3D and 4D Tumorigenesis Model for the Quantitative Analysis of Cancer Cell Behavior and Screening for Anticancer Drugs.

Cancer cells from cell lines and tumor biopsy tissue undergo aggregation and aggregate coalescence when dispersed in a 3D Matrigel™ matrix. Coalescence is a dynamic process mediated by a subset of cells within the population of cancer cells. In contrast, non-tumorigenic cells from normal cell lines and normal tissues do not aggregate or coalesce, nor do they possess the motile cell types that orchestrate coalescence of cancer cells. Therefore, coalescence is a cancer cell-specific phenotype that may drive tumor growth in vivo, especially in cases of field cancerization. Here, we describe a simple 3D tumorigenesis model that takes advantage of the coalescence capabilities of cancer cells and uses this feature as the basis for a screen for treatments that inhibit tumorigenesis. The screen is especially useful in testing monoclonal antibodies that target cell-cell interactions, cell-matrix interactions, cell adhesion molecules, cell surface receptors, and general cell surface markers. The model can also be used for 2D imaging in a 96-well plate for rapid screening and is adaptable for 3D high-resolution assessment. In the latter case, we show how the 3D model can be optically sectioned with differential interference contrast (DIC) optics, then reconstructed in 4D and quantitatively analyzed by computer-assisted methods, or, alternatively, imaged with confocal microscopy for 4D quantitative analysis of cancer cell interactions with normal cells within the tumor microenvironment. We demonstrate reconstructions and quantitative analyses using the advanced image analysis software J3D-DIAS 4.2, in order to illustrate the types of detailed phenotypic characterizations that have proven useful. Other software packages may be able to perform similar types of analyses.

Directed movement toward, translocation along, penetration into and exit from vascular networks by breast cancer cells in 3D.

We developed a computer-assisted platform using laser scanning confocal microscopy to 3D reconstruct in real-time interactions between metastatic breast cancer cells and human umbilical vein endothelial cells (HUVECs). We demonstrate that MB-231 cancer cells migrate toward HUVEC networks, facilitated by filopodia, migrate along the network surfaces, penetrate into and migrate within the HUVEC networks, exit and continue migrating along network surfaces. The system is highly amenable to 3D reconstruction and computational analyses, and assessments of the effects of potential anti-metastasis monoclonal antibodies and other drugs. We demonstrate that an anti-RHAMM antibody blocks filopodium formation and all of the behaviors that we found take place between MB-231 cells and HUVEC networks.

New monoclonal antibodies that recognize an unglycosylated, conserved, extracellular region of CD44 in vitro and in vivo, and can block tumorigenesis.

CD44 is a transmembrane glycoprotein that binds to hyaluronic acid, plays roles in a number of cellular processes and is expressed in a variety of cell types. It is up-regulated in stem cells and cancer. Anti-CD44 monoclonal antibodies affect cell motility and aggregation, and repress tumorigenesis and metastasis. Here we describe four new anti-CD44 monoclonal antibodies originating from B cells of a mouse injected with a plasmid expressing CD44 isoform 12. The four monoclonal antibodies bind to the terminal, extracellular, conserved domain of CD44 isoforms. Based on differences in western blot patterns of cancer cell lysates, the four anti-CD44 mAbs separated into three distinct categories that include P4G9, P3D2, and P3A7, and P3G4. Spot assay analysis with peptides generated in Escherichia coli support the conclusion that the monoclonal antibodies recognize unglycosylated sequences in the N-terminal conserved region between amino acid 21-220, and analyses with a peptide generated in human embryonic kidney 293 cells, demonstrate that these monoclonal antibodies bind to these peptides only after deglycosylation. Western blots with lysates from three cancer cell lines demonstrate that several CD44 isoforms are unglycosylated in the anti-CD44 target regions. The potential utility of the monoclonal antibodies in blocking tumorigenesis was tested by co-injection of cells of the breast cancer-derived tumorigenic cell line MDA-MB-231 with the anti-CD44 monoclonal antibody P3D2 into the mammary fat pads of mice. All five control mice injected with MDA-MB-231 cells plus anti-IgG formed palpable tumors, while only one of the six test mice injected with MDA-MB-231 cells plus P3D2 formed a tiny tumor, while the remaining five were tumor-free, indicating that the four anti-CD44 mAbs may be useful therapeutically.

