Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Tumor Cell Budding, Motility, and Invasion as Driving Forces for Liver Metastases

Tumor Cell Budding as a Mode of Tumor Cell Individualization

Colorectal Carcinoma (CRC) cells exert distinct growth patterns that affect disease progression and prognosis, in particular tumor cell budding at the invasive margin. Budding (or sprouting) is defined as small clusters of individualized and poorly differentiated carcinoma cells in or outside the invasion front. The presence of such dissociated, dedifferentiated tumor cells at the tumor-host interface is predictive of lymphatic involvement of CRC (Ogawa et al. 2009), lymphovascular invasion (also in early stage N0 CRC; Wang et al. 2009), predicts local and distant metastases (Nakamura et al. 2005; Zlobec et al. 2008; Suzuki et al. 2009), including lymph node metastasis (Yamauchi et al. 2008), and is correlated with an enhanced malignant potential in CRC (Okuyama et al. 2002; Prall 2007; Kanazawa et al. 2008). These budding cells have undergone epithelial-mesenchymal transition (EMT) and are growth-arrested due to increased expression of p16 and other mechanisms (Jung et al. 2001). Low cellular proliferation at the invasion front, the niche of budding cells, has been associated with poor prognosis in CRC (Palmqvist et al. 1999). Loss of expression of the adhesion molecule, E-cadherin, and APAF-1 were independent predictors of budding, and the loss was associated with more frequent peritumoral lymphocytes and tumor-infiltrating lymphocytes (Zlobec et al. 2007).

Motility, Migration, and Invasion of Tumor Cells: The Role of Invadopodia

The dissociation process of budding cells at the invasion front requires distinct modes of motility/cell locomotion linked to the tumor cells’ cytoskeleton. Individualized cancer cells are capable to migrate through the extracellular matrix (ECM) and can form ECM-degrading, actin-rich ventral membrane protrusions called invadopodia, similar to the podosomes of macrophages. These dynamic organelle-like structures adhere to, and proteolytically digest, collagens, laminin, and fibronectin located in the ECM (Yamaguchi et al. 2006; Stylli et al. 2008).

The formation of invadopodia requires WAVE2 expression but seems to be independent of the tubulin system (Kikuchi and Takahashi 2008). In the regulation pathways programming the conversion between rounded and elongated modes of cancer cell movement, DOCK3, NEDD9, WAVE2, and ARHGAP22 are key factors regulating Rac and Rho signaling (Croft and Olson 2008). The formation of diverse types of lamellipode with actin reassembly at the leading edge requires the membrane transport of WAVE2. The membrane transport of WAVE2 requires the association between the constitutive complex of betaPIX and GIT1 with WAVE2 (Morimura et al. 2009) and a small GTPase, Rac1, the motor protein kinesin, and microtubules (the latter not of importance in invadopodia). The p21-activated kinase Pak1 is a downstream effector of Rac1 and binds to WAVE2 and is transported with WAVE2 to the leading edge by stimulation with hepatocyte growth factor. Concomitantly, phosphorylation of tubulin-bound stathmin/Op18 at serine 25 and serine 38, microtubule growth, and stathmin/Op18 binding to the kinesin-WAVE2 complex are induced (Takahashi and Suzuki 2009). The human ortholog of mammalian enabled (hMena), an actin regulatory phosphoprotein involved in the control of cell motility, is expressed at the invasion margin of CRC and here particularly in budding cells (Toyoda et al. 2009).

