Chapter 5: Growth factors, receptors, and cancer
5.2: The Src protein functions as a tyrosine kinase
The first clues to how cell-to-cell signaling via growth factors operates came from biochemical analysis of the v-src oncogene and the protein that it specifies.
Src operates as a protein kinase—an enzyme that removes a high energy phosphate group from ATP and transfers it to a suitable protein substrate.
Independent of this kinase activity, Src itself is a phosphoprotein, that is, it carries phosphate groups attached covalently to one or more of its amino acid side chains.
This indicates that Src also serves as a substrate for phosphorylation by a protein kinase—either autophosphorylation or serving as the substrate of yet another kinase.
Src specifically phosphorylates certain tyrosine residues of its protein substrates.
This is pretty rare for a kinase, and transformation of cells by the v-Src oncogene saw dramatically risen phosphotyrosine levels.
Signaling through tyrosine phosphorylation is a device that is used largely by mitogenic signaling pathways in mammalian cells, whereas the kinases that are involved in regulating thousands of other processes rely almost exclusively on serine and threonine phosphorylation to convey their messages.
5.3: The EGF receptor functions as a tyrosine kinase
Epidermal growth factor (EGF) was the first of growth factor to be discovered.
It was initially characterized because of its ability to provoke premature eye opening in newborn mice.
Soon after, EGF was found to have mitogenic effects when applied to a variety of epithelial cell types.
A cell surface protein, an EGF receptor (EGF-R), was able to specifically recognize EGF in the extracellular space, bind to it, and inform the cell interior that an encounter with EGF had occurred.
EGF-R’s N terminus domain protrudes into the extracellular space (called its ectodomain) and is clearly involved in recognizing and binding the EGF ligand.
The C terminus extends into the cytoplasm.
The overall structure of the EGF receptor suggested in outline how it functions.
After its ectodomain binds EGF, a signal is transmitted through the plasma membrane to activate the cytoplasmic domain of the receptor.
Once activated, the latter emits signals that induce a cell to grow and divide.
It was found that the cytoplasmic domain had similarities with the already known sequence of the Src protein.
This indicated that once its ectodomain binds EGF, the Src-like kinase in the cytoplasmic domain becomes activated, proceeds to phosphorylate tyrosines on certain cytoplasmic proteins, and thereby causes a cell to proliferate.
5.4: An altered growth factor receptor can function as an oncoprotein
A really important discovery occurred in 1984 when the sequence of the EGF receptor was recognized to be closely related to the sequence of a known oncoprotein specified by the ErbB oncogene.
ErbB was discovered originally in the genome of avian erythroblastosis virus (AEV), a transforming retrovirus that rapidly induces a leukemia of the red blood cell precursors (an erythroleukemia).
A protein used by the cell to sense the presence of a growth factor in its surroundings had been appropriated (in its avian form) and converted into a potent retrovirus-encoded oncoprotein.
The oncoprotein was found to lack sequences present in the N-terminal ectodomain of the EGF receptor.
Without these N-terminal receptors, the ErbB oncoprotein clearly cannot recognize and bind EGF, and yet it functions as a potent stimulator of cell proliferation.
Deletion of the ectodomain enables the resulting truncated EGF receptor protein to send growth-stimulating signals in a constitutive fashion, fully independent of EGF.
The presence in breath cancers of similarly truncated versions of ErbB/EGF-R’s close cousin, termed variously ErbB2, HER2, or Neu, were found to be associated with particularly poor prognosis.
These insights into receptor function provided one solution to a long-standing problem in cancer cell biology.
The discovery of the ErbB—EGF-R connection yielded a simple, neat explanation of this particular trait of cancer cells:
The ErbB oncoprotein releases signals very similar to those emitted by a ligand-activated, wild type, EGF receptor.
However, unlike the EGF receptor, the ErbB oncoprotein can send a constant, unrelenting stream of growth-simulating signals into the cell, thereby persuading the cell that substantial amounts of EGF are present in its surroundings when none might be there at all.
In the presence of such an oncoprotein, a cell is liberated from dependence on growth factors present in its surrounding.
5.5: A growth factor gene can become an oncogene: the case of sis
In 1983, the B chain of the platelet-derived growth factor (PDGF) was found to be closely related in sequence to the oncoprotein encoded by the v-sis oncogene of the simian sarcoma virus.
The PDGF protein was discovered to be unrelated in structure to EGF and to stimulate proliferation of a different set of cells.
PDGF stimulates largely mesenchymal cells, such as fibroblasts, adipocytes, and smooth muscle cells; in the brain, it can also stimulate the growth and survival of glial cells.
By contrast, the mitogenic activities of EGF and related GF ligands are focused largely on epithelial cells.
Like the EGF receptor, PDGF-R uses a tyrosine kinase in its cytoplasmic domain.
The connection between PDGF and the sis-encoded oncoprotein suggested another important mechanism by which oncoproteins could transform cells:
When simian sarcoma virus infects a cell, its sis oncogene causes the infected cell to release copious amounts of the PDGF-like Sis protein into the surrounding extracellular space.
