Molecular Biology of the Cell - Volume 2 - Alberts B., Bray D., Lewis J., Raff M., Roberts K., Watson J. 1993

Cell Growth and Division
"Social Control" Genes in Cell Division

As we saw in the case of Yeast, genetics offers powerful tools for elucidating the molecular mechanisms regulating Cell Division, provided there are ways to select Mutations in the corresponding genes. In Multicellular animals, mutations in genes that directly participate in the "social control" of cell division (which we will refer to as social control genes) are readily isolated. A cell carrying one or more such mutations will continue to divide, ignoring the needs of the Organism as a whole, and its progeny will form a macroscopically visible tumor.

Cancers are, by definition, malignant tumors—meaning their Cells not only divide uncontrollably but also invade other Tissues, giving rise to numerous secondary tumors, or metastases. For Cancer to develop, a cell must first undergo a series of mutations that free it from various regulators of cell division, followed by the accumulation of further changes that confer the capacity for Invasion and Metastasis. These aspects of cancer will be discussed in Chapter 21. Here, however, we will not attempt to explain The Nature of cancer itself; instead, we will examine what The Study of cancer cells can reveal about the genes that normally govern cell division.

13.4.1. Cell Transformation in Culture Provides a Way to Identify Genes Involved in the Social Control of Cell Division

How can a mutant Gene (or genes) be identified when one has a clone of tumor cells presumably generated by mutation? Classical gene mapping Methods are inapplicable here because the cells do not reproduce sexually. A more direct approach consists of isolating genetic material from tumor cells and screening for fragments that, when introduced into normal cells, induce behavior resembling that of tumor cells. Methods to solve such a formidable problem were first developed only in the late 1970s, but they built upon earlier studies of very similar natural phenomena.

Some types of tumors are caused by Viruses. Viruses isolated from such tumors infect normal cells and, by introducing their own RNA or DNA, transform these cells into tumor cells. The first such tumor-inducing (oncogenic) virus discovered was the Rous Sarcoma virus, which causes Connective Tissue tumors (sarcomas) in birds. It became a crucial research model, alongside several other Oncogenic Viruses discovered later.

The nature of tumor transformation is most easily studied in cell cultures. Within a few days after The addition of an oncogenic virus, small colonies of abnormally proliferating cells appear in the culture. Each such colony is a clone derived from a single cell that was infected by the virus and integrated the viral genetic material into its genome. Released from the social control of cell division, transformed cells grow faster than normal cells in a culture dish, just as they do in the organism, making them easy to isolate. Transformed cells frequently exhibit a whole complex of abnormalities (Table 13-2): they lack contact inhibition of cell division (Section 13.3.5)—instead, piling up on top of one another as the culture grows (Fig. 13-31); they often do not require anchorage to grow and are capable of dividing even in suspension; they display a more rounded Morphology, reminiscent of normal cells in mitosis, with weak adhesion to the substrate and to one another; they can divide even in the absence of growth factors; they are immortal and do not undergo senescence in culture; and finally, when injected into a suitable host organism, they can give rise to tumors.

Class="center">Table 13-2. Some Changes Frequently Observed Upon Transformation of Normal Cells in Culture by an Oncogenic Virus

Fig. 13-31. Unlike most normal cells, cancer cells continue to grow and pile up on one another after forming a continuous monolayer.

13-20

13.4.2. Tumor-Inducing Viruses Serve as a Source of Easily Clonable Oncogenes [28]

A tumor virus disrupts normal control of cell division by irreversibly altering the genetic constitution of the host cell, causing The Cell to produce a protein immune to normal regulatory mechanisms. Such viruses thus provide a means to uncover the mechanisms normally responsible for controlling cell division. To date, the most important results have been obtained from the study of RNA-containing tumor viruses, also known as Retroviruses. Following infection of a cell by a retrovirus, DNA is synthesized from its RNA via reverse METABOLISM/31.html">Transcription and is subsequently integrated into the host cell genome. The life cycle of a retrovirus is illustrated in Fig. 5-75.

