamentals of cancer metabolism. Sci.
Adv. 10.1126/sciadv.1600200
13. Rumseys, L., I, D. F. W., and Matti, E. (1990) Cellular Energetics and the Oxygen
Dependence of Respiration Cardiac Myocytes Isolated from Adult Rat * in
14. Manuscript, A., and Structures, T. (2009) NIH Public Access. 6, 247–253
15. Currie, E., Schulze, A., Zechner, R., Walther, T. C., and Farese, R. V. (2013) Cellular
fatty acid metabolism and cancer. Cell Metab. 18, 153–161
16. Ananieva, E. A., and Wilkinson, A. C. (2018) Branched-chain amino acid metabolism in
cancer. Curr. Opin. Clin. Nutr. Metab. Care. 21, 64–70
17. Masoud, G. N., and Li, W. (2015) HIF-1α pathway: Role, regulation and intervention for
cancer therapy. Acta Pharm. Sin. B. 5, 378–389
18. Puzio-Kuter, A. M. (2011) The role of p53 in metabolic regulation. Genes and Cancer. 2,
385–391
19. Schwartzenberg-bar-yoseph, F., Armoni, M., and Karnieli, E. (2004) The Tumor
Suppressor p53 Down-Regulates Glucose Transporters GLUT1 and GLUT4 Gene
Expression The Tumor Suppressor p53 Down-Regulates Glucose Transporters GLUT1
and GLUT4 Gene Expression. Cancer Res. 10.1158/0008-5472.CAN-03-0846
20. Hu, W., Zhang, C., Wu, R., Sun, Y., Levine, A., and Feng, Z. (2010) Glutaminase 2, a
novel p53 target gene regulating energy metabolism and antioxidant function. Proc. Natl.
Acad. Sci. 107, 7455–7460
21. Zhao, Y., Coloff, J. L., Ferguson, E. C., Jacobs, S. R., Cui, K., and Rathmell, J. C. (2008)
144 Glucose metabolism attenuates p53 and Puma-dependent cell death upon growth factor
deprivation. J. Biol. Chem. 283, 36344–36353
22. Manuscript, A., and Metabolism, C. (2013) c-Myc and Cancer Metabolism. 18, 5546–
5553
23. Chen, H., Liu, H., and Qing, G. (2018) Targeting oncogenic Myc as a strategy for cancer
treatment. Signal Transduct. Target. Ther. 3, 5
24. Camarda, R., Zhou, A. Y., Kohnz, R. A., Balakrishnan, S., Mahieu, C., Anderton, B.,
Eyob, H., Kajimura, S., Tward, A., Krings, G., Nomura, D. K., and Goga, A. (2016)
Inhibition of fatty acid oxidation as a therapy for MYC-overexpressing triple-negative
breast cancer. Nat. Med. 22, 427–432
25. Wang, R., Dillon, C. P., Shi, L. Z., Milasta, S., Carter, R., Finkelstein, D., mccormick, L.
L., Fitzgerald, P., Chi, H., Munger, J., and Green, D. R. (2011) The Transcription Factor
Myc Controls Metabolic Reprogramming upon T Lymphocyte Activation. Immunity. 35,
871–882
26. Hung, C.-L., Wang, L.-Y., Yu, Y.-L., Chen, H.-W., Srivastava, S., Petrovics, G., and
Kung, H.-J. (2014) A long noncoding RNA connects c-Myc to tumor metabolism. Proc.
Natl. Acad. Sci. 111, 18697–18702
27. Catalina-Rodriguez, O., Kolukula, V. K., Tomita, Y., Preet, A., Palmieri, F., Wellstein,
A., Byers, S., Giaccia, A. J., Glasgow, E., Albanese, C., and Avantaggiati, M. L. (2012)
The mitochondrial citrate transporter, CIC, is essential for mitochondrial homeostasis.
Oncotarget. 3, 1220–1235
28. Hayes, J. D., and Ashford, M. L. J. (2012) Nrf2 orchestrates fuel partitioning for cell
proliferation. Cell Metab. 16, 139–141
145 29. Valencia, T., Kim, J. Y., Abu-Baker, S., Moscat-Pardos, J., Ahn, C. S., Reina-Campos,
M., Duran, A., Castilla, E. A., Metallo, C. M., Diaz-Meco, M. T., and Moscat, J. (2014)
Metabolic Reprogramming of Stromal Fibroblasts through p62-mtorc1 Signaling
Promotes Inflammation and Tumorigenesis. Cancer Cell. 26, 121–135
30. Bott, A. J., Peng, I. C., Fan, Y., Faubert, B., Zhao, L., Li, J., Neidler, S., Sun, Y., Jaber,
N., Krokowski, D., Lu, W., Pan, J. A., Powers, S., Rabinowitz, J., Hatzoglou, M., Murphy,
D. J., Jones, R., Wu, S., Girnun, G., and Zong, W. X. (2015) Oncogenic Myc Induces
Expression of Glutamine Synthetase through Promoter Demethylation. Cell Metab. 22,
1068–1077
31. Peng, Z., Shen, R., Li, Y. W., Teng, K. Y., Shapiro, C. L., and Lin, H. J. L. (2012)
Epigenetic Repression of RARRES1 is mediated by methylation of a proximal promoter
and a loss of CTCF binding. Plos One. 7, 1–11
32. Roy, A., Ramalinga, M., Kim, O. J., Chijioke, J., Lynch, S., Byers, S., and Kumar, D.
(2017) Multiple roles of RARRES1 in prostate cancer: Autophagy induction and
angiogenesis inhibition. Plos One. 10.1371/journal.pone.0180344
33. Teufel, A., Becker, D., Weber, S. N., Dooley, S., Breitkopf-Heinlein, K., Maass, T.,
Hochrath, K., Krupp, M., Marquardt, J. U., Kolb, M., Korn, B., Niehrs, C., Zimmermann,
T., Godoy, P., Galle, P. R., and Lammert, F. (2012) Identification of RARRES1 as a core
regulator in liver fibrosis. J. Mol. Med. 90, 1439–1447
34. Montagner, A., Polizzi, A., Fouché, E., Ducheix, S., Lippi, Y., Lasserre, F., Barquissau,
V., Régnier, M., Lukowicz, C., Benhamed, F., Iroz, A., Bertrand-Michel, J., Al Saati, T.,
Cano, P., Mselli-Lakhal, L., Mithieux, G., Rajas, F., Lagarrigue, S., Pineau, T., Loiseau,
N., Postic, C., Langin, D., Wahli, W., and Guillou, H. (2016) Liver pparα is crucial for
146 whole-body fatty acid homeostasis and is protective against NAFLD. Gut. 65, 1202–1214
35. Johansson, L. E., Danielsson, A. P. H., Parikh, H., Klintenberg, M., Norström, F., Groop,
L., and Ridderstråle, M. (2012) Differential gene expression in adipose tissue from obese
human subjects during weight loss and weight maintenance. Am. J. Clin. Nutr. 96, 196–
207
36. Nagpal, S., Patel, S., Asano, A. T., Johnson, A. T., Duvic, M., and Chandraratna, R. A. S.
(1996) Tazarotene-induced gene 1 (TIG1), a novel retinoic acid receptor-responsive gene
in skin. J. Invest. Dermatol. 106, 269–274
37. Aagaard, A., Listwan, P., Cowieson, N., Huber, T., Ravasi, T., Wells, C. A., Flanagan, J.
U., Kellie, S., Hume, D. A., Kobe, B., and Martin, J. L. (2005) An inflammatory role for
the mammalian carboxypeptidase inhibitor latexin: Relationship to cystatins and the tumor
suppressor TIG1. Structure. 13, 309–317
38. Pallarès, I., Bonet, R., García-Castellanos, R., Ventura, S., Avilés, F. X., Vendrell, J., and
Gomis-Rüth, F. X. (2005) Structure of human carboxypeptidase A4 with its endogenous
protein inhibitor, latexin. Proc. Natl. Acad. Sci. U. S. A. 102, 3978–83
39. Krebs, C. J., Jarvis, E. D., Chan, J., Lydon, J. P., Ogawa, S., and Pfaff, D. W. (2000) A
membrane-associated progesterone-binding protein, 25-Dx, is regulated by progesterone
in brain regions involved in female reproductive behaviors. Proc. Natl. Acad. Sci. U. S. A.
97, 12816–12821
40. Normant, E., Martres, M. P., Schwartz, J. C., and Gros, C. (1995) Purification, cdna
cloning, functional expression, and characterization of a 26-kda endogenous mammalian
carboxypeptidase inhibitor. Proc. Natl. Acad. Sci. U. S. A. 92, 12225–9
41. Rehli, M., Krause, S. W., and Andreesen, R. (2000) The membrane-bound ectopeptidase
147 CPM as a marker of macrophage maturation in vitro and in vivo. Adv. Exp. Med. Biol.
477, 205–216
42. Uratani, Y., Takiguchi-Hayashi, K., Miyasaka, N., Sato, M., Jin, M., and Arimatsu, Y.
(2000) Latexin, a carboxypeptidase A inhibitor, is expressed in rat peritoneal mast cells
and is associated with granular structures distinct from secretory granules and lysosomes.
Biochem. J. 346 Pt 3, 817–26
43. Reddanna, P., Prabhu, K. S., Whelan, J., and Reddy, C. C. (2003) Carboxypeptidase A-
catalyzed direct conversion of leukotriene c4to leukotriene F4. Arch. Biochem. Biophys.
413, 158–163
44. Jaffray, C., Mendez, C., Denham, W., Carter, G., and Norman, J. (2000) Specific
Pancreatic Enzymes Activate Macrophages to Produce Tumor Necrosis Factor-Alpha:
Role of Nuclear Factor Kappa B and Inhibitory Kappa B Proteins. J. Gastrointest. Surg. 4,
370–378
45. Sahab, Z. J., Hall, M. D., Sung, Y. M., Dakshanamurthy, S., Ji, Y., Kumar, D., and Byers,
S. W. (2011) Tumor suppressor RARRES1 interacts with cytoplasmic carboxypeptidase
AGBL2 to regulate the α-tubulin tyrosination cycle. Cancer Res. 71, 1219–1228
46. Boeckmann, B., Bairoch, A., Apweiler, R., Blatter, M. C., Estreicher, A., Gasteiger, E.,
Martin, M. J., Michoud, K., O’Donovan, C., Phan, I., Pilbout, S., and Schneider, M.
(2003) The SWISS-PROT protein knowledgebase and its supplement trembl in 2003.