Reciprocal signaling and direct physical interactions between fibroblasts and breast cancer cells in a 3D environment.

Tumorigenic cells undergo cell aggregation and aggregate coalescence in a 3D Matrigel environment. Here, we expanded this 3D platform to assess the interactions of normal human dermal fibroblasts (NHDFs) and human primary mammary fibroblasts (HPMFs) with breast cancer-derived, tumorigenic cells (MDA-MB-231). Medium conditioned by MDA-MB-231 cells activates both types of fibroblasts, imbuing them with the capacity to accelerate the rate of aggregation and coalescence of MDA-MB-231 cells more than four fold. Acceleration is achieved 1) by direct physical interactions with MDA-MB-231 cells, in which activated fibroblasts penetrate the MDA-MB-231/Matrigel 3D environment and function as supporting scaffolds for MDA-MB-231 aggregation and coalescence, and 2) through the release of soluble accelerating factors, including matrix metalloproteinase (MMPs) and, in the case of activated NHDFs, SDF-1α/CXCL12. Fibroblast activation includes changes in morphology, motility, and gene expression. Podoplanin (PDPN) and fibroblast activation protein (FAP) are upregulated by more than nine-fold in activated NHDFs while activated HPMFs upregulate FAP, vimentin, desmin, platelet derived growth factor receptor A and S100A4. Overexpression of PDPN, but not FAP, in NHDF cells in the absence of MDA-MB-231-conditioned medium, activates NHDFs. These results reveal that complex reciprocal signaling between fibroblasts and cancer cells, coupled with their physical interactions, occurs in a highly coordinated fashion that orchestrates aggregation and coalescence, behaviors specific to cancer cells in a 3D environment. These in vitro interactions may reflect events involved in early tumorigenesis, particularly in cases of field cancerization, and may represent a new mechanism whereby cancer-associated fibroblasts (CAFs) promote tumor growth.

Integrin α-3 ß-1's central role in breast cancer, melanoma and glioblastoma cell aggregation revealed by antibodies with blocking activity.

Breast cancer, melanoma and glioblastoma cells undergo cell-mediated aggregation and aggregate coalescence in a transparent 3D Matrigel environment. Cells from normal tissue and non-tumorigenic cell lines do not exhibit these behaviors. Here, 266 monoclonal antibodies (mAbs) demonstrated to interact with a wide variety of membrane, secreted and matrix proteins, have been screened for their capacity to block these tumorigenic cell-specific behaviors in a 3D environment. Remarkably, only six of the 266 tested mAbs exhibited blocking activity, four targeting integrin ß-1, one targeting integrin α-3 and one targeting CD44. Colocalization of integrins ß-1 and α-3 in fixed cells and in live aggregates suggests that the integrin α-3 ß-1 dimer plays a central role in cancer cell aggregation in the 3D environment provided by Matrigel. Our results suggest that blocking by anti-integrin and anti-CD44 mAbs involves interference in cell-cell interactions.

Overexpressing TPTE2 (TPIP), a homolog of the human tumor suppressor gene PTEN, rescues the abnormal phenotype of the PTEN-/- mutant.

One possible approach to normalize mutant cells that are metastatic and tumorigenic, is to upregulate a functionally similar homolog of the mutated gene. Here we have explored this hypothesis by generating an overexpressor of TPTE2 (TPIP), a homolog of PTEN, in PTEN-/- mutants, the latter generated by targeted mutagenesis of a human epithelial cell line. Overexpression of TPTE2 normalized phenotypic changes associated with the PTEN mutation. The PTEN-/- -associated changes rescued by overexpressing TPTE2 included 1) accelerated wound healing in the presence or absence of added growth factors (GFs), 2) increased division rates on a 2D substrate in the presence of GFs, 3) adhesion and viability on a 2D substrate in the absence of GFs, 4) viability in a 3D Matrigel model in the absence of GFs and substrate adhesion 5) loss of apoptosis-associated annexin V cell surface binding sites. The results justify further exploration into the possibility that upregulating TPTE2 by a drug may reverse metastatic and tumorigenic phenotypes mediated in part by a mutation in PTEN. This strategy may also be applicable to other tumorigenic mutations in which a homolog to the mutated gene is present and can substitute functionally.