Immunohistochemical staining for the cytoskeletal component, beta(III)-tubulin, was evident in budding CRC cells, suggesting that changes in tubulin isotypes could modulate the invading activity of these cells (Portyanko et al. 2009). 14-3-3 sigma is frequently expressed in budding tumor cells in the invasive area, and overexpression of the protein modulates cell migration in the presence of the ECM protein, tenascin-C (Ide et al. 2007). As motile dissociated cancer cells (including budding/sprouting cells) with a decreased level of differentiation play a crucial role in carcinoma invasion, the actin remodeling system critically participating in the locomotion machinery is studied in cancer, including CRC. At the leading edge of migrating cells (lamellipodia and invadopodia), protrusive forces are developed by the assembly of F-actin filaments organized in dendritic array at the front and a more distal lamellar linear array. Production and remodeling of F-actin depends on several proteins acting in a highly orchestrated fashion. Tropomyosin markedly inhibits lamellipodium formation and facilitates the distal lamellar linear array, in that it competes with the actin nucleator, actin-related protein 2/3 (Arp2/3) during filament branching, and counteracts the effects of the neural Wiskott-Aldrich syndrome protein, N-WASP (Bugyi et al. 2010). Cortactin is an F-actin-binding protein that stabilizes F-actin networks and promotes actin polymerization by activating the Arp 2/3 complex. Overexpression of cortactin stimulates cell migration in cancer cells, and in CRC cells, cortactin interacts with the zona occludens-1(ZO-1) protein, and expression of this complex favors an invasive and spreading/metastasizing phenotype (Hirakawa et al. 2009). Actin-related protein 2 (Arp2) colocalizes with WAVE2 in CRC cells and this is correlated with liver metastasis (Iwaya et al. 2007). Two CARMIL (capping protein, Arp2/3, and myosin-I linker) family proteins exert important regulatory functions. CARMIL1 localizes to lamellipodia and macropinosomes, and loss of its function causes loss of lamellipodial actin, along with defects in protrusion, ruffling, and macropinocytosis. CARMIL2 colocalizes with vimentin, and loss of its function causes a distinct multipolar phenotype, associated with decreased levels of myosin-IIB (Liang et al. 2009). An important group of regulating factors are members of the TOCA family of F-BAR-containing proteins. These proteins bind to and remodel lipid bilayers in a clathrin-dependent manner via their conserved F-BAR domains and regulate actin dynamics via their N-WASP binding SH3 domains. In mammalian cells, TOCA1 is associated with N-WASP and WPS-1 directly or WAVE2 indirectly via ABI-1 (Giuliani et al. 2009). Generally, BAR (Bin-Amphiphysin-Rvs) domain proteins can sense membrane curvature and recruit actin to membranes. They are also critically involved in clathrin-mediated endocytosis, in which actin polymerization mediated by the Arp2/3 complex is essential for membrane tubulation and vesicle formation and fission. The tubulated membrane between the future endocytic vesicle and the plasma membrane seems to form an arc upon scission of the endocytic vesicle, and the formation of this structure depends on arc sensing mediated by the RFC/F-BAR domain protein, FBP17 (Suetsugu 2009). In endocytosis, BAR proteins also interact with other components of the endocytic and cytoskeletal machinery, including the GTPase dynamin (which mediates vesicle fission), N-WASP (an Arp2/3 complex regulator), and synaptojanin (a phosphoinositide phosphatase). Three classes of BAR proteins have been defined, i.e., BAR, N-BAR, and F-Bar (Dawson et al. 2006).

Stromal Factors Exerting an Influence on Tumor Cell Migration and Invasion

CXCL12 (stromal cell-derived factor 1) is an important factor for the stimulation of cancer cell motility and migration. CXCL12 promotes motility in neoplastic cells that express the receptor, CXCR4 (Do Carmo et al. 2010). CXCL12+ tumor budding in CRC is a significant prognostic factor (Akishima-Fukasawa et al. 2009). CXCR4 is also known to enhance the proliferation of certain cancer cells via AKT and ERK-dependent pathways (Shen et al. 2010). The liver has high concentrations of CXCL12 (Kim et al. 2005) that could provide a specific homing target and promoting factor for CXCR4-expressing CRC cells. CXCL12 is a constitutively expressed chemokine which orchestrates a large array of functions. Six different isoforms exist in humans (α, β, γ, δ, ε, and ϕ), all differing in their C-terminal domain (Yu et al. 2006). CXCL12 binds to the receptor, CXCR4, which is transcriptionally regulated by HIF-1alpha and a constitutively activated EGFR variant, EGFRvIII (Rahimi et al. 2010). Angiogenesis is promoted by attraction of motile endothelial cells by CXCL12. The axis markedly affects the migration of tumor cells to metastatic sites, allowing cancer cells to access cellular niches expressing the CXCR4 receptor favoring cell survival and growth (“metastatic homing” of cancer cells; Murphy 2001; Burger and Kipps 2006; Dewan et al. 2006). In CRC, CXCR4 expression was found to increase the risk for recurrence, liver metastasis, and poor survival (Kim et al. 2005). The role of the microvascular bed in mechanisms of tumor cell spread surfaced in the observation that concomitant overexpression of CXCR4 and vascular endothelial growth factor (VEGF) in the tumors predicts early distant relapse in stage II-III CRC patients (Ottaiano et al. 2006).

