Epithelia provide the fundamental building block for nearly all Eumetazoan organs. These tissues provide a barrier between the organism and its surroundings, facilitating absorption and secretion of necessary macromolecules, while protecting the animal from harm. To perform these functions, epithelia polarize along their apicobasal axis, forming an apical domain that faces the environment or lumen, and a basal domain that attaches to the extracellular matrix. Maintenance of apicobasal polarity is critical for normal organ function, and defective architecture has been linked to disease. Most importantly, loss of tissue polarity is associated with increased malignancy of human tumors. However, the mechanisms coupling epithelial polarity to the suppression of tumor formation are unknown.

The identification of regulators of epithelial architecture revealed that the core apical and basolateral determinants are evolutionarily conserved. Thus, the epithelial tissues of Drosophila provide a genetically tractable model for studying the relationship between polarity and tumor suppression. In Drosophila epithelia, disruption of apical membrane regulators leads to cell death, while mutation of basolateral determinants gives rise to the strikingly opposite phenotype: the formation of massively overproliferating apolar tumors that fail to differentiate and exhibit increased metastatic potential. Due to these cancer-like phenotypes, genes encoding the basolateral regulators have been termed neoplastic tumor suppressor genes (nTSGs). Although these data demonstrate an intimate link between polarity and proliferation control, it remains mysterious how disorganization at the plasma membrane leads to the downstream transcriptional changes required for neoplastic tumor formation.

Studies of nTSG mutant tissue have identified several signaling pathways activated upon polarity loss. In Chapter 1, I review the recent findings implicating these pathways and their known downstream transcriptional targets to tumor formation. Specifically, I highlight the roles of the stress-activated Jun kinase (JNK) cascade, the apical regulator, atypical Protein Kinase C (aPKC), and the tumor suppressive Hippo (Hpo) pathway in neoplastic overgrowth. In addition, I discuss the controversies surrounding how these pathways interact to modulate downstream transcription.

A key unanswered question that follows from these findings is: how is polarity status at the plasma membrane transduced into expression alterations of oncogenic target genes in the nucleus? To address this outstanding issue, in Chapter 2, I analyze the gene expression profile of neoplastic tumors and identify transcriptional signatures of neoplasia as well as specific functional targets mediating overproliferation. To assess the interactions of downstream pathways driving transcription, I isolate a polarity-responsive enhancer element at a functionally important locus and evaluate its response to aPKC, JNK and Hpo signaling. My results demonstrate that, although JNK is necessary for neoplastic tumor formation, aPKC is a key mediator of gene expression alterations, via the Hpo pathway transcription factor Yorkie. These findings provide a model for how polarity-sensitive signaling pathways synergize to activate mitogenic target genes, leading to tumor formation.

In Chapter 3, I demonstrate additional regulation of polarity-responsive enhancer elements by the chromatin-modifying Polycomb Group genes, a recently identified TSG class. Further, I use molecular, genetic, and genomic approaches to evaluate the inter-relationship between the polarity-regulating and chromatin-modifying TSGs in modulating downstream target expression. Based upon my findings, I propose a model where misregulation of chromatin-modifying TSGs upon polarity loss potentiates target enhancers for activation by the polarity-responsive signaling pathways.

Taken together, my thesis work reveals that polarity loss triggers architecture disruption, the activation of stress signaling, dedifferentiation, and mitogenic gene expression. Interestingly, these responses, including PcG-mediated transcriptional derepression, are similarly elicited upon epithelial damage, such as wounding. After a wound response, however, the epithelium is repaired, and these signals are abrogated to maintain homeostasis. In contrast, epithelial integrity is never restored in polarity-deficient tissues and these signals persist, leading to the formation of a neoplastic tumor. These data suggest that Drosophila tumors behave like `wounds that never heal', paralleling previously described models of human tumors.