When a retrovirus transforms a normal cell into a tumor cell, the abnormal behavior is frequently driven by a gene that is introduced by the virus but is actually unnecessary for the virus's own survival and Replication. This was first discovered through the isolation of Rous sarcoma virus mutants that could reproduce normally but failed to transform the cell. It turned out that some of these non-transforming mutants lack the gene (or part of the gene) encoding a protein with a Molecular Weight of 60,000. Through other mutations in this gene, the transforming action of the virus can become Temperature-sensitive: infected cells display a transformed phenotype at 34°C, but upon raising the temperature to 39°C, they rapidly (within a few hours) revert to a normal phenotype (Fig. 13-32). Apparently, this specific gene within the oncogenic virus is responsible for cellular transformation (thereby capturing our attention) while remaining superfluous ballast from the standpoint of viral reproduction itself.

Fig. 13-32. Cells infected with a Rous sarcoma virus carrying a temperature-sensitive mutation in the transformation-related gene (the v-src oncogene) (scanning electron micrographs). A. Cells are transformed and have acquired an abnormal rounded morphology at the permissive low temperature (34°C) where the oncogene product is active. B. The same cells, firmly attached to the culture dish and having restored their normal flattened morphology after the oncogene product has been inactivated by a temperature shift to 39°C. (Courtesy of G. Steven Martin.)

The transforming gene of the Rous sarcoma virus identified in these experiments was named the v-src gene. It was classified as an oncogene (from the Greek *onkos*, meaning bulk or tumor) because, when introduced into a normal cell, it converts the cell into a tumor cell. What is THE ORIGIN OF this gene, and what is its normal function? When a radioactive DNA copy of the viral src gene was used as a probe to search for related sequences via DNA-DNA Hybridization (Section 4.6.7), it turned out that the genomes of normal vertebrate cells contain a very similar, though not identical, sequence. This homolog of the viral src gene in normal cells is designated c-src and is classified as a proto-oncogene. Presumably, the viral oncogene was once captured from the host cell genome and underwent mutation. It is likely that the proto-oncogene is one of the normal social control genes, and the retrovirus has essentially cloned it for us. Today, many other oncogenes have been identified and analyzed in the same way, and in each case, this has led to the Discovery of the corresponding proto-oncogene.

13-21

13.4.3. Tumors Arising by Diverse Mechanisms Contain Mutations in the Same Proto-oncogenes [26, 29]

Tumors frequently arise not from viral infection, but As a result of mutations occurring spontaneously or induced by chemical carcinogens or radiation. Purified DNA can be extracted from such tumor cells and tested for the presence of oncogenes by Introduction into untransformed cells in *in vitro* culture. NIH/3T3 cells are frequently used for this assay because they divide indefinitely in culture and carry mutations that make them readily transformable by the addition of just a single oncogene. The oncogene responsible for such transformation can be isolated, cloned, and sequenced using Recombinant DNA technology (Fig. 13-33). Remarkably, in most cases where this has been done, the oncogene turned out to be a mutant form of one of the very same proto-oncogenes that were isolated using retroviruses, although a few novel oncogenes were also discovered. Using similar approaches, it has been established that transformation can also result from the overexpression of certain normal gene products. Tumors frequently contain an unchanged proto-oncogene whose overexpression is driven either by the presence of an excessively high copy number or by a chromosomal rearrangement that places it under the control of an inappropriate promoter. These issues will be discussed in Chapter 21.

Fig. 13-33. A method for identifying and cloning human oncogenes. Oncogenes contained within a DNA sample extracted from a human tumor are detected by their ability to transform mouse 3T3 cells. Transformed 3T3 cells divide uncontrollably and are recognized by the colonies they form in a culture dish. Repetitive sequences of the Alu family (Section 10.5.11) are scattered throughout The Human Genome and serve as a convenient marker that can be used to identify human DNA within cells of another organism. Utilizing Alu sequences as a DNA probe allows the cloning of the human oncogene that transforms 3T3 cells. Upon retesting with this same method, the cloned DNA containing the oncogene will transform 3T3 cells with high efficiency.