Pathogenic Features of Liver Metastases: Dissemination and Evolution of Premalignant Cells (Circulating Tumor Cells, CTCs) and the “Wandering Field”/“Cancer Field” Hypotheses

In the previous paragraph, it was outlined that the settling of tumor cells and their progenitors at future metastatic sites requires the establishment of a premetastatic niche that permits circulating cells to home to tissues and to interact with the host organ (Sleeman 2012). On the other hand, the early metastatic process also requires that subpopulations of cells are released that are capable to home to a niche and to produce a metastatic growth. The capacity of tumor cells to escape from the primary tumor, individualize, migrate, invade, and spread to remote tissues was described as a phenomenon resembling Darwinian selection (Klein 2013). However, this may represent a simplistic view, as the so-called invasion and metastasis cascade does not seem to exactly work that way. Rather, there is recent evidence that tumor cells and particularly their progenitors (cancer stem cells) individualize and disseminate in very early and primordial malignant neoplasms already, having a still not well-studied status of “premalignant” cells (PMCs), single disseminated tumor cells (DTCs), or circulating tumor cells (CTCs) (Klein 2000; Diamond et al. 2012) that will evolve almost or completely independent from the primary tumor (Ansieau et al. 2008). These individualized cells have a fate of their own and possess a high degree of plasticity (Bednarz-Knoll et al. 2012). It is these PMCs that will acquire, by progressive genomic instability and selection pressures, distinct genomic signatures, and these alterations will render PMCs capable to induce, and interact with, future premetastatic niches. In this setting, spread of cancer may be described by the features of a “wandering field hypothesis,” meaning that PMCs and their induced premetastatic niches form a cancer field that is distributed within the host. The generation and distribution of tumor cells that can finally give rise to metastases in “field-like” body structure has been addressed by methods of systems biology, discussing selection-driven independence, invasion, and so-called swarm intelligence (Tarabichi et al. 2013).

Pathogenic Features of Liver Metastases: Parallel and Nonparallel Progression of Tumor and Metastases

Introduction

An older view of metastatic spread identified the primary tumor as the single important engine determining metastasis through the production and release of fully metastasis-capable cells. A second hypothesis, termed the parallel progression model (Klein 2009), proposes parallel, independent progression of metastases arising from early disseminated cells (single disseminated tumor cells (DTCs); see above). Clonal expansion mechanisms and a marked genetic disparity between primary tumors and disseminated tumor cells (Stoecklein and Klein 2010) support this hypothesis. Other findings were not compatible with the parallel progression model, but the linear progression model in this hypothesis works when it is assumed that tumors harbor two types of cells, i.e., metastasis-forming cells (MFCs), likely to derive from tumor stem cells, and a second set that is not capable to produce metastases. There is evidence that the proportion of MFCs decreases as a function of increasing tumor size, one possible explanation for the observation that at least part of large tumors have a lesser probability to produce metastases (review: Koscielny and Tubiana 2010).

Do Primary Tumors and Their Metastases Share a Parallel Evolution?

Animal experimentation had already shown that removal of a primary malignancy can be followed by rapid outgrowth of metastases. This observation illustrates that primary tumors might exert an effect on growth of their metastases and that primary tumor and metastases may not have parallel growth patterns. In human CRCs, resection of the primary tumor was associated with a modest increase in cancer cell proliferation of synchronous hepatic metastases and a significant decrease in metastatic tumor cell apoptosis. These findings suggest that primary tumors can inhibit growth of their metastases (Peeters et al. 2006).