The defined shapes and sizes of the diverse tissues that comprise multicellular organisms are intimately linked to their specialized functions. The plethora of known cellular behaviors and characteristics that contribute to tissue function and development comes in large part from any given cell's ability to receive and respond to signals from its environment. These signals are often received via transmembrane protein receptors that interpret extracellular signals and transmit them inside the cell to effect cellular changes. Cellular signaling pathways influence cellular shape, proliferative capacity, cell fate, cell movement and many other traits that are crucial for proper cellular and tissue function both during development and over the lifetime of an animal. Therefore, understanding the mechanisms by which cells regulate these signals is a central question in biology.

Drosophila is an ideal model organism in which to study the regulation of cellular signaling pathways. In addition to the ease of genetic manipulation, Drosophila contains epithelial tissues comparable to those in mammals, which have defined apico-basal polarity that is essential for tissue function. Cell polarity signaling pathways in epithelial cells can be studied by the isolation of neoplastic tumor suppressor genes (TSGs) from genetic screens. Drosophila neoplastic TSGs are genes whose functions are absolutely required for proper apico-basal polarity and the regulation of epithelial cell proliferation, though the underlying mechanisms of this regulation and the links between polarity and proliferation remain unknown. In addition to studying how these genes regulate cellular polarity, the identification of certain classes of neoplastic TSGs, such as the endocytic regulators, has given us tools to investigate the regulation of other developmentally important signaling pathways.

A number of cellular signaling pathways are regulated by endocytosis, which is commonly thought to play a role in the attenuation of signaling, but recently appreciated to provide a much more complex method of regulation that may in some contexts promote signaling as well. In general, the roles of both cargo internalization routes and subsequent trafficking down a degradative or recycling pathway in the regulation of cellular signaling pathways are not understood. In Chapter 2 I investigate the role of endocytosis on both trafficking and signaling regulation of the Notch/Delta pathway in Drosophila. I address the roles of specific components of the endocytic internalization machinery in the endocytosis of Notch and Delta, and define whether distinct paths of entry into the cell impact Notch signaling. In Chapter 3, I explore how endocytic trafficking regulates epithelial cell polarity and identify links between the endocytic and junctional scaffold classes of tumor suppressors. In particular, I investigate the possibilities that a common underlying mechanism of tumor suppression exists for these different classes.

Another mechanism through which cells regulate signaling pathways is the ubiquitin-proteasome system. In Drosophila, one well-known example of this is the negative regulation of Hedgehog and Wingless signaling pathways via the ubiquitylation of downstream effectors by the E3 ubiquitin ligase SCFSlmb. In a screen for new neoplastic TSGs, a component of this E3 ubiquitin ligase was recovered along with many endocytic regulators. In Chapter 4, I detail my unexpected discovery that slmb regulates the epithelial cell polarity signaling pathway. I further describe my progress toward finding a presumed polarity-regulating Slmb substrate via a genetic interaction screen.

In sum, my graduate work sheds new light on the ways in which basic cell biological mechanisms, including endocytosis and regulated proteolysis, regulate specific signaling pathways to control cell fate, proliferation, and polarity.

