6 Unravelling the genetic origin of cancer

Cancer has been with us throughout human development and some forms have been recognised as a specific disease for thousands of years (e.g. Figure 2.1 or (Binder et al. 2014)). Naturally, understanding of the underlying causes and drivers of cancer was slow to develop in the absence of means to study cellular function at the cellular or molecular level. While a familial susceptibility for some forms of cancer had been identified it was difficult to establish underlying mechanisms. Today, however, it is clear that many of those influences identifiable as risk factors such as smoking or UV radiation as well as hereditary risk can be linked to potential DNA changes or lack of DNA repair.

The next sections start to shed some lights on how genes are involved in cancer. Furthermore, the human genome contains around 30,000 genes. This raises the questions whether all of these genes are relevant and if not what sort of genes are most likely to play a role in cancer development.

6.1 Discovery of oncogenes

The discovery of genes as distinct units of heritable information (Gregor Mendel) predates the realisation that these were ‘stored’ in the DNA or the discovery of DNA structure, or the understanding of how genes were linked to proteins; however, Theodore and Marcella Boveri had made clear predictions about the involvment of chromatin (DNA) in 1904.

One of the first discoveries providing a clear link of DNA and, specifically, of genes as a driver of cancer comes from the discovery that specific genes can give rise to specific proteins that can lead to the development of cancer. Importantly, scientists discovered that these cancer causing genes are very similar to normal genes which have important physiological roles in the body; in fact only very subtle modifications are required to turn such genes into cancer causing genes. The cancer causing genes were called oncogenes and their normal, physiological precursors proto-oncogenes. This discovery happened in a series of steps (Greenwood 2006)15:

In 1911 Peyton Rous, a US pathologist, showed that a malignant tumour of the connective tissue (sarcoma) affecting domestic chicken could be transferred from an affected to a healthy chicken by injection of a cell free fluid extracted from the tumour. For this seminal discovery he received the Nobel Prize in Physiology or Medicine more than 40 year later (1966).16 It turned out that the fluid actually contained a microscopic virus, now known as Rous sarcoma virus (RSV), which today is know to be an enveloped retrovirus.

In retrovirus the genetic material is based on RNA (not DNA). In order to replicate in the host cells these types of virus need to transcribe their RNA into host DNA (reverse transcription). This DNA is then integrated into the cell genome using the virus’ integrase enzyme. Once that has happened this integrated DNA appears to the cell like any of its own DNA; the physiological transcription/translation machinery can be used to transcribe the integrated viral DNA into cellular RNA for subsequent translation into proteins. Once integrated into the DNA the virus will persist indefinitely in the host and its DNA will be passed on to daughter cells up on cell division. The integrated DNA, also known as provirus, can be activated at any point in time once cellular transcription/translation of its genes is triggered to make active virus.

Importantly, retrovirus not only integrate their own genome into the host DNA, but when new virus is produced in the host cells the process can also lead to some of the host genes becoming part of the viral DNA: Varmus and Bishop (‘70,’79) discovered that the viral gene that causes the Rous sarcoma, called v-src, is closely related to a gene found in healthy chicken (c-src). This suggests that at some point in the past the RSV had acquired this gene from an infected cell. This gene is now known as c-src (pronounced ‘c sarc’). However, after integration into the viral DNA subtle changes now had made this viral gene (known as v-src) able to cause cancer.

This led to the realisation that there are specific cellular genes that if transformed can contribute to driving cells towards cancer. These type of genes were called ‘proto-oncogenes’ (e.g. c-src) and their transformed counterparts that give rise to the tumour ‘oncogenes’ (e.g. v-src). Varmus and Bishop were awarded the Nobel Prize in 1989.17 These studies demonstrated subtle changes to our own genes, i.e. mutations, could lead cancer (however, this process of mutation typically does not involve virus).

However, it remained unclear which genes in the human genome are susceptible to becoming cancer-causing?

V-src was found to code for a tyrosine kinase which is able to drive downstream cell mitosis and thus proliferation. In contrast to the normal src gene (c-src) v-src has lost a domain responsible for its inhibition making the enzyme constitutively active (always on). V-src is not required for viral replication but virus which possess it are more virulent as the new cells resulting from the cell divisions already contain the protovirus.

To summarise, the viral gene that can give rise to tumours in chicken was essentially a ‘rogue’ form of the chicken’s own gene; this demonstrates that some of our normal genes when subtly modified can drive cancer. Such cancer causing transformed genes were called oncogenes; such genes typically have a physiological role in promoting cell growth and as they potetnially can drive cancers were called ‘proto-oncogenes.’

6.1.1 Other important human oncogenes

Since these inital discoveries a whole range of other genes have been discovered that when transformed can act like oncogenes.