Nucleic Acids Res. 31, 365–370
47. Altschul, S. F., Gish, W., Miller, W., Myers, E. E. W. W., and Lipman, D. J. (1990) Basic
local alignment search tool. J. Mol. Biol. 215, 403–10
48. Tian, X., Gautron, J., Monget, P., and Pascal, G. (2010) What Makes an Egg Unique?
148 Clues from Evolutionary Scenarios of Egg-Specific Genes1. Biol. Reprod. 83, 893–900
49. Wu, C.-C., Tsai, F.-M., Shyu, R.-Y., Tsai, Y.-M., Wang, C.-H., and Jiang, S.-Y. (2011) G
protein-coupled receptor kinase 5 mediates Tazarotene-induced gene 1-induced growth
suppression of human colon cancer cells. BMC Cancer. 11, 175
50. Parker, A. L., Kavallaris, M., and mccarroll, J. A. (2014) Microtubules and their role in
cellular stress in cancer. Front. Oncol. 4, 153
51. Song, Y., and Brady, S. T. (2015) Post-translational modifications of tubulin: Pathways to
functional diversity of microtubules. Trends Cell Biol. 25, 125–136
52. Janke, C., and Kneussel, M. (2010) Tubulin post-translational modifications: Encoding
functions on the neuronal microtubule cytoskeleton. Trends Neurosci. 33, 362–372
53. Westermann, S., and Weber, K. (2003) Post-translational modifications regulate
microtubule function. Nat. Rev. Mol. Cell Biol. 4, 938–947
54. Janke, C., Bulinski, J. C., and Chloë Bulinski, J. (2011) Post-translational regulation of the
microtubule cytoskeleton: mechanisms and functions. Nat. Rev. Mol. Cell Biol. 12, 773–
786
55. Lewis, S. A., Lee, M. G. S., and Cowan, N. J. (1985) Five mouse tubulin isotypes and
their regulated expression during development. J. Cell Biol. 101, 852–861
56. Dráberová, E., Sulimenko, V., Vinopal, S., Sulimenko, T., Sládková, V., D’Agostino, L.,
Sobol, M., Hozák, P., Křen, L., Katsetos, C. D., and Dráber, P. (2017) Differential
expression of human g-tubulin isotypes during neuronal development and oxidative stress
points to a g-tubulin-2 prosurvival function. FASEB J. 31, 1828–1846
57. Niwa, S., Takahashi, H., and Hirokawa, N. (2013) β-Tubulin mutations that cause severe
neuropathies disrupt axonal transport. EMBO J. 32, 1352–1364
149 58. Keays, D. A., Tian, G., Poirier, K., Huang, G. J., Siebold, C., Cleak, J., Oliver, P. L., Fray,
M., Harvey, R. J., Molnár, Z., Piñon, M. C., Dear, N., Valdar, W., Brown, S. D. M.,
Davies, K. E., Rawlins, J. N. P., Cowan, N. J., Nolan, P., Chelly, J., and Flint, J. (2007)
Mutations in α-Tubulin Cause Abnormal Neuronal Migration in Mice and Lissencephaly
in Humans. Cell. 128, 45–57
59. Million, K., Larcher, J., Laoukili, J., Bourguignon, D., Marano, F., and Tournier, F. (1999)
Polyglutamylation and polyglycylation of alpha- and beta-tubulins during in vitro ciliated
cell differentiation of human respiratory epithelial cells. J. Cell Sci. 112 ( Pt 2, 4357–4366
60. Manuscript, A., and Structures, T. (2009) NIH Public Access. 6, 247–253
61. Prota, A. E., Magiera, M. M., Kuijpers, M., Bargsten, K., Frey, D., Wieser, M., Jaussi, R.,
Hoogenraad, C. C., Kammerer, R. A., Janke, C., and Steinmetz, M. O. (2013) Structural
basis of tubulin tyrosination by tubulin tyrosine ligase. J. Cell Biol. 200, 259–270
62. Erck, C., Peris, L., Andrieux, A., Meissirel, C., Gruber, A. D., Vernet, M., Schweitzer, A.,
Saoudi, Y., Pointu, H., Bosc, C., Salin, P. A., Job, D., and Wehland, J. (2005) A vital role
of tubulin-tyrosine-ligase for neuronal organization. Proc. Natl. Acad. Sci. 102, 7853–
7858
63. Paturle-Lafanechère, L., Manier, M., Trigault, N., Pirollet, F., Mazarguil, H., and Job, D.
(1994) Accumulation of delta 2-tubulin, a major tubulin variant that cannot be tyrosinated,
in neuronal tissues and in stable microtubule assemblies. J. Cell Sci. 107 ( Pt 6, 1529–
1543
64. Bornovalova, M. A., and Levy, R. (2011) NIH Public Access. 22, 233–245
65. Peris, L., Thery, M., Fauré, J., Saoudi, Y., Lafanechère, L., Chilton, J. K., Gordon-Weeks,
P., Galjart, N., Bornens, M., Wordeman, L., Wehland, J., Andrieux, A., and Job, D. (2006)
150 Tubulin tyrosination is a major factor affecting the recruitment of CAP-Gly proteins at
microtubule plus ends. J. Cell Biol. 174, 839–849
66. Cambray-Deakin, M. A., and Burgoyne, R. D. (1987) Posttranslational modifications of α-
tubulin: Acetylated and detyrosinated forms in axons of rat cerebellum. J. Cell Biol. 104,
1569–1574
67. Drive, G., and Jolla, L. (2013) NIH Public Access. 23, 716–728
68. Peris, L., Wagenbach, M., Lafanechère, L., Brocard, J., Moore, A. T., Kozielski, F., Job,
D., Wordeman, L., and Andrieux, A. (2009) Motor-dependent microtubule disassembly
driven by tubulin tyrosination. J. Cell Biol. 185, 1159–1166
69. Nieuwenhuis J, Adamopoulos A, Bleijerveld OB, Mazouzi A, Stickel E, Celie P, Altelaar
M, Knipscheer P, Perrakis A, Blomen VA, B. T. (2017) Vasohibins encode tubulin
detyrosinating activity. Sci. Adv. 5676, 1–9
70. Aillaud, C., Bosc, C., Peris, L., Bosson, A., Heemeryck, P., Dijk, J. Van, Friec, J. Le,
Boulan, B., Vossier, F., Sanman, L. E., Syed, S., Amara, N., Couté, Y., Lafanechère, L.,
Denarier, E., Delphin, C., Pelletier, L., Humbert, S., Bogyo, M., and Andrieux, A. (2017)
Vasohibins/SVBP are tubulin carboxypeptidases (tcps) that regulate neuron
differentiation. 1453, 1448–1453
71. `Tort, O., Tanco, S., Rocha, C., Bièche, I., Seixas, C., Bosc, C., Andrieux, A., Moutin, M.-
J., Avilés, F. X., Lorenzo, J., and Janke, C. (2014) `The cytosolic carboxypeptidases CCP2
and CCP3 catalyze posttranslational removal of acidic amino acids. `\ Mol. Biol. Cell. 25,
3017–27
72. Herms, A., Bosch, M., Reddy, B. J. N., Schieber, N. L., Fajardo, A., Rupérez, C.,
Fernández-Vidal, A., Ferguson, C., Rentero, C., Tebar, F., Enrich, C., Parton, R. G.,
151 Gross, S. P., and Pol, A. (2015) AMPK activation promotes lipid droplet dispersion on
detyrosinated microtubules to increase mitochondrial fatty acid oxidation. Nat. Commun.
6, 7176
73. Maldonado, E. N., Sheldon, K. L., Dehart, D. N., Patnaik, J., Manevich, Y., Townsend, D.
M., Bezrukov, S. M., Rostovtseva, T. K., and Lemasters, J. J. (2013) Voltage-dependent
anion channels modulate mitochondrial metabolism in cancer cells: Regulation by free
tubulin and erastin. J. Biol. Chem. 288, 11920–11929
74. Rostovtseva, T. K., Sheldon, K. L., Hassanzadeh, E., Monge, C., Saks, V., Bezrukov, S.
M., and Sackett, D. L. (2008) Tubulin binding blocks mitochondrial voltage-dependent
anion channel and regulates respiration. Proc. Natl. Acad. Sci. U. S. A. 105, 18746–18751
75. Boroughs, L. K., and deberardinis, R. J. (2015) Metabolic pathways promoting cancer cell
survival and growth. Nat. Cell Biol. 17, 351–359
76. Dubey, A. K., Godbole, A., and Mathew, M. K. (2016) Regulation of VDAC trafficking
modulates cell death. Cell Death Discov. 2, 16085
77. Lemasters, J. J., and Holmuhamedov, E. (2006) Voltage-dependent anion channel
(VDAC) as mitochondrial governator - Thinking outside the box. Biochim. Biophys. Acta
- Mol. Basis Dis. 1762, 181–190
78. OUCHI (2014) Structural and Molecular Bases of Mitochondrial Ion Channel Function. In
Cardiac Electrophysiology: From Cell to Bedside, pp. 71–84, Elsevier BV
79. Camara, A. K. S., Zhou, Y. F., Wen, P. C., Tajkhorshid, E., and Kwok, W. M. (2017)
Mitochondrial VDAC1: A key gatekeeper as potential therapeutic target. Front. Physiol.
8, 1–18
80. Anflous-Pharayra, K., Lee, N., Armstrong, D. L., and Craigen, W. J. (2011) VDAC3 has
152 differing mitochondrial functions in two types of striated muscles. Biochim. Biophys. Acta
- Bioenerg. 1807, 150–156
81. Naghdi, Shamim, Hajnoczky, G. (2015) VDAC2-specific cellular functions and the
underlying structure. Biochim. Biophys. Acta - Mol. Cell Biol. Lipids. 25, 368–379
82. Brahimi-Horn, M. C., Giuliano, S., Saland, E., Lacas-Gervais, S., Sheiko, T., Pelletier, J.,
Bourget, I., Bost, F., Féral, C., Boulter, E., Tauc, M., Ivan, M., Garmy-Susini, B., Popa,
A., Mari, B., Sarry, J.-E., Craigen, W. J., Pouysségur, J., and Mazure, N. M. (2015)
Knockout of Vdac1 activates hypoxia-inducible factor through reactive oxygen species
generation and induces tumor growth by promoting metabolic reprogramming and
inflammation. Cancer Metab. 3, 8
83. Anflous, K., Armstrong, D. D., and Craigen, W. J. (2001) Altered mitochondrial
sensitivity for ADP and maintenance of creatine-stimulated respiration in oxidative
striated muscles from VDAC1-deficient mice. J. Biol. Chem. 276, 1954–1960
84. Mccommis, K. S., and Baines, C. P. (2012) The role of VDAC in cell death: Friend or
foe? Biochim. Biophys. Acta - Biomembr. 1818, 1444–1450
85. Shoshan-Barmatz, V., Keinan, N., Abu-Hamad, S., Tyomkin, D., and Aram, L. (2010)
Apoptosis is regulated by the VDAC1 N-terminal region and by VDAC oligomerization:
Release of cytochrome c, AIF and Smac/Diablo. Biochim. Biophys. Acta - Bioenerg. 1797,
1281–1291
86. Cheng, E. H. Y., Sheiko, T. V., Fisher, J. K., Craigen, W. J., and Korsmeyer, S. J. (2003)
VDAC2 inhibits BAK activation and mitochondrial apoptosis. Science (80-. ). 301, 513–
517
87. Arbel, N., and Shoshan-Barmatz, V. (2010) Voltage-dependent anion channel 1-based
153 peptides interact with Bcl-2 to prevent antiapoptotic activity. J. Biol. Chem. 285, 6053–
6062
88. Pastorino, J. G., and Hoek, J. B. (2008) Regulation of hexokinase binding to VDAC. J.
Bioenerg. Biomembr. 40, 171–182
89. Sheldon, K. L., Gurnev, P. A., Bezrukov, S. M., and Sackett, D. L. (2015) Tubulin tail
sequences and post-translational modifications regulate closure of mitochondrial voltage-
dependent anion channel (VDAC). J. Biol. Chem. 290, 26784–26789
90. Gurnev, P. A., Rostovtseva, T. K., and Bezrukov, S. M. (2011) Tubulin-blocked state of
VDAC studied by polymer and ATP partitioning. FEBS Lett. 585, 2363–2366
91. Carré, M., André, N., Carles, G., Borghi, H., Brichese, L., Briand, C., Braguer, D.,
Medicine, U. F. R., and La, U. O. (2002) Tubulin is an inherent component of
mitochondrial membranes that interacts with the voltage-dependent anion channel. J. Biol.
Chem. 277, 33664–33669
92. Noskov, S. Y., Rostovtseva, T. K., and Bezrukov, S. M. (2013) ATP transport through
VDAC and the VDAC-tubulin complex probed by equilibrium and nonequilibrium MD
simulations. Biochemistry. 52, 9246–9256
93. Magrì, A., Reina, S., and De Pinto, V. (2018) VDAC1 as Pharmacological Target in
Cancer and Neurodegeneration: Focus on Its Role in Apoptosis. Front. Chem. 6, 108
94. Oldridge, E. E., Walker, H. F., Stower, M. J., Simms, M. S., Mann, V. M., Collins, A. T.,
Pellacani, D., and Maitland, N. J. (2013) Retinoic acid represses invasion and stem cell
phenotype by induction of the metastasis suppressors RARRES1 and LXN. Oncogenesis.
2, e45
95. Ullah, M., Stich, S., Häupl, T., Eucker, J., Sittinger, M., and Ringe, J. (2013) Reverse
154 Differentiation as a Gene Filtering Tool in Genome Expression Profiling of Adipogenesis
for Fat Marker Gene Selection and Their Analysis. Plos One.
10.1371/journal.pone.0069754
96. Takeda, K., Sriram, S., Chan, X. H. D., Ong, W. K., Yeo, C. R., Tan, B., Lee, S. A., Kong,
K. V., Hoon, S., Jiang, H., Yuen, J. J., Perumal, J., Agrawal, M., Vaz, C., So, J., Shabbir,
A., Blaner, W. S., Olivo, M., Han, W., Tanavde, V., Toh, S. A., and Sugii, S. (2016)
Retinoic acid mediates visceral-specific adipogenic defects of human adipose-derived
stem cells. Diabetes. 65, 1164–1178
97. Shyu, R., Wang, C., Wu, C., Chen, M., Lee, M., Jiang, S., and Tsai, F. (2016) Tazarotene-
Induced Gene 1 Enhanced Cervical Cell Autophagy through Transmembrane Protein 192.
Mol. Cells. 39, 877–887
98. Wang, L. L., Jin, X. H., Cai, M. Y., Li, H. G., Chen, J. W., Wang, F. W., Wang, C. Y.,
Hu, W. W., Liu, F., and Xie, D. (2018) AGBL2 promotes cancer cell growth through
IRGM-regulated autophagy and enhanced Aurora A activity in hepatocellular carcinoma.
