Biology of Type 2 Phosphatidylinositol-5- Phosphate 4-Kinase
The Harvard community has made this article openly available. Please share how this access benefits you. Your story matters
Citation Shim, Hyeseok. 2015. Biology of Type 2 Phosphatidylinositol-5- Phosphate 4-Kinase. Doctoral dissertation, Harvard University, Graduate School of Arts & Sciences.
Citable link http://nrs.harvard.edu/urn-3:HUL.InstRepos:23845419
Terms of Use This article was downloaded from Harvard University’s DASH repository, and is made available under the terms and conditions applicable to Other Posted Material, as set forth at http:// nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of- use#LAA Biology of type 2 phosphatidylinositol-5-phosphate 4-kinase
A dissertation presented
by
Hyeseok Shim
to
The Division of Medical Sciences
in partial fulfillment of the requirements
for the degree of
Doctor of Philosophy
in the subject of
Biological and Biomedical Sciences
Harvard University
Cambridge, Massachusetts
August 2015
© 2015 Hyeseok Shim
All rights reserved
Professor Lewis C. Cantley Hyeseok Shim
Biology of type 2 phosphatidylinositol-5-phosphate 4-kinase
Abstract
Type 2 phosphatidylinositol-5-phosphate 4-kinase (PI5P4K) converts phosphatidylinositol-5-phosphate to phosphatidylinositol-4,5-bisphosphate.
Mammals have three genes, PIP4K2A, PIP4K2B and PIP4K2C that encode the enzymes PI5P4Kα, PI5P4Kβ and PI5P4Kγ respectively. Studies in mice showed that PI5P4Kβ is a negative regulator of insulin signaling (Lamia et al., 2004) and that co-deletion of Pip4k2b and Trp53 resulted in synthetic embryonic lethality
(Emerling et al., 2013). Also, deletion of two alleles of Pip4k2a and one allele of
Pip4k2b suppressed the appearance of tumors in Trp53-/- mice. These studies suggest that drugs targeting PI5P4Kα and β could be effective therapies for treating insulin resistance, type 2 diabetes and TP53 mutant cancers.
While less is known about PI5P4Kγ, several genome-wide association studies have revealed a SNP in front of the PIP4K2C at an autoimmunity susceptibility loci (Raychaudhuri et al., 2008). To evaluate the role of PI5P4Kγ, I generated Pip4k2c-/- mice and found an inflammatory phenotype with increased tissue immune infiltrates and pro-inflammatory cytokines, correlating with increased helper T cells and decreased regulatory T cells. Also, Pip4k2c-/- mice exhibited upregulated mammalian target of rapamycin complex 1 (mTORC1) signaling in tissues and rapamycin treatment reduced the inflammation of these mice. These studies support the concept that the SNP identified at the PIP4K2C
iii locus in human patients with autoimmunity contributes to disease by reducing expression of PI5P4Kγ and indicates that inhibition of mTORC1 would be beneficial to these patients.
Finally, in collaboration with Dr. Nathanael Gray’s laboratory we identified small molecules that covalently react with PI5P4Ks and thereby cause irreversible inhibition. These compounds, PIP4Kin1 and PIP4Kin2 mimicked the effect of shRNA mediated knockdown or knockout of PI5P4Kα and PI5P4Kβ, and impaired the growth of several TP53 mutant cancer cell lines, with little effect on most TP53 wild type cell lines. Utilizing the xenograft tumor model with BT474
(TP53 mutant) and MCF7 (TP53 wild type) cells, we showed that daily treatment of the mice with PIP4Kin2 inhibited the growth of the BT474 tumors but not the
MCF7 tumors, without causing any obvious toxicity. These results further validate
PI5P4Kα and PI5P4Kβ as targets for treating TP53 mutant cancers.
iv Table of Contents
Title i
Abstract iii
Table of Contents v
Acknowledgements vi
List of Tables viii
List of Figures ix
List of Abbreviations xii
Chapter 1: Introduction 1
Chapter 2: Deletion of a Novel Phosphatidylinositol Kinase results in 14
Hyperactivation of the Immune System
Chapter 3: Identification of PI5P4K inhibitors 46
Chapter 4: Summary and Future Directions 85
Appendix A: Depletion of a Putatively Druggable Class of 93
Phosphatidylinositol Kinases Inhibits Growth of p53-Null Tumors
Bibliography 137
Supplemental Materials 143
v Acknowledgments
First of all, I am deeply grateful to my advisor, Dr. Lewis C. Cantley, for mentoring my graduate study. His unlimited support, guidance and confidence encouraged me to move forward and complete all the steps required for achieving a Ph.D degree. I would like to thank my thesis committee, Drs. Karen
Cichowski, John Blenis and Alex Toker for giving valuable comments and suggestions on my thesis.
I thank Drs. Nathanael Gray and Tinghu Zhang for collaborating with me on developing PI5P4K inhibitors and Drs. Vijay Kuchroo and Chuan Wu on analyzing T cells in Pip4k2c -/- mice. I thank Drs. Sirano Dhe-Paganon and Hyuk-
Soo Seo for crystalizing PI5P4Kβ with PIP4Kin2 and Drs. Scott Ficarro and
Jarrod Marto for identifying modified residues of PI5P4K using mass spectrometry.
I thank members of Cantley laboratory. I thank Jihye Yun and Kaitlyn
Bosch for helping me on rapamycin treatment of Pip4k2c -/- mice, Yuxiang Zheng for performing HPLC to measure phospholipids and Gina DeNicola for measuring
ROS in cells. I give my thanks to Shivan Ramsamooj for culturing tissues and performing western blots and Zhiwei Yang and Rayman Choo-Wing for mouse genotyping. I also thank Brooke Emerling, Atsuo Sasaki, Hui Liu, Edouard
Mullarky, Jared Johnson, Florian Karreth, Jonathan Yang and Diana Wang for all their helpful suggestions and support during my PhD study.
vi Finally, I would like to thank my parents, Hyunho Shim and Dahye Lee for giving me encouragement and financial help. I thank Pom and Kumdong for staying with me. Above all, I am truly indebted to my husband, Young Kwon for his endless support and encouragement in both life and science.
vii List of Tables
Table 3.1 Metabolic pathways affect by PIP4Kin2 57-58
viii List of Figures
Figure 1.1 Compartmentalization of phosphoinositide pathways 2-3
Figure 1.2 Working Model for the role of PI5P4Kγ in immune regulation 10
Figure 2.1 Generation of Pip4k2c-/- mice 19
Figure 2.2 Validation of Pip4k2c-/- mice 20
Figure 2.3 Immune cell infiltration is increased in tissues of Pip4k2c-/- 22
mice
Figure 2.4 Immune cell infiltration in the liver of Pip4k2c-/- mice 23
Figure 2.5 Immune infiltrates in liver tissue of Pip4k2c-/- mice are mostly 23
T cells and B cells
Figure 2.6 Pro-inflammatory cytokines are increased in Pip4k2c-/- mice 24
Figure 2.7 T cell derived IFNγ and IL-17 levels are elevated in Pip4k2c- 25
/- mice
Figure 2.8 Plasma IgG3 levels are elevated in Pip4k2c-/- mice 26
Figure 2.9 Pip4k2c-/- mice exhibit an increase in CD44+ active T cells 27
and a decrease in CD62L+ naïve T cells
Figure 2.10 CD4+ and CD8+ T cells are elevated in the spleen of 27
Pip4k2c-/- mice
Figure 2.11 T cells from Pip4k2c-/- mouse have enhanced growth rates 28
Figure 2.12 Regulatory T cells are suppressed in Pip4k2c-/- mice 28
Figure 2.13 Signaling downstream of mTORC1 is upregulated in various 30
ix tissues of Pip4k2c-/- mice
Figure 2.14 Signaling downstream of mTORC1 is upregulated in the 31
spleen of Pip4k2c-/- mice
Figure 2.15 Rapamycin reduces mTORC1 signaling in Pip4k2c-/- mice 33
Figure 2.16 Changes in plasma cytokine levels after treatment with 34
rapamycin
Figure 2.17 Changes in immune cell infiltration after treatment with 35
rapamycin
Figure 2.18 Model for the role of PI5P4Kγ in regulation of immunity 36
Figure 3.1 Structures of PIP4Kin1 and PIP4Kin2 50
Figure 3.2 PIP4Kin1 and PIP4Kin2 inhibit the kinase activity of 51
PI5P4Ks
Figure 3.3 PIP4Kin1 and PIP4Kin2 inhibit proliferation of BT474 cells 52
but not MCF7 cells
Figure 3.4 PIP4Kin1 and PIP4Kin2 inhibit proliferation of BT474 cells 53
even after washout
Figure 3.5 Reducing the temperature of the culture media to 32 oC, 55
where mutant p53 in BT474 cells is partially functional,
partially rescues PIP4Kin1/2 mediated inhibition of cell
growth
Figure 3.6 PIP4Kin1 enhances Insulin-dependent activation of Akt 56
Figure 3.7 A biotin-PIP4Kin1 pull down assay indicates that PIP4Kin1 60
x and PIP4Kin2 are covalent inhibitors of the PI5P4K
enzymes
Figure 3.8 Mapping PI5P4Kβ cysteine residues targeted by PIP4Kin1 62
Figure 3.9 Co-crystal of PI5P4Kβ and PIP4Kin2 63-65
Figure 3.10 The half-life of PIP4Kin2 in serum following a single i.v. 66
injection is 9.63 hours
Figure 3.11 Maximum tolerance dose test with PIP4Kin2 67
Figure 3.12 PIP4Kin2 has significant antitumor activity in an orthotopic 69-70
BT474 xenograft model
Figure 3.13 In vivo target engagement of PIP4Kin2 72
Figure 4.1 Overexpressing PI5P4Kβ mutant (C307S C318S) partially 89-90
rescues the impaired growth of BT474 cells by PIP4Kin2
xi List of Abbreviations
PI5P4K Type 2 phosphatidylinositol-5-phosphate 4-kinase mTOR Mammalian target of rapamycin mTORC1 Mammalian target of rapamycin complex 1 p53 Tumor protein p53
Akt Protein kinase B
PI3K Phosphoinositide 3-kinase
PIKFYVE FYVE finger-containing phosphoinositide 5-kinase
PI4P5K Type 1 phosphatidylinositol-4-phosphate 5-kinase p70-S6K p70- ribosomal protein S6 kinase
SREBP1 Sterol regulatory element-binding protein 1
PI-5-P Phosphatidylinositol-5-phosphate
PI-4-P Phosphatidylinositol-4-phosphate
PI-4,5-P2 Phosphatidylinositol-4,5-bisphosphate
SNP Single nucleotide polymorphism
Th Helper T cell
Treg Regulatory T cell i.v. Intravenous p.o. Oral gavage i.p. Intraperitoneal
xii Chapter 1. Introduction
In 1997, Rameh et al. discovered that a family of enzymes that was thought to be phosphatidylinositol-4-phosphate 5-kinases (PI4P5Ks) were actually phosphatidylinositol-5-phosphate 4-kinases (PI5P4Ks). In other words, rather than producing phosphatidylinositol-4,5-P2 (PI-4,5-P2) from phosphatidylinositol-4- phosphate (PI-4-P), as had been thought for the previous 8 years, these enzymes were utilizing an impurity in bovine brain-derived PI-4-P, phosphatidylinositol-5- phosphate (PI-5-P) as a substrate to generate PI-4,5-P2 (Figure 1.1). Prior to this discovery, PI-5-P was not known to exist in vivo. In humans, three isoforms of
PI5P4Ks exist, PI5P4Kα, PI5P4Kβ and PI5P4Kγ, and these enzymes are encoded by three different genes, PIP4K2A, PIP4K2B and PIP4K2C respectively (named
Pip4k2a, Pip4k2b and Pip4k2c in mice).
While PI4P5Ks are conserved in yeast and other unicellular eukaryotes, the
PI5P4Ks appear to have evolved later from the PI4P5Ks and have only been observed in multicellular eukaryotes, from worms and flies to mammals. In mammalian cells, the substrate for the PI5P4Ks, PI-5-P, is only about 1% as abundant as the substrate for the PI4P5Ks, PI-4-P, and the vast majority of PI-4,5-
P2 is generated from PI-4-P. However, as discussed in this thesis, there is evidence that PI5P4Ks play a significant role in generating certain intracellular pools of PI-4,5-P2, in particular when cells are under stress.
1
Figure 1.1: Compartmentalization of phosphoinositide pathways
Seven different forms of phosphorylated phosphatidylinositol have been identified in mammalian cells. The most abundant species are PI-4-P and PI-4,5-P2. PI-4-P is particularly abundant in the Golgi and is transported to the plasma membrane where it is converted to PI-4,5-P2 by a family of PI4P5Ks. Although the majority of
PI-4,5-P2 is found at the plasma membrane, PI-4-P can also be converted to PI-4,5-
P2 at intracellular locations, including the lysosome, autophagosome and nucleus. The PI5P4K family of enzymes provide an alternative pathway for generating PI-
4,5-P2 at intracellular locations. The substrate for these enzymes, PI-5-P is only about 1% as abundant as PI-4-P and only a small fraction of cellular PI-4,5-P2 is generated by this pathway. The enzyme PIKFYVE, which is localized in intracellular compartments through interaction with PI-3-P, can generate PI-5-P directly from PI but also generates PI-3,5-P2 from PI-3-P. PI-3,5-P2 is converted to PI-5-P by the myotubularin family of phosphatases. Finally, a family of PI3Ks place phosphate at the 3 position of PI. The class I PI3Ks convert PI-4,5-P2 to PI-3,4,5-P3 at the plasma membrane while class II and III PI3Ks generate PI-3-P and PI-3,4-P2, mostly at intracellular membranes. Dozens of proteins involved in actin remodeling, vesicle trafficking, transmembrane transport, autophagy and signal transduction have been shown to be regulated by direct binding to specific isoforms of these various phosphoinositides (Sun et al., 2013 and Fiume et al., 2015).
2
Figure 1.1 (Continued)
2+ Actin, IP3-Ca , DAG-PKC Arf, Rac-PAK, AKT
PIP5K1A/B/C PIK3CA PI-3,4,5-P Plasma PI-4-P PI-4,5-P2 3 Membrane PTEN PI4K
ER, Golgi PI PI-4-P PIP5K1A
Late Endosome PIP4K2A/B Lysosome PI-5-P PI-4,5-P2 Autophagasome Nucleus Myotubularins
PIKFYVE
Endosome PIK3C3 PIKFYVE Lysosome PI PI-3-P PI-3,5-P2
3 The biological roles of PI5P4Ks have been elusive (Fiume et al., 2015). All three isoforms of PI5P4Ks are highly expressed in the brain while their relative expression levels differ in other tissues. For example, PI5P4Kα is highly expressed in spleen and peripheral blood cells and PI5P4Kβ is highly expressed in muscle, heart and liver tissues. In contrast, PI5P4Kγ is highly expressed in kidney and testis
(Clarke et al., 2008). As discussed in more detail below, mice with germline deletion of PIP4K2A or PIP4K2B are viable and healthy with a normal lifespan, though deletion of both genes results in perinatal lethality, suggesting some redundancy in function.
PI5P4Kβ has a nuclear localization sequence and it has been detected at the plasma membrane, in the nucleus as well as in the cytosol (Ciruela et al., 2000).
Recently it was reported that nuclear PI-5-P levels are suppressed by PI5P4Kβ and that this plays a role in control of the expression of myogenic genes during myoblast differentiation (Stijf-Bultsma et al., 2015). PI5P4Kα is mostly found in the cytosol and previous studies have suggested that heterodimerization with
PI5P4Kβ allows nuclear localization of PI5P4Kα (Bultsma et al., 2010). PI5P4Kγ is mainly detected in the cytosol and endomembrane compartments, including the ER,
Golgi and unidentified vesicular compartments (Clarke, 2009).
PI5P4Kγ is the least studied among the three isoforms since it was thought that PI5P4Kγ had no kinase activity in vitro (Clarke et al., 2008). However, the same group recently reported a low level of catalytic activity (Clarke et al., 2015) and unpublished studies from the Cantley laboratory indicate that PI5P4Kγ produced in
E. coli is active, but only about 1% as active as the other isoforms. While
4 PI5P4Kγ is far less active than PI5P4Kα and PI5P4Kβ it forms heterodimers with these isoforms and may play a role in localizing or regulating their activities.
Consistent with this model, it was shown that PI-5-P levels increased more significantly in PI5P4Kγ knockdown cells compared to PI5P4Kα or PI5P4Kβ - knockdown cells (Sarkes and Rameh, 2010).
An increasing number of papers have been suggesting that PI5P4Ks might play a role in nucleus. For example, UV irradiation resulted in an inhibitory phosphorylation of PI5P4Kβ on Ser326 by p38, which was responsible for an increase of nuclear PI-5-P (Jones et al., 2006). Gozani et al. reported that a decrease in PI-5-P by overexpressing PI5P4Kβ reduced the association of ING2 with chromatin and nuclear matrix (Gozani et al., 2003). They argued that the binding of PI-5-P to the PHD fingers of ING2 was required to activate p53 by ING2
(Gozani et al., 2003).
Like ING2, many proteins containing a PHD domain have been reported to be present in nucleus (Bienz, 2006). In some cases, the interaction of PI-5-P with these proteins affected their activities. Recently, it was reported that the binding of
PI-5-P to TAF3 regulated gene expression by changing the interaction of TAF3 to
H3K4me3 (Stijf-Bultsma et al., 2015). In another report, binding of PI-5-P to
UHRF1, ubiquitin like and ring finger containing protein, induced conformational rearrangement of UHRF1 and regulated the interaction of UHRF1 to histones
(Gelato et al., 2014). These papers are indicating that PI-5-P is involved in regulation of transcription.
5 Additionally, PI-4,5-P2 also has been suggested to play a role in the regulation of transcription in nucleus. For example, PI-4,5-P2 interacted with histone
H1 and reversed the repression of transcription (Yu et al., 1998). In another case,
PI-4,5-P2 impacted on RNA polyadenylation as the binding of PI-4,5-P2 to a non canonical poly (A) polymerase, Star-PAP stimulated its activity (Mellman et al.,
2008). Also, it was reported that incubation of the lymphocyte nuclei with PI-4,5-P2 targeted the chromatin remodeling complex BAF to chromatin (Zhao et al., 1998).
In addition to PI5P4Kα and PI5P4Kβ, we found that PI5P4Kγ localized in the nucleus as well (unpublished). We also found that many nuclear proteins co- precipitated with tagged PI5P4Kγ (unpublished). Thus, it might be intriguing to identify the role of PI5P4Ks in nucleus, suggestively in regulation of transcription.
Interestingly, PI5P4Kγ has been linked to the mammalian target of rapamycin
(mTOR) signaling complex and to cellular immunity. Mackey et al. argued that
PI5P4Kγ was negatively regulated by mTORC1 through direct phosphorylation at serine 324 (Ser324) and serine 328 (Ser328) (Mackey et al., 2014). On the other hand, it was reported that knocking out the only PI5P4K isoform in Drosophila resulted in lower body weight and that this correlated with decreased mTORC1 signaling (Gupta et al., 2013). Of particular interest, multiple studies have shown a link between a SNP in the PIP4K2C locus and familial autoimmunity (Raychaudhuri et al., 2008 and Fung et al., 2009).
The immune system is a complex network that evolved to protect organisms from invasion of various microbes. While an active immune system protects from microbes, overactivation of the immune system can result in autoimmunity. In
6 autoimmune diseases, the immune system attacks normal, uninfected tissues. An understanding of the molecular underpinnings of an overactive or underactive immune system is important for developing therapies for immune related diseases from immunodeficiency to autoimmunity and even cancers.
Mammalian target of rapamycin complex 1 (mTORC1) is a central regulator of cell survival, growth and metabolism and plays a critical role in regulation of immune cells. mTORC1 responds to intra- and extracellular signals such as growth factors, oxygen levels, energy status and amino acid levels (Laplante and Sabatini,
2013). The mTORC1 complex consists of mTOR, a serine-threonine protein kinase, regulatory associated protein of mTOR (Raptor), mammalian lethal with Sec13 protein 8 (mLST8), proline-rich Akt substrate 40 kDa (PRAS40) and DEP-domain- containing mTOR-interacting protein (Deptor). In general, mTORC1 positively regulates protein synthesis and lipid synthesis while it negatively regulates autophagy. The mTOR pathway is activated during various cellular processes, including T cell activation, insulin resistance and tumor formation.
As the name, mammalian target of rapamycin complex 1 indicates, the immunosuppressive drug rapamycin directly binds to the mTORC1 complex to suppress immune responses. Rapamycin has been used in the clinic to prevent transplant rejection. One of the identified roles of mTORC1 in the immune system is to direct T-cell fate decisions. mTORC1 positively regulates differentiation of the
Th1 and Th17 subset of Th cells (Delgoffe et al., 2011). On the other hand, mTORC1 is a negative regulator of Treg differentiation (Delgoffe et al., 2009) and at the same time required to maintain Treg function (Zheng et al., 2013).
7 To assess the physiological role of PI5P4Kγ I generated Pip4k2c-/- mice
(Chapter 2). Pip4k2c-/- mice were viable with a normal lifespan and didn’t show any specific abnormality until they were older than 8 months. But among the older mice at ages between 8 months and 14 months, Pip4k2c-/- mice displayed increased immune infiltrates in various tissues, including liver, intestine, kidney and lungs.
Moreover, plasma of Pip4k2c-/- mice contained high levels of pro-inflammatory cytokines such as interferon γ (IFNγ), interleukin 12 (IL-12) and interleukin 2 (IL-2).
Interestingly, the splenic Th population of immune cells increased while the Treg population decreased in Pip4k2c-/- mice. Thus, we looked at the mTORC1 signaling pathway to investigate if the inflammatory phenotype of Pip4k2c-/- mice correlated with the mTORC1 signaling pathway. We found that mTORC1 downstream components p70-S6K and SREBP1 were activated in Pip4k2c-/- mouse tissues, including liver, kidney, brain and muscle. Also, the pro-inflammatory cytokine levels that were high in the Pip4k2c-/- mouse plasma decreased after 2 weeks of rapamycin treatment. These results suggested that increased activation of mTORC1 signaling, with activation of Th cells in Pip4k2c-/- mice could be responsible for the chronic inflammation in these mice. These results are in agreement with the correlation between a SNP in the PIP4K2C locus and familial autoimmunity in humans and suggest that the SNP may suppress expression of
PI5P4Kγ protein. Our current model (Figure 1.2) is that loss of PI5P4Kγ results in activation of mTORC1 signaling and that this leads to upregulation of Th cells, resulting in increased inflammation.
8 Deletions of the other PI5P4K genes in mice have also been informative about physiological function. Both in vivo and ex vivo studies have indicated a role for PI5P4Kβ in suppression of insulin signaling. Overexpression of PI5P4Kβ in
CHO-IR cells resulted in a decrease in insulin-induced PIP3 levels and a significant decrease in Akt phosphorylation on threonine 308 (Thr308), but had no effect on insulin receptor autophosphorylation or insulin receptor substrate 1 (IRS1) phosphorylation (Carricaburu et al., 2003). Furthermore, insulin signaling was enhanced in skeletal muscle of Pip4k2b knockout mice as judged by enhanced clearance of glucose in a glucose tolerance test, suppression of serum insulin and enhanced Akt activation in muscle in response to insulin injection (Lamia et al.,
2004). Results in Chapter 3 with PI5P4K inhibitors support the role of these enzymes in suppressing insulin signaling. The Pip4k2b knockout mice are more lean than wild type mice, do not gain as much weight on a high fat diet, maintain insulin sensitivity on a high fat diet and live a normal lifespan suggesting that a
PI5P4Kβ inhibitor would be beneficial for treating obesity, insulin resistance and type 2 diabetes.
9 Pip4k2c+/+ mice Pip4k2c-/- mice
PI5P4Kγ PI5P4KXγ ?" ?"
PI5P4Kα/β PI5P4Kα/β ?" ?" ?"
mTORC1 mTORC1
S6K SREBP1 S6K SREBP1
Treg Th Treg Th
inflammation inflammation
Figure 1.2: Working Model for the role of PI5P4Kγ in immune regulation The mouse gene deletion studies presented in this thesis indicate that PI5P4Kγ negatively regulates the mTORC1 pathway, which is important for the differentiation and activation of Th cells. In Pip4k2c-/- mice, the mTORC1 complex is hyper- activated and T cells are also hyper-activated, leading to immune cell infiltration in multiple organs. Flies have only a single gene for PI5P4K and deletion of this gene results in decreased mTORC1 signaling (Gupta et al., 2012). Since PI5P4Kγ has very low PI5P 4-kinase activity compared to PI5P4Kα or PI5P4Kβ and can heterodimerize with PI5P4Kα and PI5P4Kβ (Bultsma et al., 2010, Rao et al., 1998, Burden et al., 1999, Wang et al., 2010). These data suggest that PI5P4Kγ either competes with PI5P4Kα and PI5P4Kβ for binding at a location critical for PI-4,5-P2 dependent activation of mTORC1 or that heterodimerization of PI5P4Kγ with PI5P4Kα and PI5P4Kβ suppresses their activities.
10 PI5P4Kβ is also a potentially interesting target for cancer therapy. As discussed in more detail in Chapter 3, we recently found that germline deletion of both alleles of Pip4K2b and both alleles of the tumor suppressor Trp53 results in early embryonic death (Emerling et al., 2013, also see Appendix), indicating a synthetic lethality between loss of PI5P4Kβ and loss of p53 protein. Since genetic aberrations in TP53 are the most frequent event in human cancers, these studies suggest that a PI5P4Kβ inhibitor would also be a useful therapeutic for treating cancers.
The physiological role of PI5P4Kα is less clear. While several studies have linked SNPs in PI5P4Kα to schizophrenia (Schwab and Wildenauer, 2009) and suggested a critical role for PI5P4Kα in platelet formation (Schulze et al., 2006), germline deletion of Pip4k2a results in no obvious phenotype (Emerling et al.,
2013). Our genetic studies indicate some redundancy in PI5P4Kα and PI5P4Kβ function with PI5P4Kβ being the more critical gene. In contrast to the synthetic lethality observed upon deletion of Pip4k2b and Trp53, deletion of both alleles of
Pip4k2a and both alleles of Trp53 results in viable mice that develop cancers at similar rates to mice with Trp53 deletion (Emerling et al., 2013). However, we found that Trp53-/-Pip4k2a-/-Pip4k2b+/- mice had dramatically longer tumor-free survival compared to Trp53-/- Pip4k2a+/+Pip4k2b+/+ mice or Trp53-/- Pip4k2a+/+Pip4k2b+/- mice. These results suggest that inhibition of both PI5P4Kα and PI5P4Kβ may be necessary to suppress tumor formation or treat existing tumors with genetic aberrations in TP53. Consistent with this idea, it was necessary to knockdown both
11 PI5P4Kα and PI5P4Kβ to suppress the growth of the human TP53 mutant breast cancer cell line, BT474 (Emerling et al., 2013).