Melanoma cells undergo aggressive coalescence in a 3D Matrigel model that is repressed by anti-CD44.

Using unique computer-assisted 3D reconstruction software, it was previously demonstrated that tumorigenic cell lines derived from breast tumors, when seeded in a 3D Matrigel model, grew as clonal aggregates which, after approximately 100 hours, underwent coalescence mediated by specialized cells, eventually forming a highly structured large spheroid. Non-tumorigenic cells did not undergo coalescence. Because histological sections of melanomas forming in patients suggest that melanoma cells migrate and coalesce to form tumors, we tested whether they also underwent coalescence in a 3D Matrigel model. Melanoma cells exiting fragments of three independent melanomas or from secondary cultures derived from them, and cells from the melanoma line HTB-66, all underwent coalescence mediated by specialized cells in the 3D model. Normal melanocytes did not. However, coalescence of melanoma cells differed from that of breast-derived tumorigenic cell lines in that they 1) coalesced immediately, 2) underwent coalescence as individual cells as well as aggregates, 3) underwent coalescence far faster and 4) ultimately formed long, flat, fenestrated aggregates that were extremely dynamic. A screen of 51 purified monoclonal antibodies (mAbs) targeting cell surface-associated molecules revealed that two mAbs, anti-beta 1 integrin/(CD29) and anti-CD44, blocked melanoma cell coalescence. They also blocked coalescence of tumorigenic cells derived from a breast tumor. These results add weight to the commonality of coalescence as a characteristic of tumorigenic cells, as well as the usefulness of the 3D Matrigel model and software for both investigating the mechanisms regulating tumorigenesis and screening for potential anti-tumorigenesis mAbs.

4D Tumorigenesis Model for Quantitating Coalescence, Directed Cell Motility and Chemotaxis, Identifying Unique Cell Behaviors, and Testing Anticancer Drugs.

A 4D high-resolution computer-assisted reconstruction and motion analysis system has been developed and applied to the long-term (14-30 days) analysis of cancer cells migrating and aggregating within a 3D matrix. 4D tumorigenesis models more closely approximate the tumor microenvironment than 2D substrates and, therefore, are improved tools for elucidating the interactions within the tumor microenvironment that promote growth and metastasis. The model we describe here can be used to analyze the growth of tumor cells, aggregate coalescence, directed cell motility and chemotaxis, matrix degradation, the effects of anticancer drugs, and the behavior of immune and endothelial cells mixed with cancer cells. The information given in this chapter is also intended to acquaint the reader with computer-assisted methods and algorithms that can be used for high-resolution 3D reconstruction and quantitative motion analysis.

Quantitative Motion Analysis in Two and Three Dimensions.

This chapter describes 2D quantitative methods for motion analysis as well as 3D motion analysis and reconstruction methods. Emphasis is placed on the analysis of dynamic cell shape changes that occur through extension and retraction of force generating structures such as pseudopodia and lamellipodia. Quantitative analysis of these structures is an underutilized tool in the field of cell migration. Our intent, therefore, is to present methods that we developed in an effort to elucidate mechanisms of basic cell motility, directed cell motion during chemotaxis, and metastasis. We hope to demonstrate how application of these methods can more clearly define alterations in motility that arise due to specific mutations or disease and hence, suggest mechanisms or pathways involved in normal cell crawling and treatment strategies in the case of disease. In addition, we present a 4D tumorigenesis model for high-resolution analysis of cancer cells from cell lines and human cancer tissue in a 3D matrix. Use of this model led to the discovery of the coalescence of cancer cell aggregates and unique cell behaviors not seen in normal cells or normal tissue. Graphic illustrations to visually display and quantify cell shape are presented along with algorithms and formulae for calculating select 2D and 3D motion analysis parameters.

A computer-assisted 3D model for analyzing the aggregation of tumorigenic cells reveals specialized behaviors and unique cell types that facilitate aggregate coalescence.