Cancer Lett. 414, 71–80
99. Huyghe, J. R., Jackson, A. U., Fogarty, M. P., Buchkovich, M. L., Stančáková, A.,
Stringham, H. M., Sim, X., Yang, L., Fuchsberger, C., Cederberg, H., Chines, P. S.,
Teslovich, T. M., Romm, J. M., Ling, H., mcmullen, I., Ingersoll, R., Pugh, E. W.,
Doheny, K. F., Neale, B. M., Daly, M. J., Kuusisto, J., Scott, L. J., Kang, H. M., Collins,
F. S., Abecasis, G. R., Watanabe, R. M., Boehnke, M., Laakso, M., and Mohlke, K. L.
(2013) Exome array analysis identifies new loci and low-frequency variants influencing
insulin processing and secretion. Nat. Genet. 45, 197–201
100. Kloth, M., Goering, W., Ribarska, T., Arsov, C., Sorensen, K. D., and Schulz, W. A.
15 5 (2012) The SNP rs6441224 influences transcriptional activity and prognostically relevant
hypermethylation of RARRES1 in prostate cancer. Int. J. Cancer. 10.1002/ijc.27628
101. Yanatatsaneejit, P., Chalermchai, T., Kerekhanjanarong, V., Shotelersuk, K., Supiyaphun,
P., Mutirangura, A., and Sriuranpong, V. (2008) Promoter hypermethylation of CCNA1,
RARRES1, and HRASLS3 in nasopharyngeal carcinoma. Oral Oncol. 44, 400–406
102. Coyle, K. M., Murphy, J. P., Vidovic, D., Vaghar-Kashani, A., Dean, C. A., Sultan, M.,
Clements, D., Wallace, M., Thomas, M. L., Hundert, A., Giacomantonio, C. A., Helyer,
L., Gujar, S. A., Lee, P. W. K., Weaver, I. C. G., and Marcato, P. (2016) Breast cancer
subtype dictates DNA methylation and ALDH1A3-mediated expression of tumor
suppressor RARRES1. Oncotarget. 10.18632/oncotarget.9858
103. Bonazzi, V. F., Irwin, D., and Hayward, N. K. (2009) Identification of candidate tumor
suppressor genes inactivated by promoter methylation in melanoma. Genes. Chromosomes
Cancer. 48, 10–21
104. Youssef, E. M., Chen, X. Q., Higuchi, E., Kondo, Y., Garcia-Manero, G., Lotan, R., and
Issa, J. P. (2004) Hypermethylation and silencing of the putative tumor suppressor
Tazarotene-induced gene 1 in human cancers. Cancer Res. 64, 2411–2417
105. Jing, C., El-Ghany, M. A., Beesley, C., Foster, C. S., Rudland, P. S., Smith, P., and Ke, Y.
(2002) Tazarotene-induced gene 1 (TIG1) expression in prostate carcinomas and its
relationship to tumorigenicity. J. Natl. Cancer Inst. 94, 482–90
106. Wu, C. C., Shyu, R. Y., Chou, J. M., Jao, S. W., Chao, P. C., Kang, J. C., Wu, S. T.,
Huang, S. L., and Jiang, S. Y. (2006) RARRES1 expression is significantly related to
tumour differentiation and staging in colorectal adenocarcinoma. Eur. J. Cancer. 42, 557–
565
15 6 107. Nakamura, T., Kuwai, T., Kitadai, Y., Sasaki, T., Fan, D., Coombes, K. R., Kim, S. J., and
Fidler, I. J. (2007) Zonal heterogeneity for gene expression in human pancreatic
carcinoma. Cancer Res. 67, 7597–7604
108. Chen, X. H., Wu, W. G., and Ding, J. (2014) Aberrant TIG1 methylation associated with
its decreased expression and clinicopathological significance in hepatocellular carcinoma.
Tumor Biol. 35, 967–971
109. Sakurai, T., and Kudo, M. (2013) Molecular Link between Liver Fibrosis and
Hepatocellular Carcinoma. Liver Cancer. 2, 365–366
110. Hernandez-Gea, V., and Friedman, S. L. (2011) Pathogenesis of Liver Fibrosis. Annu.
Rev. Pathol. Mech. Dis. 6, 425–456
111. Affo, S., Yu, L.-X., and Schwabe, R. F. (2017) The Role of Cancer-Associated Fibroblasts
and Fibrosisin Liver Cancer. Annu. Rev. Pathol. Mech. Dis. 12, 153–186
112. Vennin, C., Murphy, K. J., Morton, J. P., Cox, T. R., Pajic, M., and Timpson, P. (2018)
Reshaping the Tumor Stroma for Treatment of Pancreatic Cancer. Gastroenterology. 154,
820–838
113. Bhola, N., Balko, J., Duggar, T., and et al (2013) TGF- [ Beta ] Inhibition enhances
chemotherapy action against triple-negative breast cancer. J Clin Investig. 123, 1348–
1358
114. Zhu H, Zheng Z, Zhang J, Liu X, Liu Y, Yang W, Liu Y, Zhang T, Zhao Y, Liu Y, Su X,
G. X. (2015) Effects of AGBL2 on cell proliferation and chemotherapy resistance of
gastric cancer. Hepatogastroenterology. 62, 497–502
115. Wood, R. J., Tchack, L., Angelo, G., Pratt, R. E., and Sonna, L. A. (2004) DNA
microarray analysis of vitamin D-induced gene expression in a human colon carcinoma
15 7 cell line. Physiol. Genomics. 17, 122–129
116. Cairns, R., Harris, I., and Mak, T. (2011) Regulation of cancer cell metabolism. Nat. Rev.
Cancer. 11, 85–95
117. Sancho, P., Burgos-Ramos, E., Tavera, A., Bou Kheir, T., Jagust, P., Schoenhals, M.,
Barneda, D., Sellers, K., Campos-Olivas, R., Graña, O., Viera, C. R., Yuneva, M., Sainz,
B., and Heeschen, C. (2015) MYC/PGC-1α balance determines the metabolic phenotype
and plasticity of pancreatic cancer stem cells. Cell Metab. 22, 590–605
118. Peiris-Pagès, M., Martinez-Outschoorn, U. E., Pestell, R. G., Sotgia, F., and Lisanti, M. P.
(2016) Cancer stem cell metabolism. Breast Cancer Res. 10.1186/s13058-016-0712-6
119. Wheaton, W. W., Weinberg, S. E., Hamanaka, R. B., Soberanes, S., Sullivan, L. B., Anso,
E., Glasauer, A., Dufour, E., Mutlu, G. M., Scott Budigner, G. R., and Chandel, N. S.
(2014) Metformin inhibits mitochondrial complex I of cancer cells to reduce
tumorigenesis. Elife. 2014, 1–18
120. Shaw, R. J., Kosmatka, M., Bardeesy, N., Hurley, R. L., Witters, L. A., depinho, R. A.,
and Cantley, L. C. (2004) The tumor suppressor LKB1 kinase directly activates AMP-
activated kinase and regulates apoptosis in response to energy stress. Proc. Natl. Acad.
Sci. 101, 3329–3335
121. Liang, Y., Jansen, M., Aronow, B., Geiger, H., and Van Zant, G. (2007) The quantitative
trait gene latexin influences the size of the hematopoietic stem cell population in mice.
Nat. Genet. 39, 178–88
122. Rogowski, K., van Dijk, J., Magiera, M. M., Bosc, C., Deloulme, J. C., Bosson, A., Peris,
L., Gold, N. D., Lacroix, B., Grau, M. B., Bec, N., Larroque, C., Desagher, S., Holzer, M.,
Andrieux, A., Moutin, M. J., and Janke, C. (2010) A family of protein-deglutamylating
15 8 enzymes associated with neurodegeneration. Cell. 143, 564–578
123. Lafanechère, L., Courtay-Cahen, C., Kawakami, T., Jacrot, M., Rüdiger, M., Wehland, J.,
Job, D., and Margolis, R. L. (1998) Suppression of tubulin tyrosine ligase during tumor
growth. J. Cell Sci. 111, 171–81
124. Yu, I., Garnham, C. P., and Roll-Mecak, A. (2015) Writing and reading the tubulin code.
J. Biol. Chem. 290, 17163–17172
125. Akimenko, M. A, Johnson, S. L., Westerfield, M., and Ekker, M. (1995) Differential
induction of four msx homeobox genes during fin development and regeneration in
zebrafish. Development. 121, 347–357
126. Kimmel, C. B., Ballard, W. W., Kimmel, S. R., Ullmann, B., and Schilling, T. F. (1995)
Stages of embryonic development of the zebrafish. Dev. Dyn. An Off. Public. 203, 253–
310
127. Furutani-Seiki, M., Jiang, Y. J., Brand, M., Heisenberg, C. P., Houart, C., Beuchle, D.,
van Eeden, F. J., Granato, M., Haffter, P., Hammerschmidt, M., Kane, D. A., Kelsh, R. N.,
Mullins, M. C., Odenthal, J., and Nüsslein-Volhard, C. (1996) Neural degeneration
mutants in the zebrafish, Danio rerio. Development. 123, 229–239
128. Eaton, J. L., and Glasgow, E. (2007) Zebrafish orthopedia (otp) is required for isotocin
cell development. Dev. Genes Evol. 217, 149–158
129. Harris, A., Morgan, J. I., Pecot, M., Soumare, A., Osborne, A., and Soares, H. D. (2000)
Regenerating motor neurons express Nna1, a novel ATP/GTP-binding protein related to
zinc carboxypeptidases. Mol. Cell. Neurosci. 16, 578–96
130. Chakraborti, S., Natarajan, K., Curiel, J., Janke, C., and Liu, J. (2016) The emerging role
of the tubulin code: From the tubulin molecule to neuronal function and disease.
1 59 Cytoskeleton. 73, 521–550
131. Kalinina, E., Biswas, R., Berezniuk, I., Hermoso, A., Aviles, F. X., and Fricker, L. D.
(2007) A novel subfamily of mouse cytosolic carboxypeptidases. FASEB J. 21, 836–850
132. Valenstein, M. L., and Roll-Mecak, A. (2016) Graded Control of Microtubule Severing by
Tubulin Glutamylation. Cell. 164, 911–921
133. Bailey, M. E., Sackett, D. L., and Ross, J. L. (2015) Katanin Severing and Binding
Microtubules Are Inhibited by Tubulin Carboxy Tails. Biophys. J. 109, 2546–2561
134. Wolff, A., de Néchaud, B., Chillet, D., Mazarguil, H., Desbruyères, E., Audebert, S.,
Eddé, B., Gros, F., and Denoulet, P. (1992) Distribution of glutamylated alpha and beta-
tubulin in mouse tissues using a specific monoclonal antibody, GT335. Eur. J. Cell Biol.
59, 425—432
135. Harrow, J., Denoeud, F., Frankish, A., Reymond, A., Chen, C.-K., Chrast, J., Lagarde, J.,
Gilbert, J. G., Storey, R., Swarbreck, D., Rossier, C., Ucla, C., Hubbard, T., Antonarakis,
S. E., Guigo, R., Mattick, J., Bartel, D., Guigo, R., Flicek, P., Abril, J., Reymond, A.,
Lagarde, J., Denoeud, F., Antonarakis, S., Ashburner, M., Bajic, V., Birney, E., Deloukas,
P., Matthews, L., Ashurst, J., Burton, J., Gilbert, J., Jones, M., Stavrides, G., Almeida, J.,
Babbage, A., Bagguley, C., Will, C., Luhrmann, R., Parra, G., Blanco, E., Guigo, R.,
Burge, C., Karlin, S., Wang, M., Buhler, J., Brent, M., Wiehe, T., Gebauer-Jung, S.,
Mitchell-Olds, T., Guigo, R., Salamov, A., Solovyev, V., Eyras, E., Caccamo, M.,
Curwen, V., Clamp, M., Birney, E., Andrews, T., Bevan, P., Caccamo, M., Chen, Y.,
Clarke, L., Coates, G., Cuff, J., Curwen, V., Cutts, T., Kozak, M., Lewis, B., Green, R.,
Brenner, S., Ohler, U., Shomron, N., Burge, C., Kapranov, P., Drenkow, J., Cheng, J.,
Long, J., Helt, G., Dike, S., Gingeras, T., Shiraki, T., Kondo, S., Katayama, S., Waki, K.,
160 Kasukawa, T., Kawaji, H., Kodzius, R., Watahiki, A., Nakamura, M., Arakawa, T., Ng, P.,
Wei, C., Sung, W., Chiu, K., Lipovich, L., Ang, C., Gupta, S., Shahab, A., Ridwan, A.,
Wong, C., Potter, S., Clarke, L., Curwen, V., Keenan, S., Mongin, E., Searle, S., Stabenau,
A., Storey, R., Clamp, M., Rice, P., Longden, I., Bleasby, A., Benson, G., Mott, R.,
Birney, E., Clamp, M., Durbin, R., Lowe, T., Eddy, S., Down, T., Hubbard, T., Searle, S.,
Gilbert, J., Iyer, V., Clamp, M., Sonnhammer, E., Wootton, J., Reymond, A., Friedli, M.,
Henrichsen, C., Chapot, F., Deutsch, S., Ucla, C., Rossier, C., Lyle, R., Guipponi, M.,
Antonarakis, S., Reymond, A., Camargo, A., Deutsch, S., Stevenson, B., Parmigiani, R.,
Ucla, C., Bettoni, F., Rossier, C., Lyle, R., Guipponi, M., Guigo, R., Dermitzakis, E.,
Agarwal, P., Ponting, C., Parra, G., Reymond, A., Abril, J., Keibler, E., Lyle, R., and
Ucla, C. (2006) GENCODE: producing a reference annotation for ENCODE. Genome
Biol. 7, S4
136. Mclaren, W., Gil, L., Hunt, S. E., Riat, H. S., Ritchie, G. R. S., Thormann, A., Flicek, P.,
and Cunningham, F. (2016) The Ensembl Variant Effect Predictor. Biorxiv.