In addition there are possible links between PI5P4Ks and autophagosome function. It was recently suggested that PI-5-P synthesis by the phosphatidylinositol 5-kinase (PIKFYVE) was required for autophagosome biogenesis (Vicinanza et al., 2015). This study also showed that knocking down each PI5P4K isoform increased autophagosome formation with increased levels of
LC3-II in the presence of Bafilomycin A1 that blocks LC3-II clearance by inhibiting autophagosome-lysosome fusion. Also, overexpression of each PI5P4K isoform impaired autophagy while a PI5P4kγ-inactive mutant did not.
Based on these genetic studies we decided to focus on the development of small molecules that target PI5P4K isoforms. In a collaboration with Dr. Nathanael
Gray’s laboratory we explored compounds that had been developed to target ATP sites on protein kinases but that cross-reacted with members of the PI5P4K family of enzymes. We selected the most potent compounds in this series, including compounds that we named PIP4Kin1 and PIP4Kin2 for further studies to investigate their effects in vivo. We found that those compounds completely recapitulated the knockdown of PI5P4Kα/β in cells. PIP4Kin1 and PIP4Kin2 inhibited the growth of
TP53 mutant BT474 cells without inhibiting the growth of TP53 wild type MCF7 cells at 1 µM concentration. The growth inhibition of BT474 cells was partially rescued by culturing the cells at 32o C, the temperature at which the mutant p53 protein regains partial function. Moreover, we obtained evidence that PIP4Kin1 and PIP4Kin2, each
12 of which have acrylamide moieties, make covalent bonds with PI5P4Ks at conserved cysteine residues.
To further validate the efficacy of PIP4Kin2, we evaluated the effect of
PIP4Kin2 on mouse xenografts with BT474 and MCF7 cells. Daily i.p. injection of the mice with PIP4Kin2 decreased the growth of TP53 mutant BT474 tumors but not TP53 wild type MCF7 tumors, consistent with previously published data that knockdown of PI5P4Kα and PI5P4Kβ blocked the growth in BT474 xenografts but not MCF7 xenografts (Emerling et al., 2013).
13 Chapter 2 . Deletion of a Novel Phosphatidylinositol Kinase results in Hyperactivation of the Immune System
Abstract:
Mammals have three enzymes that generate the lipid phosphatidylinositol-
4,5-bisphosphate (PI-4,5-P2) from phosphatidylinositol-5-phosphate (PI-5-P):
PI5P4Kα, PI5P4Kβ and PI5P4Kγ. These enzymes and their substrate, PI-5-P have recently been implicated in a variety of stress responses in cell lines. Germline deletion studies in mice have also revealed a role for PI5P4Kβ in obesity and diabetes and a role for both PI5P4Kα and PI5P4Kβ in the growth of TP53-/- tumors.
The physiological role of PI5P4Kγ has been elusive, although recent studies have correlated a SNP in the gene encoding this enzyme with familial autoimmune disease. Here we show that mice with germline deletion of Pip4k2c, the gene encoding PI5P4Kγ, appear normal in regard to growth and viability but have increased inflammation and T cell activation as they age. Immune cell infiltrates increased in Pip4k2c-/- mouse tissues. Also, there was an increase in pro- inflammatory cytokines, including interferon γ, interleukin 12 and interleukin 2 in plasma of Pip4k2c-/- mice. Pip4k2c-/- mice had an increase in T helper cell populations and a decrease in regulatory T cell populations with increased proliferation of T cells. Interestingly, mTORC1 signaling was hyperactivated in several tissues from Pip4k2c-/- mice and treating Pip4k2c-/- mice with rapamycin reduced the inflammatory phenotype resulting in a decrease in mTORC1 signaling in tissues and a decrease in pro-inflammatory cytokines in plasma. These results indicate that PI5P4Kγ plays a role in T cell activation by regulating mTORC1
14 signaling and suggest that the SNP at the PIP4K2C locus in human patients with autoimmunity might cause a decrease in PI5P4Kγ expression and thereby an increase in mTORC1 signaling. These results suggest that inhibition of mTORC1 would be beneficial to these patients. These studies also suggest that agents that inhibit PIP4K2C function could be useful to enhance cancer immunotherapy.
15 Introduction:
Phosphatidylinositol-5-phosphate 4-kinase (PI5P4K) phosphorylates the 4- position of phosphatidylinositol-5-phosphate (PI-5-P) producing phosphatidylinositol-4,5-bisphosphate (PI-4,5-P2). PI5P4K was first identified in
1997 as an alternate route for the generation of PI-4,5-P2 (Rameh et al.,1997). The greater part of PI-4,5-P2 is generated by the canonical pathway in which type I phosphatidylinositol-4-phosphate 5-kinase (PI4P5K) converts phosphatidylinositol-
4-phosphate (PI-4-P) to PI-4,5-P2.
All three isoforms are highly expressed in brain while their relative expressions in other tissues vary. PI5P4Kα is highly expressed in spleen and the peripheral blood. PI5P4Kβ is highly expressed in muscle while PI5P4Kγ is highly expressed in kidney (Castellino et al., 1997). In kidney, PI5P4Kγ is mostly detected in the cortex and the outer medulla (Clarke et al., 2008). All tissues examined appear to express at least one isoform of PI5P4K.
At the cellular level, PI5P4Ks are found in several organelles, including plasma membrane, cytosol, nucleus and vesicular compartments (Ciruela et al.,
2000 and Clarke et al., 2009). It is not simple to define unique compartmentalizations of the individual enzymes because all three isoforms can homodimerize or heterodimerize with each other.
At the sequence level, PI5P4Kα and PI5P4Kβ are more homologous to each other than either is to PI5P4Kγ. Also, PI5P4Kγ is only about 1% as active as the other isoforms, raising the possibility that its major role may be to localize or regulate the activities of PI5P4Kα and PI5P4Kβ.
16 Germline deletion of both alleles of Pip4k2a or Pip4k2b in mice results in healthy mice that live a normal life span. The Pip4k2b-/- mice have increased insulin sensitivity and are protected from obesity, insulin resistance and type 2 diabetes when placed on a high fat diet (Lamia et al., 2004). Also, crossing the
Pip4k2b-/- mice with Trp53-/- mice results in early embryonic lethality for the subset of embryos that are Pip4k2b-/-, Trp53-/-, indicating a synthetic lethality relationship between these genes. In contrast, Pip4k2a-/- mice do not exhibit any of the phenotypes observed in the Pip4k2b-/- mice (Emerling et al., 2013). They are not protected from obesity or insulin resistance, do not exhibit a synthetic lethality relationship with Trp53 and in all ways examined, they resemble wild type mice
(unpublished data and Emerling et al., 2013). However, germline deletion of one allele of Pip4k2a in the context of germline Pip4k2b-/- causes enhancement of the phenotypes of the Pip4k2b-/- mice. Deletion of both alleles of Pip4k2a and Pip4k2b did not have any observable effect on embryonic development up till the time of birth, but resulted in perinatal lethality of all pups. These results indicate that these genes do not play a major role in embryonic development and that Pip4k2a provides a backup for Pip4k2b that becomes critical at the time of birth.
Here I present the first characterization of mice with germline deletion of
Pip4k2c. Surprisingly, Pip4k2c-/- mice exhibit a phenotype that is quite different from Pip4k2a-/- or Pip4k2b-/- mice. These mice appear to develop normally and are not protected from obesity, insulin resistance or diabetes, but rather develop enhanced immune responses, resulting in autoimmunity. In addition, they exhibit hyperactivation of the mammalian target of rapamycin (mTor) signaling complex
17 (mTORC1) in multiple tissues, indicating that Pip4k2c negatively regulates mTORC1. These results, along with a recent observation that the enzyme encoded by this gene is a substrate of mTORC1 (Mackey et al., 2014), suggest a close relationship between mTORC1 and Pip4k2c. The hyper-immune phenotype of the
Pip4k2c-/- mice could be partially ameliorated by treatment with the mTORC1 inhibitor, rapamycin. Importantly, a SNP located near the PIP4K2C locus has been correlated with familial autoimmunity (Raychaudhuri et al., 2008) and the results presented here suggest that loss of PIP4K2C function could explain this disease.
18 Results:
Generation of Pip4k2c-/- mice
To investigate the role of PI5P4Kγ in mammals, PI5P4Kγ knockout (Pip4k2c -
/-) mice were generated using Pip4k2c-targeted ES cell clones obtained from the knockout mouse project repository.
Wild type locus
pwtF pwtR
Ex Ex Ex Ex : loxP : FRT 2 3 4 5
Cassette inserted locus
Ex En2 IR β βact: Ex Ex Ex pA pA 2 SA ES gal :neo 3 4 5
X CMV-Cre
Knockout locus
Ex En2 IR β Ex pA 2 SA ES gal 5 pkoF pkoR
Figure 2.1: Generation of Pip4k2c-/- mice Schematic representation of the Pip4k2c locus before and after deletion of the critical exons (exon 3 and exon 4). The deletion of the critical exons was induced by Cre-lox recombination (Skarnes et al., 2011). Ex, exon; En2SA, Engrailed 2 gene splice acceptor sequence; β gal, β-galactosidase; pA, polyA signal; neo, neomycin.
19
A Mouse tail DNA B Mouse kidney +/+ +/- -/- +/+ +/+ +/+ -/- -/- -/- WT (pwtF, pwtR) PI5P4Kγ
KO β−actin (pkoF, pkoR)
Figure 2.2: Validation of Pip4k2c-/- mice (A) PCR analysis of genomic DNA prepared from mouse tail. Primers used for genotyping by PCR are pwtF, pwtR, pkoF and pkoR. (B) Western blotting of proteins prepared from mouse kidney.
In the targeting construct, the exons 3 and 4 of Pip4k2c were bracketed by loxP sequences, so deletion of these exons could remove much of the kinase domain and place the protein out of the correct reading frame (Figure 2.1). ES cell clones were karyotyped and selected for microinjection. Since the targeted clones were
C57BL/6N (agouti)-derived JM8A3.N1, the cells were injected into C57BL/6J (black) blastocysts. Three chimeric mice were born and they were backcrossed with
C57BL/6 mice for three generations. Then whole body knockout mice were obtained by crossing the transgenic mice with B6.C-Tg(CMV-cre)1Cgn/J. The
Pip4k2c +/- mice were crossed with Pip4k2c +/- mice to get Pip4k2c+/+ (wild type, WT) mice and Pip4k2c-/- (knockout, KO) mice to use for the further experiments. The wild type allele and knockout allele of each mouse was confirmed by genomic DNA PCR
(Figure 2.2).
20 Increased inflammation in Pip4k2c-/- mice
Pip4k2c-/- mice bred normally and grew into adulthood displaying no obvious growth or behavioral abnormalities. Unlike Pip4k2b-/- mice, these mice did not have enhanced insulin sensitivity or and were not protected from obesity on a high fat diet (Figure S2.4 and S2.5).
Since a SNP near the human PIP4K2C locus has been linked to autoimmunity, we examined whether Pip4k2c-/- mice exhibit any inflammatory phenotype. We carried out a complete necropsy of Pip4k2c-/- mice at different ages.
We found that immune cells formed clusters in the organs of mature Pip4k2c-/- mice
(8-14 months of age). The immune cell infiltration was observed in the liver, kidney, intestines and lungs of the mice (Figure 2.3). To measure the surface area of infiltrating immune cells in the liver, we paraformaldehyde fixed and paraffin- embedded liver tissues and performed H&E staining. The ratio of immune infiltrates per total area was significantly increased in the Pip4k2c-/- mice (Figure 2.4), which indicated that the Pip4k2c-/- mice developed chronic inflammation without a specific trigger such as infection or injuries. To identify which type of immune cells infiltrated the organs, we stained the liver tissue sections with anti-CD3, anti-B220 and anti-
Mac-2 antibodies. The infiltrating immune cells in the Pip4k2c -/- livers consisted of mostly T cells and B cells (Figure 2.5).
Moreover, the plasma levels of various pro-inflammatory cytokines increased in the Pip4k2c-/- mice, including the Th1-type cytokines interferon γ (IFNγ), interleukin 12 (IL-12) and interleukin 2 (IL-2) (Figure 2.6).
21 Mouse Liver WT KO
Mouse Kidney WT KO
Figure 2.3: Immune cell infiltration in the liver of Pip4k2c-/- mice
H&E staining readily identifies immune infiltrates in the liver tissue (left) and kidney tissue (right) of 12-14 months old Pip4k2c-/- mice. Scale bars, 100um.
22 *
(%) 2
1.5
1
0.5 totalarea
Immuneinfiltrates 0 WT KO
Figure 2.4: Immune cell infiltration in the liver of Pip4k2c-/- mice
The area of immune infiltrates of H&E stained liver tissues were quantified using ImageJ software. 25 Pip4k2c-/- mice (12-14 months old) and age matched 12 wild type mice, 4 images per mouse were examined. The Y axis indicates the ratio of the area of immune infiltrates over total area of the tissue in each image. The results are presented as means ± standard errors of the means. *p < 0.05
CD3 B220 MAC-2
KO1
KO2
Figure 2.5: Immune infiltrates in liver tissue of Pip4k2c-/- mice are mostly T cells and B cells
The Pip4k2c-/- liver sections of 2 animals (KO1 and KO2, 12 months old) were incubated with anti-CD3, B220 and Mac2 antibodies, respectively.
23 IL-15 IL-7 Eotaxin LIF KO MIG WT IL-1a G-CSF LIX IP-10 MIP-2 KC TNFa IL-1B IL-13 M-CSF (67) IL-6 MIP-1a (64) VEGF (76) IL-4 GM-CSF IL-10 IL-5 MCP-1 MIP-1B IL-3 RANTES IL-17 IL-9 IL-2 IL-12 (p70) IL-12 (p40) IFNγ 0.1 1 10 100 1000 10000 pg/ml (log scale)
Figure 2.6: Pro-inflammatory cytokines are increased in Pip4k2c-/- mice Plasma cytokines were detected using multiplex cytokine ELISA. The experiments were performed on >10 mice (12-14 months old) per each group, with 2 measurements per mouse. The Y axis is in logarithmic scale. The results are presented as means ± standard errors of the means.
24 Additionally, IL-17 and IFNγ secreted by CD4+ T cells isolated from spleen were higher in the Pip4k2c-/- mice (Figure 2.7). As T cell cytokines can affect immunoglobulin class switching, we performed immunoglobulin isotyping. It is known that IL-17 drives B cells to undergo preferential isotype class switching to
IgG3 and IgG2a (Mitsdoerffer et al., 2010) We found that the level of IgG3 increased in the plasma of Pip4k2c-/- mice, which agreed with the increase in IL-17 levels (Figure 2.8).
A
WT KO
γ IFN
IL-17
B /ml) /ml) ng ng ( γ IFN IL-17(
WT KO WT KO
Figure 2.7: T cell derived IFNγ and IL-17 levels are elevated in Pip4k2c-/- mice Flow cytometry of IL-17 and IFNγ secretion by CD4+ T cells isolated from spleen indicated groups. The data are representative of three independent experiments with n > 3 mice (12-14 months old) each group. *p < 0.05 (Student’s t test, error bars represent standard deviation).
25 2.5 *
2 WT KO 1.5
1
0.5 Absorbance450nmat 0 IgG1 IgG2a IgG2b IgG3 IgA IgM
Figure 2.8: Plasma IgG3 levels are elevated in Pip4k2c-/- mice Plasma Ig levels were measured using ELISA. 3 Pip4k2c-/- mice (12 months old) and age matched 4 wild type mice were examined. The results are presented as means ± standard errors of the means. *p < 0.05
T cells are hyperactivated in Pip4k2c-/- mice
Flow cytometry was utilized to determine the activity of T cells by counting
CD44+ and CD62L+ T cells. We found that Pip4k2c-/- mice have more CD44+ activated T cells and fewer CD62L+ naïve T cells compared to the wild type mice
(Figure 2.9). These data suggested that T cells were more activated in Pip4k2c-/- mice. Also, Pip4k2c-/- mice showed increased CD4+ and CD8+ populations, indicating an increase of Th cells and cytotoxic T cells (Figure 2.10).
To investigate proliferation of T cells, we isolated CD4+ T cells from the spleens from Pip4k2c-/- mice and wild type mice. The cells were seeded in 96 well plates coated with anti-CD3 and anti-CD28. After 48 hours of culture, 3H-thymidine was added for another 16 hours before measuring the 3H-thymidine incorporation
(Figure 2.11). The Pip4k2c-/- T cells showed significantly higher proliferation rates than T cells from wild type mice. In accordance with increased T cell activation in
26 Pip4k2c-/- mice, the number of Foxp3+CD4+ Treg cells decreased in Pip4k2c-/- mice
(Figure 2.12).
WT KO
CD62L
CD44
Figure 2.9: Pip4k2c-/- mice exhibit an increase in CD44+ active T cells and a decrease in CD62L+ naïve T cells Flow cytometry of CD44+ and CD62L+ T cells isolated from spleen indicated groups. The most representative data from three independent experiments are given, with n > 3 mice (12-14 months old) from each group.
WT KO
CD4
CD8
Figure 2.10: CD4+ and CD8+ T cells are elevated in the spleen of Pip4k2c-/- mice Flow cytometry of CD4+ and CD8+ T cells isolated from spleen indicated groups. The most representative data from three independent experiments are given, with n > 3 mice (12-14 months old) from each group.
27 80000 WT# WT ) * KO 60000 KO
c.p.m * * 40000 H Uptake (c.p.m.) HUptake (
3 20000 3
0 10 3 1 0.3 0.1 Anti-CD3anti-CD3 ( u(µg/ml)g/ml)
Figure 2.11: T cells from Pip4k2c-/- mouse have enhanced growth rates CD4+ T cells were isolated from spleens of 12 months old mice and cultured in media containing 3H-thymidine. The level of radioactivity was measured by liquid scintillation. The data are presented as mean 3H-thymidine incorporation (cpm ± SEM, performed in triplicate). *p < 0.05
WT KO
Foxp3
CD4
Figure 2.12: Regulatory T cells are suppressed in Pip4k2c-/- mice Flow cytometry of Foxp3+ and CD4+ T cells isolated from spleen indicated groups. The most representative data from three independent experiments are given, with n > 3 mice (12-14 months old) from each group.
28 Increased mTORC1 signaling in Pip4k2c-/- mice
Since there was an increase in Th cell population and a decrease in Treg cell population in Pip4k2c-/- mice, we investigated if mTORC1 signaling, which can regulate T cell differentiation and activation, was changed in Pip4k2c-/- mice. We found that phosphorylation of p70-S6K on threonine 389 (a direct substrate of mTORC1) was increased in kidney, liver, brain and muscle tissues from Pip4k2c-/- mice compared to wild type mice (Figure 2.13). Phosphorylation of threonine
(Thr389) of p70-S6K was also enhanced in spleen, the major immune system organ
(Figure 2.14). Levels of mature SREBP1 have recently been shown to be a downstream reporter for mTORC1 activity (Duvel et al., 2010). Consistent with the increased p70-S6K phosphorylation, we also found that levels of mature SREBP1 were significantly higher in various tissues from the Pip4k2c-/- mice (Figure. 2.13 and 2.14).
29 Kidney Liver +/+ +/+ +/+ -/- -/- -/- +/+ +/+ +/+ -/- -/- -/- p-p70-S6K (T389)
p70-S6K
SREBP1 (mature)
Brain Muscle +/+ +/+ +/+ -/- -/- -/- -/- -/- -/- p-p70-S6K +/+ +/+ +/+ (T389)
p70-S6K
SREBP1 (mature)
P-p70-S6K / total p70-S6K p = 0.128 1.5 p =* 0.038
1
0.5
0 WT KO WT KO WT KO WT KO kidney kidney liver liver brain brain muscle muscle
Figure 2.13: Signaling downstream of mTORC1 is upregulated in various tissues of Pip4k2c-/- mice p70-S6K Thr389 phosphorylation, total p70-S6K and SREBP1 (cleaved mature form) were blotted for in kidney, liver, brain and muscles from 12 months old wild type and Pip4k2c-/- mice (top). 3 mice per group. The bar graph (bottom) shows the ratio of p70-S6K Thr389 phosphorylation over total p70-S6K. The results are presented as means ± standard errors of the means. *p < 0.05
30 A P-p70-60K / Spleen total p70-60K +/+ +/+ +/+ -/- -/- -/- 2 p = 0.07 p-p70-S6K 1.5 (T389) 1 0.5 p70-S6K 0 WT KO spleen spleen
B Th Treg +/+ -/- +/+ -/- p-p70-S6K p-p70-S6K (T389) (T389) p70-S6K p70-S6K
Figure 2.14: Signaling downstream of mTORC1 is upregulated in the spleen of Pip4k2c-/- mice (A) p70-S6K Thr389 phosphorylation and total p70-S6K were blotted for spleen, and (B) Th cells and Treg cells isolated from the spleens of 12 months old mice. 3 mice per group. The bar graph shows the ratio of p70-S6K Thr389 phosphorylation over total p70-S6K. The results are presented as means ± standard errors of the means.
We next examined whether rapamycin, the allosteric mTORC1 inhibitor, could reduce the inflammatory phenotype of Pip4k2c-/- mice. Pip4k2c wild type and knockout mice were intraperitoneally injected with either vehicle or rapamycin
(3mg/kg/d) once a day for 2 weeks. Blood was withdrawn 2 weeks before the treatment and 24 hours after the first treatment. On the final day, blood, liver and spleen were collected. Using protein lysates from the liver and spleen, we performed SDS-PAGE and western blot. We found that Thr389 of p70-S6K was still
31 hyperphosphorylated in liver and spleen tissues of Pip4k2c-/- mice treated with vehicle control for 2 weeks. However, in the rapamycin treated Pip4k2c-/- mice, p70-
S6K phosphorylation on Thr389 was reduced to the levels seen in tissues from wild type mice, indicating that rapamycin had suppressed mTORC1 signaling (Figure
2.15). The levels of plasma cytokine were measured in the control and rapamycin treated mice and the ratio was calculated (Figure 2.16). IFNγ levels decreased in both wild type and knockout mice in 24 hours after the first treatment with rapamycin. However, after 2 weeks of treatment with rapamycin, IFNγ levels in both wild type and knockout mice returned to the basal levels (the level before any treatment). On the other hand, the plasma level of IL-12(p40) was not only suppressed by 24 hours of rapamycin treatment, but also remained suppressed after 2 weeks of treatment of the Pip4k2c-/- mice. IL-12 levels did not significantly change in the wild type mice treated with rapamycin. Also, we observed that the area of immune infiltrates in the livers of rapamycin treated Pip4k2c-/- mice was decreased after two weeks of rapamycin treatment (Figure 2.17). These data collectively indicate that inhibition of mTORC1 by rapamycin partially reduced the inflammatory phenotypes in Pip4k2c-/- mice.
32 Liver Pip4k2c-+/+ Pip4k2c-/- vehicle + + + - - - + + + - - - rapamycin - - - + + + - - - + + + p-p70-S6K (T389)
p70-S6K
actin
Spleen Pip4k2c-+/+ Pip4k2c-/- vehicle + + + - - - + + + - - - rapamycin - - - + + + - - - + + +
p-p70-S6K (T389) p70-S6K
actin
P-p70-60K / total p70-60K
p = 0.021* 2.5 1.5 p = 0.502 2 1.5 1 1 0.5 0.5 0 0 WT WT KO KO WT WT KO KO veh rapa veh rapa veh rapa veh rapa
Figure 2.15: Rapamycin reduces the activation of mTORC1 signaling in Pip4k2c-/- mice p70-S6K Thr389 phosphorylation, total p70-S6K and actin were blotted for in liver (top) and spleen (middle) from 12-14 months old wild type and Pip4k2c-/- mice treated with vehicle or rapamycin (daily i.p. injection for 2 weeks, 3mg/kg/d). 3 mice per group. The bar graph (bottom) shows the ratio of p70-S6K Thr389 phosphorylation over total p70-S6K. The results are presented as means ± standard errors of the means. *p < 0.05
33
A
3 WT veh 24 hrs 3 KO veh 24 hrs 2.5 WT veh 2 wks 2.5 KO veh 2 wks WT rapa 24h KO rapa 24h 2 2 WT rapa 2 wks KO rapa 2 wks 1.5 1.5 after treatment after
beforetreatment 1 1 0.5 0.5 0 0 ratio= IFNγ IL-12 IL-12 IL-2 IFNγ IL-12 IL-12 IL-2 (p40) (p70) (p40) (p70)
B
3 WT veh 24 hrs 3 WT rapa 24h 2.5 WT veh 2 wks 2.5 WT rapa 2 wks 2 KO veh 24 hrs 2 KO rapa 24h KO veh 2 wks 1.5 1.5 KO rapa 2 wks after treatment after beforetreatment 1 1 0.5 0.5
ratio= 0 0 IFNγ IL-12 IL-12 IL-2 IFNγ IL-12 IL-12 IL-2 (p40) (p70) (p40) (p70)
Figure 2.16: Changes in plasma cytokine levels after treatment with rapamycin (A and B) Plasma cytokines were detected using multiplex cytokine ELISA. Plasma was collected before the treatment, 24 hours after the first treatment and a day after the final treatment (2 weeks) from the following 4 groups: wild type with vehicle, wild type with rapamycin (3mg/kg/d), knockout with vehicle and knockout with rapamycin (3mg/kg/d), 4~8 mice (12-14 months old) per group, daily i.p. injection for 2 weeks. (A) Comparison of vehicle and rapamycin treatment in the same genotype groups (B) Comparison of wild type and knockout in the same treatment groups. The results are presented as means ± standard errors of the means.
34
9 * 8 7 (%)
6 5 4 3 totalarea 2 Immuneinfiltrates/ 1 0 WT WT KO KO veh rapa veh rapa
Figure 2.17: Changes in immune cell infiltration after treatment with rapamycin Mouse livers were collected after the final treatment (daily i.p. injection for 2 weeks) from the following groups: wild type with vehicle, wild type with rapamycin (3mg/kg/d), knockout with vehicle and knockout with rapamycin (3mg/kg/d), 4~8 mice (12-14 months old) per group. The area of immune infiltrates of H&E stained liver tissues were quantified using ImageJ software. 5 images per mouse were examined. The Y axis indicates the ratio of the area of immune infiltrates over total area of the tissue in each image. The results are presented as means ± standard errors of the means. *p < 0.05
35
Pip4k2c+/+ mice Pip4k2c-/- mice
PI5P4Kγ PI5P4KXγ
mTORC1 mTORC1
S6K SREBP1 S6K SREBP1
Treg Th Treg Th
inflammation inflammation
Figure 2.18: Model for the role of PI5P4Kγ in regulation of immunity PI5P4Kγ negatively regulates the mTORC1 pathway, compromising its ability to regulate Th cell differentiation and activation. The downstream signaling of mTORC1 in T cells is specifically hyperactivated in Pip4k2c-/- mice and this may be responsible for the inflammatory phenotype found in these mice.