Pathogenic Features of Liver Metastases: Tumor Self-seeding and Cross-seeding in Metastasis

Introduction

According to classical concepts, metastasis has been considered to be a unidirectional process, whereby cancer cells are released by, or actively emigrate from, the primary tumor and seed cells into lymph nodes and to remote organs. This pathogenetic view, oriented on a mode “as the river flows,” has not been questioned for a long time. There is now evidence that metastasis is a multidirectional process, whereby cancer cells can not only seed to distant sites but also back to the primary tumor. The process of recolonizing of the primary tumor and its premetastatic niche is termed, tumor self-seeding. This hypothesis is based on the presence of tumor cell-supporting noncancerous cells not only in remote tissues but also in the primary tumor. The self-seeding model may answer many open questions regarding the complexity of the metastatic process (Norton and Massagué 2006; Norton 2008; Kim et al. 2009; reviews: Leung and Brugge 2009; Aguirre-Ghiso 2010; Hahnfeldt 2010; Norton 2011; Comen et al. 2011; Sleeman et al. 2011; Comen 2012; Comen and Norton 2012). In mice tumor models, it was shown that the self-seeding process can select more aggressive cells involved in cancer progression (Kim et al. 2009). Circulating tumor cells were found to enhance primary tumor growth, probably through stromal paracrine effects (Hahnfeldt 2010).

Homing Mechanisms in Tumor Self-seeding

How are tumor cells attracted to reinfiltrate primary tumors? Circulating tumors studied in mice secrete factors (“seed-derived factors”) that favor stromal recruitment and angiogenesis also in the primary lesion, including the chemokine CXCL1 (Kim et al. 2009). It has been shown that the protein chromogranin A inhibits the shedding of cancer cells into circulation from primary tumors, as well as the reinfiltration of primary tumors, via decreased vascular leakage in tumors and inhibition of transendothelial migration of cancer cells (Dondossola et al. 2012).

Pathogenic Features in Liver Metastases: Cell Clone Ecology Hypothesis

In the cell clone ecology hypothesis, it is assumed that the survival and fate of cell clones that behave like “parasites” in the organism and result in cancer are governed by the principles of ecology. The hypothesis is based on the concept that mutated and aneuploid cell clones undergoing progressive genomic instability are prone to degradation rather than becoming “successfully progressing clones,” with an inherent risk to become extinct or enfeebled to become a benign condition. In order to establish a “fit” clone, the cells of a so-called transient clone must undergo cell-cell fusion (the cell fusion model of cancer progression and metastasis). Cell-cell fusion provides, according to this hypothesis, a sex-like mode of reproduction essential for an ecologically fit parasite organism, provides the opportunity required by tumors to rejuvenate cell lines containing abnormal and instable genomes, and serves to overcome erosion of telomeres during clone expansion (Parris 2006a, b).

Pathogenic Features in Liver Metastases: Cancer Cell-Normal Cell Fusion Theory of Metastasis

Introduction

Invasion and spread of cancer cells require an intricate interaction between the tumor cells and normal cells of their invaded tissue environment, involving complex patterns or mutual cross talk between tumor cells, stromal cells, and leukocytes (the so-called seed and soil theory; review : Mareel and Madani 2006). In this complex interplay, cells involved create close contacts through cell-to-cell adhesion and enter a complex interactomics (van der Bij et al. 2005). This may potentially lead to the secretion of factors involved in tumor-stromal interactions (e.g. the matricellular protein SPARC; Podhajcer et al. 2008) and mutual signaling via exchange of signaling substances between normal cells and tumor cells. For example, secreted activated leukocyte adhesion molecule (sALCAM) attenuates melanoma cell invasion (van Kilsdonk et al. 2008). Leukocytes infiltrating tumors facilitate dissemination in that they disrupt intercellular junctions and cell surface adhesion molecules (Man et al. 2011). Leukocytes in contact with tumor cells can create a premetastatic niche (Oppenheimer 2006), but they may also act as myeloid-derived suppressor cells (review: Nardin and Abastado 2008). Furthermore, information may be exchanged between stem cells and cancer cells via horizontal genomic transfer (large-scale transfer of DNA and chromatin; Glinsky 2005). This has been suggested for cancer cells expressing antigen of tumor-associated macrophages (TAMs) (Shabo and Svanvik 2011). However, there is evidence of a more intimate interaction between normal cells and cancer cells (Lagarde 1986). It has been found that macrophages and other bone marrow-derived cells can fuse with cancer cells forming heterologous chimeras (Vignery 2005; Vignery 2008). Cell-to-cell fusion results in a complex sharing of several cellular features by the chimeras and the eventual emergence of novel types of behavior, including a metastatic phenotype (Lagarde and Kerbel 1985; Parris 2008). Somatic cell fusion is a potent source of genetic rearrangement leading to potentially metastatic cell variants (Larizza and Schirrmacher 1984; Carter 2008). In vitro fusion of a nonmetastatic tumor cell type with a bone marrow-derived macrophage results in the tumor cell’s high metastatic capacity (Larizza et al. 1984a). In animal tumor xenograft models, this phenomenon was associated with tumor metastasis (review : Pawelek and Chakraborty 2008).

