Transcript Carcinoma Myelocytomatosis

Mco
A
c-Myc Regulated Functional Gene and Protein
Networks Involved in Tumourigenesis
Sam Robson
1: C-Myc
Deregulation of the c-Myc (Carcinoma
Myelocytomatosis) proto-oncogene is seen in a
large number of human cancers, resulting in
aggressive, poorly differentiated tumours. The
c-Myc protein is a transcription factor, and is
known to be involved largely with cell cycle
progression (G1 to S phase) and the inhibition
of terminal differentiation. Paradoxically, the
transcription factor also appears to sensitize the
cell to apoptotic activity.
c-Myc works as part of a heterodimeric complex
with the protein Max, which binds itself to the
carboxy-terminal basic-helix-loop-helix-zipper
(bHLHZ) domain of the c-Myc protein. This MycMax heterodimer is able to bind specific DNA
sequences, such as the E-box sequence
‘CACGTG’. We also find two highly conserved
elements in the N-terminal domain; the ‘Myc
boxes’ MBI and MBII. These are required for
transactivation of various target genes.
2: Pancreatic Cancer
and Diabetes
The pancreas is responsible for regulating
blood glucose levels. It contains hundreds
of colonies of cells within the exocrine
tissue, known as islets. One of the cell types
contained within the islets are β-cells, the
main source of insulin within the body.
Excessive proliferation of β-cells results in
insulinoma, while excessive apoptotic
activity has been linked to some forms of
diabetes (due to the subsequent loss of
insulin, leading to a drop in glucose
metabolism).
3: Transgenic Model
By the time cancers are detected, they are usually very advanced
and have acquired many additional genetic lesions. This makes it
difficult to identify the initial pathways involved in
tumourigenesis, making human in vivo study difficult.
Mouse strains have been genetically modified to allow direct
control of the activation of the c-Myc transcription factor within
the pancreatic islets. This allows the analysis of the genetic
changes that occur as a direct result of c-Myc activation.
To do this, c-Myc is fused at the carboxy-terminal domain to the
hormone binding domain of a mutant mouse oestrogen receptor
to form a c-MYC-ERTAM transgene. This transgene encodes a
chimeric protein c-MYC-ERTAM that can be activated by the
specific ligand 4-hydroxytamoxifen (4-OHT), which is
administered through daily intraperitoneal injections.
Activation of the c-MYC-ERTAM protein in the islets results predominately in apoptosis rather than proliferation,
leading to islet involution and onset of diabetes within 9 days. This indicates that the β-cells are only mildly
buffered against cell death in vivo. In order to study the oncogenic potential of c-Myc in islet carcinogenesis, Mycinduced apoptosis was blocked by constitutive over-expression of the anti-apoptotic protein Bcl-XL (Beta-cell
lymphoma X - large). This was achieved by cross-breeding the c-Myc-ERTAM mice with transgenic mice in which
expression of Bcl-XL is placed under the control of the rat insulin promoter (RIP7). This produces double
transgenic Ins-c-Myc-ERTAM/RIP-Bcl-XL, otherwise known as RM mice.
Activation of c-MYC-ERTAM in double transgenic RM mice results in rapid, synchronous entry of nearly all β-cells
into the cell-cycle with no discernable c-Myc induced apoptosis. The islets continue to increase in size so long as
4-OHT is continuously administered. Thus c-MYC-ERTAM activation results in grossly hyperplastic islets .
Carcinogenesis involves accumulation of several somatic lesions, leading to unmediated growth of cells, loss of
differentiation, invasion and angiogenesis. By controlling the expression of c-Myc and preventing cell death, we
can study the role that it plays in these processes.
4: Microarrays
Microarrays are high density
oligonucleotide chips that are able to
measure the expression levels of every
gene within the genome by measuring
RNA levels. Affymetrix Genechips are
commercially available microarrays with
very high reproducability.
RNA is extracted from pancreatic tissue at various timepoints from
the start of 4-OHT administration. The RNA is labelled with biotin
and washed onto the chip (MOE430 Plus2 Mouse chip). RNA strands
bind to the oligonucleotide strands on the chip. Oligonucleotides
matching to specific genes are organised together to form probes. A
fluorescent image of the chip is taken, where the intensity of each of
the probes relates to the concentration of RNA in the cell.