36 Discussion:
Here we studied the physiological role of PI5P4Kγ in immune regulation by characterizing the phenotypes caused by knockout of Pip4k2c. We found that the immune system of Pip4k2c-/- mice was hyperactivated with increased immune cell infiltration in multiple organs, including liver, kidney, intestines and lungs. These infiltrating immune cells were mostly T cells and B cells. Moreover, we found that pro-inflammatory Th1-type cytokines were elevated in the serum of Pip4k2c-/- mice.
Pip4k2c-/- mice also exhibited an increase in the CD44+ T cell population (central memory T cells), indicating an increase in activated T cells. Moreover, we found that Pip4k2c-/- mice had fewer Foxp3+CD4+ Treg cells (regulatory T cells) compared to wild type mice. In addition, Pip4k2c-/- mice had increased CD4+ and CD8+ T cells and T cells from Pip4k2c-/- mice proliferated faster than those from wild type mice.
The biochemical mechanism by which the immune system is hyperactivated in Pip4k2c-/- mice is not clear. Our studies suggest that mTORC1, that is known to regulate T cell differentiation and activation (Delgoffe et al., 2010) is required to maintain the inflammatory phenotype in Pip4k2c-/- mice. The loss of PI5P4Kγ appears to activate mTORC1 signaling in mice. Interestingly, Mackey et al. reported that PI5P4Kγ is phosphorylated by mTORC1 on Ser324 and Ser328 (Mackey et al.,
2014). These results suggest that there is a feedback loop between mTORC1 and
PI5P4Kγ. On the other hand, Mackey et al. observed a decrease in activation of mTORC1 signaling when they knocked down expression of PIP4K2C in HeLa cells that had hyperactivation of mTORC1 due to TSC2 knockdown. It is not clear why germline deletion of Pip4k2c-/- has the opposite effect on in vivo mTORC1 signaling
37 to that observed in response to shRNA mediated knockdown in a cell line. One possibility is that the knockdown of TSC2 alters the role of Pip4k2c-/- in the control of mTORC1. Another possibility is that during development the other isoforms of
PI5P4Ks (which have much higher enzymatic activity) have replaced PI5P4Kγ resulting in a more active mTORC1 signaling complex.
The genetic interaction among PI5P4Kγ, PI5P4Kα and PI5P4Kβ is particularly interesting with respect to their abilities to heterodimerize with each other (Bultsma et al., 2010, Rao et al., 1998, Burden et al., 1999, Wang et al.,
2010). Interestingly, Pip4k2a-/-Pip4k2b-/- mice are perinatal (Emerling et al., 2013) and Pip4k2b-/-Pip4k2c-/- mice are not viable (Table S2.1) while Pip4k2a-/-Pip4k2c-/- mice are viable. Thus, mice that only have Pip4k2b are viable, but if this gene is deleted both Pip4k2a and Pip4k2c are critical for viability, indicating that these genes do not have redundant functions.
The respective roles for PI5P4Kγ, PI5P4Kα and PI5P4Kβ are complex.
Pip4k2b-/- mice exhibited enhanced insulin sensitivity, smaller body size, and decreased adiposity on a high fat diet. In contrast, Pip4k2c-/- mice were not different from wild type mice in these features (Figure. S2.4 and S2.5). In addition to the synthetic lethality for loss of Pip4k2a and Pip4k2b, and for loss of Pip4k2b and
Pip4k2c, distinct phenotypes of each of the Pip4k2a-/-, Pip4k2b-/- and Pip4k2c-/- mice indicate that each isoform has a unique role in vivo.
Of particular interest is the effect of the various knockouts on signaling through the PI3K-Akt-mTORC1 pathway. Deletion of Pip4k2b enhances Akt activation but, surprisingly, does not result in enhanced mTORC1 signaling.
38 Previous studies from many laboratories have shown that impaired mTORC1 activation results in smaller cells and smaller mice (Shima et al., 1998 and Laplante and Sabatini, 2013). Consistent with the failure of Pip4k2b deletion to link Akt activation to mTORC1 activation, the Pip4k2b-/- mice are smaller than wild type littermates. Although deletion of Pip4k2a results in no observable phenotypes, deletion of a single allele of Pip4k2a in the context of deletion of both alleles of
Pip4k2b results in even smaller mice. These data indicate that Pip4k2a and
PIP4k2b suppress PI3K-Akt activation but facilitate mTORC1 activation. This model is consistent with the observation that deletion of the single form of PIP4K2 in flies causes suppression of TORC1 activation and suppression of growth (Gupta et al.,
2013). Here we show that deletion of Pip4k2c results in enhanced rather than suppressed mTORC1 signaling. The simplest explanation for this result is that, since PI5P4Kγ has very little kinase activity compared to PI5P4Kα and
PI5P4Kβ, and it functions as a suppressor of these enzymes, at least in regard to their functions in supporting mTORC1 activation. This suppression could occur by competing with PI5P4Kα and PI5P4Kβ for a binding site required for localization of these enzymes at the site required for mTORC1 activation. Alternatively, PI5P4Kγ could generally suppress the activity of PI5P4Kα and PI5P4Kβ through formation of heterodimers that are less active than homodimers. This working model for the role of PI5P4Kγ is illustrated in Figure 1.2. In order to determine the epistasis among these enzymes, phenotypes of the viable double knockout mice will need to be better examined.
39 Finally, the results that we present here support the association of a SNP in
PIP4K2C locus with autoimmunity, suggesting that PI5P4Kγ expression is probably low in the autoimmune patients with the SNP near the PIP4K2C locus. The result from our study that treating Pip4k2c-/- mice with rapamycin reduced the inflammatory phenotype by decreasing the activation of mTORC1 indicates that drugs that target mTORC1 signaling are likely to be effective on the subset of autoimmune patients with the SNP near the PIP4K2C locus.
40 Methods:
Generation of Pip4k2c-/- mice
Pip4k2c-targeted embryonic stem cell clones, BO1, BO2 and DO1, were obtained from the knockout mouse project (KOMP) repository. These cells were grown in our laboratory and the conditional knockout allele of each clone was confirmed by genomic DNA PCR. The clones were then karyotyped at the
Transgenic Core at Dana Farber and the normal clones BO1 and BO2 were selected and injected into blastocysts at the Beth Israel Deaconess Transgenic
Facility. Three chimeric male mice were obtained and each was backcrossed with
C57BL/6J mice. Cassette bearing mice was mated to Rosa-eFLP1 mice to remove the lacZ reporter. The critical exons 3 and 4 were removed by mating the lacZ reporter-deleted mice to germline CMV-Cre deleter mice. Knockout mice were verified by PCR and western blotting.
PCR – genotyping
For genotyping, four primers below were used to amplify regions of genomic
DNA present in either wild type samples or knockout samples. pwtF : TGTCCCCAGGTCTTCAGGAACCT pwtR : TGCCTTCAGTTTCGCTTGGGGG pkoF : CACACCTCCCCCTGAACCTGAAAC pkoR : AGCCGCTGGGGCCAGATGAT
The primer pair pwtF/pwtR amplifies a fragment (~0.5kb) in wild type and the primer pair pkoF/pkoR amplifies a fragment (~0.5kb) in knockout.
41 Western blotting
Tissues were collected from the euthanized mice and flash frozen in liquid nitrogen. Tissues were homogenized in pre-chilled NP-40 lysis buffer (50 mM Tris pH 7.8, 150 mM NaCl, 0.5% NP40, Roche cOmplete EDTA-free protease inhibitor cocktail tablet (1 tablet per 25ml) added). Protein was quantified using Bradford assay (BIO-RAD, CA, USA), resolved with SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes. The membranes were probed overnight at 4 ˚C with the specified primary antibody. The images densitometries of the western blots were quantified using ImageJ software. Antibodies used were as follows: PI5P4Kγ (#HPA028658) (Sigma, USA), p-p70-S6K (#9206), p70-S6K
(#2708), PI5P4Kα (#5527), (Cell Signaling Technology, Inc., MA, USA), β-actin
(ab6276) (Abcam, Cambridge, UK). In addition, PI5P4Kγ antibody generated by
Cell Signaling Technology (commercially not available) was used to detect mouse
PI5P4Kγ.
Preparation of mouse tissues for immunohistochemistry
Tissues were removed from the euthanized mice and washed with PBS. The samples were fixed in 10% buffered formalin for 24 hours and paraffin embedded.
H&E staining was performed at the Rodent Histopathology Core at Dana
Farber/Harvard Cancer Center. Staining the immune infiltrates in liver tissue was performed at the Laboratory of Comparative Pathology at Memorial Sloan Kettering
Cancer Center: The samples were microsectioned, deparaffinized, rehydrated and heated with a pressure cooker to 125 °C for 30 seconds in citrate buffer for antigen
42 retrieval and then incubated with peroxidase and protein blocking reagents respectively for 5 minutes. Sections were then incubated with anti-CD3, B220 and
Mac2 antibodies, respectively.
Blood collection from the mouse and plasma preparation
Mouse tails were cut 1 mm from the tip with scissors and blood was collected into pre-chilled 1.5 ml EDTA coated Eppendorf tubes (Microvette CB300, Sarstedt).
The samples were centrifuged for 15 minutes at 3,000 rpm at 4 oC. The supernatants were transferred to a new Eppendorf tubes and passed through 0.22 micron filters (centrifuging for 1 minute at 5,000 rpm at 4oC).
Measurement of plasma cytokines
Mouse plasma was prepared as described above. Plasma cytokine levels were measured using multiplex cytokine assay at Eve Technologies (Alberta,
Canada). To measure cytokines in our laboratory, the BD Bioscience ELISA kit for
IL-12, IFNγ and IL-2 (#M1270, #MIF00 and #M2000, respectively) was used according to the manufacturer’s protocol.
Measurement of cytokines secreted from T cells
Secreted cytokines were measured by ELISA. All cytokines antibodies were purchased from Biolegend. Flow cytometric analysis was performed using a FACS
Calibur (Becton Dickinson).
43 Immunoglobulin isotyping
Mouse plasma was prepared as described above. Immunoglobulin isotyping was performed using BD Bioscience mouse immunoglobulin isotyping ELISA kit
(#550487) according to the manufacturer’s protocol.
T cell proliferation
CD4+ T cells, isolated from the spleens of wild type or knockout mice, were seeded at 10^5/well/200 µL medium in 96 well plates coated with the indicated concentrations of anti-CD3 and 1 µg/ml anti-CD28. Cells were cultured for 48 hours and then 3H-thymidine was added for another 16 hours before the measurement.
The level of radioactivity was measured by liquid scintillation.
Rapamycin treatment of the mouse
For rapamycin treatment, stock solutions (50 mg/ml) were diluted into vehicle
(5% Tween-80, 5% PEG 400 (polyethylene glycol, molecular weight 400) in 1× phosphate-buffered saline (PBS)) for 2 week (3 mg/kg/d) treatments through intraperitoneal injections. Mice were sacrificed after 2 weeks of treatment.
44 Acknowledgements:
We thank Dr. Chuan Wu for isolating T cells from mouse spleen and performing flow cytometry and T cell proliferation assay. We thank Rayman Choo- wing and Dr. Zhiwei Yang for assistance with genotyping Pip4k2c-/- mice. We thank
Shivan Ramsamooj for assistance with western blotting. We thank Kaitlyn Bosch and Dr. Jihye Yun for assistance with rapamycin treatment of Pip4k2c-/- mice. We thank Dr. Anthony Couvillon for providing anti-PI5P4Kγ antibody.
45 Chapter 3. Identification of PI5P4K inhibitors
Abstract:
Phosphatidylinositol-5-phosphate 4-kinases (PI5P4K) are emerging therapeutic targets for treating cancers and type 2 diabetes, and based on the results in Chapter 2 might also be useful in modulating immune responses. Mice with germline deletion of the Pip4k2b gene (encoding PI5P4Kβ) are hypersensitive to insulin and remain lean on a high-fat diet (Lamia et al., 2004). Germline deletion of both Pip4k2b alleles and both Trp53 alleles results in early embryonic death.
Deletion of both Pip4k2a alleles and one Pip4k2b allele in the context of Trp53-/- results in viable mice that are protected from tumor-dependent death (Emerling et al., 2013). Also, a subset of breast cancers express high levels of PI5P4Kα and
PI5P4Kβ and these kinases are essential for cell proliferation when p53 function is lost. These data suggest that inhibitors of PI5P4Kα and PI5P4Kβ could be beneficial to patients with type 2 diabetes, obesity or TP53 mutant cancers. Here we report the identification of novel compounds (named PIP4Kin1 and PIP4Kin2) that specifically inhibit PI5P4K isoforms. Micromolar concentrations of PIP4Kin1 and PIP4Kin2 mimic shRNA mediated depletion of PI5P4Kα and β by inhibiting the growth of a TP53 mutant breast cancer cell line (BT474) with little effect on a TP53 wild type cell line (MCF7). We also show that PIP4Kin1 and PIP4Kin2 irreversibly inhibit PI5P4Kα and PI5P4Kβ by covalently reacting with cysteine (Cys) residues in a regulatory loop of these enzymes. Importantly a crystal structure of PIP4Kin2 bound to PI5P4Kβ reveals that this molecule occupies the ATP binding pocket with
46 the Cys-reactive acrylamide moiety located near the regulatory loop. Finally we show that daily intraperitoneal injection of PIP4Kin2 into mice resulted in suppression of the growth of BT474 xenograft tumors but not MCF7 tumors in the opposite flank of the same mice. These results suggest that drugs that inhibit both
PI5P4Kα and PI5P4Kβ could be effective in treating TP53 mutant cancers at doses that are tolerable in vivo.
47 Introduction:
While the majority of phosphatidylinositol-4,5-bisphosphate (PI-4,5-P2) in mammalian tissues is generated by the canonical pathway in which phosphatidylinositol-4-phosphate (PI-4-P) is phosphorylated at the 5 position of the inositol ring, a minor fraction of PI-4,5-P2 is generated by a family of PI5P4K enzymes that phosphorylate phosphatidylinositol-5-phosphate (PI-5-P) at the 4 position of the inositol ring. Germline deletion of both alleles of Pip4k2b, the gene encoding PI5P4Kβ, results in leaner mice that do not become obese on a high fat diet and that have increased insulin sensitivity, suggesting that inhibitors of
PI5P4Kβ could be useful for treating obesity, insulin resistance or type 2 diabetes
(Lamia et al., 2004). In contrast, germline deletion of both alleles of Pip4k2b along with both alleles of the tumor suppressor gene Trp53 results in early embryonic lethality, suggesting that PI5P4Ks are essential for mediating stress responses in the absence of p53 (Emerling et al., 2013). Germline deletion of both alleles of
Pip4k2a, which encodes a second member of the PI5P4K family, did not cause synthetic lethality in the context of Trp53 deletion. However, deletion of both alleles of Pip4k2a and one allele of Pip4k2b in the context of germline Trp53 deletion resulted in a dramatic tumor-free life extension compared to mice with germline deletion of Trp53 alone (Emerling et al., 2013). These observations, along with other results in Emerling et al, (2013) and in other recent publications (Jones et al.,
2006, Keune et al., 2013, Jones et al., 2013 and Fiume et al., 2015) indicate that the PI5P4K family members play an important role in cellular regulation, especially in response to cellular stress. More importantly, these studies indicate that
48 inhibitors of either PI5P4Kβ alone or of both PI5P4Kα and PI5P4Kβ could be useful for treating cancers that are driven by genetic aberrations in TP53. The data in
Chapter 2 also raise the possibility that blocking the function of PI5P4Kγ might be beneficial for stimulating the immune system to attack cancers.
Here we present a novel approach for developing covalent inhibitors of
PI5P4K family members. A fortuitous observation that PI5P4Kγ is precipitated by beads linked to a covalent JNK kinase inhibitor (Zhang et al., 2012) led us to interrogate a family of related compounds for their ability to inhibit PI5P4Kα and
PI5P4Kβ. We identified two compounds, hereafter called PIP4Kin1 and PIP4Kin2
(Figure 3.1) that inhibit both PI5P4Kα and PI5P4Kβ at sub-micromolar concentrations. We also show that the alpha-beta unsaturated carboxylate moiety of these compounds reacts covalently with Cys residues in a flexible regulatory loop of PI5P4Kα and PI5P4Kβ, resulting in irreversible inhibition. A crystal structure reveals that PIP4Kin2 occupies the ATP binding pocket with the Cys-reactive acrylamide moiety pointing toward the region of the flexible loop that contains the
Cys residues involved in the covalent reaction. We also show that PIP4Kin2 inhibits the growth of a TP53 mutant breast cancer cell line (BT474) but not a TP53 wild type breast cancer cell line (MCF7). Finally, we show that PIP4Kin2 suppresses the growth of BT474 tumors but not MCF7 tumors in mouse xenografts.
49 Results:
PIP4Kin1 and PIP4Kin2 inhibit the enzymatic activity of PI5P4Ks
A series of analogs of JNK-In-1, an acrylamide-based covalent inhibitor of
JNK (Zhang et al., 2012) were interrogated for their ability to inhibit PI5P4Ks. GST-
PI5P4Kα, GST-PI5P4Kβ and GST-PI5P4Kγ were purified from Escherichia coli and incubated with either dimethyl sulfoxide (DMSO) control or 1 µM of each compound for 30 minutes. Then a lipid kinase assay was performed by adding γ-32P-ATP and
PI-5-P. The radiolabeled PI-4,5-P2 was separated by thin layer chromatography and quantified by autoradiography. Although most of the compounds tested had little or no effect on the PI5P4Ks at 1 µM, PIP4Kin1 and PIP4Kin2 (Figure. 3.1) exhibited strong inhibition of the in vitro kinase activity of all three PI5P4K isoforms, with
PIP4Kin2 being somewhat more effective (Figure 3.2 and Figure S3.1). In contrast,
PIP4Kin2 at 5 µM had no effect on the evolutionarily-related phosphatidylinositol kinases PIKFYVE and PI4P5Kα (Figure S3.2 and S3.3)
PIP4Kin1 PIP4Kin2
Figure 3.1: Structures of PIP4Kin1 and PIP4Kin2 PIP4Kin1 and PIP4Kin2 share structural similarities, most notably an acrylamide moiety that could react with cysteine residues on the PI5P4Ks.
50 PI5P4Kα PI5P4Kβ
1µM 1µM PIP4Kin1 PIP4Kin2 DMSO PIP4Kin1 PIP4Kin2 DMSO
PI-4,5-P2 PI-4,5-P2
1 1
0.5 0.5 DMSO PIP4Kin
0 0
PIP4Kin1 PIP4Kin2 PIP4Kin1 PIP4Kin2 Figure 3.2: PIP4Kin1 and PIP4Kin2 inhibit the kinase activity of PI5P4Ks GST-PI5P4Kα or GST-PI5P4Kβ purified from E.coli was incubated with indicated compounds (1 µM, 30 minutes, 25oC) and the kinase assays were carried out as described in the Methods. The radiolabeled PI-4,5-P2 product was visualized by autoradiography and quantified with Image J software. Signal was normalized to that of the DMSO control. N=3. The results are presented as means ± standard errors of the means.
PIP4Kin1 and PIP4Kin2 inhibited the growth of TP53 mutant breast cancer cells
In a previous study we showed that shRNA mediated knockdown of
PIP4K2A and PIP4K2B, inhibits the growth of the TP53 mutant breast cancer cell line BT474 but has no effect on the growth of the TP53 wild type cell line MCF7
(Emerling et al., 2013). We examined the effects of PIP4Kin1 and PIP4Kin2 on these cells and found that they mimicked the shRNA effects. At 1 µM both
51 compounds suppressed growth of BT474 cells with no effect on MCF7 cells (Figure.
3.3).
PIP4Kin1
DMSO DMSO 10000000 PIP4Kin1, 1µM 10000000 PIP4Kin1, 1µM 9000000 PIP4Kin1, 0.1µM 9000000 PIP4Kin1, 0.1µM 8000000 8000000 7000000 7000000 6000000 6000000
RLU 5000000 5000000 4000000 4000000 3000000 3000000 2000000 2000000 1000000 MCF7 1000000 BT474 0 0 Day 0 Day 1 Day 2 Day 3 Day 4 Day 0 Day 1 Day 2 Day 3 Day 4
PIP4Kin2
DMSO DMSO 10000000 PIP4Kin2, 1µM 10000000 PIP4Kin2, 1µM 9000000 PIP4Kin2, 0.1µM 9000000 PIP4Kin2, 0.1µM 8000000 8000000 7000000 7000000 6000000 6000000 5000000 5000000 RLU 4000000 4000000 3000000 3000000 2000000 2000000 1000000 MCF7 1000000 BT474 0 0 Day 0 Day 1 Day 2 Day 3 Day 4 Day 0 Day 1 Day 2 Day 3 Day 4
Figure 3.3: PIP4Kin1 and PIP4Kin2 inhibit proliferation of BT474 cells but not MCF7 cells Luminescent cell viability assay in BT474 cells (left) and MCF7 cells (right) following incubation with the indicated compounds. N=4. The results are presented as means ± standard errors of the means.
52 Moreover, it was only necessary to have PIP4Kin2 present for the first 6 hours to suppress growth of BT474 cells over a 6 day interval, consistent with
PIP4Kin2 causing irreversible inhibition of PI5P4Kα and PI5P4Kβ (Figure 3.4).
BT474 cells recovered more quickly when PIP4Kin1 was removed from the media, consistent with PIP4Kin1 being a less effective inhibitor of PI5P4Kα and PI5P4Kβ.
PIP4Kin2 also suppressed the growth of the TP53 mutant breast cancer cells
HCC38 and HCC70, which have hot spot TP53 mutations (Figure S3.4).
9000000 5000000 4500000 8000000 4000000 7000000 3500000 6000000 3000000 5000000 2500000 RLU 4000000 2000000 3000000 1500000 2000000 1000000 1000000 500000 0 0 Day 0 Day 2 Day 4 Day 6 Day 0 Day 2 Day 4 Day 6 DMSO DMSO, wash PIP4Kin1, 1µM PIP4Kin1, 1µM, wash PIP4Kin2, 1µM PIP4Kin2, 1µM, wash
Figure 3.4: PIP4Kin1 and PIP4Kin2 continue to inhibit proliferation of BT474 cells after washout Luminescent cell viability assay in BT474 cells treated with indicated compounds. Left: Cells were treated at Day 0 and monitored for 6 days. Right: Cells were treated on Day 0 and washed with PBS 3 times 6 hours after the treatment. Fresh media containing no compound was added and cells were monitored for 6 days. N=4. The results are presented as means ± standard errors of the means.
53 Inhibition of BT474 cell growth by PIP4Kin1 and PIP4Kin2 is partially rescued by shifting to a temperature where p53 is partially functional.
The TP53 E285K mutation in BT474 cells is nonfunctional at 37 oC but has partial function at 32 oC (Dearth et al., 2007). We previously showed that partially restoring p53 activity of the BT474 cells by culturing the cells at 32 oC partially rescued the proliferation of the cells in which PIP4K2A and PIP4K2B were knocked down by shRNA. The data in Figure 3.5 shows that when BT474 cells are cultured at 32 oC they acquire partial resistance to PIP4Kin1 and PIP4Kin2. These results further support the concept that PIP4K2A and PIP4K2B are only critical for cell growth and survival in the context of defective p53 protein and also suggest that
PIP4Kin1 and PIP4Kin2 are acting on target to suppress the growth of BT474 cells.
PIP4Kin1 increases insulin-dependent activation of Akt
Previous studies showed that Pip4k2b-/- mice exhibit increased insulin sensitivity and that this correlated with enhanced phosphorylation of Akt at Thr308 in muscle in response to insulin injection (Lamia et al., 2004). In agreement with these in vivo studies, overexpression of PIP4K2B in cell culture caused suppression of insulin-dependent Akt phosphorylation at Thr308 (Carricaburu et al., 2003). The data in Figure 3.6 shows that within 10 minutes of addition of 1 µM PIP4Kin1 to
HeLa cells, there is a significant increase in insulin-stimulated phosphorylation of
Akt at Thr308 and that this enhanced phosphorylation persists for 2 hours.
Phosphorylation at Ser473 was also somewhat enhanced. These results provide evidence that PIP4Kin1 mimics the effects of PIP4K2B gene deletion. They also
54 argue that PI5P4Kβ has a relatively direct role in enhancing insulin signaling since the response occurs so quickly. Despite the dramatic increase in insulin-induced
Akt activation there was only a subtle increase in p70-S6K phosphorylation.
120 PIP4Kin1 / DMSO (37 oC) 100 PIP4Kin1 / DMSO (32 oC) % 80 60 40 20 0 Day 0 Day 2 Day 4 Day 6
120 PIP4Kin2 / DMSO (37 oC) 100 PIP4Kin2 / DMSO (32 oC) % 80 60 40 20 0 Day 0 Day 2 Day 4 Day 6
Figure 3.5: Reducing the temperature of the culture media to 32 oC, where mutant p53 in BT474 cells is partially functional, partially rescues PIP4Kin1/2 mediated inhibition of cell growth Luminescent cell viability assay for BT474 cells treated with the indicated compounds and cultured at 37 oC or 32 oC. The bar graph shows the ratio of the cell growth in 1 µM PIP4Kin1/2 over the cell growth in DMSO control. N=4. The results are presented as means ± standard errors of the means.
55
Insulin (5nM) - + + + + + + + + - + + + + - - - - DMSO 10m 30m 1h 2h + + + + PIP4Kin1 (1µM ) - - - - - 10m 30m 1h 2h p-Akt (T308)
p-Akt (S473)
Akt p-p70-S6K (T389) p-p70-S6K
Figure 3.6: PIP4Kin1 enhances Insulin-dependent activation of Akt HeLa cells were serum starved for 16 hours and treated with 5nM insulin and 1 µM PIP4Kin1. The cells were then harvested at 10 minutes, 30 minutes, 1 hour and 2 hours and the lysates were blotted for Akt phosphorylation at Thr308 and Ser473, total Akt, p70-S6K phosphorylation at Thr389 and total p70-S6K.