We have developed a 4D computer-assisted reconstruction and motion analysis system, J3D-DIAS 4.1, and applied it to the reconstruction and motion analysis of tumorigenic cells in a 3D matrix. The system is unique in that it is fast, high-resolution, acquires optical sections using DIC microscopy (hence there is no associated photoxicity), and is capable of long-term 4D reconstruction. Specifically, a z-series at 5 µm increments can be acquired in less than a minute on tissue samples embedded in a 1.5 mm thick 3D Matrigel matrix. Reconstruction can be repeated at intervals as short as every minute and continued for 30 days or longer. Images are converted to mathematical representations from which quantitative parameters can be derived. Application of this system to cancer cells from established lines and fresh tumor tissue has revealed unique behaviors and cell types not present in non-tumorigenic lines. We report here that cells from tumorigenic lines and tumors undergo rapid coalescence in 3D, mediated by specific cell types that we have named "facilitators" and "probes." A third cell type, the "dervish", is capable of rapid movement through the gel and does not adhere to it. These cell types have never before been described. Our data suggest that tumorigenesis in vitro is a developmental process involving coalescence facilitated by specialized cells that culminates in large hollow spheres with complex architecture. The unique effects of select monoclonal antibodies on these processes demonstrate the usefulness of the model for analyzing the mechanisms of anti-cancer drugs.

Mediated coalescence: a possible mechanism for tumor cellular heterogeneity.

Recently, we demonstrated that tumorigenic cell lines and fresh tumor cells seeded in a 3D Matrigel model, first grow as clonal islands (primary aggregates), then coalesce through the formation and contraction of cellular cables. Non-tumorigenic cell lines and cells from normal tissue form clonal islands, but do not form cables or coalesce. Here we show that as little as 5% tumorigenic cells will actively mediate coalescence between primary aggregates of majority non-tumorigenic or non-cancerous cells, by forming cellular cables between them. We suggest that this newly discovered, specialized characteristic of tumorigenic cells may explain, at least in part, why tumors contain primarily non-tumorigenic cells.

PTEN redundancy: overexpressing lpten, a homolog of Dictyostelium discoideum ptenA, the ortholog of human PTEN, rescues all behavioral defects of the mutant ptenA-.

Mutations in the tumor suppressor gene PTEN are associated with a significant proportion of human cancers. Because the human genome also contains several homologs of PTEN, we considered the hypothesis that if a homolog, functionally redundant with PTEN, can be overexpressed, it may rescue the defects of a PTEN mutant. We have performed an initial test of this hypothesis in the model system Dictyostelium discoideum, which contains an ortholog of human PTEN, ptenA. Deletion of ptenA results in defects in motility, chemotaxis, aggregation and multicellular morphogenesis. D. discoideum also contains lpten, a newly discovered homolog of ptenA. Overexpressing lpten completely rescues all developmental and behavioral defects of the D. discoideum mutant ptenA-. This hypothesis must now be tested in human cells.

Huntingtin regulates Ca(2+) chemotaxis and K(+)-facilitated cAMP chemotaxis, in conjunction with the monovalent cation/H(+) exchanger Nhe1, in a model developmental system: insights into its possible role in Huntington׳s disease.

Huntington׳s disease is a neurodegenerative disorder, attributable to an expanded trinucleotide repeat in the coding region of the human HTT gene, which encodes the protein huntingtin. These mutations lead to huntingtin fragment inclusions in the striatum of the brain. However, the exact function of normal huntingtin and the defect causing the disease remain obscure. Because there are indications that huntingtin plays a role in Ca(2+) homeostasis, we studied the deletion mutant of the HTT ortholog in the model developmental system Dictyostelium discoideum, in which Ca(2+) plays a role in receptor-regulated behavior related to the aggregation process that leads to multicellular morphogenesis. The D. discoideum htt(-)-mutant failed to undergo both K(+)-facilitated chemotaxis in spatial gradients of the major chemoattractant cAMP, and chemotaxis up a spatial gradient of Ca(2+), but behaved normally in Ca(2+)-facilitated cAMP chemotaxis and Ca(2+)-dependent flow-directed motility. This was the same phenotypic profile of the null mutant of Nhel, a monovalent cation/H(+)exchanger. The htt(-)-mutant also failed to orient correctly during natural aggregation, as was the case for the Nhel mutant. Moreover, in a K(+)-based buffer the normal localization of actin was similarly defective in both htt(-) and nhe1(-) cells in a K(+)-based buffer, and the normal localization of Nhe1 was disrupted in the htt(-) mutant. These observations demonstrate that Htt and Nhel play roles in the same specific cation-facilitated behaviors and that Nhel localization is directly or indirectly regulated by Htt. Similar cation-dependent behaviors and a similar relationship between Htt and Nhe1 have not been reported for mammalian neurons and deserves investigation, especially as it may relate to Huntington׳s disease.