Macrophage-Tumor Cell Chimeras

It is known that diverse types of malignant neoplastic cells acquire distinct sets of macrophage properties, including morphological appearance, surface adhesion, phagocytosis, motile behavior, and membrane lipid composition (Huysentruyt et al. 2008). This phenomenon has led to the concept of “cancer as a macrophage-mediated autoaggressive disease” (Muntarova and Kovarik 1987). It has been observed that such features may be acquired through heterologous cell-to-cell hybridization (Kerbel et al. 1983; Lagarde and Kerbel 1985). Tumor cell fusion with leukocytes is a source of myeloid traits in cancer, including the acquisition of typical motile behaviors (Rachkovsky et al. 1998; Pawelek 2005). Also the distinct features of tumor-associated macrophages (TAMs) may be acquired by tumor cells via cell fusion (review: Pawelek et al. 2006). Fusion of melanoma cells with macrophages results in acquired melanocyte stimulating hormone-inducible chemotaxis in the tumor cells (Rachkovsky and Pawelek 1999). Macrophage-melanoma cell fusion-induced hybrids upregulate N-acetylglucosaminyltransferase V (beta 1,6-branching; Gnt-V) (Chakraborty et al. 2001), acquire specific motility patterns, are highly metastatic in vivo, and express macrophage-associated traits of increased Gnt-V (Chakraborty and Pawelek 2003; Chakraborty and Pawelek 2007). Gnt-V forms beta 1,6 branching on the trimannosyl terminus of N-glycans, causing the production of beta 1,6 Glc-NAc-bearing oligosaccharide, important for the adhesion and motile behavior of both normal granulocytes and tumor cells. In macrophage-melanoma cell hybrids, these oligosaccharides colocalize with coarse melanin granules (Rupani et al. 2004). It has been proposed that mechanisms allowing escape from the extracellular matrix are important for migration and invasion, and such escape strategies may be acquired through cell fusion (review: Parris 2006b). Macrophages are capable to fuse with epithelial cells of murine intestinal tumors, resulting in nuclear reprogramming with the expression of unique subsets of transcripts (Powell et al. 2011). There is suggestive evidence that the highly metastatic variant ESb of the T cell lymphoma Eb is derived from spontaneous fusion with a host macrophage (Larizza et al., 1984b). Leukocyte-cancer cell fusion might act as the initiator for the use of aerobic glycolysis as a metabolic energy source of cancer cells, the Warburg effect (Lazova et al. 2011).

Tumor Cell-Normal Cell Chimeras

Native or damaged tumor cells can fuse with nonneoplastic host cells, e.g., following irradiation (Wiener et al. 1974a; Ber et al. 1978). Heterospecifc chimeras are a mechanism for horizontal transmission of malignancy (Goldenberg 2012). Normal diploid cells fused to malignant neoplastic cells can suppress the malignant phenotype in the chimeras (Harris et al. 1969; Wiener et al. 1974b). On the other hand, murine melanoma–macrophage hybrids showed enhanced metastatic behavior (Rachkovsky et al. 1998). A fusion-induced alteration in the cell surface structure, involving N-glycosylation of membrane proteins mediated by N-acetylglucosaminyltransferase V, is involved in this mechanism (Sodi et al. 1998; Chakraborty et al. 2001).