6.1.1.1 Ras

Other human oncogene examples include Ras. The Ras protein which acts as a ON/OFF switch in cell signalling e.g. in the MAPKinase pathway and activates proteins downstream, e.g., growth factor pathways (MAPK). Different forms of Ras, i.e.,based on h-ras, k-ras, n-ras genes are thought to be involved in around 20% of cancers, and in 95% of pancreatic cancers.

6.1.1.2 Myc

The Myc family (c-myc, l-myc, n-myc) codes for transcription factors that regulate the downstream expression of up to 15% of genes. Constitutive expression (i.e. at relatively constant levels and not relative to upstream regulation) drives continued cell proliferation and therefore cancer. When upregulated Myc can therefore drive progression of cervix, colon, lung, and stomach cancers.

6.1.1.3 Bcr-Abl

The bcr-abl gene is based on a chromosomal defect in which parts of two broken chromosomes are fused (9-22); it has also been known as Philadelphia chromosome, named after the city in which the research institutes to first identify the abnormality in 1959 is located. Bcr-abl is involved in up to 95% of chronic myeloid leukemias (CMLs) but can also be involved in other leukemias. It is the result of a chromosomal translocation fusion in which the abl tyrosine kinase fused to a region called BCR (‘break point cluster region’). As a consequence the Abl tyrosine kinase is always active and driving downstream proliferative signalling and cell division. It is considered a driver during the ‘chronic’ phase of disease; but further changes are thought to be required to trigger a blast crisis. Bcr-Abl is the target of Gleevec.

6.1.1.4 EGFR family

see ErB/HER RTK family.

There many additional genes that potentially can act as oncogenes and many of those are linked to signalling pathways linked to proliferation; however, the progression from a from a local, potentially benign tumour to the malignant and even metastasising disease requires the involvement of other signalling pathways which will be discussed in more detail later.

6.2 Discovery of tumour suppressor genes

Cellular signalling pathways typically do not only rely on having an ‘on’ switch but also consist of molecular feedback mechanisms which ensure a rapid default back to ‘neutral’ i.e. by implementing a rapid ‘off’ switch. As discussed for RTK signalling earlier (e.g. the Ras cycle or PTEN), in the same way as phosphorylation by a kinases induces a rapid structural change and activation of proteins so de-phosphoryliation by phosphatases allows rapid deactivation.

So are their also genes that counterbalance the genes driving the cell to proliferation?

Alfred Knudson, a physician and geneticist , studied the childhood cancer retinoblastoma, a cancer of the eye. The disease affects the retina and (unlike most other forms of cancer) almost exclusively occurs in children. While surgical treatment is mostly successful it means that many affected children will loose their vision in the affected eye.

The disease exists as a hereditary (familial), i.e. where other family members have a history of the dissease, and a sporadic form where no such links exist. Looking at both forms of the disease, Knudson analysed how long on average it took for the disease to occur in children (age) in either form of the disease and whether or not one eye or both eyes were eventually affected. It became clear that germ line mutation alone (i.e. in hereditary cases) are not sufficient to cause cancer e.g. affected parent could have a healthy child but potentially affected grandchildren. Furthermore, patients with the hereditary form of the disease do develop tumours in most cases eventually will go on to develop bilateral tumours. In addition, analysis of all sporadic cases (i.e. if there was no family history of retinoblastoma) showed that retinoblastoma would occur later and mostly only affect one eye.

This pattern of the disease followed a statistical distribution consistent with, in the case of the sporadic disease, two independent modifications events or mutations being required (6.1). On the other hand, for the hereditary form the earlier onset suggests that only one mutation being required to trigger the disease. Based on this data and using statistical analysis Knudson was able to show in 1971 that retinoblastoma was caused by two ‘hits’ or mutations (Knudson 2001):18

Timecourse of diagnosis of retinoblastoma tumours in a population of children with hereditary/familial and sporadic forms of the disease.

Figure 6.1: Timecourse of diagnosis of retinoblastoma tumours in a population of children with hereditary/familial and sporadic forms of the disease.

6.2.1 Two hit hypothesis summary

  • Each DNA double strand carries two independent copies or alleles of a gene. If both copies are identical this is known as homozygous, when they differ it is called heterozygous. In the hereditary cases that develop the disease over time the patient will already have one mutated allele from the parents through the germ line. This does not trigger the disease but if a second, spontatnous mutation occurs the retinoblastoma cancer is triggered; this occurs earlier and frequently bilateral.
  • Sporadic cases of retinoblastoma can occur in individuals with two functional alleles of the gene. A first mutation will not have any effect by itself but if another mutation occurs as a 2nd hit the cancer will develop.