PIP4Kin2 decreases the levels of glucose metabolites in BT474 cells
Previous studies showed that shRNA mediated knock down of PIP4K2A and
PIP4K2B in BT474 cells results in a decrease in the levels of glucose metabolites
(Emerling et al., 2013). The data in Table 3.1 reveal that within 30 minutes of adding 1 µM of PIP4Kin2 to BT474 cells there is a significant decrease in a number of intermediates of glucose metabolism and that this effect persists for 24 hours.
56
Table 3.1: The effect of PIP4Kin2 on metabolites in BT474 cells
Metabolites extracted from BT474 cells were treated with either DMSO or 1 µM PIP4Kin2 for 30 minutes or 24 hours. Metabolites were analyzed using mass spectrometry. All assays were performed in triplicate. Enriched pathways were found using metabolites that changed (increased or decreased) by more than 30% in cells treated with PIP4Kin2. Enrichment analysis was performed using MBRole (Chagoyen et al., 2011). P-values were adjusted for multiple testing using a false discovery rate (FDR) method.
57 Table 3.1 (Continued) metabolites decreased in 30 minutes
adjusted in Enriched pathway p-val p-val hit bckgnd Citrate cycle (TCA cycle) 1.12E-06 2.67E-08 7 20 Alanine, aspartate and glutamate metabolism 3.15E-06 1.13E-07 7 24 Oxidative phosphorylation 1.20E-04 5.81E-06 5 16 Pyrimidine metabolism 1.20E-04 7.13E-06 8 59 Nicotinate and nicotinamide metabolism 1.32E-04 9.41E-06 7 44 Glycolysis / Gluconeogenesis 1.59E-04 1.32E-05 6 31 metabolites increased in 30 minutes
adjusted in Enriched pathway p-val p-val hit bckgnd Purine metabolism 3.83E-06 1.24E-07 12 92 Phenylalanine, tyrosine and tryptophan biosynthesis 1.07E-05 5.19E-07 7 27 ABC transporters 1.34E-05 8.62E-07 11 90 Aminoacyl-tRNA biosynthesis 1.36E-04 1.10E-05 9 75 Pyrimidine metabolism 1.46E-04 1.41E-05 8 59 Arginine and proline metabolism 1.40E-03 1.58E-04 8 82 metabolites decreased in 24 hours
adjusted in Enriched pathway p-val p-val hit bckgnd Citrate cycle (TCA cycle) 1.12E-06 2.67E-08 7 20 Alanine, aspartate and glutamate metabolism 3.15E-06 1.13E-07 7 24 Pyrimidine metabolism 1.20E-04 7.13E-06 8 59 Oxidative phosphorylation 1.20E-04 5.81E-06 5 16 Nicotinate and nicotinamide metabolism 1.32E-04 9.41E-06 7 44 Glycolysis / Gluconeogenesis 1.59E-04 1.32E-05 6 31 Cysteine and methionine metabolism 5.06E-04 4.81E-05 7 56 metabolites increased in 24 hours
adjusted in Enriched pathway p-val p-val hit bckgnd Pyrimidine metabolism 1.10E-10 3.68E-12 16 59 Purine metabolism 5.05E-06 2.52E-07 11 92 Alanine, aspartate and glutamate metabolism 6.00E-04 4.00E-05 5 24 Pentose phosphate pathway 2.05E-03 1.71E-04 5 32 Arginine and proline metabolism 4.07E-03 4.07E-04 7 82
58 PIP4Kin1 and PIP4Kin2 are covalent inhibitors
The ability of PIP4Kin1 and PIP4Kin2 to have persistent effects on BT474 cells for several days following compound removal suggested that the acrylamide moiety of these molecules may be forming a covalent bond with residues on
PI5P4Kα and PI5P4Kβ. We generated biotinylated PIP4Kin1 (biotin-PIP4Kin1) and added this molecule to lysates of HEK293T cells, then used streptavidin beads to pull down proteins that were tightly bound to this molecule. Western blots of the proteins that were pulled down revealed that all three forms of PI5P4K were pulled down with the beads (Figure. 3.7). Incubating with 6 M urea prior to the pull down did not reduce the amount of the three proteins pulled down confirming that the association of biotinylated PIP4Kin1 with these three enzymes is covalent. More importantly, preincubating the intact HEK293T cells with 1 µM of non-biotinylated
PIP4Kin1 or PIP4Kin2 for 16 hours prior to cell lysis very effectively blocked the ability of biotinylated-PIP4Kin1 to pull down the three enzymes when added to the lysed cells. These results argue that over a period of 16 hours of continuous exposure in intact HEK293T cells to either 1 µM PIP4Kin1 or 1 µM of PIP4Kin2 all three PI5P4K enzymes in HEK293T cells are covalently inhibited and that the stoichiometry of inhibition is more than 80%.
59
Figure 3.7: A biotin-PIP4Kin1 pull down assay indicates that PIP4Kin1 and PIP4Kin2 are covalent inhibitors of the PI5P4K enzymes HEK293T cells were treated with DMSO or 1 µM PIP4Kin2 for 16 hours. Lysates were incubated with biotin-PIP4Kin1 and streptavidin-beads were added. The pellet was washed with either lysis buffer or lysis buffer containing 6M Urea. Bound proteins were analyzed by western blotting for PI5P4Kα/β, PI5P4Kβ and PI5P4Kγ.
Identification of residues in PI5P4Kα and PI5P4Kβ that react with PIP4Kin1
Using mass spectrometry, we identified cysteine residues in PI5P4Kα (C293) and PI5P4Kβ (C307 and C318) that covalently react with PIP4Kin1 (Figure 3.8).
These Cys residues are in the highly divergent sequence known as the “insert”
(Boronenkov and Anderson, 1995), a flexible region that includes a cluster of nearby Ser/Thr residues that have been shown to mediate inhibition of PI5P4K activity when phosphorylated (Figure 3.8, *, S326 in PIP4K2β in response to UV irradiation, Jones et al., 2006). In collaboration with Sirano Dhe-Paganon’s laboratory, we crystallized PI5P4Kβ in complex with PIP4Kin2 (Figure 3.9). The structure revealed PIP4Kin2 bound in the nucleotide-binding pocket of
PI5P4Kβ. The canonical hydrogen bond with the hinge region (Val204 backbone)
60 was formed with the N1 of the chloro-pyrimidinyl moiety of the inhibitor and it was complemented with two additional hydrogen bonds with the side chain amide of
Asn203 (Figure 3.9). The tryptophan-like group of the inhibitor, which had excellent electron density, approached the DFG-like loop (residues 368-371) and pointed towards the activation-loop (residues 372-398). On the other hand, the acrylamide- containing alkyl chain of the inhibitor, although poorly resolved, was within a few
Angstroms of Helix α7, which forms the beginning of the Cys307/318-containing
"insert". As with the apo structure (Rao et al, 1998), the "insert" did not resolve well in the final structure, indicating flexibility or disorder in this region of the protein. We believe that flexibility of the nucleophilic cystein might be required for reactivity and potent inhibition. Additionally, PI4P5Ks and PIKFYVE lack these Cys residues in the region that is analogous to the “insert” of PI5P4Ks. Moreover, Phe205 in
PI5P4Kβ is switched to Leu264 in PI5P4Kα. This might explain why these homologous enzymes, PI4P5Kα and PIKFYVE are not inhibited by PIP4Kin2. Thus, our mass-spec mapping and crystal structure studies are consistent with a model in which PIP4Kin2 binds in the catalytic pocket in an orientation that allows the acrylamide moiety to covalently react with Cys residues in the regulatory loop.
Although we did not map the reactive site in PI5P4Kγ, this protein also has a Cys residue at an analogous position to C318 of PI5P4Kβ that is likely to mediate the covalent reaction with PIP4Kin1 and PIP4Kin2.
61
Labeled peptide detected PI5P4Kβ AEDEE(C)ENDGVGGNLL(C)SYGTPPDSPGNLLSFPR C307 C318 *
PI5P4K sequence alignment PI5P4Kα 289 EEVECEENDGEEEGESDGTHPV-GTPP 314 PI5P4Kβ 305 EE----CE--NDGVG---GNLLCS----Y-GTPP 324 PI5P4Kγ 304 --E---D--E--SE--VD---GD---CS-----LTG-PP 319
Figure 3.8: Mapping PI5P4Kβ cysteine residues targeted by PIP4Kin1 Liquid chromatography/mass spectrometry was applied to identify specific PI5P4Kβ cysteines covalently modified with PIP4Kin1. The results are presented in the extracted ion chromatogram (XIC) for quadruply charged ions of AEDEECENDGVGGNLLCSYGTPPDSPGNLLSFPR (PI5P4Kβ 302-335) modified by PIP4Kin1. In the first peak, C307 is covalently labeled with PIP4Kin1, while the second peak corresponds to the same peptide modified at C318. The peptide doubly modified with PIP4Kin1 was not detected. * S326 as the site of phosphorylation on UV irradiation (Jones et al., 2006)
62
Figure 3.9: Co-crystal of PI5P4Kβ and PIP4Kin2 (A, B, C and D) Crystal complex of PI5P4Kβ with PIP4Kin2 reveals that PIP4Kin2 binds in the ATP pocket with the reactive moiety pointing toward the flexible regulatory loop. Data sets were integrated and scaled using XDS (Kabsch, 2009). Structures were solved by molecular replacement using the program Phaser (McCoy et al., 2007) and the search model PDB entry 1BO1. The ligand was positioned and preliminarily refined using Buster and Rhofit (Smart et al., 2011). Iterative manual model building and refinement using Phenix (Adams et al., 2009) and Coot (Emsley et al., 2004) led to a model with excellent statistics.
63 Figure 3.9 (Continued)
A Co-crystal of PI5P4Kβ and PIP4Kin2 catalytic pocket
reactive moiety of PIP4Kin2
35 AA region of flexible 35 AA region of flexible regulatory loop regulatory loop containing Cys residues containing Cys residues
B
64 Figure 3.9 (Continued)
C
D
65 Pharmacokinetics and maximum tolerated dose of PIP4Kin2
Pharmacokinetics of PIP4Kin2 were investigated in male Swiss Albino mice following a single intravenous (i.v.) bolus (2 mg/kg) and oral administration (10 mg/kg). Blood samples were collected at time points indicated in Figure. 3.10 and analyzed with mass spectrometry. Following a single i.v. bolus, PIP4Kin2 showed a plasma clearance of 138 mL/min/kg with terminal elimination half-life of 9.63 hours
(Figure 3.10). In the case of oral administration, the plasma concentration was detectable up to 4 hours with Tmax of 0.25 hours.
PIP4Kin2, i.v. (2 mg/kg) /mL) PIP4Kin2, p.o. (10 mg/kg) ng ( Plasmaconcentration
hours
T aC /C AUC AUC T clearance V Matrix Route max o max last inf 1/2 ss (hr) (ng/mL) (hr*ng/mL) (hr*ng/mL) (hr) (mL/min/kg) (L/kg) i.v. - 967.17 220.15 240.68 9.63 138.5 52.8 Plasma p.o. 0.25 8.5 20.02 36.71 3.98 - - Figure 3.10: The half-life of PIP4Kin2 in serum following a single i.v. injection is 9.63 hours Mice were treated with PIP4Kin2 (2 mg/kg for i.v. injection and 10 mg/kg for p.o., 9 mice per group) and plasma concentration of PIP4Kin2 was measured at the indicated time points. AUC, area under the concentration time curve; AUCinf, AUC from the time of dosing extrapolated to infinity; AUClast, AUC from the time zero to the last measurable concentration; Co, initial plasma concentration; Cmax, peak plasma concentration; Tmax, time for the peak plasma concentrations; Vss, Steady State Volume
66 To evaluate the maximum tolerated dose of PIP4Kin2 in mice, we used estrogen pellet implanted NOD/SCID mice since these implants would also be used for the efficacy studies in mice with xenografts of breast cancer cell lines. PIP4Kin2 was administered daily by intraperitoneal (i.p.) injection at three different doses (5,
10 and 20 mg/kg). No obvious weight loss or morbidity was observed in the mice in the 5 mg/kg dose group, while four out of six mice in the higher dose groups died during the 14 day study (Figure 3.11). Autopsies of the mice revealed intestinal adhesion or adhesion of other abdominal organs and enlargement of the intestinal track. So we concluded that the maximum tolerated dose was between 5 mg/kg and
10 mg/kg in these mice.
30 Group 1 PIP4Kin2 (5 mg/kg) 28 Group 2 PIP4Kin2 (10 mg/kg) 26 Group 3 PIP4Kin2 (20 mg/kg)
24
22
Body Weight (g) BodyWeight 20
18
16 * ** 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Days
Figure 3.11: Maximum tolerance dose test with PIP4Kin2 The mice were daily i.p. injected with PIP4Kin2 (3 groups, 5, 10, and 20 mg/kg, 3 mice per group) and body weight was measured daily for 2 weeks. *On day 12, 1 animal in Group 2 and 2 animals in Group 3 died. **On day 13, 1 animal in Group 2 died. The results are presented as means ± standard errors of the means.
67 PIP4Kin2 impaired the growth of BT474 tumors but not MCF7 tumors
To investigate the in vivo therapeutic efficacy of PIP4Kin2, we utilized orthotopic xenografts of MCF-7 and BT474 human breast cancer cell lines in female
NOD/SCID mice. Estrogen pellet-implanted mice were inoculated with MCF7 cells in the left mammary fat pad (day 1) and with BT474 cells in the right mammary fat pad (day 5). Previous studies with these cell lines indicated that MCF7 cells exhibited a lag in tumor growth after implantation compared to BT474 cells and the
5 day time differential resulted in similar tumor volume for the two xenografts at the time of therapeutic intervention. After 21 days, when the mean size of BT474 tumors reached approximately 200 (150~250) mm3, mice were intraperitoneally injected with either vehicle (group 1) or PIP4Kin2 (3 mg/kg, group 2) daily for the next 4 weeks. The dose of PIP4Kin2 was increased to 4 mg/kg on day 29 and then to 6 mg/kg on day 35. The mice were terminated on day 48. Mean body weights in both vehicle and PIP4Kin2 groups were stable throughout the course of the treatment and none of the mice died (Figure 3.12A). The percent (%) growth of
BT474 tumors on day 48 was 524.5% in the vehicle group while it was 341.7% in the PIP4Kin2 group (Figure 3.12B). This significant retardation in the growth of
BT474 tumors indicates that PIP4Kin2 is efficacious at doses that are tolerated without significant toxicity. In contrast, the growth of MCF7 tumors was not impaired by PIP4Kin2. These results are consistent with the model that PI5P4K inhibition is synthetic lethal with p53 loss.
68
Figure 3.12: PIP4Kin2 has significant antitumor activity in an orthotopic BT474 xenograft model The left and right mammary fat pads were inoculated with MCF7 and BT474 cells respectively. When the mean tumor size reached about 200 mm3, the treatments were initiated. The mice were i.p. injected daily with either vehicle or PIP4Kin2 (day 21-28: 3mg/kg, day 29-34: 4mg/kg, day 35-48: 6mg/kg, 8 mice per group). Body weight and tumor size were measured twice a week and mean volumes ± standard error plotted.
69 Figure 3.12 (Continued)
A 28 Body Weight 26
24
22 Vehicle
Body Weight (g) Body Weight 20 PIP4Kin2
18 20 22 24 26 28 30 32 34 36 38 40 42 Days after tumor inoculation
B 1000 BT474 tumor
) Vehicle 3 800 PIP4Kin2 600
400
200 Tumor Size(mm Tumor 0 20 22 24 26 28 30 32 34 36 38 40 42 Days after tumor inoculation C MCF7 tumor 1000 ) 3 Vehicle 800 PIP4Kin2 600
400
200 Tumor Size (mm Tumor 0 20 22 24 26 28 30 32 34 36 38 40 42 Days after tumor inoculation
70 Interestingly after day 37, the MCF7 xenograft in PIP4Kin2-treated mice was larger than in the control group (331% growth versus 239% growth-Figure. 3.12C).
This effect could be because the two tumors are competing for nutrients and that decreased growth of the BT474 xenograft in the opposite flank of the same mouse facilitated the growth of the MCF7 xenograft. It is also possible that PIP4Kin2 had a direct effect on stimulation of the growth of MCF7 cells. To address this issue, we evaluated the effect of PIP4Kin2 on the growth of MCF7 cells over a 4 week period and found a slight decrease in the proliferation of these cells (Figure. S3.5), suggesting that PIP4Kin2 does not directly stimulate the growth of MCF7 cells.
Target engagement of PIP4Kin2
To evaluate the in vivo target engagement of PIP4Kin2, xenograft tumors were removed from the mice on the final day of therapy and homogenized and lysed. The lysates were incubated with 1 µM biotin-PIP4Kin2 for 16 hours. The proteins linked to biotin-PIP4Kin2 were pulled down by incubating the lysates with streptavidin beads for 3 hours. The precipitated and released proteins were separated by SDS-PAGE and analyzed by western blotting (Figure 3.13). There was notably less PI5P4K2β pulled down with biotin-PIP4Kin2 in BT474 tumors and livers from the mice that had been treated with PIP4Kin2, indicating that at the dose achievable in vivo there was an inhibition of PI5P4Kβ.
71
BT474 tumor Liver Vehicle Vehicle Vehicle PIP4Kin2 PIP4Kin2 PIP4Kin2 Vehicle Vehicle Vehicle PIP4Kin2 PIP4Kin2 PIP4Kin2 PI5P4Kβ PI5P4Kβ
p=0.086 p=0.076 200 200
150 150
100 100
50 50
0 0 arbitraryintensity unit arbitraryintensity unit vehicle PIP4Kin2 vehicle PIP4Kin2
Figure 3.13: In vivo target engagement of PIP4Kin2
BT474 xenograft tumors and livers collected from the mice in the efficacy study as described in the Methods. The lysates incubated with 1 µM biotin-PIP4Kin2 and then with streptavidin beads. PI5P4Kβ was blotted. The images of the western blot were quantified using ImageJ software. N=3. The results are presented as means ± standard errors of the means.
72 Discussion:
Here we have shown that we have identified small molecules, PIP4Kin1 and
PIP4Kin2 as inhibitors of PI5P4Kα and PI5P4Kβ. We found that PIP4Kin1 and
PIP4Kin2 had similar effects to those previously observed with shRNA mediated knockdown of PI5P4Kα and PI5P4Kβ in cells. These compounds inhibited the proliferation of TP53 mutant BT474 cells with little effect on the proliferation of TP53 wild type MCF7 cells. Taking advantage of the previous observation that the E285K mutant of p53 in BT474 cells is inactive at 37 oC but functional at 32 oC, we showed that, when grown at 32 oC, BT474 cells exhibit partial resistance to PIP4Kin1 and
PIP4Kin2. Thus, consistent with previous studies with germline deletion of Pip4k2a and Pip4k2b and with shRNA knockdown of these enzymes (Emerling et al., 2013), the inhibitors we have identified preferentially suppress growth of cells with defective p53 protein.
In addition to the effects on growth of TP53 mutant cancers, the inhibitors mimicked other responses that had previously been observed in cells where
PIP4K2A and PIP4K2B were deleted or knocked down with shRNA. These compounds caused an increase in phosphorylation of Akt at Thr308 upon insulin stimulation, consistent with results observed in Pip4k2b -/- mice (Lamia et al., 2004).
They also caused a decrease in glucose metabolites in BT474 cells, consistent with responses to shRNA mediated knockdown of PIP4K2A and PIP4K2B (Emerling et al, 2013. Importantly, these effects were observed within a short time of addition of
PIP4Kin1 or PIP4Kin2 to cells (10 to 30 minutes) indicating that the PI5P4Ks play acute roles in cellular regulation.
73 As previously discussed (Emerling et al., 2013) the increased Akt activation paralleled by decreased glucose metabolism and impaired growth are paradoxical since the PI3K-Akt pathway is known to enhance glucose uptake and metabolism and promote cell growth and cell survival. A likely explanation is that the PI5P4Ks are mediating many different effects at different cellular compartments. The enhanced insulin stimulation of Akt could be because failure to generate sufficient
PI-4,5-P2 in endosomes allows the internalized insulin receptor to remain in the endosome rather than to traffic to another compartment where it could no longer activate Akt. Emilio Hirsch’s laboratory has recently shown that the internalized insulin receptor continues to signal to Akt in the endosome by activating the class
IIγ PI3K, which generates PI-3,4-P2 from PI-4-P, thereby promoting PDK1-dependt phosphorylation of Akt2 at Thr309 (analogous to Thr308 in Akt1) by a mechanism that does not require PI-3,4,5-P3 production by class Ia PI3Ks (Braccini et al.,
2015). This mechanism is consistent with our observation that the activation loop
Thr of Akt is preferentially phosphorylated upon adding PIP4Kin1 or PIP4Kin2 and also resolves the quandary that PIPKin1 and PIPKin2 are likely to lower PI-4,5-P2 in the endosome and thus impair production of PI-3,4,5-P3 by class Ia PI3Ks at this location.
The failure to propagate the enhanced insulin stimulation of Akt to a robust activation of the mTORC1 complex might be due to a distinct function of PI5P4Ks in providing PI-4,5-P2 at the lysosome where the mTORC1 complex resides. The model proposed in Chapter 1 and Chapter 2 proposes that PI5P4Kα and PI5P4Kβ play a critical role in enhancing mTORC1 signaling. This argument is primarily
74 supported by impaired mTORC1 activity in flies in response to deletion of the single
PIP4K2 gene in that species. But it is also supported by the small size of mice that have deletions in PIP4K2A and PIP4K2B (Lamia et al., 2004; Emerling et al., 2013) and by the enhanced mTORC1 activity upon deletion of PIP4K2C, an essentially inactive enzyme that is likely to suppress the function of PIP4K2A and PIP4K2B. In any event, the short term response to acute inhibition of these enzymes by inhibitors is likely to be very different from the long-term adjustment to the absence of the genes during development in regard to regulation of mTORC1.
The suppression of glycolytic intermediates upon addition of PIP4Kin2 to
BT474 cells is also a paradox. The inhibitor study produced similar results to those observed with shRNA knockdown of PIP4K2A and PIP4K2B, so this is probably an on-target response to PIP4Kin2. The decrease in glycolytic intermediates is almost certainly not a consequence of activating Akt or regulation of mTORC1. It might be explained by a role for PI5P4K enzymes in autophagosome function since the autophagosome plays a central role in mediating responses to metabolic stress.
Finally, as discussed in Chapter 1, there is increasing evidence that both PI-5-P and PI-4,5-P2 play critical roles in the nucleus by mediating responses to various types of stress (discussed in Chapter 1) and these mechanism might play a role in regulation of glucose metabolism.
Finally, there is considerable interest in developing inhibitors of the PI5P4Ks for cancer therapy and the inhibitors we have identified reveal a novel approach.
The success of ibrutinib, a covalent inhibitor of the BTK protein-Tyr kinase, in treating chronic lymphocytic leukemia (CLL) has encouraged pharmaceutical
75 companies to pursue covalent inhibitors for therapeutic intervention. The surprising observation we made is that it is possible to develop ATP site inhibitors that react with Cys residues that are quite distant from the catalytic pocket. Fortuitously, the relatively flexible regulatory loops of the PI5P4Ks have Cys residues that must periodically collide with the ATP binding pocket, allowing them to react with the acrylamide moiety of PIP4Kin1 and PIP4Kin2. This covalent reaction maintains these molecules at a location that enhances occupation of the ATP pocket, thereby causing irreversible inhibition. While it is clear that additional studies will be needed to optimize inhibitors that could be used therapeutically, these compounds have provided proof-of–concept that inhibitors of PI5P4Ks could be effective in treating cancers with TP53 mutations. PIP4Kin2 was toxic when given at high doses daily via intraperitoneal injection and the dose that we chose for efficacy studies tumor growth inhibition but did not shrink the existing tumors. It is not clear whether the toxicity observed is an on-target effect. Increasing the selectivity for PI5P4Ks might improve the efficacy/toxicity ratio. It is also possible that providing higher doses at longer intervals (e.g. over 3 days) would be more effective and less toxic since the growth of BT474 cells was suppressed as long as 6 days after removing PIP4Kin2 from the culture medium (Figure 3.4).
76 Methods:
PI5P4K kinase assay
PI5P4K in vitro kinase assay was carried out as described in Rameh et al
(Nature, 1997). Briefly, 0.1 µg of GST-PI5P4Kα or 0.4 µg of GST-
PI5P4Kβ resuspended in 70 µL of kinase buffer containing 20 mM HEPES pH 7.4,
100 mM NaCl, 0.5mM EGTA was stabilized at room temperature for 10 minutes and incubated with 1µM of DMSO or indicated compound for 30 minutes. Then the kinase reaction was carried out in a total volume of 100 µL for 10 minutes by adding
20 µL of lipids (4 µg of phosphatidylserine and 2 µg of PI-5-P) in buffer containing
30 mM HEPES pH7.4 and 1 mM EGTA, and 10 µL of ATP mix (500 µM non-
32 radiolabeled ATP, 10 uCi [γ- P]-ATP, 65 mM HEPES pH 7.4 and 100 mM MgCl2).
The reaction was terminated by adding 50 µL of HCl. Phosphoinositides were extracted by adding 100 µL methanol/chloroform (1:1, vol:vol) mix and resolved by thin layer chromatography on a heat-activated 1% potassium oxalate-coated silica gel (EMD Chemicals Inc., Billerica, MA, USA) with a 1-propanol/2 M acetic acid
(65:35, vol:vol) solvent system. The radiolabeled PI-4,5-P2 products were visualized by Phosphorimager (Molecular Dynamics, STORM840, GE Healthcare,
Waukesha, WI, USA).
Cell culture
o o Cells were incubated in a 37 C or 32 C humidified incubator in 5% CO2. All breast cancer cell lines were obtained from ATCC and cultured in Dulbecco’s modified Eagle medium (Mediatec) supplemented with 10% fetal bovine serum.
77 Measurement of metabolites
The BT474 breast cancer cell line was treated with either DMSO or 1µM
PIP4Kin2 for either 30 minutes or 24 hours. Metabolites were extracted by aspirating media and quickly adding pre-chilled 80% methanol to the cells on dry ice. Plates with 80% methanol were placed in -80°c for 30 minutes. Cells were scraped on dry ice and collected in pre-chilled tubes. Tubes were spun at maximum speed for 15 minutes at 4°c and supernatants were collected. Samples were dried down completely using a speed vacuum at room temperature. All samples were done in triplicate. Samples were then re-suspended using 20 µL HPLC grade water for mass spectrometry. Analyses of the metabolites were performed at the Mass
Spectrometry Facility at BIDMC according to the previously published protocol
(Emerling et al., 2013).