To date, more than 50 proto-oncogenes have been identified (Sections 21.2.3, 21.2.4). They apparently constitute a significant fraction of the normal cell's proto-oncogenes. However, many social control genes probably remain undiscovered. Fibroblast-like 3T3 cells commonly used in transformation assays may fail to respond to an oncogene that would transform Other types of differentiated cells. Furthermore, the cell transformation assay only allows the detection of dominant mutations in social control genes—that is, mutations that disrupt the Regulation of Cell division even in the presence of copies of the normal allele in the cell. It is possible that recessive mutations in social control genes, caused by the loss of gene function, are more common in cancer cells, yet they cannot be detected by this assay. Therefore, genes whose products normally help stimulate cell division are easier to identify using current methods than those whose products normally inhibit division. Nonetheless, evidence already exists that growth-inhibiting genes do exist and that their recessive mutations frequently underlie cellular transformation and cancer. For example, following the fusion of transformed cells with untransformed ones, the resulting hybrid cells very frequently behave as normal cells; cell division control appears to be restored due to the appearance of a protein that was absent in the transformed cell. Therefore, keeping in mind that many important social control genes remain undiscovered, we will now examine the functions of those genes that are already known.

13-16

13-22

13.4.4. Some proto-oncogenes encode growth factors or growth factor receptors [26, 30]

Once an oncogene has been cloned and sequenced, clues to the likely function of the corresponding social control gene (proto-oncogene) can often be found by comparing The nucleotide sequences of the oncogene and already known genes. It was precisely in this way that one of the proto-oncogenes, c-sis, was discovered to encode a functionally active subunit of the PDGF growth factor. A cell containing the corresponding v-sis oncogene continuously and needlessly produces this subunit, which, by binding to the cellular PDGF receptor, constantly stimulates the cell to proliferate. At least three other proto-oncogenes—c-erbB, c-fms, and c-erbA—also encode receptors for growth factors or Hormones: c-erbB encodes the epidermal growth factor receptor (Section 12.3.13), c-fms encodes the macrophage colony-stimulating factor (M-CSF) receptor (this factor promotes the proliferation of macrophage precursors, Section 17.5.8), and c-erbA encodes the thyroid hormone receptor (Section 12.2). Having been converted into oncogenes as a result of mutations, these genes will encode altered receptors that behave as though they had bound a Ligand (even in its absence) and therefore stimulate the cell when it is unnecessary (Section 12.3.14).

Fig. 13-34. Activity and cellular localization of Major Classes of known proto-oncogenes. The names of some typical proto-oncogenes in each class are highlighted in color. See also Fig. 21-27.

Although the functions of most other proto-oncogenes are not yet precisely known, it is reasonable to assume that the majority of them encode Proteins of the Intracellular Signaling network that enables growth factors to stimulate cell proliferation. We must now consider the wide range of functions that the already known groups of proto-oncogenes, presented in Fig. 13-34, can perform.

13.4.5. Some proto-oncogenes encode intracellular mediators involved in stimulating cell division

As discussed in Chapter 12, the receptors for most growth factors, including PDGF, are Tyrosine-specific protein Kinases that, upon activation, phosphorylate themselves and various other proteins. One group of oncogenes encodes abnormal forms of such receptors, including the aforementioned altered EGF and M-CSF receptors (Fig. 13-34). Another small family of similar proto-oncogenes, the c-ras genes, encodes proteins that bind and hydrolyze GTP and are likely distantly related to G proteins involved in transmitting many types of signals (Section 12.3.11). The action of mutant ras genes, which cause cell transformation, is associated with elevated concentrations or increased efficacy of the intracellular mediators Inositol trisphosphate and diacylglycerol, rendering the cell hypersensitive to certain growth factors; these factors are thought to induce The production of the aforementioned mediators. Genes homologous to ras are found in budding Yeasts, where they participate in regulating the cell division cycle in response to nutrient levels in the medium.

Fig. 13-35. This electron micrograph shows that the protein kinase encoded by the v-src oncogene of the Rous sarcoma virus is attached to the inner surface of the plasma membrane; the src protein produced by the c-src gene is believed to reside in the same Location, but is more difficult to detect because it is normally present in very small amounts. The localization of the src protein in this preparation was determined by its reaction with specific Antibodies coupled to electron-dense ferritin particles. (Courtesy of Ira Pastan; M. C. Willingham, G. Jay, I. Pastan, Cell 18: 125-134, 1979. Copyright Cell Press.)

