RNA Tumour Viruses

How are cellular oncogenes acquired?

A plausible model which can be reproduced in culture is that integration occurs close to and upstream of an oncogene. The provirus has deleted 3' LTR. Transcription will therefore go through the oncogene, introns removed and because it has a 5' packaging signal the RNA can be packaged. A second "helper virus" infects the cell to enable this RNA to be packaged into a virion particle. As retroviruses are diploid a chimaeric virus may result. During next round of infection non-homologous recombination may take place during reverse transcription giving rise to a provirus with two LTRs. The resulting retrovirus is defective because 3' coding information is lost.. But it can transform a cell. Infectious transforming virus is only released if functional helpers are present. Examples include:

Simian sarcoma virus and v-sis:
Identified in a fibrosarcoma of a woolly monkey. Characterised by DNA sequencing and antibody immune precipitations as a 38K protein product. Not phosphorylated and no kinase activity. Env-sis fusion product, involving 5' end of env. Data base comparisons showed a striking homology to human platelet derived growth factor (PDGF) beta-chain, about 88%. PDGF has a & beta-chains which forms both homo and heterodimers. SSV transformed cells produce large amounts of the homodimer which is functionally indistinguishable from the real thing. About 10% is secreted and the rest is firmly associated with cellular membrane fractions, probably due to the fusion with env. This suggests an autocrine function in which continual stimulation of the PDGF receptor results in eventual transformation. One discrepancy is that SSV transformed cells can grow in soft agar while PDGF stimulated normal cells cannot. This could be because:

1. The fusion protein has slightly different properties to the normal protein
2. The intracellular interaction of the homodimer with internal receptors might evoke a different response from that elicited by exogenous PDGF.

Cells transformed by many cytoplasmic oncogenes also produce to excess various growth factors. This suggests that the external stimulus provided by these factors is important in producing and maintaining the transformed phenotype.

Avian erythroblastosis virus and v-erbB gene (EGF receptor gene):
Induces erythroblastosis and sarcoma in chicken, transforms fibroblasts and erythroid cells in vitro. The v-erbB gene codes for an epidermal growth factor receptor (EGFR). The helper virus is a non-defective ALV

EGFR has 4 domains:

1. N-terminal extracellular EGF binding domain
2. Transmembrane segment
3. Tyrosine kinase catalytic domain
4. C-terminal kinase regulatory domain

The kinase activity is likely to be the primary effector function. EGF binding stimulates this activity resulting in autophosphorylation of EGFR and subsequent phosphorylation of cellular polypeptides. The v-erbB gene is truncated at both 5' and 3' ends. The 5' truncation alone will cause erythroblastosis; the 3' deletion is required for fibroblast transformation. The biochemical details of its activation are unknown but are presumably related to the loss of the C-terminal regulatory domain.

Known substrates for phosphorylation include GAP a GTPase activating protein; Phosphatyl-3-kinase and phospholipase C (thereby influencing inositoyl phosphate mediated signally events); and Grb2 a protein I shall discuss in a few moments. Proteins containing SH₂ domains are also substrates, see below. These can serve as molecular adaptors in signalling.

Rous Sarcoma virus and the src gene:
First isolated in 1911 by Rous, capable of transmitting a sarcoma between Plymouth Rock chickens. It usually requires a replication competent virus such as avian leukosis virus for transmission. It has an env deletion. Unusually for this class of viruses, replication competent viruses can be isolated, but they are unstable. Src was eventually identified as gene responsible for transformation by using a combination of approaches e.g. genetic analysis of t.s. mutants and non-conditional (deletion) mutants. Position of the gene in the virus was then determined by recombinational analysis and oligonucleotide mapping. The gene product was identified by immune precipitation using sera from sarcoma bearing rabbits. It is a phosphoprotein with a molecular weight of 60K. Unlike many v-oncs this is not expressed as a fusion with normal virus polypeptide. The corresponding cellular gene and protein product turns out to be a tyrosine protein kinase. The mammalian genome contains many (>30) tyrosine kinase genes in a single gene family. The catalytic domains show extensive homology. There is considerable divergence in the other receptor binding domain. The kinase activity is turned on when it associates with an activated receptor.

c-src is located on the inner surface of the plasma membrane, anchored to the membrane by a myristilated N-terminal glycine residue. It concentrates at focal points between cells, where it is associated with the cytoskeleton. src contains a number of domains N-SH₃-SH₂-Kinase domain-Regulatory domain-C
(SH stands for src homology region). These are found in a wide variety of proteins but were first recognized in src. SH₂ domains bind phosphotyrosine residues, SH₃ domains bind proline rich sequences in many different proteins enabling proteins like src to potentially bind to a very wide range of other proteins.