Western blotting
Cells were quickly rinsed with cold phosphate-buffered saline (PBS) while adherent, and then lysed in NP-40 lysis buffer (50 mM Tris pH 7.8, 150 mM NaCl,
0.5% NP40, Roche cOomplete EDTA-free protease inhibitor cocktail tablet (1 tablet per 25ml) added). Protein concentration was quantified using Bradford assay (BIO-
RAD, CA, USA), resolved with SDS-polyacrylamide gel electrophoresis and transferred on to a nitrocellulose membrane. The membranes were probed overnight at 4 ˚C with the appropriate primary antibody. The images of the western blot were quantified using ImageJ software. Antibodies used were as follows:
PI5P4Kα (#5527), PI5P4Kβ (#9694), phospho-p-Akt-S473 (#4060), p-Akt-T308
78 (#4056), Akt (#4691), p-p70-S6K (#9206), p70-S6K (#2708) (Cell Signaling
Technology, Inc., MA, USA), β-actin (ab6276) (Abcam, Cambridge, UK), PI5P4Kγ
(#HPA028658) (Sigma, USA).
Biotin-compound pull down assay
HEK293T cells incubated with either DMSO or 1µM PIP4Kin1/2 for 16 hours at 37 oC. Intact cells were quickly washed with cold phosphate-buffered saline
(PBS) for three times and then lysed in NP-40 lysis buffer (50 mM Tris pH 7.8, 150 mM NaCl, 0.5% NP40, Roche cOomplete EDTA-free protease inhibitor cocktail tablet (1 tablet per 25ml) added). Insoluble material was cleared by centrifugation at
13,000 rpm and the supernatants were incubated with 1µM biotin-PIP4Kin1 for additional 16 hours at 4 oC. Streptavidin-sepharose (# 17-5113-01, GE Healthcare,
USA) was added according to the manufacturer’s protocol. After 3 hours of incubation at 4 oC, samples were centrifuged at 5,000 rpm at 4 oC for 1 minute and the pellet underwent 6 washes with 1 mL of lysis buffer or 1 mL of lysis buffer containing 6M urea. 2X SDS buffer was added to the pellet after the final wash and the samples were boiled at 95 oC and subjected to SDS-PAGE.
Preparation of lysates from mouse tissues and xenograft tumors
After final dosing (4, 8 and 12 hours), the mice were euthanized and livers and xenograft tumors were collected. Tissues were homogenized in pre-chilled NP-
40 lysis buffer (50 mM Tris pH 7.8, 150 mM NaCl, 0.5% NP40, Roche cOmplete
EDTA-free protease inhibitor cocktail tablet (1 tablet per 25ml) added). Protein
79 content was quantified using Bradford assay (BIO-RAD, CA, USA).
Mass spectrometry analysis of PI5P4Kβ
PI5P4Kβ (2 µg) was labeled for 3 hrs at room temperature with a 25-fold excess of PIP4Kin1. Labeled protein was reduced with DTT (10 mM final concentration, 30 min 56 °C), alkylated with iodoacetamide (22.5 mM final concentration, 30 min room temperature protected from light) and digested with 100 ng trypsin overnight at 37 °C. Peptides were desalted by C18 and analyzed by nanoLC-MS using a NanoAcquity UPLC system (Waters, Framingham, MA) interfaced to a hybrid quadrupole/ion trap/Orbitrap mass spectrometer (Orbitrap
Fusion, ThermoFisher Scientific, San Jose, CA). Peptides (~300 ng) were injected onto a self-packed precolumn (4 cm POROS 10 R2, 100 µm ID), resolved on an analytical column (30 µm fused silica packed with 50 cm 5 µm Monitor C18; 5-60%
B in 40 minutes; A=0.2 M acetic acid in water, B=0.2 M acetic acid in acetonitrile; flow rate ~30 nL/min) and introduced to the mass spectrometer by electrospray ionization (spray voltage = 3.8 kV). The mass spectrometer was operated in data dependent mode such that the top 10 most abundant ions in each MS scan (image current detection, resolution = 120,000 at m/z 200) were subjected to MS/MS fragmentation by CAD (quadrupole isolation width 1.6 Da, collision energy=30%, electron multiplier detection) and HCD (quadrupole isolation width 1.6 Da, collision energy=30%, image current detection, resolution = 7500 at m/z 200). MS/MS spectra were converted to .mgf and matched to peptide sequences using Mascot version 2.2.1. Search parameters specified a precursor tolerance of 10 ppm,
80 product ion tolerances of 25 mmu and 0.6 Da for HCD and CAD MS2 spectra, respectively. Additional parameters specified cysteine residues to be modified by carbamidomethylation or PIP4Kin1 and variable oxidation of methionine.
Crystalizaton of PI5P4Kβ and PIP4Kin2
Protein expression and purification. Full-length human phosphatidylinositol 5- phosphate 4-kinase type-2 beta in the pGEX2T vector was overexpressed in E. coli
BL21 (DE3) in LB medium in the presence of 50 mg/ml of Ampicillin. Cells were grown at 37°C to an OD of 0.8, induced overnight at 17°C with 500 µM isopropyl-1- thio-D-galactopyranoside, collected by centrifugation, and stored at -80°C. Cell pellets were sonicated in buffer A (1x PBS + 10% glycerol + 1mM EDTA + 10mM
DTT) and the resulting lysate was centrifuged. GSH resin was mixed with lysate supernatant for 2 hrs and washed with buffer A. An 15 uL amount of Thrombin was added to the column, and it was rotated overnight in the cold room. The sample was then eluted and concentrated before passing through an Superdex-200 10/300 column in buffer D (20mM hepes7.5 + 500mM NaCl + 1mM EDTA + 10mM DTT).
Fractions were pooled, concentrated to 15 mg/ml, and frozen at -80°C.
Crystallization and data collection. A three-fold excess of PIP4Kin2 (in
DMSO) was mixed with 300 µM protein and crystallized by sitting-drop vapor diffusion at 20 °C in the following crystallization buffer: 10% PEG4000, 0.2M NaCl,
Mes-pH6.5. Crystals were transferred briefly into crystallization buffer containing
25% glycerol prior to flash-freezing in liquid nitrogen. Diffraction data from complex crystals were collected at beamline 24ID-E of the NE-CAT at the Advanced Photon
81 Source (Argonne National Laboratory).
Pharmacokinetics
Group 1: i.v. and Group 2: p.o. with each group comprising of nine mice.
PIP4Kin2 solution formulation in 5% NMP, 5% solutol HS-15 in normal saline was administered intravenously via tail vein to Group 1 at a dose of 2 mg/kg and was administered orally to Group 2 at a dose of 10 mg/kg. Blood samples were collected at Pre-dose, 0.08, 0.25, 0.5, 1, 2, 4, 8 and 24 hr (i.v.) and Pre-dose, 0.25, 0.5, 1, 2,
4, 6, 8 and 24 hr (p.o.). Plasma samples were separated by centrifugation of whole blood, processed for analysis by protein precipitation using acetonitrile and analyzed with fit-for-purpose LC/MS/MS method (LLOQ – 1.23 ng/mL for plasma).
Pharmacokinetic parameters were calculated using the non-compartmental analysis tool of Phoenix WinNonlin. The study was performed at Sai Life Sciences (India).
Maximum tolerance dose (MTD) study
The mice were implanted with estrogen pellets (17β-estradiol, 60 day release, Innovative Research of America, Sarasota, Florida, USA) at the right flank, and the day of estrogen pellet implantation was day 0. PIP4Kin2 was prepared in
5% NMP, 5% solutol HS-15 in normal saline. Dosing began at day 7 with 5, 10 and
20 mg/kg of PIP4Kin2 (Group 1, 2 and 3, respectively, with 3 mice per group) via intraperitoneal injection. The body weights were recorded daily during the first week, and three times during the second week. Animal death survival was monitored daily. During these routine checks, behavioral parameters, including mobility, visual
82 estimation of food and water consumption, body weight gain/loss (body weights were measured daily post randomization in the first week, and three times at the second week), eye/hair matting and any other abnormal effects, were examined.
The tolerated dose was defined as the dose that resulted in less than 20% mean body-weight loss, and no treatment related death during the study. The study was performed at Crown Bioscience (Taicing, China)
Efficacy study using orthotopic dual xenograft model
All mice were implanted with estrogen pellets (17β-estradiol, 60 day release,
Innovative Research of America, Sarasota, Florida, USA) at the right flank one day prior to tumor inoculation. Each mouse was inoculated at the left mammary fat pad with MCF-7 cells (1 x 107) in 0.1ml PBS with matrigel (1:1) and the right mammary fat pad with BT474 cells (1 x 107) in 0.1ml of PBS with matrigel (1:1) for tumor development. The treatments were initiated when the mean tumor size of BT474 tumor reached 209 mm3 and MCF-7 tumor reached 178 mm3. PIP4Kin2 solution was prepared as described above in the MTD study. Dosing schedule is shown in the following experimental design table. 3 mice in each group were terminated at day 48 for sample collection.
Dose Dosing Group Number Treatment Schedule (mg/kg) Route 1 8 Vehicle 0 i.p. 4 weeks 3 i.p. 8 days (on day 21~28) 2 8 PIP4Kin2 4 i.p. 6 days (on day 29~34) 6 i.p. 14 days (on day 35~48) The study was performed at Crown Bioscience (Taicing, China)
83 Acknowledgements:
We thank Drs. Tinghu Zhang and Nathanel Gray for synthesizing PIP4Kin1 and PIP4Kin2. We thank Drs. Sirano Dhe-Paganon and Hyuk-Soo Seo for crystalizing PI5P4Kβ and PIP4Kin2. We thank Dr. Scott Ficarro for performing mass spectrometry. We thank Dr. Yuxiang Zheng for assistance with performing HPLC to measure phospholipids and Dr. Gina DeNicola for assistance with measuring ROS in cells. We thank Shivan Ramsamooj for assistance with culturing tissues and performing western blots.
84 Chapter 4. Summary and Future Directions
The work presented in this thesis uncovers a role for the poorly understood protein, PI5P4Kγ in suppressing autoimmunity (Chapter 2) and reveals a novel approach for irreversibly inhibiting members of the PI5P4K family of enzymes through designing ATP site inhibitors with an acrylamide moiety that covalently reacts with Cys residues in the regulatory loop of these enzymes (Chapter 3).
In Chapter 2, we have shown that mice lacking PI5P4Kγ in all tissues are viable. Pip4k2c-/- mice are normal in size (Figure S2.2 and S2.3), have normal insulin sensitivity (Figure S2.4 and S2.5) and reproduce normally. However, we found that Pip4k2c-/- mice have a hyperactivated immune system with an increase in immune infiltrating cells in multiple organs with concomitant increases in plasma levels of pro-inflammatory Th1-type cytokines. T cells were more activated in these mice as displayed by an increased Th cell population and a decreased Treg population in spleen with increased mTORC1 signaling in the tissues investigated.
Pip4k2c-/- mice treated with rapamycin mitigated the inflammatory phenotype previously described in parallel with a decrease in mTORC1 signaling. Our current model is that loss of PI5P4Kγ results in enhanced mTORC1 signaling, which then activates T cells and causes an inflammatory phenotype in multiple organs (Figure
1.2). These results suggest that the autoimmune patients with the SNP in PIP4K2C might have a lower expression of PI5P4Kγ. To investigate this possibilities, blood drawn from these autoimmune patients can be processed to quantify the mRNA expression level as well as the protein level of PI5P4Kγ. If PI5P4Kγ expression is
85 indeed decreased, inhibition of mTORC1 signaling might be a beneficial option for these patients.
The molecular mechanism by which PI5P4Ks regulate mTORC1 signaling is another question that remains to be answered. Our current model is that the levels of PI-5-P and/or PI-4,5-P2 controlled by PI5P4Kα and PI5P4Kβ might affect the
PI3K-Akt-mTORC1 signaling pathway, potentially by affecting the recycling of receptor tyrosine kinases such as the insulin receptor. In addition, the previous observation that mTORC1 phosphorylates PI5P4Kγ at the regulatory loop (Mackey et al., 2014) suggests that a fraction of this enzyme is located at the lysosome where mTORC1 is localized and regulated. So it is possible that PI5P4Kγ modulates the ability of PI5P4Kα and PI5P4Kβ to suppress PI-5-P and elevate PI-
4,5-P2 in the lysosomal membrane and thereby modulate mTORC1 regulation.
Deletion of the sole PIP4K2 gene in flies resulted in a severe decrease in mTORC1 signaling (Gupta et al., 2013), supporting the model that we are proposing. PI-4,5-
P2 has been shown to be necessary for budding of vesicles from lysosomes and autophagosomes (Rong et al., 2012) but further studies are needed to clarify the role of phosphoinositides in these organelles.
To determine the genetic interactions among the PI5P4K isoforms, we crossed Pip4k2c-/- mice with Pip4k2a-/- mice and Pip4k2b-/- mice. We generated
Pip4k2a-/- Pip4k2c-/- mice while we could not recover Pip4k2b-/- Pip4k2c-/- mice
(Table S2.1, unpublished). These results suggest that although PI5P4Kγ is not essential when PI5P4Kβ is present, it becomes essential when PI5P4Kβ is deleted.
In fact, when PI5P4Kβ is absent due to gene deletion, then both PI5P4Kα and
86 PI5P4Kγ become essential for viability. Additionally, it will be intriguing to perform timed matings to investigate when the developmental arrest occurs in the Pip4k2b-/-
Pip4k2c-/- embryos and why co-deletion of Pip4k2b and Pip4k2c results in synthetic lethality during embryonic development. Interestingly, the Pip4k2a-/- Pip4k2b-/- embryos are perfectly formed at day E18 but all of the mice die within 12 hours of birth (Emerling et al., 2013). This phenotype is consistent with phenotypes of mice in which genes involved in autophagy are deleted (Kuma et al., 2004, Komatsu et al.,
2005 and Sou et al., 2008).
It should be noted that the studies described in Chapter 2 do not address whether the alterations in T cells observed in Pip4k2c-/- mice are T cell autonomous.
They could be a consequence of defects in Pip4k2c in other immune cells or even in non-immune tissues. Since we generated a floxed form of Pip4k2c it would be possible to address this question by generating mice with T cell specific deletion of
Pip4k2c (e.g. using LCK-Cre).
In Chapter 3, we identified pan-PI5P4K inhibitors, PIP4Kin1 and PIP4Kin2.
These molecules selectively inhibited the kinase activities of all three PI5P4K isoforms at 1 µM concentration without affecting the related PIKFYVE and PI4P5Ks at the same concentration. Additional work will be needed to identify other targets of these molecules and determine whether off-target effects of these molecules could contribute to the observed toxicities or could contribute to some of the responses observed. PIP4Kin1 and PIP4Kin2 mimicked the shRNA mediated depletion of PI5P4Kα and β by inhibiting the proliferation of a number of TP53 mutant breast cancer cell lines with little effect on most of the TP53 wild type cell
87 lines examined. One way to obtain more information about off-target effects is to perform quantitative mass spec analysis of the full range of proteins that are pulled down by the PIP4Kin2-biotin–streptavidin bead experiment described in Chapter 3.
By performing the experiment in the presence or absence of urea and before or after incubation of cells with non-biotinylated PIP4Kin2, it should be possible to determine which proteins are specifically bound and which of these are also covalently bound.
In Chapter 3, we showed that these PIP4Kin1 covalently reacted with cysteine residues in a regulatory loop of PI5P4Kα and PI5P4Kβ. The cysteine residues in the regulatory loop are conserved in human and mouse PI5P4Ks, but are not conserved in worms or zebrafish, suggesting that they are not critical for enzymatic activity. Indeed, we observed that mutating Cys293 of PI5P4Kα to Ser or
Cys307 and Cys318 of PI5P4Kβ to Ser did not significantly affect enzymatic activity
(Figures 4.1C). However, a preliminary experiment indicated that these mutant enzymes were not inhibited by incubation with 1 µM PIP4Kin1 or PIP4Kin2 (Figure
4.1C), indicating that these compounds have relatively low affinity for the ATP pocket and are primarily held in place through covalent reactions with the Cys residues in the regulatory loop. Further studies will need to be performed in which the concentrations and duration of incubation is varied to determine how much the reaction with Cys contributes to the inhibition.
88
Figure 4.1: Overexpressing PI5P4Kβ mutant (C307S C318S) partially rescues the impaired growth of BT474 cells by PIP4Kin2 (A) Luminescent cell viability assay in BT474 cells overexpressing indicated mutant proteins that were treated with either DMSO or 1 µM PIP4Kin2. The bar graph shows the ratio (%) of the cell growth with 1 µM PIP4Kin2 to the cell growth with DMSO in 4 days. N=7. Results are means ± standard errors of the means. WT PI5P4Kβ;, wild type PI5P4Kβ; C307S, PI5P4Kβ C307S mutant; C318S, PI5P4Kβ C318S mutant; C307S C318S, PI5P4Kβ C307S C318S double mutant. (B) Western blotting for PI5P4Kα and PI5P4Kβ with cell lysates overexpressing Cys mutants. (C) Kinase assay with Cys mutants treated with either DMSO or 1 µM PIP4Kin2.
89 Figure 4.1 (Continued)
A 40 35 (%)
30 25 20 15 DMSO
PIP4Kin2 10 5 0
β
vector C307S C318S
WT PI5P4K2 C307S C318S
B β β β α vector C318S PI5P4K C307S PI5P4K C307SC318S PI5P4K C293S PI5P4K
PI5P4Kα
PI5P4Kβ
C WT C307S C318S C307SC318S PI5P4Kβ PI5P4Kβ PI5P4Kβ PI5P4Kβ DMSO DMSO DMSO DMSO PIP4Kin2 PIP4Kin2 PIP4Kin2 PIP4Kin2
90 We also have preliminary studies addressing whether expression of the
C307S/C318S double mutant of PI5P4Kβ in BT474 cells provides resistance to
PIP4Kin2. As shown in Figure 4.1A, transient expression of this mutant protein in
BT474 cells provided partial rescue of cell growth in the presence of PIP4Kin2.
Interestingly, neither the C307S single mutant nor C318S single mutant protein provided significant resistance to PIP4Kin2. These results seem inconsistent with the ability of these mutations to provide resistance to PIP4Kin2 in an in vitro assay. This could be because the incubation with PIPKin2 in the in vitro assay was only 30 minutes while the incubation with the cells in culture was 4 days, increasing the probability of PIP4Kin2 reacting with the remaining Cys residue on the regulatory loop of the mutant PI5P4Kβ over this longer period of exposure. This result suggests that reaction of PIPKin2 with either C307 or C318 of PI5P4Kβ is sufficient to cause inhibition of activity. The failure to obtain full rescue of growth with transient expression is likely due to the fact that all the cells did not take up the plasmid, though we cannot rule out the possibility that PIP4Kin2 has toxic effects independent of the inhibition of PI5P4Ks. Cell lines with stable expression of these mutants are being generated to address these questions. It is possible, or even likely, that these mutations will allow cells to more efficiently recover from a brief exposure to over PIP4Kin2.
In summary, the results presented in Chapter 3 and the preliminary studies discussed in this chapter suggest that the effects of PIP4Kin1 and PIP4Kin2 on
TP53 mutant cell lines are due to on-target inhibition of PI5P4Ks. These results suggested that drugs that inhibit both PI5P4Kα and PI5P4Kβ might be useful in the
91 treatment of TP53 mutant cancers. Moreover, PIP4Kin1 and PIP4Kin2 could be great tools to investigate the roles of PI5P4Ks in cell biology, cell signaling and physiology.
92 APPENDIX A: Depletion of a Putatively Druggable Class of Phosphatidylinositol
Kinases Inhibits Growth of p53-Null Tumors
Brooke M. Emerling1, 2, 3, Jonathan B. Hurov4, George Poulogiannis1, 2, Kazumi S.
Tsukazawa1, 2, Rayman Choo-Wing1, 2, 3, Gerburg M. Wulf2, Eric L. Bell5, Hyeseok
Shim1, 2, Katja A. Lamia6, Lucia E. Rameh7, Gary Bellinger3, Atsuo T. Sasaki8,
John M. Asara2, 9, Xin Yuan2, Andrea Bullock2, Gina M. DeNicola1, 2, Jiaxi Song10, 11,
Victoria Brown10, 11, Sabina Signoretti10, 11, Lewis C. Cantley1, 2, 3#
1 Department of Systems Biology, Harvard Medical School, Boston, MA 02115,
USA
2 Division of Signal Transduction, Beth Israel Deaconess Medical Center, Boston,
MA 02115, USA
3 Department of Medicine, Weill Cornell Medical College, New York, NY 10065,
USA
4 Agios Pharmaceuticals, Cambridge, MA 02139, USA
5 Department of Biology, Paul F. Glenn Laboratory, Massachusetts Institute of
Technology, Cambridge, MA 02139, USA
6 Department of Chemical Physiology, The Scripps Research Institute, La Jolla, CA
92037, USA
7 Department of Medicine, Boston University School of Medicine, Boston, MA
02118, USA
93 8 Department of Internal Medicine, University of Cincinnati College of Medicine, UC
Neuroscience Institute, Brain Tumor Center, Cincinnati, OH 45267, USA
9 Department of Medicine, Harvard Medical School, Boston, MA 02115, USA
10 Department of Medical Oncology, Dana-Farber Cancer Institute, Boston, MA
02115, USA
11 Department of Pathology, Brigham and Women’s Hospital, Harvard Medical
School, Boston, MA 02115, USA
# Correspondence should be addressed to Lewis C. Cantley, [email protected]
94 ABSTRACT
Here, we show that a subset of breast cancers express high levels of the type 2 phosphatidylinositol-5-phosphate 4-kinases α and/or β (PI5P4Kα and β) and provide evidence that these kinases are essential for growth in the absence of p53.
Knocking down PI5P4Kα and β in a breast cancer cell line bearing an amplification of the gene encoding PI5P4Kβ and deficient for p53 impaired growth on plastic and in xenografts. This growth phenotype was accompanied by enhanced levels of reactive oxygen species (ROS) leading to senescence. Mice with homozygous deletion of both TP53 and PIP4K2B were not viable, indicating a synthetic lethality for loss of these two genes. Importantly however, PIP4K2A−/−, PIP4K2B+/−, and
TP53−/− mice were viable and had a dramatic reduction in tumor formation compared to TP53−/− littermates. These results indicate that inhibitors of PI5P4Ks could be effective in preventing or treating cancers with mutations in TP53.
95 INTRODUCTION
The phosphoinositide family of lipids includes seven derivatives of phosphatidylinositol (PI) that are formed through the phosphorylation of the 3-, 4-, and 5-positions on the inositol ring. Phosphoinositides have distinct biological roles and regulate many cellular processes, including proliferation, survival, glucose uptake, and migration. Phosphoinositide kinases, phosphatases, and phospholipases spatially and temporally regulate the generation of the different phosphoinositide species, which localize to different subcellular compartments.
Phosphatidylinositol-3,4,5-trisphosphate (PI-3,4,5-P3) is synthesized by phosphoinositide 3-kinase (PI3K) and serves as the plasma membrane docking site for a subset of proteins that have pleckstrin-homology (PH) domains that bind this lipid, including the serine/threonine protein kinase Akt (also known as protein kinase
B or PKB). Akt is a protooncogene that has critical regulatory roles in insulin signaling and cancer progression. Phosphatidylinositol-4,5-bisphosphate (PI-4,5-P2) is the major substrate for class I PI3Ks and has a significant role itself in mediating the localization of proteins to the plasma membrane and in nucleating cortical actin polymerization (Cantley, 2002).
Until 1997, it was thought that PI-4,5-P2 was produced exclusively by phosphorylation of phosphatidylinositol-4-phosphate (PI-4-P) at the 5 position of the inositol ring, a reaction catalyzed by the type 1 PI-4-P 5-kinases (encoded by the genes PIP5K1A, B, and C). Unexpectedly, a second highly related family of PIP kinases (called type 2) was found to produce PI-4,5-P2 by phosphorylating the 4
96 position of phosphatidylinositol-5-phosphate (PI-5-P), a lipid that had been previously overlooked due to its comigration with the much more abundant PI-4-P (
Rameh and Cantley, 1999 and Rameh et al., 1997). The type 2 PIP kinases are not present in yeast but are conserved in higher eukaryotes from worms and flies to mammals. Humans and mice have three distinct genes, PIP4K2A, B, and C encoding enzymes called PI5P4Kα, β, and γ, respectively. The bulk of PI-4,5-P2 in most tissues is almost certainly derived from the type 1 PIP5Ks, yet recent quantitative proteomic studies on cell lines have revealed a higher abundance of
PI5P4Ks than PI4P5Ks ( Nagaraj et al., 2011). This high abundance of the type 2 enzymes may, in part, explain why the substrate PI-5-P is present at very low levels. Although the type 1 PIP kinases generate PI-4,5-P2 at the plasma membrane, the type 2 kinases are located at internal membranes, including the endoplasmic reticulum (ER), Golgi, and nucleus and probably generate PI-4,5-P2 at those locations ( Fruman et al., 1998, Sarkes and Rameh, 2010, Schaletzky et al.,
2003 and Walker et al., 2001). The vast majority of PI-4,5-P2 is located at the plasma membrane, and it is not clear whether the critical function of the type 2 PIP kinases is to generate PI-4,5-P2 at intracellular sites or to maintain low levels of PI-
5-P (or both).
In a previous study, we generated mice in which one of the type 2 PIP kinase genes (PIP4K2B) was deleted in the germline. These mice were viable, exhibited enhanced insulin sensitivity, and enhanced insulin-dependent activation of Akt in skeletal muscle ( Lamia et al., 2004). Paradoxically, despite increased Akt activation the mice were smaller and had decreased adiposity on a high-fat diet.
97 Cell-based assays revealed that PI5P4Kβ (encoded by PIP4K2B) becomes phosphorylated by p38 at Ser326 in response to cellular stresses, such as UV and
H2O2, and that this causes inhibition of the PI5P 4-kinase activity and results in increased cellular PI-5-P levels ( Jones et al., 2006). These studies suggest that the type 2 PIP kinases mediate cellular stress responses downstream of p38
(presumably by altering the PI-5-P/PI-4,5-P2 ratio at intracellular locations) and that under conditions of low stress, these enzymes suppress the PI3K/Akt signaling pathway. It should be pointed out that the type 2 PIP kinases are unlikely to supply
PI-4,5-P2 as a substrate for PI3K because activation of Akt correlates with loss of
PI5P4K activity rather than gain.