Abstract

Establishing Planar Cell Polarity in an Edgeless Epithelium

By

Katherine Ruth Lipari

Doctor of Philosophy in Molecular and Cell Biology

University of California, Berkeley

Professor David Bilder, Chair

In all animals, organ function is dependent on two major processes that happen in the very beginning of embryonic development: differentiation and morphogenesis. In this study we focus on the process of morphogenesis, studying the planar cell polarity (PCP) pathway, a process that acts perpendicular to the apical-basal axis to orient the cell within the larger environment. From the elongation and closure of the entire central nervous system during neural tube closure, to the subtle orientation of hairs on the skin of a mammal, PCP dictates gross morphological events by first polarizing subcellular structures. Most of our knowledge about PCP derives from the Drosophila wing, where two well-characterized signaling pathways are currently known to regulate PCP: Fz/Vang PCP and Fat/Ds signaling. These two pathways have also been implicated in vertebrate PCP; despite vigorous investigation, there is a surprising paucity of data on how these pathways are shaping vertebrate organs, beyond the fact that they do. One major limitation is that PCP has been most deeply studied in organs with a fundamentally different topology than the tubules that make up the majority of animal organs. While planar polarization of proteins within the cells of the Drosophila wing align with the edge of the wing, causing elongation in the direction of polarization, vertebrate planar polarity proteins align perpendicular to later elongation and do not orient towards the edge of the tissue. Indeed, vertebrate tubular organs do not have similar edges to dictate orientation; they are ‘edgeless’ tissues. To study how planar polarization occurs in such structures we use the Drosophila egg chamber as a model. This tissue undergoes planar polarization perpendicular to the axis of elongation – similar to other tubule elongation – and depends upon the Fat2 cadherin for planar polarization and elongation. In this dissertation I show that planar polarization in this tissue is established early during development of the tissue – at the level of a few cells – by biased microtubule growth. This bias is then propagated by rotation of the entire tissue to create planar polarization of the overlying extracellular matrix to establish a persistent orientation as the tissue grows. I find novel genetic enhancers of the Fat2 mutant phenotype that identify genes involved in this early event and later morphogenetic events. I also study what cells within the epithelium itself are necessary for this rotation to occur. Overall, my thesis work has shown how planar polarization is established and propagated in edgeless tissues through tissue migration.

Elongation of a tissue along one body axis is a conserved morphogenetic process that happens repeatedly during embryo and organ morphogenesis. While the mechanisms of tissue elongation in 2D sheet-like tissues are increasingly well understood, the molecular and cellular mechanisms that elongate topologically distinct tubular and acinar organs may be distinct. In this dissertation, I utilize Drosophila egg chamber (follicle) morphogenesis in the adult female ovary as a model system to understand tissue elongation in the unbounded tube-like organs. The follicle grows spherically during early oogenesis and then elongates during mid-oogenesis along the anterior-posterior axis to become an ellipsoidal organ. Whole-tissue rotation of the follicle during the elongation phase is thought to provide Planar Cell Polarity (PCP) information that orients basal actin bundles and basement membrane Collagen IV fibrils around the circumference, through a process requiring the atypical cadherin Fat2. Despite its essential role in the process, how Fat2 regulates PCP symmetry-breaking to specify rotation in one chirality was unresolved.

With long-term live imaging, I showed that the follicle initiates rotation immediately after it buds off from the germarium. To trace the earliest sign of chiral symmetry-breaking, I developed an in toto 3D image analysis platform to comprehensively chart the cytoskeletal PCP in curved structures. In collaboration with my labmate Katherine Lipari, we found that microtubule growth polarity exhibits a bias within the germarium, prior to follicle budding and rotation. This germarial microtubule bias is dependent on Fat2, and provides chiral information that breaks the PCP symmetry prior to rotation. We propose a PCP relay mechanism that initiates from germarial microtubule chirality, to rotational polarity and actin bundle alignment, to finally a polarized basement membrane organization that elongates the organ.

The limitations of current ex vivo cultures for long-term morphogenetic analyses compelled me to develop a tool that can track cell motility over extended periods. I developed M-TRAIL, which utilizes cell-ECM interactions to retrospectively trace cell movements in fixed tissues. By making clones overexpressing GFP-tagged ECM proteins, cell motility trajectories are revealed based on the long-lasting traces of fluorescently-tagged ECM proteins deposited onto basement membrane. I applied M-TRAIL to compare ex vivo versus in vivo rotation and between WT and fat2 to validate M-TRAIL’s validity. Strikingly, a previously reported protein-truncating fat2 mutant allele, which does not rotate ex vivo yet produces elongated follicles, actually rotates in vivo at half the speed than WT. These results revalidate the model that tissue rotation is required to build a proper basement membrane that elongate the follicle. The M-TRAIL approach also demonstrates how reliance solely on ex vivo imaging can mislead.

By combining our previous work that chart the BM stiffness profile and by measuring BM stiffness in the slow rotating Fat2-truncated follicle, we show that rather than the absolute BM stiffness, an anterior-posteriorly patterned BM stiffness gradient elongate the follicle. By further refining the 3D image analysis platform I developed and use it to chart tissue and cellular scale morphometrics, I found three major cell behaviors that can contribute to follicle elongation, which are oriented cell division, polarized cell shape changes, and cell rearrangement. With analyses of either mutants that fail to elongate or by genetic manipulations that disrupt one particular cell behavior, I found cell shape changes and cell rearrangement play a major role in elongating the follicle. The tools I developed and my findings provide a platform and new insights in understanding the molecular, cellular, and mechanical mechanisms in organ morphogenesis.