10.1101/042374
137. Tang, X.-H., and Gudas, L. J. (2011) Retinoids, retinoic acid receptors, and cancer. Annu.
Rev. Pathol. 6, 345–64
138. Orford, K., Orford, C. C., and Byers, S. W. (1999) Exogenous expression of ??-catenin
regulates contact inhibition, anchorage-independent growth, anoikis, and radiation-
induced cell cycle arrest. J. Cell Biol. 146, 855–867
139. Robu, M. E., Larson, J. D., Nasevicius, A., Beiraghi, S., Brenner, C., Farber, S. A., and
Ekker, S. C. (2007) P53 Activation By Knockdown Technologies. Plos Genet. 3, 787–801
140. Lyons, P. J., Sapio, M. R., and Fricker, L. D. (2013) Zebrafish cytosolic
161 carboxypeptidases 1 and 5 are essential for embryonic development. J. Biol. Chem. 288,
30454–30462
141. Farris, J. C., Pifer, P. M., Zheng, L., Gottlieb, E., Denvir, J., and Frisch, S. M. (2016)
Grainyhead-like 2 Reverses the Metabolic Changes Induced by the Oncogenic Epithelial-
Mesenchymal Transition: Effects on Anoikis. Mol. Cancer Res. 14, 528–538
142. Frisch, S., and Francis, H. (1994) Disruption of epithelial cell matrix interactions induces
apoptosis. J. Cell Biol. 124, 619–626
143. Frisch, S. M., and Screaton, R. A. (2001) Anoikis mechanisms. Curr. Opin. Cell Biol. 13,
555–562
144. Yagoda, N., von Rechenberg, M., Zaganjor, E., Bauer, A. J., Yang, W. S., Fridman, D. J.,
Wolpaw, A. J., Smukste, I., Peltier, J. M., Boniface, J. J., Smith, R., Lessnick, S. L.,
Sahasrabudhe, S., and Stockwell, B. R. (2007) RAS–RAF–MEK-dependent oxidative cell
death involving voltage-dependent anion channels. Nature. 447, 865–869
145. Head, S. A., Shi, W., Zhao, L., Gorshkov, K., Pasunooti, K., Chen, Y., Deng, Z., Li, R.,
Shim, J. S., Tan, W., Hartung, T., Zhang, J., Zhao, Y., Colombini, M., and Liu, J. O.
(2015) Antifungal drug itraconazole targets VDAC1 to modulate the AMPK/mtor
signaling axis in endothelial cells. Proc. Natl. Acad. Sci. 112, E7276–E7285
146. Viticchiè, G., Agostini, M., Lena, A. M., Mancini, M., Zhou, H., Zolla, L., Dinsdale, D.,
Saintigny, G., Melino, G., and Candi, E. (2015) p63 supports aerobic respiration through
hexokinase II. Proc. Natl. Acad. Sci. 112, 11577–11582
147. Mihaylova, M. M., and Shaw, R. J. (2011) The AMPK signalling pathway coordinates cell
growth, autophagy and metabolism. Nat. Cell Biol. 13, 1016–1023
148. Komlodi-Pasztor, E., Sackett, D., Wilkerson, J., and Fojo, T. (2011) Mitosis is not a key
162 target of microtubule agents in patient tumors. Nat. Rev. Clin. Oncol. 8, 244
149. Liberti, M. V., and Locasale, J. W. (2016) The Warburg Effect: How Does it Benefit
Cancer Cells? Trends Biochem. Sci. 41, 211–218
150. Farnie, G., Sotgia, F., Lisanti, M. P., and Gillian Farnie, F. S. M. P. L. (2015) High
mitochondrial mass identifies a sub-population of stem-like cancer cells that are chemo-
resistant. Oncotarget. 6, 30472–30486
151. Coloff, J. L., Murphy, J. P., Braun, C. R., Harris, I. S., Shelton, L. M., Kami, K., Gygi, S.
P., Selfors, L. M., and Brugge, J. S. (2016) Differential Glutamate Metabolism in
Proliferating and Quiescent Mammary Epithelial Cells. Cell Metab. 23, 867–880
152. Abshagen, K., König, M., Hoppe, A., Müller, I., Ebert, M., Weng, H., holzhã¼tter, H.
G., Zanger, U. M., Bode, J., Vollmar, B., Thomas, M., and Dooley, S. (2015)
Pathobiochemical signatures of cholestatic liver disease in bile duct ligated mice. BMC
Syst. Biol. 10.1186/s12918-015-0229-0
153. Du, X., Matsumura, T., Edelstein, D., Rossetti, L., and Brownlee, M. (2003) Inhibition of
GAPDH activity. J. Clin. Invest. 112, 1049–1057
154. Wilson, J. E. (2003) Isozymes of mammalian hexokinase: structure, subcellular
localization and metabolic function. J. Exp. Biol. 206, 2049–2057
155. De La Vega Otazo, M. R., Lorenzo, J., Tort, O., Avilés, F. X., and Bautista, J. M. (2013)
Functional segregation and emerging role of cilia-related cytosolic carboxypeptidases
(ccps). FASEB J. 27, 424–431
156. Dakshanamurthy, S., Issa, N. T., Assefnia, S., Seshasayee, A., Peters, O. J., Madhavan, S.,
Uren, A., Brown, M. L., and Byers, S. W. (2012) Predicting new indications for approved
drugs using a proteochemometric method. J. Med. Chem. 55, 6832–6848
163 157. Wu, L., Zhou, B., Oshiro-Rapley, N., Li, M., Paulo, J. A., Webster, C. M., Mou, F.,
Kacergis, M. C., Talkowski, M. E., Carr, C. E., Gygi, S. P., Zheng, B., and Soukas, A. A.
(2016) An Ancient, Unified Mechanism for Metformin Growth Inhibition in C. Elegans
and Cancer. Cell. This issue, 1705–1711.e13
158. Ohnishi, S., Okabe, K., Obata, H., Otani, K., Ishikane, S., Ogino, H., Kitamura, S., and
Nagaya, N. (2009) Involvement of tazarotene-induced gene 1 in proliferation and
differentiation of human adipose tissue-derived mesenchymal stem cells. Cell Prolif. 42,
309–316
159. Montagner, A., Polizzi, A., Fouché, E., Ducheix, S., Lippi, Y., Lasserre, F., Barquissau,
V., Régnier, M., Lukowicz, C., Benhamed, F., Iroz, A., Bertrand-Michel, J., Saati, T. Al,
Cano, P., Mselli-Lakhal, L., Mithieux, G., Rajas, F., Lagarrigue, S., Pineau, T., Loiseau,
N., Postic, C., Langin, D., Wahli, W., and Guillou, H. (2016) Liver pparα is crucial for
whole-body fatty acid homeostasis and is protective against NAFLD. Gut. 65, 1202–1214
160. So, J. S., Hur, K. Y., Tarrio, M., Ruda, V., Frank-Kamenetsky, M., Fitzgerald, K.,
Koteliansky, V., Lichtman, A. H., Iwawaki, T., Glimcher, L. H., and Lee, A. H. (2012)
Silencing of lipid metabolism genes through ire1α-mediated Mrna decay lowers plasma
lipids in mice. Cell Metab. 16, 487–499
161. Gonzalez, M., Sealls, W., Jesch, E. D., Brosnan, M. J., Ladunga, I., Ding, X., Black, P. N.,
and dirusso, C. C. (2011) Defining a relationship between dietary fatty acids and the
cytochrome P450 system in a mouse model of fatty liver disease. Physiol. Genomics. 43,
121–135
162. Saheki, T., Iijima, M., Meng, X. L., Kobayashi, K., Horiuchi, M., Ushikai, M., Okumura,
F., Xiao, J. M., Inoue, I., Tajima, A., Moriyama, M., Eto, K., Kadowaki, T., Sinasac, D.
164 S., Tsui, L. C., Tsuji, M., Okano, A., and Kobayashi, T. (2007) Citrin/mitochondrial
glycerol-3-phosphate dehydrogenase double knock-out mice recapitulate features of
human citrin deficiency. J. Biol. Chem. 282, 25041–25052
163. Kosters, A., Frijters, R. J. J. M., Kunne, C., Vink, E., Schneiders, M. S., Schaap, F. G.,
Nibbering, C. P., Patel, S. B., and Groen, A. K. (2005) Diosgenin-induced biliary
cholesterol secretion in mice requires Abcg8. Hepatology. 41, 141–150
164. Hollmén, M., Roudnicky, F., Karaman, S., and Detmar, M. (2015) Characterization of
macrophage--cancer cell crosstalk in estrogen receptor positive and triple-negative breast
cancer. Sci. Rep. 5, 9188
165. Viale, A., Corti, D., and Draetta, G. F. (2015) Tumors and mitochondrial respiration: a
neglected connection. Cancer Res. 75, 3685–3686
166. Menendez, J. A., and Lupu, R. (2007) Fatty acid synthase and the lipogenic phenotype in
cancer pathogenesis. Nat. Rev. Cancer. 7, 763–777
167. Currie, E., Schulze, A., Zechner, R., Walther, T. C., and Farese, R. V. (2013) Cellular
fatty acid metabolism and cancer. Cell Metab. 18, 153–161
168. Moffat, J., Grueneberg, D. A., Yang, X., Kim, S. Y., Kloepfer, A. M., Hinkle, G., Piqani,
B., Eisenhaure, T. M., Luo, B., Grenier, J. K., Carpenter, A. E., Foo, S. Y., Stewart, S. A.,
Stockwell, B. R., Hacohen, N., Hahn, W. C., Lander, E. S., Sabatini, D. M., and Root, D.
E. (2006) A Lentiviral rnai Library for Human and Mouse Genes Applied to an Arrayed
Viral High-Content Screen. Cell. 124, 1283–1298
169. Coia, H., Ma, N., He, A. R., Kallakury, B., Berry, D. L., Permaul, E., Kepher, H., Fu, Y.,
and Chung, F. (2017) Detection of a lipid peroxidation-induced DNA adduct across liver
disease stages. 10.21037/hbsn.2017.06.01
165 170. Cui, Q., Lewis, I. A., Hegeman, A. D., Anderson, M. E., Li, J., Schulte, C. F., Westler, W.
M., Eghbalnia, H. R., Sussman, M. R., and Markley, J. L. (2008) Metabolite identification
via the Madison Metabolomics Consortium Database [3]. Nat. Biotechnol. 26, 162–164
171. Wishart, D. S., Jewison, T., Guo, A. C., Wilson, M., Knox, C., Liu, Y., Djoumbou, Y.,
Mandal, R., Aziat, F., Dong, E., Bouatra, S., Sinelnikov, I., Arndt, D., Xia, J., Liu, P.,
Yallou, F., Bjorndahl, T., Perez-Pineiro, R., Eisner, R., Allen, F., Neveu, V., Greiner, R.,
and Scalbert, A. (2013) HMDB 3.0-The Human Metabolome Database in 2013. Nucleic
Acids Res. 10.1093/nar/gks1065
172. Fahy, E., Subramaniam, S., Murphy, R. C., Nishijima, M., Raetz, C. R. H., Shimizu, T.,
Spener, F., van Meer, G., Wakelam, M. J. O., and Dennis, E. A. (2009) Update of the
LIPID MAPS comprehensive classification system for lipids. J. Lipid Res. 50, S9–S14
173. Gueuvoghlanian-Silva, B. Y., Cordeiro, F. B., Lobo, T. F., Cataldi, T. R., Lo Turco, E. G.,
Bertolla, R. P., Mattar, R., Torloni, M. R., and Daher, S. (2015) Lipid fingerprinting in
mild versus severe forms of Gestational Diabetes Mellitus. Plos One. 10, 1–13
174. Chughtai, S., Chughtai, K., Cillero-Pastor, B., Kiss, A., Agrawal, P., macaleese, L., and
Heeren, R. M. A. (2012) A multimodal mass spectrometry imaging approach for the study
of musculoskeletal tissues. Int. J. Mass Spectrom. 325–327, 150–160
175. Chansela, P., Goto-Inoue, N., Zaima, N., Hayasaka, T., Sroyraya, M., Kornthong, N.,
Engsusophon, A., Tamtin, M., Chaisri, C., Sobhon, P., and Setou, M. (2012) Composition
and localization of lipids in penaeus merguiensis ovaries during the ovarian maturation
cycle as revealed by imaging mass spectrometry. Plos One. 7, 1–11
176. Ghosh, S. P., Singh, R., Chakraborty, K., Kulkarni, S., Uppal, A., Luo, Y., Kaur, P.,
Pathak, R., Kumar, K. S., Hauer-Jensen, M., and Cheema, A. K. (2013) Metabolomic
166 changes in gastrointestinal tissues after whole body radiation in a murine model. Mol.