In this study, we have interrogated the potential role of type 2 PIP kinases in cancers. We found high levels of either PI5P4Kα or PI5P4Kβ enzymes or both in a number of breast cancer cell lines and, more importantly, found amplification of the
PIP4K2B gene and high levels of both the PI5P4Kα and PI5P4Kβ proteins in a subset of human breast tumors. We found that knocking down the levels of both
PI5P4Kα and PI5P4Kβ in a TP53-deficient breast cancer cell line blocked growth on plastic and in xenografts. This impaired growth correlated with impaired glucose metabolism and enhanced levels of reactive oxygen species (ROS) leading to senescence. The impaired glucose metabolism, despite activation of the PI3K-Akt pathway (which typically enhances glucose metabolism) was paradoxical. The results indicate that PI3K activation is not driving the ROS production but may be an inadequate feedback attempt to restore glucose uptake and metabolism.
98 Finally, to assess the role of type 2 PIP kinases in tumor formation, we generated mice with germline deletions of PIP4K2A and PIP4K2B and crossed these with TP53−/− mice and evaluated tumor formation in all the viable genotypes.
We found that mice with homozygous deletion of both TP53 and PIP4K2B were not viable, indicating a synthetic lethality for loss of these two genes. Importantly, mice with the genotype PIP4K2A−/−, PIP4K2B+/−, TP53−/− were viable and had a dramatic reduction in tumor formation compared to siblings that were TP53−/− and wild-type for PIP4K2B and/or PIP4K2A genes. These results suggest that PI5P4Kα and
PI5P4Kβ could be targets for pharmaceutical intervention in cancers that are defective in TP53.
99 RESULTS
Amplification of PIP4K2B in HER-2/Neu-Positive Breast Cancers and Co-
Occurrence with TP53 Mutation/Deletion
Gene amplification in breast cancer is associated with disease progression, adverse prognosis, and development of drug resistance. PIP4K2B is located in a chromosomal region (17q12) close to ERBB2 (HER2/Neu), which is amplified in about ∼25% of breast cancers and in a smaller fraction of nonsmall cell lung adenocarcinomas, as well as other cancer types including colorectal and renal (
Luoh et al., 2004, Slamon, 1987, Slamon et al., 1989 and Slamon et al., 2001).
Approximately half of the breast tumors that exhibit ERBB2 amplification also exhibit amplification of PIP4K2B ( Figure 1A). For the majority of tumors that have both ERBB2 and PIP4K2B amplified, the two genes are on the same amplicon (
Figure 1B). However, for a significant fraction (27/78) these two genes appear to be on distinct amplicons. Also, tumors were identified that had relatively focal amplification of PIP4K2B without amplification of ERBB2 ( Figure 1A). Amplification of PIP4K2A was only observed in a small fraction of breast cancers (data not shown). Furthermore, it is interesting to note that the genomic alteration of PIP4K2B and TP53 across 66 breast carcinomas indicate a trend of co-occurrence between
PIP4K2B gain/amplification and TP53 mutation/deletion ( Figure 1B). Interestingly, using reverse phase protein array (RPPA) a high-throughput proteomics technology we could correlate PIP4K2B gain/amplification with a significant increase in ERBB2 phosphorylation on tyrosine 1248 (Y1248) and of total ERBB2 levels ( Figure 1C),
100 as well as, with a small but significant decrease in the phosphorylation of Akt on threonine 308 (T308) with no change in total Akt protein levels ( Figure 1D). To further address the link between PIP4K2A/B and TP53 in cancers, we more deeply interrogated the rapidly growing TCGA database of breast cancers. We found that the subgroup of breast cancers that had homozygous deletion of TP53 (analogous to the TP53-null mouse breast cancers discussed below) had significantly higher
PIP4K2A mRNA compared to tumors with two alleles of TP53 or heterozygous loss of TP53 ( Figure 1E). There was also a trend toward higher expression of PIP4K2B and PIP4K2C in the tumors with homozygous deletion of TP53 ( Figure 1E), though these changes did not reach significance.
PI5P4K Expression in Breast Cancer
To evaluate PI5P4Kα and β protein levels in breast cancer, we utilized antibodies against these proteins for immunohistochemistry staining of a breast cancer tissue array (Figures 2A, 2B, and S1 available online; Tables S1 and S2).
As shown in Figure 2, PI5P4Kα expression is detectable in both normal breast and breast cancer, but high levels of expression are found in 74% of tumors and only
29% of normal breast epithelium. This high level of expression is distributed over all the major subtypes. In contrast, PI5P4Kβ was not detected in any of the normal breast epithelial tissue but was highly expressed in 38% of the breast tumors. The subset of tumors with the highest level of expression was the HER2-positive group where 62% had high levels of PI5P4Kβ. Thus, the HER2 subtype has high protein
101 expression, consistent with a high frequency of PI5P4Kβ gene amplification
(Figure 1A).
We also evaluated the total protein expression of both isoforms in a panel of breast cancer cell lines using western blots. PI5P4Kα is expressed in all the breast cancer cell lines that we investigated, whereas PI5P4Kβ is expressed at very low levels in most breast cancer cell lines, with the exception of BT474 cells, which have both ERBB2 and PIP4K2B amplified (14 and 12 alleles, respectively)
(https://cansar.icr.ac.uk/cansar/). Interestingly, the T47D cell line has four alleles of
PIP4K2B and also shows some increased expression ( Figure 2C). In contrast,
AU565 cells have ERBB2 amplified but have only two alleles of PIP4K2B and do not express high levels of this protein ( Figure 2C).
Knockdown of Both PI5P4Kα and β in BT474 Cells Abrogates Cell
Proliferation and Impairs Tumor Growth in a Xenograft Model
Because BT474 cells exhibit amplification of the PIP4K2B gene and express high levels of the protein encoded by this gene (PI5P4Kβ), we examined the effect of knocking down expression of either this gene alone or both PIP4K2B and
PIP4K2A on cell growth. Knocking down expression of either gene alone with small hairpin RNA (shRNA) had little effect on cell growth ( Figures 3 and S2; Table S3), but knocking down expression of both genes caused a dramatic inhibition of cell growth ( Figure 3). Two independent sets of shRNAs targeting these proteins resulted in effective decreases in both PI5P4Kα and β ( Figures 3A and 3B) and resulted in approximately 80% reduction in cell number over a 72 hr period of cell
102 growth ( Figure 3C). Importantly, to address any off target effects of the hairpins, we rescued cell proliferation of the double-knockdown line by expressing a flag-tagged version of mouse PI5P4Kβ that circumvents the shRNA directed against human
PI5P4Kβ ( Figures 3D and 3E).
To address whether loss of p53 function contributes to the impaired growth of BT474 cells in the context of PI5P4Kα and β knockdown, we took advantage of the previous observation that the p53 E285K mutation in these cells is nonfunctional at 37°C but functional at 32°C (Dearth et al., 2007). Consistent with the previous publication, p53 activity was at least partially restored when BT474 cells were cultured at 32 degrees, as assessed by increased expression of the p53 target gene p21. Importantly, the shift to 32°C resulted in a partial rescue of proliferation (Figure 3F). As expected, the control pLKO.1 BT474 cells grew somewhat slower when shifted to 32°C. These results are consistent with a model in which knocking down PI5P4Kα and β only impairs cell growth when p53 is defective.
We also found that knocking down both PI5P4Kα and β in BT474 cells dramatically impaired tumor formation in xenografts (Figures 3G and 3H). The striking differences in tumor formation in the xenografts were due to a dramatic decrease in viable cells in the PI5P4Kα/β knockdown xenograft tumors compared to the BT474 vector control xenograft tumors (Figure 3I), along with a slight decrease in Ki67 staining (Figure 3J). Most strikingly though is the strong increase in p27 (a marker of senescence) (Young and Kaelin, 2008 and Young et al., 2008) in the
PI5P4K α/β knockdown xenograft tumors compared to the BT474 vector control
103 xenograft tumors (Figure 3K). These results support a model in which the loss on
PI5P4K leads to a cell-cycle-arrest phenotype in vitro and a nonviable tumor in vivo.
Furthermore, these data imply that the knockdown of PI5P4K in vivo may lead to senescence.
Knockdown of Both PI5P4Kα and β in BT474 Cells Enhances PI3K Signaling,
Increases ROS and Respiration, and Triggers Senescence
Consistent with our previous observation of increased in vivo Akt activity in response to deletion of the PIP4K2B gene in mice ( Lamia et al., 2004), the knockdown of both PI5P4Kα and β resulted in increased basal Akt phosphorylation in BT474 cells ( Figures 4A and S3). This increase in Akt phosphorylation could be explained by an increase in PI-3,4,5-P3 levels in response to knockdown of both
PI5P4Kα and β ( Figure 4C). Surprisingly, we observed a dramatic decrease in PI-
3,4-P2 in the double-knockdown cells ( Figure 4C). PI-3,4-P2 can be degraded to PI-
3-P by the phosphatase, inositol polyphosphate 4-phosphatase-IIB (INPP4B) (
Gewinner et al., 2009, Norris et al., 1995 and Norris et al., 1997) so we investigated the level of this enzyme in the BT474 cells. The levels of INPP4B were virtually undetectable in BT474 cells prior to the knockdown of PI5P4Kα and β but were quite high following the knockdown, explaining the drop in PI-3,4-P2 levels ( Figures
4B and 4C). These results suggest a negative feedback loop in which INPP4B upregulation is an attempt to suppress PI-3,4-P2 activation of Akt ( Franke et al.,
1997 and Gewinner et al., 2009). We detected no significant change in PI-4,5-P2
104 levels, indicating that the majority of PI-4,5-P2 is produced by type 1 PIP kinases, as previously assumed ( Figure 4C).
The increase in PI3K/Akt signaling, yet decrease in cell proliferation in response to knockdown of both PI5P4Kα and β, seems paradoxical given the established role of PI3K as an oncogene and in promoting cell growth and cell survival. The decrease in cell number was not due to an increased rate of cell death as judged by no change in LDH release from the cells (Figure 4D). However, we observed a significant increase in reactive oxygen species (ROS), in oxygen consumption, and in β-galactosidase staining in the BT474 cells in response to knockdown of both PI5P4Kα and β consistent with ROS-induced senescence
(Figures 4E–4G).
Hyperactivation of the PI3K/Akt pathway due to loss of PTEN (and INPP4B) has previously been shown to induce senescence (Chen et al., 2005, Gewinner et al., 2009 and Nogueira et al., 2008), raising the possibility that activation of Akt in response to PI5P4Kα/β knockdown is responsible for the senescence that we observed in the BT474 cells. However, BT474 cells are defective in p53, which typically mediates PI3K/Akt pathway-dependent senescence. To evaluate the possibility that the ROS and senescence observed in response to knockdown of
PI5P4Kα/β is a consequence of PI3K/Akt activation, we treated these cells with the pan-PI3K inhibitor GDC-0941 to block this pathway (Figure S3). Rather than restoring growth in the PI5P4Kα/β knockdown BT474 cells, GDC-0941 caused a decrease in cell number (Figure 4I). GDC-0941 caused a small elevation in ROS in control BT474 cells and did not lower ROS levels in PI5P4Kα/β knockdown BT474
105 cells (Figure 4H). These data indicate that the ROS and senescence observed in
PI5P4Kα/β knockdown cells is not due to activation of the PI3K/Akt pathway and raise the possibility that activation of this pathway is an attempt to rescue the cells from ROS.
In contrast to the results with BT474 cells, knocking down PI5P4Kα and β in a breast cancer cell line (MCF7 cells) that does not have PIP4K2B amplified or overexpressed and has wild-type p53 and an activating mutation in PIK3CA had no effect on growth or Akt signaling ( Figure S2). Together, these data suggest that maintaining high levels of PI5P4Kα and β, in part through amplification of the
PIP4K2B gene, is critical to prevent senescence in specific mutational backgrounds
(e.g., HER2 amplification and p53 loss) but is not, in general, essential for the growth of cancer cell lines.
Altered Gene Expression and Metabolomic Signatures in the PI5P4Kα/β
Double-Knockdown BT474 Cells
To further explore the phenotype of knocking down PI5P4Kα and β in BT474 cells, we performed a microarray analysis. As illustrated in the heatmap, there are striking differences in gene expression between the control and PI5P4Kα/β knockdown cells (Figure S4; Table S4), particularly in relation to genes involved in the p38 MAPK pathway (Figure 5A). Consistent with this, we also observe an increase in phospho-p38 MAPK in the PI5P4Kα/β knockdown cells (Figure S3).
These microarray results support the idea that PI5P4Kα and β regulate cellular stress responses in the p38 MAPK pathway and corroborate previous studies
106 indicating that p38 MAPK modulates PI-5-P levels by phosphorylating PI5P4Kβ
(Jones et al., 2006). We next performed gene set enrichment analysis (GSEA) to show that the PI5P4Kα/β double-knockdown cells are significantly enriched in the luminal compared to the basal or mesenchymal-type gene expression signature
(Figure 5B). We also performed gene ontology analysis of the microarray data between the control versus shPI5P4Kα/β double-knockdown cells to show that the most differentially expressed hits are genes that control the rate and direction of cellular metabolism (Figure S5; Table S5).
To better understand the effect of knocking down PI5P4Kα and β on cellular metabolism in BT474 cells, we utilized targeted mass spectrometry to examine the level of 180 metabolites. We found a significant drop in intermediates in glucose metabolism in the PI5P4Kα/β knockdown cells (Figure S5; Table S5). These results are paradoxical. In most cellular contexts, an increase in PI3K/Akt signaling results in increased glycolysis and decreased oxidative phosphorylation (the Warburg effect), but in the context of PI5P4Kα/β knockdown in BT474 cells PI3K/Akt signaling is increased but glucose metabolism is decreased and oxygen consumption is increased (reversal of Warburg effect). The increased oxygen consumption (presumably due to increased oxidative phosphorylation to maintain
ATP levels at low rates of glycolysis) could explain the higher ROS levels (Figures
4E, 4F, and 4H). The metabolic imbalance and consequent high ROS levels are likely to contribute to the senescence observed.
In contrast to the effects seen in BT474 cells, knocking down PI5P4Kα/β in
MCF7 cells did not cause a decrease in glucose metabolites (Figure S2; Table S3).
107 These data suggest that it is the decreased glucose metabolism, unique to BT474 cells, that causes ROS and senescence in these cells. To further interrogate whether the ROS is responsible for the decrease in the growth of PI5P4Kα/β knockdown BT474 cells, we added the antioxidant N-acetylcysteine (NAC) to the media and found that this partially rescued the cell growth (Figure 4I). Consistent with the in vitro data, the PI5P4Kα/β knockdown xenograft tumors have higher levels of 8-hydroxydeoxyguanosine (8-oxodGuo), an indicator of ROS-dependent
DNA damage, than the control BT474 tumors, indicating that oxidative stress is higher in the PI5P4Kα/β knockdown tumors (Figure 4J).
The strong correlation between high expression of PI5P4Kα or β in human breast cancers that lack both alleles of p53 and the observation that knocking down
PI5P4Kα/β in a cell line that lacks p53 (BT474) results in senescence, whereas knocking down these genes in a cell line with wild-type p53 (MCF7) had no effect raises the possibility that expression of PI5P4Kα/β genes may only be essential under conditions of stress and that they may be particularly important for growth of cancers that lack p53. To explore this possibility, we generated mice with genetic deletions of PIP4K2A and crossed them into PIP4K2B−/− and TP53−/− backgrounds.
PIP4K2A−/− Mice Are Viable and Appear Normal
We generated mice deficient for PIP4K2A, using a conditional targeting strategy ( Figure 6A). Clones were picked and analyzed for homologous recombination into the endogenous PIP4K2A locus using a combination of
Southern blot ( Figure 6B) and PCR analysis ( Figure 6C). A representative NdeI
108 digest, followed by Southern blot analysis of one positive and one negative neomycin resistant clone, is shown using the indicated E3 probe ( Figures 6A and
6B). A representative PCR using three mice is seen in Figure 6C, indicating the expected products from germline deleted PIP4K2A−/−, PIP4K2A+/−, and PIP4K2A+/+ mice. Western blot analysis of brain tissue from PIP4K2A−/− mice using an antibody that recognizes both PI5P4Kα and β indicated that the gene product PI5P4Kα was not detected after germline deletion, nor were any truncated proteins ( Figure 6D).
PIP4K2A−/− mice were born in near Mendelian ratios and displayed no obvious histological, growth, or reproductive phenotypes (data not shown). Also, in contrast to PIP4K2B−/− mice that are growth retarded with decreased adiposity and increased insulin sensitivity ( Lamia et al., 2004), PIP4K2A−/− mice were like wild- type littermates in regard to all these characteristics (data not shown).
PIP4K2A−/−PIP4K2B−/− Mice Develop Normally, but Exhibit Neonatal Lethality
The absence of a significant phenotype for the PIP4K2A−/− mice led us to breed these mice to the PIP4K2B−/− mice in order to generate PIP4K2A−/−,
PIP4K2B−/− mice. The cross and backcross of PIP4K2A+/− mice to PIP4K2B+/− mice resulted in non-Mendelian ratios of offspring ( Figure 6F). Most notably, no pups of the PIP4K2A−/−PIP4K2B−/− genotype were viable: these pups looked normal at birth and were able to suckle but died within 12 hr. There was also a sub-Mendelian ratio of viable mice of the PIP4K2A+/−, PIP4K2B−/− genotype (19 out of 43 expected).
These mice were born at Mendelian ratios with normal appearance and weight, but approximately half died within 12 hr. The survivors were growth retarded with only
109 50% the weight of wild-type littermates at the adult stage ( Figure 6G). All other genotypes appeared at Mendelian ratios. No histological abnormalities were observed in any of the genotypes. These observations indicate that type II PI5P4Kα and β are not essential for normal embryonic development but are critical for surviving the stresses of birth and contribute to growth following birth. Consistent with normal embryonic development, PIP4K2A−/−PIP4K2B−/− mouse embryonic fibroblasts (MEFs) grow at a normal rate (data not shown). Also, consistent with enhanced PI3K/Akt signaling in the PI5P4Kα/β knockdown BT474 cell lines (above) and in muscle from PIP4K2B−/− mice ( Lamia et al., 2004), introduction of cre recombinase to PIP4K2Aflx/−PIP4K2B−/− MEFs to delete the second allele of
PIP4K2A resulted in prolonged Akt phosphorylation at Thr308 ( Figure S6).
Conversely, overexpressing PIP4K2A in the same background suppresses Thr308 phosphorylation ( Figure S6). Consistent with PI-5-P being the substrate of the
PI5P4Ks, the PIP4K2A−/−, PIP4K2B−/− MEFs had higher levels of PI-5-P than the
PIP4K2A+/−, PIP4K2B+/− MEFs derived from littermates ( Figure 6H). In contrast to the BT474 cells where knockdown of these genes caused senescence, no senescence was observed in the PIP4K2A−/−, PIP4K2B−/− MEFs.
PI5P4K Deficiency Reduces Tumor-Dependent Death in TP53−/− Mice
Previous studies have indicated a major role for p53 in mediating cellular responses to stress, especially metabolic stress and ROS stress (Vurusaner et al.,
2012). Our observation of increased metabolic stress and ROS, leading to senescence upon knockdown of PI5P4Kα and β in a cell line lacking p53 (BT474)
110 but no induction of senescence in MEFs in the context of deletion of PIP4K2A and PIP4K2B, raised the possibility that partial loss of PIP4K2A and/or PIP4K2B alleles, although viable in normal tissues, may result in senescence in the context of p53-deleted tumors. To test this idea, we crossed the PIP4K2A−/−, PIP4K2B+/− mice with TP53−/− mice and monitored tumor formation in all the viable mice that emerged from the backcrosses. In approximately 4–6 months, mice deficient in p53 develop spontaneous tumors, with the majority being lymphomas and soft tissue sarcomas ( Jacks et al., 1994). In contrast, we find that PIP4K2A−/−, PIP4K2B+/−,
TP53−/− mice that emerged from the crosses had a dramatic reduction in tumors compared to PIP4K2A+/+, PIP4K2B+/+, TP53−/− mice ( Figures 7 and S7).
Deletion of PIP4K2B and TP53 Results in Synthetic Lethality
Although PIP4K2B−/− mice are viable and appear at Mendelian ratios and have a normal lifespan in TP53 wild-type backgrounds ( Lamia et al., 2004), no mice with the PIP4K2B−/−, TP53−/− genotype emerged from the crosses, indicating synthetic lethality upon loss of TP53 and PIP4K2B. In order to investigate why
PIP4K2B−/−, TP53−/− mice are not viable we set up timed mating pairs and at E12.5, we were able to get one double-knockout embryo out of 11 from three litters.
Deletion of both alleles of the PIP4K2B and TP53 genes caused exencephaly (data not shown), a failure of neural tube closure that is occasionally observed in TP53−/− mice, but that does not usually cause early embryonic lethality ( Jacks et al., 1994).
In contrast, loss of only one allele of PIP4K2B and both alleles of PIP4K2A in the
111 context of p53 deficiency results in viable mice but an apparent synthetic lethality for the TP53−/− tumors.
112 DISCUSSION
Here, we have shown that the PIP4K2B gene is amplified in a subset of tumors, especially in HER2-positive breast tumors, and that, although this gene is often in the same amplicon with ERBB2, in many tumors it is in a separate amplicon, and in some tumors it is amplified independent of ERBB2 amplification.
Importantly, we have found that the protein product of this gene, PI5P4Kβ, is highly expressed in a majority of HER2-positive tumors. The related enzyme, PI5P4Kα, which is more broadly expressed, was also found to be elevated in many breast tumors. We showed that BT474 cells, which exhibit amplification of both ERBB2 and PIP4K2B and loss of p53, express high levels of both PI5P4Kα and β and that knocking down the expression of these proteins results in impaired glucose metabolism, increased ROS, and senescence. While this paper was in review,
Jones et al. (2013) was published, demonstrating that addition of peroxide to cells increases PI-5-P levels (especially in cells lacking p53). Furthermore, overexpression of PI5P4Kα prevented the elevation in PI-5-P and also rescued cell growth in the presence of ROS providing independent evidence of a role for
PI5P4Ks in rescuing cells from ROS toxicity.
The biochemical mechanism by which PI5P4Ks protect from ROS in the context of p53 deletion is not clear. Our studies suggest that high levels of PI5P4Ks are required to maintain high rates of glucose metabolism, thereby reducing rates of
ROS production from oxidative phosphorylation and enhancing NADPH production via the pentose phosphate pathway. The role of PI5P4Ks in maintaining glucose
113 metabolism appears to only be critical in the context of p53 loss. Cell lines (e.g.,
MCF7 cells and MEFs) with wild-type p53 do not show defects in glucose metabolism, enhanced ROS, or senescence upon loss of PI5P4Ks, and muscle tissue shows enhanced glucose uptake upon loss of PI5P4Kβ (Lamia et al., 2004).
These results suggest that, in the absence of PI5P4Ks, p53 mediates an adaptive response to ROS, by maintaining glucose uptake and flux into the pentose phosphate pathway. Indeed, several studies have indicated an important role for p53 in maintaining glucose and ROS homeostasis, in part via TIGAR (Mor et al.,
2011). Our studies support this concept and argue that when both p53 and
PI5P4Ks are absent cells are not capable of maintaining glucose and ROS homeostasis. Interestingly, some recent studies also indicate a role for p53 in enhancing ROS in some mutational backgrounds (for review, see Vurusaner et al.,
2012).
Jones et al. (2013) have proposed that the critical role of PI5P4Ks is to suppress the high levels of PI-5-P that appear in response to ROS, implying that PI-
5-P is a second messenger that mediates or enhances ROS-dependent cell damage. An alternative possibility, consistent with the data presented in this article, is that PI5P4Ks, when presented with high levels of PI-5-P (due to ROS), generate
PI-4,5-P2 at an intracellular location that facilitates cellular responses that enhance glucose metabolism and suppress ROS.
The fact that knocking down PI5P4Ks both activates the PI3K/Akt pathway and suppresses glucose metabolism (in p53-deficient BT474 cells) is paradoxical.
In other contexts, activation of the PI3K/Akt pathway enhances glucose uptake and
114 metabolism, and, in the context of wild-type p53, deletion of PIP4K2B in mice activates the PI3K/Akt pathway and enhances glucose uptake into muscle ( Lamia et al., 2004). The data presented here, using the pan PI3K inhibitor GDC0941, show that the activation of the PI3K/Akt pathway is not responsible for the ROS and senescence in BT474 cells. It is likely that the PI5P4Ks suppress PI3K/Akt signaling because they provide an alternative mechanism for enhancing glucose metabolism in response to ROS. The activation of PI3K/Akt signaling upon suppressing
PI5P4Ks is clearly not sufficient to restore glucose and ROS homeostasis when p53 is not present.
The decreased tumor incidence in the background of PIP4K2A−/−,
PIP4K2B+/−, TP53−/− compared to TP53−/− alone is particularly interesting with respect to Li-Fraumeni syndrome (germline TP53 mutations). Our results indicate that expression of PI5P4Kα and/or β is critical for the growth of tumors with TP53 mutations or deletions. Thus, coamplification of PIP4K2B with ERBB2 might explain why breast cancers in patients with Li-Fraumeni syndrome show ERBB2 amplifications (HER2 positive) in over 83% of cases as opposed to 16% of age- matched patients with wild-type TP53 ( Wilson et al., 2010).
The results that we present here suggest that PI5P4Kα and β play a critical role in mediating changes in metabolism in response to stress, and, in particular,
ROS stress that occurs in the absence of p53. Germline deletion of either PIP4K2A or PIP4K2B alone resulted in mice with normal lifespans, and germline deletion of both PIP4K2A and PIP4K2B resulted in full-term embryos of normal size and appearance at birth, indicating that these genes do not play a major role in normal
115 embryonic growth and development. Yet, the PIP4K2A−/−, PIP4K2B−/− pups die shortly after birth, consistent with these genes having a role in mediating stress responses known to occur following birth. Importantly, germline deletion of both
PIP4K2B and TP53 resulted in lethality, whereas germline deletion of either gene alone resulted in Mendelian ratios of viable pups. Thus, the genetic studies suggest that TP53 and PIP4K2B have overlapping roles in mediating cellular responses to stress and that, whereas neither gene alone is essential, loss of both genes is not tolerated.
The most exciting observation from these studies in regard to potential new therapies for p53 mutant tumors is that germline deletion of both alleles of PIP4K2A and one allele of PIP4K2B in the context of TP53−/− results in a viable mouse with a dramatic reduction in tumor-dependent death compared to TP53−/− mice that are wild-type for PIP4K2A and B. These results (and studies of the PIP4K2A−/−,
PIP4K2B+/− or PIP4K2A+/−, PIP4K2B−/− mice in the context of wild-type TP53) indicate that normal tissues tolerate well the loss of three out of four alleles of the
PIP4K2A and PIP4K2B genes, but that tumors are not viable in this context.