Activation of src weakens cell contacts and induces membrane ruffling which is associated with increased cell motility.Fibronectin a protein associated with the cytoskeleton as well as several other associated proteins are phosphorylated by src. This is due to the src mediated activation of a protein called focal adhesion kinase or FAK. src activation also increases secretion of plasminogen activator, this elevates plasmin levels which degrades proteins in the extracellular matrix again increasing cell motility:

What are the differences between c and v-src ?

• v-src has a series of point mutations but these are not conserved in various strain.
• v-src is also close to viral LTR and contains no introns but over expression of c-src gene by a heterologous promoter or homologous substitution in a RSV does not result in complete transformation.

By elimination it seems that a 900 base deletion from the 3' end of v-src is the cause. The result of this deletion is that the 19 C-terminal amino acids v-src are lost and replaced by 12 amino acids derived from the 3'UTR of c-src. This leads to a simple (but not complete) biochemical explanation. Src is normally inactive. The kinase activity of src is down regulated by phosphorylation of a tyrosine 527. This is thought to bind intramolecularly to the SH₂ domain resulting in an inactive conformation.c-src gains activity if it is dephosphorylated by in vitro phosphatase treatment. v-src lacks this tyrosine residue and is permanently activated. There is good evidence for this model from other experiments, e.g:

• Other transducing retroviruses which carry src have C-terminal truncations which result in the loss of tyrosine 527
• The activation of other tyrosine-kinase oncogenes such as v-yes and v-fgr also involve C-terminal truncations and the loss of the phosphorylatable tyrosine residue
• Site directed mutagenesis of the src tyrosine 527 gives it expression level dependent transforming ability in the 3T3 assay.

Moloney sarcoma virus and mos:
Isolated from a rhabdomyosarcoma in mice. The virus will transform cultured fibroblasts. Expressed as a env-mos fusion protein in the soluble portion of cytoplasm. It is a serine/threonine protein kinases. c-mos is normally expressed at very low levels in a cell (<1 mos MRNA per cell). c-mos gene has the same transforming power as the v-mos gene which suggests that transformation may be a consequence of overexpression by the viral LTR.

MC29 and myc:
Avian myelocytomatosis virus, of which MC29 is the prototype, caries the v-myc gene as a gag-myc fusion. c-myc is a DNA binding phosphoprotein found in the nuclear matrix. It has an N-terminal transactivating domain. A DNA binding domain which recognises a consensus DNA sequence CACGTG. And a C-terminal domain which contains two motifs often involved in protein- protein interactions. A leucine zipper and helix-lop helix. It functions as a heterodimer with another protein known as max. Max also interacts with a protein known as Mad. In a sense Myc and Mad are in competition for Max. The Myc-Max complex promotes the switch into G1. One of the pivotal genes activated by this complex is cdc25A. This protein activates a G1 cyclin dependent kinases by removing a phosphate group. This helps the push into G1. Mad-Max by contrast recruits an enzyme Histone deacetylase. This results in condensation of chromatin, transcriptional silencing and the cell remaining in GO. Myc expression is tightly controlled.A variety of cells express it during the proliferative phase. Both the mRNA and protein have a very short half life consistent with a regulatory role in the cell cycle. The gene is turned off normally; unregulated expression is important in transformation as evidenced by transformation assays and transgenic animal experiments. Max is expressed at a higher and more constant level.

Again, the properties of v-onc may be explained just by:

1. High level expression driven by the LTR
2. By a variety of point mutations found in v-myc.
3. Other nuclear oncogenes also identified including myb, jun, fos and ski.

Presumably, all are involved in switching on transcription of genes involved with cell growth. One signal transduction pathway which is particularly well understood is exemplified by:

Kirsten and Harvey murine sarcoma virus and ras:
Isolated from sarcomas that developed in rats. This virus has transduced a ras oncogene. Generally the transduced ras gene is not fused to a viral gene and contains no deletions. Activation is explained in terms of:

• High levels of expression driven by the viral LTR
• Point mutations which occur at 2 specific positions in p21 ras i.e. 12 and 59 which reduce GTPase activity

The ras family of proteins are located on cytoplasmic face of the cell membrane. They have a GTPase activity and a homology with G proteins. These proteins serve to couple extracellular signals via a receptor to intracellular enzymes. Antibodies raised to the GTP binding domain reverse the transformed phenotype when injected into cells confirming the importance of the GTPase activity. Transduction Pathway:

A growth factor binds to a receptor-resulting in receptor dimerisation and trans-phosphorylation.
A protein (GRB2-SOS) dimer complex binds to the receptor (growth factor receptor binding protein and son of sevenless).
Ras associates with SOS, a guanine exchange factor, and exchanges GDP for GTP. Ras is like a switch, it is on when bound to GTP, off when bound to GDP.
Raf-1 a protein kinase associates with SOS and is activated.
This complex phosphorylates MEK which in turn phosphorylates MAPK ( mitogen activated protein kinase.
MEK can phosphorylate several oncogenes including myc, jun and ets, which are transcriptional factors. Get co-ordinated changes in gene expression, positive factors stimulated, -ve factors inhibited.
This highly conserved pathway which is present in man Drosophila and yeast has many steps. There are alternative activators and MAPK may be phosphorylated by kinases other than MEK.