Biosyst. 9, 723
177. Xia, J., and Wishart, D. S. (2016) Using metaboanalyst 3.0 for comprehensive
metabolomics data analysis. Curr. Protoc. Bioinforma. 2016, 14.10.1-14.10.91
178. Grasso, C. S., Wu, Y. M., Robinson, D. R., Cao, X., Dhanasekaran, S. M., Khan, A. P.,
Quist, M. J., Jing, X., Lonigro, R. J., Brenner, J. C., Asangani, I. A., Ateeq, B., Chun, S.
Y., Siddiqui, J., Sam, L., Anstett, M., Mehra, R., Prensner, J. R., Palanisamy, N., Ryslik,
G. A., Vandin, F., Raphael, B. J., Kunju, L. P., Rhodes, D. R., Pienta, K. J., Chinnaiyan,
A. M., and Tomlins, S. A. (2012) The mutational landscape of lethal castration-resistant
prostate cancer. Nature. 487, 239–243
179. Koboldt, D. C., Fulton, R. S., mclellan, M. D., Schmidt, H., Kalicki-Veizer, J., mcmichael,
J. F., Fulton, L. L., Dooling, D. J., Ding, L., Mardis, E. R., Wilson, R. K., Ally, A.,
Balasundaram, M., Butterfield, Y. S. N., Carlsen, R., Carter, C., Chu, A., Chuah, E.,
Chun, H. J. E., Coope, R. J. N., Dhalla, N., Guin, R., Hirst, C., Hirst, M., Holt, R. A., Lee,
D., Li, H. I., Mayo, M., Moore, R. A., Mungall, A. J., Pleasance, E., Robertson, A. G.,
Schein, J. E., Shafiei, A., Sipahimalani, P., Slobodan, J. R., Stoll, D., Tam, A., Thiessen,
N., Varhol, R. J., Wye, N., Zeng, T., Zhao, Y., Birol, I., Jones, S. J. M., Marra, M. A.,
Cherniack, A. D., Saksena, G., Onofrio, R. C., Pho, N. H., Carter, S. L., Schumacher, S.
E., Tabak, B., Hernandez, B., Gentry, J., Nguyen, H., Crenshaw, A., Ardlie, K.,
Beroukhim, R., Winckler, W., Getz, G., Gabriel, S. B., Meyerson, M., Chin, L.,
Kucherlapati, R., Hoadley, K. A., Auman, J. T., Fan, C., Turman, Y. J., Shi, Y., Li, L.,
Topal, M. D., He, X., Chao, H. H., Prat, A., Silva, G. O., Iglesia, M. D., Zhao, W., Usary,
J., Berg, J. S., Adams, M., Booker, J., Wu, J., Gulabani, A., Bodenheimer, T., Hoyle, A.
167 P., Simons, J. V., Soloway, M. G., Mose, L. E., Jefferys, S. R., Balu, S., Parker, J. S.,
Hayes, D. N., Perou, C. M., Malik, S., Mahurkar, S., Shen, H., Weisenberger, D. J.,
Triche, T., Lai, P. H., Bootwalla, M. S., Maglinte, D. T., Berman, B. P., Van Den Berg, D.
J., Baylin, S. B., Laird, P. W., Creighton, C. J., Donehower, L. A., Noble, M., Voet, D.,
Gehlenborg, N., Di Cara, D., Zhang, J., Zhang, H., Wu, C. J., Yingchun Liu, S.,
Lawrence, M. S., Zou, L., Sivachenko, A., Lin, P., Stojanov, P., Jing, R., Cho, J., Sinha,
R., Park, R. W., Nazaire, M. D., Robinson, J., Thorvaldsdottir, H., Mesirov, J., Park, P. J.,
Reynolds, S., Kreisberg, R. B., Bernard, B., Bressler, R., Erkkila, T., Lin, J., Thorsson, V.,
Zhang, W., Shmulevich, I., Ciriello, G., Weinhold, N., Schultz, N., Gao, J., Cerami, E.,
Gross, B., Jacobsen, A., Sinha, R., Aksoy, B. A., Antipin, Y., Reva, B., Shen, R., Taylor,
B. S., Ladanyi, M., Sander, C., Anur, P., Spellman, P. T., Lu, Y., Liu, W., Verhaak, R. R.
G., Mills, G. B., Akbani, R., Zhang, N., Broom, B. M., Casasent, T. D., Wakefield, C.,
Unruh, A. K., Baggerly, K., Coombes, K., Weinstein, J. N., Haussler, D., Benz, C. C.,
Stuart, J. M., Benz, S. C., Zhu, J., Szeto, C. C., Scott, G. K., Yau, C., Paull, E. O., Carlin,
D., Wong, C., Sokolov, A., Thusberg, J., Mooney, S., Ng, S., Goldstein, T. C., Ellrott, K.,
Grifford, M., Wilks, C., Ma, S., Craft, B., Yan, C., Hu, Y., Meerzaman, D., Gastier-Foster,
J. M., Bowen, J., Ramirez, N. C., Black, A. D., Pyatt, R. E., White, P., Zmuda, E. J., Frick,
J., Lichtenberg, T. M., Brookens, R., George, M. M., Gerken, M. A., Harper, H. A.,
Leraas, K. M., Wise, L. J., Tabler, T. R., mcallister, C., Barr, T., Hart-Kothari, M., Tarvin,
K., Saller, C., Sandusky, G., Mitchell, C., Iacocca, M. V., Brown, J., Rabeno, B.,
Czerwinski, C., Petrelli, N., Dolzhansky, O., Abramov, M., Voronina, O., Potapova, O.,
Marks, J. R., Suchorska, W. M., Murawa, D., Kycler, W., Ibbs, M., Korski, K., Spychała,
A., Murawa, P., Brzeziński, J. J., Perz, H., Łaźniak, R., Teresiak, M., Tatka, H.,
168 Leporowska, E., Bogusz-Czerniewicz, M., Malicki, J., Mackiewicz, A., Wiznerowicz, M.,
Van Le, X., Kohl, B., Viet Tien, N., Thorp, R., Van Bang, N., Sussman, H., Phu, B. D.,
Hajek, R., Hung, N. P., Phuong, T. V. T., Thang, H. Q., Khan, K. Z., Penny, R., Mallery,
D., Curley, E., Shelton, C., Yena, P., Ingle, J. N., Couch, F. J., Lingle, W. L., King, T. A.,
Gonzalez-Angulo, A. M., Dyer, M. D., Liu, S., Meng, X., Patangan, M., Waldman, F.,
Stöppler, H., Rathmell, W. K., Thorne, L., Huang, M., Boice, L., Hill, A., Morrison, C.,
Gaudioso, C., Bshara, W., Daily, K., Egea, S. C., Pegram, M. D., Gomez-Fernandez, C.,
Dhir, R., Bhargava, R., Brufsky, A., Shriver, C. D., Hooke, J. A., Campbell, J. L., Mural,
R. J., Hu, H., Somiari, S., Larson, C., Deyarmin, B., Kvecher, L., Kovatich, A. J., Ellis, M.
J., Stricker, T., White, K., Olopade, O., Luo, C., Chen, Y., Bose, R., Chang, L. W., Beck,
A. H., Pihl, T., Jensen, M., Sfeir, R., Kahn, A., Chu, A., Kothiyal, P., Wang, Z., Snyder,
E., Pontius, J., Ayala, B., Backus, M., Walton, J., Baboud, J., Berton, D., Nicholls, M.,
Srinivasan, D., Raman, R., Girshik, S., Kigonya, P., Alonso, S., Sanbhadti, R., Barletta, S.,
Pot, D., Sheth, M., Demchok, J. A., Shaw, K. R. M., Yang, L., Eley, G., Ferguson, M. L.,
Tarnuzzer, R. W., Zhang, J., Dillon, L. A. L., Buetow, K., Fielding, P., Ozenberger, B. A.,
Guyer, M. S., Sofia, H. J., and Palchik, J. D. (2012) Comprehensive molecular portraits of
human breast tumours. Nature. 490, 61–70
180. Muzny, D. M., Bainbridge, M. N., Chang, K., Dinh, H. H., Drummond, J. A., Fowler, G.,
Kovar, C. L., Lewis, L. R., Morgan, M. B., Newsham, I. F., Reid, J. G., Santibanez, J.,
Shinbrot, E., Trevino, L. R., Wu, Y. Q., Wang, M., Gunaratne, P., Donehower, L. A.,
Creighton, C. J., Wheeler, D. A., Gibbs, R. A., Lawrence, M. S., Voet, D., Jing, R.,
Cibulskis, K., Sivachenko, A., Stojanov, P., mckenna, A., Lander, E. S., Gabriel, S., Ding,
L., Fulton, R. S., Koboldt, D. C., Wylie, T., Walker, J., Dooling, D. J., Fulton, L.,
169 Delehaunty, K. D., Fronick, C. C., Demeter, R., Mardis, E. R., Wilson, R. K., Chu, A.,
Chun, H. J. E., Mungall, A. J., Pleasance, E., Gordon Robertson, A., Stoll, D.,
Balasundaram, M., Birol, I., Butterfield, Y. S. N., Chuah, E., Coope, R. J. N., Dhalla, N.,
Guin, R., Hirst, C., Hirst, M., Holt, R. A., Lee, D., Li, H. I., Mayo, M., Moore, R. A.,
Schein, J. E., Slobodan, J. R., Tam, A., Thiessen, N., Varhol, R., Zeng, T., Zhao, Y.,
Jones, S. J. M., Marra, M. A., Bass, A. J., Ramos, A. H., Saksena, G., Cherniack, A. D.,
Schumacher, S. E., Tabak, B., Carter, S. L., Pho, N. H., Nguyen, H., Onofrio, R. C.,
Crenshaw, A., Ardlie, K., Beroukhim, R., Winckler, W., Meyerson, M., Protopopov, A.,
Hadjipanayis, A., Lee, E., Xi, R., Yang, L., Ren, X., Sathiamoorthy, N., Chen, P. C.,
Haseley, P., Xiao, Y., Lee, S., Seidman, J., Chin, L., Park, P. J., Kucherlapati, R., Todd
Auman, J., Hoadley, K. A., Du, Y., Wilkerson, M. D., Shi, Y., Liquori, C., Meng, S., Li,
L., Turman, Y. J., Topal, M. D., Tan, D., Waring, S., Buda, E., Walsh, J., Jones, C. D.,
Mieczkowski, P. A., Singh, D., Wu, J., Gulabani, A., Dolina, P., Bodenheimer, T., Hoyle,
A. P., Simons, J. V., Soloway, M., Mose, L. E., Jefferys, S. R., Balu, S., O’Connor, B. D.,
Prins, J. F., Chiang, D. Y., Neil Hayes, D., Perou, C. M., Hinoue, T., Weisenberger, D. J.,
Maglinte, D. T., Pan, F., Berman, B. P., Van Den Berg, D. J., Shen, H., Triche, T., Baylin,
S. B., Laird, P. W., Getz, G., Noble, M., Voat, D., Gehlenborg, N., Dicara, D., Zhang, J.,
Zhang, H., Wu, C. J., Liu, S. Y., Shukla, S., Zhou, L., Lin, P., Park, R. W., Nazaire, M.
D., Robinson, J., Thorvaldsdottir, H., Mesirov, J., Thorsson, V., Reynolds, S. M., Bernard,
B., Kreisberg, R., Lin, J., Iype, L., Bressler, R., Erkkilä, T., Gundapuneni, M., Liu, Y.,
Norberg, A., Robinson, T., Yang, D., Zhang, W., Shmulevich, I., De Ronde, J. J., Schultz,
N., Cerami, E., Ciriello, G., Goldberg, A. P., Gross, B., Jacobsen, A., Gao, J.,
Kaczkowski, B., Sinha, R., Arman Aksoy, B., Antipin, Y., Reva, B., Shen, R., Taylor, B.