PI5P4Kα and β are kinases and pharmaceutical companies have shown that it is possible to develop highly specific inhibitors of both protein kinases and lipid kinases. The synthetic lethality that we observe between TP53 loss and loss of these kinases indicates that drugs that target either the enzyme PI5P4Kβ alone or that target both PI5P4Kα and β are likely to be well tolerated and very effective on tumors that have loss of function mutations or deletions of TP53. Our observations with BT474 cells suggest that HER2-positive tumors that have amplifications of
116 PIP4K2B and mutations in TP53 may be particularly sensitive to PI5P4Kα,β inhibitors.
The ERBB2 (Her2) amplicon on chromosome 17 is variable in size and can contain a number of cancer-related genes in addition to the ERBB2 locus (
Figure 1A). Clinically, patients who have tumors with small amplicons confined to the ERBB2 locus have the greatest benefit from ERBB2-directed therapies such as
Trastuzumab, whereas tumors with wider ERBB2 amplicons have poor responses, suggesting coamplification of genes that contribute to Trastuzumab resistance (
Morrison et al., 2007). PIP4K2B may be a candidate for an adjacent coamplified gene that confers Trastuzumab resistance, and, conversely, concomitant inhibition of ERBB2 and PIP4K2B could be a highly effective treatment option for ERBB2
(Her2)-positive tumors that are p53 mutant and PIP4K2B amplified.
117 METHODS
Cell Culture
All cells were incubated in a 37°C or 32°C humidified incubator with 5% CO2.
All breast cancer cell lines were obtained from ATCC and were cultured in
Dulbecco’s modified Eagle medium (Mediatec). MEFs were cultured in Dulbecco’s modified Eagle medium (Mediatech). All the media was supplemented with 10% fetal bovine serum (Gemini Bio-Products), 100 U/mL penicillin/streptomycin
(Mediatech). Mouse embryonic fibroblasts (MEFs) were isolated from E13.5 embryos as described previously (Hurov et al., 2001).
Virus Production and Infection
293T packaging cell line was used for lentiviral amplification, and all lentiviral infections were carried out as previously described (Moffat et al., 2006). In brief, viruses were collected 48 hr after infection, filtered, and used for infecting cells in the presence of 8 µg/mL polybrene prior to puromycin selection. All lentiviral vectors were obtained from Broad Institute TRC shRNA library. PI5P4Kα pLK0.1 shRNA sequence 1 is TRCN0000006009 and sequence 2 is TRCN0000006010. PI5P4Kβ pLK0.1 shRNA sequence 1 is TRCN0000006013, and sequence 2 is
TRCN0000006017. The pLK0.1 vector was used as the control. To generate virus using the above-described transfer vector, we used pMD2.G (Addgene plasmid
12259) and psPAX2 (Addgene plasmid 12260). Mouse PI5P4Kβ (GenBank number
BC047282) was cloned into the pLNCX2 retroviral vector along with a Kozak
118 sequence and a FLAG tag at the N terminus of the cDNA using in-fusion cloning reagents and the manufacturer’s cloning protocol (Clontech Laboratories). BT474 were infected and stable lines were selected with 1 mg/mL of G418 for 2 weeks.
Post-G418 selection, knockdown of PI5P4Kα and PI5P4Kβ were performed as above. BT474 cells were selected with 2 µg/mL of puromycin.
Immunoblot Analysis and Antibodies
Total cell lysates were prepared by washing cells with cold phosphate- buffered saline, and then the cells were lysed with buffer containing 20 mM Tris/HCl
(pH 7.5), 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, and 1% Triton as well as protease and phosphatase inhibitors. Protein was measured using the Bradford assay (Bio-Rad), and at least 50 µg of total cell lysates was run on a SDS- polyacrylamide gel electrophoresis. The proteins were transferred on to a nitrocellulose membrane, and membranes were probed overnight at 4°C with the appropriate primary antibody. Antibodies used were as follows: PI5P4Kα/β (Schulze et al., 2006), Akt (Santa Cruz Biotechnology), α-tubulin (Sigma), GAPDH (Abcam),
PI5P4Kβ, phosphoAkt-473, phosphoAkt-308, phosphoPRAS40, phosphoERK, phospho-p38MAPK, ERK, p38MAPK, INPP4B, and p21 (Cell Signaling
Technology). See also Extended Experimental Procedures.
119 REFERENCE
Bardeesy, N., Sinha, M., Hezel, A.F., Signoretti, S., Hathaway, N.A., Sharpless, N.E., Loda, M., Carrasco, D.R., and DePinho, R.A. (2002). Loss of the Lkb1 tumour suppressor provokes intestinal polyposis but resistance to transformation. Nature 419, 162–167.
Cantley, L.C. (2002). The phosphoinositide 3-kinase pathway. Science 296, 1655–1657.
Chen, Z., Trotman, L.C., Shaffer, D., Lin, H.K., Dotan, Z.A., Niki, M., Koutcher, J.A., Scher, H.I., Ludwig, T., Gerald, W., et al. (2005). Crucial role of p53- dependent cellular senescence in suppression of Pten-deficient tumorigenesis. Nature 436, 725–730.
Dearth, L.R., Qian, H., Wang, T., Baroni, T.E., Zeng, J., Chen, S.W., Yi, S.Y., and Brachmann, R.K. (2007). Inactive full-length p53 mutants lacking dominant wild-type p53 inhibition highlight loss of heterozygosity as an important aspect of p53 status in human cancers. Carcinogenesis 28, 289–298.
Franke, T.F., Kaplan, D.R., Cantley, L.C., and Toker, A. (1997). Direct regulation of the Akt proto-oncogene product by phosphatidylinositol-3,4-bisphosphate. Science 275, 665–668.
Fruman, D.A., Meyers, R.E., and Cantley, L.C. (1998). Phosphoinositide kinases. Annu. Rev. Biochem. 67, 481–507.
Gewinner, C., Wang, Z.C., Richardson, A., Teruya-Feldstein, J., Etemadmoghadam, D., Bowtell, D., Barretina, J., Lin, W.M., Rameh, L., Salmena, L., et al. (2009). Evidence that inositol polyphosphate 4-phosphatase type II is a tumor suppressor that inhibits PI3K signaling. Cancer Cell 16, 115–125.
Hurov, J.B., Stappenbeck, T.S., Zmasek, C.M., White, L.S., Ranganath, S.H., Russell, J.H., Chan, A.C., Murphy, K.M., and Piwnica-Worms, H. (2001). Immune system dysfunction and autoimmune disease in mice lacking Emk (Par-1) protein kinase. Mol. Cell. Biol. 21, 3206–3219.
120 Jacks, T., Remington, L., Williams, B.O., Schmitt, E.M., Halachmi, S., Bronson, R.T., and Weinberg, R.A. (1994). Tumor spectrum analysis in p53-mutant mice. Curr. Biol. 4, 1–7.
Jones, D.R., Bultsma, Y., Keune, W.J., Halstead, J.R., Elouarrat, D., Mohammed, S., Heck, A.J., D’Santos, C.S., and Divecha, N. (2006). Nuclear PtdIns5P as a transducer of stress signaling: an in vivo role for PIP4Kbeta. Mol. Cell 23, 685–695.
Jones, D.R., Foulger, R., Keune, W.J., Bultsma, Y., and Divecha, N. (2013). PtdIns5P is an oxidative stress-induced second messenger that regulates PKB activation. FASEB J. 27, 1644–1656.
Lamia, K.A., Peroni, O.D., Kim, Y.B., Rameh, L.E., Kahn, B.B., and Cantley, L.C. (2004). Increased insulin sensitivity and reduced adiposity in phosphatidylinositol 5-phosphate 4-kinase beta-/- mice. Mol. Cell. Biol. 24, 5080– 5087.
Luoh, S.W., Venkatesan, N., and Tripathi, R. (2004). Overexpression of the amplified Pip4k2beta gene from 17q11-12 in breast cancer cells confers proliferation advantage. Oncogene 23, 1354–1363.
McClintock, D.S., Santore, M.T., Lee, V.Y., Brunelle, J., Budinger, G.R., Zong, W.X., Thompson, C.B., Hay, N., and Chandel, N.S. (2002). Bcl-2 family members and functional electron transport chain regulate oxygen deprivation- induced cell death. Mol. Cell. Biol. 22, 94–104.
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., et al. (2006). A lentiviral RNAi library for human and mouse genes applied to an arrayed viral highcontent screen. Cell 124, 1283–1298.
Mor, I., Cheung, E.C., and Vousden, K.H. (2011). Control of glycolysis through regulation of PFK1: old friends and recent additions. Cold Spring Harb. Symp. Quant. Biol. 76, 211–216. Published online Nov 17, 2011. http://dx.doi.org/10. 1101/sqb.2011.76.010868.
Morrison, L.E., Jewell, S.S., Usha, L., Blondin, B.A., Rao, R.D., Tabesh, B., Kemper, M., Batus, M., and Coon, J.S. (2007). Effects of ERBB2 amplicon size and genomic alterations of chromosomes 1, 3, and 10 on patient response
121 to trastuzumab in metastatic breast cancer. Genes Chromosomes Cancer 46, 397–405.
Nagaraj, N., Wisniewski, J.R., Geiger, T., Cox, J., Kircher, M., Kelso, J., Pa¨ a¨ bo, S., and Mann, M. (2011). Deep proteome and transcriptome mapping of a human cancer cell line. Mol. Syst. Biol. 7, 548.
Nogueira, V., Park, Y., Chen, C.C., Xu, P.Z., Chen, M.L., Tonic, I., Unterman, T., and Hay, N. (2008). Akt determines replicative senescence and oxidative or oncogenic premature senescence and sensitizes cells to oxidative apoptosis. Cancer Cell 14, 458–470.
Norris, F.A., Atkins, R.C., and Majerus, P.W. (1997). The cDNA cloning and characterization of inositol polyphosphate 4-phosphatase type II. Evidence for conserved alternative splicing in the 4-phosphatase family. J. Biol. Chem. 272, 23859–23864.
Norris, F.A., Auethavekiat, V., and Majerus, P.W. (1995). The isolation and characterization of cDNA encoding human and rat brain inositol polyphosphate 4-phosphatase. J. Biol. Chem. 270, 16128–16133.
Rameh, L.E., and Cantley, L.C. (1999). The role of phosphoinositide 3-kinase lipid products in cell function. J. Biol. Chem. 274, 8347–8350.
Rameh, L.E., Tolias, K.F., Duckworth, B.C., and Cantley, L.C. (1997). A new pathway for synthesis of phosphatidylinositol-4,5-bisphosphate. Nature 390, 192–196.
Sarkes, D., and Rameh, L.E. (2010). A novel HPLC-based approach makes possible the spatial characterization of cellular PtdIns5P and other phosphoinositides. Biochem. J. 428, 375–384.
Schaletzky, J., Dove, S.K., Short, B., Lorenzo, O., Clague, M.J., and Barr, F.A. (2003). Phosphatidylinositol-5-phosphate activation and conserved substrate specificity of the myotubularin phosphatidylinositol 3-phosphatases. Curr. Biol. 13, 504–509.
Schulze, H., Korpal, M., Hurov, J., Kim, S.W., Zhang, J., Cantley, L.C., Graf, T., and Shivdasani, R.A. (2006). Characterization of the megakaryocyte demarcation membrane system and its role in thrombopoiesis. Blood 107, 3868–3875.
122 Serunian, L.A., Auger, K.R., and Cantley, L.C. (1991). Identification and quantification of polyphosphoinositides produced in response to platelet-derived growth factor stimulation. Methods Enzymol. 198, 78–87.
Slamon, D.J. (1987). Proto-oncogenes and human cancers. N. Engl. J. Med. 317, 955–957.
Slamon, D.J., Godolphin, W., Jones, L.A., Holt, J.A., Wong, S.G., Keith, D.E., Levin, W.J., Stuart, S.G., Udove, J., Ullrich, A., et al. (1989). Studies of the HER-2/neu proto-oncogene in human breast and ovarian cancer. Science 244, 707–712.
Slamon, D.J., Leyland-Jones, B., Shak, S., Fuchs, H., Paton, V., Bajamonde, A., Fleming, T., Eiermann, W., Wolter, J., Pegram, M., et al. (2001). Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer that overexpresses HER2. N. Engl. J. Med. 344, 783–792.
Vurusaner, B., Poli, G., and Basaga, H. (2012). Tumor suppressor genes and ROS: complex networks of interactions. Free Radic. Biol. Med. 52, 7–18.
Walker, D.M., Urbe´ , S., Dove, S.K., Tenza, D., Raposo, G., and Clague, M.J. (2001). Characterization of MTMR3. an inositol lipid 3-phosphatase with novel substrate specificity. Curr. Biol. 11, 1600–1605.
Wilson, J.R., Bateman, A.C., Hanson, H., An, Q., Evans, G., Rahman, N., Jones, J.L., and Eccles, D.M. (2010). A novel HER2-positive breast cancer phenotype arising from germline TP53 mutations. J. Med. Genet. 47, 771–774.
Young, A.P., and Kaelin, W.G., Jr. (2008). Senescence triggered by the loss of the VHL tumor suppressor. Cell Cycle 7, 1709–1712.
Young, A.P., Schlisio, S., Minamishima, Y.A., Zhang, Q., Li, L., Grisanzio, C., Signoretti, S., and Kaelin, W.G., Jr. (2008). VHL loss actuates a HIF-independent senescence programme mediated by Rb and p400. Nat. Cell Biol. 10, 361–369.
123 ACKNOWLEDGEMENT
We thank Rodrick Bronson and the entire HMS Rodent Histopathology Core for technical help with the mouse histopathology and discussions concerning the project. Also, we would like to thank Jared Johnson for cloning expertise and the
Nikon Imaging Center at Harvard Medical School for help with light microscopy.
This work was supported by NIH grant R01 GM041890 to L.C.C., DF/HCC Career
Development Award to B.M.E., and by a Stand Up to Cancer Dream Team
Translational Research Grant, a Program of the Entertainment Industry Foundation
(SU2C-AACR-DT0209). G.P. is a Pfizer Fellow of the Life Sciences Research
Foundation. NIH DK R01-63219 supports L.E.R.
124 FIGURES
Figure 1. Amplification of PIP4K2B in HER-2/Neu-Positive Breast Cancers and Co-occurrence with TP53 Mutation/Deletion (A) Genomic landscape of PIP4K2B and ERBB2 (HER2) DNA copy number amplifications in cancer. Tumor samples are divided into three groups: group I: those with amplification in PIP4K2B but no or lower level amplification of ERBB2, group II: samples with amplification of both genes but likely to be derived from two different amplicons, and group III: samples with amplification in both genes derived from the same amplicon. Colored bar indicates degree of copy number gain (red). (B) Oncoprints of PIP4K2B and TP53 across 66 breast carcinomas indicating a trend of co-occurrence between PIP4K2B gain/amplification and TP53 mutation/deletion (p value: 0.006306, Fisher’s exact test). Individual genes are represented as rows, and individual cases or patients are represented as columns. Genetic alterations are color-coded with red indicating amplification; pink, DNA copy number gain; light blue, hemizygous deletion and green box, point mutation. Only cases with gain/amplification of PIP4K2B are included. These oncoprints are based on data obtained from the Stand Up to Cancer cBio portal (http://cbio.mskcc.org/su2c-portal/). (C) Box plots indicating that the levels of ERBB2 phosphorylation on tyrosine 1248 (y1248) (left) and of total ERBB2 (right) measured by RPPA are significantly higher in breast carcinomas with PIP4K2B gain/amplification (p value: 1.7 × 10−5, p value: 4.2 × 10−6, respectively). (D) Box plots indicating that the levels of AKT phosphorylation on threonine 308 (T308) (left) measured by RPPA are significantly lower in breast carcinomas with PIP4K2B gain/amplification (p value: 0.007), whereas the total AKT levels remain unchanged (p value: 0.24) Data obtained from the Stand Up to Cancer cBio portal (http://cbio.mskcc.org/su2c-portal/). (E) Upper: box plots indicating the expression of TP53 (blue) and PIP4K2A (red) across different subsets of breast carcinomas divided by the putative DNA copy number status of TP53. (Homdel: homozygous deletion [n = 0], Hetloss: heterozygous deletion (n = 1), diploid: n = 2, gain: n ≥ 3). TP53 mRNA expression is as expected progressively higher across the different subgroups of breast carcinomas with TP53 ploidy status ranging from n = 0 to n ≥ 3 (one-way ANOVA, p = 4.3 × 10−6), whereas PIP4K2A expression is only significantly higher in the subgroup of breast carcinomas with homozygous deletion of TP53 (one-way ANOVA, p = 0.009). Lower: box plots indicating the expression of PIP4K2B (orange) and PIP4K2C (yellow) across different subsets of breast carcinomas divided by the putative DNA copy number status of TP53. The expression of either PIP4K2B or PIP4K2C is not significantly different across the subgroups of breast carcinomas with different ploidy status of TP53 (one-way ANOVA, p > 0.05).
125
126
Figure 2. PI5P4K Expression in Breast Cancer (A) Histograms illustrating expression levels of PI5P4Kα (∗∗∗p = 0.000) or PI5P4Kβ (∗∗p = 0.004) in breast cancer samples. (B) Representative IHC images from breast tumor and normal samples. Scale bar, 100 µM. (C) PI5P4Kα and PI5P4Kβ expression in panel of breast cancer cell lines. See also Figure S1 and Tables S1 and S2.
127
Figure 3. Knockdown of Both PI5P4Kα and β in BT474 Cells Abrogates Cell Proliferation and Fail to Form Tumors in Xenograft Model (A and B) Stable knockdown of PI5P4Kα/β in BT474 cells. shPI5P4kα/β-1 (sequence 1) and shPI5P4kα/β-2 (sequence 2) are two independent hairpins targeted against PI5P4kα and PI5P4kβ. All single knockdowns are sequence 1 (see Experimental Procedures). (C) Luminescent cell viability assay in stable PI5P4Kα/β knockdown cells. Results are from four independent experiments and are represented as the mean value ± SEM. ∗∗∗p value < 0.0001 with two-tailed Student’s t test. (D) Stable overexpression of mouse Flag-tagged PI5P4Kβ in BT474 cells and subsequent stable PI5P4Kα/β knockdown. (E) Mouse Flag-tagged PI5P4Kβ rescues proliferation in PI5P4Kα/β knockdown cells. Results are from three independent experiments and are represented as the mean value ± SEM. (F) Right: Luminescent cell viability assay in stable PI5P4Kα/β knockdown cells cultured at restrictive (37°C) and permissive conditions (32°C) for indicated time points. Results are from three independent experiments and are represented as the mean value ± SEM. Left: Total p53, p21, and GAPDH (loading control) protein levels in BT474 cells cultured at restrictive and permissive conditions for 24 hr. (G) Tumor formation over time in nude mice injected with the BT474 cancer cell line expressing shRNA pLK0.1 control or shRNA PI5P4K α/β. Error bars are SEM. (H) Images of tumors, pLK0.1 control cells (left flank), or shPI5P4Kα/β (right flank) after mice were euthanized. (I) Histogram showing percentages of viable area in PI5P4Kα/β knockdown and control xenografts. Data from two independent experiments were considered and displayed as means ± SD. (J) Histogram displaying expression levels of Ki67-positive cells in shPI5P4Kα/β and pLK0.1 xenografts. Quantification of Ki67-positive cells in shPI5P4Kα/β and pLK0.1 xenografts. Data from two independent experiments were considered and displayed as means ± SD. (K) Upper: histogram illustrating expression levels of p27-positive cells in shPI5P4Kα/β and pLK0.1 xenografts. Quantification of p27-positive cells in shPI5P4Kα/β and pLK0.1 xenografts. Data from two independent experiments were considered and displayed as means ± SD (∗∗∗p < 0.001). Lower: representative images of p27 immunostains are shown. Scale bar, 100 µM. See also Figure S2 and Table S3.
128
129 Figure 4. Knockdown of Both PI5P4Kα and β in BT474 Cells Enhances PI3K Signaling, Increases ROS and Respiration, and Triggers Senescence (A) AKT phosphorylation at serine 473(pS473) and total AKT protein levels in pLK0.1 vector control cells and in shPI5P4kα/β double-knockdown cells. Cells were serum starved overnight and then treated with media with 10% serum for the indicated time points. (B) Total INPP4B protein levels in pLK0.1 vector control cells and in shPI5P4kα/β double-knockdown cells. The MCF7 breast cancer cell line is known to have high expression of INPP4B.(C) BT474 pLK0.1 control vector cells or PI5P4Kα/β knockdown cells labeled with [3H]-inositol for 48 hr. Deacylated lipids were analyzed by HPLC, quantified, and normalized to PI4P levels. Results are from three independent experiments and are represented as the mean value ± SEM. ∗p < 0.01 with two-tailed Student’s t test. (D) Cell death of pLK0.1 control or shRNA PI5P4K α/β double-knockdown cells was assessed by LDH release. Mean values ± SEMs from four independent experiments are shown. (E) ROS determined by incubating pLK0.1 control or shRNA PI5P4K α/β double-knockdown cells with DCFH-DA
(10 µM) for 30 min or a bolus of H2O2 (100 µM) for 15 min as positive control. n = 4 mean ± SEM. ∗∗p < 0.001 with two-tailed Student’s t test. (F) Oxygen consumption of pLK0.1 control or shRNA PI5P4K α/β double-knockdown cells. n = 4 mean ± SEM. ∗∗p < 0.001 with two-tailed Student’s t test. (G) pLK0.1 control or shRNA PI5P4K α/β double-knockdown cells were plated for senescence assay. Images presented from a representative experiment performed in triplicate. (H) ROS determined by incubating pLK0.1 control or shRNA PI5P4K a/b double-knockdown cells with DCFH-DA (10 µM) for 1 hr. Cells were untreated or treated with GDC- 0941 (1 µM) or N-acetylcysteine (NAC) (10 mM) for 24 hr. n = 3 mean ± SEM. ∗∗p < 0.001 with two-tailed Student’s t test. (I) Luminescent cell viability assay in stable (upper) pLK0.1 control cells and (lower) PI5P4Kα/β knockdown cells ± GDC- 0941 (1 µM) or NAC (10 mM) for indicated time points. Results are from three independent experiments and are represented as the mean value ± SEM. (J) Immunohistochemical detection of 8-oxo-dGuo in pLK0.1 and shPI5P4Kα/β xenografts. Two representative images of 8-oxo-dGuo immunostains are shown (mouse #6 and mouse #10). Scale bar,100 µM. See also Figure S3.
130
131
Figure 5. Distinct Gene Expression and Metabolomic Signatures in the PI5P4Kα/β Double-Knockdown Cells Expression data of PI5P4Kα/β knockdown cells (shPI5P4Kα/β) or control vector cells (pLK0.1) using Affymetrix Human Genome U133 Plus (∼40,000 genes). (A) Enrichment analysis of the curated pathways (BioCarta) following PI5P4Kα/β knockdown. The p38 MAPK and RAS signaling pathways were identified as the most significant pathways with respective p values of 1.83 × 10−4 and 0.004. (B) Gene set enrichment analysis (GSEA) signatures highlighting coordinated differential expression of gene sets that are enriched in PI5P4Kα/β knockdown cells. The CHARAFE_BREAST_CANCER_LUMINAL _VS_MESENCHYMAL_UP (genes upregulated in luminal-like breast cancer cell lines compared to the mesenchymal- like ones) and CHARAFE_BREAST_CANCER_ LUMINAL_VS_BASAL_UP (genes upregulated in luminal-like breast cancer cell lines compared to the basal-like ones) were scored among the most significantly enriched gene signatures in PI5P4K knockdown cells, p < 0.001. See also Figures S4 and S5 and Tables S4 and S5.
132
133
Figure 6. Requirement for PIP4K2A and PIP4K2B for Survival and Growth (A) Schematic representation of wild-type PIP4K2A locus, targeting vector, recombined allele and FlpR/CreR-deleted null allele. Neomycin resistance cassette flanked by Frt sites (yellow boxes), exon 2 flanked by loxP sites (red arrows), and diphtheria toxin cassette located 3′ to exon 2. N, NdeI restriction site. (B) Southern blot of embryonic stem cell clones (wild-type and targeted) using probe E3 after NdeI digestion. (C) PCR analysis of genomic DNA derived from PIP4K2A F2 littermates. (D) PI5P4Kα and β protein levels from brain homogenate of PIP4K2A F2 littermates. (E) PI5P4Kα and β protein levels in primary MEFs derived from intercrossing PIP4K2A and PIP4K2B knockout mice. (F) Genotyping results of PIP4K2A/2B double heterozygote interbreeding at weaning (observed numbers in bold, expected in parentheses). (G) Body weight (in grams) measurements of indicated mouse genotypes. Results represented as the mean value ± SEM. ∗∗∗p < 0.0001 with two-tailed Student’s t test. (H) Quantitation of PI-5-P levels in PIP4K2A/2B-deficient relative to PIP4K2A/2B heterozygous MEFs. Results are from three independent experiments and are represented as the mean value ± SEM. See also Figure S6.
134
135
Figure 7. PI5P4K Deficiency Restricts Tumor Death after p53 Deletion Kaplan-Meier plot analysis of tumor free survival (15 PI5P4Kα+/+ PI5P4Kβ+/+ and 20 −/− +/− PI5P4Kα PI5P4Kβ ). ∗p < 0.05 with two-tailed Student’s t test. See also Figure S7.
Supplemental figures and tables
Supplemental Information includes Extended Experimental Procedures, seven figures, and five tables and can be found with this article online at http://dx.doi.org/10.1016/j.cell.2013.09.057.
136 Bibliography
Adams, P. D., P. V. Afonine, G. Bunkoczi, V. B. Chen, I. W. Davis, N. Echols, J. J. Headd, L. W. Hung, G. J. Kapral, R. W. Grosse-Kunstleve, A. J. McCoy, N. W. Moriarty, R. Oeffner, R. J. Read, D. C. Richardson, J. S. Richardson, T. C. Terwilliger and P. H. Zwart (2010). "PHENIX: a comprehensive Python-based system for macromolecular structure solution." Acta Crystallogr D Biol Crystallogr 66(Pt 2): 213-221.
Bienz, M. (2006). "The PHD finger, a nuclear protein-interaction domain." Trends Biochem Sci 31(1): 35-40.
Bultsma, Y., W. J. Keune and N. Divecha (2010). "PIP4Kbeta interacts with and modulates nuclear localization of the high-activity PtdIns5P-4-kinase isoform PIP4Kalpha." Biochem J 430(2): 223-235.