In conclusions

  • Two alleles are involved, i.e. typically each gene represented with one allele from each parent.
  • In these cases the gene is only fully disfunctional when both alleles have mutations
  • This means that two working copies are required for the gene to be functional consistent (recessive allele)

Subsequently (Friend ‘86), the mutations were identified as deletions in gene region mapping to the Rb gene which as Rb protein has important functions in the regulation of the cell cycle.

6.2.2 Tumour suppressor gene functions

These studies demonstrated that there were indeed genes which normally have an anti-proliferative function and which if they are mutated so that they loose function could drive cells towards cancer (‘loss of breaks’). However, it appears these genes require two genetic events for both alleles to be ‘switched off.’ As long as one of the alleles is functional the activity of these proteins is sufficient to fulfil these tumour suppressor functions; only when both genes are mutated and none of the protein has the correct confirmation do the mutations have a functional effect (i.e. mutation followed by loss of heterozygosity leads to loss of function).

Thus, tumour suppressor genes code for proteins that prevent or delay tumour development and cancer progression. Their functions include, for example, repression of cell cycle driving genes, negative feedback in signalling pathways, DNA damage signalling or repair. Important examples of tumour suppressor genes include RB (retinoblastoma), BRCA1 (breast cancer 1, early onset), APC (colorectal).

There are exceptions where the tumour suppressor gene does not follow the ‘recessive’ rule, e.g., TP53 (p53), also known as guardian of the genome, has a ‘dominant negative’ effect , i.e. a mutant form in only one allele is sufficient to functionally disable this tumour suppressor. This is similar to the typical oncogene which also tends to have a dominant mutation.

6.2.3 Oncogenes and Tumour Suppressor Genes - Examples in the RTK pathways

The principle of many cellular regulatory process is based on providing a balance of activating (pro) and de-activating (anti) protein activities. The appropriate level of proliferation is determined by balancing activation and de-activation which act like the accelerator or brakes on a car; just as pressing the accelerator can make a car go too fast so can the absence of sufficient breaking.

Various genes in the cell change gene expression in ways that can promote cancer, e.g. by driving proliferation, or counteract it e.g. by slowing it down. Gene expression is carefully controlled to balance both effects in healthy cells. Mutation of such genes removes their normal control or function by creating imbalance. Oncogenes are genes that control cell proliferation and can mutate so that they are always ‘on’ - promoting cancer developent. Tumour suppressor genes are genes that reduce the likelihood of cacner development e.g. by slowing down cell proliferation; when mutated a loss of control or function would then fail to slow down proliferation or counteract the activity of a activated pro-proliferative protein.

The RTK signaling pathway discussed earlier provides examples of how normal regulation works and which proteins or genes are vulnerable to mutations that could facilitate cancer progression. For example, GAP would be seen as a tumour suppressor because a mutation would disable its ability to slow down proliferation whereas RAS would be seen as an oncogene when a mutations causes it to be ‘constitutively’ (always) active.

6.2.3.1 Normal tissue

In normal tissues the production and release of growth factors is tightly controlled so that entry of cells into and progression through cell cycle only occurs when required. The levels of growth factors in a tissue are not only regulated in terms of concentration but also with respect to timing and location, i.e. when and where they are released. In addition, factors in the tissue environment may also add to a differential modulation of the growth factor distribution. For example, the growth factor molecules can interact with components in the extracellular matrix (ECM), e.g. by binding to them, so that some growth factors are temporarily sequestered in the pericellular space unable to interact with the receptors. At the same time, the activity of extracellular enzymes can trigger the release of sequestered growth factors via protease, sulfatase etc. enzyme activity. In this way enabling signals are transmitted in a temporally and spatially regulated fashion (on and off switches). In cancer this carefully regulated balance can be disturbed in a number of different ways.

Dysregulation of RTK pathways occur can occur at different levels/nodes of these pathways but with the same effect of driving proliferation.

Figure 6.2: Dysregulation of RTK pathways occur can occur at different levels/nodes of these pathways but with the same effect of driving proliferation.