170 S., Ladanyi, M., Sander, C., Akbani, R., Zhang, N., Broom, B. M., Casasent, T., Unruh,
A., Wakefield, C., Hamilton, S. R., Craig Cason, R., Baggerly, K. A., Weinstein, J. N.,
Haussler, D., Benz, C. C., Stuart, J. M., Benz, S. C., Zachary Sanborn, J., Vaske, C. J.,
Zhu, J., Szeto, C., Scott, G. K., Yau, C., Ng, S., Goldstein, T., Ellrott, K., Collisson, E.,
Cozen, A. E., Zerbino, D., Wilks, C., Craft, B., Spellman, P., Penny, R., Shelton, T.,
Hatfield, M., Morris, S., Yena, P., Shelton, C., Sherman, M., Paulauskis, J., Gastier-
Foster, J. M., Bowen, J., Ramirez, N. C., Black, A., Pyatt, R., Wise, L., White, P.,
Bertagnolli, M., Brown, J., Chan, T. A., Chu, G. C., Czerwinski, C., Denstman, F., Dhir,
R., Dörner, A., Fuchs, C. S., Guillem, J. G., Iacocca, M., Juhl, H., Kaufman, A., Iii, B. K.,
Van Le, X., Mariano, M. C., Medina, E. N., Meyers, M., Nash, G. M., Paty, P. B., Petrelli,
N., Rabeno, B., Richards, W. G., Solit, D., Swanson, P., Temple, L., Tepper, J. E., Thorp,
R., Vakiani, E., Weiser, M. R., Willis, J. E., Witkin, G., Zeng, Z., Zinner, M. J., Zornig,
C., Jensen, M. A., Sfeir, R., Kahn, A. B., Chu, A. L., Kothiyal, P., Wang, Z., Snyder, E.
E., Pontius, J., Pihl, T. D., Ayala, B., Backus, M., Walton, J., Whitmore, J., Baboud, J.,
Berton, D. L., Nicholls, M. C., Srinivasan, D., Raman, R., Girshik, S., Kigonya, P. A.,
Alonso, S., Sanbhadti, R. N., Barletta, S. P., Greene, J. M., Pot, D. A., Shaw, K. R. M.,
Dillon, L. A. L., Buetow, K., Davidsen, T., Demchok, J. A., Eley, G., Ferguson, M.,
Fielding, P., Schaefer, C., Sheth, M., Yang, L., Guyer, M. S., Ozenberger, B. A., Palchik,
J. D., Peterson, J., Sofia, H. J., and Thomson., E. (2012) Comprehensive molecular
characterization of human colon and rectal cancer. Nature. 487, 330–337
181. Fabregat, A., Sidiropoulos, K., Garapati, P., Gillespie, M., Hausmann, K., Haw, R., Jassal,
B., Jupe, S., Korninger, F., mckay, S., Matthews, L., May, B., Milacic, M., Rothfels, K.,
Shamovsky, V., Webber, M., Weiser, J., Williams, M., Wu, G., Stein, L., Hermjakob, H.,
171 and D’Eustachio, P. (2016) The reactome pathway knowledgebase. Nucleic Acids Res. 44,
D481–D487
182. Kanehisa, M., Furumichi, M., Tanabe, M., Sato, Y., and Morishima, K. (2017) KEGG:
New perspectives on genomes, pathways, diseases and drugs. Nucleic Acids Res. 45,
D353–D361
183. HOLLOWAY, P. W., and WAKIL, S. J. (1964) SYNTHESIS OF FATTY ACIDS IN
ANIMAL TISSUES. II. THE OCCURRENCE AND. J. Biol. Chem. 239, 2489–2495
184. Shimomura, I., Shimano, H., Korn, B. S., Bashmakov, Y., and Horton, J. D. (1998)
Nuclear sterol regulatory element-binding proteins activate genes responsible for the
entire program of unsaturated fatty acid biosynthesis in transgenic mouse liver. J. Biol.
Chem. 273, 35299–35306
185. Listenberger, L. L., Han, X., Lewis, S. E., Cases, S., Farese, R. V., Ory, D. S., and
Schaffer, J. E. (2003) Triglyceride accumulation protects against fatty acid-induced
lipotoxicity. Proc. Natl. Acad. Sci. 100, 3077–3082
186. Sathyanarayan, A., Mashek, M. T., and Mashek, D. G. (2017) ATGL Promotes
Autophagy/Lipophagy via SIRT1 to Control Hepatic Lipid Droplet Catabolism. Cell Rep.
19, 1–9
187. Ward, C., Martinez-Lopez, N., Otten, E. G., Carroll, B., Maetzel, D., Singh, R., Sarkar, S.,
and Korolchuk, V. I. (2016) Autophagy, lipophagy and lysosomal lipid storage disorders.
Biochim. Biophys. Acta - Mol. Cell Biol. Lipids. 1861, 269–284
188. Ensenauer, R., Fingerhut, R., Schriever, S. C., Fink, B., Becker, M., Sellerer, N. C., Pagel,
P., Kirschner, A., Dame, T., Olgemöller, B., Röschinger, W., and Roscher, A. A. (2012)
In situ assay of fatty acid β-oxidation by metabolite profiling following permeabilization
172 of cell membranes. J. Lipid Res. 53, 1012–1020
189. Hoppel, C. (2003) The role of carnitine in normal and altered fatty acid metabolism. Am.
J. Kidney Dis. 41, S4–S12
190. Thupari, J. N., Landree, L. E., Ronnett, G. V., and Kuhajda, F. P. (2002) C75 increases
peripheral energy utilization and fatty acid oxidation in diet-induced obesity. Proc. Natl.
Acad. Sci. 99, 9498–9502
191. Landree, L. E., Hanlon, A. L., Strong, D. W., Rumbaugh, G., Miller, I. M., Thupari, J. N.,
Connolly, E. C., Huganir, R. L., Richardson, C., Witters, L. A., Kuhajda, F. P., and
Ronnett, G. V. (2004) C75, a Fatty Acid Synthase Inhibitor, Modulates AMP-activated
Protein Kinase to Alter Neuronal Energy Metabolism. J. Biol. Chem. 279, 3817–3827
192. Wu, X., Daniels, G., Lee, P., and Monaco, M. E. (2014) Lipid metabolism in prostate
cancer. Am. J. Clin. Exp. Urol. 2, 111–20
193. Dufort, F. J., Gumina, M. R., Ta, N. L., Tao, Y., Heyse, S. A., Scott, D. A., Richardson,
A. D., Seyfried, T. N., and Chiles, T. C. (2014) Glucose-dependent de novo lipogenesis in
B lymphocytes: A requirement for atp-citrate lyase in lipopolysaccharide-induced
differentiation. J. Biol. Chem. 289, 7011–7024
194. Sanders, F. W. B., and Griffin, J. L. (2016) De novo lipogenesis in the liver in health and
disease: More than just a shunting yard for glucose. Biol. Rev. 91, 452–468
195. Mookerjee, S. A., Nicholls, D. G., and Brand, M. D. (2016) Determining Maximum
Glycolytic Capacity Using Extracellular Flux Measurements. Plos One. 11, e0152016
196. Kuhajda, F. P., Pizer, E. S., Li, J. N., Mani, N. S., Frehywot, G. L., and Townsend, C. A.
(2000) Synthesis and antitumor activity of an inhibitor of fatty acid synthase. Proc. Natl.
Acad. Sci. U. S. A. 97, 3450–4
173 197. Fritz, V., Benfodda, Z., Henriquet, C., Hure, S., Cristol, J. P., Michel, F., Carbonneau, M.
A., Casas, F., and Fajas, L. (2013) Metabolic intervention on lipid synthesis converging
pathways abrogates prostate cancer growth. Oncogene. 32, 5101–5110
198. Bhatt, A. P., Jacobs, S. R., Freemerman, A. J., Makowski, L., Rathmell, J. C., Dittmer, D.
P., and Damania, B. (2012) Dysregulation of fatty acid synthesis and glycolysis in non-
Hodgkin lymphoma. Proc Natl Acad Sci U S A. 109, 11818–11823
199. Mcfadden, J. W., Aja, S., Li, Q., Bandaru, V. V. R., Kim, E. K., Haughey, N. J., Kuhajda,
F. P., and Ronnett, G. V. (2014) Increasing fatty acid oxidation remodels the
hypothalamic neurometabolome to mitigate stress and inflammation. Plos One.
10.1371/journal.pone.0115642
200. Hardie, D. G., Ross, F. A., and Hawley, S. A. (2012) AMPK: A nutrient and energy
sensor that maintains energy homeostasis. Nat. Rev. Mol. Cell Biol. 13, 251–262
201. Fullerton, M. D., Galic, S., Marcinko, K., Sikkema, S., Pulinilkunnil, T., Chen, Z.-P.,
O’Neill, H. M., Ford, R. J., Palanivel, R., O’Brien, M., Hardie, D. G., Macaulay, S. L.,
Schertzer, J. D., Dyck, J. R. B., van Denderen, B. J., Kemp, B. E., and Steinberg, G. R.
(2013) Single phosphorylation sites in Acc1 and Acc2 regulate lipid homeostasis and the
insulin-sensitizing effects of metformin. Nat. Med. 19, 1649–1654
202. Walther, T. C., and Farese, R. V (2012) Lipid droplets and cellular lipid metabolism.
Annu. Rev. Biochem. 81, 687–714
203. Huang, D. W., Sherman, B. T., and Lempicki, R. A. (2009) Bioinformatics enrichment
tools: Paths toward the comprehensive functional analysis of large gene lists. Nucleic
Acids Res. 37, 1–13
204. Huang, D. W., Sherman, B. T., and Lempicki, R. A. (2009) Systematic and integrative
174 analysis of large gene lists using\ndavid bioinformatics resources. Nat. Protoc. 4, 44–57
205. Ruiz, R., Jideonwo, V., Ahn, M., Surendran, S., Tagliabracci, V. S., Hou, Y., Gamble, A.,
Kerner, J., Irimia-Dominguez, J. M., Puchowicz, M. A., depaoli-Roach, A., Hoppel, C.,
Roach, P., and Morral, N. (2014) Sterol regulatory element-binding protein-1 (SREBP-1)
is required to regulate glycogen synthesis and gluconeogenic gene expression in mouse
liver. J. Biol. Chem. 289, 5510–5517
206. Vega, R. B., Huss, J. M., and Kelly, D. P. (2000) The Coactivator PGC-1 Cooperates with
Peroxisome Proliferator-Activated Receptor alpha in Transcriptional Control of Nuclear
Genes Encoding Mitochondrial Fatty Acid Oxidation Enzymes. Mol. Cell. Biol. 20, 1868–
1876
207. Mihaylova, M. M., Cheng, C. W., Cao, A. Q., Tripathi, S., Mana, M. D., Bauer-Rowe, K.
E., Abu-Remaileh, M., Clavain, L., Erdemir, A., Lewis, C. A., Freinkman, E., Dickey, A.
S., La Spada, A. R., Huang, Y., Bell, G. W., Deshpande, V., Carmeliet, P., Katajisto, P.,
Sabatini, D. M., and Yilmaz, Ö. H. (2018) Fasting Activates Fatty Acid Oxidation to
Enhance Intestinal Stem Cell Function during Homeostasis and Aging. Cell Stem Cell.
10.1016/j.stem.2018.04.001
208. Lee, J., Choi, J., Scafidi, S., and Wolfgang, M. J. (2016) Hepatic Fatty Acid Oxidation
Restrains Systemic Catabolism during Starvation. Cell Rep. 16, 201–212
209. Monaco, M. E. (2017) Fatty acid metabolism in breast cancer subtypes. Oncotarget. 8,
29487–29500
210. Park, J. H., Vithayathil, S., Kumar, S., Sung, P. L., Dobrolecki, L. E., Putluri, V., Bhat, V.
B., Bhowmik, S. K., Gupta, V., Arora, K., Wu, D., Tsouko, E., Zhang, Y., Maity, S.,
Donti, T. R., Graham, B. H., Frigo, D. E., Coarfa, C., Yotnda, P., Putluri, N., Sreekumar,
175 A., Lewis, M. T., Creighton, C. J., Wong, L. J. C., and Kaipparettu, B. A. (2016) Fatty
Acid Oxidation-Driven Src Links Mitochondrial Energy Reprogramming and Oncogenic
Properties in Triple-Negative Breast Cancer. Cell Rep. 14, 2154–2165
211. Yamashita, Y., Nishiumi, S., Kono, S., Takao, S., Azuma, T., and Yoshida, M. (2017)
Differences in elongation of very long chain fatty acids and fatty acid metabolism between
triple-negative and hormone receptor-positive breast cancer. BMC Cancer. 17, 1–21
212. Sørlie, T., Wang, Y., Xiao, C., Johnsen, H., Naume, B., Samaha, R. R., and Børresen-
Dale, A. L. (2006) Distinct molecular mechanisms underlying clinically relevant subtypes
of breast cancer: Gene expression analyses across three different platforms. BMC
Genomics. 10.1186/1471-2164-7-127
213. Kittleson, M. M. (2005) Gene expression analysis of ischemic and nonischemic
cardiomyopathy: shared and distinct genes in the development of heart failure. Physiol.