Tumor Cell-Tumor Cell Chimeras

There is strong evidence that tumor cells can form hybrids/chimeras among each other by cell fusion. The capability to fuse is not restricted by the type of cell origin or cell lineages involved, in that epithelial cells can form somatic hybrids with mesenchymal cells or germ cell tumor cells (Peterson and Weiss 1972; Litwack and Croce 1979; Fortuna et al. 1989). Fusion of tumor cells, also in vivo, can alter the biological behavior of the cells and effect invasion and spread. Spontaneous fusion of two mouse tumor cell subpopulations resulted in a more aggressive tumor cell variant (Miller et al. 1989), in part with enhanced metastatic potential (Mi et al. 2012), but loss of malignancy of hybrids derived from two murine malignant cell lines has also been observed (Jami and Ritz 1975). A predominance of a metastatic phenotype has been observed in hybrids formed by fusion of mouse and human melanoma cells (van Golen et al. 1996). Certain oncogenic features expressed in tumor cell chimeras can behave in a switchable dominant manner (Zimmermann et al. 1981).

Fusion of Intact Cancer Cells with Cancer Cell Components

Cells can not only fuse with intact partner cells but also with parts of the partner cell body, e.g., organelles and plasma membrane vesicles. Fusion of plasma membrane vesicles from highly metastatic cancer cells affect arrest and metastasis of blood-borne tumor cells (Poste and Nicolson 1980). Fusion of plasma membrane vesicles from highly metastatic cancer cells with blood-borne tumor cells can modify the arrest and mestatatic behavior of the recipient cells (Poste and Nicolson 1980).

Stem Cell-Tumor Cell Chimeras

As a mechanism of pathogenesis of cancer it has long been hypothesized that cancer cells might arise from a hybrid cell which derives from normal somatic cell fusion but obtains a proliferative advantage (Qian 1993). On the other hand, it has been suggested that distinct features of stem cells can be transferred to tumor cells through cell fusion. According to this concept, cancer can result from a fusion between an “altered” premalignant cell and a bone marrow-derived stem cell, aneuploidy as a typical feature of tumor cells being the direct result of this fusion process. This model might explain striking similarities between stem cells and cancer cells (He et al. 2005).

Organelle Chimeras of Normal Cells

Through cell fusion, cells can uptake organelles from partner cells. The resulting organelle chimera is heteroplasmic (heteroplasmy). On a genomic level, heteroplasmy denotes a situation characterized by the presence of a mixture of more than one type of organelle genomes, mainly mitochondrial DNA and plastid DNA, within a cell or individual. However, the term heteroplasmy is most often employed to describe a distinct state of the mitochondrial genome. Mitochondrial heteroplasmy results from mutations in the mitochondrial genome, whereby microheteroplasmy with numerous independent mutations is normally found in organisms. There is evidence that the most dramatic effects of heteroplasmy on the fate of the chimeric cells are obtained in case of mitochondrial chimeras, which contain two or more sets of mitochondrial DNA (Hua et al. 2012). Very low-level heteroplasmy mitochondrial DNA variations are an inherited trait in humans (Guo et al. 2013). Normal mammalian cells can even be propagated after integration of organelles as foreign as those of plants, e.g., chloroplasts (Nass 1969).

Heterosis

Heterosis or hybrid vigor (synonym: outbreeding enhancement) denotes the phenomenon that hybrids of two genetically pure cell lines grow more rapidly or give rise to more cells than the parental pure lines. The phenomenon of heterosis results in increased yield of cellular offspring and therefore increased biomass (Shull 1948; McDaniel and Sarkissian 1966; Baranwal et al. 2012; Groszmann et al. 2013). Effects of gene dosage imbalances play a significant role in heterosis. The underlying mechanisms seem to involve cumulative positive effects in the differential expression of sets of genes on yield-producing metabolic pathways (reviews: Kamb 2003; Baranwal et al. 2012). For a heterotic phenotype to occur, genetic and genomic studies have shown that gene expression in hybrids is regulated by the interactions of the two parental epigenetic systems and their genomes (review: Groszmann et al. 2013).

Cell Chimeras and Nuclear Alterations

Fusion of two contiguous tumor cells will increase the chromosome content of the resultant chimera, and this increase of ploidy could facilitate and heighten the inherent genomic instability of the lineage. This phenomenon has been observed as spontaneous fusion between metastatic mammary tumor subpopulations (Miller et al. 1988; review: Hart 1983). Murine fusion-derived hepatocytes showed ploidy reductions (Duncan et al. 2009).