Interferon regulatory factor 6 regulates keratinocyte migration.

Interferon regulatory factor 6 (Irf6) regulates keratinocyte proliferation and differentiation. In this study, we tested the hypothesis that Irf6 regulates cellular migration and adhesion. Irf6-deficient embryos at 10.5 days post-conception failed to close their wound compared with wild-type embryos. In vitro, Irf6-deficient murine embryonic keratinocytes were delayed in closing a scratch wound. Live imaging of the scratch showed deficient directional migration and reduced speed in cells lacking Irf6. To understand the underlying molecular mechanisms, cell-cell and cell-matrix adhesions were investigated. We show that wild-type and Irf6-deficient keratinocytes adhere similarly to all matrices after 60 min. However, Irf6-deficient keratinocytes were consistently larger and more spread, a phenotype that persisted during the scratch-healing process. Interestingly, Irf6-deficient keratinocytes exhibited an increased network of stress fibers and active RhoA compared with that observed in wild-type keratinocytes. Blocking ROCK, a downstream effector of RhoA, rescued the delay in closing scratch wounds. The expression of Arhgap29, a Rho GTPase-activating protein, was reduced in Irf6-deficient keratinocytes. Taken together, these data suggest that Irf6 functions through the RhoA pathway to regulate cellular migration.

Myosin heavy chain kinases play essential roles in Ca2+, but not cAMP, chemotaxis and the natural aggregation of Dictyostelium discoideum.

Behavioral analyses of the deletion mutants of the four known myosin II heavy chain (Mhc) kinases of Dictyostelium discoideum revealed that all play a minor role in the efficiency of basic cell motility, but none play a role in chemotaxis in a spatial gradient of cAMP generated in vitro. However, the two kinases MhckA and MhckC were essential for chemotaxis in a spatial gradient of Ca(2+), shear-induced directed movement, and reorientation in the front of waves of cAMP during natural aggregation. The phenotypes of the mutants mhckA(-) and mhckC(-) were highly similar to that of the Ca(2+) channel/receptor mutant iplA(-) and the myosin II phosphorylation mutant 3XALA, which produces constitutively unphosphorylated myosin II. These results demonstrate that IplA, MhckA and MhckC play a selective role in chemotaxis in a spatial gradient of Ca(2+), but not cAMP, and suggest that Ca(2+) chemotaxis plays a role in the orientation of cells in the front of cAMP waves during natural aggregation.

The IplA Ca2+ channel of Dictyostelium discoideum is necessary for chemotaxis mediated through Ca2+, but not through cAMP, and has a fundamental role in natural aggregation.

During aggregation of Dictyostelium discoideum, nondissipating, symmetrical, outwardly moving waves of cAMP direct cells towards aggregation centers. It has been assumed that the spatial and temporal characteristics of the front and back of each cAMP wave regulate both chemokinesis and chemotaxis. However, during the period preceding aggregation, cells acquire not only the capacity to chemotax in a spatial gradient of cAMP, but also in a spatial gradient of Ca(2+). The null mutant of the putative IplA Ca(2+) channel gene, iplA(-), undergoes normal chemotaxis in spatial gradients of cAMP and normal chemokinetic responses to increasing temporal gradients of cAMP, both generated in vitro. However, iplA(-) cells lose the capacity to undergo chemotaxis in response to a spatial gradient of Ca(2+), suggesting that IplA is either the Ca(2+) chemotaxis receptor or an essential component of the Ca(2+) chemotaxis regulatory pathway. In response to natural chemotactic waves generated by wild-type cells, the chemokinetic response of iplA(-) cells to the temporal dynamics of the cAMP wave is intact, but the capacity to reorient in the direction of the aggregation center at the onset of each wave is lost. These results suggest that transient Ca(2+) gradients formed between cells at the onset of each natural cAMP wave augment reorientation towards the aggregation center. If this hypothesis proves correct, it will provide a more complex contextual framework for interpreting D. discoideum chemotaxis.

