By Abby Van Voorhees, MD, June 01, 2011
In this month’s Acta Eruditorum column, Physician Editor Abby S. Van Voorhees, MD, talks with Craig Ceol, PhD, about his recent Nature article, “The histone methyltransferase SETDB1 is recurrently amplified in melanoma and accelerates its onset.”
Dr. Van Voorhees: What is the most common mutation found in melanomas? Is this mutation also found in benign nevi? How is it thought to work?
Dr. Ceol: The most common mutation in melanomas is a mutation in the BRAF gene, which is a part of the RAS signaling pathway. The RAS pathway is abnormally overactivated in many different types of cancers. BRAF is mutated in about 50-60 percent of all melanomas, and most BRAF mutations are the same they cause a single amino acid change in the BRAF protein which creates the BRAF(V600E) oncogene. The BRAF(V600E) mutation is found in roughly half of all melanomas and also in benign nevi, so it’s thought that BRAF(V600E) is probably necessary but not sufficient for melanoma formation and that BRAF(V600E) needs to cooperate with other types of genomic changes in order to make a melanoma.
These changes can take the form of single DNA nucleotide mutations in the genome or they can be amplifications or deletions of broader regions of the chromosomes. If you look at enough melanomas what you find is the same chromosomal interval can be amplified in many different tumor samples. Some of these regions have known melanoma oncogenes. For example, BRAF itself is in a region that is recurrently amplified. However, there are other recurrently amplified regions where we suspect there are oncogenes but we don’t know which gene is the one that cooperates with BRAF. Finding these new oncogenes is one of the reasons we initiated our study.
Dr. Van Voorhees: Tell us about the model that you used to study these sections of the genome. What made you consider looking further at chromosome 1?
Dr. Ceol: We have been working with a zebrafish model of melanoma; zebrafish are pretty useful for studies of melanocyte biology. They have melanocytes — that’s where they get their name, their black stripes are made up of melanocytes. We previously had put the human version of BRAF(V600E) in zebrafish and found that it caused the formation of nevi. Together with another mutation in the p53 gene, BRAF(V600E) gave rise to melanomas in fish.
So we wanted to find out, if we throw other things into the mix, can we accelerate or delay tumor onset or change any of the properties of the tumors we see in our model? We considered a number of different regions that are recurrently amplified in the human melanoma genome, in particular ones where we didn’t know the important oncogene in the interval. We focused on an interval of chromosome 1 because it was focally amplified and had no known oncogene. We screened through genes using our zebrafish model to look for an oncogene that substantially accelerated the formation of melanomas.[pagebreak]
Dr. Van Voorhees: Was a specific oncogene identified that regulates tumorigenesis in melanomas? How is it thought to function?
Dr. Ceol: We found a gene, SETDB1, that markedly accelerated melanoma onset. SETDB1 encodes an enzyme that adds methyl groups onto histone proteins and, through this activity, SETDB1 can downregulate the expression levels of hundreds of target genes. In zebrafish melanomas with excess SETDB1 we indeed found that many genes were downregulated. In looking at the most downregulated genes we found that the same set was also downregulated in human melanomas with excess SETDB1. What this says is that we are barking up the right tree, that is, there’s a strong correlation between what we’re seeing in our zebrafish model and what happens in human tumors.
Dr. Van Voorhees: Does SETDB1 work alone or in combination with other mutations?
Dr. Ceol: SETDB1 by itself does not cause melanoma formation; rather, it cooperates with BRAF to accelerate melanomas. You absolutely need BRAF there; what SETDB1 does is speed the process up. What we saw in human tissue samples is consistent with this finding. Levels of SETDB1 were quite high in melanomas but relatively low in nevi and normal melanocytes. We think this means that when a BRAF(V600E) mutation is present along with excess SETDB1, a lesion more easily becomes a tumor. However, with a low level of SETDB1, a lesion is more likely to persist in a benign state.
Dr. Van Voorhees: Is this model thought to be the case in other malignancies?
Dr. Ceol: At this point we’re not sure. However, we do know that the same region of chromosome 1 is recurrently amplified in many types of cancers, including non small cell lung, hepatocellular, ovarian, and others. So SETDB1 may have a broader role in these other cancers as well, but we have not looked at this directly.
Dr. Van Voorhees: Does there seem to be a dose correlation in those malignancies as well?
Dr. Ceol: We don’t know that yet. What we’d like to do is the same thing we did with melanomas — get tissue samples and monitor the level of SETDB1 in normal, benign, and malignant stages. But we don’t have hard evidence that SETDB1 is causing the formation of these other tumors; we just know that this small region of the genome is recurrently amplified, which is a nice correlation but doesn’t show causation.[pagebreak]
Dr. Van Voorhees: SETDB1 seems to increase the frequency of the tumors; does it also increase their aggressiveness?
Dr. Ceol: In our fish model, when SETDB1 was present at excess levels the tumors were much more aggressive. We can monitor invasion rather easily. Normally melanomas in zebrafish tend to grow in exophytic fashionoutward off the skin. However, when SETDB1 was there, the melanomas more easily invaded into the underlying musculature and often went into the spinal column of the fish, and they did so in a very short period of time.
Dr. Van Voorhees: Does this study help us better understand what allows for the promotion of cancer? Can it be used to predict the potential aggressiveness of a melanoma?
Dr. Ceol: SETDB1 appears to have a role in modulating the process of oncogene-induced senescence. In established tumors the BRAF(V600E) oncogene can promote cell division and proliferation. But when it’s initially introduced in a nave cell, it doesn’t do that; it leads to a senescent arrest. Overcoming that arrest is a key step in tumor formation. When SETDB1 was overexpressed and amplified, we found that BRAF-induced senescence was bypassed. Instead of arresting, cells with a BRAF(V600E) oncogene and excess SETDB1 began proliferating almost immediately. Through our studies we can investigate these early stages of tumor development that are very difficult to capture by looking solely at human tissue biopsies.
Whether SETDB1 is a predictor of aggressiveness is an interesting question. Currently we are assembling tissue samples, each of which is tied to a thorough clinical history, so this question can be addressed. In addition to being a prognostic marker, it is also possible that SETDB1 could be a useful therapeutic target. In well-publicized recent trials, BRAF inhibitors have shown remarkable efficacy in patients with BRAF-mutant melanomas. However, relapses are frequent and typically occur within months of initial treatment. It will be interesting to see if SETDB1 inhibitors, in combination with BRAF inhibitors, can prevent or delay relapses.
Dr. Ceol is assistant professor in the programs in molecular medicine and cell dynamics at the University of Massachusetts School of Medicine. His article was published in Nature, 471, 513517 (24 March 2011). doi:10.1038/nature09806.