Cancer Cell Chimeras and Cell Differentiation

Progression of cancer, including the establishment of a metastatic phenotype, depends on the proliferative status and the differentiation of the neoplastic cells. It is known that hybrid formation between cells affects differentiation of the fusion partners. In hybridization experiments using cells of defined and different histotypes, it turned out that both parental determinations are retained by hybrid cells and that parental differentiation modes are repressed only in a mutually exclusive fashion (Fougère and Weiss 1978). On the other hand, fusion of interspecific cells (rat and mouse) demonstrated that one specific function of one of the fusion partners can be expressed in the hybrid cells (Peterson and Weiss 1972). In hybrids between murine melanoma and teratocarcinoma cells, the formation of “supermelanotic chimeras” depends on the activation of genes involved in differentiation (Watanabe 1984).

Fusion of Human Cells with Cells of Nonhuman Organisms

It has been theorized that normal cells might acquire “dangerous” genetic information by integration of genetic information other than that of viruses. One hypothesis suggests that tumorigenic cells in part originate from fusions between normal human cells and bacteria, the fusion product termed “bactocytes” (Shaw 2003).

Pathogenic Features of Liver Metastases: Transfer of Cell-Free Genomic and Signaling Components Between Metastatic Cells and Between Normal Cells and Tumor Cells. An Extended Concept of Metastasis

Introduction

There is evidence that genetic and signaling information can spread from cancer cells to other cancer cells or normal, nontransformed cells (review: Garcia-Olmo et al. 2012). Through such mechanisms, metastasis of cancer has lived a conceptual extension from spread of malignant cells to spread and intracorporeal distribution of subcellular elements and components.

Exosomes

Exosomes are defined as small membrane-bound vesicles which cells can release into their microenvironment and from there to blood and lymph. The morphogenesis and biogenesis of exosomes involve a process of inward budding of membrane compartments into the lumen of late endosomes, followed by the fusion of the proto-multivesicular bodies with the plasma membrane. The cargo of exosomes consists of noncoding RNAs (including microRNAs and onco-miRs), signaling substances, adhesion molecules, and lipids (Rak 2010; Ludwig and Giebel 2012). As exosomes carry and transport bioactive lipids, and are enriched in cholesterol and sphingolipids in comparison with their parental cells, exosomes can serve as structures mediating intercellular communication dependent on lipid signaling (Record et al. 2014). The transfer of this exosomal cargo may induce pathologic alterations at the exosomal terminals (so-called exosomopathies; Korc 2015). Circulating exosomes and their cargo can successfully be employed as molecular biomarkers for diagnosis, classification, and monitoring of various malignant neoplasms.

Circulating Nuclear DNA (fcDNA)

In cancers, release of intact nuclear DNA or of DNA fragments from cancer cells is a common phenomenon, due to ongoing decay of neoplastic cells and their nuclei caused by apoptosis and necrosis. This DNA enters circulation in the form of free circulating DNA (fcDNA; circulating nucleic acids in plasma and serum, CNAPS; Schwarzenbach et al. 2011; Garcia-Olmo and Garcia-Olmo 2012; Garcia-Olmo et al. 2012; Gonzalez-Masia et al. 2013). In addition to cancer cells, also normal cells and cells of benign tumor release fcDNA. Assessments of fcDNA can be used as diagnostic markers, molecular characterization of primary and metastatic tumors, and design of personalized therapies (Goessl 2003; Fleischhacker and Schidt 2010; Vlassov et al. 2010; Ulivi and Silvestrini 2013).