Nhe1 is essential for potassium but not calcium facilitation of cell motility and the monovalent cation requirement for chemotactic orientation in Dictyostelium discoideum.

In Dictyostelium discoideum, extracellular K+ or Ca2+ at a concentration of 40 or 20 mM, respectively, facilitates motility in the absence or presence of a spatial gradient of chemoattractant. Facilitation results in maximum velocity, cellular elongation, persistent translocation, suppression of lateral pseudopod formation, and myosin II localization in the posterior cortex. A lower threshold concentration of 15 mM K+ or Na or 5 mM Ca2+ is required for chemotactic orientation. Although the common buffer solutions used by D. discoideum researchers to study chemotaxis contain sufficient concentrations of cations for chemotactic orientation, the majority contain insufficient levels to facilitate motility. Here it has been demonstrated that Nhe1, a plasma membrane protein, is required for K+ but not Ca2+ facilitation of cell motility and for the lower K+ but not Ca2+ requirement for chemotactic orientation.

Ca2+ chemotaxis in Dictyostelium discoideum.

Using a newly developed microfluidic chamber, we have demonstrated in vitro that Ca(2+) functions as a chemoattractant of aggregation-competent Dictyostelium discoideum amoebae, that parallel spatial gradients of cAMP and Ca(2+) are more effective than either alone, and that cAMP functions as a stronger chemoattractant than Ca(2+). Effective Ca(2+) gradients are extremely steep compared with effective cAMP gradients. This presents a paradox because there is no indication to date that steep Ca(2+) gradients are generated in aggregation territories. However, given that Ca(2+) chemotaxis is co-acquired with cAMP chemotaxis during development, we speculate on the role that Ca(2+) chemotaxis might have and the possibility that steep, transient Ca(2+) gradients are generated during natural aggregation in the interstitial regions between cells.

How a cell crawls and the role of cortical myosin II.

The movements of Dictyostelium discoideum amoebae translocating on a glass surface in the absence of chemoattractant have been reconstructed at 5-second intervals and motion analyzed by employing 3D-DIAS software. A morphometric analysis of pseudopods, the main cell body, and the uropod provides a comprehensive description of the basic motile behavior of a cell in four dimensions (4D), resulting in a list of 18 characteristics. A similar analysis of the myosin II phosphorylation mutant 3XASP reveals a role for the cortical localization of myosin II in the suppression of lateral pseudopods, formation of the uropod, cytoplasmic distribution of cytoplasm in the main cell body, and efficient motility. The results of the morphometric analysis suggest that pseudopods, the main cell body, and the uropod represent three motility compartments that are coordinated for efficient translocation. It provides a contextual framework for interpreting the effects of mutations, inhibitors, and chemoattractants on the basic motile behavior of D. discoideum. The generality of the characteristics of the basic motile behavior of D. discoideum must now be tested by similar 4D analyses of the motility of amoeboid cells of higher eukaryotic cells, in particular human polymorphonuclear leukocytes.

The effects of extracellular calcium on motility, pseudopod and uropod formation, chemotaxis, and the cortical localization of myosin II in Dictyostelium discoideum.

Extracellular Ca(++), a ubiquitous cation in the soluble environment of cells both free living and within the human body, regulates most aspects of amoeboid cell motility, including shape, uropod formation, pseudopod formation, velocity, and turning in Dictyostelium discoideum. Hence it affects the efficiency of both basic motile behavior and chemotaxis. Extracellular Ca(++) is optimal at 10 mM. A gradient of the chemoattractant cAMP generated in the absence of added Ca(++) only affects turning, but in combination with extracellular Ca(++), enhances the effects of extracellular Ca(++). Potassium, at 40 mM, can partially substitute for Ca(++). Mg(++), Mn(++), Zn(++), and Na(+) cannot. Extracellular Ca(++), or K(+), also induce the cortical localization of myosin II in a polar fashion. The effects of Ca(++), K(+) or a cAMP gradient do not appear to be similarly mediated by an increase in the general pool of free cytosolic Ca(++). These results suggest a model, in which each agent functioning through different signaling systems, converge to affect the cortical localization of myosin II, which in turn effects the behavioral changes leading to efficient cell motility and chemotaxis. Cell Motil. Cytoskeleton 2009. (c) 2009 Wiley-Liss, Inc.