The normal response of an animal cell to growth factor stimulation involves many other intracellular effects, including changes in pH, in Ca²⁺ and cyclic AMP concentrations, in protein phosphorylation, in gene transcription, in mRNA Processing and degradation, in Protein Synthesis, and in the Cytoskeleton. Many of these changes occur within seconds, whereas others require hours. Most of the protein products of proto-oncogenes involved in this complex regulatory network are still poorly classified. Some of them, such as the growth factor receptors mentioned above, are membrane-bound tyrosine-specific protein kinases. Others, such as the product of the cdc28 gene in yeast, are cytoplasmic Serine/Threonine-specific protein kinases. A third category comprises proteins located predominantly in The Nucleus (see Fig. 13-34); one of these, the c-jun protein, has been identified as the AP-1 transcription regulator (Section 10.2.8, Table 10.1); it can combine with another member of the same family, the c-fos protein, to form a DNA-binding complex.

Another nuclear group protein, corresponding to the c-myc proto-oncogene, apparently serves as an indicator of whether a cell is in a proliferative state: in rapidly dividing cells, the c-myc protein is present at a constant low concentration throughout the cycle, but as soon as the cell enters the resting G₀ state, it disappears. When growth factors are added to a medium containing resting cells, the concentration of the c-myc protein rises sharply, peaks within a few hours, and then drops to a lower non-zero level. In contrast to c-myc, the vast majority of other cellular proteins exhibit almost no change in concentration during the transition from the proliferative state to quiescence or vice versa.

13.4.6. The Effect of oncogenes on cell division regulation is closely linked to their impact on Cell Adhesion [32, 33]

One of the most intensively studied proto-oncogenes is c-src, which corresponds to the v-src oncogene of the Rous sarcoma virus. It belongs to a small family of homologous proto-oncogenes and encodes the src protein—a tyrosine-specific protein kinase with a molecular weight of 60,000 (hence also called p60^src)—which contains a covalently attached fatty acid that anchors it to the inner leaflet of the plasma membrane (Section 8.2.3). In its oncogenic form, this kinase is hyperactive, and its attachment to the membrane is essential for it to induce cell transformation (Fig. 13-35). Antibody experiments show that the src protein is concentrated in focal contacts, i.e., sites where the cell is anchored to the substrate via matrix junctions involving intracellular actin filaments (Fig. 13-36). The src protein appears to play a role in anchoring actin to the membrane, since Activation of a temperature-sensitive variant of the v-src protein (by lowering the temperature) immediately leads to increased membrane ruffling (as a result of lamellipodia movement driven by actin; see Section 11.2.11), as well as to a decrease in cell adhesion, including the disruption of focal contacts, causing the cell to round up (see Fig. 13-30). The observed alteration in adhesion is attributed to at least two independent effects of the src protein. First, the active v-src kinase phosphorylates a tyrosine residue in the intracellular "tail" of the fibronectin receptor (Section 11.2.8). In vitro studies suggest that this phosphorylation reduces the affinity of the receptor for both talin (intracellularly) and fibronectin (extracellularly). In addition, cells transformed by v-src secrete large amounts of a proteolytic enzyme called plasminogen activator. This name derives from its ability to activate another proteolytic enzyme, plasmin, by cleaving its precursor, plasminogen; however, plasminogen activator is also capable of directly degrading other proteins. Apparently, through both Direct and Indirect actions, it helps the cell weaken its attachment and migrate within the Extracellular matrix. When a monoclonal antibody against plasminogen activator is added to the culture medium, the cells become more adhesive and tend to spread out on the substrate. Thus, the excessive activity of the v-src tyrosine kinase apparently weakens cell adhesion in two ways: it phosphorylates fibronectin receptors (and other transmembrane molecules of the integrin group involved in cell-matrix attachment—see Section 14.1.3) and induces the secretion of a proteinase that degrades fibronectin (and other matrix molecules).

Fig. 13-36. The src protein is present in many Regions of the cell, but is particularly concentrated, apparently, in focal contacts and other sites of cell attachment to the extracellular matrix. A. Immunofluorescence micrograph revealing the distribution of the src protein via the binding of src-specific antibodies. B. The same cell viewed using optical Interference reflection Microscopy, which highlights areas of tight cell-substrate attachment (dark regions). The distribution of the bright spots of the src protein in photo A corresponds to the distribution of the dark attachment spots in photo B (indicated by arrows). These photographs show the distribution of the viral src protein in a cell transformed by Rous sarcoma virus. The src protein synthesized in a normal cell is presumably distributed in a similar manner, but is harder to detect because it is present in smaller amounts. (L. R. Rohrschneider, Proc. Natl. Acad. Sci. USA 77: 3514-3518, 1980.)