6: Conclusions and Future Research
Whilst many conclusions can be made from this data, it is uncertain
how relevant they will be. Much of the data is variable, suggesting
problems with the collected RNA. Also, many genes show
unexpected expression profiles. Whilst this may show novel gene
targets of c-Myc, it is more likely that this is due to problems with
the RNA (be it RNA degradation, masking of important data by
exocrine tissue or the non-comparability between the time points).
The group is currently working to solve this problem by using Laser
Capture Microscopy to isolate islet cells for microarray analysis.
Current problems exist due to RNA degredation, which we hope to
solve in the near future.
Once this problem has been overcome, a similar microarray
experiment will be run focusing on early timepoints. This will allow
the analysis of direct effects of c-Myc activation in the islet tissue,
allowing the distinction of direct c-Myc targets from indirect targets.
However it has been suggested that as many as 10% of the genes in
the genome are under the direct transcriptional control of c-Myc
(close to 4,500 genes), which may make this process difficult.
Once information on c-Myc targets, along with downstream effects,
has been discerned, gene network pathways can be set up showing
the effects of c-Myc activation on tumourigenesis. Various
mathematical techniques are currently being tested for the
validation of these pathways, most notably the Bayesian approaches
being optimised by the IPCR group. Once these networks have been
fully understood, it is hoped that they may provide possibilities for
therepeutic research in the future.
5: Microarray Data Analysis
Microarray experiments create a huge amount of data. At least three replicate chips must be
run for each time point to allow statistical analysis. Four time points were used in this
experiment; Time 0 (with no 4-OHT administered), 1 day and 14 days of 4-OHT administration,
and a tumour reversal time point, with 4-OHT administered for 14 days, then stopped for 7
days.
The Microarray data was analysed using Genespring, a very powerful piece of software from
Agilent. The first step of data analysis is to perform quality control to remove outlying
samples. Several methods can be used, such as hierarchical clustering algorithms and
principle component analysis. The data must also be normalised to allow comparisons
between genes on different chips. Normalization shifts the data to have mean 1.
The next step of analysis is to reduce the dimensionality of the data by finding ‘important’
genes. These are genes that are differentially expressed across the time course, and may
provide an insight into the pathways involved at various stages of tumourigenesis.
One of the key methods of analysis is to map the
expression profiles onto pathways available
from the Kyoto Encyclopaedia of Genes and
Genomes (KEGG). Shown here are the
expression levels of genes involved in the cell
cycle after 1 day of 4-OHT administration (red =
upregulation; blue = downregulation; yellow = no
change).
Pathways are available for many of the most
important processes of the body. By analysing
changes in gene expression levels at various
timepoints, we get an idea of the effect that cMyc has on these pathways.
One key feature being focused on is the DNA
damage pathway, and how this relates to the
apoptotic activity of c-Myc. The exact
mechanisms of how c-Myc leads to apoptosis is
currently unknown, and it is hoped that this will
shed some light on the problem.
We see much variation in the data, which suggests that the RNA quality is not as good as it
could be. The pancreas contains many RNAses (enzymes that degrade RNA), that begin to act
instantly upon making incisions into the tissue. Several methods have been used to prevent
this degradation, including snap freezing the tissue in liquid nitrogen instantly after extraction,
fixing the tissue using formalyn, and using the RNAlater stabilising reagent from the Qiagen
company. Currently, work is under way to improve these methods, such as perfusing formalyn
into the tissue through the blood system before extraction.
A further problem with the data is a result of using the entire pancreatic tissue when all that we
are interested in is the islet tissue. The islets make up only a small part of the total tissue mass
and it is probable that exocrine pancreatic tissue will mask the RNA levels of the islets.
Another problem comes from the fact that the islet to exocrine ratio is not the same at the
various time points. Thus comparisons between expression levels across the time points may
not be useful. Further studies should be performed using only RNA from the islet cells.
Acknowledgements
Special thanks to Mike Khan and Stella Pelengaris for giving me this opportunity to work with their
group, and also to David Epstein for keeping me routed in the maths department. Thanks also to
Helen Bird, Lesley and Sue Davis for their help with the Microarray work. Lastly, thanks to Linda
Cheung, Vicky Ifandi, Sevi, Göran and Sylvie Abouna for all of their help in the lab work.