Once there, the PDGF-like molecules can attach to the PDGF-R displayed by the same cell that just synthesized and released them.
The result is strong activation of this cells PDGF receptors and, in turn, a flooding of the cell with an unrelenting flux of the growth-stimulating signals released by the ligand-activated PDGF-R.
This represents an auto-stimulatory, or autocrine, signaling loop in which a cell manufactured its own mitogens.
Such autocrine signaling loops seem to represent potential perils for tissues and organisms.
In normal tissues, the proliferation of individual cells almost always depends on signals received from other cells; such interdependence ensures the stability of cell populations and the constancy of tissue architecture.
A cell that has gained the ability to control its own proliferation (by making its own mitogens) therefore creates an imminent danger, since self-reinforcing, positive-feedback loops often lead to gross physiological imbalances.
5.6: Transphosphorylation underlies the operations of many receptor tyrosine kinases
By supplying cells with a continuous flux of growth-stimulatory signals, oncoproteins like sis and ErbB are able to drive the repeated rounds of cell growth and division that are needed in order for large populations of cancer cells to accumulate and for tumors to form.
How to growth factor receptors use their tyrosine kinase domains to emit intracellular signals in response to ligand binding?
Many growth factor ligands were found to be dimeric, being composed of two identical protein subunits (homodimers) or of very similar but non identical subunits (heterodimers).
An examination of the proteins that become phosphorylated within seconds after a growth factor such as EGF is applied to cells showed that the most prominent is often the receptor molecule itself. Hence, these receptors seem to be capable of autophosphorylation.
Also, many transmembrane proteins constructed like EGF and PDGF receptors have lateral mobility in the plane of the plasma membrane. As long as the hydrophobic transmembrane domains of these receptor proteins remain embedded within the lipid bilayer of this membrane, they are relatively free to wander back and forth across the surface of the cell.
Taken together, these facts led to a simple model:
In the absence of a ligand, a growth factor receptor exists in a monomeric form, embedded in the plasma membrane. It may wander laterally in the membrane and encounter another identical RTK, with which it forms an unstable homodimeric complex.
However, when presented with the appropriate growth factor ligand, the homodimeric receptor complex will be stabilized.
Next, the cytoplasmic portions of the RTKs are pulled together. Now, each kinase domain phosphorylated tyrosine residues present in the cytoplasmic domain of the other receptor.
The receptor dimerization model explains how growth factor receptors can participate in the formation of cancers when the receptor molecules are overexpressed.
Since these receptors are free to move laterally in the plane of the plasma membrane, their high numbers can cause them to collide frequently, and these chance encounters, like the dimerization events triggered by ligand binding, can result in transphosphorylation, receptor activation, and signal emission.
The EGF receptor is overexpressed in a wide variety of human tumors, mostly carcinomas.
In such tumors, this overexpression may result in ligand-independent receptor dimerization and firing.
The most direct and compelling demonstration of the key role of receptor homodimerization comes from the mechanisms by which fusion proteins formed in many human cancers drive constitutive receptor signaling.
5.10: The Ras protein, an apparent component of the downstream signaling cascade, functions as a G protein
Ras proteins carry covalently attached lipid tails that allow it to become anchored to the cytoplasmic face of the membrane.
Ras was found to:
Bind a GDP molecule when in its inactive state
Jettison its bound GDP after receiving some stimulatory signal from upstream in a signaling cascade
Acquire a GTP molecule in place of the recently evicted GDP
Shift to an activated, signal emitting configuration while binding to this GTP
Cleave this GTP after a short period, using its own intrinsic GTPase function to do so, thereby placing itself, one again, in its non-signal emitting configuration.
In more detail, mitogenic signals activated a GEF for Ras, that induced an inactive Ras to shed its GDP and bind GTP instead.
This period of signaling would be terminated when Ras decided to hydrolyze its bound GTP with GAP.
A striking discovery was made in the course of detailed biochemical analysis of a Ras oncoprotein made by a point mutated Ras oncogene.
Like the normal Ras protein, the Ras oncoprotein could bind GTP.
However, the oncoprotein was found to have lost virtually all GTPase activity.
In such condition, it could be pushed into its active, signal emitting configuration by some upstream stimulatory signal and GEF.
But, once in this activated state, it was unable to turn itself off.
Rather than sending out short, carefully rationed pulses of growth stimulating signals, the oncoprotein emits signals for a long, possibly indefinite period of time, thereby flooding the cells with these signals.
It was found that missense mutations at the 12th, 13th, or 61st codon of the reading frame of the Ras genes cause these oncoproteins.
These residues are located around the cavitt in the Ras protein where the GTPase catalytic activity operates.
Consequently, almost all substitutions of these three amino acids, such as the glycine to valine substitution, compromise the ability of GAP to trigger hydrolysis of GTP.
Even though a point mutations affecting all the amino acid residues of a Ras protein are likely to occur with comparable frequencies, only those few conferring substantial proliferative advantage will actually be found in tumor cells.
Other cells that happen to acquire point mutations affecting other residues in the Ras protein will retain a normal phenotype or may even lose proliferative ability.