What bearing might these facts have on The regulation of cell growth? If a quiescent fibroblast is treated with PDGF, it immediately undergoes ruffling deformation, and the cell's focal contacts restructure within minutes: vinculin temporarily disappears from these contacts, and the bundles of actin filaments anchored there are transiently disrupted. Thus, in a resting cell stimulated to divide, PDGF triggers many of the same changes that are induced by v-src. Indeed, among the immediate changes caused by PDGF is enhanced phosphorylation of the c-src protein, which, by increasing src kinase activity, could directly account for this similarity in effects (Fig. 13-37). From this perspective, cell transformation by the v-src oncogene (and many other oncogenes with similar action) represents, as it were, an exaggerated effect of the normal growth-stimulation mechanism, which also involves the weakening of cell adhesion. The danger of oncogenes to the organism lies in the fact that, unlike PDGF, which stimulates the cell only briefly, proteins like v-src tend to irreversibly drive the cell out of the G₀ state and thereby keep it in a proliferative state.

Fig. 13-37. A speculative model illustrating how rapid changes in cell adhesion might occur when cells are stimulated to divide by PDGF. The binding of PDGF to its receptor leads (by an as yet unknown pathway) to the phosphorylation of the c-src protein. As a result, this plasma membrane-associated protein kinase is activated and in turn phosphorylates tyrosine residues on neighboring transmembrane cell adhesion proteins, including the fibronectin receptor. This causes focal contacts and other cell adhesion sites to partially break down, and the actin filaments linked to them lose their connection to the membrane. In part, this model is based on observations of cells transformed by Rous sarcoma virus, which contain a constitutively active modified src protein (v-src). Tyrosine protein kinases encoded by two other proto-oncogenes of the src family—c-abl and c-yes—might function in the same manner as the c-src protein in the mechanism described above. However, such Enzymes typically phosphorylate many proteins, and it remains unclear which of these play a decisive role in controlling cell division. Some important targets may be present in the cell in only one or a few copies (which is too few to detect by standard Biochemical Methods), and targets may vary among different cell types. Moreover, it is difficult to establish causal relationships within a complex network of interacting components where many factors may act in parallel and the same effect can be achieved through different pathways.

13.4.7. The relationship between cell proliferation and cell adhesion remains poorly understood [33]

Our Structure/133.html">Discussion of the control of normal vertebrate cell proliferation leads to a paradox. On the one hand, it is clear that in order to exit the G₀ state and begin dividing, normal cells must form adhesive contacts with the substrate (cell-matrix adhesion). This suggests that transmembrane proteins linking cells to the extracellular matrix (including the fibronectin receptor and other integrin group proteins) generate an intracellular signal that facilitates the division of cells in an appropriate state. On the other hand, adhesion per se is not sufficient to trigger division—growth factors are also required. The paradox here is that growth factors appear to act, in part, by temporarily weakening the adhesion upon which normal cell proliferation depends (Fig. 13-37). This brings to mind a second observation: in many cases, brief Treatment of growth-arrested normal cells with a proteolytic enzyme (such as Trypsin), which causes attached cells to lose contact with the substrate and round up, has the additional side effect of triggering a single round of cell division. Apparently, cell proliferation is transiently stimulated both by extracellular proteases that directly weaken cell-matrix adhesion by digesting adhesive extracellular proteins, and by growth factors that disrupt this adhesion indirectly by acting on focal contacts through intracellular mediators.

Studies on cancer cells heighten this paradox. Most such cells, including those transformed by the well-studied oncogenes shown in Fig. 13-34, differ from their normal counterparts in that they do not require attachment to a substrate in order to divide. Since such anchorage independence allows transformed cells to grow in novel environments where normal cell-cell and cell-matrix contacts cannot be established (Section 13.3.6), one can surmise that it arose through the natural Selection of tumor-forming cells. But why do many cancer cells not merely divide independently of anchorage, but fail to firmly attach to the extracellular matrix even when given the opportunity? A hint toward an answer comes from observations of transformed cells that are artificially forced to attach to a culture dish. As noted above, chicken embryo fibroblasts transformed by v-src secrete large amounts of plasminogen activator, which weakens their attachment to the dish. If such cells are grown in the presence of an antibody that blocks The activity of this protease, they attach more firmly to the dish and at the same time become more susceptible to the normal social control of cell division: instead of forming a multilayered structure, they tend to stop dividing upon mutual contact. Thus, firm adhesion to the extracellular matrix apparently inhibits the growth of these transformed cells.