Of the factors described in this pathway so far only ras mutations appear in human cancers.

2. Replication competent viruses and insertional mutagenesis

Exemplified by avian leukosis viruses. The initial event is thought to be integration near a cellular gene - generally myc. Viruses of this type are replication competent and the animals are chronically viraemic. Tumours appear after a long multistaged latent period (in contrast to transducing retroviruses) and are clonal in origin. So although many cells are infected, only one progresses to a tumour.

When immunologically incompetent newly hatched chicks are infected with ALV a large number of cells in the bursa of fabricus are infected. Within one month of this a preneoplastic transformation occurs with the formation of about 100 follicles containing immature blast cells. A few of these form nodules 5-10mm in size after several months. During natural bursal regression disseminated metastases derive from these nodules.

It was discovered that in most bursal lymphomas induced by ALV the cellular myc gene was activated, apparently because proviral integration occurs near the myc locus. Myc RNA is elevated 50x compared to normal cells. The myc transcripts varied in length but generally contained U5 sequences derived from the 3' LTR This suggests that the elevated myc transcription is due to the activity of the 3' LTR promoter elements. Simplest interpretation is therefore that there is abnormal activation of the c-myc gene. This is an example of promoter insertion. There are also cases where insertion is 3' of the myc gene or where insertion is in the wrong orientation in which case the activation effect is postulated to be due to enhancer insertion. This conclusion seems warranted because in all known cases of this the enhancer sequences are maintained.

These experiments helped to establish the insertional model of oncogenesis.

However the story is more complicated than you might think and provides more evidence for the multistage development of these tumours. Other experiments using the NIH3T3 system and DNA from bursal tumours showed that the "transforming DNA" was not myc (failure to detect myc activity in these cells is hardly surprising as they are already immortalized) but Blym-1. This codes for a 8,000MW protein homologous to the N-terminus of transferrin. The clear inference is that myc activation precedes Blym-1-1 activation. The function of Blym-1 is unknown.

Mouse mammary tumour virus:
MMTV also has a long induction period for tumourigenesis. The transformation event is rare despite 50% of mammary epithelial cells being infected. Three major unlinked and unrelated loci are transcriptionally activated in many MMTV tumours int1, int2 and int3. Int-1 is involved in normal mouse neural development. In normal cells int-1 transcripts are not detectable. In transformed cells there are about 10 copies of mRNA per cell. Int-1 can transform a mammary epithelial cell line. Transgenic mice expressing int-1 under the control of the MMTV LTR show a high frequency of adenocarcinomas developing from hyperplastic mammary glands. An enhancer element of the LTR is thought to cause overexpression of int-1 since it is always located downstream or in the opposite transcriptional sense upstream. Int2 shows homology to fibroblastic growth factor, potential autocrine activator. Int3 is another protein involved in cell development.

A link with human breast cancer?
Retroviral like particles can often be observed in cells isolated from human breast cancers. An epidemiological correlation between the presence of mice in human dwellings and breast cancer incidence has also been claimed. Recently it has been found that 40% of breast cancer tissues contain DNA sequences almost identical to MMTV virus sequences. A second study has found RAKalpha nucleic acid sequences in 40% of human breast tumours. The antigens are expressed and have 90% identity to HIV-1. Both of these studies need to be confirmed by other investigators but raise the possibility that retroviruses are implicated in at least some cases of breast cancer.

3. Replication competent viruses with transacting functions, e.g. HTLV-1

Transactivating non-defective oncoviruses:

• Human T Cell Leukaemia virus types I and II.
• Bovine leukaemia virus or BLV is a closely related virus.