Circulating Nucleosomes (cNUCs)

Nucleosomes as complexes of nuclear DNA and histone proteins can be released as intact particles from stressed, dying, or dead cells, including cancer cells. In patients with various malignancies, plasma and serum concentrations of nucleosomes are frequently elevated (Gezer et al. 2013). This “nucleosomics” can serve as a biomarker, e.g., for the detection of colorectal cancer (Holdenrieder et al. 2014) and the prognostication in cancer patients undergoing chemotherapy (Stoetzer et al. 2012). Nucleosomes are released from cells stimulated intrinsically or extrinsically to undergo apoptotic cell death, but in certain cells, e.g., neutrophils, nucleosomes can also be released by an active secretion process (Holdenrieder et al. 2012). Increased circulating nucleosomes (cNUCs) are also detected in autoimmune diseases, psoriasis, sepsis, trauma, and situations of large tissue necrosis. As a vehicle for DNA, circulating nucleosomes in cancer patients can be employed similar to fcDNA for diagnostic and monitoring purposes (Holdenrieder et al. 2001; Holdenrieder et al. 2008, Holdenrieder and Stieber 2009; Stoetzer et al. 2012; Wittwer et al. 2013). Nucleosomes and in particular their histones can also form the cargo of exosomes. Interestingly, cancer patients show distinct posttranslational modifications, including trimethylated histone tails, of histones carried by exosomes (Gezer et al. 2012; Leszinski et al. 2012; Gezer and Holdenrieder 2014).

Circulating Mitochondrial DNA (ccfmtDNA)

Similar to nuclear DNA, also mtDNA can escape from damaged normal and cancer cells and enter circulation, either as free mtDNA or in the setting of mitochondria. Circulating mtDNA (circulating cell-free mtDNA, ccfmtDNA) increases with age and can act as so-called damage-associated molecular pattern (DAMP) agents and are considered to be a pathogenic component for a low-grade chronic inflammatory status occurring in elderly subjects, a condition termed “inflamm-aging” (review : Pinti et al. 2014). mtDNA can undergo characteristic mutational changes in various tumors. Circulating cancer-derived mutant mtDNA can therefore be used as a predictive biomarker in several neoplasms (Ellinger et al. 2012; Uzawa et al. 2012; Yu 2012).

microRNAs (miRNAs) and Other Noncoding RNAs

Micro-RNAs play an important role in the pathogenesis of metastases, as they control the expression of stem cell molecules and molecules involved in invasion and spread (Nicoloso et al. 2009). All examined biological fluids in the body contain extracellular miRNAs that are nuclease resistant. In patients with cancer, neoplastic cells can release miRNA and via this mechanism transfer part of their molecular signature into the organism. In addition to represent a cancer marker, extracellular miRNA are considered molecules with an important cell-to-cell communication function (Turchinovich et al. 2013). Part of miRNAs released into the extracellular space enter blood and lymph circulation and thus become circulating miRNAs. As these miRNAs are highly stable, they are suitable as diagnostic cancer markers that can easily be detected and measured. Circulating miRNAs have now been utilized as markers for several solid neoplasms and hematologic malignancies (Zheng et al. 2013). In particular, plasma miRNA have been identified and evaluated for the early detection, monitoring, and outcome prediction of CRCs and their metastases (Giraldez et al. 2013; Luo et al. 2013; Kjersem et al. 2014). Certain species of miRNA are characteristic for CRCs, such as circulating miRNA-106a, −130b, −378, and −484 (Kjersem et al. 2014; Zanutto et al. 2014). MicroRNA signatures can discriminate between colorectal cancer recurrences to lymph nodes and liver and between CRC metastases and primary hepatic tumors (Drusco et al. 2014). Lymph node metastases of colorectal cancer express distinct patterns of long noncoding RNAs (lncRNAs) (Han et al. 2014), a phenomenon that might, similar to microRNAs, be transferred to remote areas in the body. Apart from microRNAs, other types of noncoding RNAs can be released into circulation, including small interfering RNA (Valiunas et al. 2015).

Exosome-Derived MicroRNAs (miRNAs)

Apart from miRNAs freely circulating in blood and lymph following their release from normal and cancer cells, miRNA can also travel in circulation as cargo located within exosomes (Camacho et al. 2013; Kosaka et al. 2013). Exosomal miRNAs (exomiRNAs) are thought to have a significant role in intercellular communication and transfer of specific signals, as circulating exomiRNA can be taken up by recipient cells to alter gene expression in these cells via RNA interference. This mechanism can induce several functional changes in recipient cells, including induction of pro-inflammatory cytokine expression (Mobergslien and Sioud 2014); exomiRNAs are increasingly employed as noninvasive biomarkers for diagnosing and screening of malignancies (Cazzoli et al. 2013).