Signalling pathways can be disrupted at various levels/nodes along the path, e.g. as illustrated for the HER RTK pathway (6.2). In cancer, proliferative signalling is autonomous, unregulated, and out of context. This has been recognised as one of the defining characteristics of cancer leading to sustained proliferative signalling (discussed in more detail under cancer hallmarks). This disturbance can conceptually happen in at a number of different levels which in the case of RTKs can be divided in those linked to the receptor and ligand and those which are downstream of those interactions:

6.2.3.2 Cancer – a) Receptor-ligand dependent interactions

At this level, the deregulation could be caused by various modifications around the ligand-receptor interactions, e.g. if increased level of growth factors reach the cancer cells. Typically, this could happen as a result of an autocrine loop where growth factor secretion and receptor response both happen in the cancer cell population. Alternatively, a paracrine mode of release can sometimes be observed, where neighbouring cells release (bystander stimulation) growth factors, e.g. via the interplay between tumour infiltrating immune cells or fibroblasts and cancer cells (e.g. {Sun et al. (2012)}). Similar enhanced proliferation can be caused without higher growth factor secretion levels if more receptor protein is present on the cell surface, this is thought to be contributing to some forms of breast cancer for example. Furthermore, structural alterations in the receptor itself could facilitate ligand-independent firing of the receptor so the cancer cells become hypersensitive even to limiting amounts of ligand.

6.2.3.3 Cancer – b) Ligand-receptor independent interactions

In addition to changes to the receptor and ligand, the quantity or structure other downstream elements of these proliferation pathways can lead to a constitutive activation and sustained proliferative signalling. For example somatic mutations occur which can lead to constitutive activation of specific proteins in the pathway. Examples for this sort of activation include mutations of B-Raf which play a role in ~40% of melanomas. Similarly, there are mutations to the PI3K catalytic subunit which then interfere with Akt/PKB signalling. In addition to mutations which lead to constant activation of signalling molecules it is also possible that sustained proliferative signalling could happen due to the inability to ‘switch off’ the signal in a timely fashion. Typically, negative feed back loops provide a physiological way to attenuate the proliferative signalling. Examples of mutations that limit the ability to ‘switch off’ include oncogenic mutation affecting ras genes, i.e. whereby the ability of Ras to be inactivated (Ras-GTP ⟶ Ras-GDP) is compromised. Another example affecting the PI3K pathway is the PTEN phosphatase which normally degrades PI3-kinase product PIP3 to PIP2. A loss-of-function would lead to enhanced PI3K activity and has been shown to be tumorigenic in experimental models but also plays a role e.g. in breast cancer {Pereira et al. (2016)}. In human tumours PTEN promoter methylations (a form of epigenetic mutation) have been observed which would interfere with PTEN expression. Such genes linked to the expression of proteins that have the ability to reduce proliferation are typically known as tumour suppressor genes.

Unravelling mutations that potentially drive the various signalling pathways towards cancer progression has immediate therapeutic relevance as each of these molecules provides a potential target for molecular targeted cancer therapies. By targeting specific molecules known to be responsible for cancer progression such therapies offer potentially much more specific ways of treating cancer with the prospect of reduction in the side effects frequently associated with conventional chemotherapeutic agents.

Molecular targeted drugs related to the HER/EGFR and VEGF pathways

Figure 6.3: Molecular targeted drugs related to the HER/EGFR and VEGF pathways

6.2.4 HER family therapeutic target genes

The various targets identified and molecular targeted drugs in the clinic or under development are illustrated for the HER pathways (6.3).

The HER family of proteins and the related proliferative signalling pathways are important and well understood drivers of cell growth. In particular HER1 (EGFR-1) and HER2 are relevant to a number of different cancer types. These pathways support the sustained proliferative signalling central to the development of tumours and pharmacological interferrence with these pathways thus provides a rationale route to selective molecular targeted cancer therapies. Molecular targeted therapies that target different levels of the pathway have been developed and the value of such compounds has been demonstrated clinically.

6.3 Cell signalling pathways vs networks

Cell signalling pathways are often discussed in a linear fashion, i.e. one molecule activates the next and so on until a particular target is hit/activated. However, it is quite important that despite the apparent complexity these are in fact typically very simplified models of signalling. The reality is that signalling pathways consist of nodes with multiple branching points and interconnections. Furthermore, the response to signalling often is more complex than simply a “on/off” switch but will require a gradual response to a stimulus. Furthermore, such responses are not necessarily linear and in fact will often depend on the state of other surrouding or even complexed molecules. The transcription of every factor is regulated and interactions happen in a non-binary fashion and will react to other changes in various feedback loops.
Cell signalling pathways are complex and in fact more appropriately thought of as networks. Modified from

Figure 6.4: Cell signalling pathways are complex and in fact more appropriately thought of as networks. Modified from

It is therefore essential to be mindful of the fact that cell signalling pathways are in fact complex networks with important proteins acting as nodes (6.4). One of the important implications from this is that often a molecule will have multiple effects or roles in terms of signalling. Furthermore, in relation to therapeutic application it will become clear that often when a molecular targeted therapy interferes with a particular part of a signalling pathway alternative pathways may be available in the widere network that may circumvent any therapeutic drug blocking a particular sections.