The genome organization is specific to this group, i.e. 5'-LTR-gag-pol-env-rex-tax-LTR-3'

HTLV-1 is associated with specific lymphoid malignancies and is endemic in Japan, the Caribbean and Africa. Discovered after intense efforts to isolate retroviruses by R. Gallo. The virus was isolated from an adult T cell malignancy. The malignancy is always a monoclonal CD4+ T cell. Molecular probes show that HTLV-1 is present in almost all cases of adult T cell leukaemia. There is also a strong sero-epidemiological correlation between the disease and the virus. Transmission is thought to occur by cell-cell contact and not by free virus during the intimate exchange of body fluids:

• Perinatal transmission via milk lymphocytes during breast feeding
• Lymphocytes contained in semen during sexual intercourse
• Unscreened blood transfusions
• Sharing of blood contaminated needles by intravenous drug users.

In Japan, about 1% population is infected. Clustering results in much higher rates of infections in some regions, e.g. Okinawa 35%, Kyushu province 10%. The virus is becoming increasing common in W.Europe and N. America particularly amongst intravenous drug users and homosexuals. About 1-3% of infected individuals will eventually develop this aggressive leukaemia after an incubation period which is usually several decades long. Following an asymptomatic period, the patient may develop pre-ATL, in which they remain asymptomatic but there is a clonal expansion of T-cells which can be morphologically-altered. 50% of patients with pre ATL progress to chronic/smouldering ATL. This is manifested by skin lesions (mycosis fungoides) and high leukocyte counts. They then progress to acute ATL within several months. The survival time is measured in months.

White blood cell counts are elevated, there is a dominant malignant clone of morphologically altered T cells. Symptoms include lymphadenopathy, hepatosplenomegaly, and hypercalcaemia. Patients are also immunocompromised and present with fungal infections, pneumocystis pneumonia and cytomegalovirus infection. Despite a wide spread disseminated immune response the infection is never cleared. Treatment is reserved for subacute and acute phases. However, there are no treatments of demonstrated value.

HTLV can transform T lymphocytes in culture. Transformation is defined as the ability to grow indefinitely in the absence of exogenous T cell growth factor (TCGF) or interleukin 2 (IL-2). HTLV-1 is not thought to have a preferred integration site which might lead to cellular gene activation. No single activated oncogene can be detected in a transformation assay. It is currently assumed that the virus augments T cell proliferation and thus provides an expanded T cell population in which a malignant event can occur.

There is some molecular data which may be relevant to transformation. Transactivation of transcription by an HTLV gene product called tax has been observed. Tax is essential for viral replication. Viruses which do not contain a functional tax gene are not infectious. They make only trace amounts of RNA suggesting a role in transcriptional activation. Direct evidence of this has been obtained using a transient transfection assay with a LTR-enzyme reporter gene. Tax can transactivate other host promoters, e.g. an NFkB-like transcription factor characteristic of activated T cells which binds to an upstream element of the TCGF gene. Mice transgenic for tax develop multiple mesenchymal proliferative disorders.

HTLV-1 positive individuals also have a 20% lifetime risk of developing tropical spastic paraparesis. This is characterized by uncoordinated motor control. The pyramidal tracts of the spinal chord which carry motor neurones house the virus and show degenerative changes. The cause of this syndrome is unclear.

Until recently the only RNA tumour viruses belonged to the family Retroviridae. Now it seems that a member of the Flaviviridae family can also cause tumours. Hepatitis C virus was identified in 1989 as the cause on 90% of non-A non- B hepatitis infections. It is an enveloped virus and contains a single stranded RNA genome of 9.5kb. The virus is undoubtedly transmitted by blood contact. Perinatal and sexual transmission may also occur. Infection may be asymptomatic for decades but the majority of infections eventually result in hepatitis and hepatocellular carcinoma. Conditions for growth in vitro have not yet been established and little is known about the pathology and cancer caused by the virus.

JSRV, Jaagsietke sheep retrovirus
This virus is the cause of ovine pulmonary carcinoma. A contagious sheep disease prevalent in many countries including the UK and S Africa. Following infectional multifocal tumours rapidly develop (compare to transducing retroviruses). But this virus is a simple retrovirus, gag-pol-env. It has not transduced an oncogene. The cloned virus can cause OPC; furthermore the cloned env gene can cause OPC. A gene called HYAL2 codes for the GPI anchored receptor protein for this virus. HYAL2 may act as a tumour supressor protein and viral binding to it may prevent this function. Alternately interaction of the virus with HYAL2 sends a growth stimulatory signal into the cell, resulting in transformation. OPC has clinical and histological similarity with human bronchioalveolar carcinoma, responsible for 25% of human lung cancers. There is no epidemiological link between JSRV and this carcinoma and at this stage no-one is suggesting a link between the two. However the HYAL2 gene is deleted in this type of lung tumour, consistent with the protein having a tumour supressor function. It seems then these two diseases share a coincidental pathology and retroviruses can also cause tumours by interfering directly with tumour supressor proteins- a model previously reserved for DNA viruses.