Genomics. 21, 299–307
214. Banerjee, P., Banerjee, P., and Dhal, S. S. (2012) International Journal of Advanced
Research in Computer Science and Software Engineering. Int. J. 2, 62–70
215. Barth, A., Kuner, R., Buness, A., Ruschhaupt, M., Merk, S., Zwermann, L., Kääb, S.,
Kreuzer, E., Steinbeck, G., Mansmann, U., Poustka, A., Nabauer, M., and Sültmann, H.
(2006) Identification of a Common Gene Expression Signature in Dilated
Cardiomyopathy Across Independent Microarray Studies, 10.1016/j.jacc.2006.07.026
216. Rome, S., Clément, K., Rabasa-Lhoret, R., Loizon, E., Poitou, C., Barsh, G. S., Riou, J. P.,
Laville, M., and Vidal, H. (2003) Microarray profiling of human skeletal muscle reveals
that insulin regulates 800 genes during a hyperinsulinemic clamp. J. Biol. Chem. 278,
18063–18068 ∼
176 217. Schwab, M. (ed.) (2011) Class II Tumor Suppressor Genes. In Encyclopedia of Cancer,
pp. 877–879, Springer Berlin Heidelberg, Berlin, Heidelberg, 10.1007/978-3-642-16483-
5_1201
218. Gropper SS, S. J. (2013) Advanced nutrition and human metabolism, 10.1016/0307-
4412(92)90040-S
219. Zirn, B., Samans, B., Spangenberg, C., Graf, N., Eilers, M., and Gessler, M. (2005) All-
trans retinoic acid treatment of Wilms tumor cells reverses expression of genes associated
with high risk and relapse in vivo. Oncogene. 24, 5246–5251
220. Bikle, D. D. (2004) Vitamin D regulated keratinocyte differentiation. J. Cell. Biochem. 92,
436–444
221. Cross, H. S. (2005) Vitamin D and Colon Cancer. In Vitamin D, pp. 1709–1725, 2, 1709–
1725
222. Herrero, C., Hu, X., Li, W. P., Samuels, S., Sharif, M. N., Kotenko, S., and Ivashkiv, L. B.
(2003) Reprogramming of IL-10 Activity and Signaling by IFN- . J. Immunol. 171, 5034–
5041
223. Wehinger, J., Gouilleux, F., Groner, B., Finke, J., Mertelsmann, R., and Weber-Nordt, R.
M. (1996) IL-10 induces DNA binding activity of three STAT proteins (Stat1, Stat3, and
Stat5) and their distinct combinatorial assembly in the promoters of selected genes. FEBS
Lett. 394, 365–370
224. Mäkelä, M. J., Kanehiro, A., Borish, L., Dakhama, A., Loader, J., Joetham, A., Xing, Z.,
Jordana, M., Larsen, G. L., and Gelfand, E. W. (2000) IL-10 is necessary for the
expression of airway hyperresponsiveness but not pulmonary inflammation after allergic
sensitization. Proc. Natl. Acad. Sci. U. S. A. 97, 6007–12
177 225. Huebner, H., Strick, R., Wachter, D. L., Kehl, S., Strissel, P. L., Schneider-Stock, R.,
Hartner, A., Rascher, W., Horn, L. C., Beckmann, M. W., Ruebner, M., and Fahlbusch, F.
B. (2017) Hypermethylation and loss of retinoic acid receptor responder 1 expression in
human choriocarcinoma. J. Exp. Clin. Cancer Res. 10.1186/s13046-017-0634-x
226. Ching, J. K., Rajguru, P., Marupudi, N., Banerjee, S., and Fisher, J. S. (2010) A role for
AMPK in increased insulin action after serum starvation. AJP Cell Physiol. 299, C1171–
C1179
227. Zielinski, R., Przytycki, P. F., Zheng, J., Zhang, D., Przytycka, T. M., and Capala, J.
(2009) The crosstalk between EGF, IGF, and Insulin cell signaling pathways -
Computational and experimental analysis. BMC Syst. Biol. 3, 88
228. Borisov, N., Aksamitiene, E., Kiyatkin, A., Legewie, S., Berkhout, J., Maiwald, T.,
Kaimachnikov, N. P., Timmer, J., Hoek, J. B., and Kholodenko, B. N. (2009) Systems-
level interactions between insulin-EGF networks amplify mitogenic signaling. Mol. Syst.
Biol. 10.1038/msb.2009.19
229. Laplante, M., and Sabatini, D. M. (2012) MTOR signaling in growth control and disease.
Cell. 149, 274–293
230. Schieke, S. M., Phillips, D., mccoy, J. P., Aponte, A. M., Shen, R. F., Balaban, R. S., and
Finkel, T. (2006) The mammalian target of rapamycin (mtor) pathway regulates
mitochondrial oxygen consumption and oxidative capacity. J. Biol. Chem. 281, 27643–
27652
231. Leontieva, O. V., and Blagosklonny, M. V. (2014) M(o)TOR of pseudo-hypoxic state in
aging: Rapamycin to the rescue. Cell Cycle. 13, 509–515
232. Margolis, R. N., Evans, R. M., and O’Malley, B. W. (2005) The Nuclear Receptor
178 Signaling Atlas: Development of a Functional Atlas of Nuclear Receptors. Mol.
Endocrinol. 19, 2433–2436
233. Messeguer, X., Escudero, R., Farré, D., Núñez, O., Martínez, J., and Albà, M. M. (2002)
PROMO: Detection of known transcription regulatory elements using species-tailored
searches. Bioinformatics. 18, 333–334
234. Blättler, S. M., Verdeguer, F., Liesa, M., Cunningham, J. T., Vogel, R. O., Chim, H., Liu,
H., Romanino, K., Shirihai, O. S., Vazquez, F., Rüegg, M. A., Shi, Y., and Puigserver, P.
(2012) Defective mitochondrial morphology and bioenergetic function in mice lacking the
transcription factor Yin Yang 1 in skeletal muscle. Mol. Cell. Biol. 32, 3333–3346
235. Cunningham, J. T., Rodgers, J. T., Arlow, D. H., Vazquez, F., Mootha, V. K., and
Puigserver, P. (2007) mtor controls mitochondrial oxidative function through a YY1-
PGC-1alpha transcriptional complex. Nature. 450, 736–740
236. Blättler, S. M., Cunningham, J. T., Verdeguer, F., Chim, H., Haas, W., Liu, H., Romanino,
K., Rüegg, M. A., Gygi, S. P., Shi, Y., and Puigserver, P. (2012) Yin Yang 1 deficiency in
skeletal muscle protects against rapamycin-induced diabetic-like symptoms through
activation of insulin/IGF signaling. Cell Metab. 15, 505–517
237. Broglie, L., Margolis, D., and Medin, J. A. (2017) Yin and Yang of mesenchymal stem
cells and aplastic anemia. World J. Stem Cells. 9, 219–226
238. Lu, Z., Hong, C. C., Kong, G., Assumpção, A. L. F. V, Ong, I. M., Bresnick, E. H.,
Zhang, J., and Pan, X. (2018) Polycomb Group Protein YY1 Is an Essential Regulator of
Hematopoietic Stem Cell Quiescence. Cell Rep. 22, 1545–1559
239. Wang, G. Z., and Goff, S. P. (2015) Regulation of Yin Yang 1 by tyrosine
phosphorylation. J. Biol. Chem. 290, 21890–21900
179 240. Kassardjian, A., Rizkallah, R., Riman, S., Renfro, S. H., Alexander, K. E., and Hurt, M.
M. (2012) The Transcription Factor YY1 Is a Novel Substrate for Aurora B Kinase at
G2/M Transition of the Cell Cycle. Plos One. 10.1371/journal.pone.0050645
241. Deng, Z., Cao, P., Wan, M., and Sui, G. (2010) Yin Yang 1. Transcription. 1, 81–84
242. Adam, R. C., and Fuchs, E. (2016) The Yin and Yang of Chromatin Dynamics In Stem
Cell Fate Selection. Trends Genet. 32, 89–100
243. Stange, D. E., and Clevers, H. (2013) Concise review: the yin and yang of intestinal
(cancer) stem cells and their progenitors. Stem Cells. 31, 2287–95
244. Yu, K. R., Yang, S. R., Jung, J. W., Kim, H., Ko, K., Han, D. W., Park, S. B., Choi, S. W.,
Kang, S. K., Schöler, H., and Kang, K. S. (2012) CD49f enhances multipotency and
maintains stemness through the direct regulation of OCT4 and SOX2. Stem Cells. 30,
876–887
245. Verdeguer, F., Blättler, S. M., Cunningham, J. T., Hall, J. A., Chim, H., and Puigserver, P.
(2014) Decreased Genetic Dosage of Hepatic Yin Yang 1 Causes Diabetic-Like
Symptoms. Mol. Endocrinol. 28, 308–316
246. Fang, Y. (1998) Regulation of epidermal growth factor expression in mammary epithelial
cells by a Yin-Yang-1-like element. J. Mol. Endocrinol. 20, 337–344
247. Kersten, S., Desvergne, B., and Wahli, W. (2000) Roles of ppars in health and disease.
Nature. 405, 421–424
248. Nunez-Anita, R. E., Cajero-Juarez, M., and Aceves, C. (2011) Peroxisome proliferator-
activated receptors: role of isoform gamma in the antineoplastic effect of iodine in
mammary cancer. Curr. Cancer Drug Targets. 11, 775–86
249. Lee, S. J., Yang, E. K., and Kim, S. G. (2006) Peroxisome proliferator-activated receptor-
180 gamma and retinoic acid X receptor alpha represses the tgfbeta1 gene via PTEN-mediated
p70 ribosomal S6 kinase-1 inhibition: role for Zf9 dephosphorylation. Mol. Pharmacol.
70, 415–425
250. Lachmann, A., Xu, H., Krishnan, J., Berger, S. I., Mazloom, A. R., and Ma’ayan, A.
(2010) chea: Transcription factor regulation inferred from integrating genome-wide chip-
X experiments. Bioinformatics. 26, 2438–2444
251. Semenkovich, C. F. (2005) tzds and diabetes: Testing the waters. Nat. Med. 11, 822–824
252. Mckeage, K., and Keating, G. M. (2011) Fenofibrate. Drugs. 71, 1917–1946
253. Srivastava, R. A. K. (2009) Fenofibrate ameliorates diabetic and dyslipidemic profiles in
kkay mice partly via down-regulation of 11β-HSD1, PEPCK and DGAT2. Comparison of
pparα, pparγ, and liver x receptor agonists. Eur. J. Pharmacol. 607, 258–263
254. Rao, M. S., and Reddy, J. K. (2001) Peroxisomal beta-oxidation and steatohepatitis.
Semin. Liver Dis. 21, 43–55
255. Hashimoto, T., Cook, W. S., Qi, C., Yeldandi, A. V., Reddy, J. K., and Rao, M. S. (2000)
Defect in peroxisome proliferator-activated receptor α-inducible fatty acid oxidation
determines the severity of hepatic steatosis in response to fasting. J. Biol. Chem. 275,
28918–28928
256. Varga, T. Are a unique set of fatty acid regulated transcription factors controlling both
lipid metabolism and inflammation., Czimmerer, Z., and Nagy, L. (2011) ppars are a
unique set of fatty acid regulated transcription factors controlling both lipid metabolism
and inflammation. Biochim. Biophys. Acta. 1812, 1007–22
257. Dormandy, J. A., al, et, and Investigators, Pro. S. (2005) Secondary prevention of
macrovascular events in patients with type 2 diabetes in the proactive Study (prospective
181 pioglitazone Clinical Trial In macrovascular Events): a randomised controlled trial.
Lancet. 366, 1279–1289
258. Saltiel, A. R., and Olefsky, J. M. (1996) Thiazolidinediones in the treatment of insulin
resistance and type II diabetes. Diabetes. 45, 1661–1669
259. Soccio, R. E., Chen, E. R., and Lazar, M. A. (2014) Thiazolidinediones and the promise of
insulin sensitization in type 2 diabetes. Cell Metab. 20, 573–591
260. Nolan, J. J., Ludvik, B., Beerdsen, P., Joyce, M., and Olefsky, J. (1994) Improvement in
glucose tolerance and insulin resistance in obese subjects treated with troglitazone. N.