The cellular and molecular mechanisms generating the diversity of animal morphologies are still a relatively little explored subject within developmental biology. While changes in cell number help with the growth of tissue, the orientation of cell divisions, migrations, cell rearrangements or changes in cell shape can explain how anisotropies in tissue shape are formed during development. It remains unclear whether the current set of described behaviors can account for all of the diverse morphologies we see in extant metazoan species.

To better understand the mechanisms underlying tissue morphogenesis, my Ph.D. dissertation focused on the development of the ellipsoid egg in Drosophila melanogaster as a new, emerging model system to study tissue elongation. Developing egg chambers (follicles) in Drosophila originate as a sphere and initially grow isotropically but elongate during oogenesis to form an ellipsoid egg. The cellular and molecular mechanisms underlying follicle elongation were not known, although genetic evidence suggested interactions between the follicle epithelium and the surrounding extracellular matrix (ECM) are somehow involved.

To elucidate how tissue elongation occurs in the Drosophila ovary, I utilized live imaging of developing follicles ex vivo and discovered that the follicle unexpectedly undergoes several revolutions of polarized, circumferential global tissue rotation - relative to the surrounding ECM - during the major elongation phase of oogenesis. Follicles with epithelia mutant for an Integrin receptor or Collagen IV, an ECM molecule, fail to rotate and result in round eggs. We found that Collagen IV fibrils become circumferentially planar polarized during the rotation phase but become misoriented in non-rotating `round egg' mutants. Furthermore, acute degradation of Collagen IV rounds previously elongated follicles, suggesting that follicle rotation polarizes a fibrillar matrix that constrains the growing egg in a `molecular corset', generating its ellipsoid shape. Global tissue rotation is thus a novel morphogenetic behavior, using polarized cell motility to propagate planar polarity information in synchrony to the ECM to control tissue shape.

Many new questions developed from the discovery of global tissue rotation. It is unclear how the follicle epithelium responds to a revolving tissue and a mechanically constraining ECM as the tissue grows and elongates. Do conventional behaviors of tissue elongation like polarized cell intercalation, cell elongation or cell division occur in the Drosophila follicle? Chapter 3 aims to further our understanding of the cellular basis of follicle elongation by developing and performing preliminary quantitative morphometric analysis on the follicle epithelium during the elongation phase. In addition, I examine other known round egg mutants to determine whether they too regulate global tissue rotation or whether these genes regulate additional morphogenetic behaviors necessary for egg elongation.

My final chapter investigates how planar cell polarity (PCP) and global tissue rotation are established in the Drosophila follicle. How does molecular planar polarity arise? Does this occur before the onset of polarized follicle rotation? Is there an activation signal to initiate global tissue rotation and from where does this signal originate? I perform an in silico enhancer trap screen to identify developmental signaling pathways and other genes that may coincide with the major elongation phase and has led to the identification of the Notch/Delta signaling pathway as a putative regulator of follicle PCP and global tissue rotation, potentially providing an activation signal from the germline.

The findings from my Ph.D. dissertation provide a new framework to think about the cellular and molecular mechanisms underlying tissue elongation of the developing egg, but moreover, challenge current perspectives on metazoan morphogenesis. Global tissue rotation is a novel polarized morphogenetic behavior required for tissue elongation, but the cellular output of this movement remains ambiguous, perhaps because of the follicle's closed topology as an epithelial chamber. It is the first collective cell migration with an obvious individual cell polarity and tissue polarity, but with no obvious collective cell polarity with leader and follower cells. It also raises the possibility that different mechanisms of PCP establishment and propagation may occur in different tissues. These and other issues indicate that continued studies of Drosophila egg elongation will provide novel and exciting insights to our understanding of tissue polarity, morphogenesis and development in a variety of organisms.