Fig. 13-38. The general nature of social control signals acting on normal and transformed cells. In both cases, various growth factors (designated here as 1, 2, and 3) act cooperatively to "lift" the cell from G₀ into a proliferative state. Because a transformed cell is maintained in a position close to the transition threshold (colored line), it can often be stimulated to proliferate by a single growth factor alone (or by a very low concentration of a growth factor mixture). However, as noted in the text, There is a fundamental difference in the effect of attachment (designated as A) between these two cell types: attachment is essential for the proliferation of normal cells, whereas in transformed cells it tends to inhibit proliferation.

Fig. 13-39. A hypothetical scheme explaining the observed dependence of normal and transformed cell proliferation on cell-matrix adhesion. The focus is on Two Types of molecules: 1) a cytoplasmic protein that serves as an intracellular division signal, and 2) a transmembrane linker protein capable of binding both to the cytoplasmic signaling molecule on one side of The cell membrane and to the extracellular matrix on the other side. This binding is a cooperative process, such that signaling molecules bound to the linker protein inside the cell stabilize the transmembrane structure and promote its binding to the extracellular matrix; conversely, binding to the extracellular matrix promotes the binding of signaling molecules to the linker protein inside the cell. For a cell to receive a division signal, the signaling molecules must be unbound in the Cytoplasm and in an active conformation in which they bind less tightly to the linker protein. It is postulated that signaling molecules are activated by phosphorylation, which is driven by the kinase activity of the linker protein.

When an anchored cell is stimulated by growth factors (A), kinase activity is switched on; signaling molecules are phosphorylated and detach from linker proteins, which signals the cell to divide and reduces adhesion. Normal cells in suspension (B) do not divide when exposed to growth factors because very few intracellular signaling molecules bind to the linker proteins that would otherwise phosphorylate them. Transformed cells (C) exhibit reduced adhesiveness and can divide even in suspension, as their regulatory system features a "shunt" that keeps signaling molecules in a permanently phosphorylated state.

The diverse conditions required for the growth of normal and transformed cells, and the contrasting effects of extracellular matrix adhesion on them (Fig. 13-38), can be logically explained. The complex structure formed at focal contacts between a cell and its substrate likely plays a crucial role in generating the intracellular signals that regulate cell division. These phenomena could be explained by assuming that three steps are required to normally trigger division: (1) cell attachment to the matrix mediated by an ordered complex of cytoskeletal proteins assembled within the cell (Section 11.2.8); (2) activation of this complex, typically by one or more growth factors, to generate a division signal; and (3) partial disruption of cell-matrix contacts as a necessary step for transmitting this signal. For a transformed cell, the first step becomes unnecessary, and the uncoupled intracellular Components of the matrix contacts are both necessary and sufficient to trigger a proliferative signal. A hypothetical model of this growth regulation mechanism is presented in Fig. 13-39.

13.4.8. Positional Signals and Autonomous Cellular Programs Control Cell Division in the Growing Organism [20, 34]

Experiments conducted under simplified, artificial cell culture conditions provide much of what we know about the molecular mechanisms controlling Cell Growth and Division in multicellular animals. So far, however, this work has revealed only some of the fundamental "nuts and bolts" of a much more complex system of social control that must operate within an intact organism to regulate the proliferation of each cell group in accordance with its spatial position and prior developmental course.

In discussing The problem of cellular senescence, we suggested that cells are often governed by long-term intracellular programs, meaning that a cell's current proliferative behavior depends on its history of exposure to certain factors many cell generations prior (Section 13.3.10). Although the relationship between long-term and short-term control mechanisms remains enigmatic, both appear to involve many of the same molecules, including growth factors and proto-oncogene products. During embryonic development, cell division programs can be remarkably complex and precise. This is vividly demonstrated, for example, in the nematode Caenorhabditis elegans, whose fertilized egg divides to produce precisely 959 somatic cell nuclei in the adult animal; the study of some of the gene products involved in executing this program has already begun (Section 16.3.3).