Engl. J. Med. 331, 1188–93
261. Seluanov, A., Hine, C., Azpurua, J., Feigenson, M., Bozzella, M., Mao, Z., Catania, K. C.,
and Gorbunova, V. (2009) Hypersensitivity to contact inhibition provides a clue to cancer
resistance of naked mole-rat. Proc. Natl. Acad. Sci. 106, 19352–19357
262. Growth, E., and Receptor, F. (1985) Epidermal Growth Factor Receptor. J. Biol. Chem.
260, 9820–9824
263. Petrides, P. E., Bock, S., Bovens, J., Hofmann, R., and Jakse, G. (1990) Modulation of
pro-epidermal growth factor, pro-transforming growth factor alpha and epidermal growth
factor receptor gene expression in human renal carcinomas. Cancer Res. 50, 3934–9
264. Henson, E. S., and Gibson, S. B. (2006) Surviving cell death through epidermal growth
factor (EGF) signal transduction pathways: Implications for cancer therapy. Cell. Signal.
18, 2089–2097
265. Kuleshov, M. V., Jones, M. R., Rouillard, A. D., Fernandez, N. F., Duan, Q., Wang, Z.,
Koplev, S., Jenkins, S. L., Jagodnik, K. M., Lachmann, A., mcdermott, M. G., Monteiro,
C. D., Gundersen, G. W., and Ma’ayan, A. (2016) Enrichr: a comprehensive gene set
182 enrichment analysis web server 2016 update. Nucleic Acids Res. 44, W90–W97
266. Ricote, M., and Glass, C. K. (2007) ppars and molecular mechanisms of transrepression.
Biochim. Biophys. Acta - Mol. Cell Biol. Lipids. 1771, 926–935
267. Devchand, P. R., Keller, H., Peters, J. M., Vazquez, M., Gonzalez, F. J., and Wahli, W.
(1996) The pparα-leukotriene B4 pathway to inflammation control. Nature. 384, 39–43
268. Lee, C. H., Chawla, A., Urbiztondo, N., Liao, D., Boisvert, W. A., and Evans, R. M.
(2003) Transcriptional Repression of Atherogenic Inflammation: Modulation by pparδ.
Science (80-. ). 302, 453–457
269. Feinstein, D. L., Spagnolo, A., Akar, C., Weinberg, G., Murphy, P., Gavrilyuk, V., and
Dello Russo, C. (2005) Receptor-independent actions of PPAR thiazolidinedione agonists:
Is mitochondrial function the key? Biochem. Pharmacol. 70, 177–188
270. Scatena, R., Bottoni, P., Vincenzoni, F., Messana, I., Martorana, G. E., Nocca, G., De
Sole, P., Maggiano, N., Castagnola, M., and Giardina, B. (2003) Bezafibrate Induces a
Mitochondrial Derangement in Human Cell Lines: A PPAR-Independent Mechanism for a
Peroxisome Proliferator. Chem. Res. Toxicol. 16, 1440–1447
271. Zhang, H., Ren, Y., Pang, D., and Liu, C. (2014) Clinical implications of AGBL2
expression and its inhibitor latexin in breast cancer. World J. Surg. Oncol. 12, 1–7
272. Snyder, V., Reed-Newman, T. C., Arnold, L., Thomas, S. M., and Anant, S. (2018)
Cancer Stem Cell Metabolism and Potential Therapeutic Targets. Front. Oncol. 8, 203
273. Gurumurthy, S., Xie, S. Z., Alagesan, B., Kim, J., Yusuf, R. Z., Saez, B., Tzatsos, A.,
Ozsolak, F., Milos, P., Ferrari, F., Park, P. J., Shirihai, O. S., Scadden, D. T., and
Bardeesy, N. (2010) The Lkb1 metabolic sensor maintains haematopoietic stem cell
survival. Nature. 468, 659–63
183 274. Peiris-Pagès, M., Martinez-Outschoorn, U. E., Pestell, R. G., Sotgia, F., and Lisanti, M. P.
(2016) Cancer stem cell metabolism. Breast Cancer Res. 18, 55
275. Attie, A. D., and Scherer, P. E. (2009) Adipocyte metabolism and obesity: Fig. 1. J. Lipid
Res. 50, S395–S399
276. Rosen, E. D., and Spiegelman, B. M. (2006) Adipocytes as regulators of energy balance
and glucose homeostasis. Nature. 444, 847–853
277. Miyoshi, H., Souza, S. C., Zhang, H. H., Strissel, K. J., Christoffolete, M. A., Kovsan, J.,
Rudich, A., Kraemer, F. B., Bianco, A. C., Obin, M. S., and Greenberg, A. S. (2006)
Perilipin promotes hormone-sensitive lipase-mediated adipocyte lipolysis via
phosphorylation-dependent and -independent mechanisms. J. Biol. Chem. 281, 15837–
15844
278. Bijland, S., Mancini, S. J., and Salt, I. P. (2013) Role of AMP-activated protein kinase in
adipose tissue metabolism and inflammation. Clin. Sci. (Lond). 124, 491–507
279. Drehmer, D. L., de Aguiar, A. M., Brandt, A. P., Petiz, L., Cadena, S. M. S. C., Rebelatto,
C. K., Brofman, P. R. S., Filipak Neto, F., Dallagiovanna, B., and Abud, A. P. R. (2016)
Metabolic switches during the first steps of adipogenic stem cells differentiation. Stem
Cell Res. 17, 413–421
280. Vicente, C. P., Fortuna, V. A, Margis, R., Trugo, L., and Borojevic, R. (1998) Retinol
uptake and metabolism, and cellular retinol binding protein expression in an in vitro
model of hepatic stellate cells. Mol Cell Biochem. 187, 11–21
281. Beaven, S. W., Wroblewski, K., Wang, J., Hong, C., Bensinger, S., Tsukamoto, H., and
Tontonoz, P. (2011) Liver X receptor signaling is a determinant of stellate cell activation
and susceptibility to fibrotic liver disease. Gastroenterology. 140, 1052–1062
184 282. Sousa, C. M., Biancur, D. E., Wang, X., Halbrook, C. J., Sherman, M. H., Zhang, L.,
Kremer, D., Hwang, R. F., Witkiewicz, A. K., Ying, H., Asara, J. M., Evans, R. M.,
Cantley, L. C., Lyssiotis, C. A., and Kimmelman, A. C. (2016) Pancreatic stellate cells
support tumour metabolism through autophagic alanine secretion. Nature. 536, 479–483
283. Sunami, Y., Rebelo, A., and Kleeff, J. (2017) Lipid Metabolism and Lipid Droplets in
Pancreatic Cancer and Stellate Cells. Cancers (Basel). 10, 3
284. Li, J., Ghazwani, M., Liu, K., Huang, Y., Chang, N., Fan, J., He, F., Li, L., Bu, S., Xie,
W., Ma, X., and Li, S. (2017) Regulation of hepatic stellate cell proliferation and
activation by glutamine metabolism. Plos One. 10.1371/journal.pone.0182679
285. Yang, L., Achreja, A., Yeung, T. L., Mangala, L. S., Jiang, D., Han, C., Baddour, J.,
Marini, J. C., Ni, J., Nakahara, R., Wahlig, S., Chiba, L., Kim, S. H., Morse, J., Pradeep,
S., Nagaraja, A. S., Haemmerle, M., Kyunghee, N., Derichsweiler, M., Plackemeier, T.,
Mercado-Uribe, I., Lopez-Berestein, G., Moss, T., Ram, P. T., Liu, J., Lu, X., Mok, S. C.,
Sood, A. K., and Nagrath, D. (2016) Targeting Stromal Glutamine Synthetase in Tumors
Disrupts Tumor Microenvironment-Regulated Cancer Cell Growth. Cell Metab. 24, 685–
700
286. Tajan, M., and Vousden, K. H. (2016) The Quid Pro Quo of the Tumor/Stromal
Interaction. Cell Metab. 24, 645–646
287. Bradbury, M. W. (2006) Lipid metabolism and liver inflammation. Am. J. Physiol.
Gastrointest. Liver Physiol. 290, G194-8
288. Deng, J., Huang, Q., Wang, Y., Shen, P., Guan, F., Li, J., Huang, H., and Shi, C. (2014)
Hypoxia-inducible factor-1alpha regulates autophagy to activate hepatic stellate cells.
Biochem. Biophys. Res. Commun. 454, 328–334
185 289. Thoen, L. F. R., Guimarães, E. L. M., Dollé, L., Mannaerts, I., Najimi, M., Sokal, E., and
Van Grunsven, L. A. (2011) A role for autophagy during hepatic stellate cell activation. J.
Hepatol. 55, 1353–1360
290. Hsu, C. C., and Schwabe, R. F. (2011) Autophagy and hepatic stellate cell activation -
Partners in crime? J. Hepatol. 55, 1176–1177
291. Hong, Y., Li, S., Wang, J., and Li, Y. (2018) In vitro inhibition of hepatic stellate cell
activation by the autophagy- related lipid droplet protein ATG2A. Sci. Rep.
10.1038/s41598-018-27686-6
292. Bround, M. J., Wambolt, R., Luciani, D. S., Kulpa, J. E., Rodrigues, B., Brownsey, R. W.,
Allard, M. F., and Johnson, J. D. (2013) Cardiomyocyte ATP production, metabolic
flexibility, and survival require calcium flux through cardiac ryanodine receptors in vivo.
J. Biol. Chem. 288, 18975–18986
293. Wang, K., Xu, Y., Sun, Q., Long, J., Liu, J., and Ding, J. (2018) Mitochondria regulate
cardiac contraction through ATP-dependent and ATP-independent mechanisms. Free
Radic. Res. 10.1080/10715762.2018.1453137
294. Tong, Z., Xie, Y., He, M., Ma, W., Zhou, Y., Lai, S., Meng, Y., and Liao, Z. (2017)
VDAC1 deacetylation is involved in the protective effects of resveratrol against
mitochondria-mediated apoptosis in cardiomyocytes subjected to anoxia/reoxygenation
injury. Biomed. Pharmacother. 95, 77–83
295. Gibb, A. A., Epstein, P. N., Uchida, S., Zheng, Y., mcnally, L. A., Obal, D., Katragadda,
K., Trainor, P., Conklin, D. J., Brittian, K. R., Tseng, M. T., Wang, J., Jones, S. P.,
Bhatnagar, A., and Hill, B. G. (2017) Exercise-Induced Changes in Glucose Metabolism
Promote Physiological Cardiac Growth. Circulation. 136, 2144–2157
186 296. Even-Ram, S., Doyle, A. D., Conti, M. A., Matsumoto, K., Adelstein, R. S., and Yamada,
K. M. (2007) Myosin IIA regulates cell motility and actomyosin-microtubule crosstalk.
Nat. Cell Biol. 9, 299–309
297. Robison, P., and Prosser, B. L. (2017) Microtubule mechanics in the working myocyte. J.
Physiol. 595, 3931–3937
298. Belmadani, S., Pous, C., Ventura-Clapier, R., Fischmeister, R., and Mery, P. F. (2002)
Post-translational modifications of cardiac tubulin during chronic heart failure in the rat.
Mol Cell Biochem. 237, 39–46
299. Robison, P., Caporizzo, M. A., Ahmadzadeh, H., Bogush, A. I., Chen, C. Y., Margulies,
K. B., Shenoy, V. B., and Prosser, B. L. (2016) Detyrosinated microtubules buckle and
bear load in contracting cardiomyocytes. Science (80-. ). 10.1126/science.aaf0659
300. Kerr, J. P., Robison, P., Shi, G., Bogush, A. I., Kempema, A. M., Hexum, J. K., Becerra,
N., Harki, D. A., Martin, S. S., Raiteri, R., Prosser, B. L., and Ward, C. W. (2015)
Detyrosinated microtubules modulate mechanotransduction in heart and skeletal muscle.
Nat. Commun. 6, 1–14
301. Chen, C. Y., Caporizzo, M. A., Bedi, K., Vite, A., Bogush, A. I., Robison, P., Heffler, J.
G., Salomon, A. K., Kelly, N. A., Babu, A., Morley, M. P., Margulies, K. B., and Prosser,
B. L. (2018) Suppression of detyrosinated microtubules improves cardiomyocyte function
in human heart failure. Nat. Med. 10.1038/s41591-018-0046-2
302. Kittleson, M. M. (2005) Gene expression analysis of ischemic and nonischemic
cardiomyopathy: shared and distinct genes in the development of heart failure. Physiol.
Genomics. 21, 299–307
303. Navani, S. (2011) The human protein atlas. J. Obstet. Gynecol. India. 61, 27–31
187