Burden, L. M., V. D. Rao, D. Murray, R. Ghirlando, S. D. Doughman, R. A. Anderson and J. H. Hurley (1999). "The flattened face of type II beta phosphatidylinositol phosphate kinase binds acidic phospholipid membranes." Biochemistry 38(46): 15141-15149.
Carricaburu, V., K. A. Lamia, E. Lo, L. Favereaux, B. Payrastre, L. C. Cantley and L. E. Rameh (2003). "The phosphatidylinositol (PI)-5-phosphate 4-kinase type II enzyme controls insulin signaling by regulating PI-3,4,5-trisphosphate degradation." Proc Natl Acad Sci U S A 100(17): 9867-9872.
Chagoyen, M. and F. Pazos (2011). "MBRole: enrichment analysis of metabolomic data." Bioinformatics 27(5): 730-731.
Ciruela, A., K. A. Hinchliffe, N. Divecha and R. F. Irvine (2000). "Nuclear targeting of the beta isoform of type II phosphatidylinositol phosphate kinase (phosphatidylinositol 5-phosphate 4-kinase) by its alpha-helix 7." Biochem J 346 Pt 3: 587-591.
Clarke, J. H., P. C. Emson and R. F. Irvine (2008). "Localization of phosphatidylinositol phosphate kinase IIgamma in kidney to a membrane trafficking compartment within specialized cells of the nephron." Am J Physiol Renal Physiol 295(5): F1422-1430.
Clarke, J. H., P. C. Emson and R. F. Irvine (2009). "Distribution and neuronal expression of phosphatidylinositol phosphate kinase IIgamma in the mouse brain." J Comp Neurol 517(3): 296-312.
137
Dearth, L. R., H. Qian, T. Wang, T. E. Baroni, J. Zeng, S. W. Chen, S. Y. Yi and R. K. Brachmann (2007). "Inactive full-length p53 mutants lacking dominant wild-type p53 inhibition highlight loss of heterozygosity as an important aspect of p53 status in human cancers." Carcinogenesis 28(2): 289-298.
Delgoffe, G. M., T. P. Kole, Y. Zheng, P. E. Zarek, K. L. Matthews, B. Xiao, P. F. Worley, S. C. Kozma and J. D. Powell (2009). "The mTOR kinase differentially regulates effector and regulatory T cell lineage commitment." Immunity 30(6): 832- 844.
Delgoffe, G. M., K. N. Pollizzi, A. T. Waickman, E. Heikamp, D. J. Meyers, M. R. Horton, B. Xiao, P. F. Worley and J. D. Powell (2011). "The kinase mTOR regulates the differentiation of helper T cells through the selective activation of signaling by mTORC1 and mTORC2." Nat Immunol 12(4): 295-303.
Duvel, K., J. L. Yecies, S. Menon, P. Raman, A. I. Lipovsky, A. L. Souza, E. Triantafellow, Q. Ma, R. Gorski, S. Cleaver, M. G. Vander Heiden, J. P. MacKeigan, P. M. Finan, C. B. Clish, L. O. Murphy and B. D. Manning (2010). "Activation of a metabolic gene regulatory network downstream of mTOR complex 1." Mol Cell 39(2): 171-183.
Elouarrat, D., Y. U. van der Velden, D. R. Jones, W. H. Moolenaar, N. Divecha and A. P. Haramis (2013). "Role of phosphatidylinositol 5-phosphate 4-kinase alpha in zebrafish development." Int J Biochem Cell Biol 45(7): 1293-1301.
Emerling, B. M., J. B. Hurov, G. Poulogiannis, K. S. Tsukazawa, R. Choo-Wing, G. M. Wulf, E. L. Bell, H. S. Shim, K. A. Lamia, L. E. Rameh, G. Bellinger, A. T. Sasaki, J. M. Asara, X. Yuan, A. Bullock, G. M. Denicola, J. Song, V. Brown, S. Signoretti and L. C. Cantley (2013). "Depletion of a putatively druggable class of phosphatidylinositol kinases inhibits growth of p53-null tumors." Cell 155(4): 844-857.
Emsley, P. and K. Cowtan (2004). "Coot: model-building tools for molecular graphics." Acta Crystallogr D Biol Crystallogr 60(Pt 12 Pt 1): 2126-2132.
Fiume, R., Y. Stijf-Bultsma, Z. H. Shah, W. J. Keune, D. R. Jones, J. G. Jude and N. Divecha (2015). "PIP4K and the role of nuclear phosphoinositides in tumour suppression." Biochim Biophys Acta 1851(6): 898-910.
Gelato, K. A., M. Tauber, M. S. Ong, S. Winter, K. Hiragami-Hamada, J. Sindlinger, A. Lemak, Y. Bultsma, S. Houliston, D. Schwarzer, N. Divecha, C. H. Arrowsmith and
138 W. Fischle (2014). "Accessibility of different histone H3-binding domains of UHRF1 is allosterically regulated by phosphatidylinositol 5-phosphate." Mol Cell 54(6): 905-919.
Gozani, O., P. Karuman, D. R. Jones, D. Ivanov, J. Cha, A. A. Lugovskoy, C. L. Baird, H. Zhu, S. J. Field, S. L. Lessnick, J. Villasenor, B. Mehrotra, J. Chen, V. R. Rao, J. S. Brugge, C. G. Ferguson, B. Payrastre, D. G. Myszka, L. C. Cantley, G. Wagner, N. Divecha, G. D. Prestwich and J. Yuan (2003). "The PHD finger of the chromatin-associated protein ING2 functions as a nuclear phosphoinositide receptor." Cell 114(1): 99-111.
Gupta, A., S. Toscano, D. Trivedi, D. R. Jones, S. Mathre, J. H. Clarke, N. Divecha and P. Raghu (2013). "Phosphatidylinositol 5-phosphate 4-kinase (PIP4K) regulates TOR signaling and cell growth during Drosophila development." Proc Natl Acad Sci U S A 110(15): 5963-5968.
Jones, D. R., Y. Bultsma, W. J. Keune, J. R. Halstead, D. Elouarrat, S. Mohammed, A. J. Heck, C. S. D'Santos and N. Divecha (2006). "Nuclear PtdIns5P as a transducer of stress signaling: an in vivo role for PIP4Kbeta." Mol Cell 23(5): 685-695.
Jones, D. R., R. Foulger, W. J. Keune, Y. Bultsma and N. Divecha (2013). "PtdIns5P is an oxidative stress-induced second messenger that regulates PKB activation." FASEB J 27(4): 1644-1656
Kabsch, W. (2010). "Integration, scaling, space-group assignment and post- refinement." Acta Crystallogr D Biol Crystallogr 66(Pt 2): 133-144.
Keune, W. J., D. R. Jones and N. Divecha (2013). "PtdIns5P and Pin1 in oxidative stress signaling." Adv Biol Regul 53(2): 179-189.
Komatsu, M., S. Waguri, T. Ueno, J. Iwata, S. Murata, I. Tanida, J. Ezaki, N. Mizushima, Y. Ohsumi, Y. Uchiyama, E. Kominami, K. Tanaka and T. Chiba (2005). "Impairment of starvation-induced and constitutive autophagy in Atg7-deficient mice." J Cell Biol 169(3): 425-434.
Kuma, A., M. Hatano, M. Matsui, A. Yamamoto, H. Nakaya, T. Yoshimori, Y. Ohsumi, T. Tokuhisa and N. Mizushima (2004). "The role of autophagy during the early neonatal starvation period." Nature 432(7020): 1032-1036.
Lamia, K. A., O. D. Peroni, Y. B. Kim, L. E. Rameh, B. B. Kahn and L. C. Cantley (2004). "Increased insulin sensitivity and reduced adiposity in phosphatidylinositol 5- phosphate 4-kinase beta-/- mice." Mol Cell Biol 24(11): 5080-5087.
139
Laplante, M. and D. M. Sabatini (2013). "Regulation of mTORC1 and its impact on gene expression at a glance." J Cell Sci 126(Pt 8): 1713-1719.
Mackey, A. M., D. A. Sarkes, I. Bettencourt, J. M. Asara and L. E. Rameh (2014). "PIP4kgamma is a substrate for mTORC1 that maintains basal mTORC1 signaling during starvation." Sci Signal 7(350): ra104.
McCoy, A. J., R. W. Grosse-Kunstleve, P. D. Adams, M. D. Winn, L. C. Storoni and R. J. Read (2007). "Phaser crystallographic software." J Appl Crystallogr 40(Pt 4): 658-674.
Mellman, D. L., M. L. Gonzales, C. Song, C. A. Barlow, P. Wang, C. Kendziorski and R. A. Anderson (2008). "A PtdIns4,5P2-regulated nuclear poly(A) polymerase controls expression of select mRNAs." Nature 451(7181): 1013-1017.
Mitsdoerffer, M., Y. Lee, A. Jager, H. J. Kim, T. Korn, J. K. Kolls, H. Cantor, E. Bettelli and V. K. Kuchroo (2010). "Proinflammatory T helper type 17 cells are effective B-cell helpers." Proc Natl Acad Sci U S A 107(32): 14292-14297.
Purkerson, J. and P. Isakson (1992). "A two-signal model for regulation of immunoglobulin isotype switching." FASEB J 6(14): 3245-3252.
Rameh, L. E., K. F. Tolias, B. C. Duckworth and L. C. Cantley (1997). "A new pathway for synthesis of phosphatidylinositol-4,5-bisphosphate." Nature 390(6656): 192-196.
Rao, V. D., S. Misra, I. V. Boronenkov, R. A. Anderson and J. H. Hurley (1998). "Structure of type IIbeta phosphatidylinositol phosphate kinase: a protein kinase fold flattened for interfacial phosphorylation." Cell 94(6): 829-839.
Raychaudhuri, S., E. F. Remmers, A. T. Lee, R. Hackett, C. Guiducci, N. P. Burtt, L. Gianniny, B. D. Korman, L. Padyukov, F. A. Kurreeman, M. Chang, J. J. Catanese, B. Ding, S. Wong, A. H. van der Helm-van Mil, B. M. Neale, J. Coblyn, J. Cui, P. P. Tak, G. J. Wolbink, J. B. Crusius, I. E. van der Horst-Bruinsma, L. A. Criswell, C. I. Amos, M. F. Seldin, D. L. Kastner, K. G. Ardlie, L. Alfredsson, K. H. Costenbader, D. Altshuler, T. W. Huizinga, N. A. Shadick, M. E. Weinblatt, N. de Vries, J. Worthington, M. Seielstad, R. E. Toes, E. W. Karlson, A. B. Begovich, L. Klareskog, P. K. Gregersen, M. J. Daly and R. M. Plenge (2008). "Common variants at CD40 and other loci confer risk of rheumatoid arthritis." Nat Genet 40(10): 1216-1223.
140 Rong, Y., M. Liu, L. Ma, W. Du, H. Zhang, Y. Tian, Z. Cao, Y. Li, H. Ren, C. Zhang, L. Li, S. Chen, J. Xi and L. Yu (2012). "Clathrin and phosphatidylinositol-4,5- bisphosphate regulate autophagic lysosome reformation." Nat Cell Biol 14(9): 924- 934.
Sarkes, D. and L. E. Rameh (2010). "A novel HPLC-based approach makes possible the spatial characterization of cellular PtdIns5P and other phosphoinositides." Biochem J 428(3): 375-384.
Schulze, H., M. Korpal, J. Hurov, S. W. Kim, J. Zhang, L. C. Cantley, T. Graf and R. A. Shivdasani (2006). "Characterization of the megakaryocyte demarcation membrane system and its role in thrombopoiesis." Blood 107(10): 3868-3875.
Schwab, S. G. and D. B. Wildenauer (2009). "Update on key previously proposed candidate genes for schizophrenia." Curr Opin Psychiatry 22(2): 147-153.
Shima, H., M. Pende, Y. Chen, S. Fumagalli, G. Thomas and S. C. Kozma (1998). "Disruption of the p70(s6k)/p85(s6k) gene reveals a small mouse phenotype and a new functional S6 kinase." EMBO J 17(22): 6649-6659.
Skarnes, W. C., B. Rosen, A. P. West, M. Koutsourakis, W. Bushell, V. Iyer, A. O. Mujica, M. Thomas, J. Harrow, T. Cox, D. Jackson, J. Severin, P. Biggs, J. Fu, M. Nefedov, P. J. de Jong, A. F. Stewart and A. Bradley (2011). "A conditional knockout resource for the genome-wide study of mouse gene function." Nature 474(7351): 337-342.
Smart, O. S., T. O. Womack, C. Flensburg, P. Keller, W. Paciorek, A. Sharff, C. Vonrhein and G. Bricogne (2012). "Exploiting structure similarity in refinement: automated NCS and target-structure restraints in BUSTER." Acta Crystallogr D Biol Crystallogr 68(Pt 4): 368-380.
Sou, Y. S., S. Waguri, J. Iwata, T. Ueno, T. Fujimura, T. Hara, N. Sawada, A. Yamada, N. Mizushima, Y. Uchiyama, E. Kominami, K. Tanaka and M. Komatsu (2008). "The Atg8 conjugation system is indispensable for proper development of autophagic isolation membranes in mice." Mol Biol Cell 19(11): 4762-4775.
Stijf-Bultsma, Y., L. Sommer, M. Tauber, M. Baalbaki, P. Giardoglou, D. R. Jones, K. A. Gelato, J. van Pelt, Z. Shah, H. Rahnamoun, C. Toma, K. E. Anderson, P. Hawkins, S. M. Lauberth, A. P. Haramis, D. Hart, W. Fischle and N. Divecha (2015). "The basal transcription complex component TAF3 transduces changes in nuclear phosphoinositides into transcriptional output." Mol Cell 58(3): 453-467.
141 Sun, Y., N. Thapa, A. C. Hedman and R. A. Anderson (2013). "Phosphatidylinositol 4,5-bisphosphate: targeted production and signaling." Bioessays 35(6): 513-522.
Vicinanza, M., V. I. Korolchuk, A. Ashkenazi, C. Puri, F. M. Menzies, J. H. Clarke and D. C. Rubinsztein (2015). "PI(5)P regulates autophagosome biogenesis." Mol Cell 57(2): 219-234.
Wang, M., N. J. Bond, A. J. Letcher, J. P. Richardson, K. S. Lilley, R. F. Irvine and J. H. Clarke (2010). "Genomic tagging reveals a random association of endogenous PtdIns5P 4-kinases IIalpha and IIbeta and a partial nuclear localization of the IIalpha isoform." Biochem J 430(2): 215-221.
Yu, H., K. Fukami, Y. Watanabe, C. Ozaki and T. Takenawa (1998). "Phosphatidylinositol 4,5-bisphosphate reverses the inhibition of RNA transcription caused by histone H1." Eur J Biochem 251(1-2): 281-287.
Zhang, T., F. Inesta-Vaquera, M. Niepel, J. Zhang, S. B. Ficarro, T. Machleidt, T. Xie, J. A. Marto, N. Kim, T. Sim, J. D. Laughlin, H. Park, P. V. LoGrasso, M. Patricelli, T. K. Nomanbhoy, P. K. Sorger, D. R. Alessi and N. S. Gray (2012). "Discovery of potent and selective covalent inhibitors of JNK." Chem Biol 19(1): 140-154.
Zhao, K., W. Wang, O. J. Rando, Y. Xue, K. Swiderek, A. Kuo and G. R. Crabtree (1998). "Rapid and phosphoinositol-dependent binding of the SWI/SNF-like BAF complex to chromatin after T lymphocyte receptor signaling." Cell 95(5): 625-636.
Zheng, Y., A. Chaudhry, A. Kas, P. deRoos, J. M. Kim, T. T. Chu, L. Corcoran, P. Treuting, U. Klein and A. Y. Rudensky (2009). "Regulatory T-cell suppressor program co-opts transcription factor IRF4 to control T(H)2 responses." Nature 458(7236): 351- 356.
142 Supplemental Materials
+/+ +/+ +/+ -/- -/- -/-
PI5P4Kγ
Figure S2.1: Pip4k2c-/- mice do not express either the full length or truncated PI5P4Kγ The expression of PI5P4Kγ in kidney of wild type mice and Pip4k2c-/- mice.
143 35
30
25
20
15 WT male Bodyweight (g) 10 KO male 5 WT female KO female 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Days (started at 4 weeks of age)
60
50
40
30
Bodyweight (g) 20 WT male, HFD KO male, HFD 10 WT female, HFD KO female, HFD 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Days (started at 4 weeks of age)
Figure S2.2: The growth curves of wild type and Pip4k2c-/- mice Mice (started at the age of 4 weeks) were weighted weekly for 16 weeks. (A) mice placed on normal diet. N>9 per group. (2) mice placed on high-fat diet (HFD). N>13 per group. The results are presented as means ± standard errors of the means.
144 0.50 0.40 0.40 0.30 0.30 0.20 0.20 0.10 0.10 Foodintake (g)/ Bodyweight (g)
Foodintake (g)/ Bodyweight (g) 0.00 0.00 WT KO WT KO male male female female
Figure S2.3: Food intake of wild type and Pip4k2c-/- mice The food intake was measured by housing one or two mice of the same genotype per cage and weighing the food in the cages every morning at 11 a.m. every 3 days for two weeks. The amount of food consumed was divided by the average body weight of the mice in each cage. N=5 per group. The results are presented as means ± standard errors of the means.
) WT male
dL 600 KO male 500 400 300 200 100
Bloodglucose (mg/ 0 0 15 30 60 90 120 Time (min) after injection
Figure S2.4: Glucose tolerance test of wild type and Pip4k2c-/- mice After weaning, the 4-weeks old mice were placed on a high fat diet for 5 months. The mice (5 WT mice and 6 KO mice) were i.p. injected with 1 mg of glucose per g or body weight from a 20 mg/ml solution of glucose in 0.9% NaCl. Blood glucose levels were measured with One Touch Basic glucose meter before injection of glucose and at indicated time points. The results are presented as means ± standard errors of the means.
145
300 WT male 250 KO male )
dL 200 150 (mg/ 100 Bloodglucose 50 0 0 15 30 45 60
Figure S2.5: Insulin tolerance test of wild type and Pip4k2c-/- mice The mice were starved from 9 a.m. to 1 p.m. and at 1 p.m the mice were i.p. injected with Novolin-R at a dose of 0.5U per kg of body weight. Blood glucose levels were measured with One Touch Basic glucose meter before injection of glucose and at indicated time points. 3 WT mice and 6 KO mice (6 months old). The results are presented as means ± standard errors of the means.
146 ALT AST Total Bilirubin 400 600 3 200 400 2 u/l u/l 0 mg/dl 200 1 WT KO -200 0 0 -400 WT KO WT KO
BUN Glucose Carbon Dioxide 30 30
400 /l 20 20 mg/dl mg/dl
200 mmol 10 10 0 0 0 WT KO WT KO WT KO
Triglycerides Total Cholesterol 100 100
50 50 mg/dl mg/dl
0 0 WT KO WT KO Figure S2.6: Plasma levels of liver function parameter proteins Plasma of the mice were collected and analyzed for the level of the metabolites indicated. 3 mice per each genotype (12 months old). The results are presented as means ± standard errors of the means.
147
0.200 0.150 0.100
Absorbance 0.050 0.000 WT KO
Figure S2.7: Levels of rheumatoid factor in plasma Rheumatoid factor was measured using mouse Rheumatoid Factor Ig ELISA kit (#6200, Alpha Diagnostic International, USA) according to the manufacturer’s protocol. 3 mice per each genotype (12 months old). The results are presented as means ± standard errors of the means.
WT KO
Figure S2.8: Gross observation of Pip4k2c-/- mouse livers Left: wild type, right: Pip4k2c-/-. 12 months old. Pale color livers were often found in Pip4k2c-/ mice older than 12 months.
Thymus +/+ -/- p-p70-S6K (T389)
p70-S6K
Figure S2.9: p70-S6K is not activated in the thymus of Pip4k2c-/- mice p70-S6K Thr389 phosphorylation and total p70-S6K were blotted for in the thymus lysates. The mice were 12 months old.
148
Pip4k2b +/+ Pip4k2b +/- Pip4k2b -/-
Pip4k2c +/+ 7.8 (3.9) % 15.6 (7.8) % 4.7 (3.9) %
Pip4k2c +/- 15.6 (12.5) % 35.9 (25.0) % 4.7 (12.5) %
Pip4k2c -/- 3.1 (8.6) % 12.5 (17.2) % 0 (8.6) %
Table S2.1: Pip4k2b-/- Pip4k2c-/- mice are not viable at the developmental stage Table showing the expected and observed genotype frequency (%) of 3-wk-old pups from breeding Pip4k2b+/- Pip4k2c+/- with Pip4k2b+/- Pip4k2c-/-, Pip4k2b+/- Pip4k2c+/- or Pip4k2b-/- Pip4k2c+/- mice. Expected % frequency in parentheses. N = 64.
PIP4Kin1 PIP4Kin2 DMSO
PI-4,5-P 2 Figure S3.1: PIP4Kin1 and PIP4Kin2 do not inhibit PI5P4Kγ PI5P4Kγ was incubated with indicated compounds (1µM, 30 minutes, 25oC) and the kinase assays were carried out as described in the Methods. The radiolabeled PI-
4,5-P2 product was visualized by autoradiography.
149
YM201636 CompoundY DMS PIP4Kin2 Compound X
PI-3,5-P 2 Figure S3.2: PIP4Kin1 and PIP4Kin2 do not inhibit PIKFYVE PIKFYVE was incubated with indicated compounds (5 µM, 30 minutes, 25oC) and the kinase assays were carried out as described in the Methods. The lipid substrate was replaced from PI-5-P to PI-3-P. PIKFYVE was a gift from Dr. Jared L. Johnson.
The radiolabeled PI-3,5-P2 product was visualized by autoradiography.
PIP4Kin1 PIP4Kin2 DMSO PI-4,5-P 2 Figure S3.3: PIP4Kin1 and PIP4Kin2 do not inhibit type I PI4P5Kα Type I PI4P5Kα was incubated with indicated compounds (5 µM, 30 minutes, 25oC) and the kinase assays were carried out as described in the Methods. The lipid substrate was replaced from PI-5-P to PI-4-P. The radiolabeled PI-4,5-P2 product was visualized by autoradiography.
150 HCC38 HCC70 8000000 9000000 DMSO 7000000 DMSO 8000000 6000000 PIP4Kin2 7000000 PIP4Kin2 5000000 6000000 5000000
RLU 4000000 4000000 3000000 3000000 2000000 2000000 1000000 1000000 0 0 Day 1 Day 3 Day 5 Day 1 Day 3 Day 5
Figure S3.4: PIP4Kin2 inhibit proliferation of HCC38 and HCC70 cells HCC38 (p53 R273L) and HCC70 (p53 R248Q) cells were treated with 1 µM PIP4Kin2. Luminescent cell viability assay following incubation with the indicated compounds was carried out as described in the Methods. N=8. The results are presented as means ± standard errors of the means.
10000000 DMSO 8000000
PIP4Kin2 6000000 RLU 4000000
2000000
0 day 0 day 3 day 5 Figure S3.5: Extended incubation with PIP4Kin2 does not affect proliferation of MCF7 cells MCF7 cells were treated with either DMSO or 1 µM PIP4Kin2 and were cultured for 1 month with changing media every 2 days and splitting every 4 days. Then the cells were seeded in 96 well plates, and were treated with the indicated compounds and cultured for 5 days to perform luminescent cell viability assay. N=8. The results are presented as means ± standard errors of the means.
151
Figure S3.6: Phosphoinositides affected by PIP4Kin2 HEK293T cells were treated with either DMSO or 1 µM PIP4Kin1 (30 minutes and 3 hours). Cells were labeled with [3H]-inositol for 48 hours. Deacylated lipids were analyzed by HPLC, quantified, and normalized to PI-4-P levels. Results are from 3 independent experiments and are represented as the mean value ± standard errors of the means.
152
PI-3-P /PI-4-P PI-5-P /PI-4-P 0.08 0.02 0.06 0.015 0.04 0.01 0.005 0.02 0 0.00
DMSO, 3h DMSO, 30m DMSO, 3h PIP4Kin1, 3h PIP4Kin1, 30m DMSO, 30m PIP4Kin1, 3h PIP4Kin1, 30m
PI-3,5-P2 /PI-4-P PI-3,4-P2 /PI-4-P PI-4,5-P2 /PI-4-P 0.0035 0.008 3 0.007 0.003 0.006 2.5 0.0025 0.005 2 0.002 0.004 1.5 0.0015 0.003 0.001 0.002 1 0.0005 0.001 0.5 0 0 0
DMSO, 3h DMSO, 3h DMSO, 30m DMSO, 3h DMSO, 30m PIP4Kin1, 3h DMSO, 30m PIP4Kin1, 3h PIP4Kin1, 30m PIP4Kin1, 3h PIP4Kin1, 30m PIP4Kin1, 30m
PI-3,4,5-P3 /PI-4-P Total PIs /PI-4-P 0.02 5 0.015 4 3 0.01 2 0.005 1 0 0
DMSO, 3h DMSO, 3h DMSO, 30m PIP4Kin1, 3h DMSO, 30m PIP4Kin1, 3h PIP4Kin1, 30m PIP4Kin1, 30m
153 24 hours 48 hours PIP4Kin1, 1µM PIP4Kin1, 1µM M M M M µ µ 1 DMSO 1 10nM DMSO 10nM 100nM 100nM p-Akt (T308) p-Akt (S473)
Figure S3.7: 24 hour incubation with PIP4Kin1 decreased Akt activation in BT474 cells BT474 cells were treated with either DMSO or 1 µM PIP4Kin1 (24 hours and 48 hours). Cells were lysed and subjected to SDS-PAGE. Akt Thr308 phosphorylation and Ser473 phosphorylation were detected by immunoblotting.
* 1400 1200 1000 800 600 400 200
Fluorescence Fluorescence intensity 0
Figure S3.8: ROS affected by PIP4Kin1 and PIP4Kin2 BT474 cells were treated with either DMSO or 1 µM PIP4Kin1 or PIP4Kin2 (30 minutes and 90 minutes). ROS determined by incubating the cells with DCFH-DA (10 µM) for 1 hour. Fluorescence quantified by flow cytometry. N=3 or more. The results are presented as means ± standard errors of the means.
154
MCF7
37oC, 48 hours
DMSO 1µM PIP4Kin1
BT474
37oC, 144 hours
DMSO 1µM PIP4Kin1
32oC 144 hours
DMSO 1µM PIP4Kin1
Figure S3.9: PIP4Kin1 did not affect senescence in MCF7 cells and BT474 cells. β-gal staining for MCF7 cells and BT474 cells cultured at 37 oC or 32 oC following incubation with either DMSO or 1 µM PIP4Kin1.
155