However, one should not assume that embryonic growth is regulated simply by counting cell divisions. This became clear, for instance, when comparing newts of varying ploidy. The cells of a pentaploid newt are roughly five times larger than those of a haploid newt, yet because there are correspondingly five times fewer cells in each tissue, the body and organ sizes of both animals are practically identical (Figs. 13-40 and 13-41). Evidently, in vertebrates, the cell division control mechanisms that determine body size are based on measuring dimensions rather than simply counting cells or division cycles. Such mechanisms require sophisticated positional control, in which diffusing growth factors could play a vital role.

Fig. 13-40. Top: Typical sections of renal tubules from axolotls of different ploidy. Pentaploid axolotls have larger cells than haploid ones; however, the animals themselves and their Organs are of identical size because each tissue in the pentaploid animal consists of fewer cells. This indicates that cell size is regulated by a mechanism based on measuring dimensions and distances rather than cell divisions or cell numbers. [G. Fankhauser, In: Analysis of Development (B. H. Willier, P. A. Weiss, V. Hamburger, eds.), pp. 126-150. Philadelphia, Saunders, 1955.]

Fig. 13-41. Micrographs of Brain sections from haploid and tetraploid axolotls (see also Fig. 13-40). A. Cross-section of the Hindbrain of a haploid axolotl. B. Corresponding brain section of a tetraploid axolotl; the reduced cell number is clearly compensated for by an increase in cell size. (G. Fankhauser, Int. Rev. Cytol. 1: 165-193, 1952.)

Positional control of cell division can operate with striking Specificity. When a piece of epithelium from a cockroach leg is transplanted to a homologous site on another, it "heals in" without noticeable cell division. However, if it is transplanted to a non-homologous site, both the transplant cells and the adjacent host cells begin to divide and subsequently differentiate to form the exact cell types that would normally lie between the donor site and the recipient site (Section 16.4.9). The molecular basis for this behavior remains entirely unknown.

In general, cell division during embryonic development is regulated by the combined participation of both autonomous cellular programs and intercellular interactions, though the relative importance of each factor varies from species to species and from one body part to another. In mature tissues, cell division is likewise governed by an intricate network of diverse mechanisms: when repairing a deep Skin wound in vertebrates, for example, approximately 12 distinct cell types—ranging from fibroblasts to Schwann cells—must regenerate in appropriate proportions to make up for the tissue loss. Moreover, the social control system features built-in redundancy with numerous parallel constraints, ensuring that the loss of a single control component in an individual cell (typically the result of a somatic mutation) does not harm the organism as a whole through The Emergence of a massive clone of aggressively dividing cells. Research on cancer indicates that four to six mutations must occur within a single cell Lineage before it gives rise to a malignant tumor (Section 21.1.4).

Elucidating the molecular details of the finely tuned social control mechanisms that allow an organ such as the Kidney to develop and function in an adult organism will likely require the efforts of several generations of cell biologists. Today, however, powerful new tools have become available, such as antibodies that block specific growth factors or receptors, and the generation of Transgenic Animals that produce signaling molecules normally foreign to certain cell types (Section 5.6.10). Thanks to these novel approaches, this formidable challenge no longer seems insurmountable.

Conclusion

Abnormal cells that defy social inhibitory constraints proliferate to form tumors within the organism; they also arise when cells undergo transformation in culture. Although this frequently leads to the death of the organism as a whole, as individual cells they acquire a selective advantage and are therefore easily isolated. Cellular transformation is frequently accompanied by the mutation or overexpression of specific oncogenes, many of which were first identified through their presence in tumor viruses (retroviruses). The normal homologs of these Viral Oncogenes in healthy cells, known as proto-oncogenes, appear to encode Key Components of the normal social control system for cell division. Some proto-oncogenes encode growth factors, others encode receptors for these factors or intracellular regulatory proteins involved in cell adhesion, and still others encode proteins that help transmit cell division signals to the Cell Nucleus. For a cell to become cancerous, multiple genes within the social control network must be altered, highlighting the redundancy of the complex regulatory systems that influence cell proliferation in tissues.